]>
matita.cs.unibo.it Git - helm.git/log
Enrico Tassi [Wed, 9 Apr 2008 08:47:07 +0000 (08:47 +0000)]
...
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 22:00:51 +0000 (22:00 +0000)]
Variables are no longer experted (cooking is now implemented).
Better reporting with indentation.
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 21:49:50 +0000 (21:49 +0000)]
Swapped arguments in error message.
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 21:31:13 +0000 (21:31 +0000)]
Variables having a body can occur in cooked terms and must be delta-expanded
during translation to the new CIC format.
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 21:18:03 +0000 (21:18 +0000)]
Fix over the previous one: Var-LetIn abstractions should be ignored and not
transformed into applications at all.
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 21:13:18 +0000 (21:13 +0000)]
Error message improved.
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 20:57:30 +0000 (20:57 +0000)]
Cooking w.r.t. variables with bodies is now implemented as delta-expansion.
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 17:51:10 +0000 (17:51 +0000)]
Fix name capture in cofix.
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 17:50:32 +0000 (17:50 +0000)]
To test the translation.
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 17:27:41 +0000 (17:27 +0000)]
Name capture fixed.
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 17:22:19 +0000 (17:22 +0000)]
Improved error messages in place of "sort elimination not allowed".
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 17:03:35 +0000 (17:03 +0000)]
Seed reset before each convert_obj.
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 17:02:04 +0000 (17:02 +0000)]
Use seed to avoid further name clashes.
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 16:59:26 +0000 (16:59 +0000)]
...
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 16:45:16 +0000 (16:45 +0000)]
Incredible, but true! Our name mangling clashed with one identifier in the
library of Coq, generating two functions that mutually called each other,
making the type-checker diverge!!!!
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 16:37:16 +0000 (16:37 +0000)]
Cooked objects are no longer well typed in the uncooked old environment.
Thus we need to fix_outtype before cooking.
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 16:10:40 +0000 (16:10 +0000)]
Improved eat_prods error message (at the cost of passing the 6th argument!)
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 15:56:31 +0000 (15:56 +0000)]
The old kernel does not accept ens whose order is different from the one
declared for the object whose reference it is applied to. Fixed.
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 15:25:46 +0000 (15:25 +0000)]
Bad alpha-conversion by Enrico fixed.
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 15:18:33 +0000 (15:18 +0000)]
Bug fixed: the ens need not be ordered w.r.t. the declared variables.
The latter order is the right one for abstraction.
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 14:54:14 +0000 (14:54 +0000)]
j off by one in error message
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 14:47:56 +0000 (14:47 +0000)]
fix_outty partially fixed: missing lifts in args and ens
Claudio Sacerdoti Coen [Tue, 8 Apr 2008 14:38:48 +0000 (14:38 +0000)]
fix_outty fixed in order to perform eta-expansion when the term is not
a sequence of lambdas.
Enrico Tassi [Tue, 8 Apr 2008 09:11:33 +0000 (09:11 +0000)]
added simplify to avoid ugly proofterm
Claudio Sacerdoti Coen [Mon, 7 Apr 2008 23:16:36 +0000 (23:16 +0000)]
Tentative (and bad!) fix of outtypes in non-dependent eliminations.
The fix does not work since there is no guarantee in Coq that the outtype
is made of a sequence of Lambdas (and indeed it is not for automatically
generated elimiantion principles). Thus I need to perform recursion on the
type of the outtype and I also need to perform some eta-expansion to insert
the additional lambda in the right position :-(
Claudio Sacerdoti Coen [Mon, 7 Apr 2008 22:09:02 +0000 (22:09 +0000)]
Cooking discriminated into cooking in types (using Prods) and cooking in bodies
(using Lambdas).
Claudio Sacerdoti Coen [Mon, 7 Apr 2008 22:04:17 +0000 (22:04 +0000)]
Pretty-printing of definitions fixed.
Claudio Sacerdoti Coen [Mon, 7 Apr 2008 22:00:51 +0000 (22:00 +0000)]
Cooking implemented (not tested yet).
Claudio Sacerdoti Coen [Mon, 7 Apr 2008 21:27:48 +0000 (21:27 +0000)]
Reports improved.
Claudio Sacerdoti Coen [Mon, 7 Apr 2008 21:05:13 +0000 (21:05 +0000)]
Debugging code fixed.
Enrico Tassi [Mon, 7 Apr 2008 16:30:11 +0000 (16:30 +0000)]
added invalidate to clear the cache and ease the comparison with the new kernel
Enrico Tassi [Mon, 7 Apr 2008 16:19:11 +0000 (16:19 +0000)]
commented out uris of objs that do not compile properly
Enrico Tassi [Mon, 7 Apr 2008 16:18:52 +0000 (16:18 +0000)]
added comparison with old kernel
Enrico Tassi [Mon, 7 Apr 2008 16:18:22 +0000 (16:18 +0000)]
fixed error message for eat prods
Enrico Tassi [Mon, 7 Apr 2008 16:17:57 +0000 (16:17 +0000)]
invalidate
Enrico Tassi [Mon, 7 Apr 2008 16:17:16 +0000 (16:17 +0000)]
uris relative to axioms are translated in the right way
Enrico Tassi [Mon, 7 Apr 2008 16:16:19 +0000 (16:16 +0000)]
added invalidate
Enrico Tassi [Mon, 7 Apr 2008 15:01:32 +0000 (15:01 +0000)]
added preliminary checks for indtys
typecheck the explicit type of letins
Enrico Tassi [Mon, 7 Apr 2008 15:00:23 +0000 (15:00 +0000)]
call get_obj instead of get_cooked_obj ~trust:false
Enrico Tassi [Mon, 7 Apr 2008 14:59:43 +0000 (14:59 +0000)]
added case of const and axiom
Enrico Tassi [Mon, 7 Apr 2008 14:59:04 +0000 (14:59 +0000)]
added # to comment
Enrico Tassi [Mon, 7 Apr 2008 14:58:39 +0000 (14:58 +0000)]
better printings
Enrico Tassi [Mon, 7 Apr 2008 14:31:03 +0000 (14:31 +0000)]
the explicit type in a LetIn must be typecheckd before convertibility is checked
Enrico Tassi [Mon, 7 Apr 2008 12:16:04 +0000 (12:16 +0000)]
opt compilation enabled, removed some pps, check has -alluris to regenerate
the uri list, alluri sorted such that the last fpart of the file fails (but not the former).
all nat/ compiles!
Enrico Tassi [Mon, 7 Apr 2008 11:57:33 +0000 (11:57 +0000)]
lefts_ad_tys properly sorted, added some metasenv here and there,
recursive args defined in case of appl
Enrico Tassi [Mon, 7 Apr 2008 11:55:43 +0000 (11:55 +0000)]
lefts_and_tys was tys@lefts, CSC claims it was working since that context is only used
by whd and it did not contain any definitions by contructions, thus the length was the only important bit
Enrico Tassi [Mon, 7 Apr 2008 11:26:54 +0000 (11:26 +0000)]
added a list of uris to ease debugging
Enrico Tassi [Mon, 7 Apr 2008 11:07:11 +0000 (11:07 +0000)]
print the excpetion and raise it again, seems to produce a reasonable backtrace
Enrico Tassi [Mon, 7 Apr 2008 11:04:14 +0000 (11:04 +0000)]
new constants have depth = max_int insted of 0 so that reducing to 0 expands them
Enrico Tassi [Mon, 7 Apr 2008 10:54:41 +0000 (10:54 +0000)]
off by one in calling count_from
Enrico Tassi [Mon, 7 Apr 2008 10:54:11 +0000 (10:54 +0000)]
fixed bug in translating Fix, recno was not properly computed
Enrico Tassi [Mon, 7 Apr 2008 10:53:05 +0000 (10:53 +0000)]
added a pretty printing of the new object and use argv[1] as uri
Enrico Tassi [Mon, 7 Apr 2008 10:22:52 +0000 (10:22 +0000)]
context, metasenv and subst made mandatory in CicPp
Enrico Tassi [Mon, 7 Apr 2008 10:06:33 +0000 (10:06 +0000)]
guarded by has a nice error message
Enrico Tassi [Mon, 7 Apr 2008 09:57:46 +0000 (09:57 +0000)]
reference type made private, added mk_constructor to be used
to create the i-th constructor of an inductive type (used in match typechecking).
fixed name capture in oCic2nCic when typing fixpoint types
Enrico Tassi [Fri, 4 Apr 2008 17:43:20 +0000 (17:43 +0000)]
some debug printings
Enrico Tassi [Fri, 4 Apr 2008 17:43:00 +0000 (17:43 +0000)]
added to the cache a boolean to state if the object is
already checked or not (a frozen list is really needed)
Enrico Tassi [Fri, 4 Apr 2008 17:41:58 +0000 (17:41 +0000)]
fix
Enrico Tassi [Fri, 4 Apr 2008 17:41:33 +0000 (17:41 +0000)]
today it seems that the substituted should be lifted by k-1+lift_args and not k+lift_args since k starts from 1
Enrico Tassi [Fri, 4 Apr 2008 16:40:05 +0000 (16:40 +0000)]
added ppobj
Enrico Tassi [Fri, 4 Apr 2008 13:10:22 +0000 (13:10 +0000)]
type_of_constant was retunrning the type of the wrong constructor (-1 missing)
fixpoints were not properly debruijnated before being checked
better error message for match branches
Enrico Tassi [Fri, 4 Apr 2008 13:08:32 +0000 (13:08 +0000)]
fixed list.nth and added some paretheses
Enrico Tassi [Fri, 4 Apr 2008 13:07:36 +0000 (13:07 +0000)]
added some printings and catched more exceptions
Enrico Tassi [Fri, 4 Apr 2008 13:06:46 +0000 (13:06 +0000)]
removed useless printing
Enrico Tassi [Fri, 4 Apr 2008 13:06:21 +0000 (13:06 +0000)]
iterator map was mapping Lambdas to Prods!!!
Enrico Tassi [Fri, 4 Apr 2008 11:46:57 +0000 (11:46 +0000)]
debugging started
Enrico Tassi [Fri, 4 Apr 2008 10:42:57 +0000 (10:42 +0000)]
type of constant ported
Enrico Tassi [Fri, 4 Apr 2008 10:26:50 +0000 (10:26 +0000)]
added add_obj to store objects in the environment, implementation still
shaky since no frozen list is there
Enrico Tassi [Fri, 4 Apr 2008 10:20:23 +0000 (10:20 +0000)]
logger added
Enrico Tassi [Fri, 4 Apr 2008 10:12:25 +0000 (10:12 +0000)]
added get_obj in nCicEnvironment that returns an object and a boolean stating
if the object is checked or not (to be used by the typechecker, function
type_of_constant
Enrico Tassi [Fri, 4 Apr 2008 10:07:22 +0000 (10:07 +0000)]
indentation fixed
Enrico Tassi [Fri, 4 Apr 2008 10:04:06 +0000 (10:04 +0000)]
returns_a_counductive implemented
Enrico Tassi [Fri, 4 Apr 2008 09:57:58 +0000 (09:57 +0000)]
type_of_branch ported and optimized to not lift the outty every Prod, keep the
number of passes binders and lift it just once
Enrico Tassi [Fri, 4 Apr 2008 09:40:10 +0000 (09:40 +0000)]
is_really_smaller ported, still to understand the case in which
whe find a Fix or and applied Fix (still there after a whd)
Enrico Tassi [Thu, 3 Apr 2008 16:55:10 +0000 (16:55 +0000)]
debujin implemented with the map recursor
Enrico Tassi [Thu, 3 Apr 2008 16:26:19 +0000 (16:26 +0000)]
bug found rewriting the kernel backported: n instead of m when pulling a def from the context, previously a meaningless index was used
Enrico Tassi [Thu, 3 Apr 2008 16:23:04 +0000 (16:23 +0000)]
non debruijned constructor may be useless elswere, for the moment we do not bind it
Enrico Tassi [Thu, 3 Apr 2008 16:16:59 +0000 (16:16 +0000)]
guarded_by_destructor ported, many auxiliary functions still to be done
Enrico Tassi [Thu, 3 Apr 2008 16:15:51 +0000 (16:15 +0000)]
removed FSF banner
Enrico Tassi [Thu, 3 Apr 2008 15:42:10 +0000 (15:42 +0000)]
added ugly test showing many many bugs in the current kernel
Enrico Tassi [Wed, 2 Apr 2008 13:15:27 +0000 (13:15 +0000)]
added slim version of does_not_occur
Enrico Tassi [Wed, 2 Apr 2008 13:14:45 +0000 (13:14 +0000)]
added translation of Set to Type0 (avoid warning)
Enrico Tassi [Wed, 2 Apr 2008 12:56:27 +0000 (12:56 +0000)]
fixes backported from the new kernel
Enrico Tassi [Wed, 2 Apr 2008 09:39:23 +0000 (09:39 +0000)]
removed dummy rewrite
Enrico Tassi [Wed, 2 Apr 2008 09:30:32 +0000 (09:30 +0000)]
removed dummy rewrite
Enrico Tassi [Wed, 2 Apr 2008 09:23:30 +0000 (09:23 +0000)]
removed dummy rewrites
Enrico Tassi [Wed, 2 Apr 2008 09:22:14 +0000 (09:22 +0000)]
fixed according to the new rewrite semantics (fails if no progress)
Enrico Tassi [Wed, 2 Apr 2008 09:18:08 +0000 (09:18 +0000)]
fixed depends after the removal of Fsub
Enrico Tassi [Tue, 1 Apr 2008 19:07:11 +0000 (19:07 +0000)]
better check for progress
Claudio Sacerdoti Coen [Tue, 1 Apr 2008 17:51:14 +0000 (17:51 +0000)]
typeof_obj0 implemented
Claudio Sacerdoti Coen [Tue, 1 Apr 2008 17:01:45 +0000 (17:01 +0000)]
1) added get_checked_indtys that returns the whole block of inductive types
2) type checking of Match almost finished
Enrico Tassi [Tue, 1 Apr 2008 14:08:11 +0000 (14:08 +0000)]
progress made better, still not perfect
Enrico Tassi [Tue, 1 Apr 2008 12:40:50 +0000 (12:40 +0000)]
added to rewrite a check to effectively do something, and fail if nothing has been rewritten.
Enrico Tassi [Tue, 1 Apr 2008 08:59:02 +0000 (08:59 +0000)]
added some comments, but the samentics of many functions is still unclear to me
Enrico Tassi [Tue, 1 Apr 2008 08:20:29 +0000 (08:20 +0000)]
added set_ppterm
Claudio Sacerdoti Coen [Mon, 31 Mar 2008 17:56:14 +0000 (17:56 +0000)]
Automatic generation of elimination and inversion principles for co-inductive
types disabled (since it does not make sense!)
Claudio Sacerdoti Coen [Mon, 31 Mar 2008 17:45:01 +0000 (17:45 +0000)]
Large amount of duplicated code (still in comments) factorized.
Claudio Sacerdoti Coen [Mon, 31 Mar 2008 17:02:48 +0000 (17:02 +0000)]
1) Impredicative sort "Set" removed everywhere.
2) Sort "CProp" is now predicative.
3) Optimized case IRL in check_metasenv_consistency.
Claudio Sacerdoti Coen [Mon, 31 Mar 2008 14:10:21 +0000 (14:10 +0000)]
1) more sharing everywhere in NCicSubstitution
2) type of iterators in nCicUtils changed to take a NCic.hypothesis when
crossing binders
Enrico Tassi [Thu, 27 Mar 2008 15:07:09 +0000 (15:07 +0000)]
more cases of the type checker honoured, still missing Match and Const.
lookup_meta added