justin = { main feed , music , code , askjf , pubkey };
Ask Justin Frankel
No reasonable question unanswered since 2009!

Suggested topics: programming, music, sleep, coffee, etc.

Note: please do not ask questions about REAPER features, bugs or scheduling, use the forums instead.

Name: Ask: Human (enter yes):
[back to index] | [unreplied] | [replied] | [recent comments] | [all]

Question: Can you explain your workflow of creating branches for specific features and then making custom builds that include some of those features. I don't really understand how you merge all stuff and keep everything sane. Do you have some automation system for this? How does the naming of branches / builds work? Thanks
Asked by nikolalkc (88.172.223.x) on August 6 2022, 11:32pm
Reply on August 7 2022, 3:16am:
    We use pre_ for branches that go in prereleases. Then we keep a branch called dev-pub which when we want to do +dev builds include all of the pre_ branches.

    When we want to do a build, we put a special commit on the end (WANT_BUILD version_number), and push that to a branch named want_build, which has machines watching it and ready to build on demand.

    Once a pre_ branch is ready-ish to go to release, we merge it to our main branch. Sometiems two pre_ branches end up intertwined, we'll merge them, or keep one dependent on the other, in order to avoid continual conflicts.

  • Posted by Gio (79.130.175.x) on August 7 2022, 1:26pm:
    When a release for a particular arch is ready; do you to test it individually or you rely on the outcome from the development arch that things went fine? And to expand the question a little bit more, do you use real machines for each arch or VMs?

  • Posted by Justin on August 7 2022, 4:09pm:
    We test the release as it will be in the form of release candidates. We use a mix of real machines and VMs, running on stack of small and power-efficient-when-idle mac minis and intel NUCs.

  • Posted by Gio (79.130.175.x) on August 7 2022, 4:50pm:
    Very nice Justin! In the past I used VMs for x86 but I realized that they were a waste of time, mainly because I couldn't disable specific SSE instructions to mimic real hardware (e.g. Pentium Pro).

    Your Name:   -- Site Owner's Name:  (for human-verification)


[back to index] | [unreplied] | [replied] | [recent comments] | [all]
Copyright 2022 Justin Frankel. | RSS