- 02 Jun, 2020 6 commits
-
-
Cyril 'kibi' Brulebois authored
It's better to have sudo ask for a password upfront than letting the wrapper start children which might lead to entangled sudo prompt and log lines.
-
Cyril 'kibi' Brulebois authored
Wrong indentation was making the closing ')' hang below the code block.
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
Needed after the last few successive merges. :)
-
intrigeri authored
-
- 01 Jun, 2020 3 commits
-
-
Cyril 'kibi' Brulebois authored
A recent example is the switch to GitLab: one might have forgotten to update the “Build” step in the parallel_build_IUKs job to point at the new remote.
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
v2: Use names agreed on in #17658
-
- 24 May, 2020 1 commit
-
-
intrigeri authored
-
- 06 May, 2020 5 commits
-
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
Apparently, clocking release management work and manual testing separately is no longer desired.
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
Let's show some progress.
-
- 05 May, 2020 1 commit
-
-
Cyril 'kibi' Brulebois authored
There are 2+ GB of images to push through git-annex, which is a huge bottleneck at the moment. To compensate for that, start the git annex copy (almost) as soon as the image reproducibility has been verified. Keep that step after the Torrent file creation step though, as the latter already moves images (+ their signatures) around a little, paving the way for the git-annex dance for those who might have $ISOS pointing to an ISO history checkout. While at it, include some suggested commands to speed things up. Once the transfer is started, the release manager can start building IUKs locally (which might need some adjusting if parallel builds are desired, until #17657 is acted on).
-
- 01 May, 2020 1 commit
-
-
Cyril 'kibi' Brulebois authored
-
- 27 Apr, 2020 23 commits
-
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
It might get overlooked otherwise, and closing it with the right target version after a roadmap entry is closed… is a pain. Hopefully things will be better with GitLab anyway.
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
Trying to remove PREVIOUS_VERSION from the stable channel after having published a final major release wouldn't work too well.
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
As pointed out by Stephanie.
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
One needs to be in a checkout anyway to call ./bin/idf-content so let's get back to that directory explicitly.
-
Cyril 'kibi' Brulebois authored
There are inconsistencies regarding the directory changes, and subshells are only sometimes used. Get rid of some of them. [ Commit best viewed ignoring space-related changes. ]
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
Any of those branches might be stuck a release or two ago, so let's make sure to catch up with origin before doing anything else.
-
Cyril 'kibi' Brulebois authored
This seems in line with wiki/src/contribute/how/code/HACKING.mdwn
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
It might be better to remember where we started and go back but oh well; let's highlight being left in an unusual directory and let the RM figure out what they prefer.
-
Cyril 'kibi' Brulebois authored
-
Cyril 'kibi' Brulebois authored
Also expanding the list address and mentioning it's a public list while we're at it.
-
Cyril 'kibi' Brulebois authored
Previous operations are about dealing with other Git branches, so highlight that one needs to go back before touching the changelog.
-