aboutsummaryrefslogtreecommitdiff
path: root/tactics
AgeCommit message (Collapse)Author
2004-10-06Leibniz equality is now a quantified relation.sacerdot
This means that you can declare a morphism signature that has an argument (or its output type) that is just eq. E.g.: Add Morphism incl with signature incl --> eq ++> impl. is a correct signature for a morphism property of type forall A, list A -> list A -> Prop git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6183 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-10-05* code simplificationsacerdot
* error message improved * bug fixed: it was not checked whether the carrier of a relation of class Leibniz matched the expected type git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6180 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-10-05* [ bug fixed ]: when a subterm (c x1 ... xn) it is checked whethersacerdot
there exists an i such that (c x1 ... xi) is a morphism. Previously only the case c was tried. * [ bug fixed ]: the term that must be replaced must not occur in c (i.e. in (c x1 ... xi)) git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6179 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-10-05* Bug fix: in case of non dependent implications the second argument wassacerdot
not correctly de-lifted. * [EXPERIMENTAL]: the Add Relation and Add Morphism commands now accept also quantified relations and quantified morphisms. [ Add Setoid doesn't do that yet. ] However, in case of quantified relations the matching between an argument type and the (quantified) carrier expected for it is quite weak and is complete only not modulus conversion. * Many bugs have probably been introduced by the experimental feature. However, the bugs should manifest only in the case of quantified relations. In particular Leibniz has a strange status and its management should be revised. * Open problem: should the data type for relations and morphisms be changed to explicitly show the quantifications? git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6178 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-10-01eq_constr replaced with a conversion test where possible.sacerdot
This is particular useful in case of modules (when it often happens to have two setoids that can be composed only up to convertibility of their input/output types). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6172 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-10-011. added new parameter "as ..." to Add Setoid. Used to synthesize the namesacerdot
of the morphism. Previously the name was automatically generated, but it was impossible to generate the same name in a module type and in a module. 2. behaviour of the relation/morphism tables w.r.t. module types and functors fixed. 3. relations/setoids/morphisms can now happear in a module type. Add Morphism in a module type declares an axiom instead of starting a new proof. Add Morphism/Add Relation/Add Setoid in a module now generate objects that match those generated by the same commands in a module type. 4. error messages improved/fixed. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6171 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-10-01add_setoid has a new parameter used to synthesize the morphism name.sacerdot
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6170 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-10-01The "Add Setoid" command now has a new argument "as name" that is usedsacerdot
to generate the name of the morphism. Previously the name was automatically generated, but this behaviour was not compatible with module typing: it was not possible to generate the same identifier in the module type and in the module. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6168 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-30New tacticsacerdot
setoid_replace ... with ... in ... [using relation ...] [generate side conditions ...] git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6166 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-30New tactic [setoid_]rewrite ... in ... [generate side conditions ...].sacerdot
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6165 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-30Proof term size optimization: setoid_rewrite H where H is an application ofsacerdot
Leibniz equality and no side conditions are imposed by the user simply calls the rewrite tactic. This was already done for setoid_replace/replace. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6164 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-30cutrewrite does not work with relations that are not Coq-like equalities.sacerdot
Thus it does not work for setoid relations and it can replace setoid_replace when the user wants to specify what the relation should be. To solve the problem this commit enables the syntax setoid_replace E1 with E2 using relation R ... git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6163 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-30setoid_replace E1 with E2sacerdot
now replaces E1 with E2 either generating the goal E1 R E2 or the goal E2 R E1 (as expected). Note: the relation R is guessed when more than one is applicable. This is not very nice, since it can guess the wrong one. Since making the tactic compute R requires much code refactoring, I would suggest the user to always use cutrewrite in place of setoid_rewrite. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6162 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-29* error messages improvedsacerdot
* msg_warning no longer used (since it pretty prints the messages on the stderr and coqide does not show them). Is this a Coq bug??? git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6160 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-29Add Relation is now able to detect non well typed relation registrations.sacerdot
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6159 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-29The Add Morphism command is now able to detect in advance:sacerdot
1) if the proposed morphism named has already been used 2) if the proposed signature matches the function type git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6158 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-29New syntaxsacerdot
"setoid_rewrite dir term generate side conditions constr_list" to specify a list of side conditions that must be a subset of the generated new goals. Used to disambiguate in case of multiple set of generated side conditions. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6157 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-29The Prod special case works only when the premise is of type Prop.sacerdot
This is now checked. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6156 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-291. major code clean-up for the Prod casesacerdot
2. cleaner behaviour: occurrences of the pattern in the type of a variable bound by a dependent product are no longer git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6155 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-29impl is a reflexive relation (it used to be areflexive).sacerdot
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6154 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-291) bug fixed: new goals where compared according to the relation argumentssacerdot
only. Now the head of the relation is used to. 2) much nicer pretty-printing of the multiple set of side conditions that the tactic can now produce git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6152 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-29The list of possible solutions proposed by the tactic cannot contain anysacerdot
longer two set of goals such that the first is a subset of the second. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6151 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-29Hugly temporary notationsacerdot
Add Morphism constr @ arugments_list @ output as name replaced with the nicer (stable?) notation Add Morphism constr with signature signature as name git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6149 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-29The quoting function is now 100% complete.sacerdot
[ Note: untested, since the source of uncompleteness fixed is extremely unlikely to happen and it never happens in my test cases. ] git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6148 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-28The quoting function is now almost complete (in the sense that it can findsacerdot
multiple solutions). Since several solutions are observationally equal (i.e. they open the same new set of goals), they are identified by the tactic. In case of multiple non observationally equal solutions, the list of them is presented to the user and the first one is randomically chosen. The user should be given a possibility to choose. Still todo: making the quoting function completely complete. Note: the wf_unifty_to_subterm is now called twice. The first time is not modulus delta conversion (the reasonable choice). The second time is modulus delta conversion and the call is made iff the first call failed. This is required by Ring that exploits the delta conversion ;-( git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6145 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-27The quoting function of the reflexive tactic is now sound: all the termssacerdot
that it generates are well-typed Coq terms. A limited form of backtracking is used to avoid non well-typed choices. The function is not complete yet (since the backtracking algorithm is incomplete). This will be fixed shortly. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6143 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-27corrected bug when lhs is matched w.r.t deltabarras
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6140 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-25- Prise en compte de Fail n (n>0) dans plusieurs cas qui avaientherbelin
disparus dès la version 1.11 de proofs/tacinterp.ml - Prise en compte du contexte de filtrage sous-terme du but dans 'match goal with' quand des hypothèses aussi sont filtrées ce qui avait disparu dans la version 1.56 de proofs/tacinterp.ml - Restauration du filtrage sous-terme dans 'match c with' qui avait disparu dans la version 1.27 de tactics/tacinterp.ml - Nettoyage git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6132 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-24New: (temporary) concrete syntax to specify the morphism signature:sacerdot
"Add Morphism m @ arg1 ... argn @ out as ident" where argi = constr arrow and arrow = "-->" | "++>" | "==>" (for contravariant, covariant and bi-variant morphisms). The syntax should be improved by getting rid of the "@" and maybe choosing better symbols to represent the arrows. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6129 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-24Simplifications concommitantes à la correction du bug #855herbelin
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6126 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-24Correction bug report #855herbelin
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6125 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-17restructuration des printers: proofs passe avant parsingbarras
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6113 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-15hiding the meta_map in evar_defsbarras
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6109 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-15put empty_env in hint clause (vo were becoming huge!)barras
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6108 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-14evar tactic bugfixcorbinea
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6105 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-13* The ML tactic is now also aware of rewriting directions.sacerdot
* Code clean-up. * Preparation for the last implementation phase. The last implementation phase will consist in the algorithm that, performing backtracking, chooses the right combination of morphisms in the syntactic representation of the goal. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6103 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-13The ML part of the tactic now knows about covariant and contravariant morphismsacerdot
arguments. However, it still does not know about rewrite directions. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6102 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-12inclusion de meta_map dans evar_defsbarras
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6099 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-10simplification de clenvbarras
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6096 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-10Dead code removed.sacerdot
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6095 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-101. add_new_morphism now has a new optional argument that is the signaturesacerdot
2. partial setoids (a.k.a. areflexive relations) are now properly supported 3. the first two arguments of proj2 in the proof term have been simplified to be just the mutual coimplication of the old and the new goal. 4. code simplification git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6094 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-10add_new_morphism has now a new argument that is the signaturesacerdot
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6093 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-10Add_new_morphism has now a new optional argument that is the signature ofsacerdot
the morphism. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6092 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-08unification encore...barras
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6085 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-08* cleaning/renaming in Setoids.vsacerdot
* the data type for relations has been extended to cover all the cases git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6084 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-08The Coq part of the reflexive part can now deal with irreflexive relations too.sacerdot
In particular it can deal with partial setoids (also called typoids). The latter feature was request by Marco Maggesi. This commit is just a porting of the ML part of the tactic to the new Coq part. It does not allow to exploit the new potentialities, yet. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6080 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-07deuxieme vague de modifs: evar_defs fonctionnelbarras
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6071 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-07* The Coq part of the reflexive tactic setoid_rewrite is generalized tosacerdot
asymmetric relations thanks to the introduction of morphisms that are covariant or contravariant in their arguments. * The ML part of the tactic is updated only for backward compatibility: it is still not possible to benefit from the asymmetric relations and relative morphisms. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6070 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-03premiere reorganisation de l\'unificationbarras
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6057 85f007b7-540e-0410-9357-904b9bb8a0f7
2004-09-03New reflexive implementation of setoid_rewrite. The new implementationsacerdot
introduces a few backward incompatibilities: 1. setoid_replace used to try to close the new goal using full_trivial. The new implementation does not try to close the goal. To fix the script it is sufficient to add a "; [ idtac | trivial ]." after every application of setoid_replace that used to generate just one goal. 2. (setoid_replace c1 c2) used to replace either c1 with c2 or c2 with c1 (if c1 did not occur in the goal). The new implementation only replaces c1 with c2. To fix the script it is sufficient to replace (setoid_replace c1 with c2) with (setoid_replace c1 with c2) || (setoid_replace c2 with c1) in those cases where it is unknown what should be substituted with what. 3. the goal generated by the "Add Morphism" command has the hypothesis permutated w.r.t. the old implementation. I.e. the old implementation used to generate: forall x1 x2, forall x3 x4, x1=x2 -> x3=x4 -> ... whereas the new implementation generates forall x1 x2, x1=x2 -> forall x3 x4, x3 = x4 -> ... Fixing the script is usually trivial. 4. the "Add Morphism P" command used to generate a non-standard goal in the case of P being a predicate (i.e. a function whose codomain is Prop). In that case the goal ended with an implication (instead of ending with a coimplication). The new implementation generates a standard goal that ends with a coimplication. To fix the script you can add the following line at the very beginning: cut <<old_goal>>; intro Hold; split; apply Hold; assumption || (symmetry;assumption). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6049 85f007b7-540e-0410-9357-904b9bb8a0f7
generated by cgit v1.2.3 (git 2.25.1) at 2026-06-12 21:45:38 +0000