aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2010-06-13Fixing bug 2295 (omission of option "as" in return clause of "match" was notherbelin
working with section/goal variables). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13131 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-13Fixing bug 2300 (ltac pattern-matching returning terms with concrete universes).herbelin
By the way, there is an open problem of which conversion to use (conv, evarconv, with or w/o universes levels) when trying to unify multiple instances of the same variable in ltac pattern-matching. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13130 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-13Addressed bug #2320 in v8.2 and v8.3 branches ("refine" problem withherbelin
metas in return predicate of "match"); propagated protection for #2310 ("refine" problem with dependent metas of higher-order type) in v8.3 to v8.2. Everything goes well in trunk thanks for the new proof engine. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13129 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-13Fixing definition of set_map (bug report #2111) which was actually alreadyherbelin
incorrect when it was introduced in 1998. Proofs about it remain to be done... git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13128 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-13Addressed bug #2310 and replaced anomaly "unknown meta" raised by "refine"herbelin
by a regular error in v8.3. Example behaves better in trunk thanks to new proof engine. In fact, there are two distinct solutions to a unification problem: should "refine" commit to one of them or leave the problem open? For trunk, improved the unification error message by enforcing nf_evar (but at some time, nf_evar in error messages should move to himsg because it is costly when errors are used for backtracking). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13127 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-13Fixed bug #2314 (inversion using not checking the correctness of its argumentsherbelin
enough) git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13126 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-13Made intros_until and onInductionArg a bit stricter and robustherbelin
The tolerance for overloading "id" as quantified hypothesis and as declared variable is kept - because induction_arg entry is not available to extended tactics, e.g. "discriminate", and these extensions do not know a priori if a name is quantified or declared. However, an upstream check is done to ensure that an induction argument exists as term if not quantified so that the tactics do not have to check this individually by themselves. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13125 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-13Fixed a bug in pretty-printing "induction" and "destruct" due to aherbelin
swap in the evar flags and the isrec flag. (e.g. "induction" was printed "edestruct"). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13124 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-12Fixed bug #2135 (second-order unification was raising cryptic message)herbelin
- made the example work (a call to whd_meta was missing) - replaced the internal error messages of w_unify_to_subterm_list into user-understandable messages - incidentally fixed the meaning of whd_meta (which now takes an evd) and meta_name (which now does what it means and do not treat differently the instantiated metas) git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13122 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-12Added regression tests for bugs + miscellaneous improvementsherbelin
- #2095 (not fixed in v8.2 but fixed in v8.3 and trunk) - #2108 (fixed in v8.2, v8.3, trunk) - #2102 (moved warning to msg_warning to ensure flushing, but still open enhancement) git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13121 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-12Applying François' patch fixing grammar of uniform inheritance condition ↵herbelin
message (see bug report #2331) git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13120 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-12Fixing spelling: pr_coma -> pr_commaherbelin
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13119 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-12Improved the inference of the return predicate in dependent pattern-matching.herbelin
More precisely, the mecanism used to automatically infer return predicates of the form "as x in I y1..yn match y1..yn x with u1(z1)..un(zn) => P(z1..zn) | _ => ID end" now computes the dependencies in the types of y1..yn and x. This allows it to benefit of the generalisation mechanism of the pattern-matching compilation algorithm ("Abstract") and to infer more sophisticated return predicates (e.g. when working with "vector"). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13118 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-12Added rudimentary support for using constructors from the explicitherbelin
substitution of evars when solving equation "?n[subst] = t": this is a quite common useful heuristic for inferring the return predicate of "match". Made incidentally a minor simplification of expand_full_opt. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13117 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-12Fixed a bug in evarutil.ml (wrong env of a postponed conversion problem).herbelin
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13116 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-12Added debugging printer for the idmap used at evar definition time forherbelin
compacting the explicit substitution of the evars. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13115 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-11Reverted commit 13110 about headers.sty that I wrongly thought was buggy. Sorry.herbelin
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13114 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-11Mainly made that evarconv is able to solve "?n = (fun x => x) ?n" (sic).herbelin
Use two ways to solve it: - added a whd_betaiota in solve_simple_eqn (since evarconv itself refuses beta to preserve the opportunities of first-order-matching expressions of the form "(fun x => P) t"; an advantage of this whd_betaiota is also that it may simplify K-redexes. - also added a last-chance test in case of failing occur-check by trying to fully head-normalize (with delta) the right-hand-side (allows to solve for instance "?n = id ?n" where id is a constant (a bridled form of solve_refl that use fconv instead of evar_conv_x). Incidentally improved a bit the rendering of the type of generalized terms in pattern-matching by using whd_betaiota. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13113 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-10Fixed two bugs in pattern-matching compilationherbelin
- first bug was a missing lift when the return predicate was assumed to be a simple evar - second problem was in the inference of a custom inversion predicate in case of matching on a term in a constrained type (like "vect (S n)"): the rel bound to local definitions in the instance of evars where instantiated (by other evars) which is wrong per se but which contradicted the assumptions of find_projectable_vars in evarutil.ml which assumed that this did not occur; solved the problem by not instantiating references to local definition which (probably) will not loose some opportunity of unification in presence of non unfolded local definitions, ... - also cleaned a bit prepare_predicate git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13112 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-10Fixed a very old (from V6.3) typo in headers.styherbelin
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13110 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-10Extraction Implicits: can accept argument names instead of just positionsletouzey
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13109 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-10Fix build with OCaml 3.12glondu
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13107 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-09Fixed bug # 2303 in r 13087.msozeau
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13104 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-09Tentative fix for bug #2226: put inj_pair2 and eq_dep_eq hints in amsozeau
different (eqdep) database. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13103 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-09Fix bug #2317: setoid_rewrite ignored binding lists. Slightlymsozeau
generalize the interface of Clenv to be able to use the existing treatment of bindings. Clenv functions did not use goals conclusions but insisted on getting goals anyway (which is even more problematic as goals appear in evar maps now). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13102 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-09Fix bug #2262: bad implicit argument number by avoiding countingmsozeau
products as implicit if they're part of a term and not a type (issue a warning instead). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13101 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-09Keep description of Automatic Introduction at only one place of CHANGES.herbelin
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13100 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-09Relaxed the freshness constraint in "intro H" (with "H" explicit):herbelin
don't forbid to overwrite a global reference of same basename. Should hopefully be more convenient. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13099 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-09Allowing to use an ordering different than Lt with measurejforest
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13098 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-09Automatic introduction of names given before ":" in Lemma's andherbelin
Definition's is not so painless. It seems to however generally provide "nicer" scripts so let us keep it and update the contribs and test-suite accordingly. Also enforced that the actual introduced names to be exactly as given in the statements. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13097 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-09Backported r13080 (support for open terms in ltac matching) from trunk to v8.3.herbelin
Also updated perf-analysis file (the part of the commit that delays typing of ltac instances seems to slightly improve a few contributions) git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13096 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-08Fixing commit r13090 (forgot to commit the file generating Nmake_gen.v).herbelin
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13095 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-08Using vernac parsing for Functionjforest
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13094 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-08Extraction with implicits: perform the occur-check after optimisationsletouzey
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13093 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-08Fixed wrong spelling in a warning.herbelin
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13092 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-08Added documentation: "Theorem id x1..xn : T" and "Set Automatic Introduction".herbelin
Updated documentation in many ways: - merged binder and binderlet as done for a while in implementation - fixed a few technical problems (bad indexation of {x:A|P x}, missing spaces in html code in many situations due to missing curly brackets around TeX macros - e.g. around \ldots -, missing dots at end of sentences, ...) - renamed "statement" commands into "assertion" commands and "declaration" commands into "assumption" commands - moved Goal and Save out of the Gallina chapter - avoid some recovering in the Gallina and proof mode chapters - slight smoothing of some hard-to-read paragraphs of the manual git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13091 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-08Made option "Automatic Introduction" active by default before too manyherbelin
people use the undocumented "Lemma foo x : t" feature in a way incompatible with this activation. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13090 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-08Typo in ExtrOcamlString: list char instead of char listletouzey
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13089 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-08Fix unfolding tactic for well-founded Programs.msozeau
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13088 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-08Tentative fix for typeclass resolution raising Evd.define exceptions.msozeau
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13087 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-08Fix treatment of {struct x} annotations in presence of generalizedmsozeau
binders. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13086 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-07Fix commentglondu
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13085 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-07Document grammar.cma->unix.cma dependency in CHANGESglondu
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13084 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-07Replace ld by gcc in ocamlopt_shared_os5fix.sh (Closes: #2325)glondu
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13083 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-07fixing error message display.vgross
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13082 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-07Backporting part of r12970 to trunk (deprecation of double induction).herbelin
Backport of the part of r12970 about Fixpoint doc will come in a further commit. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13081 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-06Added support for Ltac-matching terms with variables bound in the patternherbelin
- Instances found by matching.ml now collect the set of bound variables they possibly depend on in the pattern (see type Pattern.extended_patvar_map); the variables names are canonically ordered so that non-linear matching takes actual names into account. - Removed typing of matching constr instances in advance (in tacinterp.ml) and did it only at use time (in pretyping.ml). Drawback is that we may have to re-type several times the same term but it is necessary for considering terms with locally bound variables of which we do not keep the type (and if even we had kept the type, we would have to adjust the indices to the actual context the term occurs). - A bit of documentation of pattern.mli, matching.mli and pretyping.mli. - Incidentally add env while printing idtac messages. It seems more correct and I hope I did not break some intended existing behavior. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13080 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-06Updated performance analysis fileherbelin
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13079 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-04Extraction: attempt to provide nice extraction of chars and strings for Ocamlletouzey
When Requiring ExtrOcamlString : * ascii is mapped to Ocaml's char, the ugly translation of constructor and pattern-match should hopefully be seen very rarely (never ?). We add a hack in ocaml.ml for recognizing constant chars. * string is mapped to (list char). Extracting to Ocaml's string could be done, but would be really nasty (lots of non-trivial Extract Constant to add). For now, (list char) seems a good compromise. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13078 85f007b7-540e-0410-9357-904b9bb8a0f7
2010-06-04Ascii: simplier ascii_of_pos, conversion from/to N, proofs about ↵letouzey
nat-->ascii-->nat - ascii_of_pos isn't tail-recursive anymore, but recursivity is at most of depth 8. - (brutal) proof that N_of_ascii (ascii_of_N n) = n for all n<256, same for nat. Ideally, we could say someday that N_of_ascii (ascii_of_N n) = n mod 256 git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13077 85f007b7-540e-0410-9357-904b9bb8a0f7
generated by cgit v1.2.3 (git 2.25.1) at 2026-07-18 05:16:24 +0000