[mir-inlining] Don't inline virtual calls #55046
Conversation
|
r? @varkor (rust_highfive has picked a reviewer for you, use r? to override) |
rust-highfive
added
the
S-waiting-on-review
src/librustc_mir/transform/inline.rs
Outdated
| let is_virtual = | ||
| if let InstanceDef::Virtual(..) = instance.def { | ||
| true | ||
| } else { false }; |
There was a problem hiding this comment.
Nit: false should be on its own line.
| @@ -0,0 +1,41 @@ | |||
| // Copyright 2018 The Rust Project Developers. See the COPYRIGHT | |||
| println!("{}", test(&())); | ||
| } | ||
|
|
||
| fn test(x: &X) -> u32 { |
|
Was it inlining the default function implementation, or the implementation for |
|
It was inlining the default implementation so prior to the patch, the test would output |
wesleywiser
force-pushed
the
no_virtual_call_inlining
branch
from
bc47afd to
6c2dace
Compare
|
Feedback resolved. Thanks @varkor! |
Prior to this change, the test case would output `1` instead of `2` like it should.
wesleywiser
force-pushed
the
no_virtual_call_inlining
branch
from
6c2dace to
69eaa11
Compare
|
@bors r=varkor |
|
📌 Commit 69eaa11 has been approved by |
bors
added
S-waiting-on-bors
[mir-inlining] Don't inline virtual calls
|
☀️ Test successful - status-appveyor, status-travis |
|
Just, something strange is going on, and while the tests pass locally they fail on CI?!? See #55086. |
|
@RalfJung Is there anything that needs to be done related to this bug fix? I see you closed the other pull request. |
|
Yeah, turns out I was wrong when I thought that the ICE got fixed. :( Sorry for the noise. |
|
No worries :) |
No description provided.