Am 2017-12-16 14:49, schrieb Andreas Jaeger:
Looking at failures, I see http://logs.openstack.org/periodic/git.openstack.org/openstack/networking-bg...
Problem is: + git status On branch proposals Changes not staged for commit: (use "git add <file>..." to update what will be committed) (use "git checkout -- <file>..." to discard changes in working directory)
modified: networking_bgpvpn_heat/locale/en_GB/LC_MESSAGES/networking_bgpvpn_heat.po
and then git review fails because of modified local files.
I'm puzzled why the logic is not working with this repository.
Could anybody else dig into this, please?
Okay, my first idea was the function compress_po_files which is called after the directory was added to git which caused in a content change [1][2]. Then I observed double brackets in line 151 propose_translation_update.sh but that doesn't matter, the function was already called there. My next finding: zanata-cli pull was called a second time in [3] after horizon installation to build releasenotes. This changed the file in networking_bgpvpn_heat/locale/ again, because only releasenotes added to git in function propose_releasenotes (first run was for python module and second run for django). Anyway, I proposed a fix to prevent the duplicate pull_from_zanata on the same repo [4]. Maybe it's easier to implement a new function without python/django switches to pull from Zanata one time. Frank [1] http://logs.openstack.org/periodic/git.openstack.org/openstack/networking-bg... [2] http://logs.openstack.org/periodic/git.openstack.org/openstack/networking-bg... [3] http://logs.openstack.org/periodic/git.openstack.org/openstack/networking-bg... [4] https://review.openstack.org/#/c/528518/