Commit graph

797 commits

Author SHA1 Message Date
Ayaz Hafiz
54bb258aac
Remove stale uniqueness inference comment 2022-12-27 09:26:08 -06:00
Ayaz Hafiz
1e847efbfe
Correctly choose specialized shapes for anonymous closures 2022-12-27 09:18:41 -06:00
Ayaz Hafiz
593344f5c5
Determine the proc layout based on captures of an individual lambda
Previously, we determined whether a closure argument should be added to
the proc layout of a compiled function by checking whether the lambda
set was material at all. But, the extension for a single function should
be added if and only if the function itself captures.
2022-12-27 09:15:21 -06:00
Ayaz Hafiz
877714ff91
We prefer slices 2022-12-25 16:56:00 -06:00
Ayaz Hafiz
a8693e6102
Order list-min-size tests in descending order
Some of the head-constructor tests we generate can be supersets of other tests.
Edges must be ordered so that more general tests always happen after their
specialized variants.

For example, patterns

  [1, ..] -> ...
  [2, 1, ..] -> ...

may generate the edges

  ListLen(>=1) -> <rest>
  ListLen(>=2) -> <rest>

but evaluated in exactly this order, the second edge is never reachable.
The necessary ordering is

  ListLen(>=2) -> <rest>
  ListLen(>=1) -> <rest>

Closes #4732
2022-12-25 16:55:58 -06:00
Ayaz Hafiz
a295e7ac3d
Allow any int in switch condition layout 2022-12-25 16:55:37 -06:00
Folkert de Vries
4022b44ff9
Merge pull request #4798 from JTeeuwissen/main
Replace DeclarationToIndex elements with a hash map.
2022-12-22 17:19:37 +01:00
J.Teeuwissen
1caafbc131
Feedback 2022-12-22 15:42:26 +01:00
J.Teeuwissen
5229edd742
Use hash instead of binary search 2022-12-22 14:08:41 +01:00
Folkert
4ef541395a
remove dbg lowlevel 2022-12-22 02:45:09 +01:00
Folkert
f76df8a356
make dbg transparent to refcounting 2022-12-22 01:03:03 +01:00
J Teeuwissen
dc65298f9d
Merge branch 'roc-lang:main' into main 2022-12-20 10:19:35 +01:00
Joshua Warner
174f7d5e4d
Fix bug in unifying records
This was leading us to have an infinitely-recursive type, which eventually causes layout to stack-overflow

Fixes #4739
2022-12-17 11:37:19 -08:00
Folkert de Vries
f550f049db
Merge pull request #4768 from roc-lang/expects-store-specialized-variable
Support using dbg/expect in polymorphic functions
2022-12-16 00:10:40 +01:00
Richard Feldman
e9e82ddc08
Merge pull request #4766 from roc-lang/remove-f128
remove F128
2022-12-15 17:59:48 -05:00
Ayaz Hafiz
4315a36e90
Refine imported ZAT/function compilation 2022-12-14 17:13:59 -06:00
Ayaz Hafiz
02eec1f60c
Simplify the model of handling imported/unloaded symbol specializations 2022-12-14 17:13:58 -06:00
Ayaz Hafiz
b8aa8df100
Make sure to resolve lambda sets for imported top-levels correctly
Part of #4733
2022-12-14 17:13:43 -06:00
Ayaz Hafiz
a96225e92e
Store specialized variable for use in dbg as fake symbol in lowlevel call 2022-12-14 17:00:37 -06:00
Ayaz Hafiz
fe90355265
Store specialized variables of expect lookups in expect frames 2022-12-14 17:00:36 -06:00
Ayaz Hafiz
16209ef866
Store specialized lookup variables in expects 2022-12-14 16:59:05 -06:00
Ayaz Hafiz
9d5f2aa138
Store specialized var of lookups in storage subs 2022-12-14 16:55:14 -06:00
Ayaz Hafiz
66c5e1292e
Thread expectations through in phase result structs
And provide expectation subs when expectations should be compiled. This
will be useful for getting the specialized type of an expect lookup when
it's available.
2022-12-14 16:55:14 -06:00
Folkert
d287eafa3a
remove F128 2022-12-14 23:28:38 +01:00
Folkert de Vries
6f2e14cf18
Merge pull request #4765 from roc-lang/expects-store-layout-in-lookaside
Remove layouts from the mono AST for expects, and add regtest for #4749
2022-12-14 23:22:03 +01:00
Ayaz Hafiz
a5ecded934
Don't collect layouts in mono of expects anymore 2022-12-14 14:12:57 -06:00
Ayaz Hafiz
1962f2045e
Remove layouts from the mono AST for expects
This actually isn't needed, because the backends must lookup the layout
from the environment anyway. So it's enough to lookup the symbol and
find its layout, there is no need to additionally store it.
2022-12-14 14:10:02 -06:00
Ayaz Hafiz
f2ffda6d05
Revert "Add a LayoutBuffer to store layouts aside and thread it through"
This reverts commit 3d73e33b49.
2022-12-14 14:05:56 -06:00
Ayaz Hafiz
897b69b072
Revert "Store layouts in a layout-buffer for expects"
This reverts commit bba6e36a18.
2022-12-14 14:05:46 -06:00
Ayaz Hafiz
bba6e36a18
Store layouts in a layout-buffer for expects 2022-12-14 14:05:28 -06:00
Ayaz Hafiz
3d73e33b49
Add a LayoutBuffer to store layouts aside and thread it through 2022-12-14 13:48:58 -06:00
J.Teeuwissen
e3dcaa2141
typo 2022-12-14 14:02:09 +01:00
Ayaz Hafiz
00ff6dcad7
Print similar-layout procs when borrow fails to find one 2022-12-12 14:13:07 -06:00
Folkert
a72556b927
refactor entry point 2022-12-09 15:10:25 +01:00
Ayaz Hafiz
759127660d
Spread list tests can never touch exact-sized bounds tests
When compiling a pattern match like

```
[] -> ..
[_] -> ..
[_, ..] -> ..
```

to a decision tree, we must make sure that the last test (len >= 1)
does not touch the branch reached by the second test (len == 1). It is
enough to ban (len >=) tests from ever touching exact-sized list
patterns, because a spread test (len >=) can never reach an exact-sized
test.

On the other hand, an exact-sized test can reach a spread pattern,
because in

```
[_, _] -> ..
[..] -> ..
```

the last branch generates tests for patterns `[]` and `[_]`, and we would
like those patterns to be covered by the spread test (len >= 0)!

Closes #4685
2022-12-05 13:45:35 -06:00
Ayaz Hafiz
d660016b53
Clippy 2022-12-03 23:43:47 -06:00
Ayaz Hafiz
01dea13b49
Pretty print symbols 2022-12-03 22:33:45 -06:00
Ayaz Hafiz
b0300e4e90
Fix bug in resolving layout 2022-12-03 22:18:15 -06:00
Ayaz Hafiz
7b21936d4f
Add mono checker to tests 2022-12-03 22:10:37 -06:00
Ayaz Hafiz
6e72307736
Format problems reported by the IR checker 2022-12-03 21:35:01 -06:00
Ayaz Hafiz
a84aebf2d3
Implement a type checker for the mono IR 2022-12-03 18:24:03 -06:00
Ayaz
b5a195eaf0
Merge pull request #4650 from roc-lang/detect-reachable-by-recursion-only
Provide warning for defs that are used only in (mutual) recursion
2022-12-02 13:39:52 -06:00
Ayaz Hafiz
5cfbc49ebf
Remove needless hole translation 2022-12-02 12:21:44 -06:00
Ayaz Hafiz
52c858dc53
Remove aliases in partial procs
This is no longer needed
2022-12-02 12:19:13 -06:00
Ayaz Hafiz
0a807dc43e
Provide warning for defs that are used only in (mutual) recursion
This patch provides errors for defs that are used only in
possibly-mutual recursion, and are not reachable outside of their
recursive closures. For example:

```
test_report!(
    mutual_recursion_not_reached_nested,
    indoc!(
        r#"
        app "test" provides [main] to "./platform"

        main =
            f = \{} -> if Bool.true then "" else g {}
            g = \{} -> if Bool.true then "" else f {}
            ""
        "#
    ),
@r###"
── DEFINITIONs ONLY USED IN RECURSION ──────────────────── /code/proj/Main.roc ─

These 2 definitions are only used in mutual recursion with themselves:

4│>      f = \{} -> if Bool.true then "" else g {}
5│>      g = \{} -> if Bool.true then "" else f {}

If you don't intend to use or export any of them, they should all be
removed!
"###
);
```
2022-12-02 11:50:12 -06:00
Ayaz Hafiz
42e8e409d4
Address lint 2022-12-01 15:59:01 -06:00
Ayaz Hafiz
9181ed8092
Correctly compile rvalue closures defined in nested defines to lvalues
Previously, a program like

```
main =
  f =
    n = 1
    \{} -[#lam]-> n  # suppose lambda set = #lam
  f {}
```

would be transformed to

```
main =
  n = 1
  f = \{} -[#lam]-> n
  f {}
```

However, the IR lowering procedure is such that we would then associate
`f` as definining the procedure given the lambda set `#lam`. This is not
correct, as `f` is really a function pointer in this circumstance,
rather than the definer of `#lam`.

Instead, the transformation we want to perform is

```
main =
  n = 1
  #lam = \{} -[#lam]-> n
  f = #lam
  f {}
```

Which is what this patch does

Closes #2403
2022-12-01 15:47:18 -06:00
Ayaz Hafiz
68e364d897
Do not attempt to handle aliasing of procs in variable assignments
Please see the comment in the diff to explain the rationale of this
change.

Closes #4636
2022-12-01 15:20:58 -06:00
Ayaz Hafiz
2dfe0276e4
Make sure to assign and update program holes correctly when specializing symbols
Previously this was a bit hacky, we worked around having to update
symbol substitutions by special-casing the rest of the program. There's
not need to do that.
2022-12-01 14:57:01 -06:00
Richard Feldman
58fad36f9d
Merge pull request #4460 from roc-lang/crash
Crash
2022-11-25 17:18:21 -05:00