Fix ICE 6840 - make is_normalizable more strict #6866
Conversation
|
r? @flip1995 (rust-highfive has picked a reviewer for you, use r? to override) |
rust-highfive
added
the
S-waiting-on-review
| is_normalizable_helper(cx, param_env, ty, &mut HashMap::new()) | ||
| } | ||
|
|
||
| fn is_normalizable_helper<'tcx>( |
There was a problem hiding this comment.
One last thing: This function should not have to exist, if the normalize function in rustc would work as intended. Can you add a FIXME linking to the normalize documentation, explaining that this helper can be removed in the future, once normalize doesn't panic anymore?
|
@bors r+ Thanks! |
|
📌 Commit 9707599 has been approved by |
Fix ICE 6840 - make is_normalizable more strict fixes #6840 make `is_normalizable` more strict, which should catch this ICE and related cases changelog:
|
💔 Test failed - checks-action_test |
|
@bors retry |
|
☀️ Test successful - checks-action_dev_test, checks-action_remark_test, checks-action_test |
|
Hey folks! Thanks for the fix. Did this miss the boat for |
|
This unfortunately did miss 1.51. Looks like this fix is available in the current Rust beta release (1.52) |
|
Yeah, it missed it. To catch such things earlier and backport them, we're heavily relying on Clippy nightly/beta users, reporting that they want to have those fixes backported, before the new beta is branched (~1 week before release). |
|
Ok, thanks: we'll try to consume beta to get this.
It doesn't look like it's in the latest beta toolchain either, afaict: |
|
The new beta wasn't released yet. It should be in beta tomorrow. (see the version number of beta, it still says |
fixes #6840
make
is_normalizablemore strict, which should catch this ICE and related caseschangelog: Fix ICE in [
zero_sized_map_values]