Travis matrix adjustments to avoid duplication and frequent timeouts#11468
Travis matrix adjustments to avoid duplication and frequent timeouts#11468t8m wants to merge 1 commit intoopenssl:masterfrom
Conversation
dabfb5f to
ac60c4b
Compare
- do not exclude all clang builds on Linux - exclude the constantly timeouting -fsanitize=address build on OS/X - drop some mostly duplicate builds - change the base linux distro to Bionic - drop sudo as that is no longer needed - always on - drop -D__NO_STRING_INLINES where not needed - memleak test is not working with old clang
|
As you can see here https://travis-ci.com/github/t8m/openssl?utm_medium=notification&utm_source=email the only build not passing within the extended tests is the one with external krb5 test. That is no regression. |
I thought we had the krb5 test working fairly recently, to while it may not be a local regression I am still interested to know how it came to be. I will go kick off a local build with the krb5 tests, at least... |
The last commit where krb5 test was passing is: f33ca11 However there were a few (rather innocently looking) commits in between where Travis did not run. |
|
Actually the commits are not that innocent I suppose - most probably the krb5 build is broken due to some mix-up of system and local openssl after the change from shlib_wrap.sh to wrap.pl. |
|
Thanks! That does feel like a likely culprit, given what things look like in my local testing. |
|
24 hours has passed since 'approval: done' was set, but as this PR has been updated in that time the label 'approval: ready to merge' is not being automatically set. Please review the updates and set the label manually. |
- do not exclude all clang builds on Linux - exclude the constantly timeouting -fsanitize=address build on OS/X - drop some mostly duplicate builds - change the base linux distro to Bionic - drop sudo as that is no longer needed - always on - drop -D__NO_STRING_INLINES where not needed - memleak test is not working with old clang Reviewed-by: Richard Levitte <[email protected]> (Merged from #11468)
|
Merged to master as fbc6efb |
The intention is to re-enable some clang build on linux and replace the constantly timeouting osx build with it.