From 61dc740ed1c3780cccaec00d059a28f0d31d0052 Mon Sep 17 00:00:00 2001 From: Stephane Glondu Date: Mon, 4 Jun 2012 12:07:52 +0200 Subject: Imported Upstream version 8.4~gamma0+really8.4beta2 --- doc/refman/RefMan-tac.tex | 6168 +++++++++++++++++++++++---------------------- 1 file changed, 3140 insertions(+), 3028 deletions(-) (limited to 'doc/refman/RefMan-tac.tex') diff --git a/doc/refman/RefMan-tac.tex b/doc/refman/RefMan-tac.tex index 198f8f30..b630772b 100644 --- a/doc/refman/RefMan-tac.tex +++ b/doc/refman/RefMan-tac.tex @@ -6,15 +6,15 @@ A deduction rule is a link between some (unique) formula, that we call the {\em conclusion} and (several) formulas that we call the {\em -premises}. Indeed, a deduction rule can be read in two ways. The first -one has the shape: {\it ``if I know this and this then I can deduce +premises}. A deduction rule can be read in two ways. The first +one says: {\it ``if I know this and this then I can deduce this''}. For instance, if I have a proof of $A$ and a proof of $B$ then I have a proof of $A \land B$. This is forward reasoning from premises to conclusion. The other way says: {\it ``to prove this I have to prove this and this''}. For instance, to prove $A \land B$, I have to prove $A$ and I have to prove $B$. This is backward reasoning -which proceeds from conclusion to premises. We say that the conclusion -is {\em the goal}\index{goal} to prove and premises are {\em the +from conclusion to premises. We say that the conclusion +is the {\em goal}\index{goal} to prove and premises are the {\em subgoals}\index{subgoal}. The tactics implement {\em backward reasoning}. When applied to a goal, a tactic replaces this goal with the subgoals it generates. We say that a tactic reduces a goal to its @@ -25,29 +25,23 @@ Each (sub)goal is denoted with a number. The current goal is numbered address a particular goal in the list by writing {\sl n:\tac} which means {\it ``apply tactic {\tac} to goal number {\sl n}''}. We can show the list of subgoals by typing {\tt Show} (see -Section~\ref{Show}). +Section~\ref{Show}). Since not every rule applies to a given statement, every tactic cannot be used to reduce any goal. In other words, before applying a tactic to a given goal, the system checks that some {\em preconditions} are satisfied. If it is not the case, the tactic raises an error message. -Tactics are build from atomic tactics and tactic expressions (which +Tactics are built from atomic tactics and tactic expressions (which extends the folklore notion of tactical) to combine those atomic tactics. This chapter is devoted to atomic tactics. The tactic language will be described in Chapter~\ref{TacticLanguage}. -There are, at least, three levels of atomic tactics. The simplest one -implements basic rules of the logical framework. The second level is -the one of {\em derived rules} which are built by combination of other -tactics. The third one implements heuristics or decision procedures to -build a complete proof of a goal. - \section{Invocation of tactics \label{tactic-syntax} \index{tactic@{\tac}}} -A tactic is applied as an ordinary command. If the tactic does not +A tactic is applied as an ordinary command. If the tactic is not meant to address the first subgoal, the command may be preceded by the wished subgoal number as shown below: @@ -56,11 +50,107 @@ subgoal number as shown below: & $|$ & {\tac} {\tt .} \end{tabular} -\section{Explicit proof as a term} +\subsection{Bindings list +\index{Binding list} +\label{Binding-list}} + +Tactics that take a term as argument may also support a bindings list, so +as to instantiate some parameters of the term by name or position. +The general form of a term equipped with a bindings list is {\tt +{\term} with {\bindinglist}} where {\bindinglist} may be of two +different forms: + +\begin{itemize} +\item In a bindings list of the form {\tt (\vref$_1$ := \term$_1$) + \dots\ (\vref$_n$ := \term$_n$)}, {\vref} is either an {\ident} or a + {\num}. The references are determined according to the type of + {\term}. If \vref$_i$ is an identifier, this identifier has to be + bound in the type of {\term} and the binding provides the tactic + with an instance for the parameter of this name. If \vref$_i$ is + some number $n$, this number denotes the $n$-th non dependent + premise of the {\term}, as determined by the type of {\term}. + + \ErrMsg \errindex{No such binder} + +\item A bindings list can also be a simple list of terms {\tt + \term$_1$ \dots\ \term$_n$}. In that case the references to + which these terms correspond are determined by the tactic. In case + of {\tt induction}, {\tt destruct}, {\tt elim} and {\tt case} (see + Section~\ref{elim}) the terms have to provide instances for all the + dependent products in the type of \term\ while in the case of {\tt + apply}, or of {\tt constructor} and its variants, only instances for + the dependent products that are not bound in the conclusion of the + type are required. + + \ErrMsg \errindex{Not the right number of missing arguments} +\end{itemize} + +\subsection{Occurrences sets and occurrences clauses} +\label{Occurrences clauses} +\index{Occurrences clauses} + +An occurrences clause is a modifier to some tactics that obeys the +following syntax: + +\begin{tabular}{lcl} +{\occclause} & ::= & {\tt in} {\occgoalset} \\ +{\occgoalset} & ::= & + \zeroone{{\ident$_1$} \zeroone{\atoccurrences} {\tt ,} \\ +& & {\dots} {\tt ,}\\ +& & {\ident$_m$} \zeroone{\atoccurrences}}\\ +& & \zeroone{{\tt |-} \zeroone{{\tt *} \zeroone{\atoccurrences}}}\\ +& | & + {\tt *} {\tt |-} \zeroone{{\tt *} \zeroone{\atoccurrences}}\\ +& | & + {\tt *}\\ +{\atoccurrences} & ::= & {\tt at} {\occlist}\\ +{\occlist} & ::= & \zeroone{{\tt -}} {\num$_1$} \dots\ {\num$_n$} +\end{tabular} + +The role of an occurrence clause is to select a set of occurrences of +a {\term} in a goal. In the first case, the {{\ident$_i$} +\zeroone{{\tt at} {\num$_1^i$} \dots\ {\num$_{n_i}^i$}}} parts +indicate that occurrences have to be selected in the hypotheses named +{\ident$_i$}. If no numbers are given for hypothesis {\ident$_i$}, +then all the occurrences of {\term} in the hypothesis are selected. If +numbers are given, they refer to occurrences of {\term} when the term +is printed using option {\tt Set Printing All} (see +Section~\ref{SetPrintingAll}), counting from left to right. In +particular, occurrences of {\term} in implicit arguments (see +Section~\ref{Implicit Arguments}) or coercions (see +Section~\ref{Coercions}) are counted. + +If a minus sign is given between {\tt at} and the list of occurrences, +it negates the condition so that the clause denotes all the occurrences except +the ones explicitly mentioned after the minus sign. + +As an exception to the left-to-right order, the occurrences in the +{\tt return} subexpression of a {\tt match} are considered {\em +before} the occurrences in the matched term. + +In the second case, the {\tt *} on the left of {\tt |-} means that +all occurrences of {\term} are selected in every hypothesis. + +In the first and second case, if {\tt *} is mentioned on the right of +{\tt |-}, the occurrences of the conclusion of the goal have to be +selected. If some numbers are given, then only the occurrences denoted +by these numbers are selected. In no numbers are given, all +occurrences of {\term} in the goal are selected. + +Finally, the last notation is an abbreviation for {\tt * |- *}. Note +also that {\tt |-} is optional in the first case when no {\tt *} is +given. + +Here are some tactics that understand occurrences clauses: +{\tt set}, {\tt remember}, {\tt induction}, {\tt destruct}. + +\SeeAlso~Sections~\ref{tactic:set}, \ref{Tac-induction}, \ref{SetPrintingAll}. + +\section{Applying theorems} -\subsection{\tt exact \term +\subsection{\tt exact \term} \tacindex{exact} -\label{exact}} +\label{exact} This tactic applies to any goal. It gives directly the exact proof term of the goal. Let {\T} be our goal, let {\tt p} be a term of type @@ -72,48 +162,17 @@ convertible (see Section~\ref{conv-rules}). \end{ErrMsgs} \begin{Variants} - \item \texttt{eexact \term}\tacindex{eexact} - - This tactic behaves like \texttt{exact} but is able to handle terms with meta-variables. - -\end{Variants} + \item \texttt{eexact \term}\tacindex{eexact} + This tactic behaves like \texttt{exact} but is able to handle terms + and goals with meta-variables. -\subsection{\tt refine \term -\tacindex{refine} -\label{refine} -\index{?@{\texttt{?}}}} - -This tactic allows to give an exact proof but still with some -holes. The holes are noted ``\texttt{\_}''. - -\begin{ErrMsgs} -\item \errindex{invalid argument}: - the tactic \texttt{refine} doesn't know what to do - with the term you gave. -\item \texttt{Refine passed ill-formed term}: the term you gave is not - a valid proof (not easy to debug in general). - This message may also occur in higher-level tactics, which call - \texttt{refine} internally. -\item \errindex{Cannot infer a term for this placeholder} - there is a hole in the term you gave - which type cannot be inferred. Put a cast around it. -\end{ErrMsgs} - -An example of use is given in Section~\ref{refine-example}. - -\section{Basics -\index{Typing rules}} - -Tactics presented in this section implement the basic typing rules of -{\CIC} given in Chapter~\ref{Cic}. +\end{Variants} -\subsection{{\tt assumption} -\tacindex{assumption}} +\subsection{\tt assumption} +\tacindex{assumption} -This tactic applies to any goal. It implements the -``Var''\index{Typing rules!Var} rule given in -Section~\ref{Typed-terms}. It looks in the local context for an +This tactic looks in the local context for an hypothesis which type is equal to the goal. If it is the case, the subgoal is proved. Otherwise, it fails. @@ -130,134 +189,470 @@ subgoal is proved. Otherwise, it fails. \end{Variants} +\subsection{\tt refine \term} +\tacindex{refine} +\label{refine} +\label{refine-example} +\index{?@{\texttt{?}}} -\subsection{\tt clear {\ident} -\tacindex{clear} -\label{clear}} - -This tactic erases the hypothesis named {\ident} in the local context -of the current goal. Then {\ident} is no more displayed and no more -usable in the proof development. +This tactic applies to any goal. It behaves like {\tt exact} with a big +difference: the user can leave some holes (denoted by \texttt{\_} or +{\tt (\_:\type)}) in the term. {\tt refine} will generate as +many subgoals as there are holes in the term. The type of holes must be +either synthesized by the system or declared by an +explicit cast like \verb|(_:nat->Prop)|. This low-level +tactic can be useful to advanced users. -\begin{Variants} +\Example -\item {\tt clear {\ident$_1$} {\ldots} {\ident$_n$}} - - This is equivalent to {\tt clear {\ident$_1$}. {\ldots} clear - {\ident$_n$}.} - -\item {\tt clearbody {\ident}}\tacindex{clearbody} +\begin{coq_example*} +Inductive Option : Set := + | Fail : Option + | Ok : bool -> Option. +\end{coq_example} +\begin{coq_example} +Definition get : forall x:Option, x <> Fail -> bool. +refine + (fun x:Option => + match x return x <> Fail -> bool with + | Fail => _ + | Ok b => fun _ => b + end). +intros; absurd (Fail = Fail); trivial. +\end{coq_example} +\begin{coq_example*} +Defined. +\end{coq_example*} - This tactic expects {\ident} to be a local definition then clears - its body. Otherwise said, this tactic turns a definition into an - assumption. +\begin{ErrMsgs} +\item \errindex{invalid argument}: + the tactic \texttt{refine} does not know what to do + with the term you gave. +\item \texttt{Refine passed ill-formed term}: the term you gave is not + a valid proof (not easy to debug in general). + This message may also occur in higher-level tactics that call + \texttt{refine} internally. +\item \errindex{Cannot infer a term for this placeholder}: + there is a hole in the term you gave + which type cannot be inferred. Put a cast around it. +\end{ErrMsgs} -\item \texttt{clear - {\ident$_1$} {\ldots} {\ident$_n$}} +\subsection{\tt apply \term} +\tacindex{apply} +\label{apply} - This tactic clears all hypotheses except the ones depending in - the hypotheses named {\ident$_1$} {\ldots} {\ident$_n$} and in the - goal. +This tactic applies to any goal. The argument {\term} is a term +well-formed in the local context. The tactic {\tt apply} tries to +match the current goal against the conclusion of the type of {\term}. +If it succeeds, then the tactic returns as many subgoals as the number +of non-dependent premises of the type of {\term}. If the conclusion of +the type of {\term} does not match the goal {\em and} the conclusion +is an inductive type isomorphic to a tuple type, then each component +of the tuple is recursively matched to the goal in the left-to-right +order. -\item \texttt{clear} +The tactic {\tt apply} relies on first-order unification with +dependent types unless the conclusion of the type of {\term} is of the +form {\tt ($P$ $t_1$ \dots\ $t_n$)} with $P$ to be instantiated. In +the latter case, the behavior depends on the form of the goal. If the +goal is of the form {\tt (fun $x$ => $Q$)~$u_1$~\ldots~$u_n$} and the +$t_i$ and $u_i$ unifies, then $P$ is taken to be {\tt (fun $x$ => $Q$)}. +Otherwise, {\tt apply} tries to define $P$ by abstracting over +$t_1$~\ldots ~$t_n$ in the goal. See {\tt pattern} in +Section~\ref{pattern} to transform the goal so that it gets the form +{\tt (fun $x$ => $Q$)~$u_1$~\ldots~$u_n$}. - This tactic clears all hypotheses except the ones depending in - goal. +\begin{ErrMsgs} +\item \errindex{Impossible to unify \dots\ with \dots} -\item {\tt clear dependent \ident \tacindex{clear dependent}} + The {\tt apply} + tactic failed to match the conclusion of {\term} and the current goal. + You can help the {\tt apply} tactic by transforming your + goal with the {\tt change} or {\tt pattern} tactics (see + sections~\ref{pattern},~\ref{change}). - This clears the hypothesis \ident\ and all hypotheses - which depend on it. +\item \errindex{Unable to find an instance for the variables +{\ident} \dots\ {\ident}} -\end{Variants} + This occurs when some instantiations of the premises of {\term} are not + deducible from the unification. This is the case, for instance, when + you want to apply a transitivity property. In this case, you have to + use one of the variants below: -\begin{ErrMsgs} -\item \errindex{{\ident} not found} -\item \errindexbis{{\ident} is used in the conclusion}{is used in the - conclusion} -\item \errindexbis{{\ident} is used in the hypothesis {\ident'}}{is - used in the hypothesis} \end{ErrMsgs} -\subsection{\tt move {\ident$_1$} after {\ident$_2$} -\tacindex{move} -\label{move}} +\begin{Variants} -This moves the hypothesis named {\ident$_1$} in the local context -after the hypothesis named {\ident$_2$}. +\item{\tt apply {\term} with {\term$_1$} \dots\ {\term$_n$}} + \tacindex{apply \dots\ with} -If {\ident$_1$} comes before {\ident$_2$} in the order of dependences, -then all hypotheses between {\ident$_1$} and {\ident$_2$} which -(possibly indirectly) depend on {\ident$_1$} are moved also. + Provides {\tt apply} with explicit instantiations for all dependent + premises of the type of {\term} which do not occur in the conclusion + and consequently cannot be found by unification. Notice that + {\term$_1$} \dots\ {\term$_n$} must be given according to the order + of these dependent premises of the type of {\term}. -If {\ident$_1$} comes after {\ident$_2$} in the order of dependences, -then all hypotheses between {\ident$_1$} and {\ident$_2$} which -(possibly indirectly) occur in {\ident$_1$} are moved also. + \ErrMsg \errindex{Not the right number of missing arguments} -\begin{Variants} +\item{\tt apply {\term} with ({\vref$_1$} := {\term$_1$}) \dots\ ({\vref$_n$} + := {\term$_n$})} -\item {\tt move {\ident$_1$} before {\ident$_2$}} + This also provides {\tt apply} with values for instantiating + premises. Here, variables are referred by names and non-dependent + products by increasing numbers (see syntax in Section~\ref{Binding-list}). -This moves {\ident$_1$} towards and just before the hypothesis named {\ident$_2$}. +\item {\tt apply} {\term$_1$} {\tt ,} \ldots {\tt ,} {\term$_n$} -\item {\tt move {\ident} at top} + This is a shortcut for {\tt apply} {\term$_1$} {\tt ; [ ..~|} + \ldots~{\tt ; [ ..~| {\tt apply} {\term$_n$} ]} \ldots~{\tt ]}, i.e. for the + successive applications of {\term$_{i+1}$} on the last subgoal + generated by {\tt apply} {\term$_i$}, starting from the application + of {\term$_1$}. -This moves {\ident} at the top of the local context (at the beginning of the context). +\item {\tt eapply \term}\tacindex{eapply}\label{eapply} -\item {\tt move {\ident} at bottom} + The tactic {\tt eapply} behaves like {\tt apply} but it does not fail + when no instantiations are deducible for some variables in the + premises. Rather, it turns these variables into so-called + existential variables which are variables still to instantiate. An + existential variable is identified by a name of the form {\tt ?$n$} + where $n$ is a number. The instantiation is intended to be found + later in the proof. -This moves {\ident} at the bottom of the local context (at the end of the context). +\item {\tt simple apply {\term}} \tacindex{simple apply} -\end{Variants} + This behaves like {\tt apply} but it reasons modulo conversion only + on subterms that contain no variables to instantiate. For instance, + the following example does not succeed because it would require the + conversion of {\tt id ?1234} and {\tt O}. -\begin{ErrMsgs} +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{coq_example*} +Definition id (x : nat) := x. +Hypothesis H : forall y, id y = y. +Goal O = O. +\end{coq_example*} +\begin{coq_example} +simple apply H. +\end{coq_example} -\item \errindex{{\ident$_i$} not found} + Because it reasons modulo a limited amount of conversion, {\tt + simple apply} fails quicker than {\tt apply} and it is then + well-suited for uses in used-defined tactics that backtrack often. + Moreover, it does not traverse tuples as {\tt apply} does. -\item \errindex{Cannot move {\ident$_1$} after {\ident$_2$}: - it occurs in {\ident$_2$}} +\item \zeroone{{\tt simple}} {\tt apply} {\term$_1$} \zeroone{{\tt with} + {\bindinglist$_1$}} {\tt ,} \ldots {\tt ,} {\term$_n$} \zeroone{{\tt with} + {\bindinglist$_n$}}\\ + \zeroone{{\tt simple}} {\tt eapply} {\term$_1$} \zeroone{{\tt with} + {\bindinglist$_1$}} {\tt ,} \ldots {\tt ,} {\term$_n$} \zeroone{{\tt with} + {\bindinglist$_n$}} -\item \errindex{Cannot move {\ident$_1$} after {\ident$_2$}: - it depends on {\ident$_2$}} + This summarizes the different syntaxes for {\tt apply} and {\tt eapply}. -\end{ErrMsgs} +\item {\tt lapply {\term}} \tacindex{lapply} -\subsection{\tt rename {\ident$_1$} into {\ident$_2$} -\tacindex{rename}} + This tactic applies to any goal, say {\tt G}. The argument {\term} + has to be well-formed in the current context, its type being + reducible to a non-dependent product {\tt A -> B} with {\tt B} + possibly containing products. Then it generates two subgoals {\tt + B->G} and {\tt A}. Applying {\tt lapply H} (where {\tt H} has type + {\tt A->B} and {\tt B} does not start with a product) does the same + as giving the sequence {\tt cut B. 2:apply H.} where {\tt cut} is + described below. -This renames hypothesis {\ident$_1$} into {\ident$_2$} in the current -context\footnote{but it does not rename the hypothesis in the - proof-term...} + \Warning When {\term} contains more than one non + dependent product the tactic {\tt lapply} only takes into account the + first product. -\begin{Variants} +\end{Variants} -\item {\tt rename {\ident$_1$} into {\ident$_2$}, \ldots, - {\ident$_{2k-1}$} into {\ident$_{2k}$}} +\Example +Assume we have a transitive relation {\tt R} on {\tt nat}: +\label{eapply-example} - Is equivalent to the sequence of the corresponding atomic {\tt rename}. +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{coq_example*} +Variable R : nat -> nat -> Prop. +Hypothesis Rtrans : forall x y z:nat, R x y -> R y z -> R x z. +Variables n m p : nat. +Hypothesis Rnm : R n m. +Hypothesis Rmp : R m p. +\end{coq_example*} -\end{Variants} +Consider the goal {\tt (R n p)} provable using the transitivity of +{\tt R}: + +\begin{coq_example*} +Goal R n p. +\end{coq_example*} + +The direct application of {\tt Rtrans} with {\tt apply} fails because +no value for {\tt y} in {\tt Rtrans} is found by {\tt apply}: + +%\begin{coq_eval} +%Set Printing Depth 50. +%(********** The following is not correct and should produce **********) +%(**** Error: generated subgoal (R n ?17) has metavariables in it *****) +%\end{coq_eval} +\begin{coq_example} +apply Rtrans. +\end{coq_example} + +A solution is to apply {\tt (Rtrans n m p)} or {\tt (Rtrans n m)}. + +\begin{coq_example} +apply (Rtrans n m p). +\end{coq_example} + +\begin{coq_eval} +Undo. +\end{coq_eval} + +Note that {\tt n} can be inferred from the goal, so the following would +work too. + +\begin{coq_example*} +apply (Rtrans _ m). +\end{coq_example*} +\begin{coq_eval} +Undo. +\end{coq_eval} + +More elegantly, {\tt apply Rtrans with (y:=m)} allows to only mention +the unknown {\tt m}: + +\begin{coq_example*} +apply Rtrans with (y := m). +\end{coq_example*} +\begin{coq_eval} +Undo. +\end{coq_eval} + +Another solution is to mention the proof of {\tt (R x y)} in {\tt +Rtrans} \ldots + +\begin{coq_example} +apply Rtrans with (1 := Rnm). +\end{coq_example} +\begin{coq_eval} +Undo. +\end{coq_eval} + +\ldots or the proof of {\tt (R y z)}. + +\begin{coq_example} +apply Rtrans with (2 := Rmp). +\end{coq_example} +\begin{coq_eval} +Undo. +\end{coq_eval} + +On the opposite, one can use {\tt eapply} which postpone the problem +of finding {\tt m}. Then one can apply the hypotheses {\tt Rnm} and {\tt +Rmp}. This instantiates the existential variable and completes the proof. + +\begin{coq_example} +eapply Rtrans. +apply Rnm. +apply Rmp. +\end{coq_example} + +\begin{coq_eval} +Reset R. +\end{coq_eval} + +\subsection{\tt apply {\term} in {\ident}} +\tacindex{apply \dots\ in} + +This tactic applies to any goal. The argument {\term} is a term +well-formed in the local context and the argument {\ident} is an +hypothesis of the context. The tactic {\tt apply {\term} in {\ident}} +tries to match the conclusion of the type of {\ident} against a +non-dependent premise of the type of {\term}, trying them from right to +left. If it succeeds, the statement of hypothesis {\ident} is +replaced by the conclusion of the type of {\term}. The tactic also +returns as many subgoals as the number of other non-dependent premises +in the type of {\term} and of the non-dependent premises of the type +of {\ident}. If the conclusion of the type of {\term} does not match +the goal {\em and} the conclusion is an inductive type isomorphic to a +tuple type, then the tuple is (recursively) decomposed and the first +component of the tuple of which a non-dependent premise matches the +conclusion of the type of {\ident}. Tuples are decomposed in a +width-first left-to-right order (for instance if the type of {\tt H1} +is a \verb=A <-> B= statement, and the type of {\tt H2} is \verb=A= +then {\tt apply H1 in H2} transforms the type of {\tt H2} into {\tt + B}). The tactic {\tt apply} relies on first-order pattern-matching +with dependent types. \begin{ErrMsgs} +\item \errindex{Statement without assumptions} -\item \errindex{{\ident$_1$} not found} +This happens if the type of {\term} has no non dependent premise. -\item \errindexbis{{\ident$_2$} is already used}{is already used} +\item \errindex{Unable to apply} + +This happens if the conclusion of {\ident} does not match any of the +non dependent premises of the type of {\term}. +\end{ErrMsgs} + +\begin{Variants} +\item {\tt apply \nelist{\term}{,} in {\ident}} + +This applies each of {\term} in sequence in {\ident}. + +\item {\tt apply \nelist{{\term} with {\bindinglist}}{,} in {\ident}} + +This does the same but uses the bindings in each {\bindinglist} to +instantiate the parameters of the corresponding type of {\term} +(see syntax of bindings in Section~\ref{Binding-list}). + +\item {\tt eapply \nelist{{\term} with {\bindinglist}}{,} in {\ident}} +\tacindex{eapply \dots\ in} + +This works as {\tt apply \nelist{{\term} with {\bindinglist}}{,} in +{\ident}} but turns unresolved bindings into existential variables, if +any, instead of failing. + +\item {\tt apply \nelist{{\term}{,} with {\bindinglist}}{,} in {\ident} as {\disjconjintropattern}} + +This works as {\tt apply \nelist{{\term}{,} with {\bindinglist}}{,} in +{\ident}} then destructs the hypothesis {\ident} along +{\disjconjintropattern} as {\tt destruct {\ident} as +{\disjconjintropattern}} would. + +\item {\tt eapply \nelist{{\term}{,} with {\bindinglist}}{,} in {\ident} as {\disjconjintropattern}} + +This works as {\tt apply \nelist{{\term}{,} with {\bindinglist}}{,} in {\ident} as {\disjconjintropattern}} but using {\tt eapply}. + +\item {\tt simple apply {\term} in {\ident}} +\tacindex{simple apply \dots\ in} +\tacindex{simple eapply \dots\ in} + +This behaves like {\tt apply {\term} in {\ident}} but it reasons +modulo conversion only on subterms that contain no variables to +instantiate. For instance, if {\tt id := fun x:nat => x} and {\tt H : + forall y, id y = y -> True} and {\tt H0 :\ O = O} then {\tt simple + apply H in H0} does not succeed because it would require the +conversion of {\tt id ?1234} and {\tt O} where {\tt ?1234} is a variable to +instantiate. Tactic {\tt simple apply {\term} in {\ident}} does not +either traverse tuples as {\tt apply {\term} in {\ident}} does. + +\item {\tt \zeroone{simple} apply \nelist{{\term} \zeroone{with {\bindinglist}}}{,} in {\ident} \zeroone{as {\disjconjintropattern}}}\\ +{\tt \zeroone{simple} eapply \nelist{{\term} \zeroone{with {\bindinglist}}}{,} in {\ident} \zeroone{as {\disjconjintropattern}}} + +This summarizes the different syntactic variants of {\tt apply {\term} + in {\ident}} and {\tt eapply {\term} in {\ident}}. +\end{Variants} + +\subsection{\tt constructor \num} +\label{constructor} +\tacindex{constructor} + +This tactic applies to a goal such that its conclusion is +an inductive type (say {\tt I}). The argument {\num} must be less +or equal to the numbers of constructor(s) of {\tt I}. Let {\tt ci} be +the {\tt i}-th constructor of {\tt I}, then {\tt constructor i} is +equivalent to {\tt intros; apply ci}. +\begin{ErrMsgs} +\item \errindex{Not an inductive product} +\item \errindex{Not enough constructors} \end{ErrMsgs} -\subsection{\tt intro +\begin{Variants} +\item \texttt{constructor} + + This tries \texttt{constructor 1} then \texttt{constructor 2}, + \dots\ , then \texttt{constructor} \textit{n} where \textit{n} is + the number of constructors of the head of the goal. + +\item {\tt constructor \num~with} {\bindinglist} + + Let {\tt ci} be the {\tt i}-th constructor of {\tt I}, then {\tt + constructor i with \bindinglist} is equivalent to {\tt intros; + apply ci with \bindinglist}. + + \Warning the terms in the \bindinglist\ are checked + in the context where {\tt constructor} is executed and not in the + context where {\tt apply} is executed (the introductions are not + taken into account). + +% To document? +% \item {\tt constructor {\tactic}} + +\item {\tt split}\tacindex{split} + + This applies only if {\tt I} has a single constructor. It is then + equivalent to {\tt constructor 1}. It is typically used in the case + of a conjunction $A\land B$. + + \ErrMsg \errindex{Not an inductive goal with 1 constructor} + +\item {\tt exists {\bindinglist}}\tacindex{exists} + + This applies only if {\tt I} has a single constructor. It is then + equivalent to {\tt intros; constructor 1 with \bindinglist}. It is + typically used in the case of an existential quantification $\exists + x, P(x)$. + + \ErrMsg \errindex{Not an inductive goal with 1 constructor} + +\item {\tt exists \nelist{\bindinglist}{,}} + + This iteratively applies {\tt exists {\bindinglist}}. + +\item {\tt left}\tacindex{left}\\ + {\tt right}\tacindex{right} + + These tactics apply only if {\tt I} has two constructors, for instance + in the case of a + disjunction $A\lor B$. Then, they are respectively equivalent to {\tt + constructor 1} and {\tt constructor 2}. + + \ErrMsg \errindex{Not an inductive goal with 2 constructors} + +\item {\tt left with \bindinglist}\\ + {\tt right with \bindinglist}\\ + {\tt split with \bindinglist} + + As soon as the inductive type has the right number of constructors, + these expressions are equivalent to calling {\tt + constructor $i$ with \bindinglist} for the appropriate $i$. + +\item \texttt{econstructor}\tacindex{econstructor}\\ + \texttt{eexists}\tacindex{eexists}\\ + \texttt{esplit}\tacindex{esplit}\\ + \texttt{eleft}\tacindex{eleft}\\ + \texttt{eright}\tacindex{eright} + + These tactics and their variants behave like \texttt{constructor}, + \texttt{exists}, \texttt{split}, \texttt{left}, \texttt{right} and + their variants but they introduce existential variables instead of + failing when the instantiation of a variable cannot be found (cf + \texttt{eapply} and Section~\ref{eapply-example}). + +\end{Variants} + +\section{Managing the local context} + +\subsection{\tt intro} \tacindex{intro} -\label{intro}} +\label{intro} -This tactic applies to a goal which is either a product or starts with +This tactic applies to a goal that is either a product or starts with a let binder. If the goal is a product, the tactic implements the ``Lam''\index{Typing rules!Lam} rule given in Section~\ref{Typed-terms}\footnote{Actually, only the second subgoal will be generated since the other one can be automatically checked.}. If the -goal starts with a let binder then the tactic implements a mix of the +goal starts with a let binder, then the tactic implements a mix of the ``Let''\index{Typing rules!Let} and ``Conv''\index{Typing rules!Conv}. -If the current goal is a dependent product {\tt forall $x$:$T$, $U$} (resp {\tt +If the current goal is a dependent product $\forall x:T,~U$ (resp {\tt let $x$:=$t$ in $U$}) then {\tt intro} puts {\tt $x$:$T$} (resp {\tt $x$:=$t$}) in the local context. % Obsolete (quantified names already avoid hypotheses names): @@ -265,15 +660,15 @@ let $x$:=$t$ in $U$}) then {\tt intro} puts {\tt $x$:$T$} (resp {\tt $x$:=$t$}) % {\tt x}{\it n}{\tt :T} where {\it n} is such that {\tt x}{\it n} is a %fresh name. The new subgoal is $U$. -% If the {\tt x} has been renamed {\tt x}{\it n} then it is replaced -% by {\tt x}{\it n} in {\tt U}. +% If the {\tt x} has been renamed {\tt x}{\it n} then it is replaced +% by {\tt x}{\it n} in {\tt U}. -If the goal is a non dependent product {\tt $T$ -> $U$}, then it puts +If the goal is a non-dependent product $T \to U$, then it puts in the local context either {\tt H}{\it n}{\tt :$T$} (if $T$ is of type {\tt Set} or {\tt Prop}) or {\tt X}{\it n}{\tt :$T$} (if the type of $T$ is {\tt Type}). The optional index {\it n} is such that {\tt H}{\it n} or {\tt X}{\it n} is a fresh identifier. -In both cases the new subgoal is $U$. +In both cases, the new subgoal is $U$. If the goal is neither a product nor starting with a let definition, the tactic {\tt intro} applies the tactic {\tt red} until the tactic @@ -288,12 +683,12 @@ the tactic {\tt intro} applies the tactic {\tt red} until the tactic \item {\tt intros}\tacindex{intros} - Repeats {\tt intro} until it meets the head-constant. It never reduces + This repeats {\tt intro} until it meets the head-constant. It never reduces head-constants and it never fails. \item {\tt intro {\ident}} - Applies {\tt intro} but forces {\ident} to be the name of the + This applies {\tt intro} but forces {\ident} to be the name of the introduced hypothesis. \ErrMsg \errindex{name {\ident} is already used} @@ -302,28 +697,28 @@ the tactic {\tt intro} applies the tactic {\tt red} until the tactic constant then the latter can still be referred to by a qualified name (see \ref{LongNames}). -\item {\tt intros \ident$_1$ \dots\ \ident$_n$} - - Is equivalent to the composed tactic {\tt intro \ident$_1$; \dots\ ; +\item {\tt intros \ident$_1$ \dots\ \ident$_n$} + + This is equivalent to the composed tactic {\tt intro \ident$_1$; \dots\ ; intro \ident$_n$}. More generally, the \texttt{intros} tactic takes a pattern as argument in order to introduce names for components of an inductive - definition or to clear introduced hypotheses; This is explained + definition or to clear introduced hypotheses. This is explained in~\ref{intros-pattern}. \item {\tt intros until {\ident}} \tacindex{intros until} - - Repeats {\tt intro} until it meets a premise of the goal having form + + This repeats {\tt intro} until it meets a premise of the goal having form {\tt (} {\ident}~{\tt :}~{\term} {\tt )} and discharges the variable named {\ident} of the current goal. \ErrMsg \errindex{No such hypothesis in current goal} - + \item {\tt intros until {\num}} \tacindex{intros until} - - Repeats {\tt intro} until the {\num}-th non-dependent product. For - instance, on the subgoal % + + This repeats {\tt intro} until the {\num}-th non-dependent product. For + instance, on the subgoal % \verb+forall x y:nat, x=y -> y=x+ the tactic \texttt{intros until 1} is equivalent to \texttt{intros x y H}, as \verb+x=y -> y=x+ is the first non-dependent product. And on the subgoal % @@ -335,7 +730,7 @@ the tactic {\tt intro} applies the tactic {\tt red} until the tactic \ErrMsg \errindex{No such hypothesis in current goal} - Happens when {\num} is 0 or is greater than the number of non-dependent + This happens when {\num} is 0 or is greater than the number of non-dependent products of the goal. \item {\tt intro after \ident} \tacindex{intro after}\\ @@ -343,7 +738,7 @@ the tactic {\tt intro} applies the tactic {\tt red} until the tactic {\tt intro at top} \tacindex{intro at top}\\ {\tt intro at bottom} \tacindex{intro at bottom} - Applies {\tt intro} and moves the freshly introduced hypothesis + These tactics apply {\tt intro} and move the freshly introduced hypothesis respectively after the hypothesis \ident{}, before the hypothesis \ident{}, at the top of the local context, or at the bottom of the local context. All hypotheses on which the new hypothesis depends @@ -351,337 +746,412 @@ the tactic {\tt intro} applies the tactic {\tt red} until the tactic hypotheses. Note that {\tt intro at bottom} is a synonym for {\tt intro} with no argument. -\begin{ErrMsgs} -\item \errindex{No product even after head-reduction} -\item \errindex{No such hypothesis} : {\ident} -\end{ErrMsgs} + \ErrMsg \errindex{No such hypothesis} : {\ident} \item {\tt intro \ident$_1$ after \ident$_2$}\\ {\tt intro \ident$_1$ before \ident$_2$}\\ {\tt intro \ident$_1$ at top}\\ {\tt intro \ident$_1$ at bottom} - - Behaves as previously but naming the introduced hypothesis + + These tactics behave as previously but naming the introduced hypothesis \ident$_1$. It is equivalent to {\tt intro \ident$_1$} followed by the appropriate call to {\tt move}~(see Section~\ref{move}). -\begin{ErrMsgs} -\item \errindex{No product even after head-reduction} -\item \errindex{No such hypothesis} : {\ident} -\end{ErrMsgs} - \end{Variants} -\subsection{\tt apply \term -\tacindex{apply} -\label{apply}} - -This tactic applies to any goal. The argument {\term} is a term -well-formed in the local context. The tactic {\tt apply} tries to -match the current goal against the conclusion of the type of {\term}. -If it succeeds, then the tactic returns as many subgoals as the number -of non dependent premises of the type of {\term}. If the conclusion of -the type of {\term} does not match the goal {\em and} the conclusion -is an inductive type isomorphic to a tuple type, then each component -of the tuple is recursively matched to the goal in the left-to-right -order. - -The tactic {\tt apply} relies on first-order unification with -dependent types unless the conclusion of the type of {\term} is of the -form {\tt ($P$~ $t_1$~\ldots ~$t_n$)} with $P$ to be instantiated. In -the latter case, the behavior depends on the form of the goal. If the -goal is of the form {\tt (fun $x$ => $Q$)~$u_1$~\ldots~$u_n$} and the -$t_i$ and $u_i$ unifies, then $P$ is taken to be (fun $x$ => $Q$). -Otherwise, {\tt apply} tries to define $P$ by abstracting over -$t_1$~\ldots ~$t_n$ in the goal. See {\tt pattern} in -Section~\ref{pattern} to transform the goal so that it gets the form -{\tt (fun $x$ => $Q$)~$u_1$~\ldots~$u_n$}. - -\begin{ErrMsgs} -\item \errindex{Impossible to unify \dots\ with \dots} - - The {\tt apply} - tactic failed to match the conclusion of {\term} and the current goal. - You can help the {\tt apply} tactic by transforming your - goal with the {\tt change} or {\tt pattern} tactics (see - sections~\ref{pattern},~\ref{change}). - -\item \errindex{Unable to find an instance for the variables -{\ident} \ldots {\ident}} +\subsection{\tt intros {\intropattern} {\ldots} {\intropattern}} +\label{intros-pattern} +\tacindex{intros \intropattern} +\index{Introduction patterns} +\index{Naming introduction patterns} +\index{Disjunctive/conjunctive introduction patterns} - This occurs when some instantiations of the premises of {\term} are not - deducible from the unification. This is the case, for instance, when - you want to apply a transitivity property. In this case, you have to - use one of the variants below: +This extension of the tactic {\tt intros} combines introduction of +variables or hypotheses and case analysis. An {\em introduction pattern} is +either: +\begin{itemize} +\item a {\em naming introduction pattern}, i.e. either one of: + \begin{itemize} + \item the pattern \texttt{?} + \item the pattern \texttt{?\ident} + \item an identifier + \end{itemize} +\item a {\em disjunctive/conjunctive introduction pattern}, i.e. either one of: + \begin{itemize} + \item a disjunction of lists of patterns: + {\tt [$p_{11}$ \dots\ $p_{1m_1}$ | \dots\ | $p_{11}$ \dots\ $p_{nm_n}$]} + \item a conjunction of patterns: {\tt ($p_1$ , \dots\ , $p_n$)} + \item a list of patterns {\tt ($p_1$ \&\ \dots\ \&\ $p_n$)} + for sequence of right-associative binary constructs + \end{itemize} +\item the wildcard: {\tt \_} +\item the rewriting orientations: {\tt ->} or {\tt <-} +\end{itemize} -\end{ErrMsgs} +Assuming a goal of type $Q \to P$ (non-dependent product), or +of type $\forall x:T,~P$ (dependent product), the behavior of +{\tt intros $p$} is defined inductively over the structure of the +introduction pattern~$p$: +\begin{itemize} +\item introduction on \texttt{?} performs the introduction, and lets {\Coq} + choose a fresh name for the variable; +\item introduction on \texttt{?\ident} performs the introduction, and + lets {\Coq} choose a fresh name for the variable based on {\ident}; +\item introduction on \texttt{\ident} behaves as described in + Section~\ref{intro}; +\item introduction over a disjunction of list of patterns {\tt + [$p_{11}$ \dots\ $p_{1m_1}$ | \dots\ | $p_{11}$ \dots\ $p_{nm_n}$]} + expects the product to be over an inductive type + whose number of constructors is $n$ (or more generally over a type + of conclusion an inductive type built from $n$ constructors, + e.g. {\tt C -> A\textbackslash/B if $n=2$}): it destructs the introduced + hypothesis as {\tt destruct} (see Section~\ref{destruct}) would and + applies on each generated subgoal the corresponding tactic; + \texttt{intros}~$p_{i1}$ {\ldots} $p_{im_i}$; if the disjunctive + pattern is part of a sequence of patterns and is not the last + pattern of the sequence, then {\Coq} completes the pattern so that all + the argument of the constructors of the inductive type are + introduced (for instance, the list of patterns {\tt [$\;$|$\;$] H} + applied on goal {\tt forall x:nat, x=0 -> 0=x} behaves the same as + the list of patterns {\tt [$\,$|$\,$?$\,$] H}); +\item introduction over a conjunction of patterns {\tt ($p_1$, \ldots, + $p_n$)} expects the goal to be a product over an inductive type $I$ with a + single constructor that itself has at least $n$ arguments: it + performs a case analysis over the hypothesis, as {\tt destruct} + would, and applies the patterns $p_1$~\ldots~$p_n$ to the arguments + of the constructor of $I$ (observe that {\tt ($p_1$, {\ldots}, + $p_n$)} is an alternative notation for {\tt [$p_1$ {\ldots} + $p_n$]}); +\item introduction via {\tt ($p_1$ \& \dots\ \& $p_n$)} + is a shortcut for introduction via + {\tt ($p_1$,(\ldots,(\dots,$p_n$)\ldots))}; it expects the + hypothesis to be a sequence of right-associative binary inductive + constructors such as {\tt conj} or {\tt ex\_intro}; for instance, an + hypothesis with type {\tt A\verb|/\|exists x, B\verb|/\|C\verb|/\|D} can be + introduced via pattern {\tt (a \& x \& b \& c \& d)}; +\item introduction on the wildcard depends on whether the product is + dependent or not: in the non-dependent case, it erases the + corresponding hypothesis (i.e. it behaves as an {\tt intro} followed + by a {\tt clear}, cf Section~\ref{clear}) while in the dependent + case, it succeeds and erases the variable only if the wildcard is + part of a more complex list of introduction patterns that also + erases the hypotheses depending on this variable; +\item introduction over {\tt ->} (respectively {\tt <-}) expects the + hypothesis to be an equality and the right-hand-side (respectively + the left-hand-side) is replaced by the left-hand-side (respectively + the right-hand-side) in both the conclusion and the context of the goal; + if moreover the term to substitute is a variable, the hypothesis is + removed. +\end{itemize} -\begin{Variants} +\Example -\item{\tt apply {\term} with {\term$_1$} \dots\ {\term$_n$}} - \tacindex{apply \dots\ with} - - Provides {\tt apply} with explicit instantiations for all dependent - premises of the type of {\term} which do not occur in the conclusion - and consequently cannot be found by unification. Notice that - {\term$_1$} \dots\ {\term$_n$} must be given according to the order - of these dependent premises of the type of {\term}. +\begin{coq_example} +Goal forall A B C:Prop, A \/ B /\ C -> (A -> C) -> C. +intros A B C [a| [_ c]] f. +apply (f a). +exact c. +Qed. +\end{coq_example} - \ErrMsg \errindex{Not the right number of missing arguments} +\Rem {\tt intros $p_1~\ldots~p_n$} is not equivalent to \texttt{intros + $p_1$;\ldots; intros $p_n$} for the following reasons: +\begin{itemize} +\item A wildcard pattern never succeeds when applied isolated on a + dependent product, while it succeeds as part of a list of + introduction patterns if the hypotheses that depends on it are + erased too. +\item A disjunctive or conjunctive pattern followed by an introduction + pattern forces the introduction in the context of all arguments of + the constructors before applying the next pattern while a terminal + disjunctive or conjunctive pattern does not. Here is an example -\item{\tt apply {\term} with ({\vref$_1$} := {\term$_1$}) \dots\ ({\vref$_n$} - := {\term$_n$})} - - This also provides {\tt apply} with values for instantiating - premises. Here, variables are referred by names and non-dependent - products by increasing numbers (see syntax in Section~\ref{Binding-list}). +\begin{coq_example} +Goal forall n:nat, n = 0 -> n = 0. +intros [ | ] H. +Show 2. +Undo. +intros [ | ]; intros H. +Show 2. +\end{coq_example} -\item {\tt apply} {\term$_1$} {\tt ,} \ldots {\tt ,} {\term$_n$} +\end{itemize} - This is a shortcut for {\tt apply} {\term$_1$} {\tt ; [ ..~|} - \ldots~{\tt ; [ ..~| {\tt apply} {\term$_n$} ]} \ldots~{\tt ]}, i.e. for the - successive applications of {\term$_{i+1}$} on the last subgoal - generated by {\tt apply} {\term$_i$}, starting from the application - of {\term$_1$}. +\subsection{\tt clear \ident} +\tacindex{clear} +\label{clear} -\item {\tt eapply \term}\tacindex{eapply}\label{eapply} - - The tactic {\tt eapply} behaves as {\tt apply} but does not fail - when no instantiation are deducible for some variables in the - premises. Rather, it turns these variables into so-called - existential variables which are variables still to instantiate. An - existential variable is identified by a name of the form {\tt ?$n$} - where $n$ is a number. The instantiation is intended to be found - later in the proof. +This tactic erases the hypothesis named {\ident} in the local context +of the current goal. As a consequence, {\ident} is no more displayed and no more +usable in the proof development. - An example of use of {\tt eapply} is given in - Section~\ref{eapply-example}. +\begin{ErrMsgs} +\item \errindex{No such hypothesis} +\item \errindexbis{{\ident} is used in the conclusion}{is used in the + conclusion} +\item \errindexbis{{\ident} is used in the hypothesis {\ident'}}{is + used in the hypothesis} +\end{ErrMsgs} -\item {\tt simple apply {\term}} \tacindex{simple apply} +\begin{Variants} - This behaves like {\tt apply} but it reasons modulo conversion only - on subterms that contain no variables to instantiate. For instance, - if {\tt id := fun x:nat => x} and {\tt H : forall y, id y = y} then - {\tt simple apply H} on goal {\tt O = O} does not succeed because it - would require the conversion of {\tt f ?y} and {\tt O} where {\tt - ?y} is a variable to instantiate. Tactic {\tt simple apply} does not - either traverse tuples as {\tt apply} does. +\item {\tt clear {\ident$_1$} \dots\ {\ident$_n$}} - Because it reasons modulo a limited amount of conversion, {\tt - simple apply} fails quicker than {\tt apply} and it is then - well-suited for uses in used-defined tactics that backtrack often. + This is equivalent to {\tt clear {\ident$_1$}. {\ldots} clear + {\ident$_n$}.} -\item \zeroone{{\tt simple}} {\tt apply} {\term$_1$} \zeroone{{\tt with} - {\bindinglist$_1$}} {\tt ,} \ldots {\tt ,} {\term$_n$} \zeroone{{\tt with} - {\bindinglist$_n$}}\\ - \zeroone{{\tt simple}} {\tt eapply} {\term$_1$} \zeroone{{\tt with} - {\bindinglist$_1$}} {\tt ,} \ldots {\tt ,} {\term$_n$} \zeroone{{\tt with} - {\bindinglist$_n$}} +\item {\tt clearbody {\ident}}\tacindex{clearbody} - This summarizes the different syntaxes for {\tt apply} and {\tt eapply}. + This tactic expects {\ident} to be a local definition then clears + its body. Otherwise said, this tactic turns a definition into an + assumption. -\item {\tt lapply {\term}} \tacindex{lapply} + \ErrMsg \errindexbis{{\ident} is not a local definition}{is not a local definition} - This tactic applies to any goal, say {\tt G}. The argument {\term} - has to be well-formed in the current context, its type being - reducible to a non-dependent product {\tt A -> B} with {\tt B} - possibly containing products. Then it generates two subgoals {\tt - B->G} and {\tt A}. Applying {\tt lapply H} (where {\tt H} has type - {\tt A->B} and {\tt B} does not start with a product) does the same - as giving the sequence {\tt cut B. 2:apply H.} where {\tt cut} is - described below. +\item \texttt{clear - {\ident$_1$} \dots\ {\ident$_n$}} - \Warning When {\term} contains more than one non - dependent product the tactic {\tt lapply} only takes into account the - first product. + This tactic clears all the hypotheses except the ones depending in + the hypotheses named {\ident$_1$} {\ldots} {\ident$_n$} and in the + goal. + +\item \texttt{clear} + + This tactic clears all the hypotheses except the ones the goal depends on. + +\item {\tt clear dependent \ident \tacindex{clear dependent}} + + This clears the hypothesis \ident\ and all the hypotheses + that depend on it. \end{Variants} -\subsection{{\tt apply {\term} in {\ident}} -\tacindex{apply \ldots\ in}} +\subsection{\tt revert \ident$_1$ \dots\ \ident$_n$} +\tacindex{revert} +\label{revert} -This tactic applies to any goal. The argument {\term} is a term -well-formed in the local context and the argument {\ident} is an -hypothesis of the context. The tactic {\tt apply {\term} in {\ident}} -tries to match the conclusion of the type of {\ident} against a non -dependent premise of the type of {\term}, trying them from right to -left. If it succeeds, the statement of hypothesis {\ident} is -replaced by the conclusion of the type of {\term}. The tactic also -returns as many subgoals as the number of other non dependent premises -in the type of {\term} and of the non dependent premises of the type -of {\ident}. If the conclusion of the type of {\term} does not match -the goal {\em and} the conclusion is an inductive type isomorphic to a -tuple type, then the tuple is (recursively) decomposed and the first -component of the tuple of which a non dependent premise matches the -conclusion of the type of {\ident}. Tuples are decomposed in a -width-first left-to-right order (for instance if the type of {\tt H1} -is a \verb=A <-> B= statement, and the type of {\tt H2} is \verb=A= -then {\tt apply H1 in H2} transforms the type of {\tt H2} into {\tt - B}). The tactic {\tt apply} relies on first-order pattern-matching -with dependent types. +This applies to any goal with variables \ident$_1$ \dots\ \ident$_n$. +It moves the hypotheses (possibly defined) to the goal, if this respects +dependencies. This tactic is the inverse of {\tt intro}. \begin{ErrMsgs} -\item \errindex{Statement without assumptions} +\item \errindex{No such hypothesis} +\item \errindexbis{{\ident} is used in the hypothesis {\ident'}}{is + used in the hypothesis} +\end{ErrMsgs} -This happens if the type of {\term} has no non dependent premise. +\begin{Variants} +\item {\tt revert dependent \ident \tacindex{revert dependent}} -\item \errindex{Unable to apply} + This moves to the goal the hypothesis \ident\ and all hypotheses + which depend on it. -This happens if the conclusion of {\ident} does not match any of the -non dependent premises of the type of {\term}. -\end{ErrMsgs} +\end{Variants} + +\subsection{\tt move {\ident$_1$} after {\ident$_2$}} +\tacindex{move} +\label{move} + +This moves the hypothesis named {\ident$_1$} in the local context +after the hypothesis named {\ident$_2$}. The proof term is not changed. + +If {\ident$_1$} comes before {\ident$_2$} in the order of dependences, +then all hypotheses between {\ident$_1$} and {\ident$_2$} that +(possibly indirectly) depend on {\ident$_1$} are moved also. + +If {\ident$_1$} comes after {\ident$_2$} in the order of dependences, +then all hypotheses between {\ident$_1$} and {\ident$_2$} that +(possibly indirectly) occur in {\ident$_1$} are moved also. \begin{Variants} -\item {\tt apply \nelist{\term}{,} in {\ident}} -This applies each of {\term} in sequence in {\ident}. +\item {\tt move {\ident$_1$} before {\ident$_2$}} -\item {\tt apply \nelist{{\term} with {\bindinglist}}{,} in {\ident}} +This moves {\ident$_1$} towards and just before the hypothesis named {\ident$_2$}. -This does the same but uses the bindings in each {\bindinglist} to -instantiate the parameters of the corresponding type of {\term} -(see syntax of bindings in Section~\ref{Binding-list}). +\item {\tt move {\ident} at top} -\item {\tt eapply \nelist{{\term} with {\bindinglist}}{,} in {\ident}} -\tacindex{eapply {\ldots} in} +This moves {\ident} at the top of the local context (at the beginning of the context). -This works as {\tt apply \nelist{{\term} with {\bindinglist}}{,} in -{\ident}} but turns unresolved bindings into existential variables, if -any, instead of failing. +\item {\tt move {\ident} at bottom} -\item {\tt apply \nelist{{\term}{,} with {\bindinglist}}{,} in {\ident} as {\disjconjintropattern}} +This moves {\ident} at the bottom of the local context (at the end of the context). -This works as {\tt apply \nelist{{\term}{,} with {\bindinglist}}{,} in -{\ident}} then destructs the hypothesis {\ident} along -{\disjconjintropattern} as {\tt destruct {\ident} as -{\disjconjintropattern}} would. +\end{Variants} -\item {\tt eapply \nelist{{\term}{,} with {\bindinglist}}{,} in {\ident} as {\disjconjintropattern}} +\begin{ErrMsgs} -This works as {\tt apply \nelist{{\term}{,} with {\bindinglist}}{,} in {\ident} as {\disjconjintropattern}} but using {\tt eapply}. +\item \errindex{No such hypothesis} -\item {\tt simple apply {\term} in {\ident}} -\tacindex{simple apply {\ldots} in} -\tacindex{simple eapply {\ldots} in} +\item \errindex{Cannot move {\ident$_1$} after {\ident$_2$}: + it occurs in {\ident$_2$}} -This behaves like {\tt apply {\term} in {\ident}} but it reasons -modulo conversion only on subterms that contain no variables to -instantiate. For instance, if {\tt id := fun x:nat => x} and {\tt H : - forall y, id y = y -> True} and {\tt H0 : O = O} then {\tt simple - apply H in H0} does not succeed because it would require the -conversion of {\tt f ?y} and {\tt O} where {\tt ?y} is a variable to -instantiate. Tactic {\tt simple apply {\term} in {\ident}} does not -either traverse tuples as {\tt apply {\term} in {\ident}} does. +\item \errindex{Cannot move {\ident$_1$} after {\ident$_2$}: + it depends on {\ident$_2$}} -\item {\tt \zeroone{simple} apply \nelist{{\term} \zeroone{with {\bindinglist}}}{,} in {\ident} \zeroone{as {\disjconjintropattern}}}\\ -{\tt \zeroone{simple} eapply \nelist{{\term} \zeroone{with {\bindinglist}}}{,} in {\ident} \zeroone{as {\disjconjintropattern}}} +\end{ErrMsgs} + +\subsection{\tt rename {\ident$_1$} into {\ident$_2$}} +\tacindex{rename} + +This renames hypothesis {\ident$_1$} into {\ident$_2$} in the current +context. The name of the hypothesis in the proof-term, however, is left +unchanged. + +\begin{Variants} + +\item {\tt rename {\ident$_1$} into {\ident$_2$}, \ldots, + {\ident$_{2k-1}$} into {\ident$_{2k}$}} + +This is equivalent to the sequence of the corresponding atomic {\tt rename}. -This summarizes the different syntactic variants of {\tt apply {\term} - in {\ident}} and {\tt eapply {\term} in {\ident}}. \end{Variants} -\subsection{{\tt set ( {\ident} {\tt :=} {\term} \tt )} +\begin{ErrMsgs} +\item \errindex{No such hypothesis} +\item \errindexbis{{\ident$_2$} is already used}{is already used} +\end{ErrMsgs} + +\subsection{\tt set ( {\ident} := {\term} )} \label{tactic:set} \tacindex{set} -\tacindex{pose} -\tacindex{remember}} -This replaces {\term} by {\ident} in the conclusion or in the -hypotheses of the current goal and adds the new definition {\ident -{\tt :=} \term} to the local context. The default is to make this -replacement only in the conclusion. +This replaces {\term} by {\ident} in the conclusion of the current goal +and adds the new definition {\tt {\ident} := \term} to the local context. If {\term} has holes (i.e. subexpressions of the form ``\_''), the tactic first checks that all subterms matching the pattern are compatible before doing the replacement using the leftmost subterm matching the pattern. +\begin{ErrMsgs} +\item \errindex{The variable {\ident} is already defined} +\end{ErrMsgs} + \begin{Variants} -\item {\tt set (} {\ident} {\tt :=} {\term} {\tt ) in {\occgoalset}} +\item {\tt set ( {\ident} := {\term} ) in {\occgoalset}} This notation allows to specify which occurrences of {\term} have to be substituted in the context. The {\tt in {\occgoalset}} clause is an -occurrence clause whose syntax and behavior is described in +occurrence clause whose syntax and behavior are described in Section~\ref{Occurrences clauses}. -\item {\tt set (} {\ident} \nelist{\binder}{} {\tt :=} {\term} {\tt )} +\item {\tt set ( {\ident} \nelist{\binder}{} := {\term} )} - This is equivalent to {\tt set (} {\ident} {\tt :=} {\tt fun} - \nelist{\binder}{} {\tt =>} {\term} {\tt )}. + This is equivalent to {\tt set ( {\ident} := fun + \nelist{\binder}{} => {\term} )}. -\item {\tt set } {\term} +\item {\tt set \term} This behaves as {\tt set (} {\ident} := {\term} {\tt )} but {\ident} is generated by {\Coq}. This variant also supports an occurrence clause. -\item {\tt set (} {\ident$_0$} \nelist{\binder}{} {\tt :=} {\term} - {\tt ) in {\occgoalset}}\\ +\item {\tt set ( {\ident$_0$} \nelist{\binder}{} := {\term} ) in {\occgoalset}}\\ {\tt set {\term} in {\occgoalset}} These are the general forms which combine the previous possibilities. -\item {\tt remember {\term} {\tt as} {\ident}} +\item {\tt remember {\term} as {\ident}}\tacindex{remember} - This behaves as {\tt set (} {\ident} := {\term} {\tt ) in *} and using a + This behaves as {\tt set ( {\ident} := {\term} ) in *} and using a logical (Leibniz's) equality instead of a local definition. -\item {\tt remember {\term} {\tt as} {\ident} in {\occgoalset}} +\item {\tt remember {\term} as {\ident} in {\occgoalset}} This is a more general form of {\tt remember} that remembers the occurrences of {\term} specified by an occurrences set. -\item {\tt pose ( {\ident} := {\term} )} - +\item {\tt pose ( {\ident} := {\term} )}\tacindex{pose} + This adds the local definition {\ident} := {\term} to the current context without performing any replacement in the goal or in the hypotheses. It is equivalent to {\tt set ( {\ident} {\tt :=} {\term} {\tt ) in |-}}. -\item {\tt pose (} {\ident} \nelist{\binder}{} {\tt :=} {\term} {\tt )} +\item {\tt pose ( {\ident} \nelist{\binder}{} := {\term} )} This is equivalent to {\tt pose (} {\ident} {\tt :=} {\tt fun} \nelist{\binder}{} {\tt =>} {\term} {\tt )}. \item{\tt pose {\term}} - This behaves as {\tt pose (} {\ident} := {\term} {\tt )} but + This behaves as {\tt pose ( {\ident} := {\term} )} but {\ident} is generated by {\Coq}. \end{Variants} -\subsection{{\tt assert ( {\ident} : {\form} \tt )} -\tacindex{assert}} - -This tactic applies to any goal. {\tt assert (H : U)} adds a new -hypothesis of name \texttt{H} asserting \texttt{U} to the current goal -and opens a new subgoal \texttt{U}\footnote{This corresponds to the - cut rule of sequent calculus.}. The subgoal {\texttt U} comes first -in the list of subgoals remaining to prove. - -\begin{ErrMsgs} -\item \errindex{Not a proposition or a type} - - Arises when the argument {\form} is neither of type {\tt Prop}, {\tt - Set} nor {\tt Type}. +\subsection{\tt decompose [ {\qualid$_1$} \dots\ {\qualid$_n$} ] \term} +\label{decompose} +\tacindex{decompose} -\end{ErrMsgs} +This tactic allows to recursively decompose a +complex proposition in order to obtain atomic ones. -\begin{Variants} +\Example + +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{coq_example} +Goal forall A B C:Prop, A /\ B /\ C \/ B /\ C \/ C /\ A -> C. +intros A B C H; decompose [and or] H; assumption. +\end{coq_example} +\begin{coq_example*} +Qed. +\end{coq_example*} + +{\tt decompose} does not work on right-hand sides of implications or products. + +\begin{Variants} + +\item {\tt decompose sum \term}\tacindex{decompose sum} + + This decomposes sum types (like \texttt{or}). + +\item {\tt decompose record \term}\tacindex{decompose record} + + This decomposes record types (inductive types with one constructor, + like \texttt{and} and \texttt{exists} and those defined with the + \texttt{Record} macro, see Section~\ref{Record}). + +\end{Variants} + +\section{Controlling the proof flow} + +\subsection{\tt assert ( {\ident} :\ {\form} )} +\tacindex{assert} + +This tactic applies to any goal. {\tt assert (H : U)} adds a new +hypothesis of name \texttt{H} asserting \texttt{U} to the current goal +and opens a new subgoal \texttt{U}\footnote{This corresponds to the + cut rule of sequent calculus.}. The subgoal {\texttt U} comes first +in the list of subgoals remaining to prove. + +\begin{ErrMsgs} +\item \errindex{Not a proposition or a type} + + Arises when the argument {\form} is neither of type {\tt Prop}, {\tt + Set} nor {\tt Type}. + +\end{ErrMsgs} + +\begin{Variants} \item{\tt assert {\form}} - - This behaves as {\tt assert (} {\ident} : {\form} {\tt )} but + + This behaves as {\tt assert ( {\ident} :\ {\form} )} but {\ident} is generated by {\Coq}. -\item{\tt assert (} {\ident} := {\term} {\tt )} - - This behaves as {\tt assert ({\ident} : {\type});[exact +\item{\tt assert ( {\ident} := {\term} )} + + This behaves as {\tt assert ({\ident} :\ {\type});[exact {\term}|idtac]} where {\type} is the type of {\term}. -\item {\tt cut {\form}}\tacindex{cut} - - This tactic applies to any goal. It implements the non dependent + \ErrMsg \errindex{Variable {\ident} is already declared} + +\item {\tt cut {\form}}\tacindex{cut} + + This tactic applies to any goal. It implements the non-dependent case of the ``App''\index{Typing rules!App} rule given in Section~\ref{Typed-terms}. (This is Modus Ponens inference rule.) {\tt cut U} transforms the current goal \texttt{T} into the two @@ -689,16 +1159,18 @@ in the list of subgoals remaining to prove. -> T} comes first in the list of remaining subgoal to prove. \item \texttt{assert {\form} by {\tac}}\tacindex{assert by} - + This tactic behaves like \texttt{assert} but applies {\tac} to solve the subgoals generated by \texttt{assert}. + \ErrMsg \errindex{Proof is not complete} + \item \texttt{assert {\form} as {\intropattern}\tacindex{assert as}} If {\intropattern} is a naming introduction pattern (see Section~\ref{intros-pattern}), the hypothesis is named after this introduction pattern (in particular, if {\intropattern} is {\ident}, - the tactic behaves like \texttt{assert ({\ident} : {\form})}). + the tactic behaves like \texttt{assert ({\ident} :\ {\form})}). If {\intropattern} is a disjunctive/conjunctive introduction pattern, the tactic behaves like \texttt{assert {\form}} then destructing the @@ -714,12 +1186,11 @@ in the list of subgoals remaining to prove. exact {\term}} where \texttt{T} is the type of {\term}. In particular, \texttt{pose proof {\term} as {\ident}} behaves as - \texttt{assert ({\ident}:T) by exact {\term}} (where \texttt{T} is - the type of {\term}) and \texttt{pose proof {\term} as + \texttt{assert ({\ident} := {\term})} and \texttt{pose proof {\term} as {\disjconjintropattern}\tacindex{pose proof}} behaves like \texttt{destruct {\term} as {\disjconjintropattern}}. -\item {\tt specialize ({\ident} \term$_1$ {\ldots} \term$_n$)\tacindex{specialize}} \\ +\item {\tt specialize ({\ident} \term$_1$ \dots\ \term$_n$)\tacindex{specialize}} \\ {\tt specialize {\ident} with \bindinglist} The tactic {\tt specialize} works on local hypothesis \ident. @@ -730,30 +1201,37 @@ in the list of subgoals remaining to prove. Section~\ref{Binding-list} for more about bindings lists). In the second form, all instantiation elements must be given, whereas in the first form the application to \term$_1$ {\ldots} - \term$_n$ can be partial. The first form is equivalent to - {\tt assert (\ident':=\ident \term$_1$ {\ldots} \term$_n$); - clear \ident; rename \ident' into \ident}. + \term$_n$ can be partial. The first form is equivalent to + {\tt assert (\ident' := {\ident} {\term$_1$} \dots\ \term$_n$); + clear \ident; rename \ident' into \ident}. The name {\ident} can also refer to a global lemma or hypothesis. In this case, for compatibility reasons, the behavior of {\tt specialize} is close to that of {\tt - generalize}: the instantiated statement becomes an additional - premise of the goal. + generalize}: the instantiated statement becomes an additional + premise of the goal. + + \begin{ErrMsgs} + \item \errindexbis{{\ident} is used in hypothesis \ident'}{is used in hypothesis} + \item \errindexbis{{\ident} is used in conclusion}{is used in conclusion} + \end{ErrMsgs} -%% Moreover, the old syntax allows the use of a number after {\tt specialize} -%% for controlling the number of premises to instantiate. Giving this +%% Moreover, the old syntax allows the use of a number after {\tt specialize} +%% for controlling the number of premises to instantiate. Giving this %% number should not be mandatory anymore (automatic detection of how -%% many premises can be eaten without leaving meta-variables). Hence +%% many premises can be eaten without leaving meta-variables). Hence %% no documentation for this integer optional argument of specialize \end{Variants} -\subsection{\tt generalize \term +\subsection{\tt generalize \term} \tacindex{generalize} -\label{generalize}} +\label{generalize} + +This tactic applies to any goal. It generalizes the conclusion with +respect to one of its subterms. -This tactic applies to any goal. It generalizes the conclusion w.r.t. -one subterm of it. For example: +\Example \begin{coq_eval} Goal forall x y:nat, (0 <= x + y + y). @@ -775,185 +1253,49 @@ where $G'$ is obtained from $G$ by replacing all occurrences of $t$ by \begin{Variants} \item {\tt generalize {\term$_1$ , \dots\ , \term$_n$}} - - Is equivalent to {\tt generalize \term$_n$; \dots\ ; generalize + + This is equivalent to {\tt generalize \term$_n$; \dots\ ; generalize \term$_1$}. Note that the sequence of \term$_i$'s are processed from $n$ to $1$. \item {\tt generalize {\term} at {\num$_1$ \dots\ \num$_i$}} - - Is equivalent to {\tt generalize \term} but generalizing only over + + This is equivalent to {\tt generalize \term} but it generalizes only over the specified occurrences of {\term} (counting from left to right on the expression printed using option {\tt Set Printing All}). \item {\tt generalize {\term} as {\ident}} - - Is equivalent to {\tt generalize \term} but use {\ident} to name the + + This is equivalent to {\tt generalize \term} but it uses {\ident} to name the generalized hypothesis. -\item {\tt generalize {\term$_1$} at {\num$_{11}$ \dots\ \num$_{1i_1}$} +\item {\tt generalize {\term$_1$} at {\num$_{11}$ \dots\ \num$_{1i_1}$} as {\ident$_1$} , {\ldots} , {\term$_n$} at {\num$_{n1}$ \dots\ \num$_{ni_n}$} as {\ident$_2$}} - + This is the most general form of {\tt generalize} that combines the previous behaviors. - -\item {\tt generalize dependent \term} \tacindex{generalize dependent} - - This generalizes {\term} but also {\em all} hypotheses which depend - on {\term}. It clears the generalized hypotheses. - -\end{Variants} - - -\subsection{\tt revert \ident$_1$ \dots\ \ident$_n$ -\tacindex{revert} -\label{revert}} - -This applies to any goal with variables \ident$_1$ \dots\ \ident$_n$. -It moves the hypotheses (possibly defined) to the goal, if this respects -dependencies. This tactic is the inverse of {\tt intro}. - -\begin{ErrMsgs} -\item \errindexbis{{\ident} is used in the hypothesis {\ident'}}{is - used in the hypothesis} -\end{ErrMsgs} - -\begin{Variants} -\item {\tt revert dependent \ident \tacindex{revert dependent}} - - This moves to the goal the hypothesis \ident\ and all hypotheses - which depend on it. - -\end{Variants} - -\subsection{\tt change \term -\tacindex{change} -\label{change}} - -This tactic applies to any goal. It implements the rule -``Conv''\index{Typing rules!Conv} given in Section~\ref{Conv}. {\tt - change U} replaces the current goal \T\ with \U\ providing that -\U\ is well-formed and that \T\ and \U\ are convertible. - -\begin{ErrMsgs} -\item \errindex{Not convertible} -\end{ErrMsgs} - -\tacindex{change \dots\ in} -\begin{Variants} -\item {\tt change \term$_1$ with \term$_2$} - - This replaces the occurrences of \term$_1$ by \term$_2$ in the - current goal. The terms \term$_1$ and \term$_2$ must be - convertible. - -\item {\tt change \term$_1$ at \num$_1$ \dots\ \num$_i$ with \term$_2$} - - This replaces the occurrences numbered \num$_1$ \dots\ \num$_i$ of - \term$_1$ by \term$_2$ in the current goal. - The terms \term$_1$ and \term$_2$ must be convertible. - - \ErrMsg {\tt Too few occurrences} - -\item {\tt change {\term} in {\ident}} - -\item {\tt change \term$_1$ with \term$_2$ in {\ident}} - -\item {\tt change \term$_1$ at \num$_1$ \dots\ \num$_i$ with \term$_2$ in - {\ident}} - - This applies the {\tt change} tactic not to the goal but to the - hypothesis {\ident}. - -\end{Variants} - -\SeeAlso \ref{Conversion-tactics} - -\subsection{\tt fix {\ident} {\num} -\tacindex{fix} -\label{tactic:fix}} - -This tactic is a primitive tactic to start a proof by induction. In -general, it is easier to rely on higher-level induction tactics such -as the ones described in Section~\ref{Tac-induction}. - -In the syntax of the tactic, the identifier {\ident} is the name given -to the induction hypothesis. The natural number {\num} tells on which -premise of the current goal the induction acts, starting -from 1 and counting both dependent and non dependent -products. Especially, the current lemma must be composed of at least -{\num} products. - -Like in a {\tt fix} expression, the induction -hypotheses have to be used on structurally smaller arguments. -The verification that inductive proof arguments are correct is done -only at the time of registering the lemma in the environment. To know -if the use of induction hypotheses is correct at some -time of the interactive development of a proof, use the command {\tt - Guarded} (see Section~\ref{Guarded}). - -\begin{Variants} - \item {\tt fix} {\ident}$_1$ {\num} {\tt with (} {\ident}$_2$ - \nelist{{\binder}$_{2}$}{} \zeroone{{\tt \{ struct {\ident$'_2$} - \}}} {\tt :} {\type}$_2$ {\tt )} {\ldots} {\tt (} {\ident}$_1$ - \nelist{{\binder}$_n$}{} \zeroone{{\tt \{ struct {\ident$'_n$} \}}} - {\tt :} {\type}$_n$ {\tt )} - -This starts a proof by mutual induction. The statements to be -simultaneously proved are respectively {\tt forall} - \nelist{{\binder}$_2$}{}{\tt ,} {\type}$_2$, {\ldots}, {\tt forall} - \nelist{{\binder}$_n$}{}{\tt ,} {\type}$_n$. The identifiers -{\ident}$_1$ {\ldots} {\ident}$_n$ are the names of the induction -hypotheses. The identifiers {\ident}$'_2$ {\ldots} {\ident}$'_n$ are the -respective names of the premises on which the induction is performed -in the statements to be simultaneously proved (if not given, the -system tries to guess itself what they are). - -\end{Variants} - -\subsection{\tt cofix {\ident} -\tacindex{cofix} -\label{tactic:cofix}} - -This tactic starts a proof by coinduction. The identifier {\ident} is -the name given to the coinduction hypothesis. Like in a {\tt cofix} -expression, the use of induction hypotheses have to guarded by a -constructor. The verification that the use of coinductive hypotheses -is correct is done only at the time of registering the lemma in the -environment. To know if the use of coinduction hypotheses is correct -at some time of the interactive development of a proof, use the -command {\tt Guarded} (see Section~\ref{Guarded}). - -\begin{Variants} - \item {\tt cofix} {\ident}$_1$ {\tt with (} {\ident}$_2$ - \nelist{{\binder}$_2$}{} {\tt :} {\type}$_2$ {\tt )} {\ldots} {\tt - (} {\ident}$_1$ \nelist{{\binder}$_1$}{} {\tt :} {\type}$_n$ - {\tt )} +\item {\tt generalize dependent \term} \tacindex{generalize dependent} -This starts a proof by mutual coinduction. The statements to be -simultaneously proved are respectively {\tt forall} -\nelist{{\binder}$_2$}{}{\tt ,} {\type}$_2$, {\ldots}, {\tt forall} - \nelist{{\binder}$_n$}{}{\tt ,} {\type}$_n$. The identifiers - {\ident}$_1$ {\ldots} {\ident}$_n$ are the names of the - coinduction hypotheses. + This generalizes {\term} but also {\em all} hypotheses that depend + on {\term}. It clears the generalized hypotheses. \end{Variants} -\subsection{\tt evar (\ident:\term) +\subsection{\tt evar ( {\ident} :\ {\term} )} \tacindex{evar} -\label{evar}} +\label{evar} The {\tt evar} tactic creates a new local definition named \ident\ with type \term\ in the context. The body of this binding is a fresh existential variable. -\subsection{\tt instantiate (\num:= \term) +\subsection{\tt instantiate ( {\num} := {\term} )} \tacindex{instantiate} -\label{instantiate}} +\label{instantiate} The {\tt instantiate} tactic allows to refine (see Section~\ref{refine}) an existential variable @@ -963,14 +1305,14 @@ the number of the existential variable since this number is different in every session. \begin{Variants} - \item {\tt instantiate (\num:=\term) in \ident} - - \item {\tt instantiate (\num:=\term) in (Value of \ident)} - - \item {\tt instantiate (\num:=\term) in (Type of \ident)} + \item {\tt instantiate ( {\num} := {\term} ) in \ident} + + \item {\tt instantiate ( {\num} := {\term} ) in ( Value of {\ident} )} -These allow to refer respectively to existential variables occurring in -a hypothesis or in the body or the type of a local definition. + \item {\tt instantiate ( {\num} := {\term} ) in ( Type of {\ident} )} + +These allow to refer respectively to existential variables occurring in +a hypothesis or in the body or the type of a local definition. \item {\tt instantiate} @@ -983,9 +1325,9 @@ a hypothesis or in the body or the type of a local definition. \end{Variants} -\subsection{\tt admit +\subsection{\tt admit} \tacindex{admit} -\label{admit}} +\label{admit} The {\tt admit} tactic ``solves'' the current subgoal by an axiom. This typically allows to temporarily skip a subgoal so as to @@ -995,1422 +1337,1165 @@ Assumptions} (see Section~\ref{PrintAssumptions}). Admitted subgoals have names of the form {\ident}\texttt{\_admitted} possibly followed by a number. -\subsection{\tt constr\_eq \term$_1$ \term$_2$ -\tacindex{constr\_eq} -\label{constreq}} +\subsection{\tt absurd \term} +\tacindex{absurd} +\label{absurd} -This tactic applies to any goal. It checks whether its arguments are -equal modulo alpha conversion and casts. +This tactic applies to any goal. The argument {\term} is any +proposition {\tt P} of type {\tt Prop}. This tactic applies {\tt + False} elimination, that is it deduces the current goal from {\tt + False}, and generates as subgoals {\tt $\sim$P} and {\tt P}. It is +very useful in proofs by cases, where some cases are impossible. In +most cases, \texttt{P} or $\sim$\texttt{P} is one of the hypotheses of +the local context. -\ErrMsg \errindex{Not equal} +\subsection{\tt contradiction} +\label{contradiction} +\tacindex{contradiction} -\subsection{\tt is\_evar \term -\tacindex{is\_evar} -\label{isevar}} +This tactic applies to any goal. The {\tt contradiction} tactic +attempts to find in the current context (after all {\tt intros}) one +hypothesis that is equivalent to {\tt False}. It permits to prune +irrelevant cases. This tactic is a macro for the tactics sequence +{\tt intros; elimtype False; assumption}. -This tactic applies to any goal. It checks whether its argument is an -existential variable. Existential variables are uninstantiated -variables generated by e.g. {\tt eapply} (see Section~\ref{apply}). +\begin{ErrMsgs} +\item \errindex{No such assumption} +\end{ErrMsgs} -\ErrMsg \errindex{Not an evar} +\begin{Variants} +\item {\tt contradiction \ident} -\subsection{\tt has\_evar \term -\tacindex{has\_evar} -\label{hasevar}} +The proof of {\tt False} is searched in the hypothesis named \ident. +\end{Variants} -This tactic applies to any goal. It checks whether its argument has an -existential variable as a subterm. Unlike {\tt context} patterns -combined with {\tt is\_evar}, this tactic scans all subterms, -including those under binders. +\subsection{\tt contradict \ident} +\label{contradict} +\tacindex{contradict} -\ErrMsg \errindex{No evars} +This tactic allows to manipulate negated hypothesis and goals. The +name \ident\ should correspond to a hypothesis. With +{\tt contradict H}, the current goal and context is transformed in +the following way: +\begin{itemize} +\item {\tt H:$\neg$A $\vd$ B} \ becomes \ {\tt $\vd$ A} +\item {\tt H:$\neg$A $\vd$ $\neg$B} \ becomes \ {\tt H: B $\vd$ A } +\item {\tt H: A $\vd$ B} \ becomes \ {\tt $\vd$ $\neg$A} +\item {\tt H: A $\vd$ $\neg$B} \ becomes \ {\tt H: B $\vd$ $\neg$A} +\end{itemize} -\subsection{\tt is\_var \term -\tacindex{is\_var} -\label{isvar}} +\subsection{\tt exfalso} +\label{exfalso} +\tacindex{exfalso} -This tactic applies to any goal. It checks whether its argument is a -variable or hypothesis in the current goal context or in the opened sections. +This tactic implements the ``ex falso quodlibet'' logical principle: +an elimination of {\tt False} is performed on the current goal, and the +user is then required to prove that {\tt False} is indeed provable in +the current context. This tactic is a macro for {\tt elimtype False}. -\ErrMsg \errindex{Not a variable or hypothesis} +\section{Case analysis and induction} -\subsection{Bindings list -\index{Binding list} -\label{Binding-list}} +The tactics presented in this section implement induction or case +analysis on inductive or co-inductive objects (see +Section~\ref{Cic-inductive-definitions}). -Tactics that take a term as argument may also support a bindings list, so -as to instantiate some parameters of the term by name or position. -The general form of a term equipped with a bindings list is {\tt -{\term} with {\bindinglist}} where {\bindinglist} may be of two -different forms: +\subsection{\tt destruct \term} +\tacindex{destruct} +\label{destruct} -\begin{itemize} -\item In a bindings list of the form {\tt (\vref$_1$ := \term$_1$) - \dots\ (\vref$_n$ := \term$_n$)}, {\vref} is either an {\ident} or a - {\num}. The references are determined according to the type of - {\term}. If \vref$_i$ is an identifier, this identifier has to be - bound in the type of {\term} and the binding provides the tactic - with an instance for the parameter of this name. If \vref$_i$ is - some number $n$, this number denotes the $n$-th non dependent - premise of the {\term}, as determined by the type of {\term}. +This tactic applies to any goal. The argument {\term} must be of +inductive or co-inductive type and the tactic generates subgoals, one +for each possible form of {\term}, i.e. one for each constructor of +the inductive or co-inductive type. Unlike {\tt induction}, no +induction hypothesis is generated by {\tt destruct}. - \ErrMsg \errindex{No such binder} +If the argument is dependent in either the conclusion or some +hypotheses of the goal, the argument is replaced by the appropriate +constructor form in each of the resulting subgoals, thus performing +case analysis. If non-dependent, the tactic simply exposes the +inductive or co-inductive structure of the argument. -\item A bindings list can also be a simple list of terms {\tt - \term$_1$ \dots\term$_n$}. In that case the references to - which these terms correspond are determined by the tactic. In case - of {\tt induction}, {\tt destruct}, {\tt elim} and {\tt case} (see - Section~\ref{elim}) the terms have to provide instances for all the - dependent products in the type of \term\ while in the case of {\tt - apply}, or of {\tt constructor} and its variants, only instances for - the dependent products which are not bound in the conclusion of the - type are required. +There are special cases: - \ErrMsg \errindex{Not the right number of missing arguments} -\end{itemize} +\begin{itemize} -\subsection{Occurrences sets and occurrences clauses} -\label{Occurrences clauses} -\index{Occurrences clauses} +\item If {\term} is an identifier {\ident} denoting a quantified +variable of the conclusion of the goal, then {\tt destruct {\ident}} +behaves as {\tt intros until {\ident}; destruct {\ident}}. -An occurrences clause is a modifier to some tactics that obeys the -following syntax: +\item If {\term} is a {\num}, then {\tt destruct {\num}} behaves as +{\tt intros until {\num}} followed by {\tt destruct} applied to the +last introduced hypothesis. Remark: For destruction of a numeral, use +syntax {\tt destruct ({\num})} (not very interesting anyway). -$\!\!\!$\begin{tabular}{lcl} -{\occclause} & ::= & {\tt in} {\occgoalset} \\ -{\occgoalset} & ::= & - \zeroone{{\ident$_1$} \zeroone{\atoccurrences} {\tt ,} \\ -& & {\dots} {\tt ,}\\ -& & {\ident$_m$} \zeroone{\atoccurrences}}\\ -& & \zeroone{{\tt |-} \zeroone{{\tt *} \zeroone{\atoccurrences}}}\\ -& | & - {\tt *} {\tt |-} \zeroone{{\tt *} \zeroone{\atoccurrences}}\\ -& | & - {\tt *}\\ -{\atoccurrences} & ::= & {\tt at} {\occlist}\\ -{\occlist} & ::= & \zeroone{{\tt -}} {\num$_1$} \dots\ {\num$_n$} -\end{tabular} +\item The argument {\term} can also be a pattern of which holes are + denoted by ``\_''. In this case, the tactic checks that all subterms + matching the pattern in the conclusion and the hypotheses are + compatible and performs case analysis using this subterm. -The role of an occurrence clause is to select a set of occurrences of -a {\term} in a goal. In the first case, the {{\ident$_i$} -\zeroone{{\tt at} {\num$_1^i$} \dots\ {\num$_{n_i}^i$}}} parts -indicate that occurrences have to be selected in the hypotheses named -{\ident$_i$}. If no numbers are given for hypothesis {\ident$_i$}, -then all occurrences of {\term} in the hypothesis are selected. If -numbers are given, they refer to occurrences of {\term} when the term -is printed using option {\tt Set Printing All} (see -Section~\ref{SetPrintingAll}), counting from left to right. In -particular, occurrences of {\term} in implicit arguments (see -Section~\ref{Implicit Arguments}) or coercions (see -Section~\ref{Coercions}) are counted. +\end{itemize} -If a minus sign is given between {\tt at} and the list of occurrences, -it negates the condition so that the clause denotes all the occurrences except -the ones explicitly mentioned after the minus sign. +\begin{Variants} +\item{\tt destruct {\term} as {\disjconjintropattern}} -As an exception to the left-to-right order, the occurrences in the -{\tt return} subexpression of a {\tt match} are considered {\em -before} the occurrences in the matched term. + This behaves as {\tt destruct {\term}} but uses the names in + {\intropattern} to name the variables introduced in the context. + The {\intropattern} must have the form {\tt [} $p_{11}$ \ldots + $p_{1n_1}$ {\tt |} {\ldots} {\tt |} $p_{m1}$ \ldots $p_{mn_m}$ {\tt + ]} with $m$ being the number of constructors of the type of + {\term}. Each variable introduced by {\tt destruct} in the context + of the $i^{th}$ goal gets its name from the list $p_{i1}$ \ldots + $p_{in_i}$ in order. If there are not enough names, {\tt destruct} + invents names for the remaining variables to introduce. More + generally, the $p_{ij}$ can be any disjunctive/conjunctive + introduction pattern (see Section~\ref{intros-pattern}). This + provides a concise notation for nested destruction. -In the second case, the {\tt *} on the left of {\tt |-} means that -all occurrences of {\term} are selected in every hypothesis. +% It is recommended to use this variant of {\tt destruct} for +% robust proof scripts. -In the first and second case, if {\tt *} is mentioned on the right of -{\tt |-}, the occurrences of the conclusion of the goal have to be -selected. If some numbers are given, then only the occurrences denoted -by these numbers are selected. In no numbers are given, all -occurrences of {\term} in the goal are selected. +\item{\tt destruct {\term} as {\disjconjintropattern} \_eqn} -Finally, the last notation is an abbreviation for {\tt * |- *}. Note -also that {\tt |-} is optional in the first case when no {\tt *} is -given. + This behaves as {\tt destruct {\term}} but adds an equation between + {\term} and the value that {\term} takes in each of the possible + cases. The name of the equation is chosen by Coq. If + {\disjconjintropattern} is simply {\tt []}, it is automatically considered + as a disjunctive pattern of the appropriate size. -Here are some tactics that understand occurrences clauses: -{\tt set}, {\tt remember}, {\tt induction}, {\tt destruct}. +\item{\tt destruct {\term} as {\disjconjintropattern} \_eqn:~{\namingintropattern}} -\SeeAlso~Sections~\ref{tactic:set}, \ref{Tac-induction}, \ref{SetPrintingAll}. + This behaves as {\tt destruct {\term} as + {\disjconjintropattern} \_eqn} but use {\namingintropattern} to + name the equation (see Section~\ref{intros-pattern}). Note that spaces + can generally be removed around {\tt \_eqn}. +\item{\tt destruct {\term} with \bindinglist} -\section{Negation and contradiction} + This behaves like \texttt{destruct {\term}} providing explicit + instances for the dependent premises of the type of {\term} (see + syntax of bindings in Section~\ref{Binding-list}). -\subsection{\tt absurd \term -\tacindex{absurd} -\label{absurd}} +\item{\tt edestruct {\term}\tacindex{edestruct}} -This tactic applies to any goal. The argument {\term} is any -proposition {\tt P} of type {\tt Prop}. This tactic applies {\tt - False} elimination, that is it deduces the current goal from {\tt - False}, and generates as subgoals {\tt $\sim$P} and {\tt P}. It is -very useful in proofs by cases, where some cases are impossible. In -most cases, \texttt{P} or $\sim$\texttt{P} is one of the hypotheses of -the local context. + This tactic behaves like \texttt{destruct {\term}} except that it + does not fail if the instance of a dependent premises of the type of + {\term} is not inferable. Instead, the unresolved instances are left + as existential variables to be inferred later, in the same way as + {\tt eapply} does (see Section~\ref{eapply-example}). -\subsection{\tt contradiction -\label{contradiction} -\tacindex{contradiction}} +\item{\tt destruct {\term$_1$} using {\term$_2$}}\\ + {\tt destruct {\term$_1$} using {\term$_2$} with {\bindinglist}} -This tactic applies to any goal. The {\tt contradiction} tactic -attempts to find in the current context (after all {\tt intros}) one -hypothesis which is equivalent to {\tt False}. It permits to prune -irrelevant cases. This tactic is a macro for the tactics sequence -{\tt intros; elimtype False; assumption}. + These are synonyms of {\tt induction {\term$_1$} using {\term$_2$}} and + {\tt induction {\term$_1$} using {\term$_2$} with {\bindinglist}}. -\begin{ErrMsgs} -\item \errindex{No such assumption} -\end{ErrMsgs} +\item \texttt{destruct {\term} in {\occgoalset}} -\begin{Variants} -\item {\tt contradiction \ident} + This syntax is used for selecting which occurrences of {\term} the + case analysis has to be done on. The {\tt in {\occgoalset}} clause is an + occurrence clause whose syntax and behavior is described in + Section~\ref{Occurrences clauses}. -The proof of {\tt False} is searched in the hypothesis named \ident. -\end{Variants} +% When an occurrence clause is given, an equation between {\term} and +% the value it gets in each case of the analysis is added to the +% context of the subgoals corresponding to the cases (even +% if no clause {\tt as {\namingintropattern}} is given). -\subsection {\tt contradict \ident} -\label{contradict} -\tacindex{contradict} +\item{\tt destruct {\term$_1$} with {\bindinglist$_1$} as {\disjconjintropattern} \_eqn:~{\namingintropattern} using {\term$_2$} with {\bindinglist$_2$} in {\occgoalset}}\\ + {\tt edestruct {\term$_1$} with {\bindinglist$_1$} as {\disjconjintropattern} \_eqn:~{\namingintropattern} using {\term$_2$} with {\bindinglist$_2$} in {\occgoalset}} -This tactic allows to manipulate negated hypothesis and goals. The -name \ident\ should correspond to a hypothesis. With -{\tt contradict H}, the current goal and context is transformed in -the following way: -\begin{itemize} -\item {\tt H:$\neg$A $\vd$ B} \ becomes \ {\tt $\vd$ A} -\item {\tt H:$\neg$A $\vd$ $\neg$B} \ becomes \ {\tt H: B $\vd$ A } -\item {\tt H: A $\vd$ B} \ becomes \ {\tt $\vd$ $\neg$A} -\item {\tt H: A $\vd$ $\neg$B} \ becomes \ {\tt H: B $\vd$ $\neg$A} -\end{itemize} + These are the general forms of {\tt destruct} and {\tt edestruct}. + They combine the effects of the {\tt with}, {\tt as}, {\tt using}, + and {\tt in} clauses. -\subsection{\tt exfalso} -\label{exfalso} -\tacindex{exfalso} +\item{\tt case \term}\label{case}\tacindex{case} -This tactic implements the ``ex falso quodlibet'' logical principle: -an elimination of {\tt False} is performed on the current goal, and the -user is then required to prove that {\tt False} is indeed provable in -the current context. This tactic is a macro for {\tt elimtype False}. + The tactic {\tt case} is a more basic tactic to perform case + analysis without recursion. It behaves as {\tt elim \term} but using + a case-analysis elimination principle and not a recursive one. -\section{Conversion tactics -\index{Conversion tactics} -\label{Conversion-tactics}} +\item {\tt case {\term} with {\bindinglist}} -This set of tactics implements different specialized usages of the -tactic \texttt{change}. + Analogous to {\tt elim {\term} with {\bindinglist}} above. -All conversion tactics (including \texttt{change}) can be -parameterized by the parts of the goal where the conversion can -occur. This is done using \emph{goal clauses} which consists in a list -of hypotheses and, optionally, of a reference to the conclusion of the -goal. For defined hypothesis it is possible to specify if the -conversion should occur on the type part, the body part or both -(default). +\item{\tt ecase {\term}\tacindex{ecase}}\\ + {\tt ecase {\term} with {\bindinglist}} -\index{Clauses} -\index{Goal clauses} -Goal clauses are written after a conversion tactic (tactics -\texttt{set}~\ref{tactic:set}, \texttt{rewrite}~\ref{rewrite}, -\texttt{replace}~\ref{tactic:replace} and -\texttt{autorewrite}~\ref{tactic:autorewrite} also use goal clauses) and -are introduced by the keyword \texttt{in}. If no goal clause is provided, -the default is to perform the conversion only in the conclusion. + In case the type of {\term} has dependent premises, or dependent + premises whose values are not inferable from the {\tt with + {\bindinglist}} clause, {\tt ecase} turns them into existential + variables to be resolved later on. -The syntax and description of the various goal clauses is the following: -\begin{description} -\item[]\texttt{in {\ident}$_1$ $\ldots$ {\ident}$_n$ |- } only in hypotheses {\ident}$_1$ - \ldots {\ident}$_n$ -\item[]\texttt{in {\ident}$_1$ $\ldots$ {\ident}$_n$ |- *} in hypotheses {\ident}$_1$ \ldots - {\ident}$_n$ and in the conclusion -\item[]\texttt{in * |-} in every hypothesis -\item[]\texttt{in *} (equivalent to \texttt{in * |- *}) everywhere -\item[]\texttt{in (type of {\ident}$_1$) (value of {\ident}$_2$) $\ldots$ |-} in - type part of {\ident}$_1$, in the value part of {\ident}$_2$, etc. -\end{description} +\item {\tt simple destruct \ident}\tacindex{simple destruct} -For backward compatibility, the notation \texttt{in}~{\ident}$_1$\ldots {\ident}$_n$ -performs the conversion in hypotheses {\ident}$_1$\ldots {\ident}$_n$. + This tactic behaves as {\tt intros until + {\ident}; case {\tt {\ident}}} when {\ident} is a quantified + variable of the goal. -%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% -%voir reduction__conv_x : histoires d'univers. -%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +\item {\tt simple destruct {\num}} -\subsection[{\tt cbv \flag$_1$ \dots\ \flag$_n$}, {\tt lazy \flag$_1$ -\dots\ \flag$_n$} and {\tt compute}] -{{\tt cbv \flag$_1$ \dots\ \flag$_n$}, {\tt lazy \flag$_1$ -\dots\ \flag$_n$} and {\tt compute} -\tacindex{cbv} -\tacindex{lazy} -\tacindex{compute} -\tacindex{vm\_compute}\label{vmcompute}} + This tactic behaves as {\tt intros until + {\num}; case {\tt {\ident}}} where {\ident} is the name given by + {\tt intros until {\num}} to the {\num}-th non-dependent premise of + the goal. -These parameterized reduction tactics apply to any goal and perform -the normalization of the goal according to the specified flags. In -correspondence with the kinds of reduction considered in \Coq\, namely -$\beta$ (reduction of functional application), $\delta$ (unfolding of -transparent constants, see \ref{Transparent}), $\iota$ (reduction of -pattern-matching over a constructed term, and unfolding of {\tt fix} -and {\tt cofix} expressions) and $\zeta$ (contraction of local -definitions), the flag are either {\tt beta}, {\tt delta}, {\tt iota} -or {\tt zeta}. The {\tt delta} flag itself can be refined into {\tt -delta [\qualid$_1$\ldots\qualid$_k$]} or {\tt delta --[\qualid$_1$\ldots\qualid$_k$]}, restricting in the first case the -constants to unfold to the constants listed, and restricting in the -second case the constant to unfold to all but the ones explicitly -mentioned. Notice that the {\tt delta} flag does not apply to -variables bound by a let-in construction inside the term itself (use -here the {\tt zeta} flag). In any cases, opaque constants are not -unfolded (see Section~\ref{Opaque}). +\item{\tt case\_eq \term}\label{case_eq}\tacindex{case\_eq} -The goal may be normalized with two strategies: {\em lazy} ({\tt lazy} -tactic), or {\em call-by-value} ({\tt cbv} tactic). The lazy strategy -is a call-by-need strategy, with sharing of reductions: the arguments of a -function call are partially evaluated only when necessary, and if an -argument is used several times then it is computed only once. This -reduction is efficient for reducing expressions with dead code. For -instance, the proofs of a proposition {\tt exists~$x$. $P(x)$} reduce to a -pair of a witness $t$, and a proof that $t$ satisfies the predicate -$P$. Most of the time, $t$ may be computed without computing the proof -of $P(t)$, thanks to the lazy strategy. + The tactic {\tt case\_eq} is a variant of the {\tt case} tactic that + allow to perform case analysis on a term without completely + forgetting its original form. This is done by generating equalities + between the original form of the term and the outcomes of the case + analysis. -The call-by-value strategy is the one used in ML languages: the -arguments of a function call are evaluated first, using a weak -reduction (no reduction under the $\lambda$-abstractions). Despite the -lazy strategy always performs fewer reductions than the call-by-value -strategy, the latter is generally more efficient for evaluating purely -computational expressions (i.e. with few dead code). - -\begin{Variants} -\item {\tt compute} \tacindex{compute}\\ - {\tt cbv} - - These are synonyms for {\tt cbv beta delta iota zeta}. +% The effect of this tactic is similar to the effect of {\tt +% destruct {\term} in |- *} with the exception that no new hypotheses +% are introduced in the context. -\item {\tt lazy} - - This is a synonym for {\tt lazy beta delta iota zeta}. - -\item {\tt compute [\qualid$_1$\ldots\qualid$_k$]}\\ - {\tt cbv [\qualid$_1$\ldots\qualid$_k$]} +\end{Variants} - These are synonyms of {\tt cbv beta delta - [\qualid$_1$\ldots\qualid$_k$] iota zeta}. - -\item {\tt compute -[\qualid$_1$\ldots\qualid$_k$]}\\ - {\tt cbv -[\qualid$_1$\ldots\qualid$_k$]} +\subsection{\tt induction \term} +\tacindex{induction} +\label{Tac-induction} - These are synonyms of {\tt cbv beta delta - -[\qualid$_1$\ldots\qualid$_k$] iota zeta}. +This tactic applies to any goal. The argument {\term} must be of +inductive type and the tactic {\tt induction} generates subgoals, +one for each possible form of {\term}, i.e. one for each constructor +of the inductive type. -\item {\tt lazy [\qualid$_1$\ldots\qualid$_k$]}\\ - {\tt lazy -[\qualid$_1$\ldots\qualid$_k$]} +If the argument is dependent in either the conclusion or some +hypotheses of the goal, the argument is replaced by the appropriate +constructor form in each of the resulting subgoals and induction +hypotheses are added to the local context using names whose prefix is +{\tt IH}. - These are respectively synonyms of {\tt lazy beta delta - [\qualid$_1$\ldots\qualid$_k$] iota zeta} and {\tt lazy beta delta - -[\qualid$_1$\ldots\qualid$_k$] iota zeta}. +There are particular cases: -\item {\tt vm\_compute} \tacindex{vm\_compute} +\begin{itemize} - This tactic evaluates the goal using the optimized call-by-value - evaluation bytecode-based virtual machine. This algorithm is - dramatically more efficient than the algorithm used for the {\tt - cbv} tactic, but it cannot be fine-tuned. It is specially - interesting for full evaluation of algebraic objects. This includes - the case of reflexion-based tactics. +\item If {\term} is an identifier {\ident} denoting a quantified +variable of the conclusion of the goal, then {\tt induction {\ident}} +behaves as {\tt intros until {\ident}; induction {\ident}}. -\end{Variants} +\item If {\term} is a {\num}, then {\tt induction {\num}} behaves as +{\tt intros until {\num}} followed by {\tt induction} applied to the +last introduced hypothesis. Remark: For simple induction on a numeral, +use syntax {\tt induction ({\num})} (not very interesting anyway). -% Obsolete? Anyway not very important message -%\begin{ErrMsgs} -%\item \errindex{Delta must be specified before} -% -% A list of constants appeared before the {\tt delta} flag. -%\end{ErrMsgs} +\item The argument {\term} can also be a pattern of which holes are + denoted by ``\_''. In this case, the tactic checks that all subterms + matching the pattern in the conclusion and the hypotheses are + compatible and performs induction using this subterm. +\end{itemize} -\subsection{{\tt red} -\tacindex{red}} +\Example -This tactic applies to a goal which has the form {\tt - forall (x:T1)\dots(xk:Tk), c t1 \dots\ tn} where {\tt c} is a constant. If -{\tt c} is transparent then it replaces {\tt c} with its definition -(say {\tt t}) and then reduces {\tt (t t1 \dots\ tn)} according to -$\beta\iota\zeta$-reduction rules. +\begin{coq_example} +Lemma induction_test : forall n:nat, n = n -> n <= n. +intros n H. +induction n. +\end{coq_example} \begin{ErrMsgs} -\item \errindex{Not reducible} -\end{ErrMsgs} +\item \errindex{Not an inductive product} +\item \errindex{Unable to find an instance for the variables +{\ident} \ldots {\ident}} -\subsection{{\tt hnf} -\tacindex{hnf}} + Use in this case + the variant {\tt elim \dots\ with \dots} below. +\end{ErrMsgs} -This tactic applies to any goal. It replaces the current goal with its -head normal form according to the $\beta\delta\iota\zeta$-reduction -rules, i.e. it reduces the head of the goal until it becomes a -product or an irreducible term. +\begin{Variants} +\item{\tt induction {\term} as {\disjconjintropattern}} -\Example -The term \verb+forall n:nat, (plus (S n) (S n))+ is not reduced by {\tt hnf}. + This behaves as {\tt induction {\term}} but uses the names in + {\disjconjintropattern} to name the variables introduced in the context. + The {\disjconjintropattern} must typically be of the form + {\tt [} $p_{11}$ \ldots + $p_{1n_1}$ {\tt |} {\ldots} {\tt |} $p_{m1}$ \ldots $p_{mn_m}$ {\tt + ]} with $m$ being the number of constructors of the type of + {\term}. Each variable introduced by {\tt induction} in the context + of the $i^{th}$ goal gets its name from the list $p_{i1}$ \ldots + $p_{in_i}$ in order. If there are not enough names, {\tt induction} + invents names for the remaining variables to introduce. More + generally, the $p_{ij}$ can be any disjunctive/conjunctive + introduction pattern (see Section~\ref{intros-pattern}). For instance, + for an inductive type with one constructor, the pattern notation + {\tt ($p_{1}$,\ldots,$p_{n}$)} can be used instead of + {\tt [} $p_{1}$ \ldots $p_{n}$ {\tt ]}. -\Rem The $\delta$ rule only applies to transparent constants -(see Section~\ref{Opaque} on transparency and opacity). +%\item{\tt induction {\term} as {\namingintropattern}} -\subsection{\tt simpl -\tacindex{simpl}} +% This behaves as {\tt induction {\term}} but adds an equation between +% {\term} and the value that {\term} takes in each of the induction +% case. The name of the equation is built according to +% {\namingintropattern} which can be an identifier, a ``?'', etc, as +% indicated in Section~\ref{intros-pattern}. -This tactic applies to any goal. The tactic {\tt simpl} first applies -$\beta\iota$-reduction rule. Then it expands transparent constants -and tries to reduce {\tt T'} according, once more, to $\beta\iota$ -rules. But when the $\iota$ rule is not applicable then possible -$\delta$-reductions are not applied. For instance trying to use {\tt -simpl} on {\tt (plus n O)=n} changes nothing. Notice that only -transparent constants whose name can be reused as such in the -recursive calls are possibly unfolded. For instance a constant defined -by {\tt plus' := plus} is possibly unfolded and reused in the -recursive calls, but a constant such as {\tt succ := plus (S O)} is -never unfolded. +%\item{\tt induction {\term} as {\namingintropattern} {\disjconjintropattern}} -The behaviour of {\tt simpl} can be tuned using the {\tt Arguments} vernacular -command as follows: -\begin{itemize} -\item -A constant can be marked to be never unfolded by {\tt simpl}: -\begin{coq_example*} -Arguments minus x y : simpl never -\end{coq_example*} -After that command an expression like {\tt (minus (S x) y)} is left untouched by -the {\tt simpl} tactic. -\item -A constant can be marked to be unfolded only if applied to enough arguments. -The number of arguments required can be specified using -the {\tt /} symbol in the arguments list of the {\tt Arguments} vernacular -command. -\begin{coq_example*} -Definition fcomp A B C f (g : A -> B) (x : A) : C := f (g x). -Notation "f \o g" := (fcomp f g) (at level 50). -Arguments fcomp {A B C} f g x /. -\end{coq_example*} -After that command the expression {\tt (f \verb+\+o g)} is left untouched by -{\tt simpl} while {\tt ((f \verb+\+o g) t)} is reduced to {\tt (f (g t))}. -The same mechanism can be used to make a constant volatile, i.e. always -unfolded by {\tt simpl}. -\begin{coq_example*} -Definition volatile := fun x : nat => x. -Arguments volatile / x. -\end{coq_example*} -\item -A constant can be marked to be unfolded only if an entire set of arguments -evaluates to a constructor. The {\tt !} symbol can be used to mark such -arguments. -\begin{coq_example*} -Arguments minus !x !y. -\end{coq_example*} -After that command, the expression {\tt (minus (S x) y)} is left untouched by -{\tt simpl}, while {\tt (minus (S x) (S y))} is reduced to {\tt (minus x y)}. -\item -A special heuristic to determine if a constant has to be unfolded can be -activated with the following command: -\begin{coq_example*} -Arguments minus x y : simpl nomatch -\end{coq_example*} -The heuristic avoids to perform a simplification step that would -expose a {\tt match} construct in head position. For example the -expression {\tt (minus (S (S x)) (S y))} is simplified to -{\tt (minus (S x) y)} even if an extra simplification is possible. -\end{itemize} +% This combines the two previous forms. -\tacindex{simpl \dots\ in} -\begin{Variants} -\item {\tt simpl {\term}} - - This applies {\tt simpl} only to the occurrences of {\term} in the - current goal. +\item{\tt induction {\term} with \bindinglist} -\item {\tt simpl {\term} at \num$_1$ \dots\ \num$_i$} - - This applies {\tt simpl} only to the \num$_1$, \dots, \num$_i$ - occurrences of {\term} in the current goal. + This behaves like \texttt{induction {\term}} providing explicit + instances for the premises of the type of {\term} (see the syntax of + bindings in Section~\ref{Binding-list}). - \ErrMsg {\tt Too few occurrences} +\item{\tt einduction {\term}\tacindex{einduction}} -\item {\tt simpl {\ident}} - - This applies {\tt simpl} only to the applicative subterms whose head - occurrence is {\ident}. + This tactic behaves like \texttt{induction {\term}} excepts that it + does not fail if some dependent premise of the type of {\term} is + not inferable. Instead, the unresolved premises are posed as + existential variables to be inferred later, in the same way as {\tt + eapply} does (see Section~\ref{eapply-example}). -\item {\tt simpl {\ident} at \num$_1$ \dots\ \num$_i$} - - This applies {\tt simpl} only to the \num$_1$, \dots, \num$_i$ -applicative subterms whose head occurrence is {\ident}. +\item {\tt induction {\term$_1$} using {\term$_2$}} -\end{Variants} + This behaves as {\tt induction {\term$_1$}} but using {\term$_2$} as + induction scheme. It does not expect the conclusion of the type of + {\term$_1$} to be inductive. -\subsection{\tt unfold \qualid -\tacindex{unfold} -\label{unfold}} +\item {\tt induction {\term$_1$} using {\term$_2$} with {\bindinglist}} -This tactic applies to any goal. The argument {\qualid} must denote a -defined transparent constant or local definition (see Sections~\ref{Basic-definitions} and~\ref{Transparent}). The tactic {\tt - unfold} applies the $\delta$ rule to each occurrence of the constant -to which {\qualid} refers in the current goal and then replaces it -with its $\beta\iota$-normal form. + This behaves as {\tt induction {\term$_1$} using {\term$_2$}} but + also providing instances for the premises of the type of {\term$_2$}. -\begin{ErrMsgs} -\item {\qualid} \errindex{does not denote an evaluable constant} +\item \texttt{induction {\term}$_1$ $\ldots$ {\term}$_n$ using {\qualid}} -\end{ErrMsgs} + This syntax is used for the case {\qualid} denotes an induction principle + with complex predicates as the induction principles generated by + {\tt Function} or {\tt Functional Scheme} may be. -\begin{Variants} -\item {\tt unfold {\qualid}$_1$, \dots, \qualid$_n$} - \tacindex{unfold \dots\ in} - - Replaces {\em simultaneously} {\qualid}$_1$, \dots, {\qualid}$_n$ - with their definitions and replaces the current goal with its - $\beta\iota$ normal form. +\item \texttt{induction {\term} in {\occgoalset}} -\item {\tt unfold {\qualid}$_1$ at \num$_1^1$, \dots, \num$_i^1$, -\dots,\ \qualid$_n$ at \num$_1^n$ \dots\ \num$_j^n$} - - The lists \num$_1^1$, \dots, \num$_i^1$ and \num$_1^n$, \dots, - \num$_j^n$ specify the occurrences of {\qualid}$_1$, \dots, - \qualid$_n$ to be unfolded. Occurrences are located from left to - right. + This syntax is used for selecting which occurrences of {\term} the + induction has to be carried on. The {\tt in \occgoalset} clause is an + occurrence clause whose syntax and behavior is described in + Section~\ref{Occurrences clauses}. - \ErrMsg {\tt bad occurrence number of {\qualid}$_i$} +% When an occurrence clause is given, an equation between {\term} and +% the value it gets in each case of the induction is added to the +% context of the subgoals corresponding to the induction cases (even +% if no clause {\tt as {\namingintropattern}} is given). - \ErrMsg {\qualid}$_i$ {\tt does not occur} +\item {\tt induction {\term$_1$} with {\bindinglist$_1$} as {\namingintropattern} {\disjconjintropattern} using {\term$_2$} with {\bindinglist$_2$} in {\occgoalset}}\\ + {\tt einduction {\term$_1$} with {\bindinglist$_1$} as {\namingintropattern} {\disjconjintropattern} using {\term$_2$} with {\bindinglist$_2$} in {\occgoalset}} -\item {\tt unfold {\qstring}} + These are the most general forms of {\tt induction} and {\tt + einduction}. It combines the effects of the {\tt with}, {\tt as}, + {\tt using}, and {\tt in} clauses. - If {\qstring} denotes the discriminating symbol of a notation (e.g. {\tt - "+"}) or an expression defining a notation (e.g. \verb!"_ + _"!), and - this notation refers to an unfoldable constant, then the tactic - unfolds it. +\item {\tt elim \term}\label{elim} -\item {\tt unfold {\qstring}\%{\delimkey}} + This is a more basic induction tactic. Again, the type of the + argument {\term} must be an inductive type. Then, according to + the type of the goal, the tactic {\tt elim} chooses the appropriate + destructor and applies it as the tactic {\tt apply} + would do. For instance, if the proof context contains {\tt + n:nat} and the current goal is {\tt T} of type {\tt + Prop}, then {\tt elim n} is equivalent to {\tt apply nat\_ind with + (n:=n)}. The tactic {\tt elim} does not modify the context of + the goal, neither introduces the induction loading into the context + of hypotheses. - This is variant of {\tt unfold {\qstring}} where {\qstring} gets its - interpretation from the scope bound to the delimiting key - {\delimkey} instead of its default interpretation (see - Section~\ref{scopechange}). - -\item {\tt unfold \qualidorstring$_1$ at \num$_1^1$, \dots, \num$_i^1$, -\dots,\ \qualidorstring$_n$ at \num$_1^n$ \dots\ \num$_j^n$} + More generally, {\tt elim \term} also works when the type of {\term} + is a statement with premises and whose conclusion is inductive. In + that case the tactic performs induction on the conclusion of the + type of {\term} and leaves the non-dependent premises of the type as + subgoals. In the case of dependent products, the tactic tries to + find an instance for which the elimination lemma applies and fails + otherwise. - This is the most general form, where {\qualidorstring} is either a - {\qualid} or a {\qstring} referring to a notation. +\item {\tt elim {\term} with {\bindinglist}} -\end{Variants} + Allows to give explicit instances to the premises of the type + of {\term} (see Section~\ref{Binding-list}). -\subsection{{\tt fold} \term -\tacindex{fold}} +\item{\tt eelim {\term}\tacindex{eelim}} -This tactic applies to any goal. The term \term\ is reduced using the {\tt red} -tactic. Every occurrence of the resulting term in the goal is then -replaced by \term. + In case the type of {\term} has dependent premises, this turns them into + existential variables to be resolved later on. -\begin{Variants} -\item {\tt fold} \term$_1$ \dots\ \term$_n$ - - Equivalent to {\tt fold} \term$_1${\tt;}\ldots{\tt; fold} \term$_n$. -\end{Variants} +\item{\tt elim {\term$_1$} using {\term$_2$}}\\ + {\tt elim {\term$_1$} using {\term$_2$} with {\bindinglist}\tacindex{elim \dots\ using}} -\subsection{{\tt pattern {\term}} -\tacindex{pattern} -\label{pattern}} +Allows the user to give explicitly an elimination predicate +{\term$_2$} which is not the standard one for the underlying inductive +type of {\term$_1$}. The {\bindinglist} clause allows to +instantiate premises of the type of {\term$_2$}. -This command applies to any goal. The argument {\term} must be a free -subterm of the current goal. The command {\tt pattern} performs -$\beta$-expansion (the inverse of $\bt$-reduction) of the current goal -(say \T) by -\begin{enumerate} -\item replacing all occurrences of {\term} in {\T} with a fresh variable -\item abstracting this variable -\item applying the abstracted goal to {\term} -\end{enumerate} +\item{\tt elim {\term$_1$} with {\bindinglist$_1$} using {\term$_2$} with {\bindinglist$_2$}}\\ + {\tt eelim {\term$_1$} with {\bindinglist$_1$} using {\term$_2$} with {\bindinglist$_2$}} -For instance, if the current goal $T$ is expressible has $\phi(t)$ -where the notation captures all the instances of $t$ in $\phi(t)$, -then {\tt pattern $t$} transforms it into {\tt (fun x:$A$ => $\phi(${\tt -x}$)$) $t$}. This command can be used, for instance, when the tactic -{\tt apply} fails on matching. + These are the most general forms of {\tt elim} and {\tt eelim}. It + combines the effects of the {\tt using} clause and of the two uses + of the {\tt with} clause. -\begin{Variants} -\item {\tt pattern {\term} at {\num$_1$} \dots\ {\num$_n$}} - - Only the occurrences {\num$_1$} \dots\ {\num$_n$} of {\term} are - considered for $\beta$-expansion. Occurrences are located from left - to right. +\item {\tt elimtype \form}\tacindex{elimtype} -\item {\tt pattern {\term} at - {\num$_1$} \dots\ {\num$_n$}} - - All occurrences except the occurrences of indexes {\num$_1$} \dots\ - {\num$_n$} of {\term} are considered for - $\beta$-expansion. Occurrences are located from left to right. + The argument {\form} must be inductively defined. {\tt elimtype I} + is equivalent to {\tt cut I. intro H{\rm\sl n}; elim H{\rm\sl n}; + clear H{\rm\sl n}}. Therefore the hypothesis {\tt H{\rm\sl n}} will + not appear in the context(s) of the subgoal(s). Conversely, if {\tt + t} is a term of (inductive) type {\tt I} and which does not occur + in the goal then {\tt elim t} is equivalent to {\tt elimtype I; 2: + exact t.} -\item {\tt pattern {\term$_1$}, \dots, {\term$_m$}} - - Starting from a goal $\phi(t_1 \dots\ t_m)$, the tactic - {\tt pattern $t_1$, \dots,\ $t_m$} generates the equivalent goal {\tt - (fun (x$_1$:$A_1$) \dots\ (x$_m$:$A_m$) => $\phi(${\tt x$_1$\dots\ - x$_m$}$)$) $t_1$ \dots\ $t_m$}.\\ If $t_i$ occurs in one of the - generated types $A_j$ these occurrences will also be considered and - possibly abstracted. +\item {\tt simple induction \ident}\tacindex{simple induction} -\item {\tt pattern {\term$_1$} at {\num$_1^1$} \dots\ {\num$_{n_1}^1$}, \dots, - {\term$_m$} at {\num$_1^m$} \dots\ {\num$_{n_m}^m$}} - - This behaves as above but processing only the occurrences \num$_1^1$, - \dots, \num$_i^1$ of \term$_1$, \dots, \num$_1^m$, \dots, \num$_j^m$ - of \term$_m$ starting from \term$_m$. + This tactic behaves as {\tt intros until + {\ident}; elim {\tt {\ident}}} when {\ident} is a quantified + variable of the goal. -\item {\tt pattern} {\term$_1$} \zeroone{{\tt at \zeroone{-}} {\num$_1^1$} \dots\ {\num$_{n_1}^1$}} {\tt ,} \dots {\tt ,} - {\term$_m$} \zeroone{{\tt at \zeroone{-}} {\num$_1^m$} \dots\ {\num$_{n_m}^m$}} - - This is the most general syntax that combines the different variants. +\item {\tt simple induction {\num}} -\end{Variants} + This tactic behaves as {\tt intros until + {\num}; elim {\tt {\ident}}} where {\ident} is the name given by + {\tt intros until {\num}} to the {\num}-th non-dependent premise of + the goal. -\subsection{Conversion tactics applied to hypotheses} +%% \item {\tt simple induction {\term}}\tacindex{simple induction} -{\convtactic} {\tt in} \ident$_1$ \dots\ \ident$_n$ +%% If {\term} is an {\ident} corresponding to a quantified variable of +%% the goal then the tactic behaves as {\tt intros until {\ident}; elim +%% {\tt {\ident}}}. If {\term} is a {\num} then the tactic behaves as +%% {\tt intros until {\ident}; elim {\tt {\ident}}}. Otherwise, it is +%% a synonym for {\tt elim {\term}}. -Applies the conversion tactic {\convtactic} to the -hypotheses \ident$_1$, \ldots, \ident$_n$. The tactic {\convtactic} is -any of the conversion tactics listed in this section. +%% \Rem For simple induction on a numeral, use syntax {\tt simple +%% induction ({\num})}. -If \ident$_i$ is a local definition, then \ident$_i$ can be replaced -by (Type of \ident$_i$) to address not the body but the type of the -local definition. Example: {\tt unfold not in (Type of H1) (Type of H3).} +\end{Variants} -\begin{ErrMsgs} -\item \errindex{No such hypothesis} : {\ident}. -\end{ErrMsgs} +%\subsection[\tt FixPoint \dots]{\tt FixPoint \dots\tacindex{Fixpoint}} +%Not yet documented. +\subsection{\tt double induction \ident$_1$ \ident$_2$} +\tacindex{double induction} -\section{Introductions} +This tactic is deprecated and should be replaced by {\tt induction \ident$_1$; induction \ident$_2$} (or {\tt induction \ident$_1$; destruct \ident$_2$} depending on the exact needs). -Introduction tactics address goals which are inductive constants. -They are used when one guesses that the goal can be obtained with one -of its constructors' type. +%% This tactic applies to any goal. If the variables {\ident$_1$} and +%% {\ident$_2$} of the goal have an inductive type, then this tactic +%% performs double induction on these variables. For instance, if the +%% current goal is \verb+forall n m:nat, P n m+ then, {\tt double induction n +%% m} yields the four cases with their respective inductive hypotheses. -\subsection{\tt constructor \num -\label{constructor} -\tacindex{constructor}} +%% In particular, for proving \verb+(P (S n) (S m))+, the generated induction +%% hypotheses are \verb+(P (S n) m)+ and \verb+(m:nat)(P n m)+ (of the latter, +%% \verb+(P n m)+ and \verb+(P n (S m))+ are derivable). -This tactic applies to a goal such that the head of its conclusion is -an inductive constant (say {\tt I}). The argument {\num} must be less -or equal to the numbers of constructor(s) of {\tt I}. Let {\tt ci} be -the {\tt i}-th constructor of {\tt I}, then {\tt constructor i} is -equivalent to {\tt intros; apply ci}. +%% \Rem When the induction hypothesis \verb+(P (S n) m)+ is not +%% needed, {\tt induction \ident$_1$; destruct \ident$_2$} produces +%% more concise subgoals. -\begin{ErrMsgs} -\item \errindex{Not an inductive product} -\item \errindex{Not enough constructors} -\end{ErrMsgs} +\begin{Variant} -\begin{Variants} -\item \texttt{constructor} - - This tries \texttt{constructor 1} then \texttt{constructor 2}, - \dots\ , then \texttt{constructor} \textit{n} where \textit{n} if - the number of constructors of the head of the goal. +\item {\tt double induction \num$_1$ \num$_2$} -\item {\tt constructor \num~with} {\bindinglist} - - Let {\tt ci} be the {\tt i}-th constructor of {\tt I}, then {\tt - constructor i with \bindinglist} is equivalent to {\tt intros; - apply ci with \bindinglist}. +This tactic is deprecated and should be replaced by {\tt induction + \num$_1$; induction \num$_3$} where \num$_3$ is the result of +\num$_2$-\num$_1$. - \Warning the terms in the \bindinglist\ are checked - in the context where {\tt constructor} is executed and not in the - context where {\tt apply} is executed (the introductions are not - taken into account). +%% This tactic applies to any goal. If the variables {\ident$_1$} and -% To document? -% \item {\tt constructor {\tactic}} +%% This applies double induction on the \num$_1^{th}$ and \num$_2^{th}$ {\it +%% non dependent} premises of the goal. More generally, any combination of an +%% {\ident} and a {\num} is valid. -\item {\tt split}\tacindex{split} +\end{Variant} - Applies if {\tt I} has only one constructor, typically in the case - of conjunction $A\land B$. Then, it is equivalent to {\tt constructor 1}. +\subsection{\tt dependent induction \ident} +\tacindex{dependent induction} +\label{DepInduction} -\item {\tt exists {\bindinglist}}\tacindex{exists} +The \emph{experimental} tactic \texttt{dependent induction} performs +induction-inversion on an instantiated inductive predicate. +One needs to first require the {\tt Coq.Program.Equality} module to use +this tactic. The tactic is based on the BasicElim tactic by Conor +McBride \cite{DBLP:conf/types/McBride00} and the work of Cristina Cornes +around inversion \cite{DBLP:conf/types/CornesT95}. From an instantiated +inductive predicate and a goal, it generates an equivalent goal where the +hypothesis has been generalized over its indexes which are then +constrained by equalities to be the right instances. This permits to +state lemmas without resorting to manually adding these equalities and +still get enough information in the proofs. - Applies if {\tt I} has only one constructor, for instance in the - case of existential quantification $\exists x\cdot P(x)$. - Then, it is equivalent to {\tt intros; constructor 1 with \bindinglist}. +\Example -\item {\tt exists \nelist{\bindinglist}{,}} +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{coq_example} +Lemma le_minus : forall n:nat, n < 1 -> n = 0. +intros n H ; induction H. +\end{coq_example} - This iteratively applies {\tt exists {\bindinglist}}. +Here we did not get any information on the indexes to help fulfill this +proof. The problem is that, when we use the \texttt{induction} tactic, +we lose information on the hypothesis instance, notably that the second +argument is \texttt{1} here. Dependent induction solves this problem by +adding the corresponding equality to the context. -\item {\tt left}\tacindex{left}\\ - {\tt right}\tacindex{right} +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{coq_example} +Require Import Coq.Program.Equality. +Lemma le_minus : forall n:nat, n < 1 -> n = 0. +intros n H ; dependent induction H. +\end{coq_example} - Apply if {\tt I} has two constructors, for instance in the case of - disjunction $A\lor B$. Then, they are respectively equivalent to {\tt - constructor 1} and {\tt constructor 2}. - -\item {\tt left with \bindinglist}\\ - {\tt right with \bindinglist}\\ - {\tt split with \bindinglist} - - As soon as the inductive type has the right number of constructors, - these expressions are equivalent to calling {\tt - constructor $i$ with \bindinglist} for the appropriate $i$. +The subgoal is cleaned up as the tactic tries to automatically +simplify the subgoals with respect to the generated equalities. +In this enriched context, it becomes possible to solve this subgoal. +\begin{coq_example} +reflexivity. +\end{coq_example} -\item \texttt{econstructor}\tacindex{econstructor}\\ - \texttt{eexists}\tacindex{eexists}\\ - \texttt{esplit}\tacindex{esplit}\\ - \texttt{eleft}\tacindex{eleft}\\ - \texttt{eright}\tacindex{eright}\\ +Now we are in a contradictory context and the proof can be solved. +\begin{coq_example} +inversion H. +\end{coq_example} - These tactics and their variants behave like \texttt{constructor}, - \texttt{exists}, \texttt{split}, \texttt{left}, \texttt{right} and - their variants but they introduce existential variables instead of - failing when the instantiation of a variable cannot be found (cf - \texttt{eapply} and Section~\ref{eapply-example}). +This technique works with any inductive predicate. +In fact, the \texttt{dependent induction} tactic is just a wrapper around +the \texttt{induction} tactic. One can make its own variant by just +writing a new tactic based on the definition found in +\texttt{Coq.Program.Equality}. -\end{Variants} +\begin{Variants} +\item {\tt dependent induction {\ident} generalizing {\ident$_1$} \dots + {\ident$_n$}}\tacindex{dependent induction \dots\ generalizing} -\section[Induction and Case Analysis]{Induction and Case Analysis -\label{Tac-induction}} + This performs dependent induction on the hypothesis {\ident} but first + generalizes the goal by the given variables so that they are + universally quantified in the goal. This is generally what one wants + to do with the variables that are inside some constructors in the + induction hypothesis. The other ones need not be further generalized. -The tactics presented in this section implement induction or case -analysis on inductive or coinductive objects (see -Section~\ref{Cic-inductive-definitions}). +\item {\tt dependent destruction {\ident}}\tacindex{dependent destruction} -\subsection{\tt induction \term -\tacindex{induction}} + This performs the generalization of the instance {\ident} but uses {\tt destruct} + instead of {\tt induction} on the generalized hypothesis. This gives + results equivalent to {\tt inversion} or {\tt dependent inversion} if + the hypothesis is dependent. +\end{Variants} -This tactic applies to any goal. The argument {\term} must be of -inductive type and the tactic {\tt induction} generates subgoals, -one for each possible form of {\term}, i.e. one for each constructor -of the inductive type. +\SeeAlso \ref{dependent-induction-example} for a larger example of +dependent induction and an explanation of the underlying technique. -If the argument is dependent in either the conclusion or some -hypotheses of the goal, the argument is replaced by the appropriate -constructor form in each of the resulting subgoals and induction -hypotheses are added to the local context using names whose prefix is -{\tt IH}. +\subsection{\tt functional induction (\qualid\ \term$_1$ \dots\ \term$_n$)} +\tacindex{functional induction} +\label{FunInduction} -There are particular cases: +The \emph{experimental} tactic \texttt{functional induction} performs +case analysis and induction following the definition of a function. It +makes use of a principle generated by \texttt{Function} +(see Section~\ref{Function}) or \texttt{Functional Scheme} +(see Section~\ref{FunScheme}). -\begin{itemize} - -\item If {\term} is an identifier {\ident} denoting a quantified -variable of the conclusion of the goal, then {\tt induction {\ident}} -behaves as {\tt intros until {\ident}; induction {\ident}}. +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{coq_example} +Functional Scheme minus_ind := Induction for minus Sort Prop. +Check minus_ind. +Lemma le_minus (n m:nat) : n - m <= n. +functional induction (minus n m); simpl; auto. +\end{coq_example} +\begin{coq_example*} +Qed. +\end{coq_example*} -\item If {\term} is a {\num}, then {\tt induction {\num}} behaves as -{\tt intros until {\num}} followed by {\tt induction} applied to the -last introduced hypothesis. Remark: For simple induction on a numeral, -use syntax {\tt induction ({\num})} (not very interesting anyway). +\Rem \texttt{(\qualid\ \term$_1$ \dots\ \term$_n$)} must be a correct +full application of \qualid. In particular, the rules for implicit +arguments are the same as usual. For example use \texttt{@\qualid} if +you want to write implicit arguments explicitly. -\item The argument {\term} can also be a pattern of which holes are - denoted by ``\_''. In this case, the tactic checks that all subterms - matching the pattern in the conclusion and the hypotheses are - compatible and performs induction using this subterm. +\Rem Parentheses over \qualid \dots \term$_n$ are mandatory. -\end{itemize} +\Rem \texttt{functional induction (f x1 x2 x3)} is actually a wrapper +for \texttt{induction x1 x2 x3 (f x1 x2 x3) using \qualid} followed by +a cleaning phase, where {\qualid} is the induction principle +registered for $f$ (by the \texttt{Function} (see Section~\ref{Function}) +or \texttt{Functional Scheme} (see Section~\ref{FunScheme}) command) +corresponding to the sort of the goal. Therefore \texttt{functional + induction} may fail if the induction scheme {\qualid} is +not defined. See also Section~\ref{Function} for the function terms +accepted by \texttt{Function}. -\Example +\Rem There is a difference between obtaining an induction scheme for a +function by using \texttt{Function} (see Section~\ref{Function}) and by +using \texttt{Functional Scheme} after a normal definition using +\texttt{Fixpoint} or \texttt{Definition}. See \ref{Function} for +details. -\begin{coq_example} -Lemma induction_test : forall n:nat, n = n -> n <= n. -intros n H. -induction n. -\end{coq_example} +\SeeAlso{\ref{Function},\ref{FunScheme},\ref{FunScheme-examples}, + \ref{sec:functional-inversion}} \begin{ErrMsgs} -\item \errindex{Not an inductive product} -\item \errindex{Unable to find an instance for the variables -{\ident} \ldots {\ident}} - - Use in this case - the variant {\tt elim \dots\ with \dots} below. +\item \errindex{Cannot find induction information on \qualid} +\item \errindex{Not the right number of induction arguments} \end{ErrMsgs} \begin{Variants} -\item{\tt induction {\term} as {\disjconjintropattern}} - - This behaves as {\tt induction {\term}} but uses the names in - {\disjconjintropattern} to name the variables introduced in the context. - The {\disjconjintropattern} must typically be of the form - {\tt [} $p_{11}$ \ldots - $p_{1n_1}$ {\tt |} {\ldots} {\tt |} $p_{m1}$ \ldots $p_{mn_m}$ {\tt - ]} with $m$ being the number of constructors of the type of - {\term}. Each variable introduced by {\tt induction} in the context - of the $i^{th}$ goal gets its name from the list $p_{i1}$ \ldots - $p_{in_i}$ in order. If there are not enough names, {\tt induction} - invents names for the remaining variables to introduce. More - generally, the $p_{ij}$ can be any disjunctive/conjunctive - introduction pattern (see Section~\ref{intros-pattern}). For instance, - for an inductive type with one constructor, the pattern notation - {\tt ($p_{1}$,\ldots,$p_{n}$)} can be used instead of - {\tt [} $p_{1}$ \ldots $p_{n}$ {\tt ]}. +\item {\tt functional induction (\qualid\ \term$_1$ \dots\ \term$_n$) + as {\disjconjintropattern} using \term$_{m+1}$ with \bindinglist} -\item{\tt induction {\term} as {\namingintropattern}} + Similarly to \texttt{Induction} and \texttt{elim} + (see Section~\ref{Tac-induction}), this allows to give explicitly the + name of the introduced variables, the + induction principle, and the values of dependent premises of the + elimination scheme, including \emph{predicates} for mutual induction + when {\qualid} is part of a mutually recursive definition. - This behaves as {\tt induction {\term}} but adds an equation between - {\term} and the value that {\term} takes in each of the induction - case. The name of the equation is built according to - {\namingintropattern} which can be an identifier, a ``?'', etc, as - indicated in Section~\ref{intros-pattern}. +\end{Variants} -\item{\tt induction {\term} as {\namingintropattern} {\disjconjintropattern}} +\subsection{\tt discriminate \term} +\label{discriminate} +\tacindex{discriminate} - This combines the two previous forms. -\item{\tt induction {\term} with \bindinglist} +This tactic proves any goal from an assumption stating that two +structurally different terms of an inductive set are equal. For +example, from {\tt (S (S O))=(S O)} we can derive by absurdity any +proposition. - This behaves like \texttt{induction {\term}} providing explicit - instances for the premises of the type of {\term} (see the syntax of - bindings in Section~\ref{Binding-list}). +The argument {\term} is assumed to be a proof of a statement +of conclusion {\tt{\term$_1$} = {\term$_2$}} with {\term$_1$} and +{\term$_2$} being elements of an inductive set. To build the proof, +the tactic traverses the normal forms\footnote{Reminder: opaque + constants will not be expanded by $\delta$ reductions.} of +{\term$_1$} and {\term$_2$} looking for a couple of subterms {\tt u} +and {\tt w} ({\tt u} subterm of the normal form of {\term$_1$} and +{\tt w} subterm of the normal form of {\term$_2$}), placed at the same +positions and whose head symbols are two different constructors. If +such a couple of subterms exists, then the proof of the current goal +is completed, otherwise the tactic fails. -\item{\tt einduction {\term}\tacindex{einduction}} +\Rem The syntax {\tt discriminate {\ident}} can be used to refer to a +hypothesis quantified in the goal. In this case, the quantified +hypothesis whose name is {\ident} is first introduced in the local +context using \texttt{intros until \ident}. - This tactic behaves like \texttt{induction {\term}} excepts that it - does not fail if some dependent premise of the type of {\term} is - not inferable. Instead, the unresolved premises are posed as - existential variables to be inferred later, in the same way as {\tt - eapply} does (see Section~\ref{eapply-example}). +\begin{ErrMsgs} +\item \errindex{No primitive equality found} +\item \errindex{Not a discriminable equality} +\end{ErrMsgs} -\item {\tt induction {\term$_1$} using {\term$_2$}} +\begin{Variants} +\item \texttt{discriminate \num} - This behaves as {\tt induction {\term$_1$}} but using {\term$_2$} as - induction scheme. It does not expect the conclusion of the type of - {\term$_1$} to be inductive. + This does the same thing as \texttt{intros until \num} followed by + \texttt{discriminate \ident} where {\ident} is the identifier for + the last introduced hypothesis. -\item {\tt induction {\term$_1$} using {\term$_2$} with {\bindinglist}} +\item \texttt{discriminate {\term} with \bindinglist} - This behaves as {\tt induction {\term$_1$} using {\term$_2$}} but - also providing instances for the premises of the type of {\term$_2$}. + This does the same thing as \texttt{discriminate {\term}} but using +the given bindings to instantiate parameters or hypotheses of {\term}. -\item \texttt{induction {\term}$_1$ $\ldots$ {\term}$_n$ using {\qualid}} +\item \texttt{ediscriminate \num}\tacindex{ediscriminate}\\ + \texttt{ediscriminate {\term} \zeroone{with \bindinglist}} - This syntax is used for the case {\qualid} denotes an induction principle - with complex predicates as the induction principles generated by - {\tt Function} or {\tt Functional Scheme} may be. + This works the same as {\tt discriminate} but if the type of {\term}, + or the type of the hypothesis referred to by {\num}, has uninstantiated + parameters, these parameters are left as existential variables. -\item \texttt{induction {\term} in {\occgoalset}} +\item \texttt{discriminate} - This syntax is used for selecting which occurrences of {\term} the - induction has to be carried on. The {\tt in {\atoccurrences}} clause is an - occurrence clause whose syntax and behavior is described in - Section~\ref{Occurrences clauses}. + This behaves like {\tt discriminate {\ident}} if {\ident} is the + name of an hypothesis to which {\tt discriminate} is applicable; if + the current goal is of the form {\term$_1$} {\tt <>} {\term$_2$}, + this behaves as {\tt intro {\ident}; injection {\ident}}. - When an occurrence clause is given, an equation between {\term} and - the value it gets in each case of the induction is added to the - context of the subgoals corresponding to the induction cases (even - if no clause {\tt as {\namingintropattern}} is given). + \ErrMsg \errindex{No discriminable equalities} +\end{Variants} -\item {\tt induction {\term$_1$} with {\bindinglist$_1$} as {\namingintropattern} {\disjconjintropattern} using {\term$_2$} with {\bindinglist$_2$} in {\occgoalset}}\\ - {\tt einduction {\term$_1$} with {\bindinglist$_1$} as {\namingintropattern} {\disjconjintropattern} using {\term$_2$} with {\bindinglist$_2$} in {\occgoalset}} +\subsection{\tt injection \term} +\label{injection} +\tacindex{injection} - These are the most general forms of {\tt induction} and {\tt - einduction}. It combines the effects of the {\tt with}, {\tt as}, - {\tt using}, and {\tt in} clauses. +The {\tt injection} tactic is based on the fact that constructors of +inductive sets are injections. That means that if $c$ is a constructor +of an inductive set, and if $(c~\vec{t_1})$ and $(c~\vec{t_2})$ are two +terms that are equal then $~\vec{t_1}$ and $~\vec{t_2}$ are equal +too. -\item {\tt elim \term}\label{elim} - - This is a more basic induction tactic. Again, the type of the - argument {\term} must be an inductive type. Then, according to - the type of the goal, the tactic {\tt elim} chooses the appropriate - destructor and applies it as the tactic {\tt apply} - would do. For instance, if the proof context contains {\tt - n:nat} and the current goal is {\tt T} of type {\tt - Prop}, then {\tt elim n} is equivalent to {\tt apply nat\_ind with - (n:=n)}. The tactic {\tt elim} does not modify the context of - the goal, neither introduces the induction loading into the context - of hypotheses. +If {\term} is a proof of a statement of conclusion + {\tt {\term$_1$} = {\term$_2$}}, +then {\tt injection} applies injectivity as deep as possible to +derive the equality of all the subterms of {\term$_1$} and {\term$_2$} +placed in the same positions. For example, from {\tt (S + (S n))=(S (S (S m)))} we may derive {\tt n=(S m)}. To use this +tactic {\term$_1$} and {\term$_2$} should be elements of an inductive +set and they should be neither explicitly equal, nor structurally +different. We mean by this that, if {\tt n$_1$} and {\tt n$_2$} are +their respective normal forms, then: +\begin{itemize} +\item {\tt n$_1$} and {\tt n$_2$} should not be syntactically equal, +\item there must not exist any pair of subterms {\tt u} and {\tt w}, + {\tt u} subterm of {\tt n$_1$} and {\tt w} subterm of {\tt n$_2$} , + placed in the same positions and having different constructors as + head symbols. +\end{itemize} +If these conditions are satisfied, then, the tactic derives the +equality of all the subterms of {\term$_1$} and {\term$_2$} placed in +the same positions and puts them as antecedents of the current goal. - More generally, {\tt elim \term} also works when the type of {\term} - is a statement with premises and whose conclusion is inductive. In - that case the tactic performs induction on the conclusion of the - type of {\term} and leaves the non-dependent premises of the type as - subgoals. In the case of dependent products, the tactic tries to - find an instance for which the elimination lemma applies and fails - otherwise. +\Example Consider the following goal: -\item {\tt elim {\term} with {\bindinglist}} - - Allows to give explicit instances to the premises of the type - of {\term} (see Section~\ref{Binding-list}). +\begin{coq_example*} +Inductive list : Set := + | nil : list + | cons : nat -> list -> list. +Variable P : list -> Prop. +\end{coq_example*} +\begin{coq_eval} +Lemma ex : + forall (l:list) (n:nat), P nil -> cons n l = cons 0 nil -> P l. +intros l n H H0. +\end{coq_eval} +\begin{coq_example} +Show. +injection H0. +\end{coq_example} +\begin{coq_eval} +Abort. +\end{coq_eval} -\item{\tt eelim {\term}\tacindex{eelim}} +Beware that \texttt{injection} yields always an equality in a sigma type +whenever the injected object has a dependent type. - In case the type of {\term} has dependent premises, this turns them into - existential variables to be resolved later on. +\Rem There is a special case for dependent pairs. If we have a decidable +equality over the type of the first argument, then it is safe to do +the projection on the second one, and so {\tt injection} will work fine. +To define such an equality, you have to use the {\tt Scheme} command +(see \ref{Scheme}). -\item{\tt elim {\term$_1$} using {\term$_2$}}\\ - {\tt elim {\term$_1$} using {\term$_2$} with {\bindinglist}\tacindex{elim \dots\ using}} +\Rem If some quantified hypothesis of the goal is named {\ident}, then +{\tt injection {\ident}} first introduces the hypothesis in the local +context using \texttt{intros until \ident}. -Allows the user to give explicitly an elimination predicate -{\term$_2$} which is not the standard one for the underlying inductive -type of {\term$_1$}. The {\bindinglist} clause allows to -instantiate premises of the type of {\term$_2$}. +\begin{ErrMsgs} +\item \errindex{Not a projectable equality but a discriminable one} +\item \errindex{Nothing to do, it is an equality between convertible terms} +\item \errindex{Not a primitive equality} +\end{ErrMsgs} -\item{\tt elim {\term$_1$} with {\bindinglist$_1$} using {\term$_2$} with {\bindinglist$_2$}}\\ - {\tt eelim {\term$_1$} with {\bindinglist$_1$} using {\term$_2$} with {\bindinglist$_2$}} +\begin{Variants} +\item \texttt{injection \num} - These are the most general forms of {\tt elim} and {\tt eelim}. It - combines the effects of the {\tt using} clause and of the two uses - of the {\tt with} clause. + This does the same thing as \texttt{intros until \num} followed by +\texttt{injection \ident} where {\ident} is the identifier for the last +introduced hypothesis. -\item {\tt elimtype \form}\tacindex{elimtype} - - The argument {\form} must be inductively defined. {\tt elimtype I} - is equivalent to {\tt cut I. intro H{\rm\sl n}; elim H{\rm\sl n}; - clear H{\rm\sl n}}. Therefore the hypothesis {\tt H{\rm\sl n}} will - not appear in the context(s) of the subgoal(s). Conversely, if {\tt - t} is a term of (inductive) type {\tt I} and which does not occur - in the goal then {\tt elim t} is equivalent to {\tt elimtype I; 2: - exact t.} +\item \texttt{injection {\term} with \bindinglist} -\item {\tt simple induction \ident}\tacindex{simple induction} - - This tactic behaves as {\tt intros until - {\ident}; elim {\tt {\ident}}} when {\ident} is a quantified - variable of the goal. + This does the same as \texttt{injection {\term}} but using + the given bindings to instantiate parameters or hypotheses of {\term}. -\item {\tt simple induction {\num}} - - This tactic behaves as {\tt intros until - {\num}; elim {\tt {\ident}}} where {\ident} is the name given by - {\tt intros until {\num}} to the {\num}-th non-dependent premise of - the goal. +\item \texttt{einjection \num}\tacindex{einjection}\\ + \texttt{einjection {\term} \zeroone{with \bindinglist}} -%% \item {\tt simple induction {\term}}\tacindex{simple induction} - -%% If {\term} is an {\ident} corresponding to a quantified variable of -%% the goal then the tactic behaves as {\tt intros until {\ident}; elim -%% {\tt {\ident}}}. If {\term} is a {\num} then the tactic behaves as -%% {\tt intros until {\ident}; elim {\tt {\ident}}}. Otherwise, it is -%% a synonym for {\tt elim {\term}}. + This works the same as {\tt injection} but if the type of {\term}, + or the type of the hypothesis referred to by {\num}, has uninstantiated + parameters, these parameters are left as existential variables. -%% \Rem For simple induction on a numeral, use syntax {\tt simple -%% induction ({\num})}. +\item{\tt injection} + + If the current goal is of the form {\term$_1$} {\tt <>} {\term$_2$}, + this behaves as {\tt intro {\ident}; injection {\ident}}. + + \ErrMsg \errindex{goal does not satisfy the expected preconditions} + +\item \texttt{injection {\term} \zeroone{with \bindinglist} as \nelist{\intropattern}{}}\\ +\texttt{injection {\num} as {\intropattern} \dots\ \intropattern}\\ +\texttt{injection as {\intropattern} \dots\ \intropattern}\\ +\texttt{einjection {\term} \zeroone{with \bindinglist} as \nelist{\intropattern}{}}\\ +\texttt{einjection {\num} as {\intropattern} \dots\ \intropattern}\\ +\texttt{einjection as {\intropattern} \dots\ \intropattern} +\tacindex{injection \dots\ as} + +These variants apply \texttt{intros} \nelist{\intropattern}{} after +the call to \texttt{injection} or \texttt{einjection}. \end{Variants} -\subsection{\tt destruct \term -\tacindex{destruct}} -\label{destruct} +\subsection{\tt inversion \ident} +\tacindex{inversion} -This tactic applies to any goal. The argument {\term} must be of -inductive or coinductive type and the tactic generates subgoals, one -for each possible form of {\term}, i.e. one for each constructor of -the inductive or coinductive type. Unlike {\tt induction}, no -induction hypothesis is generated by {\tt destruct}. +Let the type of {\ident} in the local context be $(I~\vec{t})$, +where $I$ is a (co)inductive predicate. Then, +\texttt{inversion} applied to \ident~ derives for each possible +constructor $c_i$ of $(I~\vec{t})$, {\bf all} the necessary +conditions that should hold for the instance $(I~\vec{t})$ to be +proved by $c_i$. -If the argument is dependent in either the conclusion or some -hypotheses of the goal, the argument is replaced by the appropriate -constructor form in each of the resulting subgoals, thus performing -case analysis. If non dependent, the tactic simply exposes the -inductive or coinductive structure of the argument. +\Rem If {\ident} does not denote a hypothesis in the local context +but refers to a hypothesis quantified in the goal, then the +latter is first introduced in the local context using +\texttt{intros until \ident}. -There are special cases: +\Rem As inversion proofs may be large in size, we recommend the user to +stock the lemmas whenever the same instance needs to be inverted +several times. See Section~\ref{Derive-Inversion}. -\begin{itemize} +\begin{Variants} +\item \texttt{inversion \num} -\item If {\term} is an identifier {\ident} denoting a quantified -variable of the conclusion of the goal, then {\tt destruct {\ident}} -behaves as {\tt intros until {\ident}; destruct {\ident}}. + This does the same thing as \texttt{intros until \num} then + \texttt{inversion \ident} where {\ident} is the identifier for the + last introduced hypothesis. -\item If {\term} is a {\num}, then {\tt destruct {\num}} behaves as -{\tt intros until {\num}} followed by {\tt destruct} applied to the -last introduced hypothesis. Remark: For destruction of a numeral, use -syntax {\tt destruct ({\num})} (not very interesting anyway). +\item \tacindex{inversion\_clear} \texttt{inversion\_clear \ident} -\item The argument {\term} can also be a pattern of which holes are - denoted by ``\_''. In this case, the tactic checks that all subterms - matching the pattern in the conclusion and the hypotheses are - compatible and performs case analysis using this subterm. + This behaves as \texttt{inversion} and then erases \ident~ from the + context. -\end{itemize} +\item \tacindex{inversion \dots\ as} \texttt{inversion {\ident} as \intropattern} -\begin{Variants} -\item{\tt destruct {\term} as {\disjconjintropattern}} - - This behaves as {\tt destruct {\term}} but uses the names in - {\intropattern} to name the variables introduced in the context. - The {\intropattern} must have the form {\tt [} $p_{11}$ \ldots - $p_{1n_1}$ {\tt |} {\ldots} {\tt |} $p_{m1}$ \ldots $p_{mn_m}$ {\tt - ]} with $m$ being the number of constructors of the type of - {\term}. Each variable introduced by {\tt destruct} in the context - of the $i^{th}$ goal gets its name from the list $p_{i1}$ \ldots - $p_{in_i}$ in order. If there are not enough names, {\tt destruct} - invents names for the remaining variables to introduce. More - generally, the $p_{ij}$ can be any disjunctive/conjunctive - introduction pattern (see Section~\ref{intros-pattern}). This - provides a concise notation for nested destruction. + This behaves as \texttt{inversion} but using names in + {\intropattern} for naming hypotheses. The {\intropattern} must have + the form {\tt [} $p_{11}$ \ldots $p_{1n_1}$ {\tt |} {\ldots} {\tt |} + $p_{m1}$ \ldots $p_{mn_m}$ {\tt ]} with $m$ being the number of + constructors of the type of {\ident}. Be careful that the list must + be of length $m$ even if {\tt inversion} discards some cases (which + is precisely one of its roles): for the discarded cases, just use an + empty list (i.e. $n_i=0$). -% It is recommended to use this variant of {\tt destruct} for -% robust proof scripts. + The arguments of the $i^{th}$ constructor and the + equalities that {\tt inversion} introduces in the context of the + goal corresponding to the $i^{th}$ constructor, if it exists, get + their names from the list $p_{i1}$ \ldots $p_{in_i}$ in order. If + there are not enough names, {\tt induction} invents names for the + remaining variables to introduce. In case an equation splits into + several equations (because {\tt inversion} applies {\tt injection} + on the equalities it generates), the corresponding name $p_{ij}$ in + the list must be replaced by a sublist of the form {\tt [$p_{ij1}$ + \ldots $p_{ijq}$]} (or, equivalently, {\tt ($p_{ij1}$, + \ldots, $p_{ijq}$)}) where $q$ is the number of subequalities + obtained from splitting the original equation. Here is an example. -\item{\tt destruct {\term} as {\disjconjintropattern} \_eqn} +\begin{coq_eval} +Require Import List. +\end{coq_eval} - This behaves as {\tt destruct {\term}} but adds an equation between - {\term} and the value that {\term} takes in each of the possible - cases. The name of the equation is chosen by Coq. If - {\disjconjintropattern} is simply {\tt []}, it is automatically considered - as a disjunctive pattern of the appropriate size. +\begin{coq_example} +Inductive contains0 : list nat -> Prop := + | in_hd : forall l, contains0 (0 :: l) + | in_tl : forall l b, contains0 l -> contains0 (b :: l). +Goal forall l:list nat, contains0 (1 :: l) -> contains0 l. +intros l H; inversion H as [ | l' p Hl' [Heqp Heql'] ]. +\end{coq_example} -\item{\tt destruct {\term} as {\disjconjintropattern} \_eqn: {\namingintropattern}} +\begin{coq_eval} +Abort. +\end{coq_eval} - This behaves as {\tt destruct {\term} as - {\disjconjintropattern} \_eqn} but use {\namingintropattern} to - name the equation (see Section~\ref{intros-pattern}). Note that spaces - can generally be removed around {\tt \_eqn}. +\item \texttt{inversion {\num} as \intropattern} -\item{\tt destruct {\term} with \bindinglist} + This allows to name the hypotheses introduced by + \texttt{inversion \num} in the context. - This behaves like \texttt{destruct {\term}} providing explicit - instances for the dependent premises of the type of {\term} (see - syntax of bindings in Section~\ref{Binding-list}). +\item \tacindex{inversion\_clear \dots\ as} \texttt{inversion\_clear + {\ident} as \intropattern} -\item{\tt edestruct {\term}\tacindex{edestruct}} + This allows to name the hypotheses introduced by + \texttt{inversion\_clear} in the context. - This tactic behaves like \texttt{destruct {\term}} excepts that it - does not fail if the instance of a dependent premises of the type of - {\term} is not inferable. Instead, the unresolved instances are left - as existential variables to be inferred later, in the same way as - {\tt eapply} does (see Section~\ref{eapply-example}). +\item \tacindex{inversion \dots\ in} \texttt{inversion {\ident} + in \ident$_1$ \dots\ \ident$_n$} -\item{\tt destruct {\term$_1$} using {\term$_2$}}\\ - {\tt destruct {\term$_1$} using {\term$_2$} with {\bindinglist}} + Let \ident$_1$ \dots\ \ident$_n$, be identifiers in the local context. This + tactic behaves as generalizing \ident$_1$ \dots\ \ident$_n$, and + then performing \texttt{inversion}. - These are synonyms of {\tt induction {\term$_1$} using {\term$_2$}} and - {\tt induction {\term$_1$} using {\term$_2$} with {\bindinglist}}. +\item \tacindex{inversion \dots\ as \dots\ in} \texttt{inversion + {\ident} as {\intropattern} in \ident$_1$ \dots\ + \ident$_n$} -\item \texttt{destruct {\term} in {\occgoalset}} + This allows to name the hypotheses introduced in the context by + \texttt{inversion {\ident} in \ident$_1$ \dots\ \ident$_n$}. - This syntax is used for selecting which occurrences of {\term} the - case analysis has to be done on. The {\tt in {\occgoalset}} clause is an - occurrence clause whose syntax and behavior is described in - Section~\ref{Occurrences clauses}. +\item \tacindex{inversion\_clear \dots\ in} \texttt{inversion\_clear + {\ident} in \ident$_1$ \dots\ \ident$_n$} - When an occurrence clause is given, an equation between {\term} and - the value it gets in each case of the analysis is added to the - context of the subgoals corresponding to the cases (even - if no clause {\tt as {\namingintropattern}} is given). + Let \ident$_1$ \dots\ \ident$_n$, be identifiers in the local context. This + tactic behaves as generalizing \ident$_1$ \dots\ \ident$_n$, and + then performing {\tt inversion\_clear}. -\item{\tt destruct {\term$_1$} with {\bindinglist$_1$} as {\disjconjintropattern} \_eqn: {\namingintropattern} using {\term$_2$} with {\bindinglist$_2$} in {\occgoalset}}\\ - {\tt edestruct {\term$_1$} with {\bindinglist$_1$} as {\disjconjintropattern} \_eqn: {\namingintropattern} using {\term$_2$} with {\bindinglist$_2$} in {\occgoalset}} +\item \tacindex{inversion\_clear \dots\ as \dots\ in} + \texttt{inversion\_clear {\ident} as {\intropattern} + in \ident$_1$ \dots\ \ident$_n$} - These are the general forms of {\tt destruct} and {\tt edestruct}. - They combine the effects of the {\tt with}, {\tt as}, {\tt using}, - and {\tt in} clauses. + This allows to name the hypotheses introduced in the context by + \texttt{inversion\_clear {\ident} in \ident$_1$ \dots\ \ident$_n$}. -\item{\tt case \term}\label{case}\tacindex{case} - - The tactic {\tt case} is a more basic tactic to perform case - analysis without recursion. It behaves as {\tt elim \term} but using - a case-analysis elimination principle and not a recursive one. +\item \tacindex{dependent inversion} \texttt{dependent inversion \ident} -\item{\tt case\_eq \term}\label{case_eq}\tacindex{case\_eq} + That must be used when \ident\ appears in the current goal. It acts + like \texttt{inversion} and then substitutes \ident\ for the + corresponding term in the goal. - The tactic {\tt case\_eq} is a variant of the {\tt case} tactic that - allow to perform case analysis on a term without completely - forgetting its original form. This is done by generating equalities - between the original form of the term and the outcomes of the case - analysis. The effect of this tactic is similar to the effect of {\tt - destruct {\term} in |- *} with the exception that no new hypotheses - are introduced in the context. +\item \tacindex{dependent inversion \dots\ as } \texttt{dependent + inversion {\ident} as \intropattern} -\item {\tt case {\term} with {\bindinglist}} + This allows to name the hypotheses introduced in the context by + \texttt{dependent inversion} {\ident}. - Analogous to {\tt elim {\term} with {\bindinglist}} above. +\item \tacindex{dependent inversion\_clear} \texttt{dependent + inversion\_clear \ident} -\item{\tt ecase {\term}\tacindex{ecase}}\\ - {\tt ecase {\term} with {\bindinglist}} - - In case the type of {\term} has dependent premises, or dependent - premises whose values are not inferable from the {\tt with - {\bindinglist}} clause, {\tt ecase} turns them into existential - variables to be resolved later on. + Like \texttt{dependent inversion}, except that {\ident} is cleared + from the local context. -\item {\tt simple destruct \ident}\tacindex{simple destruct} - - This tactic behaves as {\tt intros until - {\ident}; case {\tt {\ident}}} when {\ident} is a quantified - variable of the goal. +\item \tacindex{dependent inversion\_clear \dots\ as} + \texttt{dependent inversion\_clear {\ident} as \intropattern} -\item {\tt simple destruct {\num}} - - This tactic behaves as {\tt intros until - {\num}; case {\tt {\ident}}} where {\ident} is the name given by - {\tt intros until {\num}} to the {\num}-th non-dependent premise of - the goal. + This allows to name the hypotheses introduced in the context by + \texttt{dependent inversion\_clear} {\ident}. + +\item \tacindex{dependent inversion \dots\ with} \texttt{dependent + inversion {\ident} with \term} + + This variant allows you to specify the generalization of the goal. It + is useful when the system fails to generalize the goal automatically. If + {\ident} has type $(I~\vec{t})$ and $I$ has type + $\forall (\vec{x}:\vec{T}), s$, then \term~ must be of type + $I:\forall (\vec{x}:\vec{T}), I~\vec{x}\to s'$ where $s'$ is the + type of the goal. + +\item \tacindex{dependent inversion \dots\ as \dots\ with} + \texttt{dependent inversion {\ident} as {\intropattern} + with \term} + + This allows to name the hypotheses introduced in the context by + \texttt{dependent inversion {\ident} with \term}. + +\item \tacindex{dependent inversion\_clear \dots\ with} + \texttt{dependent inversion\_clear {\ident} with \term} + + Like \texttt{dependent inversion \dots\ with} but clears {\ident} from + the local context. + +\item \tacindex{dependent inversion\_clear \dots\ as \dots\ with} + \texttt{dependent inversion\_clear {\ident} as + {\intropattern} with \term} + + This allows to name the hypotheses introduced in the context by + \texttt{dependent inversion\_clear {\ident} with \term}. + +\item \tacindex{simple inversion} \texttt{simple inversion \ident} + It is a very primitive inversion tactic that derives all the necessary + equalities but it does not simplify the constraints as + \texttt{inversion} does. + +\item \tacindex{simple inversion \dots\ as} \texttt{simple inversion + {\ident} as \intropattern} + + This allows to name the hypotheses introduced in the context by + \texttt{simple inversion}. + +\item \tacindex{inversion \dots\ using} \texttt{inversion {\ident} + using \ident$'$} + + Let {\ident} have type $(I~\vec{t})$ ($I$ an inductive + predicate) in the local context, and \ident$'$ be a (dependent) inversion + lemma. Then, this tactic refines the current goal with the specified + lemma. + +\item \tacindex{inversion \dots\ using \dots\ in} \texttt{inversion + {\ident} using \ident$'$ in \ident$_1$\dots\ \ident$_n$} + + This tactic behaves as generalizing \ident$_1$\dots\ \ident$_n$, + then doing \texttt{inversion {\ident} using \ident$'$}. \end{Variants} -\subsection{\tt intros {\intropattern} {\ldots} {\intropattern} -\label{intros-pattern} -\tacindex{intros \intropattern}} -\index{Introduction patterns} -\index{Naming introduction patterns} -\index{Disjunctive/conjunctive introduction patterns} +\firstexample +\example{Non-dependent inversion} +\label{inversion-examples} -This extension of the tactic {\tt intros} combines introduction of -variables or hypotheses and case analysis. An {\em introduction pattern} is -either: -\begin{itemize} -\item A {\em naming introduction pattern}, i.e. either one of: - \begin{itemize} - \item the pattern \texttt{?} - \item the pattern \texttt{?\ident} - \item an identifier - \end{itemize} -\item A {\em disjunctive/conjunctive introduction pattern}, i.e. either one of: - \begin{itemize} - \item a disjunction of lists of patterns: - {\tt [$p_{11}$ {\ldots} $p_{1m_1}$ | {\ldots} | $p_{11}$ {\ldots} $p_{nm_n}$]} - \item a conjunction of patterns: {\tt (} $p_1$ {\tt ,} {\ldots} {\tt ,} $p_n$ {\tt )} - \item a list of patterns {\tt (} $p_1$\ {\tt \&}\ {\ldots}\ {\tt \&}\ $p_n$ {\tt )} - for sequence of right-associative binary constructs - \end{itemize} -\item the wildcard: {\tt \_} -\item the rewriting orientations: {\tt ->} or {\tt <-} -\end{itemize} +Let us consider the relation \texttt{Le} over natural numbers and the +following variables: -Assuming a goal of type {\tt $Q$ -> $P$} (non dependent product), or -of type {\tt forall $x$:$T$, $P$} (dependent product), the behavior of -{\tt intros $p$} is defined inductively over the structure of the -introduction pattern $p$: -\begin{itemize} -\item introduction on \texttt{?} performs the introduction, and lets {\Coq} - choose a fresh name for the variable; -\item introduction on \texttt{?\ident} performs the introduction, and - lets {\Coq} choose a fresh name for the variable based on {\ident}; -\item introduction on \texttt{\ident} behaves as described in - Section~\ref{intro}; -\item introduction over a disjunction of list of patterns {\tt - [$p_{11}$ {\ldots} $p_{1m_1}$ | {\ldots} | $p_{11}$ {\ldots} - $p_{nm_n}$]} expects the product to be over an inductive type - whose number of constructors is $n$ (or more generally over a type - of conclusion an inductive type built from $n$ constructors, - e.g. {\tt C -> A$\backslash$/B if $n=2$}): it destructs the introduced - hypothesis as {\tt destruct} (see Section~\ref{destruct}) would and - applies on each generated subgoal the corresponding tactic; - \texttt{intros}~$p_{i1}$ {\ldots} $p_{im_i}$; if the disjunctive - pattern is part of a sequence of patterns and is not the last - pattern of the sequence, then {\Coq} completes the pattern so as all - the argument of the constructors of the inductive type are - introduced (for instance, the list of patterns {\tt [$\;$|$\;$] H} - applied on goal {\tt forall x:nat, x=0 -> 0=x} behaves the same as - the list of patterns {\tt [$\,$|$\,$?$\,$] H}); -\item introduction over a conjunction of patterns {\tt ($p_1$, \ldots, - $p_n$)} expects the goal to be a product over an inductive type $I$ with a - single constructor that itself has at least $n$ arguments: it - performs a case analysis over the hypothesis, as {\tt destruct} - would, and applies the patterns $p_1$~\ldots~$p_n$ to the arguments - of the constructor of $I$ (observe that {\tt ($p_1$, {\ldots}, - $p_n$)} is an alternative notation for {\tt [$p_1$ {\ldots} - $p_n$]}); -\item introduction via {\tt ( $p_1$ \& \ldots \& $p_n$ )} - is a shortcut for introduction via - {\tt ($p_1$,(\ldots,(\dots,$p_n$)\ldots))}; it expects the - hypothesis to be a sequence of right-associative binary inductive - constructors such as {\tt conj} or {\tt ex\_intro}; for instance, an - hypothesis with type {\tt A\verb|/\|exists x, B\verb|/\|C\verb|/\|D} can be - introduced via pattern {\tt (a \& x \& b \& c \& d)}; -\item introduction on the wildcard depends on whether the product is - dependent or not: in the non dependent case, it erases the - corresponding hypothesis (i.e. it behaves as an {\tt intro} followed - by a {\tt clear}, cf Section~\ref{clear}) while in the dependent - case, it succeeds and erases the variable only if the wildcard is - part of a more complex list of introduction patterns that also - erases the hypotheses depending on this variable; -\item introduction over {\tt ->} (respectively {\tt <-}) expects the - hypothesis to be an equality and the right-hand-side (respectively - the left-hand-side) is replaced by the left-hand-side (respectively - the right-hand-side) in both the conclusion and the context of the goal; - if moreover the term to substitute is a variable, the hypothesis is - removed. -\end{itemize} +\begin{coq_eval} +Reset Initial. +\end{coq_eval} -\Rem {\tt intros $p_1~\ldots~p_n$} is not equivalent to \texttt{intros - $p_1$;\ldots; intros $p_n$} for the following reasons: -\begin{itemize} -\item A wildcard pattern never succeeds when applied isolated on a - dependent product, while it succeeds as part of a list of - introduction patterns if the hypotheses that depends on it are - erased too. -\item A disjunctive or conjunctive pattern followed by an introduction - pattern forces the introduction in the context of all arguments of - the constructors before applying the next pattern while a terminal - disjunctive or conjunctive pattern does not. Here is an example +\begin{coq_example*} +Inductive Le : nat -> nat -> Set := + | LeO : forall n:nat, Le 0 n + | LeS : forall n m:nat, Le n m -> Le (S n) (S m). +Variable P : nat -> nat -> Prop. +Variable Q : forall n m:nat, Le n m -> Prop. +\end{coq_example*} -\begin{coq_example} -Goal forall n:nat, n = 0 -> n = 0. -intros [ | ] H. -Show 2. -Undo. -intros [ | ]; intros H. -Show 2. -\end{coq_example} +Let us consider the following goal: -\end{itemize} +\begin{coq_eval} +Lemma ex : forall n m:nat, Le (S n) m -> P n m. +intros. +\end{coq_eval} \begin{coq_example} -Lemma intros_test : forall A B C:Prop, A \/ B /\ C -> (A -> C) -> C. -intros A B C [a| [_ c]] f. -apply (f a). -exact c. -Qed. +Show. \end{coq_example} -%\subsection[\tt FixPoint \dots]{\tt FixPoint \dots\tacindex{Fixpoint}} -%Not yet documented. - -\subsection{\tt double induction \ident$_1$ \ident$_2$} -%\tacindex{double induction}} -This tactic is deprecated and should be replaced by {\tt induction \ident$_1$; induction \ident$_2$} (or {\tt induction \ident$_1$; destruct \ident$_2$} depending on the exact needs). - -%% This tactic applies to any goal. If the variables {\ident$_1$} and -%% {\ident$_2$} of the goal have an inductive type, then this tactic -%% performs double induction on these variables. For instance, if the -%% current goal is \verb+forall n m:nat, P n m+ then, {\tt double induction n -%% m} yields the four cases with their respective inductive hypotheses. - -%% In particular, for proving \verb+(P (S n) (S m))+, the generated induction -%% hypotheses are \verb+(P (S n) m)+ and \verb+(m:nat)(P n m)+ (of the latter, -%% \verb+(P n m)+ and \verb+(P n (S m))+ are derivable). - -%% \Rem When the induction hypothesis \verb+(P (S n) m)+ is not -%% needed, {\tt induction \ident$_1$; destruct \ident$_2$} produces -%% more concise subgoals. +To prove the goal, we may need to reason by cases on \texttt{H} and to +derive that \texttt{m} is necessarily of +the form $(S~m_0)$ for certain $m_0$ and that $(Le~n~m_0)$. +Deriving these conditions corresponds to prove that the +only possible constructor of \texttt{(Le (S n) m)} is +\texttt{LeS} and that we can invert the +\texttt{->} in the type of \texttt{LeS}. +This inversion is possible because \texttt{Le} is the smallest set closed by +the constructors \texttt{LeO} and \texttt{LeS}. -\begin{Variant} +\begin{coq_example} +inversion_clear H. +\end{coq_example} -\item {\tt double induction \num$_1$ \num$_2$} +Note that \texttt{m} has been substituted in the goal for \texttt{(S m0)} +and that the hypothesis \texttt{(Le n m0)} has been added to the +context. -This tactic is deprecated and should be replaced by {\tt induction - \num$_1$; induction \num$_3$} where \num$_3$ is the result of -\num$_2$-\num$_1$. +Sometimes it is +interesting to have the equality \texttt{m=(S m0)} in the +context to use it after. In that case we can use \texttt{inversion} that +does not clear the equalities: -%% This tactic applies to any goal. If the variables {\ident$_1$} and +\begin{coq_eval} +Undo. +\end{coq_eval} -%% This applies double induction on the \num$_1^{th}$ and \num$_2^{th}$ {\it -%% non dependent} premises of the goal. More generally, any combination of an -%% {\ident} and a {\num} is valid. +\begin{coq_example} +inversion H. +\end{coq_example} -\end{Variant} +\begin{coq_eval} +Abort. +\end{coq_eval} -\subsection{\tt dependent induction \ident - \tacindex{dependent induction} - \label{DepInduction}} +\example{Dependent inversion} -The \emph{experimental} tactic \texttt{dependent induction} performs -induction-inversion on an instantiated inductive predicate. -One needs to first require the {\tt Coq.Program.Equality} module to use -this tactic. The tactic is based on the BasicElim tactic by Conor -McBride \cite{DBLP:conf/types/McBride00} and the work of Cristina Cornes -around inversion \cite{DBLP:conf/types/CornesT95}. From an instantiated -inductive predicate and a goal it generates an equivalent goal where the -hypothesis has been generalized over its indexes which are then -constrained by equalities to be the right instances. This permits to -state lemmas without resorting to manually adding these equalities and -still get enough information in the proofs. -A simple example is the following: +Let us consider the following goal: \begin{coq_eval} -Reset Initial. +Lemma ex_dep : forall (n m:nat) (H:Le (S n) m), Q (S n) m H. +intros. \end{coq_eval} + \begin{coq_example} -Lemma le_minus : forall n:nat, n < 1 -> n = 0. -intros n H ; induction H. +Show. \end{coq_example} -Here we didn't get any information on the indexes to help fulfill this -proof. The problem is that when we use the \texttt{induction} tactic -we lose information on the hypothesis instance, notably that the second -argument is \texttt{1} here. Dependent induction solves this problem by -adding the corresponding equality to the context. +As \texttt{H} occurs in the goal, we may want to reason by cases on its +structure and so, we would like inversion tactics to +substitute \texttt{H} by the corresponding term in constructor form. +Neither \texttt{Inversion} nor {\tt Inversion\_clear} make such a +substitution. +To have such a behavior we use the dependent inversion tactics: -\begin{coq_eval} -Reset Initial. -\end{coq_eval} \begin{coq_example} -Require Import Coq.Program.Equality. -Lemma le_minus : forall n:nat, n < 1 -> n = 0. -intros n H ; dependent induction H. +dependent inversion_clear H. \end{coq_example} -The subgoal is cleaned up as the tactic tries to automatically -simplify the subgoals with respect to the generated equalities. -In this enriched context it becomes possible to solve this subgoal. -\begin{coq_example} -reflexivity. -\end{coq_example} +Note that \texttt{H} has been substituted by \texttt{(LeS n m0 l)} and +\texttt{m} by \texttt{(S m0)}. -Now we are in a contradictory context and the proof can be solved. -\begin{coq_example} -inversion H. -\end{coq_example} +\subsection{\tt fix {\ident} {\num}} +\tacindex{fix} +\label{tactic:fix} -This technique works with any inductive predicate. -In fact, the \texttt{dependent induction} tactic is just a wrapper around -the \texttt{induction} tactic. One can make its own variant by just -writing a new tactic based on the definition found in -\texttt{Coq.Program.Equality}. Common useful variants are the following, -defined in the same file: +This tactic is a primitive tactic to start a proof by induction. In +general, it is easier to rely on higher-level induction tactics such +as the ones described in Section~\ref{Tac-induction}. + +In the syntax of the tactic, the identifier {\ident} is the name given +to the induction hypothesis. The natural number {\num} tells on which +premise of the current goal the induction acts, starting +from 1 and counting both dependent and non dependent +products. Especially, the current lemma must be composed of at least +{\num} products. + +Like in a {\tt fix} expression, the induction +hypotheses have to be used on structurally smaller arguments. +The verification that inductive proof arguments are correct is done +only at the time of registering the lemma in the environment. To know +if the use of induction hypotheses is correct at some +time of the interactive development of a proof, use the command {\tt + Guarded} (see Section~\ref{Guarded}). \begin{Variants} -\item {\tt dependent induction {\ident} generalizing {\ident$_1$} \dots - {\ident$_n$}}\tacindex{dependent induction \dots\ generalizing} - - Does dependent induction on the hypothesis {\ident} but first - generalizes the goal by the given variables so that they are - universally quantified in the goal. This is generally what one wants - to do with the variables that are inside some constructors in the - induction hypothesis. The other ones need not be further generalized. + \item {\tt fix \ident$_1$ {\num} with ( \ident$_2$ + \nelist{\binder$_2$}{} \zeroone{\{ struct \ident$'_2$ + \}} :~\type$_2$ ) \dots\ ( \ident$_n$ + \nelist{\binder$_n$}{} \zeroone{\{ struct \ident$'_n$ \}} :~\type$_n$ )} + +This starts a proof by mutual induction. The statements to be +simultaneously proved are respectively {\tt forall} + \nelist{{\binder}$_2$}{}{\tt ,} {\type}$_2$, {\ldots}, {\tt forall} + \nelist{{\binder}$_n$}{}{\tt ,} {\type}$_n$. The identifiers +{\ident}$_1$ {\ldots} {\ident}$_n$ are the names of the induction +hypotheses. The identifiers {\ident}$'_2$ {\ldots} {\ident}$'_n$ are the +respective names of the premises on which the induction is performed +in the statements to be simultaneously proved (if not given, the +system tries to guess itself what they are). -\item {\tt dependent destruction {\ident}}\tacindex{dependent destruction} - - Does the generalization of the instance {\ident} but uses {\tt destruct} - instead of {\tt induction} on the generalized hypothesis. This gives - results equivalent to {\tt inversion} or {\tt dependent inversion} if - the hypothesis is dependent. \end{Variants} -A larger example of dependent induction and an explanation of the -underlying technique are developed in section~\ref{dependent-induction-example}. +\subsection{\tt cofix \ident} +\tacindex{cofix} +\label{tactic:cofix} -\subsection{\tt decompose [ {\qualid$_1$} \dots\ {\qualid$_n$} ] \term -\label{decompose} -\tacindex{decompose}} +This tactic starts a proof by coinduction. The identifier {\ident} is +the name given to the coinduction hypothesis. Like in a {\tt cofix} +expression, the use of induction hypotheses have to guarded by a +constructor. The verification that the use of co-inductive hypotheses +is correct is done only at the time of registering the lemma in the +environment. To know if the use of coinduction hypotheses is correct +at some time of the interactive development of a proof, use the +command {\tt Guarded} (see Section~\ref{Guarded}). -This tactic allows to recursively decompose a -complex proposition in order to obtain atomic ones. -Example: -\begin{coq_eval} -Reset Initial. -\end{coq_eval} -\begin{coq_example} -Lemma ex1 : forall A B C:Prop, A /\ B /\ C \/ B /\ C \/ C /\ A -> C. -intros A B C H; decompose [and or] H; assumption. -\end{coq_example} -\begin{coq_example*} -Qed. -\end{coq_example*} +\begin{Variants} + \item {\tt cofix \ident$_1$ with ( \ident$_2$ + \nelist{\binder$_2$}{} :~\type$_2$ ) \dots\ ( + \ident$_n$ \nelist{\binder$_n$}{} :~\type$_n$ )} -{\tt decompose} does not work on right-hand sides of implications or products. +This starts a proof by mutual coinduction. The statements to be +simultaneously proved are respectively {\tt forall} +\nelist{{\binder}$_2$}{}{\tt ,} {\type}$_2$, {\ldots}, {\tt forall} + \nelist{{\binder}$_n$}{}{\tt ,} {\type}$_n$. The identifiers + {\ident}$_1$ {\ldots} {\ident}$_n$ are the names of the + coinduction hypotheses. -\begin{Variants} - -\item {\tt decompose sum \term}\tacindex{decompose sum} - This decomposes sum types (like \texttt{or}). -\item {\tt decompose record \term}\tacindex{decompose record} - This decomposes record types (inductive types with one constructor, - like \texttt{and} and \texttt{exists} and those defined with the - \texttt{Record} macro, see Section~\ref{Record}). \end{Variants} +\section{Rewriting expressions} -\subsection{\tt functional induction (\qualid\ \term$_1$ \dots\ \term$_n$). -\tacindex{functional induction} -\label{FunInduction}} - -The \emph{experimental} tactic \texttt{functional induction} performs -case analysis and induction following the definition of a function. It -makes use of a principle generated by \texttt{Function} -(see Section~\ref{Function}) or \texttt{Functional Scheme} -(see Section~\ref{FunScheme}). -\begin{coq_eval} -Reset Initial. -\end{coq_eval} -\begin{coq_example} -Functional Scheme minus_ind := Induction for minus Sort Prop. +These tactics use the equality {\tt eq:forall A:Type, A->A->Prop} +defined in file {\tt Logic.v} (see Section~\ref{Equality}). The +notation for {\tt eq}~$T~t~u$ is simply {\tt $t$=$u$} dropping the +implicit type of $t$ and $u$. -Lemma le_minus : forall n m:nat, (n - m <= n). -intros n m. -functional induction (minus n m); simpl; auto. -\end{coq_example} -\begin{coq_example*} -Qed. -\end{coq_example*} +\subsection{\tt rewrite \term +\label{rewrite} +\tacindex{rewrite}} -\Rem \texttt{(\qualid\ \term$_1$ \dots\ \term$_n$)} must be a correct -full application of \qualid. In particular, the rules for implicit -arguments are the same as usual. For example use \texttt{@\qualid} if -you want to write implicit arguments explicitly. +This tactic applies to any goal. The type of {\term} +must have the form -\Rem Parenthesis over \qualid \dots \term$_n$ are mandatory. +\texttt{forall (x$_1$:A$_1$) \dots\ (x$_n$:A$_n$)}\texttt{eq} \term$_1$ \term$_2$. -\Rem \texttt{functional induction (f x1 x2 x3)} is actually a wrapper -for \texttt{induction x1 x2 x3 (f x1 x2 x3) using \qualid} followed by -a cleaning phase, where $\qualid$ is the induction principle -registered for $f$ (by the \texttt{Function} (see Section~\ref{Function}) -or \texttt{Functional Scheme} (see Section~\ref{FunScheme}) command) -corresponding to the sort of the goal. Therefore \texttt{functional - induction} may fail if the induction scheme (\texttt{\qualid}) is -not defined. See also Section~\ref{Function} for the function terms -accepted by \texttt{Function}. - -\Rem There is a difference between obtaining an induction scheme for a -function by using \texttt{Function} (see Section~\ref{Function}) and by -using \texttt{Functional Scheme} after a normal definition using -\texttt{Fixpoint} or \texttt{Definition}. See \ref{Function} for -details. - -\SeeAlso{\ref{Function},\ref{FunScheme},\ref{FunScheme-examples}, - \ref{sec:functional-inversion}} - -\begin{ErrMsgs} -\item \errindex{Cannot find induction information on \qualid} - - ~ - -\item \errindex{Not the right number of induction arguments} -\end{ErrMsgs} - -\begin{Variants} -\item {\tt functional induction (\qualid\ \term$_1$ \dots\ \term$_n$) - using \term$_{m+1}$ with {\term$_{n+1}$} \dots {\term$_m$}} - - Similar to \texttt{Induction} and \texttt{elim} - (see Section~\ref{Tac-induction}), allows to give explicitly the - induction principle and the values of dependent premises of the - elimination scheme, including \emph{predicates} for mutual induction - when {\qualid} is part of a mutually recursive definition. - -\item {\tt functional induction (\qualid\ \term$_1$ \dots\ \term$_n$) - using \term$_{m+1}$ with {\vref$_1$} := {\term$_{n+1}$} \dots\ - {\vref$_m$} := {\term$_n$}} - - Similar to \texttt{induction} and \texttt{elim} - (see Section~\ref{Tac-induction}). - -\item All previous variants can be extended by the usual \texttt{as - \intropattern} construction, similar for example to - \texttt{induction} and \texttt{elim} (see Section~\ref{Tac-induction}). - -\end{Variants} - - - -\section{Equality} - -These tactics use the equality {\tt eq:forall A:Type, A->A->Prop} -defined in file {\tt Logic.v} (see Section~\ref{Equality}). The -notation for {\tt eq}~$T~t~u$ is simply {\tt $t$=$u$} dropping the -implicit type of $t$ and $u$. - -\subsection{\tt rewrite \term -\label{rewrite} -\tacindex{rewrite}} - -This tactic applies to any goal. The type of {\term} -must have the form - -\texttt{forall (x$_1$:A$_1$) \dots\ (x$_n$:A$_n$)}\texttt{eq} \term$_1$ \term$_2$. - -\noindent where \texttt{eq} is the Leibniz equality or a registered -setoid equality. +\noindent where \texttt{eq} is the Leibniz equality or a registered +setoid equality. \noindent Then {\tt rewrite \term} finds the first subterm matching \term$_1$ in the goal, resulting in instances \term$_1'$ and \term$_2'$ @@ -2419,7 +2504,7 @@ Hence, some of the variables x$_i$ are solved by unification, and some of the types \texttt{A}$_1$, \dots, \texttt{A}$_n$ become new subgoals. -% \Rem In case the type of +% \Rem In case the type of % \term$_1$ contains occurrences of variables bound in the % type of \term, the tactic tries first to find a subterm of the goal % which matches this term in order to find a closed instance \term$'_1$ @@ -2449,13 +2534,13 @@ This happens if \term$_1$ does not occur in the goal. \texttt{H1} instead of the current goal. \item \texttt{rewrite H in H1 at 1, H2 at - 2 |- *} means \texttt{rewrite H; rewrite H in H1 at 1; rewrite H in H2 at - 2}. In particular a failure will happen if any of - these three simpler tactics fails. + these three simpler tactics fails. \item \texttt{rewrite H in * |- } will do \texttt{rewrite H in H$_i$} for all hypothesis \texttt{H$_i$ <> H}. A success will happen as soon as at least one of these simpler tactics succeeds. - \item \texttt{rewrite H in *} is a combination of \texttt{rewrite H} + \item \texttt{rewrite H in *} is a combination of \texttt{rewrite H} and \texttt{rewrite H in * |-} that succeeds if at - least one of these two tactics succeeds. + least one of these two tactics succeeds. \end{itemize} Orientation {\tt ->} or {\tt <-} can be inserted before the term to rewrite. @@ -2466,7 +2551,7 @@ This happens if \term$_1$ does not occur in the goal. Rewrite only the given occurrences of \term$_1'$. Occurrences are specified from left to right as for \texttt{pattern} (\S \ref{pattern}). The rewrite is always performed using setoid - rewriting, even for Leibniz's equality, so one has to + rewriting, even for Leibniz's equality, so one has to \texttt{Import Setoid} to use this variant. \item {\tt rewrite {\term} by {\tac}} @@ -2481,7 +2566,7 @@ This happens if \term$_1$ does not occur in the goal. generated by the previous one. Orientation {\tt ->} or {\tt <-} can be inserted before each term to rewrite. One unique \textit{clause} - can be added at the end after the keyword {\tt in}; it will + can be added at the end after the keyword {\tt in}; it will then affect all rewrite operations. \item In all forms of {\tt rewrite} described above, a term to rewrite @@ -2489,13 +2574,13 @@ This happens if \term$_1$ does not occur in the goal. \begin{itemize} \item {\tt ?} : the tactic {\tt rewrite ?$\term$} performs the rewrite of $\term$ as many times as possible (perhaps zero time). - This form never fails. - \item {\tt $n$?} : works similarly, except that it will do at most - $n$ rewrites. - \item {\tt !} : works as {\tt ?}, except that at least one rewrite - should succeed, otherwise the tactic fails. - \item {\tt $n$!} (or simply {\tt $n$}) : precisely $n$ rewrites - of $\term$ will be done, leading to failure if these $n$ rewrites are not possible. + This form never fails. + \item {\tt $n$?} : works similarly, except that it will do at most + $n$ rewrites. + \item {\tt !} : works as {\tt ?}, except that at least one rewrite + should succeed, otherwise the tactic fails. + \item {\tt $n$!} (or simply {\tt $n$}) : precisely $n$ rewrites + of $\term$ will be done, leading to failure if these $n$ rewrites are not possible. \end{itemize} \item {\tt erewrite {\term}\tacindex{erewrite}} @@ -2542,10 +2627,10 @@ n}| assumption || symmetry; try assumption]}. \item {\tt replace <- {\term}}\\ Replace {\term} with {\term'} using the first assumption whose type has the form {\tt \term'=\term} \item {\tt replace {\term$_1$} with {\term$_2$} \textit{clause} }\\ - {\tt replace {\term$_1$} with {\term$_2$} \textit{clause} by \tac }\\ - {\tt replace {\term} \textit{clause}}\\ - {\tt replace -> {\term} \textit{clause}}\\ - {\tt replace <- {\term} \textit{clause}}\\ + {\tt replace {\term$_1$} with {\term$_2$} \textit{clause} by \tac }\\ + {\tt replace {\term} \textit{clause}}\\ + {\tt replace -> {\term} \textit{clause}}\\ + {\tt replace <- {\term} \textit{clause}}\\ Act as before but the replacements take place in \textit{clause}~(see Section~\ref{Conversion-tactics}) and not only in the conclusion of the goal.\\ @@ -2578,7 +2663,7 @@ the tactic changes it to {\tt u=t}. \subsection{\tt transitivity \term \tacindex{transitivity}} This tactic applies to a goal which has the form {\tt t=u} -and transforms it into the two subgoals +and transforms it into the two subgoals {\tt t={\term}} and {\tt {\term}=u}. \subsection{\tt subst {\ident} @@ -2586,13 +2671,13 @@ and transforms it into the two subgoals This tactic applies to a goal which has \ident\ in its context and (at least) one hypothesis, say {\tt H}, of type {\tt - \ident=t} or {\tt t=\ident}. Then it replaces -\ident\ by {\tt t} everywhere in the goal (in the hypotheses + \ident=t} or {\tt t=\ident}. Then it replaces +\ident\ by {\tt t} everywhere in the goal (in the hypotheses and in the conclusion) and clears \ident\ and {\tt H} from the context. -\Rem +\Rem When several hypotheses have the form {\tt \ident=t} or {\tt - t=\ident}, the first one is used. + t=\ident}, the first one is used. \begin{Variants} \item {\tt subst \ident$_1$ \dots \ident$_n$} \\ @@ -2612,7 +2697,7 @@ is typically a setoid equality. The application of {\tt stepl {\term}} then replaces the goal by ``$R$ {\term} {\term}$_2$'' and adds a new goal stating ``$eq$ {\term} {\term}$_1$''. -Lemmas are added to the database using the command +Lemmas are added to the database using the command \comindex{Declare Left Step} \begin{quote} {\tt Declare Left Step {\term}.} @@ -2640,710 +2725,1060 @@ and are registered using the command \end{quote} \end{Variants} +\subsection{\tt change \term +\tacindex{change} +\label{change}} -\subsection{\tt f\_equal -\label{f-equal} -\tacindex{f\_equal}} +This tactic applies to any goal. It implements the rule +``Conv''\index{Typing rules!Conv} given in Section~\ref{Conv}. {\tt + change U} replaces the current goal \T\ with \U\ providing that +\U\ is well-formed and that \T\ and \U\ are convertible. -This tactic applies to a goal of the form $f\ a_1\ \ldots\ a_n = f'\ -a'_1\ \ldots\ a'_n$. Using {\tt f\_equal} on such a goal leads to -subgoals $f=f'$ and $a_1=a'_1$ and so on up to $a_n=a'_n$. Amongst -these subgoals, the simple ones (e.g. provable by -reflexivity or congruence) are automatically solved by {\tt f\_equal}. +\begin{ErrMsgs} +\item \errindex{Not convertible} +\end{ErrMsgs} +\tacindex{change \dots\ in} +\begin{Variants} +\item {\tt change \term$_1$ with \term$_2$} -\section{Equality and inductive sets} + This replaces the occurrences of \term$_1$ by \term$_2$ in the + current goal. The terms \term$_1$ and \term$_2$ must be + convertible. -We describe in this section some special purpose tactics dealing with -equality and inductive sets or types. These tactics use the equality -{\tt eq:forall (A:Type), A->A->Prop}, simply written with the -infix symbol {\tt =}. +\item {\tt change \term$_1$ at \num$_1$ \dots\ \num$_i$ with \term$_2$} -\subsection{\tt decide equality -\label{decideequality} -\tacindex{decide equality}} + This replaces the occurrences numbered \num$_1$ \dots\ \num$_i$ of + \term$_1$ by \term$_2$ in the current goal. + The terms \term$_1$ and \term$_2$ must be convertible. -This tactic solves a goal of the form -{\tt forall $x$ $y$:$R$, \{$x$=$y$\}+\{\verb|~|$x$=$y$\}}, where $R$ -is an inductive type such that its constructors do not take proofs or -functions as arguments, nor objects in dependent types. -It solves goals of the form {\tt \{$x$=$y$\}+\{\verb|~|$x$=$y$\}} as well. + \ErrMsg {\tt Too few occurrences} -\subsection{\tt compare \term$_1$ \term$_2$ -\tacindex{compare}} +\item {\tt change {\term} in {\ident}} -This tactic compares two given objects \term$_1$ and \term$_2$ -of an inductive datatype. If $G$ is the current goal, it leaves the sub-goals -\term$_1${\tt =}\term$_2$ {\tt ->} $G$ and \verb|~|\term$_1${\tt =}\term$_2$ -{\tt ->} $G$. The type -of \term$_1$ and \term$_2$ must satisfy the same restrictions as in the tactic -\texttt{decide equality}. +\item {\tt change \term$_1$ with \term$_2$ in {\ident}} -\subsection{\tt discriminate {\term} -\label{discriminate} -\tacindex{discriminate} -\tacindex{ediscriminate}} +\item {\tt change \term$_1$ at \num$_1$ \dots\ \num$_i$ with \term$_2$ in + {\ident}} -This tactic proves any goal from an assumption stating that two -structurally different terms of an inductive set are equal. For -example, from {\tt (S (S O))=(S O)} we can derive by absurdity any -proposition. + This applies the {\tt change} tactic not to the goal but to the + hypothesis {\ident}. -The argument {\term} is assumed to be a proof of a statement -of conclusion {\tt{\term$_1$} = {\term$_2$}} with {\term$_1$} and -{\term$_2$} being elements of an inductive set. To build the proof, -the tactic traverses the normal forms\footnote{Reminder: opaque - constants will not be expanded by $\delta$ reductions} of -{\term$_1$} and {\term$_2$} looking for a couple of subterms {\tt u} -and {\tt w} ({\tt u} subterm of the normal form of {\term$_1$} and -{\tt w} subterm of the normal form of {\term$_2$}), placed at the same -positions and whose head symbols are two different constructors. If -such a couple of subterms exists, then the proof of the current goal -is completed, otherwise the tactic fails. +\end{Variants} -\Rem The syntax {\tt discriminate {\ident}} can be used to refer to a -hypothesis quantified in the goal. In this case, the quantified -hypothesis whose name is {\ident} is first introduced in the local -context using \texttt{intros until \ident}. +\SeeAlso \ref{Conversion-tactics} -\begin{ErrMsgs} -\item \errindex{No primitive equality found} -\item \errindex{Not a discriminable equality} -\end{ErrMsgs} -\begin{Variants} -\item \texttt{discriminate} \num +\section{Performing computations +\index{Conversion tactics} +\label{Conversion-tactics}} - This does the same thing as \texttt{intros until \num} followed by - \texttt{discriminate \ident} where {\ident} is the identifier for - the last introduced hypothesis. +This set of tactics implements different specialized usages of the +tactic \texttt{change}. -\item \texttt{discriminate} {\term} {\tt with} {\bindinglist} +All conversion tactics (including \texttt{change}) can be +parameterized by the parts of the goal where the conversion can +occur. This is done using \emph{goal clauses} which consists in a list +of hypotheses and, optionally, of a reference to the conclusion of the +goal. For defined hypothesis it is possible to specify if the +conversion should occur on the type part, the body part or both +(default). - This does the same thing as \texttt{discriminate {\term}} but using -the given bindings to instantiate parameters or hypotheses of {\term}. +\index{Clauses} +\index{Goal clauses} +Goal clauses are written after a conversion tactic (tactics +\texttt{set}~\ref{tactic:set}, \texttt{rewrite}~\ref{rewrite}, +\texttt{replace}~\ref{tactic:replace} and +\texttt{autorewrite}~\ref{tactic:autorewrite} also use goal clauses) and +are introduced by the keyword \texttt{in}. If no goal clause is provided, +the default is to perform the conversion only in the conclusion. -\item \texttt{ediscriminate} \num\\ - \texttt{ediscriminate} {\term} \zeroone{{\tt with} {\bindinglist}} +The syntax and description of the various goal clauses is the following: +\begin{description} +\item[]\texttt{in {\ident}$_1$ $\ldots$ {\ident}$_n$ |- } only in hypotheses {\ident}$_1$ + \ldots {\ident}$_n$ +\item[]\texttt{in {\ident}$_1$ $\ldots$ {\ident}$_n$ |- *} in hypotheses {\ident}$_1$ \ldots + {\ident}$_n$ and in the conclusion +\item[]\texttt{in * |-} in every hypothesis +\item[]\texttt{in *} (equivalent to \texttt{in * |- *}) everywhere +\item[]\texttt{in (type of {\ident}$_1$) (value of {\ident}$_2$) $\ldots$ |-} in + type part of {\ident}$_1$, in the value part of {\ident}$_2$, etc. +\end{description} - This works the same as {\tt discriminate} but if the type of {\term}, - or the type of the hypothesis referred to by {\num}, has uninstantiated - parameters, these parameters are left as existential variables. +For backward compatibility, the notation \texttt{in}~{\ident}$_1$\ldots {\ident}$_n$ +performs the conversion in hypotheses {\ident}$_1$\ldots {\ident}$_n$. -\item \texttt{discriminate} +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +%voir reduction__conv_x : histoires d'univers. +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% - This behaves like {\tt discriminate {\ident}} if {\ident} is the - name of an hypothesis to which {\tt discriminate} is applicable; if - the current goal is of the form {\term$_1$} {\tt <>} {\term$_2$}, - this behaves as {\tt intro {\ident}; injection {\ident}}. +\subsection[{\tt cbv \flag$_1$ \dots\ \flag$_n$}, {\tt lazy \flag$_1$ +\dots\ \flag$_n$} and {\tt compute}] +{{\tt cbv \flag$_1$ \dots\ \flag$_n$}, {\tt lazy \flag$_1$ +\dots\ \flag$_n$} and {\tt compute} +\tacindex{cbv} +\tacindex{lazy} +\tacindex{compute} +\tacindex{vm\_compute}\label{vmcompute}} - \begin{ErrMsgs} - \item \errindex{No discriminable equalities} \\ - occurs when the goal does not verify the expected preconditions. - \end{ErrMsgs} -\end{Variants} +These parameterized reduction tactics apply to any goal and perform +the normalization of the goal according to the specified flags. In +correspondence with the kinds of reduction considered in \Coq\, namely +$\beta$ (reduction of functional application), $\delta$ (unfolding of +transparent constants, see \ref{Transparent}), $\iota$ (reduction of +pattern-matching over a constructed term, and unfolding of {\tt fix} +and {\tt cofix} expressions) and $\zeta$ (contraction of local +definitions), the flag are either {\tt beta}, {\tt delta}, {\tt iota} +or {\tt zeta}. The {\tt delta} flag itself can be refined into {\tt +delta [\qualid$_1$\ldots\qualid$_k$]} or {\tt delta +-[\qualid$_1$\ldots\qualid$_k$]}, restricting in the first case the +constants to unfold to the constants listed, and restricting in the +second case the constant to unfold to all but the ones explicitly +mentioned. Notice that the {\tt delta} flag does not apply to +variables bound by a let-in construction inside the term itself (use +here the {\tt zeta} flag). In any cases, opaque constants are not +unfolded (see Section~\ref{Opaque}). -\subsection{\tt injection {\term} -\label{injection} -\tacindex{injection} -\tacindex{einjection}} +The goal may be normalized with two strategies: {\em lazy} ({\tt lazy} +tactic), or {\em call-by-value} ({\tt cbv} tactic). The lazy strategy +is a call-by-need strategy, with sharing of reductions: the arguments of a +function call are partially evaluated only when necessary, and if an +argument is used several times then it is computed only once. This +reduction is efficient for reducing expressions with dead code. For +instance, the proofs of a proposition {\tt exists~$x$. $P(x)$} reduce to a +pair of a witness $t$, and a proof that $t$ satisfies the predicate +$P$. Most of the time, $t$ may be computed without computing the proof +of $P(t)$, thanks to the lazy strategy. -The {\tt injection} tactic is based on the fact that constructors of -inductive sets are injections. That means that if $c$ is a constructor -of an inductive set, and if $(c~\vec{t_1})$ and $(c~\vec{t_2})$ are two -terms that are equal then $~\vec{t_1}$ and $~\vec{t_2}$ are equal -too. +The call-by-value strategy is the one used in ML languages: the +arguments of a function call are evaluated first, using a weak +reduction (no reduction under the $\lambda$-abstractions). Despite the +lazy strategy always performs fewer reductions than the call-by-value +strategy, the latter is generally more efficient for evaluating purely +computational expressions (i.e. with few dead code). -If {\term} is a proof of a statement of conclusion - {\tt {\term$_1$} = {\term$_2$}}, -then {\tt injection} applies injectivity as deep as possible to -derive the equality of all the subterms of {\term$_1$} and {\term$_2$} -placed in the same positions. For example, from {\tt (S - (S n))=(S (S (S m)))} we may derive {\tt n=(S m)}. To use this -tactic {\term$_1$} and {\term$_2$} should be elements of an inductive -set and they should be neither explicitly equal, nor structurally -different. We mean by this that, if {\tt n$_1$} and {\tt n$_2$} are -their respective normal forms, then: -\begin{itemize} -\item {\tt n$_1$} and {\tt n$_2$} should not be syntactically equal, -\item there must not exist any pair of subterms {\tt u} and {\tt w}, - {\tt u} subterm of {\tt n$_1$} and {\tt w} subterm of {\tt n$_2$} , - placed in the same positions and having different constructors as - head symbols. -\end{itemize} -If these conditions are satisfied, then, the tactic derives the -equality of all the subterms of {\term$_1$} and {\term$_2$} placed in -the same positions and puts them as antecedents of the current goal. +\begin{Variants} +\item {\tt compute} \tacindex{compute}\\ + {\tt cbv} -\Example Consider the following goal: + These are synonyms for {\tt cbv beta delta iota zeta}. + +\item {\tt lazy} + + This is a synonym for {\tt lazy beta delta iota zeta}. + +\item {\tt compute [\qualid$_1$\ldots\qualid$_k$]}\\ + {\tt cbv [\qualid$_1$\ldots\qualid$_k$]} + + These are synonyms of {\tt cbv beta delta + [\qualid$_1$\ldots\qualid$_k$] iota zeta}. + +\item {\tt compute -[\qualid$_1$\ldots\qualid$_k$]}\\ + {\tt cbv -[\qualid$_1$\ldots\qualid$_k$]} + + These are synonyms of {\tt cbv beta delta + -[\qualid$_1$\ldots\qualid$_k$] iota zeta}. + +\item {\tt lazy [\qualid$_1$\ldots\qualid$_k$]}\\ + {\tt lazy -[\qualid$_1$\ldots\qualid$_k$]} + + These are respectively synonyms of {\tt lazy beta delta + [\qualid$_1$\ldots\qualid$_k$] iota zeta} and {\tt lazy beta delta + -[\qualid$_1$\ldots\qualid$_k$] iota zeta}. + +\item {\tt vm\_compute} \tacindex{vm\_compute} + + This tactic evaluates the goal using the optimized call-by-value + evaluation bytecode-based virtual machine. This algorithm is + dramatically more efficient than the algorithm used for the {\tt + cbv} tactic, but it cannot be fine-tuned. It is specially + interesting for full evaluation of algebraic objects. This includes + the case of reflexion-based tactics. + +\end{Variants} + +% Obsolete? Anyway not very important message +%\begin{ErrMsgs} +%\item \errindex{Delta must be specified before} +% +% A list of constants appeared before the {\tt delta} flag. +%\end{ErrMsgs} + + +\subsection{{\tt red} +\tacindex{red}} + +This tactic applies to a goal which has the form {\tt + forall (x:T1)\dots(xk:Tk), c t1 \dots\ tn} where {\tt c} is a constant. If +{\tt c} is transparent then it replaces {\tt c} with its definition +(say {\tt t}) and then reduces {\tt (t t1 \dots\ tn)} according to +$\beta\iota\zeta$-reduction rules. + +\begin{ErrMsgs} +\item \errindex{Not reducible} +\end{ErrMsgs} + +\subsection{{\tt hnf} +\tacindex{hnf}} + +This tactic applies to any goal. It replaces the current goal with its +head normal form according to the $\beta\delta\iota\zeta$-reduction +rules, i.e. it reduces the head of the goal until it becomes a +product or an irreducible term. + +\Example +The term \verb+forall n:nat, (plus (S n) (S n))+ is not reduced by {\tt hnf}. + +\Rem The $\delta$ rule only applies to transparent constants +(see Section~\ref{Opaque} on transparency and opacity). + +\subsection{\tt simpl +\tacindex{simpl}} + +This tactic applies to any goal. The tactic {\tt simpl} first applies +$\beta\iota$-reduction rule. Then it expands transparent constants +and tries to reduce {\tt T'} according, once more, to $\beta\iota$ +rules. But when the $\iota$ rule is not applicable then possible +$\delta$-reductions are not applied. For instance trying to use {\tt +simpl} on {\tt (plus n O)=n} changes nothing. Notice that only +transparent constants whose name can be reused as such in the +recursive calls are possibly unfolded. For instance a constant defined +by {\tt plus' := plus} is possibly unfolded and reused in the +recursive calls, but a constant such as {\tt succ := plus (S O)} is +never unfolded. + +The behavior of {\tt simpl} can be tuned using the {\tt Arguments} vernacular +command as follows: +\comindex{Arguments} +\begin{itemize} +\item +A constant can be marked to be never unfolded by {\tt simpl}: +\begin{coq_example*} +Arguments minus x y : simpl never +\end{coq_example*} +After that command an expression like {\tt (minus (S x) y)} is left untouched by +the {\tt simpl} tactic. +\item +A constant can be marked to be unfolded only if applied to enough arguments. +The number of arguments required can be specified using +the {\tt /} symbol in the arguments list of the {\tt Arguments} vernacular +command. +\begin{coq_example*} +Definition fcomp A B C f (g : A -> B) (x : A) : C := f (g x). +Notation "f \o g" := (fcomp f g) (at level 50). +Arguments fcomp {A B C} f g x /. +\end{coq_example*} +After that command the expression {\tt (f \verb+\+o g)} is left untouched by +{\tt simpl} while {\tt ((f \verb+\+o g) t)} is reduced to {\tt (f (g t))}. +The same mechanism can be used to make a constant volatile, i.e. always +unfolded by {\tt simpl}. +\begin{coq_example*} +Definition volatile := fun x : nat => x. +Arguments volatile / x. +\end{coq_example*} +\item +A constant can be marked to be unfolded only if an entire set of arguments +evaluates to a constructor. The {\tt !} symbol can be used to mark such +arguments. +\begin{coq_example*} +Arguments minus !x !y. +\end{coq_example*} +After that command, the expression {\tt (minus (S x) y)} is left untouched by +{\tt simpl}, while {\tt (minus (S x) (S y))} is reduced to {\tt (minus x y)}. +\item +A special heuristic to determine if a constant has to be unfolded can be +activated with the following command: +\begin{coq_example*} +Arguments minus x y : simpl nomatch +\end{coq_example*} +The heuristic avoids to perform a simplification step that would +expose a {\tt match} construct in head position. For example the +expression {\tt (minus (S (S x)) (S y))} is simplified to +{\tt (minus (S x) y)} even if an extra simplification is possible. +\end{itemize} + +\tacindex{simpl \dots\ in} +\begin{Variants} +\item {\tt simpl {\term}} + + This applies {\tt simpl} only to the occurrences of {\term} in the + current goal. + +\item {\tt simpl {\term} at \num$_1$ \dots\ \num$_i$} + + This applies {\tt simpl} only to the \num$_1$, \dots, \num$_i$ + occurrences of {\term} in the current goal. + + \ErrMsg {\tt Too few occurrences} + +\item {\tt simpl {\ident}} + + This applies {\tt simpl} only to the applicative subterms whose head + occurrence is {\ident}. + +\item {\tt simpl {\ident} at \num$_1$ \dots\ \num$_i$} + + This applies {\tt simpl} only to the \num$_1$, \dots, \num$_i$ +applicative subterms whose head occurrence is {\ident}. + +\end{Variants} + +\subsection{\tt unfold \qualid +\tacindex{unfold} +\label{unfold}} + +This tactic applies to any goal. The argument {\qualid} must denote a +defined transparent constant or local definition (see Sections~\ref{Basic-definitions} and~\ref{Transparent}). The tactic {\tt + unfold} applies the $\delta$ rule to each occurrence of the constant +to which {\qualid} refers in the current goal and then replaces it +with its $\beta\iota$-normal form. + +\begin{ErrMsgs} +\item {\qualid} \errindex{does not denote an evaluable constant} + +\end{ErrMsgs} + +\begin{Variants} +\item {\tt unfold {\qualid}$_1$, \dots, \qualid$_n$} + \tacindex{unfold \dots\ in} + + Replaces {\em simultaneously} {\qualid}$_1$, \dots, {\qualid}$_n$ + with their definitions and replaces the current goal with its + $\beta\iota$ normal form. + +\item {\tt unfold {\qualid}$_1$ at \num$_1^1$, \dots, \num$_i^1$, +\dots,\ \qualid$_n$ at \num$_1^n$ \dots\ \num$_j^n$} + + The lists \num$_1^1$, \dots, \num$_i^1$ and \num$_1^n$, \dots, + \num$_j^n$ specify the occurrences of {\qualid}$_1$, \dots, + \qualid$_n$ to be unfolded. Occurrences are located from left to + right. + + \ErrMsg {\tt bad occurrence number of {\qualid}$_i$} + + \ErrMsg {\qualid}$_i$ {\tt does not occur} + +\item {\tt unfold {\qstring}} + + If {\qstring} denotes the discriminating symbol of a notation (e.g. {\tt + "+"}) or an expression defining a notation (e.g. \verb!"_ + _"!), and + this notation refers to an unfoldable constant, then the tactic + unfolds it. + +\item {\tt unfold {\qstring}\%{\delimkey}} + + This is variant of {\tt unfold {\qstring}} where {\qstring} gets its + interpretation from the scope bound to the delimiting key + {\delimkey} instead of its default interpretation (see + Section~\ref{scopechange}). + +\item {\tt unfold \qualidorstring$_1$ at \num$_1^1$, \dots, \num$_i^1$, +\dots,\ \qualidorstring$_n$ at \num$_1^n$ \dots\ \num$_j^n$} + + This is the most general form, where {\qualidorstring} is either a + {\qualid} or a {\qstring} referring to a notation. + +\end{Variants} + +\subsection{{\tt fold} \term +\tacindex{fold}} + +This tactic applies to any goal. The term \term\ is reduced using the {\tt red} +tactic. Every occurrence of the resulting term in the goal is then +replaced by \term. + +\begin{Variants} +\item {\tt fold} \term$_1$ \dots\ \term$_n$ + + Equivalent to {\tt fold} \term$_1${\tt;}\ldots{\tt; fold} \term$_n$. +\end{Variants} + +\subsection{{\tt pattern {\term}} +\tacindex{pattern} +\label{pattern}} + +This command applies to any goal. The argument {\term} must be a free +subterm of the current goal. The command {\tt pattern} performs +$\beta$-expansion (the inverse of $\bt$-reduction) of the current goal +(say \T) by +\begin{enumerate} +\item replacing all occurrences of {\term} in {\T} with a fresh variable +\item abstracting this variable +\item applying the abstracted goal to {\term} +\end{enumerate} + +For instance, if the current goal $T$ is expressible has $\phi(t)$ +where the notation captures all the instances of $t$ in $\phi(t)$, +then {\tt pattern $t$} transforms it into {\tt (fun x:$A$ => $\phi(${\tt +x}$)$) $t$}. This command can be used, for instance, when the tactic +{\tt apply} fails on matching. + +\begin{Variants} +\item {\tt pattern {\term} at {\num$_1$} \dots\ {\num$_n$}} + + Only the occurrences {\num$_1$} \dots\ {\num$_n$} of {\term} are + considered for $\beta$-expansion. Occurrences are located from left + to right. + +\item {\tt pattern {\term} at - {\num$_1$} \dots\ {\num$_n$}} + + All occurrences except the occurrences of indexes {\num$_1$} \dots\ + {\num$_n$} of {\term} are considered for + $\beta$-expansion. Occurrences are located from left to right. + +\item {\tt pattern {\term$_1$}, \dots, {\term$_m$}} + + Starting from a goal $\phi(t_1 \dots\ t_m)$, the tactic + {\tt pattern $t_1$, \dots,\ $t_m$} generates the equivalent goal {\tt + (fun (x$_1$:$A_1$) \dots\ (x$_m$:$A_m$) => $\phi(${\tt x$_1$\dots\ + x$_m$}$)$) $t_1$ \dots\ $t_m$}.\\ If $t_i$ occurs in one of the + generated types $A_j$ these occurrences will also be considered and + possibly abstracted. + +\item {\tt pattern {\term$_1$} at {\num$_1^1$} \dots\ {\num$_{n_1}^1$}, \dots, + {\term$_m$} at {\num$_1^m$} \dots\ {\num$_{n_m}^m$}} + + This behaves as above but processing only the occurrences \num$_1^1$, + \dots, \num$_i^1$ of \term$_1$, \dots, \num$_1^m$, \dots, \num$_j^m$ + of \term$_m$ starting from \term$_m$. + +\item {\tt pattern} {\term$_1$} \zeroone{{\tt at \zeroone{-}} {\num$_1^1$} \dots\ {\num$_{n_1}^1$}} {\tt ,} \dots {\tt ,} + {\term$_m$} \zeroone{{\tt at \zeroone{-}} {\num$_1^m$} \dots\ {\num$_{n_m}^m$}} + + This is the most general syntax that combines the different variants. + +\end{Variants} + +\subsection{Conversion tactics applied to hypotheses} + +{\convtactic} {\tt in} \ident$_1$ \dots\ \ident$_n$ + +Applies the conversion tactic {\convtactic} to the +hypotheses \ident$_1$, \ldots, \ident$_n$. The tactic {\convtactic} is +any of the conversion tactics listed in this section. + +If \ident$_i$ is a local definition, then \ident$_i$ can be replaced +by (Type of \ident$_i$) to address not the body but the type of the +local definition. Example: {\tt unfold not in (Type of H1) (Type of H3).} + +\begin{ErrMsgs} +\item \errindex{No such hypothesis} : {\ident}. +\end{ErrMsgs} + + +\section{Automation} +\subsection{\tt auto +\label{auto} +\tacindex{auto}} + +This tactic implements a Prolog-like resolution procedure to solve the +current goal. It first tries to solve the goal using the {\tt + assumption} tactic, then it reduces the goal to an atomic one using +{\tt intros} and introducing the newly generated hypotheses as hints. +Then it looks at the list of tactics associated to the head symbol of +the goal and tries to apply one of them (starting from the tactics +with lower cost). This process is recursively applied to the generated +subgoals. + +By default, \texttt{auto} only uses the hypotheses of the current goal and the +hints of the database named {\tt core}. + +\begin{Variants} + +\item {\tt auto \num} + + Forces the search depth to be \num. The maximal search depth is 5 by + default. + +\item {\tt auto with \ident$_1$ \dots\ \ident$_n$} + + Uses the hint databases $\ident_1$ \dots\ $\ident_n$ in addition to + the database {\tt core}. See Section~\ref{Hints-databases} for the + list of pre-defined databases and the way to create or extend a + database. This option can be combined with the previous one. + +\item {\tt auto with *} + + Uses all existing hint databases, minus the special database + {\tt v62}. See Section~\ref{Hints-databases} + +\item \texttt{auto using \nterm{lemma}$_1$ , \ldots , \nterm{lemma}$_n$} + + Uses \nterm{lemma}$_1$, \ldots, \nterm{lemma}$_n$ in addition to + hints (can be combined with the \texttt{with \ident} option). If + $lemma_i$ is an inductive type, it is the collection of its + constructors which is added as hints. + +\item \texttt{auto using \nterm{lemma}$_1$ , \ldots , \nterm{lemma}$_n$ with \ident$_1$ \dots\ \ident$_n$} + + This combines the effects of the {\tt using} and {\tt with} options. + +\item {\tt trivial}\tacindex{trivial} + + This tactic is a restriction of {\tt auto} that is not recursive and + tries only hints which cost 0. Typically it solves trivial + equalities like $X=X$. + +\item \texttt{trivial with \ident$_1$ \dots\ \ident$_n$} + +\item \texttt{trivial with *} + +\end{Variants} + +\Rem {\tt auto} either solves completely the goal or else leaves it +intact. \texttt{auto} and \texttt{trivial} never fail. + +\SeeAlso Section~\ref{Hints-databases} + +\subsection{\tt eauto +\tacindex{eauto} +\label{eauto}} + +This tactic generalizes {\tt auto}. In contrast with +the latter, {\tt eauto} uses unification of the goal +against the hints rather than pattern-matching +(in other words, it uses {\tt eapply} instead of +{\tt apply}). +As a consequence, {\tt eauto} can solve such a goal: -\begin{coq_example*} -Inductive list : Set := - | nil : list - | cons : nat -> list -> list. -Variable P : list -> Prop. -\end{coq_example*} -\begin{coq_eval} -Lemma ex : - forall (l:list) (n:nat), P nil -> cons n l = cons 0 nil -> P l. -intros l n H H0. -\end{coq_eval} \begin{coq_example} -Show. -injection H0. +Hint Resolve ex_intro. +Goal forall P:nat -> Prop, P 0 -> exists n, P n. +eauto. \end{coq_example} \begin{coq_eval} Abort. \end{coq_eval} -Beware that \texttt{injection} yields always an equality in a sigma type -whenever the injected object has a dependent type. +Note that {\tt ex\_intro} should be declared as an +hint. -\Rem There is a special case for dependent pairs. If we have a decidable -equality over the type of the first argument, then it is safe to do -the projection on the second one, and so {\tt injection} will work fine. -To define such an equality, you have to use the {\tt Scheme} command -(see \ref{Scheme}). +\SeeAlso Section~\ref{Hints-databases} -\Rem If some quantified hypothesis of the goal is named {\ident}, then -{\tt injection {\ident}} first introduces the hypothesis in the local -context using \texttt{intros until \ident}. +\subsection{\tt autounfold with \ident$_1$ \dots\ \ident$_n$ +\tacindex{autounfold} +\label{autounfold}} -\begin{ErrMsgs} -\item \errindex{Not a projectable equality but a discriminable one} -\item \errindex{Nothing to do, it is an equality between convertible terms} -\item \errindex{Not a primitive equality} -\end{ErrMsgs} +This tactic unfolds constants that were declared through a {\tt Hint + Unfold} in the given databases. + +\begin{Variants} +\item {\tt autounfold with \ident$_1$ \dots\ \ident$_n$ in \textit{clause}} + + Perform the unfolding in the given clause. + +\item {\tt autounfold with *} + + Uses the unfold hints declared in all the hint databases. +\end{Variants} + + +\subsection{\tt autorewrite with \ident$_1$ \dots \ident$_n$. +\label{tactic:autorewrite} +\tacindex{autorewrite}} + +This tactic \footnote{The behavior of this tactic has much changed compared to +the versions available in the previous distributions (V6). This may cause +significant changes in your theories to obtain the same result. As a drawback +of the re-engineering of the code, this tactic has also been completely revised +to get a very compact and readable version.} carries out rewritings according +the rewriting rule bases {\tt \ident$_1$ \dots \ident$_n$}. + +Each rewriting rule of a base \ident$_i$ is applied to the main subgoal until +it fails. Once all the rules have been processed, if the main subgoal has +progressed (e.g., if it is distinct from the initial main goal) then the rules +of this base are processed again. If the main subgoal has not progressed then +the next base is processed. For the bases, the behavior is exactly similar to +the processing of the rewriting rules. + +The rewriting rule bases are built with the {\tt Hint~Rewrite} vernacular +command. + +\Warning{} This tactic may loop if you build non terminating rewriting systems. + +\begin{Variant} +\item {\tt autorewrite with \ident$_1$ \dots \ident$_n$ using \tac}\\ +Performs, in the same way, all the rewritings of the bases {\tt \ident$_1$ $...$ +\ident$_n$} applying {\tt \tac} to the main subgoal after each rewriting step. + +\item \texttt{autorewrite with {\ident$_1$} \dots \ident$_n$ in {\qualid}} + + Performs all the rewritings in hypothesis {\qualid}. +\item \texttt{autorewrite with {\ident$_1$} \dots \ident$_n$ in {\qualid} using \tac} + + Performs all the rewritings in hypothesis {\qualid} applying {\tt + \tac} to the main subgoal after each rewriting step. + +\item \texttt{autorewrite with {\ident$_1$} \dots \ident$_n$ in \textit{clause}} + Performs all the rewritings in the clause \textit{clause}. \\ + The \textit{clause} argument must not contain any \texttt{type of} nor \texttt{value of}. + +\end{Variant} + +\SeeAlso Section~\ref{HintRewrite} for feeding the database of lemmas used by {\tt autorewrite}. + +\SeeAlso Section~\ref{autorewrite-example} for examples showing the use of +this tactic. + +% En attente d'un moyen de valoriser les fichiers de demos +%\SeeAlso file \texttt{contrib/Rocq/DEMOS/Demo\_AutoRewrite.v} + +\section{Controlling automation} + +\subsection{The hints databases for {\tt auto} and {\tt eauto} +\index{Hints databases} +\label{Hints-databases} +\comindex{Hint}} + +The hints for \texttt{auto} and \texttt{eauto} are stored in +databases. Each database maps head symbols to a list of hints. One can +use the command \texttt{Print Hint \ident} to display the hints +associated to the head symbol \ident{} (see \ref{PrintHint}). Each +hint has a cost that is an nonnegative integer, and an optional pattern. +The hints with lower cost are tried first. A hint is tried by +\texttt{auto} when the conclusion of the current goal +matches its pattern or when it has no pattern. + +\subsubsection*{Creating Hint databases + \label{CreateHintDb}\comindex{CreateHintDb}} + +One can optionally declare a hint database using the command +\texttt{Create HintDb}. If a hint is added to an unknown database, it +will be automatically created. + +\medskip +\texttt{Create HintDb} {\ident} [\texttt{discriminated}] +\medskip + +This command creates a new database named \ident. +The database is implemented by a Discrimination Tree (DT) that serves as +an index of all the lemmas. The DT can use transparency information to decide +if a constant should be indexed or not (c.f. \ref{HintTransparency}), +making the retrieval more efficient. +The legacy implementation (the default one for new databases) uses the +DT only on goals without existentials (i.e., auto goals), for non-Immediate +hints and do not make use of transparency hints, putting more work on the +unification that is run after retrieval (it keeps a list of the lemmas +in case the DT is not used). The new implementation enabled by +the {\tt discriminated} option makes use of DTs in all cases and takes +transparency information into account. However, the order in which hints +are retrieved from the DT may differ from the order in which they were +inserted, making this implementation observationally different from the +legacy one. + +\begin{Variants} +\item\texttt{Local Hint} \textsl{hint\_definition} \texttt{:} + \ident$_1$ \ldots\ \ident$_n$ + + This is used to declare a hint database that must not be exported to the other + modules that require and import the current module. Inside a + section, the option {\tt Local} is useless since hints do not + survive anyway to the closure of sections. + +\end{Variants} + +The general +command to add a hint to some database \ident$_1$, \dots, \ident$_n$ is: +\begin{tabbing} + \texttt{Hint} \textsl{hint\_definition} \texttt{:} \ident$_1$ \ldots\ \ident$_n$ +\end{tabbing} +where {\sl hint\_definition} is one of the following expressions: + +\begin{itemize} +\item \texttt{Resolve} {\term} + \comindex{Hint Resolve} + + This command adds {\tt apply {\term}} to the hint list + with the head symbol of the type of \term. The cost of that hint is + the number of subgoals generated by {\tt apply {\term}}. + + In case the inferred type of \term\ does not start with a product the + tactic added in the hint list is {\tt exact {\term}}. In case this + type can be reduced to a type starting with a product, the tactic {\tt + apply {\term}} is also stored in the hints list. -\begin{Variants} -\item \texttt{injection} \num{} + If the inferred type of \term\ contains a dependent + quantification on a predicate, it is added to the hint list of {\tt + eapply} instead of the hint list of {\tt apply}. In this case, a + warning is printed since the hint is only used by the tactic {\tt + eauto} (see \ref{eauto}). A typical example of a hint that is used + only by \texttt{eauto} is a transitivity lemma. - This does the same thing as \texttt{intros until \num} followed by -\texttt{injection \ident} where {\ident} is the identifier for the last -introduced hypothesis. + \begin{ErrMsgs} + \item \errindex{Bound head variable} -\item \texttt{injection} \term{} {\tt with} {\bindinglist} + The head symbol of the type of {\term} is a bound variable such + that this tactic cannot be associated to a constant. - This does the same as \texttt{injection {\term}} but using - the given bindings to instantiate parameters or hypotheses of {\term}. + \item \term\ \errindex{cannot be used as a hint} -\item \texttt{einjection} \num\\ - \texttt{einjection} \term{} \zeroone{{\tt with} {\bindinglist}} + The type of \term\ contains products over variables which do not + appear in the conclusion. A typical example is a transitivity axiom. + In that case the {\tt apply} tactic fails, and thus is useless. - This works the same as {\tt injection} but if the type of {\term}, - or the type of the hypothesis referred to by {\num}, has uninstantiated - parameters, these parameters are left as existential variables. + \end{ErrMsgs} -\item{\tt injection} - - If the current goal is of the form {\term$_1$} {\tt <>} {\term$_2$}, - this behaves as {\tt intro {\ident}; injection {\ident}}. - - \ErrMsg \errindex{goal does not satisfy the expected preconditions} + \begin{Variants} -\item \texttt{injection} \term{} \zeroone{{\tt with} {\bindinglist}} \texttt{as} \nelist{\intropattern}{}\\ -\texttt{injection} \num{} \texttt{as} {\intropattern} {\ldots} {\intropattern}\\ -\texttt{injection} \texttt{as} {\intropattern} {\ldots} {\intropattern}\\ -\texttt{einjection} \term{} \zeroone{{\tt with} {\bindinglist}} \texttt{as} \nelist{\intropattern}{}\\ -\texttt{einjection} \num{} \texttt{as} {\intropattern} {\ldots} {\intropattern}\\ -\texttt{einjection} \texttt{as} {\intropattern} {\ldots} {\intropattern}\\ -\tacindex{injection \ldots{} as} - -These variants apply \texttt{intros} \nelist{\intropattern}{} after -the call to \texttt{injection} or \texttt{einjection}. + \item \texttt{Resolve} {\term$_1$} \dots {\term$_m$} -\end{Variants} + Adds each \texttt{Resolve} {\term$_i$}. -\subsection{\tt simplify\_eq {\term} -\tacindex{simplify\_eq} -\tacindex{esimplify\_eq} -\label{simplify-eq}} + \end{Variants} -Let {\term} be the proof of a statement of conclusion {\tt - {\term$_1$}={\term$_2$}}. If {\term$_1$} and -{\term$_2$} are structurally different (in the sense described for the -tactic {\tt discriminate}), then the tactic {\tt simplify\_eq} behaves as {\tt - discriminate {\term}}, otherwise it behaves as {\tt injection - {\term}}. +\item \texttt{Immediate {\term}} +\comindex{Hint Immediate} -\Rem If some quantified hypothesis of the goal is named {\ident}, then -{\tt simplify\_eq {\ident}} first introduces the hypothesis in the local -context using \texttt{intros until \ident}. + This command adds {\tt apply {\term}; trivial} to the hint list + associated with the head symbol of the type of {\ident} in the given + database. This tactic will fail if all the subgoals generated by + {\tt apply {\term}} are not solved immediately by the {\tt trivial} + tactic (which only tries tactics with cost $0$). -\begin{Variants} -\item \texttt{simplify\_eq} \num + This command is useful for theorems such as the symmetry of equality + or $n+1=m+1 \to n=m$ that we may like to introduce with a + limited use in order to avoid useless proof-search. - This does the same thing as \texttt{intros until \num} then -\texttt{simplify\_eq \ident} where {\ident} is the identifier for the last -introduced hypothesis. + The cost of this tactic (which never generates subgoals) is always 1, + so that it is not used by {\tt trivial} itself. -\item \texttt{simplify\_eq} \term{} {\tt with} {\bindinglist} + \begin{ErrMsgs} - This does the same as \texttt{simplify\_eq {\term}} but using - the given bindings to instantiate parameters or hypotheses of {\term}. + \item \errindex{Bound head variable} -\item \texttt{esimplify\_eq} \num\\ - \texttt{esimplify\_eq} \term{} \zeroone{{\tt with} {\bindinglist}} + \item \term\ \errindex{cannot be used as a hint} - This works the same as {\tt simplify\_eq} but if the type of {\term}, - or the type of the hypothesis referred to by {\num}, has uninstantiated - parameters, these parameters are left as existential variables. + \end{ErrMsgs} -\item{\tt simplify\_eq} + \begin{Variants} -If the current goal has form $t_1\verb=<>=t_2$, it behaves as -\texttt{intro {\ident}; simplify\_eq {\ident}}. -\end{Variants} + \item \texttt{Immediate} {\term$_1$} \dots {\term$_m$} -\subsection{\tt dependent rewrite -> {\ident} -\tacindex{dependent rewrite ->} -\label{dependent-rewrite}} + Adds each \texttt{Immediate} {\term$_i$}. -This tactic applies to any goal. If \ident\ has type -\verb+(existT B a b)=(existT B a' b')+ -in the local context (i.e. each term of the -equality has a sigma type $\{ a:A~ \&~(B~a)\}$) this tactic rewrites -\verb+a+ into \verb+a'+ and \verb+b+ into \verb+b'+ in the current -goal. This tactic works even if $B$ is also a sigma type. This kind -of equalities between dependent pairs may be derived by the injection -and inversion tactics. + \end{Variants} -\begin{Variants} -\item{\tt dependent rewrite <- {\ident}} -\tacindex{dependent rewrite <-} \\ -Analogous to {\tt dependent rewrite ->} but uses the equality from -right to left. -\end{Variants} +\item \texttt{Constructors} {\ident} +\comindex{Hint Constructors} -\section{Inversion -\label{inversion}} + If {\ident} is an inductive type, this command adds all its + constructors as hints of type \texttt{Resolve}. Then, when the + conclusion of current goal has the form \texttt{({\ident} \dots)}, + \texttt{auto} will try to apply each constructor. -\subsection{\tt inversion {\ident} -\tacindex{inversion}} + \begin{ErrMsgs} -Let the type of \ident~ in the local context be $(I~\vec{t})$, -where $I$ is a (co)inductive predicate. Then, -\texttt{inversion} applied to \ident~ derives for each possible -constructor $c_i$ of $(I~\vec{t})$, {\bf all} the necessary -conditions that should hold for the instance $(I~\vec{t})$ to be -proved by $c_i$. + \item {\ident} \errindex{is not an inductive type} -\Rem If {\ident} does not denote a hypothesis in the local context -but refers to a hypothesis quantified in the goal, then the -latter is first introduced in the local context using -\texttt{intros until \ident}. + \item {\ident} \errindex{not declared} -\begin{Variants} -\item \texttt{inversion} \num - - This does the same thing as \texttt{intros until \num} then - \texttt{inversion \ident} where {\ident} is the identifier for the - last introduced hypothesis. + \end{ErrMsgs} -\item \tacindex{inversion\_clear} \texttt{inversion\_clear} \ident + \begin{Variants} - This behaves as \texttt{inversion} and then erases \ident~ from the - context. + \item \texttt{Constructors} {\ident$_1$} \dots {\ident$_m$} -\item \tacindex{inversion \dots\ as} \texttt{inversion} {\ident} \texttt{as} {\intropattern} - - This behaves as \texttt{inversion} but using names in - {\intropattern} for naming hypotheses. The {\intropattern} must have - the form {\tt [} $p_{11}$ \ldots $p_{1n_1}$ {\tt |} {\ldots} {\tt |} - $p_{m1}$ \ldots $p_{mn_m}$ {\tt ]} with $m$ being the number of - constructors of the type of {\ident}. Be careful that the list must - be of length $m$ even if {\tt inversion} discards some cases (which - is precisely one of its roles): for the discarded cases, just use an - empty list (i.e. $n_i=0$). + Adds each \texttt{Constructors} {\ident$_i$}. - The arguments of the $i^{th}$ constructor and the - equalities that {\tt inversion} introduces in the context of the - goal corresponding to the $i^{th}$ constructor, if it exists, get - their names from the list $p_{i1}$ \ldots $p_{in_i}$ in order. If - there are not enough names, {\tt induction} invents names for the - remaining variables to introduce. In case an equation splits into - several equations (because {\tt inversion} applies {\tt injection} - on the equalities it generates), the corresponding name $p_{ij}$ in - the list must be replaced by a sublist of the form {\tt [$p_{ij1}$ - \ldots $p_{ijq}$]} (or, equivalently, {\tt ($p_{ij1}$, - \ldots, $p_{ijq}$)}) where $q$ is the number of subequalities - obtained from splitting the original equation. Here is an example. + \end{Variants} -\begin{coq_eval} -Require Import List. -\end{coq_eval} +\item \texttt{Unfold} {\qualid} +\comindex{Hint Unfold} -\begin{coq_example} -Inductive contains0 : list nat -> Prop := - | in_hd : forall l, contains0 (0 :: l) - | in_tl : forall l b, contains0 l -> contains0 (b :: l). -Goal forall l:list nat, contains0 (1 :: l) -> contains0 l. -intros l H; inversion H as [ | l' p Hl' [Heqp Heql'] ]. -\end{coq_example} + This adds the tactic {\tt unfold {\qualid}} to the hint list that + will only be used when the head constant of the goal is \ident. Its + cost is 4. -\begin{coq_eval} -Abort. -\end{coq_eval} + \begin{Variants} -\item \texttt{inversion} {\num} {\tt as} {\intropattern} + \item \texttt{Unfold} {\ident$_1$} \dots {\ident$_m$} - This allows to name the hypotheses introduced by - \texttt{inversion} {\num} in the context. + Adds each \texttt{Unfold} {\ident$_i$}. -\item \tacindex{inversion\_cleardots\ as} \texttt{inversion\_clear} - {\ident} {\tt as} {\intropattern} + \end{Variants} - This allows to name the hypotheses introduced by - \texttt{inversion\_clear} in the context. - -\item \tacindex{inversion \dots\ in} \texttt{inversion } {\ident} - \texttt{in} \ident$_1$ \dots\ \ident$_n$ +\item \texttt{Transparent}, \texttt{Opaque} {\qualid} +\label{HintTransparency} +\comindex{Hint Transparent} +\comindex{Hint Opaque} - Let \ident$_1$ \dots\ \ident$_n$, be identifiers in the local context. This - tactic behaves as generalizing \ident$_1$ \dots\ \ident$_n$, and - then performing \texttt{inversion}. - -\item \tacindex{inversion \dots\ as \dots\ in} \texttt{inversion } - {\ident} {\tt as} {\intropattern} \texttt{in} \ident$_1$ \dots\ - \ident$_n$ - - This allows to name the hypotheses introduced in the context by - \texttt{inversion} {\ident} \texttt{in} \ident$_1$ \dots\ - \ident$_n$. - -\item \tacindex{inversion\_clear \dots\ in} \texttt{inversion\_clear} - {\ident} \texttt{in} \ident$_1$ \ldots \ident$_n$ - - Let \ident$_1$ \dots\ \ident$_n$, be identifiers in the local context. This - tactic behaves as generalizing \ident$_1$ \dots\ \ident$_n$, and - then performing {\tt inversion\_clear}. - -\item \tacindex{inversion\_clear \dots\ as \dots\ in} - \texttt{inversion\_clear} {\ident} \texttt{as} {\intropattern} - \texttt{in} \ident$_1$ \ldots \ident$_n$ + This adds a transparency hint to the database, making {\tt {\qualid}} + a transparent or opaque constant during resolution. This information + is used during unification of the goal with any lemma in the database + and inside the discrimination network to relax or constrain it in the + case of \texttt{discriminated} databases. - This allows to name the hypotheses introduced in the context by - \texttt{inversion\_clear} {\ident} \texttt{in} \ident$_1$ \ldots - \ident$_n$. + \begin{Variants} -\item \tacindex{dependent inversion} \texttt{dependent inversion} - {\ident} - - That must be used when \ident\ appears in the current goal. It acts - like \texttt{inversion} and then substitutes \ident\ for the - corresponding term in the goal. - -\item \tacindex{dependent inversion \dots\ as } \texttt{dependent - inversion} {\ident} \texttt{as} {\intropattern} - - This allows to name the hypotheses introduced in the context by - \texttt{dependent inversion} {\ident}. + \item \texttt{Transparent}, \texttt{Opaque} {\ident$_1$} \dots {\ident$_m$} -\item \tacindex{dependent inversion\_clear} \texttt{dependent - inversion\_clear} {\ident} - - Like \texttt{dependent inversion}, except that {\ident} is cleared - from the local context. + Declares each {\ident$_i$} as a transparent or opaque constant. -\item \tacindex{dependent inversion\_clear \dots\ as} - \texttt{dependent inversion\_clear} {\ident}\texttt{as} {\intropattern} - - This allows to name the hypotheses introduced in the context by - \texttt{dependent inversion\_clear} {\ident}. + \end{Variants} -\item \tacindex{dependent inversion \dots\ with} \texttt{dependent - inversion } {\ident} \texttt{ with } \term - - This variant allows you to specify the generalization of the goal. It - is useful when the system fails to generalize the goal automatically. If - {\ident} has type $(I~\vec{t})$ and $I$ has type - $forall (\vec{x}:\vec{T}), s$, then \term~ must be of type - $I:forall (\vec{x}:\vec{T}), I~\vec{x}\to s'$ where $s'$ is the - type of the goal. +\item \texttt{Extern \num\ [\pattern]\ => }\textsl{tactic} +\comindex{Hint Extern} -\item \tacindex{dependent inversion \dots\ as \dots\ with} - \texttt{dependent inversion } {\ident} \texttt{as} {\intropattern} - \texttt{ with } \term - - This allows to name the hypotheses introduced in the context by - \texttt{dependent inversion } {\ident} \texttt{ with } \term. + This hint type is to extend \texttt{auto} with tactics other than + \texttt{apply} and \texttt{unfold}. For that, we must specify a + cost, an optional pattern and a tactic to execute. Here is an example: -\item \tacindex{dependent inversion\_clear \dots\ with} - \texttt{dependent inversion\_clear } {\ident} \texttt{ with } \term - - Like \texttt{dependent inversion \dots\ with} but clears {\ident} from - the local context. +\begin{quotation} +\begin{verbatim} +Hint Extern 4 (~(_ = _)) => discriminate. +\end{verbatim} +\end{quotation} -\item \tacindex{dependent inversion\_clear \dots\ as \dots\ with} - \texttt{dependent inversion\_clear } {\ident} \texttt{as} - {\intropattern} \texttt{ with } \term - - This allows to name the hypotheses introduced in the context by - \texttt{dependent inversion\_clear } {\ident} \texttt{ with } \term. + Now, when the head of the goal is a disequality, \texttt{auto} will + try \texttt{discriminate} if it does not manage to solve the goal + with hints with a cost less than 4. -\item \tacindex{simple inversion} \texttt{simple inversion} {\ident} - - It is a very primitive inversion tactic that derives all the necessary - equalities but it does not simplify the constraints as - \texttt{inversion} does. + One can even use some sub-patterns of the pattern in the tactic + script. A sub-pattern is a question mark followed by an ident, like + \texttt{?X1} or \texttt{?X2}. Here is an example: -\item \tacindex{simple inversion \dots\ as} \texttt{simple inversion} - {\ident} \texttt{as} {\intropattern} - - This allows to name the hypotheses introduced in the context by - \texttt{simple inversion}. +% Require EqDecide. +\begin{coq_example*} +Require Import List. +\end{coq_example*} +\begin{coq_example} +Hint Extern 5 ({?X1 = ?X2} + {?X1 <> ?X2}) => + generalize X1, X2; decide equality : eqdec. +Goal +forall a b:list (nat * nat), {a = b} + {a <> b}. +info auto with eqdec. +\end{coq_example} +\begin{coq_eval} +Abort. +\end{coq_eval} -\item \tacindex{inversion \dots\ using} \texttt{inversion} \ident - \texttt{ using} \ident$'$ - - Let {\ident} have type $(I~\vec{t})$ ($I$ an inductive - predicate) in the local context, and \ident$'$ be a (dependent) inversion - lemma. Then, this tactic refines the current goal with the specified - lemma. +\end{itemize} -\item \tacindex{inversion \dots\ using \dots\ in} \texttt{inversion} - {\ident} \texttt{using} \ident$'$ \texttt{in} \ident$_1$\dots\ \ident$_n$ - - This tactic behaves as generalizing \ident$_1$\dots\ \ident$_n$, - then doing \texttt{inversion} {\ident} \texttt{using} \ident$'$. +\Rem One can use an \texttt{Extern} hint with no pattern to do +pattern-matching on hypotheses using \texttt{match goal with} inside +the tactic. -\end{Variants} +\begin{Variants} +\item \texttt{Hint} \textsl{hint\_definition} -\SeeAlso~\ref{inversion-examples} for detailed examples + No database name is given: the hint is registered in the {\tt core} + database. -\subsection{\tt Derive Inversion {\ident} with - ${\tt forall (}\vec{x}{\tt :}\vec{T}{\tt),} I~\vec{t}$ Sort \sort -\label{Derive-Inversion} -\comindex{Derive Inversion}} +\item\texttt{Hint Local} \textsl{hint\_definition} \texttt{:} + \ident$_1$ \ldots\ \ident$_n$ -This command generates an inversion principle for the -\texttt{inversion \dots\ using} tactic. -Let $I$ be an inductive predicate and $\vec{x}$ the variables -occurring in $\vec{t}$. This command generates and stocks the -inversion lemma for the sort \sort~ corresponding to the instance -$forall (\vec{x}:\vec{T}), I~\vec{t}$ with the name {\ident} in the {\bf -global} environment. When applied it is equivalent to have inverted -the instance with the tactic {\tt inversion}. + This is used to declare hints that must not be exported to the other + modules that require and import the current module. Inside a + section, the option {\tt Local} is useless since hints do not + survive anyway to the closure of sections. -\begin{Variants} -\item \texttt{Derive Inversion\_clear} {\ident} \texttt{with} - \comindex{Derive Inversion\_clear} - $forall (\vec{x}:\vec{T}), I~\vec{t}$ \texttt{Sort} \sort~ \\ - \index{Derive Inversion\_clear \dots\ with} - When applied it is equivalent to having - inverted the instance with the tactic \texttt{inversion} - replaced by the tactic \texttt{inversion\_clear}. -\item \texttt{Derive Dependent Inversion} {\ident} \texttt{with} - $forall (\vec{x}:\vec{T}), I~\vec{t}$ \texttt{Sort} \sort~\\ - \comindex{Derive Dependent Inversion} - When applied it is equivalent to having - inverted the instance with the tactic \texttt{dependent inversion}. -\item \texttt{Derive Dependent Inversion\_clear} {\ident} \texttt{with} - $forall (\vec{x}:\vec{T}), I~\vec{t}$ \texttt{Sort} \sort~\\ - \comindex{Derive Dependent Inversion\_clear} - When applied it is equivalent to having - inverted the instance with the tactic \texttt{dependent inversion\_clear}. -\end{Variants} +\item\texttt{Hint Local} \textsl{hint\_definition} -\SeeAlso \ref{inversion-examples} for examples + Idem for the {\tt core} database. +\end{Variants} +% There are shortcuts that allow to define several goal at once: -\subsection[\tt functional inversion \ident]{\tt functional inversion \ident\label{sec:functional-inversion}} +% \begin{itemize} +% \item \comindex{Hints Resolve}\texttt{Hints Resolve \ident$_1$ \dots\ \ident$_n$ : \ident.}\\ +% This command is a shortcut for the following ones: +% \begin{quotation} +% \noindent\texttt{Hint \ident$_1$ : \ident\ := Resolve \ident$_1$}\\ +% \dots\\ +% \texttt{Hint \ident$_1$ : \ident := Resolve \ident$_1$} +% \end{quotation} +% Notice that the hint name is the same that the theorem given as +% hint. +% \item \comindex{Hints Immediate}\texttt{Hints Immediate \ident$_1$ \dots\ \ident$_n$ : \ident.}\\ +% \item \comindex{Hints Unfold}\texttt{Hints Unfold \qualid$_1$ \dots\ \qualid$_n$ : \ident.}\\ +% \end{itemize} -\texttt{functional inversion} is a \emph{highly} experimental tactic -which performs inversion on hypothesis \ident\ of the form -\texttt{\qualid\ \term$_1$\dots\term$_n$\ = \term} or \texttt{\term\ = - \qualid\ \term$_1$\dots\term$_n$} where \qualid\ must have been -defined using \texttt{Function} (see Section~\ref{Function}). +%\begin{Warnings} +% \item \texttt{Overriding hint named \dots\ in database \dots} +%\end{Warnings} -\begin{ErrMsgs} -\item \errindex{Hypothesis {\ident} must contain at least one Function} -\item \errindex{Cannot find inversion information for hypothesis \ident} - This error may be raised when some inversion lemma failed to be - generated by Function. -\end{ErrMsgs} -\begin{Variants} -\item {\tt functional inversion \num} - This does the same thing as \texttt{intros until \num} then - \texttt{functional inversion \ident} where {\ident} is the - identifier for the last introduced hypothesis. -\item {\tt functional inversion \ident\ \qualid}\\ - {\tt functional inversion \num\ \qualid} +\subsection{Hint databases defined in the \Coq\ standard library} - In case the hypothesis {\ident} (or {\num}) has a type of the form - \texttt{\qualid$_1$\ \term$_1$\dots\term$_n$\ =\ \qualid$_2$\ - \term$_{n+1}$\dots\term$_{n+m}$} where \qualid$_1$ and \qualid$_2$ - are valid candidates to functional inversion, this variant allows to - choose which must be inverted. -\end{Variants} +Several hint databases are defined in the \Coq\ standard library. The +actual content of a database is the collection of the hints declared +to belong to this database in each of the various modules currently +loaded. Especially, requiring new modules potentially extend a +database. At {\Coq} startup, only the {\tt core} and {\tt v62} +databases are non empty and can be used. +\begin{description} +\item[\tt core] This special database is automatically used by + \texttt{auto}, except when pseudo-database \texttt{nocore} is + given to \texttt{auto}. The \texttt{core} database contains + only basic lemmas about negation, + conjunction, and so on from. Most of the hints in this database come + from the \texttt{Init} and \texttt{Logic} directories. -\subsection{\tt quote \ident -\tacindex{quote} -\index{2-level approach}} +\item[\tt arith] This database contains all lemmas about Peano's + arithmetic proved in the directories \texttt{Init} and + \texttt{Arith} -This kind of inversion has nothing to do with the tactic -\texttt{inversion} above. This tactic does \texttt{change (\ident\ - t)}, where \texttt{t} is a term built in order to ensure the -convertibility. In other words, it does inversion of the function -\ident. This function must be a fixpoint on a simple recursive -datatype: see~\ref{quote-examples} for the full details. +\item[\tt zarith] contains lemmas about binary signed integers from + the directories \texttt{theories/ZArith}. When required, the module + {\tt Omega} also extends the database {\tt zarith} with a high-cost + hint that calls {\tt omega} on equations and inequalities in {\tt + nat} or {\tt Z}. -\begin{ErrMsgs} -\item \errindex{quote: not a simple fixpoint}\\ - Happens when \texttt{quote} is not able to perform inversion properly. -\end{ErrMsgs} +\item[\tt bool] contains lemmas about booleans, mostly from directory + \texttt{theories/Bool}. -\begin{Variants} -\item \texttt{quote {\ident} [ \ident$_1$ \dots \ident$_n$ ]}\\ - All terms that are built only with \ident$_1$ \dots \ident$_n$ will be - considered by \texttt{quote} as constants rather than variables. -\end{Variants} +\item[\tt datatypes] is for lemmas about lists, streams and so on that + are mainly proved in the \texttt{Lists} subdirectory. -% En attente d'un moyen de valoriser les fichiers de demos -% \SeeAlso file \texttt{theories/DEMOS/DemoQuote.v} in the distribution +\item[\tt sets] contains lemmas about sets and relations from the + directories \texttt{Sets} and \texttt{Relations}. -\section[Classical tactics]{Classical tactics\label{ClassicalTactics}} +\item[\tt typeclass\_instances] contains all the type class instances + declared in the environment, including those used for \texttt{setoid\_rewrite}, + from the \texttt{Classes} directory. +\end{description} -In order to ease the proving process, when the {\tt Classical} module is loaded. A few more tactics are available. Make sure to load the module using the \texttt{Require Import} command. +There is also a special database called {\tt v62}. It collects all +hints that were declared in the versions of {\Coq} prior to version +6.2.4 when the databases {\tt core}, {\tt arith}, and so on were +introduced. The purpose of the database {\tt v62} is to ensure +compatibility with further versions of {\Coq} for developments done in +versions prior to 6.2.4 ({\tt auto} being replaced by {\tt auto with v62}). +The database {\tt v62} is intended not to be extended (!). It is not +included in the hint databases list used in the {\tt auto with *} tactic. -\subsection{{\tt classical\_left, classical\_right} \tacindex{classical\_left} \tacindex{classical\_right}} +Furthermore, you are advised not to put your own hints in the +{\tt core} database, but use one or several databases specific to your +development. -The tactics \texttt{classical\_left} and \texttt{classical\_right} are the analog of the \texttt{left} and \texttt{right} but using classical logic. They can only be used for disjunctions. -Use \texttt{classical\_left} to prove the left part of the disjunction with the assumption that the negation of right part holds. -Use \texttt{classical\_right} to prove the right part of the disjunction with the assumption that the negation of left part holds. +\subsection{\tt Print Hint +\label{PrintHint} +\comindex{Print Hint}} -\section{Automatizing -\label{Automatizing}} +This command displays all hints that apply to the current goal. It +fails if no proof is being edited, while the two variants can be used at +every moment. -\subsection{\tt auto -\label{auto} -\tacindex{auto}} +\begin{Variants} -This tactic implements a Prolog-like resolution procedure to solve the -current goal. It first tries to solve the goal using the {\tt - assumption} tactic, then it reduces the goal to an atomic one using -{\tt intros} and introducing the newly generated hypotheses as hints. -Then it looks at the list of tactics associated to the head symbol of -the goal and tries to apply one of them (starting from the tactics -with lower cost). This process is recursively applied to the generated -subgoals. +\item {\tt Print Hint {\ident} } -By default, \texttt{auto} only uses the hypotheses of the current goal and the -hints of the database named {\tt core}. + This command displays only tactics associated with \ident\ in the + hints list. This is independent of the goal being edited, so this + command will not fail if no goal is being edited. -\begin{Variants} +\item {\tt Print Hint *} -\item {\tt auto \num} + This command displays all declared hints. - Forces the search depth to be \num. The maximal search depth is 5 by - default. +\item {\tt Print HintDb {\ident} } +\label{PrintHintDb} +\comindex{Print HintDb} -\item {\tt auto with \ident$_1$ \dots\ \ident$_n$} - - Uses the hint databases $\ident_1$ \dots\ $\ident_n$ in addition to - the database {\tt core}. See Section~\ref{Hints-databases} for the - list of pre-defined databases and the way to create or extend a - database. This option can be combined with the previous one. + This command displays all hints from database \ident. -\item {\tt auto with *} +\end{Variants} - Uses all existing hint databases, minus the special database - {\tt v62}. See Section~\ref{Hints-databases} +\subsection{\tt Hint Rewrite \term$_1$ \dots \term$_n$ : \ident +\label{HintRewrite} +\comindex{Hint Rewrite}} -\item \texttt{auto using \nterm{lemma}$_1$ , \ldots , \nterm{lemma}$_n$} +This vernacular command adds the terms {\tt \term$_1$ \dots \term$_n$} +(their types must be equalities) in the rewriting base {\tt \ident} +with the default orientation (left to right). Notice that the +rewriting bases are distinct from the {\tt auto} hint bases and that +{\tt auto} does not take them into account. - Uses \nterm{lemma}$_1$, \ldots, \nterm{lemma}$_n$ in addition to - hints (can be combined with the \texttt{with \ident} option). If - $lemma_i$ is an inductive type, it is the collection of its - constructors which is added as hints. +This command is synchronous with the section mechanism (see \ref{Section}): +when closing a section, all aliases created by \texttt{Hint Rewrite} in that +section are lost. Conversely, when loading a module, all \texttt{Hint Rewrite} +declarations at the global level of that module are loaded. -\item \texttt{auto using \nterm{lemma}$_1$ , \ldots , \nterm{lemma}$_n$ with \ident$_1$ \dots\ \ident$_n$} +\begin{Variants} +\item {\tt Hint Rewrite -> \term$_1$ \dots \term$_n$ : \ident}\\ +This is strictly equivalent to the command above (we only make explicit the +orientation which otherwise defaults to {\tt ->}). - This combines the effects of the {\tt using} and {\tt with} options. +\item {\tt Hint Rewrite <- \term$_1$ \dots \term$_n$ : \ident}\\ +Adds the rewriting rules {\tt \term$_1$ \dots \term$_n$} with a right-to-left +orientation in the base {\tt \ident}. -\item {\tt trivial}\tacindex{trivial} +\item {\tt Hint Rewrite \term$_1$ \dots \term$_n$ using {\tac} : {\ident}}\\ +When the rewriting rules {\tt \term$_1$ \dots \term$_n$} in {\tt \ident} will +be used, the tactic {\tt \tac} will be applied to the generated subgoals, the +main subgoal excluded. - This tactic is a restriction of {\tt auto} that is not recursive and - tries only hints which cost 0. Typically it solves trivial - equalities like $X=X$. +%% \item +%% {\tt Hint Rewrite [ \term$_1$ \dots \term$_n$ ] in \ident}\\ +%% {\tt Hint Rewrite [ \term$_1$ \dots \term$_n$ ] in {\ident} using {\tac}}\\ +%% These are deprecated syntactic variants for +%% {\tt Hint Rewrite \term$_1$ \dots \term$_n$ : \ident} and +%% {\tt Hint Rewrite \term$_1$ \dots \term$_n$ using {\tac} : {\ident}}. -\item \texttt{trivial with \ident$_1$ \dots\ \ident$_n$} +\item \texttt{Print Rewrite HintDb {\ident}} -\item \texttt{trivial with *} + This command displays all rewrite hints contained in {\ident}. \end{Variants} -\Rem {\tt auto} either solves completely the goal or else leaves it -intact. \texttt{auto} and \texttt{trivial} never fail. +\subsection{Hints and sections +\label{Hint-and-Section}} -\SeeAlso Section~\ref{Hints-databases} +Hints provided by the \texttt{Hint} commands are erased when closing a +section. Conversely, all hints of a module \texttt{A} that are not +defined inside a section (and not defined with option {\tt Local}) become +available when the module {\tt A} is imported (using +e.g. \texttt{Require Import A.}). -\subsection{\tt eauto -\tacindex{eauto} -\label{eauto}} +\subsection{Setting implicit automation tactics} -This tactic generalizes {\tt auto}. In contrast with -the latter, {\tt eauto} uses unification of the goal -against the hints rather than pattern-matching -(in other words, it uses {\tt eapply} instead of -{\tt apply}). -As a consequence, {\tt eauto} can solve such a goal: +\subsubsection[\tt Proof with {\tac}.]{\tt Proof with {\tac}.\label{ProofWith} +\comindex{Proof with}} -\begin{coq_example} -Hint Resolve ex_intro. -Goal forall P:nat -> Prop, P 0 -> exists n, P n. -eauto. -\end{coq_example} -\begin{coq_eval} -Abort. -\end{coq_eval} + This command may be used to start a proof. It defines a default + tactic to be used each time a tactic command {\tac$_1$} is ended by + ``\verb#...#''. In this case the tactic command typed by the user is + equivalent to \tac$_1$;{\tac}. -Note that {\tt ex\_intro} should be declared as an -hint. +\SeeAlso {\tt Proof.} in Section~\ref{BeginProof}. -\SeeAlso Section~\ref{Hints-databases} +\begin{Variants} +\item {\tt Proof with {\tac} using {\ident$_1$ \dots {\ident$_n$}}} + Combines in a single line {\tt Proof with} and {\tt Proof using}, + see~\ref{ProofUsing} +\item {\tt Proof using {\ident$_1$ \dots {\ident$_n$}} with {\tac}} + Combines in a single line {\tt Proof with} and {\tt Proof using}, + see~\ref{ProofUsing} -\subsection{\tt autounfold with \ident$_1$ \dots\ \ident$_n$ -\tacindex{autounfold} -\label{autounfold}} +\end{Variants} -This tactic unfolds constants that were declared through a {\tt Hint - Unfold} in the given databases. +\subsubsection[\tt Declare Implicit Tactic {\tac}.]{\tt Declare Implicit Tactic {\tac}.\comindex{Declare Implicit Tactic}} + +This command declares a tactic to be used to solve implicit arguments +that {\Coq} does not know how to solve by unification. It is used +every time the term argument of a tactic has one of its holes not +fully resolved. -\begin{Variants} -\item {\tt autounfold with \ident$_1$ \dots\ \ident$_n$ in \textit{clause}} - - Perform the unfolding in the given clause. +Here is an example: -\item {\tt autounfold with *} - - Uses the unfold hints declared in all the hint databases. -\end{Variants} +\begin{coq_example} +Parameter quo : nat -> forall n:nat, n<>0 -> nat. +Notation "x // y" := (quo x y _) (at level 40). +Declare Implicit Tactic assumption. +Goal forall n m, m<>0 -> { q:nat & { r | q * m + r = n } }. +intros. +exists (n // m). +\end{coq_example} -% EXISTE ENCORE ? -% -% \subsection{\tt Prolog [ \term$_1$ \dots\ \term$_n$ ] \num} -% \tacindex{Prolog}\label{Prolog} -% This tactic, implemented by Chet Murthy, is based upon the concept of -% existential variables of Gilles Dowek, stating that resolution is a -% kind of unification. It tries to solve the current goal using the {\tt -% Assumption} tactic, the {\tt intro} tactic, and applying hypotheses -% of the local context and terms of the given list {\tt [ \term$_1$ -% \dots\ \term$_n$\ ]}. It is more powerful than {\tt auto} since it -% may apply to any theorem, even those of the form {\tt (x:A)(P x) -> Q} -% where {\tt x} does not appear free in {\tt Q}. The maximal search -% depth is {\tt \num}. +The tactic {\tt exists (n // m)} did not fail. The hole was solved by +{\tt assumption} so that it behaved as {\tt exists (quo n m H)}. -% \begin{ErrMsgs} -% \item \errindex{Prolog failed}\\ -% The Prolog tactic was not able to prove the subgoal. -% \end{ErrMsgs} +\section{Decision procedures} \subsection{\tt tauto \tacindex{tauto} @@ -3402,7 +3837,7 @@ instantiation of \verb=x= is necessary. The tactic \texttt{intuition} takes advantage of the search-tree built by the decision procedure involved in the tactic {\tt tauto}. It uses this information to generate a set of subgoals equivalent to the -original one (but simpler than it) and applies the tactic +original one (but simpler than it) and applies the tactic {\tac} to them \cite{Mun94}. If this tactic fails on some goals then {\tt intuition} fails. In fact, {\tt tauto} is simply {\tt intuition fail}. @@ -3438,17 +3873,17 @@ incompatibilities. \tacindex{rtauto} \label{rtauto}} -The {\tt rtauto} tactic solves propositional tautologies similarly to what {\tt tauto} does. The main difference is that the proof term is built using a reflection scheme applied to a sequent calculus proof of the goal. The search procedure is also implemented using a different technique. +The {\tt rtauto} tactic solves propositional tautologies similarly to what {\tt tauto} does. The main difference is that the proof term is built using a reflection scheme applied to a sequent calculus proof of the goal. The search procedure is also implemented using a different technique. -Users should be aware that this difference may result in faster proof-search but slower proof-checking, and {\tt rtauto} might not solve goals that {\tt tauto} would be able to solve (e.g. goals involving universal quantifiers). +Users should be aware that this difference may result in faster proof-search but slower proof-checking, and {\tt rtauto} might not solve goals that {\tt tauto} would be able to solve (e.g. goals involving universal quantifiers). \subsection{{\tt firstorder} \tacindex{firstorder} \label{firstorder}} The tactic \texttt{firstorder} is an {\it experimental} extension of -\texttt{tauto} to -first-order reasoning, written by Pierre Corbineau. +\texttt{tauto} to +first-order reasoning, written by Pierre Corbineau. It is not restricted to usual logical connectives but instead may reason about any first-order class inductive definition. @@ -3478,910 +3913,587 @@ instead may reason about any first-order class inductive definition. \end{Variants} Proof-search is bounded by a depth parameter which can be set by typing the -{\nobreak \tt Set Firstorder Depth $n$} \comindex{Set Firstorder Depth} +{\nobreak \tt Set Firstorder Depth $n$} \comindex{Set Firstorder Depth} vernacular command. -%% \subsection{{\tt jp} {\em (Jprover)} -%% \tacindex{jp} -%% \label{jprover}} - -%% The tactic \texttt{jp}, due to Huang Guan-Shieng, is an experimental -%% port of the {\em Jprover}\cite{SLKN01} semi-decision procedure for -%% first-order intuitionistic logic implemented in {\em -%% NuPRL}\cite{Kre02}. - -%% The tactic \texttt{jp}, due to Huang Guan-Shieng, is an {\it -%% experimental} port of the {\em Jprover}\cite{SLKN01} semi-decision -%% procedure for first-order intuitionistic logic implemented in {\em -%% NuPRL}\cite{Kre02}. - -%% Search may optionnaly be bounded by a multiplicity parameter -%% indicating how many (at most) copies of a formula may be used in -%% the proof process, its absence may lead to non-termination of the tactic. - -%% %\begin{coq_eval} -%% %Variable S:Set. -%% %Variables P Q:S->Prop. -%% %Variable f:S->S. -%% %\end{coq_eval} - -%% %\begin{coq_example*} -%% %Lemma example: (exists x |P x\/Q x)->(exists x |P x)\/(exists x |Q x). -%% %jp. -%% %Qed. - -%% %Lemma example2: (forall x ,P x->P (f x))->forall x,P x->P (f(f x)). -%% %jp. -%% %Qed. -%% %\end{coq_example*} - -%% \begin{Variants} -%% \item {\tt jp $n$}\\ -%% \tacindex{jp $n$} -%% Tries the {\em Jprover} procedure with multiplicities up to $n$, -%% starting from 1. -%% \item {\tt jp}\\ -%% Tries the {\em Jprover} procedure without multiplicity bound, -%% possibly running forever. -%% \end{Variants} - -%% \begin{ErrMsgs} -%% \item \errindex{multiplicity limit reached}\\ -%% The procedure tried all multiplicities below the limit and -%% failed. Goal might be solved by increasing the multiplicity limit. -%% \item \errindex{formula is not provable}\\ -%% The procedure determined that goal was not provable in -%% intuitionistic first-order logic, no matter how big the -%% multiplicity is. -%% \end{ErrMsgs} - - -% \subsection[\tt Linear]{\tt Linear\tacindex{Linear}\label{Linear}} -% The tactic \texttt{Linear}, due to Jean-Christophe Filli{\^a}atre -% \cite{Fil94}, implements a decision procedure for {\em Direct -% Predicate Calculus}, that is first-order Gentzen's Sequent Calculus -% without contraction rules \cite{KeWe84,BeKe92}. Intuitively, a -% first-order goal is provable in Direct Predicate Calculus if it can be -% proved using each hypothesis at most once. - -% Unlike the previous tactics, the \texttt{Linear} tactic does not belong -% to the initial state of the system, and it must be loaded explicitly -% with the command - -% \begin{coq_example*} -% Require Linear. -% \end{coq_example*} - -% For instance, assuming that \texttt{even} and \texttt{odd} are two -% predicates on natural numbers, and \texttt{a} of type \texttt{nat}, the -% tactic \texttt{Linear} solves the following goal - -% \begin{coq_eval} -% Variables even,odd : nat -> Prop. -% Variable a:nat. -% \end{coq_eval} - -% \begin{coq_example*} -% Lemma example : (even a) -% -> ((x:nat)((even x)->(odd (S x)))) -% -> (EX y | (odd y)). -% \end{coq_example*} - -% You can find examples of the use of \texttt{Linear} in -% \texttt{theories/DEMOS/DemoLinear.v}. -% \begin{coq_eval} -% Abort. -% \end{coq_eval} - -% \begin{Variants} -% \item {\tt Linear with \ident$_1$ \dots\ \ident$_n$}\\ -% \tacindex{Linear with} -% Is equivalent to apply first {\tt generalize \ident$_1$ \dots -% \ident$_n$} (see Section~\ref{generalize}) then the \texttt{Linear} -% tactic. So one can use axioms, lemmas or hypotheses of the local -% context with \texttt{Linear} in this way. -% \end{Variants} - -% \begin{ErrMsgs} -% \item \errindex{Not provable in Direct Predicate Calculus} -% \item \errindex{Found $n$ classical proof(s) but no intuitionistic one}\\ -% The decision procedure looks actually for classical proofs of the -% goals, and then checks that they are intuitionistic. In that case, -% classical proofs have been found, which do not correspond to -% intuitionistic ones. -% \end{ErrMsgs} \subsection{\tt congruence -\tacindex{congruence} -\label{congruence}} - -The tactic {\tt congruence}, by Pierre Corbineau, implements the standard Nelson and Oppen -congruence closure algorithm, which is a decision procedure for ground -equalities with uninterpreted symbols. It also include the constructor theory -(see \ref{injection} and \ref{discriminate}). -If the goal is a non-quantified equality, {\tt congruence} tries to -prove it with non-quantified equalities in the context. Otherwise it -tries to infer a discriminable equality from those in the context. Alternatively, congruence tries to prove that a hypothesis is equal to the goal or to the negation of another hypothesis. - -{\tt congruence} is also able to take advantage of hypotheses stating quantified equalities, you have to provide a bound for the number of extra equalities generated that way. Please note that one of the members of the equality must contain all the quantified variables in order for {\tt congruence} to match against it. - -\begin{coq_eval} -Reset Initial. -Variable A:Set. -Variables a b:A. -Variable f:A->A. -Variable g:A->A->A. -\end{coq_eval} - -\begin{coq_example} -Theorem T: - a=(f a) -> (g b (f a))=(f (f a)) -> (g a b)=(f (g b a)) -> (g a b)=a. -intros. -congruence. -\end{coq_example} - -\begin{coq_eval} -Reset Initial. -Variable A:Set. -Variables a c d:A. -Variable f:A->A*A. -\end{coq_eval} - -\begin{coq_example} -Theorem inj : f = pair a -> Some (f c) = Some (f d) -> c=d. -intros. -congruence. -\end{coq_example} - -\begin{Variants} - \item {\tt congruence {\sl n}}\\ - Tries to add at most {\tt \sl n} instances of hypotheses stating quantified equalities to the problem in order to solve it. A bigger value of {\tt \sl n} does not make success slower, only failure. You might consider adding some lemmas as hypotheses using {\tt assert} in order for congruence to use them. - -\end{Variants} - -\begin{Variants} -\item {\tt congruence with \term$_1$ \dots\ \term$_n$}\\ - Adds {\tt \term$_1$ \dots\ \term$_n$} to the pool of terms used by - {\tt congruence}. This helps in case you have partially applied - constructors in your goal. -\end{Variants} - -\begin{ErrMsgs} - \item \errindex{I don't know how to handle dependent equality} \\ - The decision procedure managed to find a proof of the goal or of - a discriminable equality but this proof couldn't be built in {\Coq} - because of dependently-typed functions. - \item \errindex{I couldn't solve goal} \\ - The decision procedure didn't find any way to solve the goal. - \item \errindex{Goal is solvable by congruence but some arguments are missing. Try "congruence with \dots", replacing metavariables by arbitrary terms.} \\ - The decision procedure could solve the goal with the provision - that additional arguments are supplied for some partially applied - constructors. Any term of an appropriate type will allow the - tactic to successfully solve the goal. Those additional arguments - can be given to {\tt congruence} by filling in the holes in the - terms given in the error message, using the {\tt with} variant - described above. -\end{ErrMsgs} - -\subsection{\tt omega -\tacindex{omega} -\label{omega}} - -The tactic \texttt{omega}, due to Pierre Cr{\'e}gut, -is an automatic decision procedure for Presburger -arithmetic. It solves quantifier-free -formulas built with \verb|~|, \verb|\/|, \verb|/\|, -\verb|->| on top of equalities, inequalities and disequalities on -both the type \texttt{nat} of natural numbers and \texttt{Z} of binary -integers. This tactic must be loaded by the command \texttt{Require Import - Omega}. See the additional documentation about \texttt{omega} -(see Chapter~\ref{OmegaChapter}). - -\subsection{{\tt ring} and {\tt ring\_simplify \term$_1$ \dots\ \term$_n$} -\tacindex{ring} -\tacindex{ring\_simplify} -\comindex{Add Ring}} - -The {\tt ring} tactic solves equations upon polynomial expressions of -a ring (or semi-ring) structure. It proceeds by normalizing both hand -sides of the equation (w.r.t. associativity, commutativity and -distributivity, constant propagation) and comparing syntactically the -results. - -{\tt ring\_simplify} applies the normalization procedure described -above to the terms given. The tactic then replaces all occurrences of -the terms given in the conclusion of the goal by their normal -forms. If no term is given, then the conclusion should be an equation -and both hand sides are normalized. - -See Chapter~\ref{ring} for more information on the tactic and how to -declare new ring structures. - -\subsection{{\tt field}, {\tt field\_simplify \term$_1$\dots\ \term$_n$} - and {\tt field\_simplify\_eq} -\tacindex{field} -\tacindex{field\_simplify} -\tacindex{field\_simplify\_eq} -\comindex{Add Field}} - -The {\tt field} tactic is built on the same ideas as {\tt ring}: this -is a reflexive tactic that solves or simplifies equations in a field -structure. The main idea is to reduce a field expression (which is an -extension of ring expressions with the inverse and division -operations) to a fraction made of two polynomial expressions. - -Tactic {\tt field} is used to solve subgoals, whereas {\tt - field\_simplify \term$_1$\dots\term$_n$} replaces the provided terms -by their reduced fraction. {\tt field\_simplify\_eq} applies when the -conclusion is an equation: it simplifies both hand sides and multiplies -so as to cancel denominators. So it produces an equation without -division nor inverse. +\tacindex{congruence} +\label{congruence}} -All of these 3 tactics may generate a subgoal in order to prove that -denominators are different from zero. +The tactic {\tt congruence}, by Pierre Corbineau, implements the standard Nelson and Oppen +congruence closure algorithm, which is a decision procedure for ground +equalities with uninterpreted symbols. It also include the constructor theory +(see \ref{injection} and \ref{discriminate}). +If the goal is a non-quantified equality, {\tt congruence} tries to +prove it with non-quantified equalities in the context. Otherwise it +tries to infer a discriminable equality from those in the context. Alternatively, congruence tries to prove that a hypothesis is equal to the goal or to the negation of another hypothesis. -See Chapter~\ref{ring} for more information on the tactic and how to -declare new field structures. +{\tt congruence} is also able to take advantage of hypotheses stating quantified equalities, you have to provide a bound for the number of extra equalities generated that way. Please note that one of the members of the equality must contain all the quantified variables in order for {\tt congruence} to match against it. -\Example -\begin{coq_example*} -Require Import Reals. -Goal forall x y:R, - (x * y > 0)%R -> - (x * (1 / x + x / (x + y)))%R = - ((- 1 / y) * y * (- x * (x / (x + y)) - 1))%R. -\end{coq_example*} +\begin{coq_eval} +Reset Initial. +Variable A:Set. +Variables a b:A. +Variable f:A->A. +Variable g:A->A->A. +\end{coq_eval} \begin{coq_example} -intros; field. +Theorem T: + a=(f a) -> (g b (f a))=(f (f a)) -> (g a b)=(f (g b a)) -> (g a b)=a. +intros. +congruence. \end{coq_example} \begin{coq_eval} Reset Initial. +Variable A:Set. +Variables a c d:A. +Variable f:A->A*A. \end{coq_eval} -\SeeAlso file {\tt plugins/setoid\_ring/RealField.v} for an example of instantiation,\\ -\phantom{\SeeAlso}theory {\tt theories/Reals} for many examples of use of {\tt -field}. +\begin{coq_example} +Theorem inj : f = pair a -> Some (f c) = Some (f d) -> c=d. +intros. +congruence. +\end{coq_example} -\subsection{\tt fourier -\tacindex{fourier}} +\begin{Variants} + \item {\tt congruence {\sl n}}\\ + Tries to add at most {\tt \sl n} instances of hypotheses stating quantified equalities to the problem in order to solve it. A bigger value of {\tt \sl n} does not make success slower, only failure. You might consider adding some lemmas as hypotheses using {\tt assert} in order for congruence to use them. -This tactic written by Lo{\"\i}c Pottier solves linear inequalities on -real numbers using Fourier's method~\cite{Fourier}. This tactic must -be loaded by {\tt Require Import Fourier}. +\end{Variants} -\Example -\begin{coq_example*} -Require Import Reals. -Require Import Fourier. -Goal forall x y:R, (x < y)%R -> (y + 1 >= x - 1)%R. -\end{coq_example*} +\begin{Variants} +\item {\tt congruence with \term$_1$ \dots\ \term$_n$}\\ + Adds {\tt \term$_1$ \dots\ \term$_n$} to the pool of terms used by + {\tt congruence}. This helps in case you have partially applied + constructors in your goal. +\end{Variants} -\begin{coq_example} -intros; fourier. -\end{coq_example} +\begin{ErrMsgs} + \item \errindex{I don't know how to handle dependent equality} \\ + The decision procedure managed to find a proof of the goal or of + a discriminable equality but this proof couldn't be built in {\Coq} + because of dependently-typed functions. + \item \errindex{I couldn't solve goal} \\ + The decision procedure didn't find any way to solve the goal. + \item \errindex{Goal is solvable by congruence but some arguments are missing. Try "congruence with \dots", replacing metavariables by arbitrary terms.} \\ + The decision procedure could solve the goal with the provision + that additional arguments are supplied for some partially applied + constructors. Any term of an appropriate type will allow the + tactic to successfully solve the goal. Those additional arguments + can be given to {\tt congruence} by filling in the holes in the + terms given in the error message, using the {\tt with} variant + described above. +\end{ErrMsgs} -\begin{coq_eval} -Reset Initial. -\end{coq_eval} -\subsection{\tt autorewrite with \ident$_1$ \dots \ident$_n$. -\label{tactic:autorewrite} -\tacindex{autorewrite}} +\section{Things that do not fit other sections} -This tactic \footnote{The behavior of this tactic has much changed compared to -the versions available in the previous distributions (V6). This may cause -significant changes in your theories to obtain the same result. As a drawback -of the re-engineering of the code, this tactic has also been completely revised -to get a very compact and readable version.} carries out rewritings according -the rewriting rule bases {\tt \ident$_1$ \dots \ident$_n$}. -Each rewriting rule of a base \ident$_i$ is applied to the main subgoal until -it fails. Once all the rules have been processed, if the main subgoal has -progressed (e.g., if it is distinct from the initial main goal) then the rules -of this base are processed again. If the main subgoal has not progressed then -the next base is processed. For the bases, the behavior is exactly similar to -the processing of the rewriting rules. -The rewriting rule bases are built with the {\tt Hint~Rewrite} vernacular -command. -\Warning{} This tactic may loop if you build non terminating rewriting systems. -\begin{Variant} -\item {\tt autorewrite with \ident$_1$ \dots \ident$_n$ using \tac}\\ -Performs, in the same way, all the rewritings of the bases {\tt \ident$_1$ $...$ -\ident$_n$} applying {\tt \tac} to the main subgoal after each rewriting step. -\item \texttt{autorewrite with {\ident$_1$} \dots \ident$_n$ in {\qualid}} - Performs all the rewritings in hypothesis {\qualid}. -\item \texttt{autorewrite with {\ident$_1$} \dots \ident$_n$ in {\qualid} using \tac} - Performs all the rewritings in hypothesis {\qualid} applying {\tt - \tac} to the main subgoal after each rewriting step. -\item \texttt{autorewrite with {\ident$_1$} \dots \ident$_n$ in \textit{clause}} - Performs all the rewritings in the clause \textit{clause}. \\ - The \textit{clause} argument must not contain any \texttt{type of} nor \texttt{value of}. -\end{Variant} -\SeeAlso Section~\ref{HintRewrite} for feeding the database of lemmas used by {\tt autorewrite}. -\SeeAlso Section~\ref{autorewrite-example} for examples showing the use of -this tactic. -% En attente d'un moyen de valoriser les fichiers de demos -%\SeeAlso file \texttt{contrib/Rocq/DEMOS/Demo\_AutoRewrite.v} -\section{Controlling automation} -\subsection{The hints databases for {\tt auto} and {\tt eauto} -\index{Hints databases} -\label{Hints-databases} -\comindex{Hint}} -The hints for \texttt{auto} and \texttt{eauto} are stored in -databases. Each database maps head symbols to a list of hints. One can -use the command \texttt{Print Hint \ident} to display the hints -associated to the head symbol \ident{} (see \ref{PrintHint}). Each -hint has a cost that is an nonnegative integer, and an optional pattern. -The hints with lower cost are tried first. A hint is tried by -\texttt{auto} when the conclusion of the current goal -matches its pattern or when it has no pattern. -\subsubsection*{Creating Hint databases - \label{CreateHintDb}\comindex{CreateHintDb}} -One can optionally declare a hint database using the command -\texttt{Create HintDb}. If a hint is added to an unknown database, it -will be automatically created. -\medskip -\texttt{Create HintDb} {\ident} [\texttt{discriminated}] -\medskip -This command creates a new database named \ident. -The database is implemented by a Discrimination Tree (DT) that serves as -an index of all the lemmas. The DT can use transparency information to decide -if a constant should be indexed or not (c.f. \ref{HintTransparency}), -making the retrieval more efficient. -The legacy implementation (the default one for new databases) uses the -DT only on goals without existentials (i.e., auto goals), for non-Immediate -hints and do not make use of transparency hints, putting more work on the -unification that is run after retrieval (it keeps a list of the lemmas -in case the DT is not used). The new implementation enabled by -the {\tt discriminated} option makes use of DTs in all cases and takes -transparency information into account. However, the order in which hints -are retrieved from the DT may differ from the order in which they were -inserted, making this implementation observationaly different from the -legacy one. -\begin{Variants} -\item\texttt{Local Hint} \textsl{hint\_definition} \texttt{:} - \ident$_1$ \ldots\ \ident$_n$ - - This is used to declare a hint database that must not be exported to the other - modules that require and import the current module. Inside a - section, the option {\tt Local} is useless since hints do not - survive anyway to the closure of sections. -\end{Variants} -The general -command to add a hint to some database \ident$_1$, \dots, \ident$_n$ is: -\begin{tabbing} - \texttt{Hint} \textsl{hint\_definition} \texttt{:} \ident$_1$ \ldots\ \ident$_n$ -\end{tabbing} -where {\sl hint\_definition} is one of the following expressions: -\begin{itemize} -\item \texttt{Resolve} {\term} - \comindex{Hint Resolve} - - This command adds {\tt apply {\term}} to the hint list - with the head symbol of the type of \term. The cost of that hint is - the number of subgoals generated by {\tt apply {\term}}. - - In case the inferred type of \term\ does not start with a product the - tactic added in the hint list is {\tt exact {\term}}. In case this - type can be reduced to a type starting with a product, the tactic {\tt - apply {\term}} is also stored in the hints list. - - If the inferred type of \term\ contains a dependent - quantification on a predicate, it is added to the hint list of {\tt - eapply} instead of the hint list of {\tt apply}. In this case, a - warning is printed since the hint is only used by the tactic {\tt - eauto} (see \ref{eauto}). A typical example of a hint that is used - only by \texttt{eauto} is a transitivity lemma. - \begin{ErrMsgs} - \item \errindex{Bound head variable} - The head symbol of the type of {\term} is a bound variable such - that this tactic cannot be associated to a constant. - \item \term\ \errindex{cannot be used as a hint} - The type of \term\ contains products over variables which do not - appear in the conclusion. A typical example is a transitivity axiom. - In that case the {\tt apply} tactic fails, and thus is useless. - \end{ErrMsgs} - \begin{Variants} - \item \texttt{Resolve} {\term$_1$} \dots {\term$_m$} - Adds each \texttt{Resolve} {\term$_i$}. - \end{Variants} -\item \texttt{Immediate {\term}} -\comindex{Hint Immediate} - - This command adds {\tt apply {\term}; trivial} to the hint list - associated with the head symbol of the type of {\ident} in the given - database. This tactic will fail if all the subgoals generated by - {\tt apply {\term}} are not solved immediately by the {\tt trivial} - tactic (which only tries tactics with cost $0$). - - This command is useful for theorems such as the symmetry of equality - or $n+1=m+1 \to n=m$ that we may like to introduce with a - limited use in order to avoid useless proof-search. - - The cost of this tactic (which never generates subgoals) is always 1, - so that it is not used by {\tt trivial} itself. - \begin{ErrMsgs} - \item \errindex{Bound head variable} - \item \term\ \errindex{cannot be used as a hint} +\section{Everything after this point has yet to be sorted} - \end{ErrMsgs} - \begin{Variants} +\subsection{\tt constr\_eq \term$_1$ \term$_2$ +\tacindex{constr\_eq} +\label{constreq}} - \item \texttt{Immediate} {\term$_1$} \dots {\term$_m$} +This tactic applies to any goal. It checks whether its arguments are +equal modulo alpha conversion and casts. - Adds each \texttt{Immediate} {\term$_i$}. +\ErrMsg \errindex{Not equal} - \end{Variants} +\subsection{\tt unify \term$_1$ \term$_2$ +\tacindex{unify} +\label{unify}} -\item \texttt{Constructors} {\ident} -\comindex{Hint Constructors} - - If {\ident} is an inductive type, this command adds all its - constructors as hints of type \texttt{Resolve}. Then, when the - conclusion of current goal has the form \texttt{({\ident} \dots)}, - \texttt{auto} will try to apply each constructor. +This tactic applies to any goal. It checks whether its arguments are +unifiable, potentially instantiating existential variables. - \begin{ErrMsgs} +\ErrMsg \errindex{Not unifiable} - \item {\ident} \errindex{is not an inductive type} +\begin{Variants} +\item {\tt unify \term$_1$ \term$_2$ with \ident} - \item {\ident} \errindex{not declared} + Unification takes the transparency information defined in the + hint database {\tt \ident} into account (see Section~\ref{HintTransparency}). +\end{Variants} - \end{ErrMsgs} +\subsection{\tt is\_evar \term +\tacindex{is\_evar} +\label{isevar}} - \begin{Variants} +This tactic applies to any goal. It checks whether its argument is an +existential variable. Existential variables are uninstantiated +variables generated by e.g. {\tt eapply} (see Section~\ref{apply}). - \item \texttt{Constructors} {\ident$_1$} \dots {\ident$_m$} +\ErrMsg \errindex{Not an evar} - Adds each \texttt{Constructors} {\ident$_i$}. +\subsection{\tt has\_evar \term +\tacindex{has\_evar} +\label{hasevar}} - \end{Variants} +This tactic applies to any goal. It checks whether its argument has an +existential variable as a subterm. Unlike {\tt context} patterns +combined with {\tt is\_evar}, this tactic scans all subterms, +including those under binders. -\item \texttt{Unfold} {\qualid} -\comindex{Hint Unfold} - - This adds the tactic {\tt unfold {\qualid}} to the hint list that - will only be used when the head constant of the goal is \ident. Its - cost is 4. +\ErrMsg \errindex{No evars} + +\subsection{\tt is\_var \term +\tacindex{is\_var} +\label{isvar}} - \begin{Variants} +This tactic applies to any goal. It checks whether its argument is a +variable or hypothesis in the current goal context or in the opened sections. - \item \texttt{Unfold} {\ident$_1$} \dots {\ident$_m$} +\ErrMsg \errindex{Not a variable or hypothesis} - Adds each \texttt{Unfold} {\ident$_i$}. +\section{Equality} - \end{Variants} +\subsection{\tt f\_equal +\label{f-equal} +\tacindex{f\_equal}} -\item \texttt{Transparent}, \texttt{Opaque} {\qualid} -\label{HintTransparency} -\comindex{Hint Transparent} -\comindex{Hint Opaque} - - This adds a transparency hint to the database, making {\tt {\qualid}} - a transparent or opaque constant during resolution. This information - is used during unification of the goal with any lemma in the database - and inside the discrimination network to relax or constrain it in the - case of \texttt{discriminated} databases. - - \begin{Variants} +This tactic applies to a goal of the form $f\ a_1\ \ldots\ a_n = f'\ +a'_1\ \ldots\ a'_n$. Using {\tt f\_equal} on such a goal leads to +subgoals $f=f'$ and $a_1=a'_1$ and so on up to $a_n=a'_n$. Amongst +these subgoals, the simple ones (e.g. provable by +reflexivity or congruence) are automatically solved by {\tt f\_equal}. - \item \texttt{Transparent}, \texttt{Opaque} {\ident$_1$} \dots {\ident$_m$} - Declares each {\ident$_i$} as a transparent or opaque constant. - - \end{Variants} +\section{Equality and inductive sets} -\item \texttt{Extern \num\ [\pattern]\ => }\textsl{tactic} -\comindex{Hint Extern} +We describe in this section some special purpose tactics dealing with +equality and inductive sets or types. These tactics use the equality +{\tt eq:forall (A:Type), A->A->Prop}, simply written with the +infix symbol {\tt =}. - This hint type is to extend \texttt{auto} with tactics other than - \texttt{apply} and \texttt{unfold}. For that, we must specify a - cost, an optional pattern and a tactic to execute. Here is an example: +\subsection{\tt decide equality +\label{decideequality} +\tacindex{decide equality}} -\begin{quotation} -\begin{verbatim} -Hint Extern 4 (~(_ = _)) => discriminate. -\end{verbatim} -\end{quotation} +This tactic solves a goal of the form +{\tt forall $x$ $y$:$R$, \{$x$=$y$\}+\{\verb|~|$x$=$y$\}}, where $R$ +is an inductive type such that its constructors do not take proofs or +functions as arguments, nor objects in dependent types. +It solves goals of the form {\tt \{$x$=$y$\}+\{\verb|~|$x$=$y$\}} as well. - Now, when the head of the goal is a disequality, \texttt{auto} will - try \texttt{discriminate} if it does not manage to solve the goal - with hints with a cost less than 4. - - One can even use some sub-patterns of the pattern in the tactic - script. A sub-pattern is a question mark followed by an ident, like - \texttt{?X1} or \texttt{?X2}. Here is an example: +\subsection{\tt compare \term$_1$ \term$_2$ +\tacindex{compare}} -% Require EqDecide. -\begin{coq_example*} -Require Import List. -\end{coq_example*} -\begin{coq_example} -Hint Extern 5 ({?X1 = ?X2} + {?X1 <> ?X2}) => - generalize X1, X2; decide equality : eqdec. -Goal -forall a b:list (nat * nat), {a = b} + {a <> b}. -info auto with eqdec. -\end{coq_example} -\begin{coq_eval} -Abort. -\end{coq_eval} +This tactic compares two given objects \term$_1$ and \term$_2$ +of an inductive datatype. If $G$ is the current goal, it leaves the sub-goals +\term$_1${\tt =}\term$_2$ {\tt ->} $G$ and \verb|~|\term$_1${\tt =}\term$_2$ +{\tt ->} $G$. The type +of \term$_1$ and \term$_2$ must satisfy the same restrictions as in the tactic +\texttt{decide equality}. -\end{itemize} +\subsection{\tt simplify\_eq {\term} +\tacindex{simplify\_eq} +\tacindex{esimplify\_eq} +\label{simplify-eq}} -\Rem One can use an \texttt{Extern} hint with no pattern to do -pattern-matching on hypotheses using \texttt{match goal with} inside -the tactic. +Let {\term} be the proof of a statement of conclusion {\tt + {\term$_1$}={\term$_2$}}. If {\term$_1$} and +{\term$_2$} are structurally different (in the sense described for the +tactic {\tt discriminate}), then the tactic {\tt simplify\_eq} behaves as {\tt + discriminate {\term}}, otherwise it behaves as {\tt injection + {\term}}. + +\Rem If some quantified hypothesis of the goal is named {\ident}, then +{\tt simplify\_eq {\ident}} first introduces the hypothesis in the local +context using \texttt{intros until \ident}. \begin{Variants} -\item \texttt{Hint} \textsl{hint\_definition} - - No database name is given: the hint is registered in the {\tt core} - database. +\item \texttt{simplify\_eq} \num -\item\texttt{Hint Local} \textsl{hint\_definition} \texttt{:} - \ident$_1$ \ldots\ \ident$_n$ + This does the same thing as \texttt{intros until \num} then +\texttt{simplify\_eq \ident} where {\ident} is the identifier for the last +introduced hypothesis. - This is used to declare hints that must not be exported to the other - modules that require and import the current module. Inside a - section, the option {\tt Local} is useless since hints do not - survive anyway to the closure of sections. +\item \texttt{simplify\_eq} \term{} {\tt with} {\bindinglist} + + This does the same as \texttt{simplify\_eq {\term}} but using + the given bindings to instantiate parameters or hypotheses of {\term}. -\item\texttt{Hint Local} \textsl{hint\_definition} +\item \texttt{esimplify\_eq} \num\\ + \texttt{esimplify\_eq} \term{} \zeroone{{\tt with} {\bindinglist}} - Idem for the {\tt core} database. - + This works the same as {\tt simplify\_eq} but if the type of {\term}, + or the type of the hypothesis referred to by {\num}, has uninstantiated + parameters, these parameters are left as existential variables. + +\item{\tt simplify\_eq} + +If the current goal has form $t_1\verb=<>=t_2$, it behaves as +\texttt{intro {\ident}; simplify\_eq {\ident}}. \end{Variants} -% There are shortcuts that allow to define several goal at once: +\subsection{\tt dependent rewrite -> {\ident} +\tacindex{dependent rewrite ->} +\label{dependent-rewrite}} -% \begin{itemize} -% \item \comindex{Hints Resolve}\texttt{Hints Resolve \ident$_1$ \dots\ \ident$_n$ : \ident.}\\ -% This command is a shortcut for the following ones: -% \begin{quotation} -% \noindent\texttt{Hint \ident$_1$ : \ident\ := Resolve \ident$_1$}\\ -% \dots\\ -% \texttt{Hint \ident$_1$ : \ident := Resolve \ident$_1$} -% \end{quotation} -% Notice that the hint name is the same that the theorem given as -% hint. -% \item \comindex{Hints Immediate}\texttt{Hints Immediate \ident$_1$ \dots\ \ident$_n$ : \ident.}\\ -% \item \comindex{Hints Unfold}\texttt{Hints Unfold \qualid$_1$ \dots\ \qualid$_n$ : \ident.}\\ -% \end{itemize} +This tactic applies to any goal. If \ident\ has type +\verb+(existT B a b)=(existT B a' b')+ +in the local context (i.e. each term of the +equality has a sigma type $\{ a:A~ \&~(B~a)\}$) this tactic rewrites +\verb+a+ into \verb+a'+ and \verb+b+ into \verb+b'+ in the current +goal. This tactic works even if $B$ is also a sigma type. This kind +of equalities between dependent pairs may be derived by the injection +and inversion tactics. -%\begin{Warnings} -% \item \texttt{Overriding hint named \dots\ in database \dots} -%\end{Warnings} +\begin{Variants} +\item{\tt dependent rewrite <- {\ident}} +\tacindex{dependent rewrite <-} \\ +Analogous to {\tt dependent rewrite ->} but uses the equality from +right to left. +\end{Variants} +\section{Inversion +\label{inversion}} +\subsection[\tt functional inversion \ident]{\tt functional inversion \ident\label{sec:functional-inversion}} -\subsection{Hint databases defined in the \Coq\ standard library} +\texttt{functional inversion} is a \emph{highly} experimental tactic +which performs inversion on hypothesis \ident\ of the form +\texttt{\qualid\ \term$_1$\dots\term$_n$\ = \term} or \texttt{\term\ = + \qualid\ \term$_1$\dots\term$_n$} where \qualid\ must have been +defined using \texttt{Function} (see Section~\ref{Function}). -Several hint databases are defined in the \Coq\ standard library. The -actual content of a database is the collection of the hints declared -to belong to this database in each of the various modules currently -loaded. Especially, requiring new modules potentially extend a -database. At {\Coq} startup, only the {\tt core} and {\tt v62} -databases are non empty and can be used. +\begin{ErrMsgs} +\item \errindex{Hypothesis {\ident} must contain at least one Function} +\item \errindex{Cannot find inversion information for hypothesis \ident} + This error may be raised when some inversion lemma failed to be + generated by Function. +\end{ErrMsgs} -\begin{description} +\begin{Variants} +\item {\tt functional inversion \num} -\item[\tt core] This special database is automatically used by - \texttt{auto}, except when pseudo-database \texttt{nocore} is - given to \texttt{auto}. The \texttt{core} database contains - only basic lemmas about negation, - conjunction, and so on from. Most of the hints in this database come - from the \texttt{Init} and \texttt{Logic} directories. + This does the same thing as \texttt{intros until \num} then + \texttt{functional inversion \ident} where {\ident} is the + identifier for the last introduced hypothesis. +\item {\tt functional inversion \ident\ \qualid}\\ + {\tt functional inversion \num\ \qualid} -\item[\tt arith] This database contains all lemmas about Peano's - arithmetic proved in the directories \texttt{Init} and - \texttt{Arith} + In case the hypothesis {\ident} (or {\num}) has a type of the form + \texttt{\qualid$_1$\ \term$_1$\dots\term$_n$\ =\ \qualid$_2$\ + \term$_{n+1}$\dots\term$_{n+m}$} where \qualid$_1$ and \qualid$_2$ + are valid candidates to functional inversion, this variant allows to + choose which must be inverted. +\end{Variants} -\item[\tt zarith] contains lemmas about binary signed integers from - the directories \texttt{theories/ZArith}. When required, the module - {\tt Omega} also extends the database {\tt zarith} with a high-cost - hint that calls {\tt omega} on equations and inequalities in {\tt - nat} or {\tt Z}. -\item[\tt bool] contains lemmas about booleans, mostly from directory - \texttt{theories/Bool}. -\item[\tt datatypes] is for lemmas about lists, streams and so on that - are mainly proved in the \texttt{Lists} subdirectory. +\subsection{\tt quote \ident +\tacindex{quote} +\index{2-level approach}} -\item[\tt sets] contains lemmas about sets and relations from the - directories \texttt{Sets} and \texttt{Relations}. +This kind of inversion has nothing to do with the tactic +\texttt{inversion} above. This tactic does \texttt{change (\ident\ + t)}, where \texttt{t} is a term built in order to ensure the +convertibility. In other words, it does inversion of the function +\ident. This function must be a fixpoint on a simple recursive +datatype: see~\ref{quote-examples} for the full details. -\item[\tt typeclass\_instances] contains all the type class instances - declared in the environment, including those used for \texttt{setoid\_rewrite}, - from the \texttt{Classes} directory. -\end{description} +\begin{ErrMsgs} +\item \errindex{quote: not a simple fixpoint}\\ + Happens when \texttt{quote} is not able to perform inversion properly. +\end{ErrMsgs} -There is also a special database called {\tt v62}. It collects all -hints that were declared in the versions of {\Coq} prior to version -6.2.4 when the databases {\tt core}, {\tt arith}, and so on were -introduced. The purpose of the database {\tt v62} is to ensure -compatibility with further versions of {\Coq} for developments done in -versions prior to 6.2.4 ({\tt auto} being replaced by {\tt auto with v62}). -The database {\tt v62} is intended not to be extended (!). It is not -included in the hint databases list used in the {\tt auto with *} tactic. +\begin{Variants} +\item \texttt{quote {\ident} [ \ident$_1$ \dots \ident$_n$ ]}\\ + All terms that are built only with \ident$_1$ \dots \ident$_n$ will be + considered by \texttt{quote} as constants rather than variables. +\end{Variants} -Furthermore, you are advised not to put your own hints in the -{\tt core} database, but use one or several databases specific to your -development. +% En attente d'un moyen de valoriser les fichiers de demos +% \SeeAlso file \texttt{theories/DEMOS/DemoQuote.v} in the distribution -\subsection{\tt Print Hint -\label{PrintHint} -\comindex{Print Hint}} +\section[Classical tactics]{Classical tactics\label{ClassicalTactics}} -This command displays all hints that apply to the current goal. It -fails if no proof is being edited, while the two variants can be used at -every moment. +In order to ease the proving process, when the {\tt Classical} module is loaded. A few more tactics are available. Make sure to load the module using the \texttt{Require Import} command. -\begin{Variants} +\subsection{{\tt classical\_left, classical\_right} \tacindex{classical\_left} \tacindex{classical\_right}} -\item {\tt Print Hint {\ident} } +The tactics \texttt{classical\_left} and \texttt{classical\_right} are the analog of the \texttt{left} and \texttt{right} but using classical logic. They can only be used for disjunctions. +Use \texttt{classical\_left} to prove the left part of the disjunction with the assumption that the negation of right part holds. +Use \texttt{classical\_right} to prove the right part of the disjunction with the assumption that the negation of left part holds. - This command displays only tactics associated with \ident\ in the - hints list. This is independent of the goal being edited, so this - command will not fail if no goal is being edited. +\section{Automatizing +\label{Automatizing}} -\item {\tt Print Hint *} +% EXISTE ENCORE ? +% +% \subsection{\tt Prolog [ \term$_1$ \dots\ \term$_n$ ] \num} +% \tacindex{Prolog}\label{Prolog} +% This tactic, implemented by Chet Murthy, is based upon the concept of +% existential variables of Gilles Dowek, stating that resolution is a +% kind of unification. It tries to solve the current goal using the {\tt +% Assumption} tactic, the {\tt intro} tactic, and applying hypotheses +% of the local context and terms of the given list {\tt [ \term$_1$ +% \dots\ \term$_n$\ ]}. It is more powerful than {\tt auto} since it +% may apply to any theorem, even those of the form {\tt (x:A)(P x) -> Q} +% where {\tt x} does not appear free in {\tt Q}. The maximal search +% depth is {\tt \num}. - This command displays all declared hints. +% \begin{ErrMsgs} +% \item \errindex{Prolog failed}\\ +% The Prolog tactic was not able to prove the subgoal. +% \end{ErrMsgs} -\item {\tt Print HintDb {\ident} } -\label{PrintHintDb} -\comindex{Print HintDb} - This command displays all hints from database \ident. +%% \subsection{{\tt jp} {\em (Jprover)} +%% \tacindex{jp} +%% \label{jprover}} -\end{Variants} +%% The tactic \texttt{jp}, due to Huang Guan-Shieng, is an experimental +%% port of the {\em Jprover}\cite{SLKN01} semi-decision procedure for +%% first-order intuitionistic logic implemented in {\em +%% NuPRL}\cite{Kre02}. -\subsection{\tt Hint Rewrite \term$_1$ \dots \term$_n$ : \ident -\label{HintRewrite} -\comindex{Hint Rewrite}} +%% The tactic \texttt{jp}, due to Huang Guan-Shieng, is an {\it +%% experimental} port of the {\em Jprover}\cite{SLKN01} semi-decision +%% procedure for first-order intuitionistic logic implemented in {\em +%% NuPRL}\cite{Kre02}. -This vernacular command adds the terms {\tt \term$_1$ \dots \term$_n$} -(their types must be equalities) in the rewriting base {\tt \ident} -with the default orientation (left to right). Notice that the -rewriting bases are distinct from the {\tt auto} hint bases and that -{\tt auto} does not take them into account. +%% Search may optionnaly be bounded by a multiplicity parameter +%% indicating how many (at most) copies of a formula may be used in +%% the proof process, its absence may lead to non-termination of the tactic. -This command is synchronous with the section mechanism (see \ref{Section}): -when closing a section, all aliases created by \texttt{Hint Rewrite} in that -section are lost. Conversely, when loading a module, all \texttt{Hint Rewrite} -declarations at the global level of that module are loaded. +%% %\begin{coq_eval} +%% %Variable S:Set. +%% %Variables P Q:S->Prop. +%% %Variable f:S->S. +%% %\end{coq_eval} -\begin{Variants} -\item {\tt Hint Rewrite -> \term$_1$ \dots \term$_n$ : \ident}\\ -This is strictly equivalent to the command above (we only make explicit the -orientation which otherwise defaults to {\tt ->}). +%% %\begin{coq_example*} +%% %Lemma example: (exists x |P x\/Q x)->(exists x |P x)\/(exists x |Q x). +%% %jp. +%% %Qed. -\item {\tt Hint Rewrite <- \term$_1$ \dots \term$_n$ : \ident}\\ -Adds the rewriting rules {\tt \term$_1$ \dots \term$_n$} with a right-to-left -orientation in the base {\tt \ident}. +%% %Lemma example2: (forall x ,P x->P (f x))->forall x,P x->P (f(f x)). +%% %jp. +%% %Qed. +%% %\end{coq_example*} -\item {\tt Hint Rewrite \term$_1$ \dots \term$_n$ using {\tac} : {\ident}}\\ -When the rewriting rules {\tt \term$_1$ \dots \term$_n$} in {\tt \ident} will -be used, the tactic {\tt \tac} will be applied to the generated subgoals, the -main subgoal excluded. +%% \begin{Variants} +%% \item {\tt jp $n$}\\ +%% \tacindex{jp $n$} +%% Tries the {\em Jprover} procedure with multiplicities up to $n$, +%% starting from 1. +%% \item {\tt jp}\\ +%% Tries the {\em Jprover} procedure without multiplicity bound, +%% possibly running forever. +%% \end{Variants} -%% \item -%% {\tt Hint Rewrite [ \term$_1$ \dots \term$_n$ ] in \ident}\\ -%% {\tt Hint Rewrite [ \term$_1$ \dots \term$_n$ ] in {\ident} using {\tac}}\\ -%% These are deprecated syntactic variants for -%% {\tt Hint Rewrite \term$_1$ \dots \term$_n$ : \ident} and -%% {\tt Hint Rewrite \term$_1$ \dots \term$_n$ using {\tac} : {\ident}}. +%% \begin{ErrMsgs} +%% \item \errindex{multiplicity limit reached}\\ +%% The procedure tried all multiplicities below the limit and +%% failed. Goal might be solved by increasing the multiplicity limit. +%% \item \errindex{formula is not provable}\\ +%% The procedure determined that goal was not provable in +%% intuitionistic first-order logic, no matter how big the +%% multiplicity is. +%% \end{ErrMsgs} -\item \texttt{Print Rewrite HintDb {\ident}} - This command displays all rewrite hints contained in {\ident}. +% \subsection[\tt Linear]{\tt Linear\tacindex{Linear}\label{Linear}} +% The tactic \texttt{Linear}, due to Jean-Christophe Filli{\^a}atre +% \cite{Fil94}, implements a decision procedure for {\em Direct +% Predicate Calculus}, that is first-order Gentzen's Sequent Calculus +% without contraction rules \cite{KeWe84,BeKe92}. Intuitively, a +% first-order goal is provable in Direct Predicate Calculus if it can be +% proved using each hypothesis at most once. -\end{Variants} +% Unlike the previous tactics, the \texttt{Linear} tactic does not belong +% to the initial state of the system, and it must be loaded explicitly +% with the command -\subsection{Hints and sections -\label{Hint-and-Section}} +% \begin{coq_example*} +% Require Linear. +% \end{coq_example*} -Hints provided by the \texttt{Hint} commands are erased when closing a -section. Conversely, all hints of a module \texttt{A} that are not -defined inside a section (and not defined with option {\tt Local}) become -available when the module {\tt A} is imported (using -e.g. \texttt{Require Import A.}). +% For instance, assuming that \texttt{even} and \texttt{odd} are two +% predicates on natural numbers, and \texttt{a} of type \texttt{nat}, the +% tactic \texttt{Linear} solves the following goal -\subsection{Setting implicit automation tactics} +% \begin{coq_eval} +% Variables even,odd : nat -> Prop. +% Variable a:nat. +% \end{coq_eval} -\subsubsection[\tt Proof with {\tac}.]{\tt Proof with {\tac}.\label{ProofWith} -\comindex{Proof with}} +% \begin{coq_example*} +% Lemma example : (even a) +% -> ((x:nat)((even x)->(odd (S x)))) +% -> (EX y | (odd y)). +% \end{coq_example*} - This command may be used to start a proof. It defines a default - tactic to be used each time a tactic command {\tac$_1$} is ended by - ``\verb#...#''. In this case the tactic command typed by the user is - equivalent to \tac$_1$;{\tac}. +% You can find examples of the use of \texttt{Linear} in +% \texttt{theories/DEMOS/DemoLinear.v}. +% \begin{coq_eval} +% Abort. +% \end{coq_eval} -\SeeAlso {\tt Proof.} in Section~\ref{BeginProof}. +% \begin{Variants} +% \item {\tt Linear with \ident$_1$ \dots\ \ident$_n$}\\ +% \tacindex{Linear with} +% Is equivalent to apply first {\tt generalize \ident$_1$ \dots +% \ident$_n$} (see Section~\ref{generalize}) then the \texttt{Linear} +% tactic. So one can use axioms, lemmas or hypotheses of the local +% context with \texttt{Linear} in this way. +% \end{Variants} -\begin{Variants} -\item {\tt Proof with {\tac} using {\ident$_1$ \dots {\ident$_n$}}} - Combines in a single line {\tt Proof with} and {\tt Proof using}, - see~\ref{ProofUsing} -\item {\tt Proof using {\ident$_1$ \dots {\ident$_n$}} with {\tac}} - Combines in a single line {\tt Proof with} and {\tt Proof using}, - see~\ref{ProofUsing} +% \begin{ErrMsgs} +% \item \errindex{Not provable in Direct Predicate Calculus} +% \item \errindex{Found $n$ classical proof(s) but no intuitionistic one}\\ +% The decision procedure looks actually for classical proofs of the +% goals, and then checks that they are intuitionistic. In that case, +% classical proofs have been found, which do not correspond to +% intuitionistic ones. +% \end{ErrMsgs} -\end{Variants} -\subsubsection[\tt Declare Implicit Tactic {\tac}.]{\tt Declare Implicit Tactic {\tac}.\comindex{Declare Implicit Tactic}} +\subsection{\tt omega +\tacindex{omega} +\label{omega}} -This command declares a tactic to be used to solve implicit arguments -that {\Coq} does not know how to solve by unification. It is used -every time the term argument of a tactic has one of its holes not -fully resolved. +The tactic \texttt{omega}, due to Pierre Cr{\'e}gut, +is an automatic decision procedure for Presburger +arithmetic. It solves quantifier-free +formulas built with \verb|~|, \verb|\/|, \verb|/\|, +\verb|->| on top of equalities, inequalities and disequalities on +both the type \texttt{nat} of natural numbers and \texttt{Z} of binary +integers. This tactic must be loaded by the command \texttt{Require Import + Omega}. See the additional documentation about \texttt{omega} +(see Chapter~\ref{OmegaChapter}). -Here is an example: +\subsection{{\tt ring} and {\tt ring\_simplify \term$_1$ \dots\ \term$_n$} +\tacindex{ring} +\tacindex{ring\_simplify} +\comindex{Add Ring}} -\begin{coq_example} -Parameter quo : nat -> forall n:nat, n<>0 -> nat. -Notation "x // y" := (quo x y _) (at level 40). +The {\tt ring} tactic solves equations upon polynomial expressions of +a ring (or semi-ring) structure. It proceeds by normalizing both hand +sides of the equation (w.r.t. associativity, commutativity and +distributivity, constant propagation) and comparing syntactically the +results. -Declare Implicit Tactic assumption. -Goal forall n m, m<>0 -> { q:nat & { r | q * m + r = n } }. -intros. -exists (n // m). -\end{coq_example} +{\tt ring\_simplify} applies the normalization procedure described +above to the terms given. The tactic then replaces all occurrences of +the terms given in the conclusion of the goal by their normal +forms. If no term is given, then the conclusion should be an equation +and both hand sides are normalized. -The tactic {\tt exists (n // m)} did not fail. The hole was solved by -{\tt assumption} so that it behaved as {\tt exists (quo n m H)}. +See Chapter~\ref{ring} for more information on the tactic and how to +declare new ring structures. -\section{Generation of induction principles with {\tt Scheme} -\label{Scheme} -\index{Schemes} -\comindex{Scheme}} +\subsection{{\tt field}, {\tt field\_simplify \term$_1$\dots\ \term$_n$} + and {\tt field\_simplify\_eq} +\tacindex{field} +\tacindex{field\_simplify} +\tacindex{field\_simplify\_eq} +\comindex{Add Field}} -The {\tt Scheme} command is a high-level tool for generating -automatically (possibly mutual) induction principles for given types -and sorts. Its syntax follows the schema: -\begin{quote} -{\tt Scheme {\ident$_1$} := Induction for \ident'$_1$ Sort {\sort$_1$} \\ - with\\ - \mbox{}\hspace{0.1cm} \dots\\ - with {\ident$_m$} := Induction for {\ident'$_m$} Sort - {\sort$_m$}} -\end{quote} -where \ident'$_1$ \dots\ \ident'$_m$ are different inductive type -identifiers belonging to the same package of mutual inductive -definitions. This command generates {\ident$_1$}\dots{} {\ident$_m$} -to be mutually recursive definitions. Each term {\ident$_i$} proves a -general principle of mutual induction for objects in type {\term$_i$}. +The {\tt field} tactic is built on the same ideas as {\tt ring}: this +is a reflexive tactic that solves or simplifies equations in a field +structure. The main idea is to reduce a field expression (which is an +extension of ring expressions with the inverse and division +operations) to a fraction made of two polynomial expressions. -\begin{Variants} -\item {\tt Scheme {\ident$_1$} := Minimality for \ident'$_1$ Sort {\sort$_1$} \\ - with\\ - \mbox{}\hspace{0.1cm} \dots\ \\ - with {\ident$_m$} := Minimality for {\ident'$_m$} Sort - {\sort$_m$}} +Tactic {\tt field} is used to solve subgoals, whereas {\tt + field\_simplify \term$_1$\dots\term$_n$} replaces the provided terms +by their reduced fraction. {\tt field\_simplify\_eq} applies when the +conclusion is an equation: it simplifies both hand sides and multiplies +so as to cancel denominators. So it produces an equation without +division nor inverse. - Same as before but defines a non-dependent elimination principle more - natural in case of inductively defined relations. +All of these 3 tactics may generate a subgoal in order to prove that +denominators are different from zero. -\item {\tt Scheme Equality for \ident$_1$\comindex{Scheme Equality}} +See Chapter~\ref{ring} for more information on the tactic and how to +declare new field structures. - Tries to generate a boolean equality and a proof of the - decidability of the usual equality. If \ident$_i$ involves - some other inductive types, their equality has to be defined first. +\Example +\begin{coq_example*} +Require Import Reals. +Goal forall x y:R, + (x * y > 0)%R -> + (x * (1 / x + x / (x + y)))%R = + ((- 1 / y) * y * (- x * (x / (x + y)) - 1))%R. +\end{coq_example*} -\item {\tt Scheme Induction for \ident$_1$ Sort {\sort$_1$} \\ - with\\ - \mbox{}\hspace{0.1cm} \dots\\ - with Induction for {\ident$_m$} Sort - {\sort$_m$}} +\begin{coq_example} +intros; field. +\end{coq_example} - If you do not provide the name of the schemes, they will be automatically - computed from the sorts involved (works also with Minimality). +\begin{coq_eval} +Reset Initial. +\end{coq_eval} -\end{Variants} +\SeeAlso file {\tt plugins/setoid\_ring/RealField.v} for an example of instantiation,\\ +\phantom{\SeeAlso}theory {\tt theories/Reals} for many examples of use of {\tt +field}. -\SeeAlso Section~\ref{Scheme-examples} - -\subsection{Automatic declaration of schemes} -\comindex{Set Equality Schemes} -\comindex{Set Elimination Schemes} -It is possible to deactivate the automatic declaration of the induction - principles when defining a new inductive type with the - {\tt Unset Elimination Schemes} command. It may be -reactivated at any time with {\tt Set Elimination Schemes}. -\\ - -You can also activate the automatic declaration of those boolean equalities -(see the second variant of {\tt Scheme}) with the {\tt Set Equality Schemes} - command. However you have to be careful with this option since -\Coq~ may now reject well-defined inductive types because it cannot compute -a boolean equality for them. - -\subsection{\tt Combined Scheme\label{CombinedScheme} -\comindex{Combined Scheme}} -The {\tt Combined Scheme} command is a tool for combining -induction principles generated by the {\tt Scheme} command. -Its syntax follows the schema : - -\noindent -{\tt Combined Scheme {\ident$_0$} from {\ident$_1$}, .., {\ident$_n$}}\\ -\ident$_1$ \ldots \ident$_n$ are different inductive principles that must belong to -the same package of mutual inductive principle definitions. This command -generates {\ident$_0$} to be the conjunction of the principles: it is -built from the common premises of the principles and concluded by the -conjunction of their conclusions. - -\SeeAlso Section~\ref{CombinedScheme-examples} - -\section{Generation of induction principles with {\tt Functional Scheme} -\label{FunScheme} -\comindex{Functional Scheme}} - -The {\tt Functional Scheme} command is a high-level experimental -tool for generating automatically induction principles -corresponding to (possibly mutually recursive) functions. Its -syntax follows the schema: -\begin{quote} -{\tt Functional Scheme {\ident$_1$} := Induction for \ident'$_1$ Sort {\sort$_1$} \\ - with\\ - \mbox{}\hspace{0.1cm} \dots\ \\ - with {\ident$_m$} := Induction for {\ident'$_m$} Sort - {\sort$_m$}} -\end{quote} -where \ident'$_1$ \dots\ \ident'$_m$ are different mutually defined function -names (they must be in the same order as when they were defined). -This command generates the induction principles -\ident$_1$\dots\ident$_m$, following the recursive structure and case -analyses of the functions \ident'$_1$ \dots\ \ident'$_m$. +\subsection{\tt fourier +\tacindex{fourier}} +This tactic written by Lo{\"\i}c Pottier solves linear inequalities on +real numbers using Fourier's method~\cite{Fourier}. This tactic must +be loaded by {\tt Require Import Fourier}. -\paragraph{\texttt{Functional Scheme}} -There is a difference between obtaining an induction scheme by using -\texttt{Functional Scheme} on a function defined by \texttt{Function} -or not. Indeed \texttt{Function} generally produces smaller -principles, closer to the definition written by the user. +\Example +\begin{coq_example*} +Require Import Reals. +Require Import Fourier. +Goal forall x y:R, (x < y)%R -> (y + 1 >= x - 1)%R. +\end{coq_example*} +\begin{coq_example} +intros; fourier. +\end{coq_example} -\SeeAlso Section~\ref{FunScheme-examples} +\begin{coq_eval} +Reset Initial. +\end{coq_eval} \section{Simple tactic macros @@ -4389,7 +4501,7 @@ principles, closer to the definition written by the user. \comindex{Tactic Definition} \label{TacticDefinition}} -A simple example has more value than a long explanation: +A simple example has more value than a long explanation: \begin{coq_example} Ltac Solve := simpl; intros; auto. @@ -4399,7 +4511,7 @@ Ltac ElimBoolRewrite b H1 H2 := The tactics macros are synchronous with the \Coq\ section mechanism: a tactic definition is deleted from the current environment -when you close the section (see also \ref{Section}) +when you close the section (see also \ref{Section}) where it was defined. If you want that a tactic macro defined in a module is usable in the modules that require it, you should put it outside of any section. @@ -4408,8 +4520,8 @@ Chapter~\ref{TacticLanguage} gives examples of more complex user-defined tactics. -%%% Local Variables: +%%% Local Variables: %%% mode: latex %%% TeX-master: "Reference-Manual" %%% TeX-master: "Reference-Manual" -%%% End: +%%% End: -- cgit v1.2.3