Comment 9 for bug 2020641

Revision history for this message
Andreas Hasenack (ahasenack) wrote (last edit ):

Please include a "normal" test for gdm3 in the test case. Something like testing that logins keep working after applying this update, a switch between "init 3" and "init 5", and that user switching works (while logged in as user A, with open apps, login as user B, launch an app, and come back to user A and you still have your stuff running).

And if the upstream MP changes significantly, please update the patch here as well.

Please take the following into consideration for the next time. It's not blocking this upload, but helps with the SRU processing and allows us to process more SRUs per day:
- use the same patch filename in all affected releases: it speeds up considerably the review of the diff, as git can show they are identical, or mostly identical. But with different filenames, and content (see below), not even git can do it
- try to use the same patch content-wise, where possible of course (most of the time the affected versions will be different). But the DEP3 headers at least can be very similar, if not identical
- complete DEP3 headers: one patch has an ubuntu bug reference but no upstream reference, while the other one has an upstream reference, but no ubuntu bug. And the code change is the same, but with a different diff.
- avoid including unrelated changes in the upload, even if they are no-ops. For example, the lunar upload is changing the Uploaders field in d/control. It makes us wonder if that was on purpose, if it was a mistake, perhaps we ask the uploader if they want to keep it or reupload, etc. Unecessary back-and-forth.