Merge lp:~cjwatson/isitdeployable-dependencies/pygit2-0.24.2 into lp:isitdeployable-dependencies

Proposed by Colin Watson on 2018-01-30
Status: Merged
Approved by: Colin Watson on 2018-01-30
Approved revision: 16
Merge reported by: Otto Co-Pilot
Merged at revision: not available
Proposed branch: lp:~cjwatson/isitdeployable-dependencies/pygit2-0.24.2
Merge into: lp:isitdeployable-dependencies
Diff against target: 6 lines (+0/-0)
0 files modified
To merge this branch: bzr merge lp:~cjwatson/isitdeployable-dependencies/pygit2-0.24.2
Reviewer Review Type Date Requested Status
Daniel Manrique 2018-01-30 Approve on 2018-01-30
Review via email: mp+336856@code.launchpad.net

Commit message

Add pygit2 0.24.2 and dependencies.

Description of the change

This is to prepare for installing pygit2 with pip rather than using the .deb. I went for the newest in the 0.24 series rather than 0.24.1 which is currently in our dependencies PPA for trusty, but the changes are basically trivial (https://github.com/libgit2/pygit2/blob/master/CHANGELOG.rst).

To post a comment you must log in.
Daniel Manrique (roadmr) wrote :

There was some concern about the *mu* wheels, which appear to have been built with a newer pip than the one we have in Trusty (and we must use the system pip rather than updating it because otherwise we get in trouble with system-installed packages, which upstream pip doesn't know about, while Ubuntu pip does). The SOP is to build wheels from source, when needed, on a more or less pristine Trusty host (lxc, usually) with non-upgraded pip and wheel.

That said, part of the work Colin is doing is upgrading the stack to Xenial, which has a newer pip which should handle those *mu* wheels properly. In any case, Colin says this ran fine on a build and test of the docker image, which is what we're using these days for this service. So it seems safe to approve these, since 1) they do seem to work well when building a new asset, and 2) if said new asset barfs, k8s should allow us to roll back easily, or maybe even protect us from deploying a bad asset entirely.

I think it's worth merging these and seeing what docker/k8s say.

review: Approve

Preview Diff

[H/L] Next/Prev Comment, [J/K] Next/Prev File, [N/P] Next/Prev Hunk
1=== added file 'cffi-1.11.4-cp27-cp27mu-manylinux1_x86_64.whl'
2Binary files cffi-1.11.4-cp27-cp27mu-manylinux1_x86_64.whl 1970-01-01 00:00:00 +0000 and cffi-1.11.4-cp27-cp27mu-manylinux1_x86_64.whl 2018-01-30 15:42:36 +0000 differ
3=== added file 'pycparser-2.18-py2.py3-none-any.whl'
4Binary files pycparser-2.18-py2.py3-none-any.whl 1970-01-01 00:00:00 +0000 and pycparser-2.18-py2.py3-none-any.whl 2018-01-30 15:42:36 +0000 differ
5=== added file 'pygit2-0.24.2-cp27-cp27mu-linux_x86_64.whl'
6Binary files pygit2-0.24.2-cp27-cp27mu-linux_x86_64.whl 1970-01-01 00:00:00 +0000 and pygit2-0.24.2-cp27-cp27mu-linux_x86_64.whl 2018-01-30 15:42:36 +0000 differ

Subscribers

People subscribed via source and target branches