| Commit message (Collapse) | Author | Age |
|---|
| | |
|
| |
|
|
|
| |
Function is_constructor was not properly fixed. Additionally, this fixes
a problem with the 8.5 interpretation of in-pattern (see Cases.v).
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Clarification of the existence of three algorithms for solving
unconstrained evars:
- the type-class mechanism
- the heuristics for solving pending conversion problems and multi-candidates
- Declare Implicit Tactic (when called from tactics)
Main function for solving unconstrained evars (when not using
understand): Pretyping.solve_remaining_evars
- Clarification of the existence of three corresponding kinds of
errors when reporting about unsolved evars:
Main function for checking resolution of evars independently of the
understand functions: Pretyping.check_evars_are_solved
- Introduction of inference flags in pretyping for governing which
combination of the algorithms to use when calling some understand
function; there is also a flag of expanding or not evars and for
requiring or not the resolution of all evars
- Less hackish way of managing Pretyping.type_constraint: all three
different possibilities are now represented by three different
constructors
- Main semantical changes done:
- solving unconstrained evars and reporting is not any longer mixed:
one first tries to find unconstrained evars by any way possible;
one eventually reports on the existence of unsolved evars using
check_evars_are_solved
- checking unsolved evars is now done by looking at the evar map,
not by looking at the evars occurring in the terms to pretype; the
only observed consequence so far is in Cases.v because of subterms
(surprisingly) disappering after compilation of pattern-matching
- the API changed, see dev/doc/changes.txt
Still to do:
- Find more uniform naming schemes:
- for distinguishing when sigma is passed as a reference or as a value
(are used: suffix _evars, prefix e_)
- for distinguishing when evars are allowed to remain uninstantiated or not
(are used: suffix _evars, again, suffix _tcc, infix _open_)
- be more consistent on the use of names evd/sigma/evars or evdref/evars
- By the way, shouldn't "understand" be better renamed into "infer" or
"preinfer", or "pretype". Grammatically, "understanding a term" looks
strange.
- Investigate whether the inference flags in tacinterp.ml are really
what we want (e.g. do we really want that heuristic remains
activated when typeclasses are explicitly deactivated, idem in
Tacinterp.interp_open_constr where flags are strange).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16499 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16129 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
|
|
|
|
| |
an evar in typing.ml. Thanks to HoTT people for noticing the problem.
Fixed behavior of e_typing
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16092 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
|
|
|
|
| |
Reset and the other backtracking commands (Back, BackTo, Backtrack)
are now allowed only during interactive session, not in compiled
or loaded scripts.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15087 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14700 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@12337 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
help the type _checking_ anymore (they don't become typing constraints)
but they permit to coerce a subterm in a type. In particular, when
using a VM cast we avoid unneeded, unexpected conversions using the
default machine (oops!). Also remove the corresponding comment in
pretyping and fix the wrong use of casts in toplevel/command: accept
the trouble of using evars. This has the somewhat adverse effect that
when typing casted object we now have no typing constraints (see
e.g. examples in Cases.v)!
Probably, this will be backtracked partially tomorrow as many contribs
can rely on it and the change could make some unifications fail (in
particular with deep coercions). Let's try anyway!
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11558 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
du filtrage. Cela permet de détecter les cas impossibles et de simuler
les contraintes d'inversion exprimables sous la forme d'un assignement
des arguments du constructeurs (cf le cas de Vtail dans Bvector.v).
Si l'on filtre sur t:I u1 .. un, et que chaque ui a la forme vi(wi)
avec vi composé uniquement de constructeurs, et que le résultat final
est P(w1,...,wn) (qui est éventuellement lui-même une evar) alors on
construit le prédicat
Q:=fun x1 .. xn y =>
match x1 .. xn y with
| v1(z) .. vn(z) t => P(z)
| _ .. _ _ => ?evar-speciale-cas-impossible
end
qui vérifiera bien que Q u1 .. un = P(w1,..,wp).
En raison de limitations de l'unification (on aurait besoin d'eta
conversion pour résoudre des problèmes du genre
"terme rigide == match x with _ => ?evar end", et besoin d'instanciation par
constructeurs pour des cas comme "A(y) = match ?evar with C x => A(x) end"),
je n'ai pas réussi à traiter le cas général.
Aussi, on adopte une stratégie pragmatique consistant à tester
plusieurs prédicats possibles :
- si un type final est donné, on essaie d'abord l'algorithme de
Matthieu et sinon le nouvel algorithme (permet par exemple de traiter
certains cas d'élimination dépendante de Bvector.v),
- s'il n'y a pas de type final, on essaie d'abord le nouvel algo et
sinon, on essaie avec un prédicat sans dépendance (permet de traiter
des cas compliqués comme celui de par cas sur I' dans le fichier
Case13.v de la test-suite).
Dans la pratique, il y a beaucoup de changement dans le code de compile_case.
- Par exemple, la compilation est maintenant toujours appelé avec un
prédicat (là où l'on pouvait avoir None, on a maintenant toujours au
moins une evar).
- En revanche, le membre droit des clauses est maintenant
optionnel. Si c'est None, c'est qu'on se trouve dans le cas d'une
branche impossible au moment du calcul du prédicat de retour.
- Aussi, on renonce aux PrLetIn et PrProd dans l'expression du
predicat de retour mais il faut savoir que c'est maintenant la liste
des tomatchs qui spécifie le contexte exact dans lequel le prédicat
de retour est bien typé.
- Et d'autres...
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10883 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@7693 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@4891 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
|
|
| |
PeanoSyntax.v)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@3163 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@2222 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@2095 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
|
|
| |
actuellement non implantés
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@1996 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@1651 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@1469 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@1462 85f007b7-540e-0410-9357-904b9bb8a0f7
|
| |
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@1375 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@1099 85f007b7-540e-0410-9357-904b9bb8a0f7
|
Hugo Herbelin
herbelin
pboutill
letouzey
glondu
msozeau
mohring
filliatr