On Mon, Oct 28, 2024 at 08:21:01PM +0000, Gavin Smith wrote: > I was only opposed to skipping the tests, not to allowing customizing the > shell. I haven't got an opinion on the use of SHELL as mentioned at > the referenced message,(*) as I haven't researched the implications of this. > It seems like a valid goal to allow the user to override the shell > when running Makefile rules that launch shell scripts, and SHELL is > already used in some places.
Ok. I will add to the main TODO. I think that it is low priority, though. > > If we drop omnios from the platforms that are run, I think that it would > > be better to leave it in the file and retest from time to time. > > Note that the CI system on Github is not used or controlled by the GNU > project. Github is not especially recommended for use by GNU packages; see > <https://www.gnu.org/software/repo-criteria-evaluation.html#github-com>. > > If people want to use Github to run tests of Texinfo, that is up to them. > They are not discouraged from doing so and reports from such tests are > welcome. In the case of https://github.com/gnu-texinfo/ci-check this is actually Bruno Haible and myself managing it, so we have control over whether omnios is enabled or not in that particular CI, and we (Bruno and me) also have a responsibility if our use of github is not ethical. I do not have a good understanding, actually, whether the CI system we use is transparent in that it could be reproduced somewhere else or not. In general the recipes we use come from code that seems to be free software, but we also use some containers, I don't know where they come from. Also there is the SaaSS issue, which is unclear to me in general, and in particular for CI (that I do not want to run on my computer for diverse reasons). -- Pat
