Commit graph

1807 commits

Author SHA1 Message Date
Leonardo de Moura
092c8d05b9 feat(frontends/lean,library): rename '[rewrite]' to '[simp]' 2015-07-22 09:01:42 -07:00
Leonardo de Moura
23dd47d27f fix(tests/lean): adjust tests to recent changes to the standard library 2015-07-19 21:32:42 -07:00
Leonardo de Moura
5112232d6d fix(tests/lean/run/finset): adjust test to recent changes to the
standard library
2015-07-19 11:53:21 -07:00
Leonardo de Moura
3ab0e07ba9 feat(frontends/lean): add simp tactic frontend stub
This commit also removes the fake_simplifier. It doesn't work anymore
because simp is now a reserved word.
2015-07-14 09:54:53 -04:00
Leonardo de Moura
589f9df103 fix(tests/lean): adjust tests to reflect recent changes in the standard library 2015-07-13 21:46:09 -04:00
Leonardo de Moura
267545ca0c feat(frontends/lean): parse 'with_options' tactical
see issue #492
2015-07-13 19:13:41 -04:00
Leonardo de Moura
f8d472c9f1 feat(frontends/lean/parse_rewrite_tactic): change the semantics of rewrite[↑f] when f is recursive
After this commit it behaves like 'unfold f'.
That is, it will unfold f even if it fails to fold recursive
applications. Now, only 'esimp[f]' will not unfold f-applications when
it cannot fold the recursive applications.

This commit also closes #692. It is part of a series of commits that
addresses this issue.

closes #692
2015-07-12 13:20:21 -04:00
Leonardo de Moura
4c0a656ecc fix(library/tactic/unfold_rec): support indexed families + brec_on at unfold_rec
see issue #692
2015-07-12 12:45:05 -04:00
Leonardo de Moura
7fa5c3e5da feat(library/tactic/unfold_rec): take '[recursor]' annotations into account at unfold_rec 2015-07-12 11:33:40 -04:00
Leonardo de Moura
f0e3eef38a test(tests/lean/unfold_rec2): add example from #692 2015-07-11 19:45:27 -04:00
Leonardo de Moura
8e8e08cfe7 feat(library/tactic): use occurrence object in unfold tactic family 2015-07-11 18:53:45 -04:00
Leonardo de Moura
554a42b407 fix(library/tactic/unfold_rec): add annother brec pattern that should be checked in the unfold recursive definition tactic 2015-07-10 22:16:23 -04:00
Leonardo de Moura
a9515ac7a4 feat(library/tactic/rewrite_tactic): try to fold nested recursive applications after unfolding a recursive function
See issue #692.
The implementation still has some rough spots.
It is not clear what the right semantic is.
Moreover, the folds in e_closure could not be eliminated automatically.
2015-07-08 21:19:18 -04:00
Leonardo de Moura
26574e29a9 feat(library/normalize,frontends/lean): allow multiple arguments in [unfold] hint
closes #693
2015-07-07 18:01:57 -07:00
Leonardo de Moura
a27b20cd9c feat(frontends/lean/notation_cmd): allow local notation to override reserved notation
closes #712
2015-07-07 17:30:46 -07:00
Leonardo de Moura
4b1b3e277f feat(frontends/lean): rename '[unfold-c]' to '[unfold]' and '[unfold-f]' to '[unfold-full]'
see issue #693
2015-07-07 16:37:06 -07:00
Leonardo de Moura
b0c56273e2 fix(frontends/lean/elaborator): fixes #724 2015-07-06 15:19:19 -07:00
Leonardo de Moura
7e0844a9e6 fix(tests): to reflect recent changes in the standard library 2015-07-06 15:05:01 -07:00
Leonardo de Moura
c843690d27 fix(frontends/lean/elaborator): fixes #719 2015-07-03 12:37:28 -07:00
Leonardo de Moura
aa338f6002 test(tests/lean/run): add test showing new coercion module addresses issue #668 2015-07-01 16:41:19 -07:00
Leonardo de Moura
4ae9f3ea81 feat(library/coercion): new coercion manager
closes #668
2015-07-01 16:32:34 -07:00
Leonardo de Moura
0f5b7a36f5 chore(library/coercion): remove lua bindings for coercion module
Reason: we will refactor the coercion module.
See issue #668
2015-07-01 14:08:49 -07:00
Leonardo de Moura
cf574d0127 feat(library): assign priorities to notation declarations in the standard library
Now, even if the user opens the namespaces in the "wrong" order, the
notation + coercions will behave as expected.
2015-06-30 17:38:13 -07:00
Leonardo de Moura
cabe30ba71 feat(frontends/lean): allow user to assign priorities to notation declarations 2015-06-30 17:10:27 -07:00
Leonardo de Moura
e635d9be9f refactor(kernel): rename get_weight to get_height at declaration
Motivation:
- It is the standard name for the concept: declaration height
- Avoid confusion with the expression weight
2015-06-30 12:59:10 -07:00
Leonardo de Moura
0fc2efe88e fix(library/tactic/rewrite_tactic): fixes #702 2015-06-28 20:37:17 -07:00
Leonardo de Moura
95720b1670 fix(frontends/lean/elaborator): fixes #687 2015-06-28 19:58:57 -07:00
Leonardo de Moura
ecfc01b2d0 test(tests/lean/run): add test for <d notation
see issue #695

remark: we have to fix the tutorial the ASCII notation for fold is `<d`
instead of `<D`
2015-06-28 13:10:15 -07:00
Leonardo de Moura
1b864a838f fix(library/tactic/induction_tactic.cpp): condition for checking whether 'induction' tatic is applicable or not
fixes #690
2015-06-28 13:07:02 -07:00
Leonardo de Moura
2f75768243 feat(library/tactic/rewrite_tactic): fail when nothing is rewritten
fixes #686
2015-06-28 12:05:00 -07:00
Leonardo de Moura
d1eaa7bcda feat(frontends/lean/parse_rewrite_tactic): accept trailing comman in rewrite tactic
see issue #695
2015-06-28 11:45:30 -07:00
Leonardo de Moura
4a4ef48344 fix(frontends/lean/parse_rewrite_tactic): fixes #691 2015-06-28 11:28:05 -07:00
Leonardo de Moura
869ad261c5 fix(frontends/lean/elaborator): fixes #689 2015-06-27 16:19:38 -07:00
Leonardo de Moura
97b3fd45ce fix(library/logic/axioms/prop_decidable): fixes #704 2015-06-26 19:14:36 -07:00
Leonardo de Moura
2aa64034df fix(tests/lean): adjust tests to reflect changes in the elaboration process 2015-06-26 17:18:30 -07:00
Floris van Doorn
9ef1ae0848 fix(tests): update tests to reflect the change of notation from \~ to ~ 2015-06-25 22:55:05 -04:00
Leonardo de Moura
1b414d36e7 refactor(library/init): define prod as an inductive datatype
Motivation: prod is used internally in the definitional package.
If we define prod as a structure, then Lean will tag pr1 and pr2 as
projections. This creates problems when we add special support for
projections in the elaborator. The heuristics avoid some case-splits
that are currently performed, and without them some files break.
2015-06-25 17:59:06 -07:00
Floris van Doorn
fa1979c128 feat(datatypes): let the type of unit be the lowest non-Prop universe
The definitional package (brec_on and cases_on) now use poly_unit instead of unit

closes #698
2015-06-25 17:33:46 -07:00
Leonardo de Moura
a4c0699e81 feat(library/tactic/constructor_tactic): restore 'constructor' tactic old semantics, add 'fconstructor' tactic
See issue #676

Add new test demonstrating why it is useful to have the old semantics
for 'constructor'
2015-06-17 23:48:54 -07:00
Leonardo de Moura
bf71d9f342 fix(library/tactic/rewrite_tactic): fixes #682 2015-06-17 18:49:02 -07:00
Leonardo de Moura
ce8f2a1674 feat(library/class): allow any constant to be marked as a class
closes #679
2015-06-17 16:26:45 -07:00
Leonardo de Moura
9e6e406f73 feat(frontends/lean): add '#compose' command for testing composition manager 2015-06-17 14:42:25 -07:00
Leonardo de Moura
0ae24faae3 feat(library/tactic/constructor_tactic): use 'fapply' in 'constructor' tactic
closes #676
2015-06-16 12:03:31 -07:00
Leonardo de Moura
0dc674e350 fix(tests/lean/630): adjust test to reflect changes in the standard library 2015-06-16 11:33:26 -07:00
Jeremy Avigad
a4a8253f50 refactor(library,hott,tests): rename succ_inj to succ.inj, add abbreviation eq_of_succ_eq_succ 2015-06-15 22:52:38 +10:00
Leonardo de Moura
d8620ef4c9 fix(kernel,library,frontends/lean): improve error messages
see #669
2015-06-14 19:44:00 -07:00
Leonardo de Moura
8a85e4ee87 feat(frontends/lean/builtin_cmds): improve print <id> when <id> is a not yet revealed theorem
We add a remark saying the command `reveal <id>` should be used to
access `<id>` definition.
2015-06-13 12:12:22 -07:00
Leonardo de Moura
591fb91f34 fix(frontends/lean/builtin_cmds): fixes #671 2015-06-13 11:35:03 -07:00
Leonardo de Moura
9fbf267a3f feat(library/tactic/rewrite_tactic): when 'rewrite' step fails apply esimp and try again
closes #670
2015-06-12 19:48:58 -07:00
Leonardo de Moura
3d9b557cfd feat(frontends/lean): allow the user to mark subterms that should be automatically abstracted into new definitions
closes #484
2015-06-12 17:49:26 -07:00