Commit graph

4541 commits

Author SHA1 Message Date
Leonardo de Moura
065a1f7501 feat(library/tactic): add 'induction' tactic skeleton 2015-05-12 20:21:25 -07:00
Leonardo de Moura
2014a4a672 chore(script/gen_tokens_cpp): mark automatically generated global variables as static 2015-05-12 17:15:31 -07:00
Leonardo de Moura
358afcf42c fix(script/gen_tokens_cpp): automatically generated header 2015-05-12 17:11:27 -07:00
Leonardo de Moura
c60f11ab05 refactor(frontends/lean): add script for automatically generating tokens.h and tokens.cpp 2015-05-12 17:07:08 -07:00
Leonardo de Moura
99084dce1c chore(library/user_recursors): fix style 2015-05-12 15:48:24 -07:00
Leonardo de Moura
f403ea984b feat(frontends/lean): add 'print [recursor]' command for debugging purposes 2015-05-12 15:48:24 -07:00
Leonardo de Moura
750f6d5a43 feat(library,frontends/lean): validate user defined recursors and add attribute to mark them
see issue #492

The user-defined recursors will also be used to implement the blast tactic
2015-05-12 15:48:01 -07:00
Leonardo de Moura
b6fff9fbe1 chore(tests/library/blast/union_find): fix style 2015-05-12 06:24:58 -07:00
Leonardo de Moura
fa70930ef4 feat(library/blast): add union-find datastructure 2015-05-11 16:19:51 -07:00
Leonardo de Moura
9649d540c0 fix(library/tactic/rewrite_tactic.cpp): memory leak 2015-05-11 16:19:17 -07:00
Leonardo de Moura
3d8586088b fix(frontends/lean/elaborator): problem with 'calc' proofs discussed at issue #592 2015-05-11 13:22:39 -07:00
Leonardo de Moura
4905233eb2 fix(CMakeLists.txt): warning on OSX 2015-05-11 12:55:52 -07:00
Jeremy Avigad
95dae11670 feat(src/emacs/lean-input.el): add abbreviations 'mem', 'subeq', etc. 2015-05-11 09:03:56 -07:00
Leonardo de Moura
c5fb3ec6d0 fix(library/definitional/equations): fixes #541
This commit allows recursive applications to have less or more arguments
than the equation left-hand-side.
We add two tests
   - 541a.lean  recursive call with more arguments
   - 542b.lean  recursive call with less arguments
2015-05-10 20:37:44 -07:00
Leonardo de Moura
207e8e86da fix(library/definitional/equations): do not clear eliminated hypotheses when invoking 'cases' tactic from definitional package 2015-05-10 11:08:02 -07:00
Leonardo de Moura
299fd5c919 feat(frontends/lean/pp): pp inaccessible patterns 2015-05-10 11:08:02 -07:00
Leonardo de Moura
e9c8de7bdf feat(frontends/lean): remove unnecessary option 2015-05-09 11:49:55 -07:00
Leonardo de Moura
bd28396be0 feat(kernel): transparent theorems
closes #576
2015-05-09 11:42:29 -07:00
Leonardo de Moura
f63c2d9393 feat(kernel/inductive/inductive): for datatypes that support K, we should try K before normalizing the major premise 2015-05-09 11:23:10 -07:00
Leonardo de Moura
d217655a62 fix(kernel/quotient/quotient): bug in reduction rule for quot.ind 2015-05-09 09:07:03 -07:00
Leonardo de Moura
4e35afedcc feat(frontends/lean): rename 'wait' to 'reveal'
Signed-off-by: Leonardo de Moura <leonardo@microsoft.com>
2015-05-08 20:54:16 -07:00
Leonardo de Moura
f8e2f68ce0 feat(frontends/lean): add 'wait' command
This commit also fixes several problems with -j option (parallel
compilation). The .olean files were not missing data when -j was used

see issue #576
2015-05-08 20:05:21 -07:00
Leonardo de Moura
e6566e5814 fix(frontends/lean/parser): incorrect error message 2015-05-08 18:13:44 -07:00
Leonardo de Moura
e59456c19f feat(frontends/lean): remove 'opaque' keyword
see issue #576
2015-05-08 16:45:13 -07:00
Leonardo de Moura
cf7e60e5a6 refactor(kernel): remove "opaque" field from kernel declarations
see issue #576
2015-05-08 16:06:16 -07:00
Leonardo de Moura
6c958a25e7 refactor(library/tactic/expr_to_tactic): make sure builtin tactics don't need to be marked opaque
This modification is needed since we will remove opaque definitions from
the kernel.

see issue #576
2015-05-08 16:06:16 -07:00
Leonardo de Moura
57ea660963 refactor(*): start process for eliminating of opaque definitions from the kernel
see issue #576
2015-05-08 16:06:04 -07:00
Leonardo de Moura
061e26157e fix(kernel,library): make sure macros check relevant arguments when kernel is performing full type checking 2015-05-08 12:41:23 -07:00
Leonardo de Moura
72663e8a06 feat(library/tactic/rewrite_tactic): take hypotheses into account when checking rewrite step 2015-05-08 11:34:50 -07:00
Leonardo de Moura
12bad8794b feat(frontends/lean/pp): use 'assert' instead of 'have ... [visible]' 2015-05-08 10:02:15 -07:00
Jeremy Avigad
9beb70b292 feat(src/emacs/lean-input.el): add 'empty' as abbreviation for emptyset 2015-05-08 20:04:55 +10:00
Leonardo de Moura
0b57f7d00a refactor(library/tactic): refine interface between tactic and proof-term modes
Some constraints were being lost with the previous interface.
This is why we had a workaround in fintype.lean.

We can also remove some hacks we have used in the past.
2015-05-07 18:02:51 -07:00
Leonardo de Moura
12b818b7d3 fix(library/util): "empty" is not universe polymorphic anymore 2015-05-07 16:43:58 -07:00
Floris van Doorn
20e62c9623 fix(lean-syntax): add all metaclasses and the state tactic to syntax 2015-05-07 16:39:03 -07:00
Leonardo de Moura
eb3a236119 fix(library/unifier): typo
fixes #588
2015-05-07 16:30:02 -07:00
Leonardo de Moura
88cd6e9a63 chore(library/unifier): remove dead code 2015-05-07 15:49:33 -07:00
Leonardo de Moura
45c8cdc626 fix(kernel/default_converter): discard constraints when optimization fails 2015-05-07 15:02:29 -07:00
Leonardo de Moura
f6a1d1c864 fix(frontends/lean/parser): fixes #584 2015-05-07 14:24:30 -07:00
Leonardo de Moura
aff9257c72 feat(frontends/lean): allow → to be used in calc proofs
see issue #586
2015-05-07 12:28:47 -07:00
Leonardo de Moura
b03266be70 feat(library/normalize,frontends/lean): rename '[unfold-m]' hint to '[constructor]', and allow it to be attached to constants
closes #587
2015-05-07 12:00:34 -07:00
Leonardo de Moura
5fdf140096 refactor(frontends/lean): simplify local_decls data-structure
This is the first step for fixing #584
2015-05-07 11:10:15 -07:00
Leonardo de Moura
ac24f19210 fix(library/tactic/rewrite_tactic): use relaxed type checker when processing rewrite "proof step"
That is, the restricted type checker should only be used in the
matching/unification step.

fixes #583
2015-05-07 09:57:23 -07:00
Leonardo de Moura
e0b7017435 feat(frontends/lean): make sure the proof state bit "report errors" is
set in the beginning of each begin-end element

see discussion at #583
2015-05-07 09:43:39 -07:00
Leonardo de Moura
5798ac43de fix(frontends/lean/structure_cmd): 'structure' command must set unfold-c attribute for auxiliary recursors
fixes #582
2015-05-07 09:09:07 -07:00
Leonardo de Moura
171530d5cc fix(frontends/lean/notation_cmd): fixes #585 2015-05-07 08:36:37 -07:00
Leonardo de Moura
16b7bc3922 fix(kernel/default_converter): broken optimization
we must also check the universe levels when applying the optimization for
constraints of the form:

            f.{l_1 ... l_k} a_1 ... a_n  =?= f.{l_1' ... l_k'} b_1 ... b_n

The optimization tries to avoid unfolding f if we can establish that
a_i is definitionally equal to b_i for each i in [1, n]

closes #581
2015-05-06 18:32:41 -07:00
Leonardo de Moura
e841852be4 feat(library/max_sharing): maximize sharing in universe level expressions 2015-05-06 14:26:12 -07:00
Leonardo de Moura
210cae7d6c fix(library/tactic): we must unfold untrusted macros when type checking intermediate results
Example: before this commit, the file librata/data/list/perm.lean would
not type check if the option -t 0 (trust level zero) was provided.
Reason: the intermediate term contained a macro, and macros are not
allowed at trust level zero.
2015-05-06 14:23:58 -07:00
Leonardo de Moura
b5619c2109 fix(library/export): remove workaround from 'export' procedure
The workaround was needed due to a bug in the max_sharing procedure
2015-05-06 14:00:57 -07:00
Leonardo de Moura
2ca16a099b fix(library/max_sharing): conflict with the kernel approximate caching 2015-05-06 13:57:36 -07:00