Mario Carneiro
d3d5982544
fix(init/meta/mk_dec_eq_instance): bug for dependent structures
2017-06-25 14:55:47 -07:00
Leonardo de Moura
00092c293a
fix(library/check): improve ensure_type
2017-06-25 14:51:09 -07:00
Leonardo de Moura
41a1faa131
fix(frontends/lean/elaborator): check resulting type at visit_convoy
2017-06-25 10:50:40 -07:00
Leonardo de Moura
2075ec8d61
fix(library/equations_compiler/elim_match): bad trace message
2017-06-25 10:50:40 -07:00
Gabriel Ebner
af3f6fd181
fix(util/hash): add missing cast to unsigned
2017-06-25 14:58:18 +02:00
Gabriel Ebner
19e83d413d
fix(.codecov.yml): do not fail github ci if coverage drops by 0.01%
2017-06-25 10:35:02 +02:00
Gabriel Ebner
ed27334557
refactor(frontends/lean/scanner): use uchar for scanning
2017-06-25 10:06:29 +02:00
Leonardo de Moura
04dfd55f94
perf(kernel/expr): fix hash consing performance problem
...
The kernel support opportunistic hash consing.
At commit e63c79c81e#1646 was 10x
slower after the commit above).
After analyzing the performance logs, I conjecture the hash_consing
may produce a substantial performance overhead if the following property
doesn't hold.
- Let `cache` be the hash_consing cache, then for each `e` in `cache`,
all children of `e` are also in the `cache`.
If the property above is not true, we may have problem whenever
we add an expression `t` containing multiple copies of a big term `T`.
For example, support we insert `f T_1 T_2`, where `T_1` and `T_2` are
structurally equal but are not pointer equal. Then, if we try to insert
`f T_2 T_1`, we will have to compare the huge terms twice, and the
comparison will be proportional to the size of `T_i`.
This commit tries to address this performance problem by enforcing
the property above. This is not a perfect solution since we may keep
trying to create terms using big terms created before hash_consing
has been enabled. After this commit, the example at issue #1646
is only 1.4x slower. It didn't impact the standard library
compilation (memory nor time).
I'm using this commit is a temporary workaround. We should probably
remove the hash_consing support from the kernel, and implement it
in key places. For example, for Certigrad, we can control memory
consumption by using hash_consing only during `simp`,
`instantiate_mvars` and `elaborator` finalization procedure.
@gebner @kha Any ideas/suggestions for this hash_consing issue?
2017-06-23 12:41:35 -07:00
Gabriel Ebner
16fb5ade58
fix(library/string): correctly escape non-printable characters
...
This also fixes a compiler warning on ARM, where `0 <= c` is always
true.
2017-06-23 20:39:41 +02:00
Gabriel Ebner
7342741263
fix(CMakeLists): do not use -fsigned-char
2017-06-23 18:21:30 +02:00
Gabriel Ebner
4a6513e5f5
refactor(util/serializer,library/module): use basic_ostream::write for the olean code
2017-06-23 15:13:40 +02:00
Gabriel Ebner
31162df650
fix(frontends/lean/structure_instance): compiler warning
2017-06-23 08:31:04 +02:00
Leonardo de Moura
e63c79c81e
feat(frontends/lean): enable hash_consing during tactic execution
...
This commit is trying to address a memory consumption problem in
@dselsam project.
2017-06-22 17:24:27 -07:00
Leonardo de Moura
9cb94847cb
fix(library/equations_compiler): propagate m_ignore_if_unused flag
2017-06-22 17:01:31 -07:00
Leonardo de Moura
58f0561f41
fix(library/equations_compiler/equations): is_equation
2017-06-22 17:01:08 -07:00
Leonardo de Moura
e905e83e9e
chore(library, frontends/lean): use override
2017-06-22 16:15:21 -07:00
Leonardo de Moura
096b437c11
fix(library/equations_compiler, frontends/lean): missing operator== for macro_definition_cell subclasses
2017-06-22 16:13:29 -07:00
Gabriel Ebner
30a9217a78
feat(library/type_context): unfold lemmas in major premise of acc.rec
2017-06-22 08:33:11 -07:00
Sebastian Ullrich
4d5d2abcba
fix(init/meta): fix build
2017-06-22 08:24:36 -07:00
Mario Carneiro
47c3c0ba84
fix(tests/*): fix tests
2017-06-22 08:24:19 -07:00
Mario Carneiro
09af93186a
fix(frontends/lean/elaborator): @applications don't make thunks
2017-06-22 08:24:11 -07:00
Leonardo de Moura
d666742219
doc(changes): update
2017-06-22 08:07:25 -07:00
Sebastian Ullrich
0a48809469
refactor(frontends/lean/tactic_notation): rename note/define tactics to have/let
2017-06-22 08:03:23 -07:00
Mario Carneiro
495093f6c0
fix(tests/run): fix tests
2017-06-22 08:03:23 -07:00
Mario Carneiro
9f31096892
refactor(init/meta/interactive): rename pose -> define
2017-06-22 08:03:23 -07:00
Mario Carneiro
b775a01fba
refactor(init/meta/interactive): merge assert -> note
2017-06-22 08:03:23 -07:00
Leonardo de Moura
671b6240c1
fix(library/init/meta): make sure rw still produces useful error messages
2017-06-21 20:58:59 -07:00
Leonardo de Moura
14d768ffa2
feat(library/init/meta/interactive): add simp_all tactic
...
@Armael I added the simp_all tactic. See new test for an example.
2017-06-21 20:43:56 -07:00
Leonardo de Moura
058d073cba
feat(library/init/meta/interactive): allow user to write rw [f] to rewrite using the equational lemmas for f
...
The tactic succeeds if the expression can be rewritten using one of the
equational lemmas associated with `f`.
See discussion at #1680
2017-06-21 18:56:18 -07:00
Leonardo de Moura
35b99aafb2
feat(library/init/category/combinators): add list.mfirst
2017-06-21 18:53:22 -07:00
Leonardo de Moura
51a28828d6
chore(tests/lean/interactive/info_tactic): fix test
2017-06-21 17:03:24 -07:00
Leonardo de Moura
3195a800f4
feat(library/tactic/simp_lemmas): use proof for refl lemmas at simp_lemma_rewrite too
...
I forgot to do it at 9fcb3ae4b5
2017-06-21 16:53:35 -07:00
Leonardo de Moura
b9dee04fdb
feat(library/tactic/simplify): add single_pass simplifier option (default is false)
2017-06-21 16:41:54 -07:00
Leonardo de Moura
9fcb3ae4b5
feat(library/tactic/simplify): store proof for refl lemmas and use them in simp
...
Before this commit, simp would not silently apply refl-lemmas, and use
reflexivity. This strategy produces compact proofs but may generate
performance problems. For example, the new test timeouts without this
commit.
I believe a similar performance problem is affecting the Certigrad
project developed by @dselsam.
2017-06-21 16:21:11 -07:00
Leonardo de Moura
fd17a19a23
chore(library/tactic/simplify): remove dead option
2017-06-21 15:32:08 -07:00
Scott Morrison
629d5ebbe8
feat(init/meta/simp_tactic) tactics that revert hypotheses should allow elet expressions
...
Conflicts:
library/init/meta/simp_tactic.lean
2017-06-21 13:58:46 -07:00
Scott Morrison
6aba80d389
feat(library/init/meta/simp_tactic,library/init/meta/interactive) interactive tactics for unfold_projection
2017-06-21 13:51:52 -07:00
Sebastian Ullrich
565e583915
chore(doc/changes): fix link
2017-06-21 13:23:02 -07:00
Sebastian Ullrich
676e8eee6c
chore(doc/changes): update changelog
2017-06-21 13:22:47 -07:00
Gabriel Ebner
814a5edaf1
fix(library/module_mgr): check for errors when writing olean files
2017-06-21 11:25:25 +02:00
Gabriel Ebner
a0a70a678a
fix(library/module): check for end-of-file in deserialization
2017-06-21 11:25:25 +02:00
Gabriel Ebner
942f7bf8b5
fix(library/module): unify hash computation
2017-06-21 10:50:54 +02:00
Mario Carneiro
e6f3c5cc22
fix(tests/lean/interactive/mk_input): strip \r from input files (win)
2017-06-21 08:53:11 +02:00
Leonardo de Moura
eef4d95410
feat(frontends/lean/inductive_cmds): closes #1655
2017-06-20 16:25:18 -07:00
Leonardo de Moura
e089fe6ee7
fix(library/expr_lt): make sure find_local_decl is not tested in the Meta case, and add /* fall-thru */
...
See comments at commit b25291c5c9
2017-06-20 13:47:30 -07:00
Leonardo de Moura
5a07b3b27c
test(tests/lean/run/1688): add test for PR #1688
2017-06-20 12:36:31 -07:00
Mario Carneiro
636e65a216
fix(init/meta/interactive): elab terms of change-with using same type
2017-06-20 12:34:29 -07:00
Leonardo de Moura
b25291c5c9
fix(library/tactic/simplify): fixes #1685
2017-06-20 12:27:46 -07:00
Leonardo de Moura
ce3387b246
fix(library/tactic/change_tactic): fixes #1686
2017-06-20 12:05:21 -07:00
Leonardo de Moura
3c306d0a7b
fix(library/equations_compiler/elim_match): forward dependency checking
...
See comment at #1594
This commit is not fixing the issue, but a problem described in one of
the comments.
2017-06-20 11:29:23 -07:00
