| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
Following a remark by Jason Gross and Daniel Grayson.
|
|
monad.
|
|
|
|
hypothesis.
|
|
|
|
|
|
|
|
|
|
|
|
To make this possible the state id has to reach the kernel.
Hence definition_entry has an extra field, and many files had
to be fixed.
|
|
|
|
|
|
|
|
|
|
internalization time.
|
|
|
|
interpreter.
|
|
It triggered nonsensical behaviour of list-using tactic
notation. Hopefully or not, nobody uses such notations out of
the test-suite...
|
|
|
|
Actually, this was wrong, as evars should not appear until interpretation.
Evarmaps were only passed around uselessly, and often fed with dummy or
irrelevant values.
|
|
If these messages are still relevent to somebody, feel free to
restore them in -debug or any non-default mode of your choice.
|
|
|
|
|
|
I doubt [catching_error] is performance critical in any way. But it looked silly to allocate a block to [(inner_trace,e)] when [e] was known in advance (and was already named [e]).
|
|
synchrone.
The previous heuristic is to check whether the proof ends with Qed or not. This modification allows for commands which start proof but may produce transparent term even when the function ends with Qed.
|
|
|
|
|
|
|
|
and instantiate. Each of these tactics uses it at a different
parsing level, thus actually triggering a parsing error after
merging them. This fix implies some code duplication, which is
a pity. The separation between genargs and parsing entries should
be made one day.
|
|
|
|
refine tactic, which now uses plain glob_constr's. Now there
is no real need to depend on goal when interpreting genargs.
Possible minor incompatibilities:
1. The interpretation of glob_constr to constr is now done by
Goal.constr_of_raw, which may be slightly dumbier than the dedicated
Tacinterp.interp_open_constr which tries harder. Stdlib and test-suite
do go through, though.
2. I had to change the parsing level of wit_glob in Extraargs
from lconstr to constr. It may break ML notations using glob, but
as they are only used inside Coq code and all well-parenthezised,
it should be OK.
|
|
|
|
|
|
parsing is plugged.
|
|
provided tactic environment to handle open terms.
|
|
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.
|
|
There are many functions in Tacinterp which use a goal as a proxy for the pair of an environment and an evar_map. But do not build LCF tactics with them. They don't play well with the new style of tactics.
I've changed some of them to explicitely take an environment and an evar_map instead.
|
|
|
|
anomaly is raised. As there are very few tags defined in Coq code, this is
very unlikely to appear, and can be fixed by tweaking the name of the
dynamic argument.
This should be more efficient, as we did compare equal strings each time.
|
|
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@17094 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
multiple successes.
The semantics was a bit strange: an immediate failure of a branch would cause the pattern matching to backtrack, and failure after the match could cause the selected tactic to backtrack. However, the pattern matching process won't backtrack after it has selected a tactic with at least one success.
To avoid changing the semantics of pattern matching with single-success tactics, I made it so that it is impossible to backtrack into a pattern matching once done.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@17090 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
The implementation was partly based on IStream and partly on a control flow with exception. The latter does not mix well with the monadic tactics.
I've moved the algorithmic part of pattern-matching to a new file (tactics/tacticMatching.ml), in order to de-entangle the pattern-matching procedure from the interpretation. This shaves off 300 lines of code from Tacinterp, which is still over 2000 lines of code. It is a first step towards refactoring tacinterp. To be fair, part of what disapeared are lines which sent messages to the debugger. I was not too concerned with them because I understand people found the debugger much too fine-grain on Ltac's pattern matching. But conversely, there may be too few debugging hooks now. This is worth looking into.
In TacticMatching itself, I used a monadic style to express the pattern-matching procedure concisely. I implemented the monad in a fairly brute-force way, using the existing primitive of IStream. It may be worth experimenting with specialized primitive. I am less worried about the monadic style than about the number of allocation of list cells that the primitives entail.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@17089 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@17087 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@17082 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@17080 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
Instead of putting the body directly in the AST, we register it in a table.
This time it should work properly. Tactic notation are given kernel names to
ensure the unicity of their contents.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@17079 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
Tactics notation interpretation was messed up because of the use of
identical keys for different notations. All my tentative fixes were
unsuccessful, so better blankly revert the commit for now.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@17078 85f007b7-540e-0410-9357-904b9bb8a0f7
|
