Commit graph

4738 commits

Author SHA1 Message Date
Leonardo de Moura
897289599a fix: ignore auxiliary declarations at subst and assumption 2020-09-13 09:59:37 -07:00
Leonardo de Moura
ac6fd1382d fix: collectDeps issue 2020-09-13 09:47:00 -07:00
Leonardo de Moura
e181c1adee refactor: add DefEqM
The idea is to make clear that the field `posponed` is transient
state. It is only used during `isDefEq`.
The refactoring was motivated by a bug I found where the `posponed`
constraints were not being handled correctly. For example,
the `check (e : Expr)` method was returning `true`, but leaving pending
universe constraints at `postponed`.

cc @Kha
2020-09-12 16:42:16 -07:00
Leonardo de Moura
0967ad7e91 feat: extend/cleanup isLevelDefEqAux
`tests/lean/run/toExpr.lean` could not be processed by the new
frontend without these extensions.
2020-09-12 15:38:48 -07:00
Leonardo de Moura
ab44cd28f1 chore: move more tests to new frontend 2020-09-12 09:04:39 -07:00
Leonardo de Moura
e1e65e83da fix: try to synthesize instance for substructures 2020-09-12 08:58:13 -07:00
Leonardo de Moura
a784005729 chore: move more tests to new frontend 2020-09-12 07:54:12 -07:00
Leonardo de Moura
09244c73f4 fix: lval resolution 2020-09-12 07:45:36 -07:00
Leonardo de Moura
cc520d422c fix: ensure no unassigned mvars at #eval 2020-09-12 07:18:43 -07:00
Leonardo de Moura
f242a958fc chore: move more tests to new frontend 2020-09-11 17:36:37 -07:00
Leonardo de Moura
a8de3c5115 fix: isDefEq configuration in the elaborator 2020-09-11 17:36:28 -07:00
Leonardo de Moura
bbc3bec53b chore: move more tests to new frontend 2020-09-11 16:28:57 -07:00
Leonardo de Moura
bf703c9cab fix: bug at inductive datatype resulting universe inference 2020-09-11 16:07:22 -07:00
Leonardo de Moura
b02cda0408 chore: move more tests to new frontend 2020-09-11 15:31:14 -07:00
Leonardo de Moura
2c2a357f28 chore: remove disabled test 2020-09-11 15:05:00 -07:00
Leonardo de Moura
151012cb39 feat: remove emptyc elaboration hack
@Kha I removed the hack. We know get a nice error message.
2020-09-11 14:41:44 -07:00
Leonardo de Moura
ff9b562e66 chore: move test 2020-09-11 14:40:21 -07:00
Leonardo de Moura
cc6104b369 fix: bug in the old frontend
It had to be fixed because we have code that needs to be compiled with
the old frontend, and the hacks we used to workaround this issue are
incompatible with the new frontend.
2020-09-11 14:12:03 -07:00
Leonardo de Moura
7950413526 feat: improve choice nodes + postpone 2020-09-11 12:45:38 -07:00
Leonardo de Moura
81987e59e7 fix: expected type propagation issue 2020-09-11 11:03:59 -07:00
Leonardo de Moura
1731962ea9 feat: elaborate emptyC notation
@Kha this is a little bit ugly. As in Lean3, the notation `{ }` is
overloaded. It can be `HasEmptyc.emptyc` or the empty structure.
Moreover, if we have an instance `HasEmptyc A` for a structure `A`,
then `{ }` is notation for `HasEmptyc.emptyc`. I am not very happy
about this, but it seems users use this feature. It is even used in
our own Lean4 code base :)

BTW, I will try a different alternative later.
First, move the notation `{ }` as a proper parser for `emptyc`.
Then, make sure the elaborator throws an error when
`{ }` is ambiguous. That is, `{}` can be elaborated as the empty
structure `A` and there is an `HasEmptyc A` instance.
Example:
```
structure A :=
(x : Nat := 0)

instance : HasEmptyc A :=
⟨{ x := 10 }⟩
```

The main issue here is that the support for choice+postpone is
currently broken.
2020-09-11 10:42:39 -07:00
Leonardo de Moura
4ddd4c4657 chore: move more tests to new frontend 2020-09-11 10:28:00 -07:00
Leonardo de Moura
290cd5cf0f chore: delete disabled tests 2020-09-11 08:38:12 -07:00
Leonardo de Moura
5db5cf7734 fix: do not lift (<- action) nested in quotations 2020-09-11 08:06:20 -07:00
Leonardo de Moura
54e006dc53 chore: move test to new frontend 2020-09-11 07:47:19 -07:00
Leonardo de Moura
6cf652606a chore: use new frontend 2020-09-10 19:34:05 -07:00
Leonardo de Moura
e4a3b434d7 chore: moving tests to new frontend
@Kha The transition has begun :)
I found and fixed a few bugs, but it is going well so far.
2020-09-10 18:00:34 -07:00
Leonardo de Moura
ba4fdce508 feat: expand helper macros 2020-09-10 14:25:07 -07:00
Leonardo de Moura
6b7088e71a fix: too restrictive condition 2020-09-10 11:20:09 -07:00
Leonardo de Moura
20152c1192 chore: change by precedence
@Kha it now uses the same precedence of `fun`.
The main motivation is to allow us to write `@by { ... }` instead of
`@(by { ... })`.

BTW, I am considering disabling implicit lambdas for `by ...` expressions.
It is automatically "intro"ducing the implicit variables without
giving an opportunity for users to pick their names.
2020-09-10 11:17:42 -07:00
Leonardo de Moura
2214d81e84 test: add match with proofs example 2020-09-09 16:59:45 -07:00
Leonardo de Moura
d3de12fa09 test: another dependent pattern matching test 2020-09-09 16:51:46 -07:00
Leonardo de Moura
b136c519e2 fix: scope and improve error message 2020-09-09 16:44:43 -07:00
Leonardo de Moura
20bc004c70 feat: improve subst tactic 2020-09-09 16:20:15 -07:00
Leonardo de Moura
83410083a8 fix: do not generate code for noncomputable definitions 2020-09-09 15:28:41 -07:00
Leonardo de Moura
c78d0cf1f8 feat: use | position for reporting dependent-match elimination errors 2020-09-09 14:07:23 -07:00
Leonardo de Moura
854cc3418e feat: improve error message for dependent-match elimination failures
@Kha This is a first attempt to improve the error message for examples
like the one Andrew Kent posted on Zulip.
I created a simpler example using "vectors".
2020-09-09 13:43:06 -07:00
Leonardo de Moura
c9f4f858b1 feat: ellipsis in constructor application patterns
Given
```
inductive Foo
| mk₁ (x y z w : Nat)
| mk₂ (x y z w : Nat)
```
We can now write
```
def Foo.z : Foo → Nat
| mk₁ (z := z) .. => z
| mk₂ (z := z) .. => z
```
instead of
```
def Foo.z : Foo → Nat
| mk₁ _ _ z _ => z
| mk₂ _ _ z _ => z
```

cc @Kha
2020-09-09 10:21:06 -07:00
Leonardo de Moura
276e6a55ba feat: improve collectPatternVars
Add support for `@` and named parameters.
Fix how ctor fields are processed.
2020-09-09 09:52:25 -07:00
Leonardo de Moura
2e11053eb5 fix: matchAlt macros should not consume implicit arguments
As in Lean3, the following example is a valid definition

```
def head : {α : Type} → List α → Option α
| _, a::as => some a
| _, _     => none
```
2020-09-08 17:31:57 -07:00
Leonardo de Moura
ea2e86afba feat: add Array.allDiff 2020-09-08 16:16:14 -07:00
Leonardo de Moura
77f8fe0e41 fix: mkLetRecClosureFor 2020-09-08 15:14:46 -07:00
Leonardo de Moura
fc1e4cb533 feat: add Array.isPrefixOf 2020-09-08 14:40:43 -07:00
Leonardo de Moura
ecda364985 feat: add Expr.forEach 2020-09-08 13:03:53 -07:00
Leonardo de Moura
9151fef49d feat: abstract proofs occurring in binders 2020-09-08 12:29:35 -07:00
Leonardo de Moura
0a853b2c44 feat: abstract nested proofs in definitions 2020-09-08 11:59:26 -07:00
Leonardo de Moura
ebe2cf272e feat: add MonadCacheT 2020-09-08 10:48:47 -07:00
Leonardo de Moura
ef87b33cdc test: add another missing hole test 2020-09-07 17:18:20 -07:00
Leonardo de Moura
fc0be5e391 feat: improve error message position when definition type is metavar 2020-09-07 17:16:44 -07:00
Leonardo de Moura
18a9cb9b43 feat: improve error message position for binders with implicit types 2020-09-07 17:02:43 -07:00