Ferruccio Guidi [Thu, 26 Feb 2009 21:15:13 +0000 (21:15 +0000)]
cicInspect: now we can choose not to count the Cic.Implicit constructors
proceduralTypes, proceduralOptimizer: we do not count the Cic.Implict nodes
proceduralTeX: bug fix
Ferruccio Guidi [Tue, 17 Feb 2009 19:37:22 +0000 (19:37 +0000)]
- Coq/preamble: missing alias added
- nUri: head comment fixed
- Procedural: the cases tactic is now detected and generated
but does not compile :))
- grafiteAstPp: output syntax of the cases tactic fixed
- grafiteParser: impovement in the code for the elim tactic
Enrico Tassi [Thu, 5 Feb 2009 10:48:28 +0000 (10:48 +0000)]
a non necessary but morally required change. The matched term in
are_convertible is compared with test_eq_only=true (not really needed since
if it reduces to something it is a constructor an thus an application that already honors the test_eq_only=true check for arguments
Enrico Tassi [Mon, 2 Feb 2009 13:28:08 +0000 (13:28 +0000)]
CicTypeChecker.typecheck now takes an additional parameter:
- trust:bool, set to false only by the proofChecker daemon, while
the refiner (for example) passes true
Enrico Tassi [Fri, 30 Jan 2009 14:51:27 +0000 (14:51 +0000)]
fix convertibility in case of application test_eq_only is =true for
arguments, since we dont know if reduction eventually moves them in
contravariant position
Enrico Tassi [Mon, 26 Jan 2009 17:12:38 +0000 (17:12 +0000)]
maction support added to notation, adopted for = AKA = \sub t
deactivated the library/formal_topology directory since it is completely broken and outdated
Enrico Tassi [Mon, 19 Jan 2009 18:02:22 +0000 (18:02 +0000)]
- new notation.ma file with local and common notation
- o-bp have now a O as prefix, the same for o-cs
- BP_to_OBP proved to be a functor, not faitfull nor full (simply stated)
1. new coercion(s) from CPropi to CProp
2. fixed notation for subsetS
3. coercions still have problems: the coercion from objs1 REL is not
accepted because it is identified with the coercion from objs1 SET
Enrico Tassi [Thu, 15 Jan 2009 15:38:23 +0000 (15:38 +0000)]
the new coercion behaviour (variants + composition with ID) and the new
discipline of declaring hints for carrier of structures (setoids and categories) and no
other hints simplified many passages
Enrico Tassi [Thu, 15 Jan 2009 15:32:52 +0000 (15:32 +0000)]
- name mangling changed, added __ to separate additional number
- composing a coercion with an identity does not affect the body, and a variant is generated
Enrico Tassi [Tue, 13 Jan 2009 10:54:34 +0000 (10:54 +0000)]
many changes regarding coercions:
- new `prefer coercion foo` command to reorder the list of coercions associated with a src >-> tgt pair
- rewritten coercions undoing mechanism
- coercion composition fixed (generates more coercions)
- librarySync rewritten, procedures to generate derived lemmas (eliminators, invertion lemmas, ...) are hooks registerd by other modules that can thus live anywhere (also in tactics/ like the one for inversion)
- manual fixed to talk about the new command
- coercions graph is drawn together with a list of coercions that makes their precedence visible (and can be altered with the prefer coercion command)
- dump-moo fixed to print coercions
- removed `Coercion attribute in XML files (YOU NEED TO RECOMPILE)