Skip to content

Skip some PDBUsingTests tests to unblock CI#25753

Merged
jcouv merged 1 commit intodotnet:dev15.7.xfrom
jcouv:skip-pdbtests
Mar 27, 2018
Merged

Skip some PDBUsingTests tests to unblock CI#25753
jcouv merged 1 commit intodotnet:dev15.7.xfrom
jcouv:skip-pdbtests

Conversation

@jcouv
Copy link
Copy Markdown
Member

@jcouv jcouv commented Mar 27, 2018

Tagging @JieCarolHu @tmat @jaredpar

I don't know the root cause for this problem. But from thread, my understanding is that we'll skip those tests to unblock, and proceed with investigation of the issue to re-enable.

Bugs this fixes

Works around #25737

@jcouv jcouv added this to the 15.7 milestone Mar 27, 2018
@jcouv jcouv requested a review from a team as a code owner March 27, 2018 17:17
@tmat
Copy link
Copy Markdown
Member

tmat commented Mar 27, 2018

So far it looks like a bug in DiaSymReader.Native (64-bit vs 32-bit). All tests pass using a newer DSRN build locally.

@jcouv
Copy link
Copy Markdown
Member Author

jcouv commented Mar 27, 2018

Any idea how the bug got past CI and into the branch? I don't think those tests are new.

@jcouv jcouv added the Test Test failures in roslyn-CI label Mar 27, 2018
@jaredpar
Copy link
Copy Markdown
Member

No. We are pretty confused how this happen at this point.

@jaredpar
Copy link
Copy Markdown
Member

Failure looks to be random. See this run where the tests passed (other tests failude)

https://ci.dot.net/job/dotnet_roslyn/job/dev15.7.x/job/windows_debug_unit32_prtest/870/console

@tmat
Copy link
Copy Markdown
Member

tmat commented Mar 27, 2018

The bug only reproduces when running all (many) emit tests at once. It does not repro when running tests individually, not even when running the entire PDBUsingTests class. It might be some native heap corruption problem. It also only repros when running in a 32-bit process.

@jaredpar Any chance we changed bitness of the test runner recently?

@jaredpar
Copy link
Copy Markdown
Member

@tmat

Any chance we changed bitness of the test runner recently?

No. Nothing has changed in this area in some time.

@jcouv
Copy link
Copy Markdown
Member Author

jcouv commented Mar 27, 2018

@tmat Sounds like we have a proper fix. Would you rather upgrade the DSRN package version directly, or should I move forward with this mitigation PR?

@jaredpar
Copy link
Copy Markdown
Member

Need to take this PR. This branch is on the floor due to the massive flakiness in tests. Not just PDB. PeekTests also causing major issues. If there is a fix later the skip can be reverted.

@tmat
Copy link
Copy Markdown
Member

tmat commented Mar 27, 2018

@jcouv I do not have an official build yet. Proceed with this PR please.

@JieCarolHu JieCarolHu self-requested a review March 27, 2018 17:49
@jcouv jcouv merged commit 4b3405e into dotnet:dev15.7.x Mar 27, 2018
@jcouv jcouv deleted the skip-pdbtests branch March 27, 2018 18:08
333fred added a commit to 333fred/roslyn that referenced this pull request Mar 28, 2018
…features/dataflow

* dotnet/dev15.7.x:
  Fix inference when tuple names differ (dotnet#25738)
  Avoid crash with 'new ref[]' (dotnet#25505)
  Skip some PDBUsingTests tests to unblock CI (dotnet#25753)
  Disable Spanish queue
  Tuple equality: Optimize cases where nullable always has value (dotnet#25644)
  Bump prerelease version to beta4
  Fix failure to initialize RemoteWorkspace in the out-of-process host
  remove whitespace
  Removing package dependency for Microsoft.VisualStudio.Text.UI.Wpf in EditorFeatures
  Removing package dependency for Microsoft.VisualStudio.Text.UI.Wpf in CSharpEditorFeatures
jaredpar pushed a commit to jaredpar/roslyn that referenced this pull request Mar 28, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Area-Compilers Test Test failures in roslyn-CI

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants