Don't explain erroneous bounds #19338
Conversation
som-snytt
force-pushed
the
issue/19334-OOM-erroring
branch
from
174800f to
4c2d5d8
Compare
| case skolem: SkolemType => true | ||
| case sym: Symbol => | ||
| case skolem: SkolemType => true | ||
| case sym: Symbol => |
There was a problem hiding this comment.
| case sym: Symbol => | |
| case sym: Symbol => |
There is nothing on the right of => to align.
There was a problem hiding this comment.
Thanks, I go back & forth on this style. The question is, in mix of one-liners, is the odd arrow also aligned?
Also default case: case _ => ??? where it's always ugly to align.
I'm not a huge fan of formatters, but for this use case, it would restore my decision budget.
|
|
||
| @main def main = foo: | ||
| def f() = () | ||
| f(_) // error was OOM formatting TypeVar(TypeParamRef(T)) when offering explanations |
There was a problem hiding this comment.
it might be worth adding a tests/neg/i19334.check file for this one.
There was a problem hiding this comment.
footnote I had one locally. I'm unaccustomed to the dotty-local customs around check files.
shell history
touch tests/neg/i19334.check
git add tests/neg/i19334.*
git rm tests/neg/i19334.check
I need shell history with comments, to tell me why I decided to do that.
| case skolem: SkolemType => true | ||
| case sym: Symbol => | ||
| case skolem: SkolemType => true | ||
| case sym: Symbol => | ||
| ctx.gadt.contains(sym) && ctx.gadt.fullBounds(sym) != TypeBounds.empty |
There was a problem hiding this comment.
This equality check looks like the previous tweak to use =:= for this comparison. I have not yet investigated / learned about it.
|
@som-snytt could you squash the commits before we merge? |
som-snytt
force-pushed
the
issue/19334-OOM-erroring
branch
from
14e0f90 to
23b3d7f
Compare
|
Thanks, squashed. |
som-snytt
force-pushed
the
issue/19334-OOM-erroring
branch
from
23b3d7f to
a227ce2
Compare
Fixes #19334