| Age | Commit message (Collapse) | Author |
|---|
|
This is a fairly large commit (around 140 files and 7000 lines of code
impacted), it will cause some troubles for sure (I've listed the know
regressions below, there is bound to be more).
At this state of developpement it brings few features to the user, as
the old tactics were
ported with no change. Changes are on the side of the developer mostly.
Here comes a list of the major changes. I will stay brief, but the code
is hopefully well documented so that it is reasonably easy to infer the
details from it.
Feature developer-side:
* Primitives for a "real" refine tactic (generating a goal for each
evar).
* Abstract type of tactics, goals and proofs
* Tactics can act on several goals (formally all the focused goals). An
interesting consequence of this is that the tactical (. ; [ . | ... ])
can be separated in two
tacticals (. ; .) and ( [ . | ... ] ) (although there is a conflict for
this particular syntax). We can also imagine a tactic to reorder the
goals.
* Possibility for a tactic to pass a value to following tactics (a
typical example is
an intro function which tells the following tactics which name it
introduced).
* backtracking primitives for tactics (it is now possible to implement a
tactical '+'
with (a+b);c equivalent to (a;c+b;c) (itself equivalent to
(a;c||b;c)). This is a valuable
tool to implement tactics like "auto" without nowing of the
implementation of tactics.
* A notion of proof modes, which allows to dynamically change the parser
for tactics. It is controlled at user level with the keywords Set
Default Proof Mode (this is the proof mode which is loaded at the start
of each proof) and Proof Mode (switches the proof mode of the current
proof) to control them.
* A new primitive Evd.fold_undefined which operates like an Evd.fold,
except it only goes through the evars whose body is Evar_empty. This is
a common operation throughout the code,
some of the fold-and-test-if-empty occurences have been replaced by
fold_undefined. For now,
it is only implemented as a fold-and-test, but we expect to have some
optimisations coming some day, as there can be a lot of evars in an
evar_map with this new implementation (I've observed a couple of
thousands), whereas there are rarely more than a dozen undefined ones.
Folding being a linear operation, this might result in a significant
speed-up.
* The declarative mode has been moved into the plugins. This is made
possible by the proof mode feature. I tried to document it so that it
can serve as a tutorial for a tactic mode plugin.
Features user-side:
* Unfocus does not go back to the root of the proof if several Focus-s
have been performed.
It only goes back to the point where it was last focused.
* experimental (non-documented) support of keywords
BeginSubproof/EndSubproof:
BeginSubproof focuses on first goal, one can unfocus only with
EndSubproof, and only
if the proof is completed for that goal.
* experimental (non-documented) support for bullets ('+', '-' and '*')
they act as hierarchical BeginSubproof/EndSubproof:
First time one uses '+' (for instance) it focuses on first goal, when
the subproof is
completed, one can use '+' again which unfocuses and focuses on next
first goal.
Meanwhile, one cas use '*' (for instance) to focus more deeply.
Known regressions:
* The xml plugin had some functions related to proof trees. As the
structure of proof changed significantly, they do not work anymore.
* I do not know how to implement info or show script in this new engine.
Actually I don't even know what they were suppose to actually mean in
earlier versions either. I wager they would require some calm thinking
before going back to work.
* Declarative mode not entirely working (in particular proofs by
induction need to be restored).
* A bug in the inversion tactic (observed in some contributions)
* A bug in Program (observed in some contributions)
* Minor change in the 'old' type of tactics causing some contributions
to fail.
* Compilation time takes about 10-15% longer for unknown reasons (I
suspect it might be linked to the fact that I don't perform any
reduction at QED-s, and also to some linear operations on evar_map-s
(see Evd.fold_undefined above)).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@12961 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
We renounced to distribute evars to constr and bindings and to let
tactics do the merge. There are now two disciplines:
- the general case is that the holes in tactic arguments are pushed to
the general sigma of the goal so that tactics have no such low-level
tclEVARS, Evd.merge, or check_evars to do:
- what takes tclEVARS and check_evars in charge is now a new
tactical of name tclWITHHOLES (this tactical has a flag to support
tactics in either the "e"- mode and the non "e"- mode);
- the merge of goal evars and holes is now done generically at
interpretation time (in tacinterp) and as a side-effect it also
anticipates the possibility to refer to evars of the goal in the
arguments;
- with this approach, we don't need such constr/open_constr or
bindings/ebindings variants and we can get rid of all ugly
inj_open-style coercions;
- some tactics however needs to have the exact subset of holes known;
this is the case e.g. of "rewrite !c" which morally reevaluates c at
each new rewriting step; this kind of tactics still receive a
specific sigma around their arguments and they have to merge evars
and call tclWITHHOLES by themselves.
Changes so that each specific tactics can take benefit of this generic
support remain to be done.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@12603 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@12502 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@12337 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
s'est avéré ralentir la compilation des user-contribs au final, sans
compter aussi le bug 1980 apparemment introduit par ce commit).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11505 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
abusivement sur les clauses.
Nettoyage au passage de metamap qui était utilisé à la fois pour les
substitutions de meta et pour les contextes de typage de meta.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11467 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
des types des with-bindings dans la 8.1 [ceux-ci étaient déjà
utilisés et ce qui est vraiment nouveau est que l'unification est
maintenant celle de evarconv alors que c'était avant un mélange
d'unify_0 (sans delta) et de coercion sur les termes sans evars]. Je
renonce à maintenir la compatibilité (on se retrouve donc avec un
exemple qui fonctionne différement dans TermsConv.v de CoLoR).
- Robustesse accrue pour la nouvelle facilité de syntaxe de binding
avec paramètre pour pose/set.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10856 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
"pose as" de Program.
- Report des modifs de coercion.ml (révision 10840) dans subtac_coercion.ml.
- Comportement de "simple apply" rendu plus proche de celui du apply 8.1
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10854 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
- vérification de la cohérence des ident pour éviter une option -R
avec des noms non parsables (la vérification est faite dans
id_of_string ce qui est très exigeant; faudrait-il une solution plus
souple ?)
- correction message d'erreur inapproprié dans le apply qui descend dans les
conjonctions
- nettoyage autour de l'échec en présence de métas dans le prim_refiner
- nouveau message d'erreur quand des variables ne peuvent être instanciées
- quelques simplifications et davantage de robustesse dans inversion
- factorisation du code de constructor and co avec celui de econstructor and co
Documentation des tactiques
- edestruct/einduction/ecase/eelim et nouveautés apply
- nouvelle sémantique des intropatterns disjonctifs et documentation des
pattern -> et <-
- relecture de certaines parties du chapitre tactique
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10785 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
n'était pas encore fait
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10750 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
names, gives the ability to specify qualified classes in instance
declarations. Use that in the class_tactics code.
Refine the implementation of classes. For singleton classes the
implementation of the class becomes a regular definition (into Type or
Prop). The single method becomes a 'trivial' projection that allows to
launch typeclass resolution. Each instance is just a definition as
usual. Examples in theories/Classes/RelationClasses. This permits to
define [Class reflexive A (R : relation A) := refl : forall x, R x
x.]. The definition of [reflexive] that is generated is the same as the
original one. We just need a way to declare arbitrary lemmas as
instances of a particular class to retrofit existing reflexivity lemmas
as typeclass instances of the [reflexive] class.
Also debug rewriting under binders in setoid_rewrite to allow rewriting
with lemmas which capture the bound variables when applied (works only
with setoid_rewrite, as rewrite first matches the lemma with the entire,
closed term). One can rewrite with [H : forall x, R (f x) (g x)] in the goal
[exists x, P (f x)].
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10697 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
tombé au cours du temps
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10544 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
apply : si on a trouvé une méta, alors, on l'utilise pour instancier
les trous lors de la tentative de conversion modulo delta. Cela permet
ainsi de résoudre de petits cas d'unification, tel que celui annoncé
échouant dans le "beginner question" du 6 fevrier 2008 de coq-club.
Solde au passage de modifs cosmétiques de setoid_replace.ml avant
abandon probable du code.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10523 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
filtrage d'ordre 2 échoue à trouver un prédicat de réécriture qui
n'est pas une K-abstraction, les deux rewrite essaie alors le filtrage
d'ordre 1. Ce n'est pas le plus élégant mais c'est la solution
uniforme permettant d'être conservatif.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10221 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
Ajout de l'option with à (e)destruct et (e)induction.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10169 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
apply afin de reculer au plus tard les décisions irréversibles et en
particulier de pouvoir typer les with-bindings modulo coercions :
- l'unification des types des métas données en with-bindings est
retardé à après l'unification (unify_0) de telle sorte que les
instances trouvées par unify_0 soient prioritaires et que la
décision d'insérer éventuellement des coercions autour des valeurs
données en with-bindings se fasse au dernier moment;
- toujours pour permettre d'insérer ultimement des coercions,
l'instantiation des with-bindings ne se fait plus
l'appel unify_0 (cf clenv_unique_resolver);
- pour permettre ce retardement sans limiter le test de conversion
que unify_0 fait sur les termes clos, on transmet à unify_0 les
métas données en with-bindings (ainsi l'instantiation de ces métas
peut être faite dynamiquement au moment du test de clôture);
- parce que les métas données en with-bindings qui sont en position
de rédex (cas d'un "apply f_equal with (f:=fun ...)" peuvent
simplifier le problème d'unification (et elles ne sont pas de
toutes façons pas réinférables au premier ordre), on continue à
les substituer avant l'appel à unify_0 (cf meta_reducible_instance);
- pour l'unification du second-ordre, on continue d'instancier les
with-bindings et d'unifier les types des with-bindings avant
unification;
- reste à régler un problème de compatibilité lorsque le résultat de
l'unification des types des with-bindings est utilisé pour rendre
un terme clos et pour permettre à unify_0 d'utiliser la conversion.
+ meilleure compatibilité de apply, split, left, right pour le code
qui l'utilise avec des bindings clos
+ nettoyage et uniformisation des clenv_match_args, clenv_missing, et assimilés
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@9850 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
de résoudre des buts comme celui-ci :
Record nat_retract : Type :=
{f1 : nat -> nat; f2 : nat -> nat; f1_o_f2 : forall x, f1 (f2 x) = x}.
Goal nat_retract.
exists (fun x => x) (fun x => x).
- Nouvelle tentative d'utilisation des types des metas/evars pour
inférer de nouvelles instances de metas/evars; permet par exemple
d'utiliser f_equal sans option with, mais aussi, avec la modif
précédente, de résoudre des buts tels que
Goal exists f:bool->Prop, f true = True.
exists (fun x => True).
[Les expériences passées avaient montré qu'en prenant en compte les
types dans l'unification, on peut unifier trop tôt une evars à une
mauvaise sorte; à défaut de mécanisme de prise en compte des problème
d'unification avec sous-typage, on s'est interdit ici d'unifier des
types qui sont des arités.]
- Tout les constr de tactic_expr deviennent des open_constr (même si seul
with_bindings les accepte au final... c'est pas l'idéal).
- Renommage env -> evd et templenv -> env dans clausenv.
- Renommage closed_generic_argument -> typed_generic_argument.
- Renommage closed_abstract_argument_type -> typed_abstract_argument_type.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@9842 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
Fusion des syntaxes de "apply" et "eapply". Ajout de "eapply in",
"erewrite" et "erewrite in". Correction au passage des bugs #1461 et
#1522).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@9802 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
"R ?1 ... ?n <= T". Utilisation de cette fonction dans Setoid_replace au
au lieu de w_unify (suggestion de GG).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@9673 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
du rôle de l'argument (-1) de make_clenv_binding_apply; nouvelle
correction qui évite ce codage.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@9277 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
Au passage, déplacement des tactiques cut and co plus en amont + commentaires.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@9266 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@7659 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
generating evars in place of metas. Notice that thanks to this changement
unification can be more effective (expecially after reduction) since evars
have non-empty contexts whereas metas have not.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@7069 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
convertibility is now tried in setoid_rewrite. As a consequence it is now
possible to declare relations over the function space (fun A B: Type => A -> B).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@7039 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6616 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
the new module kernel/mod_subst.ml.
MOTIVATION: mod_subst is compiled after kernel/term.ml; thus it is now
possible to define substitutions that also delta-expand constants
(by associating the delta-expanded form to the constant name).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6304 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6113 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6109 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6099 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6096 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6085 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6071 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6058 85f007b7-540e-0410-9357-904b9bb8a0f7
|
