Fix for issue #80350 tracking runtime deficiencies in the presence of static virtual method reabstraction#80987
Conversation
|
/azp run coreclr-runtime outerloop |
|
No pipelines are associated with this pull request. |
|
/azp run runtime-coreclr outerloop |
|
Azure Pipelines successfully started running 1 pipeline(s). |
|
@SamMonoRT - could you please advise someone who might be able to help me fix the Mono issues in this PR? I have updated the |
@BrzVlad will take a look at the failing Mono CI lanes later this week. |
This change fixes the issue dotnet#80350 that uncovered several inconsistencies in the CoreCLR runtime w.r.t. static virtual methods. In particular, reabstraction (declaring an abstract explicit override of a static virtual method in a derived interface) turned out to be causing runtime issues. As part of the change I'm also fixing the reabstraction unit test svm_diamondshape as it turns out Roslyn emits the method flags in a slightly different manner than I originally expected in the hand-written IL test. Thanks Tomas
Based on offline discussion with Vlad Brezae I have temporarily disabled the svm_diamondshape test on Mono where it is causing crashes in VTable construction that might take a bit to fix. Thanks Tomas
|
Based on offline discussion with @BrzVlad I have updated this PR by adding a temporary exclusion of the Thanks Tomas |
This change fixes the issue
#80350
that uncovered several inconsistencies in the CoreCLR runtime
w.r.t. static virtual methods. In particular, reabstraction (declaring
an abstract explicit override of a static virtual method in a derived
interface) turned out to be causing runtime issues. As part of
the change I'm also fixing the reabstraction unit test
svm_diamondshape as it turns out Roslyn emits the method flags
in a slightly different manner than I originally expected in the
hand-written IL test.
Thanks
Tomas
P.S. This PR would be normally an obvious adept for nominating @davidwrighton as the reviewer. As David is on parental leave right now, I am designating a larger reviewer group to get more pairs of eyes to see the change and also ideally proliferate CoreCLR typesystem knowledge to improve our agility the next time David's not available.