Commit graph

778 commits

Author SHA1 Message Date
Leonardo de Moura
a26618e0f2 feat(frontends/lean): add '[]' notation for marking arguments where class-instance resolution should be applied 2014-10-12 13:06:00 -07:00
Leonardo de Moura
549f24335e feat(frontends/lean): do not allow coercion definition in sections 2014-10-11 18:41:17 -07:00
Leonardo de Moura
158682219f feat(frontends/lean): allow parameters only in contexts 2014-10-11 17:13:56 -07:00
Leonardo de Moura
f984b51291 feat(frontends/lean/notation_cmd): remove the cleanup notation hack 2014-10-11 16:40:26 -07:00
Leonardo de Moura
ca632cca13 feat(frontends/lean): add 'universe variable' command
We can declare variables anywhere. So, we must also be able do declare
"universe" variables anywhere. Here is a minimal example that requires
this feature

```
-- We want A and B to be in the same universe
universe variable l
variable A : Type.{l}
variable B : Type.{l}
definition tst := A = B
```

The following doesn't work because A and B are in different universes
```
variable A : Type
variable B : Type
definition tst := A = B
```

The following works, but tst is not universe polymorphic, since l is
one *fixed* global universe
```
universe l
variable A : Type.{l}
variable B : Type.{l}
definition tst := A = B
```
2014-10-11 14:22:33 -07:00
Leonardo de Moura
b0f8d86f26 feat(frontends/lean/parser): reject ambiguous parameter declaration, closes #242 2014-10-10 18:08:03 -07:00
Leonardo de Moura
5e6ff3eef3 fix(frontends/lean/elaborator): bug when reporting error position 2014-10-10 17:31:12 -07:00
Leonardo de Moura
f0523a3465 feat(frontends/lean): namespaces also define scope for variables 2014-10-10 16:21:30 -07:00
Leonardo de Moura
0641ee33ce feat(frontends/lean): allow variables anywhere 2014-10-10 16:16:19 -07:00
Leonardo de Moura
3d65b1c25c fix(frontends/lean/elaborator): incorrect type information being reports in lean-mode, fixes #241 2014-10-10 15:41:55 -07:00
Leonardo de Moura
235b8975d2 feat(kernel/inductive): K-like reduction in the kernel
Given (H_1 : a = a), we have that
      eq.rec H_2 H_1
reduces to H_2

This is not exclusive to equality.
It applies to any inductive datatype in Prop, containing only one
constructor with zero "arguments" (we say they are nullary).

BTW, the restriction to only one constructor is not needed, but it is
does not buy much to support multiple nullary constructors since Prop is
proof irrelevant.
2014-10-10 14:37:45 -07:00
Leonardo de Moura
402a351937 feat(frontends/lean): add 'universes' command 2014-10-10 08:45:59 -07:00
Leonardo de Moura
9ba59c6b25 feat(library/universe): improve support for universe level constraints in the unifier 2014-10-09 20:28:39 -07:00
Leonardo de Moura
8f1b6178a7 chore(*): minimize the use of parameters 2014-10-09 07:13:06 -07:00
Leonardo de Moura
8c5d3392c7 fix(library/algebra/category): minor fixes to reflect recent changes, and fix tests 2014-10-08 23:44:09 -07:00
Leonardo de Moura
8947bf4347 feat(frontends/lean): display type of binders, closes #238 2014-10-08 22:54:10 -07:00
Leonardo de Moura
25fd370c51 fix(frontends/lean): bug when using nested sections and parameters
see tests/lean/run/section4.lean
2014-10-08 22:23:20 -07:00
Leonardo de Moura
57c85221fe fix(frontends/lean): collect used universe levels after elaboration in the check command 2014-10-08 22:23:19 -07:00
Leonardo de Moura
eab9321a3b fix(frontends/lean): make all variables/parameters visible for 'variables' command 2014-10-08 22:23:19 -07:00
Leonardo de Moura
757cb46579 test(tests/lean/run): add no_confusion test 2014-10-08 22:23:19 -07:00
Leonardo de Moura
3b23bec25b feat(frontends/lean): variables/parameters and check commands have access to all section variables/parameters, closes #231 2014-10-08 08:40:55 -07:00
Leonardo de Moura
744cee8dd8 feat(frontends/lean): force 'classes' to be declared before instances are declared, closes #228 2014-10-07 18:02:15 -07:00
Leonardo de Moura
d8572e249d feat(frontends/lean/builtin_cmds): add 'print classes' command 2014-10-07 17:30:57 -07:00
Leonardo de Moura
90ece4dd1b feat(frontends/lean): remove tactic hints for specific classes
The idea is to separate class-instance resolution and tactic framework
as two independent engines.
2014-10-07 09:44:01 -07:00
Soonho Kong
e41573afc4 chore(tests/*/test.sh): change working dir; avoid using ls in for-loop 2014-10-06 11:20:13 -07:00
Leonardo de Moura
f7e1b67f6c test(tests/lean/run/print): add tests for 'print' cmd 2014-10-05 19:06:29 -07:00
Leonardo de Moura
a0d4d82f3f refactor(data/nat/order): use new policy for marking implicit arguments and '!' operator 2014-10-05 11:36:39 -07:00
Leonardo de Moura
73aa024c31 refactor(library/logic): remove 'core' subdirectory 2014-10-05 10:50:13 -07:00
Leonardo de Moura
317e910054 refactor(library/data/bool): use new style 2014-10-05 09:50:55 -07:00
Leonardo de Moura
60d8369688 fix(library/unifier): missing justification when updating choice constraints
The bug was not producing incorrect results, but really bad error
messages.
See: empty.lean.expected.out
2014-10-04 10:40:53 -07:00
Leonardo de Moura
64f6601fe3 fix(frontends/lean/proof_qed_elaborator): information about synthesized variables in a proof-qed block was being lost 2014-10-04 09:15:42 -07:00
Leonardo de Moura
7073f036d8 test(tests/lean/run): add additional tests for section parameter/variable and strict implicit argument 2014-10-04 08:15:43 -07:00
Leonardo de Moura
bf01cfeec5 fix(frontends/lean): avoid superfluous local references of the form @-1 (@ f),
This kind of term also confuses the elaborator
2014-10-04 07:55:32 -07:00
Leonardo de Moura
e71d4548de fix(frontends/lean): universe levels associated with section variables should not be fixed in the section 2014-10-04 07:13:19 -07:00
Leonardo de Moura
938e12baa1 feat(frontends/lean/notation_cmd): allow local numeric notation 2014-10-03 13:55:57 -07:00
Leonardo de Moura
fd013e8d07 chore(tests/lean/run/group3): cleanup test 2014-10-03 13:55:57 -07:00
Leonardo de Moura
c0725d1934 refactor(frontends/lean): remove 'including' expressions 2014-10-03 09:09:27 -07:00
Leonardo de Moura
01d4644026 fix(frontends/lean): bug in include/omit commands: in the end of section/context, the configuration must be restored 2014-10-03 08:52:35 -07:00
Leonardo de Moura
284f300440 feat(frontends/lean): add 'include' and 'omit' commands, closes #184 2014-10-03 07:23:24 -07:00
Leonardo de Moura
6950b4aa9b fix(frontends/lean/decl_cmds): do not allow section parameters/variables to shadow existing parameters/variables 2014-10-02 18:29:41 -07:00
Leonardo de Moura
f006d93794 feat(frontends/lean): section variables occur after section parameters 2014-10-02 17:55:34 -07:00
Leonardo de Moura
0a13e7863a feat(frontends/lean): enforce rule section parameters cannot depend on section variables 2014-10-02 17:55:34 -07:00
Leonardo de Moura
bf081ed431 refactor(kernel): rename var_decl to constant_assumption
Motivation: it matches the notation used to declare it.
2014-10-02 17:55:34 -07:00
Leonardo de Moura
4946f55290 refactor(frontends/lean): constant/axiom are top-level commands, parameter/variable/hypothesis/conjecture are section/context-level commands
Signed-off-by: Leonardo de Moura <leonardo@microsoft.com>
2014-10-02 17:55:34 -07:00
Leonardo de Moura
f78d831de3 refactor(frontends/lean): remove hardcoded Type', and define it using notation 2014-10-02 14:29:51 -07:00
Leonardo de Moura
4cb54ac825 feat(frontends/lean/elaborator): more strict test for bad universe solution 2014-10-02 14:29:51 -07:00
Leonardo de Moura
98e66586e9 feat(frontends/lean/elaborator): elaborator rejects 'Type' if the universe is explicit 2014-10-02 14:29:51 -07:00
Leonardo de Moura
153e3927ac feat(frontends/lean/elaborator): modify '!' semantics: it stops consuming arguments as soon it finds an argument that does not occur in the rest of the type. 2014-10-01 18:50:17 -07:00
Leonardo de Moura
e64d5c4a4a refactor(library/data/nat): use new operator '!' 2014-10-01 18:39:47 -07:00
Leonardo de Moura
716cd4d651 refactor(library): rename namespace eq_ops to eq.ops 2014-10-01 17:51:17 -07:00