| Commit message (Collapse) | Author | Age |
| |
|
| |
|
|
|
|
|
| |
any length with a [None] representation and ensure that this representation
is canonical through the restricted interface.
|
|
|
|
|
| |
This completes the port of the native compiler to retroknowledge.
However, some testing and optimizations are still to be done.
|
|
|
|
| |
dependencies.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
(hopefully), and forbids generic equality. Still, it allows generic hash.
|
|
|
|
|
|
|
|
|
|
|
| |
the provided type to come with a hashing function. The internal representation
is changed, such that values are first compared w.r.t. to their hash.
This effectively saves a lot of comparisons which may be far more expensive
than O(1), as in the string case, hence resulting in an overall speedup.
CAVEAT: everything is not implemented yet, and order-sensitive functions
now do not respect the provided order anymore.
|
| |
|
|
|
|
|
|
|
| |
There are currently two other clashs : Lexer and Errors, but for
the moment these ones haven't impacted my experiments with extraction
and compiler-libs, while this Matching issue had. And anyway the new
name is more descriptive, in the spirit of the recent TacticMatching.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Anyway, a few syntactic features of 3.12 were already used here and
there (e.g. local opening via Foo.(...), or the record shortcut
{ field; ... }). Hence compiling with 3.11 wasn't working anymore.
Already take advantage of the following 3.12.1 features :
- "module type of ..." in CArray, CList, CString ...
- "ocamldep -ml-synonym" : no need anymore to hack the ocamldep output
via our coqdep to localize the .ml4 modules :-)
The -ml-synonym option (+ various bugfixes) is the reason for asking
3.12.1 directly and not just 3.12.0. After all, if debian stable is
providing 3.12.1, then everybody has it ;-)
|
|
|
|
|
| |
NB: new file miscprint.ml deserves to be part of printing.cma,
but should be part of proofs.cma for the moment, due to use in logic.ml
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Yet another revision of the build system. We avoid relying on the awkward
include-which-fails-but-works-finally-after-a-retry feature of gnu make.
This was working, but was quite hard to understand. Instead, we reuse
the idea of two explicit phases (as in 2007 and its stage{1,2,3}), but
in a lighter way. The main Makefile calls Makefile.build twice :
- first for building grammar.cma (and q_constr.cmo), with a restricted
set of .ml4 to consider (see variables BUILDGRAMMAR and ML4BASEFILES).
- then on the true target(s) asked by the user (using the special variable
MAKECMDGOALS).
In pratice, this should change very little to the concrete developper's life,
let me know otherwise. A few more messages of make due to the explicit
first sub-call, but no more strange "not ready yet" messages...
Btw: we should handle correctly now the parallel compilation of multiple
targets (e.g. make -jX foo bar). As reported by Pierre B, this was
triggering earlier two separate sub-calls to Makefile.build, one
for foo, the other for bar, with possibly nasty interactions in case
of parallelism.
In addition, some cleanup of Makefile.build, removal of useless :: rules,
etc etc.
|
| |
|
|
|
| |
This commit also introduces a module Monad to generate monadic combinators (currently, only List.map).
|
|
|
|
|
| |
Make this module deal only with opaque proofs.
Make discharging/substitution invariant more explicit via a third constructor.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For a file dir/a.v the corresponding aux file dir/.a.aux can store
arbitrary data. It maps a "Loc.t * string" (key) to a "string" (value).
Pretty much anything can fit in this schema, but ATM I see only the
following possible uses:
1) record inferred data, like the set of section variable used, so that
one can later use this info to process proofs asynchronously (i.e.
compute their discharged type without knowing the proof term).
2) record timings (how long it takes to build a proof term or check it),
so that one can take smarter scheduling decisions
3) record a bloated proof trace for automatic tactics, so that one can
replay it faster (a sort of cache). For that to be useful an Ltac
API is required.
The .aux file contains the path of the .v and its md5 hash. When loaded
it defaults to the empty map is the file is not there or if the .v file
changed.
Not finding some data in the .aux file cannot be a failure, but finding
it may help in many ways.
The current file format is very simple but human readable. It is
generated/parsed using printf/scanf and in particular the %S formatter
for the value string. The file format is private to the Aux_file
module: only an abstract interface is provided.
The current file format is not robust in face of local changes.
Any change invalidates the md5 hash (and the Loc.t is very likely to
change too).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Experimental. Turned out to be much harder to implement than I thought. The main
issue is that the reification in the native compiler and the VM is not quite
untyped. Indeed, type annotations for lambdas have to be reconstructed. Hence,
when reifying an application u = t a1 ... an, the type of t has to be known or
reconstructed. It is always possible to do so in plain CIC, when u is in normal
form and its type is known. However, with partial terms this may no longer be
the case, as in: ?1 a1 ... an. So we also compile and evaluate the type of
evars and metas.
This still has to be tested more extensively, but the correction of the kernel
native conversion (on terms without evars or metas) should not be impacted.
Much of this could be reused for the VM.
|
|
|
|
| |
internalization time.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
configure is now just a minimal wrapper around the new configure.ml.
This configure.ml is runned with the same ocaml used during
compilation, and starts with a #load "unix.cma".
For now, this new configure script is meant to be 99% compatible
with the old one. Known incompatibilities : the --foo option format
(with two --) isn't supported anymore, use -foo options instead.
Let me know if you encounter any other changes.
Internals:
- We use our own "run" command (based on Unix.create_process) to avoid
relying on some specific shell (/bin/sh or cmd.exe).
- We should have far less issues with filename quoting under windows
since we almost never rely on (cygwin) shell anymore. This remains
to be fully tested, though.
- dev/ocamldebug-coq is slightly different now, to ease its generation
|
| |
|
|
|
|
| |
parsing is plugged.
|
|
|
|
|
|
|
|
| |
Since the new proof engine, Hiddentac has been essentially trivial.
Here is what happened to the functions defined there
- Aliases, or tactics that were trivial to inline were systematically inlined
- Tactics used only in tacinterp have been moved to tacinterp
- Other tactics have been moved to a new module Tactics.Simple.
|
|
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@17095 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
by extraction.
The goal was to use Coq's partial evaluation capabilities to do manually some
inlining that Ocaml couldn't do. It may be critical as we are defining higher
order combinators in term of others and no inlining means a lot of
unnecessary, short-lived closures built.
With this modification we get back some (but not all) of the loss of performance introduced by threading the monadic type all over the place.
I still have an estimated 15% longer compilation time for Coq.
Makes use of Set Extraction Conservative Types and Set Extraction File Comment
to maintain the relationship between the functions and their types.
Uses an intermediate layer Proofview_monad between Proofview_gen and
Proofview in order to use a hand-written mli to catch potential errors in the
generated file (it uses Extract Constant a lot).
A bug in the extraction of signatures forces to remove the generated
proofview_gen.mli which does not have the correct types.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16981 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16979 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
|
|
| |
But for vm, the kernel should be functional now
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16961 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
|
|
|
|
| |
- make unification try canonical structures before expansion as in evar_conv
- add a fast path to evar inversion (patch from B. Ziliani).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16945 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
|
|
|
|
|
| |
Ideally all unmarshallable content in the state should be
stocked using Ephemeron keys. In this way the state becomes
always marshallable (because the unmarshallable content is magically
dropped). The mli contains more detailed doc.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16891 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16802 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
|
|
|
|
|
|
|
|
| |
casts of ints to evars.
- 2 in Evarutil and Goal which are really needed, even though the Goal
one could (and should) be removed;
- 2 in G_xml and Detyping that are there for completeness sake, but
that might be made anomalies altogether;
- 1 in Newring which is quite dubious at best, and should be fixed.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16786 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16765 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
|
|
|
|
|
|
| |
The extended signature is defined in CMap, and should be compatible
with the old one, except that module arguments have to be explicitely
named. The implementation itself is quite unsafe, as it relies on the
current implementation of OCaml maps, even though that should not be
a problem (it has not changed in ages).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16735 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16717 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The process_transaction function adds a new edge to the Dag without
executing the transaction (when possible).
The observe id function runs the transactions necessary to reach to the
state id. Transaction being on a merged branch are not executed but
stored into a future.
The finish function calls observe on the tip of the current branch.
Imperative modifications to the environment made by some tactics are
now explicitly declared by the tactic and modeled as let-in/beta-redexes
at the root of the proof term. An example is the abstract tactic.
This is the work described in the Coq Workshop 2012 paper.
Coq is compile with thread support from now on.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16674 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since the whole system is imperative, futures are run protecting
the global state, and the final state is also saved to let the user
freely chain futures.
Futures can represent local (lazy) computations or remote ones
(delegated). Delegating a future lets a third party assign its
value at some poit in the future; in the meanwhile accessing the
future value raises an exception.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16669 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
|
|
| |
Now we at least print the type of the offending object.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16657 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16618 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
|
|
|
|
|
|
| |
level of generic arguments. This only matters at parsing time.
TODO: the current status is not satisfactory enough, as rule
emptyness is still decided w.r.t. generic arguments. This should be
done on a grammar entry basis instead.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16617 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
|
|
|
|
|
|
|
|
| |
through a unique generic argument, and the level is only considered
at parsing time.
This may introduce unnecessary parentheses in Ltac printing though,
as every tactic argument is collapsed at the lowest level. I assume
this does not matter that much, and anyway Ltac printing is quite
bugged as of today.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16616 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16610 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
1. Genarg itself which only defines the abstract datatypes needed.
2. Genintern, first file of interp/, defining the intern and subst
functions.
3. Geninterp, first file of tactics/, defining the interp function.
4. Genprint, first file of printing/, dealing with the printers.
The Genarg file has no dependency and is in lib/, so that we can put
generic arguments everywhere, and in particular in ASTs.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16601 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
|
|
|
| |
related types. This will ultimately allow putting genargs into
these ASTs.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16600 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
|
|
| |
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16594 85f007b7-540e-0410-9357-904b9bb8a0f7
|