You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
On a buildbot, a job failed because a test worker stdout cannot be decoded from UTF-8:
Cannot read process stdout: 'utf-8' codec can't decode byte 0xdd in position 2360: invalid continuation byte
In February, @gpshead reported a similar issue: issue #101634. I fixed it by adding a try/except: commit 2ac3eec.
The try/except stops regrtest in a more clean fashion, but it has a big issue: the worker stdout is lost and cannot be displayed.
IMO regrtest should use any mean to display the stdout, especially if it's corrupted. It's too important, since some bugs only occur on some platforms. For example, Windows can log assertion errors in UTF-16 whereas regrtest uses a 8-bit encoding (issue #108989 closed as "WONTFIX").
libregrtest now decodes stdout of test worker processes with the
"backslashreplace" error handler to log corrupted stdout, instead of
failing with an error and not logging the stdout.
libregrtest now decodes stdout of test worker processes with the
"backslashreplace" error handler to log corrupted stdout, instead of
failing with an error and not logging the stdout.
vstinner
added a commit
to vstinner/cpython
that referenced
this issue
Sep 14, 2023
libregrtest now decodes stdout of test worker processes with the
"backslashreplace" error handler to log corrupted stdout, instead of
failing with an error and not logging the stdout.
)
libregrtest now decodes stdout of test worker processes with the
"backslashreplace" error handler to log corrupted stdout, instead of
failing with an error and not logging the stdout.
On a buildbot, a job failed because a test worker stdout cannot be decoded from UTF-8:
In February, @gpshead reported a similar issue: issue #101634. I fixed it by adding a try/except: commit 2ac3eec.
The try/except stops regrtest in a more clean fashion, but it has a big issue: the worker stdout is lost and cannot be displayed.
IMO regrtest should use any mean to display the stdout, especially if it's corrupted. It's too important, since some bugs only occur on some platforms. For example, Windows can log assertion errors in UTF-16 whereas regrtest uses a 8-bit encoding (issue #108989 closed as "WONTFIX").
build: https://buildbot.python.org/all/#/builders/435/builds/3612
Linked PRs
The text was updated successfully, but these errors were encountered: