smoke-test failures: Win32 #258
Comments
jkeenan
added
the
blocker-to-strict-by-default
[snip]
[snip] My very preliminary analysis of the failure output from the Win32 smoke-test run suggests that some of the failures are for reasons similar to what we've encountered in our development on *nix.
Some diagnostic hunts that may be useful. |
|
Poke at it with a few minor fixes |
|
This was my result on a interactive run: |
|
note that |
What do you mean by an "interactive" run?
If you can post verbose output for any of those, I can stare at them overnight. Thanks for your investigation. |
Results look good. See attachment. The bottom report is what I believe is the most recent smoke-test run on Perl 5 blead run on this OS/OS version/compiler/compiler version combination. The middle report is the next-to-most-recent smoke-test run on the simulation of Perl 7 branch run on this host. cpan/ExtUtils-MakeMaker/t/02-xsdynamic.t and dist/Net-Ping/t/450_service.t are frequently observed failures on this host, but all the other are attributable to our previous lack of testing on Win32. The top report is the most recent smoke-test report on the Perl 7 simulation branch. Note that now the only files failing are ones also always or intermittently failing on Perl 5 blead. We, like Perl 5, are handicapped by the fact that this host is the only one steadily submitting smoke-test reports on MSWin32. Thank you very much. |
|
That's not really true. All of these actions test windows. https://github.com/perl/perl5/actions?query=workflow%3Asmoke-windows-mingw64 https://github.com/perl/perl5/actions?query=workflow%3Asmoke-windows-msvc100-x86 https://github.com/perl/perl5/actions?query=workflow%3Asmoke-windows-msvc142-amd64 |
|
On 9/4/20 7:07 PM, Todd Rinaldo wrote:
That's not really true. All of these actions test windows.
https://github.com/perl/perl5/actions?query=workflow%3Asmoke-windows-mingw64
https://github.com/perl/perl5/actions?query=workflow%3Asmoke-windows-msvc100-x86
https://github.com/perl/perl5/actions?query=workflow%3Asmoke-windows-msvc142-amd64
Todd, at all 3 of those URLs the last report was two months ago. So
there don't appear to be any reports on our simulation branch:
https://github.com/perl/perl5/actions?query=workflow%3Asmoke-windows-mingw64+branch%3Asmoke-me%2Fjkeenan%2Fcumberland-blues
|
|
@jkeenan can we close this case or are we viewing more failures on windows? |
|
On 9/14/20 5:03 PM, James E Keenan wrote:
@jkeenan <https://github.com/jkeenan> can we close this case or are
we viewing more failures on windows?
@atoomic <https://github.com/atoomic> I suspect we aren't going to get
any more smoke-test reports for this particular Merge Candidate tag. I'd
like to do that after we resolve #267
<#267>.
An additional consideration here: kl
|
|
On 9/14/20 5:03 PM, James E Keenan wrote:
@jkeenan <https://github.com/jkeenan> can we close this case or are
we viewing more failures on windows?
@atoomic <https://github.com/atoomic> I suspect we aren't going to get
any more smoke-test reports for this particular Merge Candidate tag. I'd
like to do that after we resolve #267
<#267>.
An additional consideration: https://perl5.test-smoke.org has been down
since approx Sep 06 -- so our ability to get reports is hampered.
jimk
|
We pay for our failure/inability to test on Windows so far.
Smoke test report: http://perl5.test-smoke.org/report/117911
Log file: http://m-l.org/~perl/smoke/perl/mswin32-win2000/smoke-me/James%20E%20Keenan/log005643c4ee4b5d0cd729c01372d3b3fc59752d3a.log.gz
A gzipped file containing the relevant part of the log for the non-threaded, non-debugging configuration can be found here.
@atoomic, @toddr please review.
Thank you very much.
Jim Keenan
The text was updated successfully, but these errors were encountered: