| Age | Commit message (Collapse) | Author |
|---|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@12502 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
branch
and remove equations stuff which moves to a separate plugin.
Classes:
- Ability to define classes post-hoc from constants or inductive types.
- Correctly rebuild the hint database associated to local hypotheses when
they are changed by a [Hint Extern] in typeclass resolution.
Tactics and proofs:
- Change [revert] so that it keeps let-ins (but not [generalize]).
- Various improvements to the [generalize_eqs] tactic to make it more robust
and produce the smallest proof terms possible.
Move [specialize_hypothesis] in tactics.ml as it goes hand in hand with
[generalize_eqs].
- A few new general purpose tactics in Program.Tactics like [revert_until]
- Make transitive closure well-foundedness proofs transparent.
- More uniform testing for metas/evars in pretyping/unification.ml
(might introduce a few changes in the contribs).
Program:
- Better sorting of dependencies in obligations.
- Ability to start a Program definition from just a type and no obligations,
automatically adding an obligation for this type.
- In compilation of Program's well-founded definitions, make the functional a
separate definition for easier reasoning.
- Add a hint database for every Program populated by [Hint Unfold]s for
every defined obligation constant.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@12440 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@12337 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
rewriting (thanks to Georges Gonthier for pointing it out).
We try to find a declared rewrite relation when the equation does not
look like an equality and otherwise try to reduce it to find a leibniz equality
but don't backtrack on generalized rewriting if this fails. This new
behavior make two fsets scripts fail as they are trying to use an
underlying leibniz equality for a declared rewrite relation, a [red]
fixes it.
Do some renaming from "setoid" to "rewrite".
Fix [is_applied_rewrite_relation]'s bad handling of evars and the
environment.
Fix some [dual] hints in RelationClasses.v and assert that any declared
[Equivalence] can be considered a [RewriteRelation].
Fix minor tex output problem in coqdoc.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@12310 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
more often but respects the spec better. The changes in the stdlib are
reduced to adding a few explicit [unfold]s in FMapFacts (exact was doing
conversion with delta on open terms in that case).
Also fix a minor bug in typeclasses not seeing typeclass evars when
their type was a (defined) evar itself.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@12238 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
unused methods object. Matthieu please review this change (after
monday), I might have introduced a bug in rebuild_instance.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11659 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
instances and the corresponding evar's type if it contains existentials
to avoid dangling evars. No noticeable performance impact (at least on the
stdlib). Subsumes (and fixes) the (broken) fix in unification.ml that was
previously patched by M. Puech.
Improve error messages related to existential variables and type
classes.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11543 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
It solves feature request 1852, makes me and Arnaud happy and
will permit to factor some more code in typeclasses.
- Records are introduced using the syntax "{| x := t; y := foo |}" and
"with" clauses are currently parsed but not yet supported in the
elaboration. You are invited to suggest other syntaxes :)
- Missing fields are turned into holes, extra fields cause an error
message. The current implementation finds the type of the record
at pretyping time, from the typing constraint alone (and just expects
an inductive with one constructor). It is then impossible to use
scope information to parse the bodies: that may be wrong. The other
solution I see is using the fields to detect the type earlier, before
internalisation of the bodies, but then we get in name clash hell.
- In funind/contrib/interface I mostly put [assert false] everywhere to
avoid warnings.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11496 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
cases where coercion could not occur as well.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11414 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11405 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11403 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
class constraints of the form Π x1 ... xn, Class args.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11278 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
Now there is a class "SetoidRelation" for registering relations that
should always be considered as setoids and never unfolded. Every "Add
Relation" command adds an instance and impl,iff are there by
default. Now the test is: if there is a SetoidRelation instance, use it
; otherwise, allow unfolding to find an eq or fallback on
setoid_rewrite. To avoid searching for SetoidRelation instances
repeateadly we check that it is really needed first by unfolding the
hyp. Only two scripts relied on the now-forbidden semantics of rewriting
by an @eq inside a setoid relation, in Numbers.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11269 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
- Now [ id : Class foo ] makes id an explicit argument,
and [ Class foo ] is equivalent to [ {someid} : Class foo ].
This makes declarations such as "Class Ord [ eq : Eq a ]" have
sensible implicit args.
- Better handling of {} in class and record declarations, refactorize
code for declaring structures and classes.
- Fix merging of implicit arguments information on section closing.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11204 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
- Cases on multiple objects
- Avoid dangerous coercion with evars in subtac_coercion
- Resolve typeclasses method-by-method to get better error messages.
- Correct merging of instance databases (and add debug printer)
- Fix a script in NOrder where a setoid_replace was not working before.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11198 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
Change from named_context to rel_context for class params and fields.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11163 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11160 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
components and resolving them separately, reporting more precise failures.
Improve error messages.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11105 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
typeclass code.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11047 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
files, about definitions and type of references.
- Add missing location information on fixpoints/cofixpoint in topconstr and
syntactic definitions in vernacentries for correct dumping.
- Dump definition information in vernacentries: defs, constructors,
projections etc...
- Modify coqdoc/index.mll to use this information instead of trying to
scan the file.
- Use the type information in latex output, update coqdoc.sty accordingly.
- Use the hyperref package to do crossrefs between definition and
references to coq objects in latex.
Next step is to test and debug it on bigger developments.
On the side:
- Fix Program Let which was adding a Global definition.
- Correct implicits for well-founded Program Fixpoints.
- Add new [Method] declaration kind.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11024 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
- Fix typeclass interface: instance_constructor now takes the instance
constrs as argument to build and return the corresponding term and
type.
- Better typeclass error reporting when defining fixpoints.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10975 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
- More meaningful argument name for resolve_typeclasses.
- Try to remove unneeded instance declarations in
Morphisms.
- Allow the proofs of reflexivity arising from unchanged
arguments in setoid_rewrite to be fullfiled by Morphism instances and
not necessariy because the relation is reflexive (needed by
higher-order morphisms). Use a new "MorphismProxy" class to implement
that efficiently.
- Fix a bug in abstract_generalize not delta-reducing a type where it should.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10807 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
semantics.
- Add an Equivalence instance for pointwise equality from an
Equivalence on the codomain of a function type, used by default when
comparing functions with the Setoid's ===/equiv.
- Partially fix the auto hint database "add" function where the exact
same lemma could be added twice (happens when doing load for example).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10797 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
whether or not to keep them regardless of the actual dependencies (in
order to implement the proper discharge behavior for type classes).
This means adding an argument to rebuild_function in libobject, giving
this information on variables after a section's constants have been
discharged (discharge_function is too early). Surface syntax for
Variable not added yet.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10741 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
pas correctes
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10739 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
|
|
resolution. Add [relation] and Setoid's [equiv] as such objects.
Considerably simplify resolve_all_evars for typeclass resolution, adding
a further refinement (and hack): evars get classified as non-resolvable
(using the evar_extra dynamic field) if they are turned into a
goal. This makes it possible to perform nested typeclass resolution
without looping. We
take advantage of that in Classes/Morphisms where [subrelation_tac] is
added to the [Morphism] search procedure and calls the apply tactic which
itself triggers typeclass resolution. Having [subrelation_tac] as a tactic
instead of an instance, we can actually force that it is applied only
once in each search branch and avoid looping.
We could get rid of the hack when we have real goals-as-evars
functionality (hint hint).
Also fix some test-suite scripts which were still calling [refl]
instead of [reflexivity].
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10687 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
binding, add "!" syntax for the new binders which require parameters and
not superclasses. Change backquotes for curly braces for user-given
implicit arguments, following tradition. This requires a hack a la
lpar-id-coloneq. Change ident to global for typeclass names in class
binders. Also requires a similar hack to distinguish between [ C t1 tn ]
and [ c : C t1 tn ]. Update affected theories.
While hacking the parsing of { wf }, factorized the two versions of fix
annotation parsing that were present in g_constr and g_vernac.
Add the possibility of the user optionaly giving the priority for resolve and
exact hints (used by type classes). Syntax not fixed yet: a natural
after the list of lemmas in "Hint Resolve" syntax, a natural after a "|"
after the instance constraint in Instance declarations (ex in
Morphisms.v).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10628 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
eauto instead of an arbitrary tactic. Export more from eauto to allow
easier debugging.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10534 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
Tabareau:
- first pass: generation of the Morphism constraints with metavariables
for unspecified relations by one fold over the term.
This builds a "respect" proof term for the whole term with holes.
- second pass: constraint solving of the evars, taking care of finding a
solution for all the evars at once.
- third step: normalize proof term by found evars, apply it, done!
Works with any relation, currently not as efficient as it could be due
to bad handling of evars. Also needs some fine tuning of the instances
declared in Morphisms.v that are used during proof search, e.g. using
priorities.
Reorganize Classes.* accordingly, separating the setoids in
Classes.SetoidClass from the general morphisms in Classes.Morphisms
and the generally applicable relation theory in Classes.Relations.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10515 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
hole kind info for topconstrs.
Derive eta_expansion from functional extensionality axiom.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10439 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
implementation is much less add-hoc. Opens possibility of arbitrary prefixes in Class and Instance declarations. Current implementation with eauto is a bit more dangerous... next patch will fix it.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10432 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
extensionality tactic to apply the axiom properly and fix test-suite.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10415 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
10358-10362,10365,10371-10373,10377,10383-10384,10394-10395,10398,10403-10408 via svnmerge from
svn+ssh://msozeau@scm.gforge.inria.fr/svn/coq/branches/TypeClasses
........
r10358 | msozeau | 2007-12-10 15:42:53 +0100 (Mon, 10 Dec 2007) | 1 line
Comment grammar error
........
r10359 | msozeau | 2007-12-10 16:04:09 +0100 (Mon, 10 Dec 2007) | 7 lines
The initial Type Classes patch.
This patch introduces type classes and instance definitions a la Haskell.
Technically, it uses the implicit arguments mechanism which was extended a bit.
The patch also introduces a notation for explicitely marking implicit, maximally inserted parameters.
It includes the tactic redefinition code too (Ltac tac ::= foo redefines tac).
........
r10360 | msozeau | 2007-12-10 16:14:30 +0100 (Mon, 10 Dec 2007) | 1 line
Fix interface
........
r10361 | msozeau | 2007-12-10 16:28:19 +0100 (Mon, 10 Dec 2007) | 1 line
Fix more xlate code
........
r10362 | msozeau | 2007-12-11 02:00:53 +0100 (Tue, 11 Dec 2007) | 3 lines
Update coqdoc for type classes, fix proof state not being displayed on Next Obligation.
........
r10365 | msozeau | 2007-12-11 14:22:35 +0100 (Tue, 11 Dec 2007) | 3 lines
Bug fixes in Instance decls.
........
r10371 | msozeau | 2007-12-12 21:17:30 +0100 (Wed, 12 Dec 2007) | 3 lines
Streamline typeclass context implementation, prepare for class binders in proof statements.
........
r10372 | msozeau | 2007-12-12 22:03:38 +0100 (Wed, 12 Dec 2007) | 1 line
Minor cosmetic fixes: allow sorts as typeclass param instances without parens and infer more types in class definitions
........
r10373 | msozeau | 2007-12-13 00:35:09 +0100 (Thu, 13 Dec 2007) | 2 lines
Better names in g_vernac, binders in Lemmas and Context [] to introduce a typeclass context.
........
r10377 | msozeau | 2007-12-13 18:34:33 +0100 (Thu, 13 Dec 2007) | 1 line
Stupid bug
........
r10383 | msozeau | 2007-12-16 00:04:48 +0100 (Sun, 16 Dec 2007) | 1 line
Bug fixes in name handling and implicits, new syntax for using implicit mode in typeclass constraints
........
r10384 | msozeau | 2007-12-16 15:53:24 +0100 (Sun, 16 Dec 2007) | 1 line
Streamlined implementation of instances again, the produced typeclass is a typeclass constraint. Added corresponding implicit/explicit behaviors
........
r10394 | msozeau | 2007-12-18 23:42:56 +0100 (Tue, 18 Dec 2007) | 4 lines
Various fixes for implicit arguments, new "Enriching" kw to just enrich existing sets of impl args. New syntax !a to force an argument, even if not dependent.
New tactic clrewrite using a setoid typeclass implementation to do setoid_rewrite under compatible morphisms... very experimental.
Other bugs related to naming in typeclasses fixed.
........
r10395 | msozeau | 2007-12-19 17:11:55 +0100 (Wed, 19 Dec 2007) | 3 lines
Progress on setoids using type classes, recognize setoid equalities in hyps better.
Streamline implementation to return more information when resolving setoids (return the results setoid).
........
r10398 | msozeau | 2007-12-20 10:18:19 +0100 (Thu, 20 Dec 2007) | 1 line
Syntax change, more like Coq
........
r10403 | msozeau | 2007-12-21 22:30:35 +0100 (Fri, 21 Dec 2007) | 1 line
Add right-to-left rewriting in class_setoid, fix some discharge/substitution bug, adapt test-suite to latest syntax
........
r10404 | msozeau | 2007-12-24 21:47:58 +0100 (Mon, 24 Dec 2007) | 2 lines
Work on type classes based rewrite tactic.
........
r10405 | msozeau | 2007-12-27 18:51:32 +0100 (Thu, 27 Dec 2007) | 2 lines
Better evar handling in pretyping, reorder theories/Program and add some tactics for dealing with subsets.
........
r10406 | msozeau | 2007-12-27 18:52:05 +0100 (Thu, 27 Dec 2007) | 1 line
Forgot to add a file
........
r10407 | msozeau | 2007-12-29 17:19:54 +0100 (Sat, 29 Dec 2007) | 4 lines
Generalize usage of implicit arguments in terms, up to rawconstr. Binders are decorated with binding info, either Implicit or Explicit for rawconstr. Factorizes code for typeclasses, topconstrs decorations are Default (impl|expl) or TypeClass (impl|expl) and
implicit quantification is resolve at internalization time, getting rid of the arbitrary prenex restriction on contexts.
........
r10408 | msozeau | 2007-12-31 00:58:50 +0100 (Mon, 31 Dec 2007) | 4 lines
Fix parsing of subset binders, bugs in subtac_cases and handling of mutual defs obligations.
Add useful tactics to Program.Subsets.
........
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10410 85f007b7-540e-0410-9357-904b9bb8a0f7
|
