| Age | Commit message (Collapse) | Author |
|---|
|
now dumb wrappers around Zsqrt_plain. Wanted (dead or alive): better
implemntations _and_ their proofs.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11013 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
transformed into generic functions, and aren't anymore fields of
records znz_op/znz_spec.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11012 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
- Divers code mort (evarutil.ml, Bvector.v)
- MAJ perf-analysis
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11004 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11003 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
proving addmuldiv31
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11002 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
- Suite à une modification faite maladroitement, on ne se contentait pas
de comparer le nom de la supposition quand on l'insérait dans l'ensemble
des suppositions utilisées, mais aussi son type, ce qui était
inutilement long (mais pas le facteur principal)
- L'environnement était parcouru deux fois pour chaque variable de
section. Ce n'était pas très grave vu qu'en général on a assez peu de
variables de sections sous la main. Mais ça restait inutile.
- Les noms qui ont déjà étés explorés sont maintenant memoizés, ce qui
gagne dans le cas les pires (comme les théorèmes sur les réels
typiquement) une exponentiel dans le temps de recherche (si on visualise
l'espace de recherche comme un DAG, l'ancienne procédure le parcourais
comme si il était un arbre, ce qui a une complexité exponentielle en la
taille du DAG).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11001 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10999 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10997 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10995 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10993 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
one.
Conversion to lists of digits is really the Right Way (TM). Maybe other parts
can also benefit from this idea. To be continued...
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10978 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
CyclicAxioms specs
Currently, 8 lemmas remains to tackle. One proof is done via a _very_
brute-force ugly approach. The all story for proving composition of
phi and phi_inv (and the other way around) is surprisingly long and
tricky. In both cases, comments are welcome, I may have missed an
easier road (?)
As a consequence of the above, we have a additional time-eager
file in the stdlib (about a minute to compile here).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10973 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
idea :-(
(some bad interaction with Arnaud's framework ??)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10969 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
Vouillon)
As said at the beginning of the file:
This file contains basic definitions of a 31-bit integer
arithmetic. In fact it is more general than that. The only reason
for this use of 31 is the underlying mecanism for hardware-efficient
computations by A. Spiwack. Apart from this, a switch to, say,
63-bit integers is now just a matter of replacing every occurences
of 31 by 63. This is actually made possible by the use of
dependently-typed n-ary constructions for the inductive type
[int31], its constructor [I31] and any pattern matching on it.
If you modify this file, please preserve this genericity.
From the user point-of-view, almost nothing has changed: functions
like On, In, shiftr, shiftl and a few others now have a
syntactically-different definition, but thanks to Eval compute in their
definition, this leads to the exact same coq objects as before. The only
difference is "Check I31" that shows the compact n-ary version
(nfun digits 31 int31) instead of (digits -> ... -> digits -> int31).
But even "Print int31" shows the same answer as before (the above type
of I31 is shown after expansion).
Arnaud, could you check whether all this works fine with your
retroknowledge ?
Notice the new file NaryFunction that contain generic stuff about
n-ary dependent functions. It should end some day in another place
than theories/Numbers, but I cant figure where for now. This file is
also quite skinny yet, but it's a start.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10967 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
for helping rewriting under Quantifiers and binders, but Matthieu's
setoid rewrite now has the same kind of capabilities by default.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10966 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
part).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10964 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10944 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
no longer
needed to compile theories/Numbers. QRewrite will probably be removed in a latter commit
(need to check 2-3 things first with Matthieu and Evgeny)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10943 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
This isn't useful for BigN et BigZ, but it can't hurt; and moreover
it's a simple way to understand CyclicAxioms. Next step: proving that
Int31 is also an implementation of CyclicAxiom.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10942 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
a non-dependent product under a lambda. Now qiff can be replaced
by a simple setoid_rewrite in NumPrelude.
Change configure to not do stripping if compiling with -g...
Add -g / CAMLDEBUG flags to the native compilation command too.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10941 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
(n _must_ in fact be a power of 2). Worse: Z_31Z is just plain wrong
since it is Z/(2^31)Z and not Z/31Z (my fault).
As a consequence, switch to CyclicAxioms, Cyclic31, DoubleCyclic, etc
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10940 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10939 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
In fact, for the moment, it was only containing a proof that Z/nZ
implements the NatInt NZAxiomsSig. We move it to a more meaningful
place and name.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10937 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
* View of Int31 as a Z/nZ moved to file Z31Z.v (TO FINISH: specs are still admitted!)
* Modular specification of Z/nZ moved to ZnZ and renamed CyclicType
* More isolation between Cyclic/Abstract and Cyclic/DoubleCyclic
* A few comments
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10936 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
using the ad-hoc qsetoid_rewrite. Could QRewrite.v be made completely obsolete ?
For the moment rewrite under fun and exists don't work.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10935 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10934 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
- Fix a typo in lowercase_utf8
- Fix generation of signatures in subtac_cases not working for dependent
inductive types with dependent indices.
- Fix coercion of inductive types generating ill-typed terms.
- Fix test script using new syntax for Instances.
- Move simpl_existTs to Program.Equality and use it in simpl_depind.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10932 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10923 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
change the default pretty-printing to use Π, λ instead of
forall and fun (and allow "," as well as "=>" for "fun" to be more
consistent with the standard forall and exists syntax). Parsing allows
theses new forms too, even if not in -unicode, and does not make Π or
λ keywords. As usual, criticism and suggestions are welcome :)
Not sure what to do about "->"/"→" ?
- [setoid_replace by] now uses tactic3() to get the right parsing level
for tactics.
- Type class [Instance] names are now mandatory.
- Document [rewrite at/by] and fix parsing of occs to support their
combination.
- Backtrack on [Enriching] modifier, now used exclusively in the
implementation of implicit arguments.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10921 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
binders.
- Change syntax of type class instances to better match the usual syntax of
lemmas/definitions with name first, then arguments ":" instance.
Update theories/Classes accordingly.
- Correct globalization of tactic references when doing Ltac :=/::=, update
documentation.
- Remove the not so useful "(x &)" and "{{x}}" syntaxes from
Program.Utils, and subset_scope as well.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10919 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
complexité exponentielle dans la machine lazy depuis que l'algo de
compilation du filtrage évite systématiquement d'expanser quand le
filtrage n'est pas dépendant.
- Un peu plus de colorisation dans coqide.
- Utilisation de formats pour améliorer de l'affichage des notations Utf8.
- Systématisation paire Local/Global dans g_vernac.ml4 (même si le
défaut n'est pas toujours le même)
- Bug Makefile
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10918 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
pour les listes (trop contraignant)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10913 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10908 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
the very handy Printf.ifprintf was not available on earlier ocaml.
This file now uses a very dirty compatibility hack.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10906 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
generation of NMake.v
- genN.ml becomes NMake_gen.ml
- no need to produce the corresponding binary: we use ocaml NMake_gen.ml > NMake.v
- beware! redoing a ./configure is mandatory after this commit
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10903 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
of NMake.v
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10902 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
For the moment, the Ints files are simply moved into directories in
theories/Numbers with meaningful names. No filenames changed, apart
from:
Zaux.v -> theories/Numbers/BigNumPrelude.v
MemoFn.v -> theories/Lists/StreamMemo.v
More to come...
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10899 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
name after internalisation, to get the correct behavior with typeclass
binders. This simplifies the pretty printing and translation
of the recursive argument name in various places too. Use this
opportunity to factorize the different internalization and
interpretation functions of binders as well.
This definitely fixes part 2 of bug
#1846 and makes it possible to use fixpoint definitions with typeclass arguments in
program too, with an example given in EquivDec.
At the same time, one fix and one enhancement in Program:
- fix a de Bruijn bug in subtac_cases
- introduce locations of obligations and use them in case the obligation tactic
raises a failure when tried on a particular obligation, as suggested by
Sean Wilson.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10889 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
morphismes). Les mettre au niveau 90 est peut-être plus naturel mais
entraîne trop d'incompatibilités (cf commit 10813 et 10825)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10888 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10887 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
du filtrage. Cela permet de détecter les cas impossibles et de simuler
les contraintes d'inversion exprimables sous la forme d'un assignement
des arguments du constructeurs (cf le cas de Vtail dans Bvector.v).
Si l'on filtre sur t:I u1 .. un, et que chaque ui a la forme vi(wi)
avec vi composé uniquement de constructeurs, et que le résultat final
est P(w1,...,wn) (qui est éventuellement lui-même une evar) alors on
construit le prédicat
Q:=fun x1 .. xn y =>
match x1 .. xn y with
| v1(z) .. vn(z) t => P(z)
| _ .. _ _ => ?evar-speciale-cas-impossible
end
qui vérifiera bien que Q u1 .. un = P(w1,..,wp).
En raison de limitations de l'unification (on aurait besoin d'eta
conversion pour résoudre des problèmes du genre
"terme rigide == match x with _ => ?evar end", et besoin d'instanciation par
constructeurs pour des cas comme "A(y) = match ?evar with C x => A(x) end"),
je n'ai pas réussi à traiter le cas général.
Aussi, on adopte une stratégie pragmatique consistant à tester
plusieurs prédicats possibles :
- si un type final est donné, on essaie d'abord l'algorithme de
Matthieu et sinon le nouvel algorithme (permet par exemple de traiter
certains cas d'élimination dépendante de Bvector.v),
- s'il n'y a pas de type final, on essaie d'abord le nouvel algo et
sinon, on essaie avec un prédicat sans dépendance (permet de traiter
des cas compliqués comme celui de par cas sur I' dans le fichier
Case13.v de la test-suite).
Dans la pratique, il y a beaucoup de changement dans le code de compile_case.
- Par exemple, la compilation est maintenant toujours appelé avec un
prédicat (là où l'on pouvait avoir None, on a maintenant toujours au
moins une evar).
- En revanche, le membre droit des clauses est maintenant
optionnel. Si c'est None, c'est qu'on se trouve dans le cas d'une
branche impossible au moment du calcul du prédicat de retour.
- Aussi, on renonce aux PrLetIn et PrProd dans l'expression du
predicat de retour mais il faut savoir que c'est maintenant la liste
des tomatchs qui spécifie le contexte exact dans lequel le prédicat
de retour est bien typé.
- Et d'autres...
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10883 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
syntaxe interne de ring_lookup et field_lookup qui n'était pas assez
robuste pour supporter une syntaxe [ ... ] dans constr.
Déplacement de now_show de List.v vers Tactics.v, déplacement de "[ _ ]"
au niveau 0.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10872 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
in reasonable time), add (unfinished) documentation on type classes.
Put some classes into Prop explicitely as singleton inductive types are
no longer in Prop by default even if all the arguments are (is that
really what we want? roconnor says no).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10868 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
now. Fix proof scripts that failed correspondingly. Should make many
contribs compile again...
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10863 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10858 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
fichiers en attendant une intégration à theories/Numbers
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10857 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
substitution. Makes unification succeed a bit more often, hence auto
works better in some cases.
- Backtrack the changes of auto using Hint Unfold to do more delta and add
a new tactic "new auto" which does that, for compatibility.
The first fix may have a big impact on the contribs, whereas the second
should make them compile again... we'll see.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10855 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
failures in proof search. Catch Refiner.FailError in typeclasses eauto
to indicate that an extern tactic failed.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10853 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
account.
- Add test case for bug #1844 on Combined Scheme.
- Change Reflexive_partial_app_morphism to require a Reflexive proof to
cut down search earlier, without losing anything.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10846 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
debugging and printing hint databases
- Typeclasses unfold now correctly adds _global_ unfold hints.
- New tactic autosimpl to do simplification using the declared unfold
hints in given hint databases.
- Work on auto-modulo-some-delta (the declared Unfold constants),
actually used mostly if the goal contains evars, as Hint_db.map_auto
does not work up-to any conversions (yet).
- Fix GenMul which was using the old semantics of failing early because
of variance checks, which is not possible in the new implementation.
- Restrict when reflexive_morphism may be used using an extern tactic.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10842 85f007b7-540e-0410-9357-904b9bb8a0f7
|
