This bumps the upper version of flake8 to the latest minor version,
which also introduces a new version of pycodestyle.
Change-Id: I40967a4905490c0ad43839b4d225ae92ff2a9e4a
Signed-off-by: Stephen Finucane <sfinucan@redhat.com>
This resolves an issue introduced in 2.1.0, which will be blacklisted,
whereby Sphinx's 'latex_engine' config opt was no longer configured,
breaking various projects.
Change-Id: I6b3467a4cba99bba939f983493a9df83f0093680
Signed-off-by: Stephen Finucane <sfinucan@redhat.com>
Handles redirection in the link checking for artifacts.
Also drops looking for py2 wheels since we now only have py3.
Change-Id: I34b201902faa60795dae44ebf059af0f2f9242c6
Signed-off-by: Sean McGinnis <sean.mcginnis@gmail.com>
Check that the new jobs that mirror on tags work correctly, before
pushing the big red button.
Depends-On: https://review.opendev.org/727665
Change-Id: I8b9be10e7a090bdf87b205cacae76f8d62621800
This includes a fix for pinning of flake8 to avoid bringing in a new
version with new checks.
Change-Id: I6e498aeef82fe3cabe45ac73e764c5a7cbbacec6
Signed-off-by: Stephen Finucane <sfinucan@redhat.com>
These are the final release taggings for all cycle-with-rc projects.
While not required, having PTLs ack this patch would be nice to include
that record in the merged metadata.
Release activity should now be frozen until this is processed on May 13.
Any bugfixes merged to stable/ussuri can be queued up and ready to go as
a follow on stable release after the coordinated release date.
Since it always comes up - yes, the diff-start is right. That's the
calculated starting point where the last stable branch was created.
Change-Id: I66602ca7a6ab1af35bf33d4d82fe34191f2e3865
Signed-off-by: Sean McGinnis <sean.mcginnis@gmail.com>
In order to tell pbr not to build release notes and create CHANGELOG and
AUTHORS files, we need to pass flags in as environment variables. We
can't just pass these flags though, as that will exclude the rest of the
normal environment set in. So copy the current environment, append our
updates, and pass that in explicitly to the subprocess call.
Change-Id: Ia47e73b51b3216572d51399338e95bd4269a8d1d
Signed-off-by: Sean McGinnis <sean.mcginnis@gmail.com>
We have a couple validations that need to build an sdist now. But once
it has been built in our temporary pristine clone, we don't need to
build it again. Because we create a clean clone, we also don't need to
worry about any stale dist builds. This adds a check to the sdist build
logic to recognize if an sdist has already been built and skip it.
There are also a few extra steps in the build process that are not
needed for our purposes. The biggest is the new(ish) pbr behavior of
building all release notes in the git tree if reno is present. This adds
an environment variable flag to tell pbr not to do that. While we're at
it, we don't need AUTHORS or CHANGELOG created either, so set the flags
telling pbr to skip those as well.
Change-Id: I15be9a12ff6684729980c40afae9625a8b8e550a
Signed-off-by: Sean McGinnis <sean.mcginnis@gmail.com>
During validation we need to build a bdist_wheel which is only available
as an extension when wheel is present in the environment. This adds
wheel and makes sure all requirements are installed for the validation
job.
Change-Id: Ied3b1ff4accf66e358cec4b209f4d3a65ae88d26
Signed-off-by: Sean McGinnis <sean.mcginnis@gmail.com>
This can take a long time now with doing the final release tag
validation when all deliverables are included in one patch. Part of the
delay appears to be that pbr was updated to build release notes if reno
is present. We may look for other ways to optimize the validation jobs
to speed things up, but for now we can just make sure we have a very
large window.
Change-Id: I248968bca9a90a7b06d4d6dd0644c50d926e5f45
Signed-off-by: Sean McGinnis <sean.mcginnis@gmail.com>
This updates the script to not require a specific topic to make it
easier to automate a job that periodically makes sure all open reviews
have had the PTL and liaison added.
Also refactored the adding of teams because bash array handling can be a
bit wonky sometimes and not always append values. This makes sure no
teams are skipped.
This is slightly less optimized for large reviews like the final series
release since it loops through teams based on the deliverable files. As
a slight optimization, it will skip a team if it was the same team as
the last deliverable file. It just might be possible that the same team
will be processed multiple times if their deliverable files are not in
alphabetical order, but there is no harm in that other than taking
slightly longer to complete.
Change-Id: I9a9194ea276aa0120aef21a321f48a93419fb848
Signed-off-by: Sean McGinnis <sean.mcginnis@gmail.com>
