| Age | Commit message (Collapse) | Author |
|---|
|
-async-proofs off
the system behaves as in 8.4
-async-proofs lazy
proofs are delayed (when possible) but never processed in parallel
-async-proofs on
proofs are processed in parallel (when possible). The number of
workers is 1, can be changed with -async-proofs-j. Extra options to
the worker process can be given with -async-proofs-worker-flags.
The default for batch compilation used to be "lazy", now it is "off".
The "lazy" default was there to test the machinery, but it makes very
little sense in a batch scenario. If you process things sequentially,
you'd better do them immediately instead of accumulating everything in
memory until the end of the file and only then force all lazy computations.
The default for -ideslave was and still is "on". It becomes dynamically
"lazy" on a per task (proof) basis if the worker dies badly.
Note that by passing "-async-proofs on" to coqc one can produce a .vo
exploiting multiple workers. But this is rarely profitable given
that master-to-worker communication is inefficient (i.e. it really
depends on the size of proofs v.s. size of system state).
|
|
Experimental. Turned out to be much harder to implement than I thought. The main
issue is that the reification in the native compiler and the VM is not quite
untyped. Indeed, type annotations for lambdas have to be reconstructed. Hence,
when reifying an application u = t a1 ... an, the type of t has to be known or
reconstructed. It is always possible to do so in plain CIC, when u is in normal
form and its type is known. However, with partial terms this may no longer be
the case, as in: ?1 a1 ... an. So we also compile and evaluate the type of
evars and metas.
This still has to be tested more extensively, but the correction of the kernel
native conversion (on terms without evars or metas) should not be impacted.
Much of this could be reused for the VM.
|
|
test-suite pass.
|
|
I chose n to be 10000 iterations. It might be big, but a slave, to check
for a termination request, has to pass the ball to the thread that
sends "regularly" Ticks to the master process. Thread.yield is a
system call, so we have to do it very rarely.
|
|
But for vm, the kernel should be functional now
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16961 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
casts of ints to evars.
- 2 in Evarutil and Goal which are really needed, even though the Goal
one could (and should) be removed;
- 2 in G_xml and Detyping that are there for completeness sake, but
that might be made anomalies altogether;
- 1 in Newring which is quite dubious at best, and should be fixed.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16786 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
1. sorts.ml: A small file utility for sorts;
2. constr.ml: Really low-level terms, essentially kind_of_constr, smart
constructor and basic operators;
3. vars.ml: Everything related to term variables, that is, occurences
and substitution;
4. context.ml: Rel/Named context and all that;
5. term.ml: derived utility operations on terms; also includes constr.ml
up to some renaming, and acts as a compatibility layer, to be deprecated.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16462 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16165 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
native OCaml code.
Warning: the "retroknowledge" mechanism has not been ported to the native
compiler, because integers and persistent arrays will ultimately be defined as
primitive constructions. Until then, computation on numbers may be faster using
the VM, since it takes advantage of machine integers.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16136 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16071 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15992 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15969 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
the new Int module. Only the most obvious were removed, so there
are a lot more in the wild.
This may sound heavyweight, but it has two advantages:
1. Monomorphization is explicit, hence we do not miss particular
optimizations of equality when doing it carelessly with the generic
equality.
2. When we have removed all the generic equalities on integers, we
will be able to write something like "let (=) = ()" to retrieve all
its other uses (mostly faulty) spread throughout the code, statically.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15957 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15804 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
compiler warnings).
I was afraid that such a brutal refactoring breaks some obscure
invariant about linking order and side-effects but the standard
library still compiles.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15800 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15715 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
Same for check_leq instead of check_geq
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15081 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
Util only depends on Ocaml stdlib and Utf8 tables.
Generic pretty printing and loc functions are in Pp.
Generic errors are in Errors.
+ Training white-spaces, useless open, prlist copies random erasure.
Too many "open Errors" on the contrary.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15020 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
that the kernel conversion solves the delta/delta critical pair the
same way the tactics did. This allows to improve Qed time when slow
down is due to conversion having (arbitrarily) made the wrong choice.
Propagation is done thanks to a new kind of cast called REVERTcast.
Notes:
- Vm conversion not modified
- size of vo generally grows because of additional casts
- this remains a heuristic... for the record, when a reduction tactic
is applied on the goal t leading to new goal t', this is translated
in the kernel in a conversion t' <= t where, hence, reducing in t'
must be preferred; what the propagation of reduction cast to the
kernel does not do is whether it is preferable to first unfold c or
to first compare u' and u in "c u' = c u"; in particular,
intermediate casts are sometimes useful to solve this kind of issues
(this is the case e.g. in Nijmegen/LinAlg/subspace_dim.v where the
combination "simpl;red" needs the intermediate cast to ensure Qed
answers quickly); henceforth the merge of nested casts in mkCast is
deactivated
- for tactic "change", REVERTcast should be used when conversion is in
the hypotheses, but convert_hyp does not (yet) support this (would
require e.g. that convert_hyp overwrite some given hyp id with a
body-cleared let-binding new_id := Cast(old_id,REVERTCast,t))
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14407 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
Allows to be sure that we apply the smart constructors.
Propagate the change to Closure, Reduction, Term, Cbv and Newring
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14386 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
- Be careful with consider_remaining_unif_problems: it might instantiate an evar, including the current goal!
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13995 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
Note: I'm unsure about some subtyping error case apparently involving
aliases of inductive types (middle of Subtyping.check_inductive); I
bound it to some NotEqualInductiveAliases error, but this has to be
checked.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13885 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
No need to tell the world about the fact that constraints are
implemented via caml's Set. Other modules just need to know about
the empty and union functions (and addition functions "enforce_geq"
and "enforce_eq" that were already there).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13725 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13494 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
This commit fixes a bug that made the system inconsistent with proof
irrelevance (the main idea being that Set = Prop by reflexivity).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13450 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
governed in the latter case by a flag since (useful e.g. for setoid
rewriting which otherwise loops as it is implemented).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13443 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
terms, but an anomaly instead. It is caught in pretyping
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13353 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13323 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
betaiota-reducing it automatically (this allows for instance to
directly obtain the expected type for "match" expressions that have a
"in I x return match x with ... end" automatically inferred return
predicate feature (see e.g. Vhead and Vtail in Bvector.v).
The need for this "optimization" was not noticed in V8.2 because in
Bvector.v, betaiota was applied peremptorily at the end of sections.
The need for it has been revealed by the removal of reduction at section
closing when Arnaud introduced the new proof engine (should in
particular make CoLoR compile).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13068 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13005 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
- Many of them were broken, some of them after Pierre B's rework
of mli for ocamldoc, but not only (many bad annotation, many files
with no svn property about Id, etc)
- Useless for those of us that work with git-svn (and a fortiori
in a forthcoming git-only setting)
- Even in svn, they seem to be of little interest
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@12972 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
1- Management of the name-space in a modular development / sharing of non-logical objects.
2- Performance of atomic module operations (adding a module to the environment, subtyping ...).
1-
There are 3 module constructions which derive equalities on fields from a module to another:
Let P be a module path and foo a field of P
Module M := P.
Module M.
Include P.
...
End M.
Declare Module K : S with Module M := P.
In this 3 cases we don't want to be bothered by the duplication of names.
Of course, M.foo delta reduce to P.foo but many non-logical features of coq
do not work modulo conversion (they use eq_constr or constr_pat object).
To engender a transparent name-space (ie using P.foo or M.foo is the same thing)
we quotient the name-space by the equivalence relation on names induced by the
3 constructions above.
To implement this, the types constant and mutual_inductive are now couples of
kernel_names. The first projection correspond to the name used by the user and the second
projection to the canonical name, for example the internal name of M.foo is
(M.foo,P.foo).
So:
*************************************************************************************
* Use the eq_(con,mind,constructor,gr,egr...) function and not = on names values *
*************************************************************************************
Map and Set indexed on names are ordered on user name for the kernel side
and on canonical name outside. Thus we have sharing of notation, hints... for free
(also for a posteriori declaration of them, ex: a notation on M.foo will be
avaible on P.foo). If you want to use this, use the appropriate compare function
defined in name.ml or libnames.ml.
2-
No more time explosion (i hoppe) when using modules i have re-implemented atomic
module operations so that they are all linear in the size of the module. We also
have no more unique identifier (internal module names) for modules, it is now based
on a section_path like mechanism => we have less substitutions to perform at require,
module closing and subtyping but we pre-compute more information hence if we instanciate
several functors then we have bigger vo.
Last thing, the checker will not work well on vo(s) that contains one of the 3 constructions
above, i will work on it soon...
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@12406 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@12337 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
very useful in practice but as soon as let-in's were not forbidden in
the internal data structure, better to do it. Moreover, this gets
closer to the view were inductive definitions are uniformly built from
"contexts". (checker not changed!)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@12273 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11889 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
splay_prod_n, lam_it -> it_mkLambda, splay_lambda -> splay_lam). Added
shortcuts for "fst (decompose_prod t)" and co.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11727 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10930 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10926 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
univers, suite à discussion avec Bruno : on franchit le cap et on
ajoute le sous-typage Prop <= Set. On n'a donc plus besoin d'utiliser
l'image de Prop dans la hiérarchie en dehors de la zone de calcul de
la sorte la plus basse d'un inductif polymorphe (au passage, nous
avons décidé de renommer Type -1 en Type 0-, pour bien indiquer qu'il
se trouve au même niveau que Type 0).
Coq se retrouve donc avec la hiérarchie Prop <= Set <= Type i et avec
une copie de Prop (Type 0-) et une copie de Set (Type 0) dans la
hiérarchie Type. En théorie, on pourrait donc supprimer "Prop Null" et
"Prop Pos" de l'implémentation et ne travailler qu'avec "Type".
L'ajout de Prop <= Set vaut à la fois dans le cas Set prédicatif et
dans le cas Set imprédicatif (Prop et Set étant en bas de la
hiérarchie, il n'y a pas d'incohérence connue). Dans le modéle
ensembliste, Prop et Type 0- sont interprétés par exemple comme
{{},{o}}, où "o" est un objet particulier interprétant les preuves, et
il n'y a pas de Set imprédicatif. Dans un modèle de réalisabilité,
Set imprédicatif est interprétable et Prop peut au choix s'interpréter
comme Set ou comme booléen (cf la thèse de Miquel). Le sous-typage du
côté ensembliste s'obtient en mettant au moins l'ensemble {{},{o}}
dans l'interprétation de Set (ce qu'on fait de la même manière que
Prop <= Type 1, avec conversion typée), et du côté réalisabilité en
mettant l'ensemble {Typ(vide),Typ(unit)} dans l'interprétation de Set
("Typ" étant la coercion faisant d'un ensemble un terme), ce qui est
fait dans la section 6.2.4 de la thèse d'Alexandre Miquel (modèle du
CC implicite sans types inductifs).
Il reste un problème pratique. Lorsqu'on donne
Inductive unit:Type := tt:unit.
Coq dit que unit est dans Prop. C'est correct parce qu'il n'y a pas de
contraintes d'univers mais un peu déroutant même si la coercion
"unit : Set" reste valide. Une suggestion est de ne rendre polymorphe
que les inductifs dont on ne donne pas la sorte explicitement, comme dans
Inductive unit := tt:unit.
mais alors, comment indiquer l'absence de sorte explicite si le type a
des paramètres réels (comme "vect") ??
PS: modification de sort_cmp dans checker/inductive.ml faite.
--This line, and those below, will be ignored--
M kernel/univ.ml
M kernel/univ.mli
M kernel/inductive.ml
M kernel/reduction.ml
M kernel/indtypes.ml
M checker/inductive.ml
M checker/reduction.ml
M pretyping/reductionops.ml
M pretyping/termops.ml
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10920 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
(i.e. "Inductive unit := tt." conduisait à "t:Prop" alors que le
principe de la hiérarchie d'univers est d'être cumulative -- et que
Set en soit le niveau 0).
Une solution aurait été de poser Prop <= Set mais on adopte une autre
solution. Pour éviter le côté contre-intuitif d'avoir unit dans Type
et Prop <= Set, on garde la représentation de Prop au sein de la
hiérarchie prédicative sous la forme "Type (max ([],[])" (le niveau
sans aucune contrainte inférieure, appelons Type -1) et on adapte les
fonctions de sous-typage et de typage pour qu'elle prenne en compte la
règle Type -1 <= Prop (cf reduction.ml, reductionops.ml, et effets
incidents dans Termops.refresh_universes et Univ.super).
Petite uniformisation des noms d'univers et de sortes au passage
(univ.ml, univ.mli, term.ml, term.mli et les autres fichiers).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10859 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
de l'argument donné contient des métavariables (souhait
#1408). Beaucoup d'infrastructure autour des constantes pour cela mais
qu'on devrait pouvoir récupérer pour analyser plus finement le
comportement des constantes en général :
1- Pour insérer les coercions, on utilise une transformation
(expérimentale) de Metas vers Evars le temps d'appeler coercion.ml.
2- Pour la compatibilité, on s'interdit d'insérer une coercion entre
classes flexibles parce que sinon l'insertion de coercion peut prendre
précédence sur la résolution des evars ce qui peut changer les
comportements (comme dans la preuve de fmg_cs_inv dans CFields de CoRN).
3- Pour se souvenir rapidement de la nature flexible ou rigide du
symbole de tête d'une constante vis à vis de l'évaluation, on met en
place une table associant à chaque constante sa constante de tête (heads.ml)
4- Comme la table des constantes de tête a besoin de connaître
l'opacité des variables de section, la partie tables de declare.ml va
dans un nouveau decls.ml.
Au passage, simplification de coercion.ml, correction de petits bugs
(l'interface de Gset.fold n'était pas assez générale; specialize
cherchait à typer un terme dans un mauvais contexte d'evars [tactics.ml];
whd_betaiotazeta avait un argument env inutile [reduction.ml, inductive.ml])
et nettoyage (declare.ml, decl_kinds.ml, avec incidence sur class.ml,
classops.ml et autres ...; uniformisation noms tables dans autorewrite.ml).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10840 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
parameterize what should be unfolded or not, by default unfolding
everything.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@10819 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
vis à vis de l'équivalence engendrées par les modules non génératifs
(cf bug #1242)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@9215 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
- prise en compte du niveau à la déclaration du type comme une fonction
des sortes des conclusions des paramètres uniformes
- suppression du retypage de chaque instance de type inductif (trop coûteux)
et donc abandon de l'idée de calculer une sorte minimale même dans
des cas comme Inductive t (b:bool) := c : (if b then Prop else Type) -> t.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@8845 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@8793 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@8782 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
Ajout de cast indiquant au kernel la strategie a suivre
Resolution du bug sur les coinductifs
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@7639 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6338 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@6245 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@5920 85f007b7-540e-0410-9357-904b9bb8a0f7
|
