[PATCHES rtems, source-builder] Add GitHub Actions scripts
gedare at rtems.org
Thu Jan 19 14:42:35 UTC 2023
Nice. I would like some time to look at this and think about it a
little more. What would be the plan for removing this capability? Will
it leave any artifacts behind in the RTEMS github mirror?
On Mon, Jan 16, 2023 at 6:42 AM Christian Mauderer
<christian.mauderer at embedded-brains.de> wrote:
> some weeks ago I created a GitHub Actions based CI script that we
> (embedded brains) wanted to use to test patches (see
> https://github.com/embedded-brains/rtems/tree/ci). I don't think much of
> the RTEMS community noted these. I would like to suggest adding the
> scripts to the official RTEMS repositories so that the actions are
> executed in the official GitHub RTEMS mirrors.
> To make sure that GitHub pull requests are not perceived at the official
> way to make RTEMS contributions, an auto-responder action notifies the
> pull request user that the current way to make contributions is sending
> patch sets to devel at rtems.org.
> This step will allow users to easily test patches on a number of
> simulators before they send them to the mailing list. No one is forced
> to do it, but everyone can try it. For RTEMS, it has the advantage that
> the patches are at least guaranteed to be compile-clean on a selected
> number of BSPs and that they survived a test run on a simulator.
> Please note: With this patch I do not intent to push GitHub as the RTEMS
> CI or to move from mailing list patches to push-requests. My idea is to
> allow everyone to experiment with a proof of concept prototype. Based on
> your experiences in this test phase, I would suggest that we have a
> review discussion in a month or two to select a suitable way forward for
> RTEMS CI. I think after that test phase we all know better what we want
> or expect which helps selecting the best CI system that then can replace
> this proof of concept system with GitHub.
> But now to make make it more clear what we will get with merging these
> You can find a (not yet cleaned up) version of the patches in these
> The results from a current run on RTEMS are here:
> If you scroll down on that page, you get a summary that shows which
> tests fail on three (mostly) randomly selected simulator BSPs. GR740
> usually can run all tests but currently jffs2_fsrdwr fails. The full
> output of the rtems tester is in the Artifacts in case you want to take
> a look at the test output.
> If you want to try the CI with some of your patches before we merge this
> to the official repositories, feel free to create a pull request to the
> ci branches of the embedded-brains/rtems repositories. See the github
> manual for guidance how to create a pull request:
> Note: It is important that you somewhen forked from the official RTEMS
> repositories or from one of the forks using (for example) the fork
> button in the github web interface. If you just pushed a repo to an
> empty one, github doesn't recognize the link and won't allow you to
> create a pull-request towards the embedded-brains repository.
> Best regards
> devel mailing list
> devel at rtems.org
More information about the devel