> On 9 Jun 2016, at 10:06 AM, Doug Hellmann <d...@doughellmann.com> wrote: > > Excerpts from Sam Morrison's message of 2016-06-09 08:16:11 +1000: >> >>> On 8 Jun 2016, at 10:51 PM, Doug Hellmann <d...@doughellmann.com> wrote: >>> >>> Excerpts from Sam Morrison's message of 2016-06-08 16:59:37 +1000: >>>> We have been building our own packages for a few years now and I’ve just >>>> started to build packages for liberty. >>>> >>>> I got an error from pbr telling me my version was incompatible. An example >>>> version we were producing was 12.0.3+a74~gea340da+trusty >>>> >>>> So had a read of http://docs.openstack.org/developer/pbr/semver.html >>>> <http://docs.openstack.org/developer/pbr/semver.html> and have fixed it to >>>> be this format however I’m still getting an error: >>>> >>>> Exception occurred: >>>> File "/usr/lib/python2.7/dist-packages/pbr/version.py", line 215, in >>>> from_pip_string >>>> % (remainder, version_string)) >>>> ValueError: Unknown remainder ['gea340da'] in '12.0.3.a74.gea340da' >>>> >>>> From the website mentioned above this seems like it should be valid. I’ve >>>> had a look in the code and it looks like this type of version isn’t >>>> supported. >>>> >>>> So is this a bug in pbr or am I doing something wrong? We build packages >>>> based on git versions so I want to keep the hash in there. >>>> >>>> Thanks, >>>> Sam >>> >>> This is working as designed. pbr follows PEP-440 [1] versioning, >>> which IIRC doesn't include version control references like git SHAs. >>> >>> The version you describe appears to be the 74th alpha pre-release of >>> 12.0.3. Is that what you intend? That's a lot of pre-releases. >>> >>> pbr will automatically append a numbered dev suffix to versions, >>> so the 74th commit after 12.0.3 would be something like 12.0.4.dev74, >>> indicating that it is a development version leading to 12.0.4. >>> >>> Which project are you packaging, and in what context? >> >> Thanks Doug, >> >> So is that openstack document wrong? > >> I'm packaging nova, we have a bunch of custom commits and extra back ports >> in our version. Having the git hash in there is extremely helpful and I also >> don't want to conflict with the official openstack release. >> I don't care really what the version number is as long as when I add a new >> commit the version number increases. >> >> Having a look at that document I think I should do something like >> >> 12.0.3.post74+gea340da > > According to the comments in pbr/version.py it looks like you want > > 12.0.3.74.gea340da
OK great, yes that works, although that looks like it is invalid according to PEP440? From my understanding of that PEP I should do 12.0.3.post74+gea340da no? > I've added a patch with a test that shows that works, as well as a fix > for the syntax including the "post" in https://review.openstack.org/327424 > > If you're doing your own packaging, how are you feeding the version > number into pbr? Via the environment variable? > We’re building it based on the output of `git describe` We use GIT_DESCRIBE_VERSION_REGEX = re.compile( r""" ^v?(?P<major>\d+)\. (?P<minor>\d+) (?:\.(?P<patch>\d+(\.\d+){0,1})){0,1} (?:-(?P<commits>\d+) -g(?P<revision>[0-9a-f]+)){0,1}$""", re.VERBOSE) We should also update http://docs.openstack.org/developer/pbr/semver.html as that isn’t correct. Thanks for your help. Cheer, Sam > Doug > >> >> Any ideas? >> >> Thanks, >> Sam >> >>> >>> Doug >>> >>> [1] https://www.python.org/dev/peps/pep-0440/ >>> >>> _______________________________________________ >>> Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack >>> Post to : openstack@lists.openstack.org >>> Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack _______________________________________________ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
