Stream: t-compiler

Topic: #66983


weiznich (Dec 06 2019 at 13:27, on Zulip):

I'm looking for some feedback on #66983, especially how to handle the failing test case there.

oli (Dec 06 2019 at 13:31, on Zulip):

I think you need to delete test/ui/lint/issue-47775-nested-macro-unnecessary-parens-arg.stderr, we should really fix that --bless bug -.-

weiznich (Dec 06 2019 at 13:54, on Zulip):

My question is more: Is it OK to remove that error there?

oli (Dec 06 2019 at 13:55, on Zulip):

yea, isn't this the whole point of your PR? stop showing the diagnostic when the parens have not been written by the user

weiznich (Dec 06 2019 at 14:05, on Zulip):

It's just that this is an existing test that I've not written and where I don't know how important it is.

weiznich (Dec 06 2019 at 14:09, on Zulip):

Anyway I've pushed a fixed version.

oli (Dec 06 2019 at 14:10, on Zulip):

:+1:

pnkfelix (Dec 06 2019 at 14:19, on Zulip):

My question is more: Is it OK to remove that error there?

Its a reasonable point: after all, the comments in that test case make it sound like it was certainly intended that we would catch the "unused parens" in the macro definition.

pnkfelix (Dec 06 2019 at 14:20, on Zulip):

// ↓ But let's test that we still lint for unused parens around
// function args inside of simple, one-deep macros.

... but this also strikes me as a silly thing to make a hard requirement of this lint's specification.

pnkfelix (Dec 06 2019 at 14:22, on Zulip):

cc @Zack M. Davis , who wrote the original code (for this test, in PR #47896)

weiznich (Dec 10 2019 at 09:02, on Zulip):

What's the right way to move forward on that PR?

Last update: Jan 21 2020 at 08:25UTC