This is an archive of the discontinued LLVM Phabricator instance.

[llvm-readobj] Update tests because of changes at llvm-readobj behavior
ClosedPublic

Authored by Elvina on Jul 15 2020, 3:00 PM.

Details

Summary

This patch updates tests for llvm-readobj and llvm-readelf, because now reading from stdin can be achieved only via a '-' as described here https://bugs.llvm.org/show_bug.cgi?id=46400. Patch with changes of llvm-readobj behavior is here: https://reviews.llvm.org/D83704

Yours,
Elvina
Software Engineer
Advanced Software Technology Lab, Huawei

Diff Detail

Event Timeline

Elvina created this revision.Jul 15 2020, 3:00 PM
Elvina edited the summary of this revision. (Show Details)Jul 15 2020, 3:11 PM
jhenderson accepted this revision.Jul 16 2020, 12:50 AM

Thanks for putting in the effort for this! I didn't realise there were quite so many tests that needed updating! LGTM.

This revision is now accepted and ready to land.Jul 16 2020, 12:50 AM
MaskRay accepted this revision.Jul 16 2020, 8:14 AM

Thanks!

grimar added a subscriber: grimar.Jul 17 2020, 1:48 AM

I've looked at LLD tests and found no tests that needs to be updated there. So this LG for me too.

lenary removed a subscriber: lenary.Jul 17 2020, 3:21 AM
Elvina updated this revision to Diff 278836.Jul 17 2020, 10:27 AM

Sync with master

Thanks to everyone for looking at this!

This revision was automatically updated to reflect the committed changes.
Herald added a project: Restricted Project. · View Herald TranscriptJul 20 2020, 2:40 AM
Herald added a subscriber: cfe-commits. · View Herald Transcript

Hi @Elvina,

Just to let you know that I had to fix up three clang tests that were using the old behaviour too, prior to committing. I also noticed that you've got the stack the wrong way around - you needed to fix the tests before changing the behaviour in this case to avoid breaking the codebase, even if it was only briefly.

Anyway, these are now committed. I'll update the bug.

Hi @Elvina,

Just to let you know that I had to fix up three clang tests that were using the old behaviour too, prior to committing. I also noticed that you've got the stack the wrong way around - you needed to fix the tests before changing the behaviour in this case to avoid breaking the codebase, even if it was only briefly.

Anyway, these are now committed. I'll update the bug.

Hi James,
Thank you for your help and suggestions! I really appreciate this. I'll be more attentive next time :)