Strong +1 on moving to GitHub as soon as possible. Very important to convert the complete Trac ticket history to GitHub issues/PRs.
On Friday, September 9, 2022 at 2:54:06 AM UTC-7 Dima Pasechnik wrote: > I am resurrecting this thread, as in addition of trac continuing to eat up > funds (at a rate of over US$ 10 per day at the moment), it has gotten > increasingly broken. In particular, in the last 2 weeks no new developers > can really join the project, as there is no normal* way to add new ssh keys > into trac accounts, and it's not possible to push/pull with "new" github > ssh keys, i.e. keys that were not already "known" to trac, i.e. added to > the trac store of ssh keys before the last breakage happened. > > As far as funding is concerned, attempts to bring trac to a "free" hosting > stalled (see earlier messages in this thread). > > A further longer term issue is that trac software is basically on life > support, and it's only matter of time it will become totally obsolete. > > Such a move will allow a considerable simplification of our devops, and > free up quite a bit of developer time > to do interesting work rather than messing around with semi-obsolete stuff > such as trac, gitolite, etc. > > Importantly, Volker, the release manager, is willing to proceed with the > move. > > Also, various Sage upstream (and downstream) projects have moved away from > trac to github, e.g. Cython, or away from another system to github, e.g. > CPython, GAP, jupyter, etc... > > There is a trac ticket to manage the proposed move, > https://trac.sagemath.org/ticket/30363 tentatively set for Sage 9.8. > > I've conducted few experiments with a tool to import trac sites to github: > https://github.com/svigerske/trac-to-github, which in particular allows > to import trac tickets as github issues; a result of running it on few > tickets > may be inspected here: > https://github.com/dimpase/trac_to_gh/issues?q=is%3Aissue+is%3Aclosed > (Here issues 1-10 correspond to trac tickets one to one :-)) > Further work on trac-to-github will be needed, in particular to properly > link branches in our git tree, but it's doable, > and we have volunteers to do it. > > We'd like to hear about serious objections to the move, if any. > > > > *) normal - i.e. using trac interface; we (probably) still have a way to > modify the repository of ssh keys used by trac manually. > > On Thursday, March 18, 2021 at 10:53:54 AM UTC Frédéric Chapoton wrote: > >> Erik, did you stop the Orsay runners for gitlab ? It seems that the >> docker build there for 9.3.b9 is stuck by lack of runners. >> >> https://gitlab.com/sagemath/sage/-/pipelines >> >> Frédéric >> >> Le jeudi 11 mars 2021 à 13:25:52 UTC+1, erik....@gmail.com a écrit : >> >>> On Thu, Mar 11, 2021 at 1:20 PM E. Madison Bray <erik....@gmail.com> >>> wrote: >>> > >>> > On Thu, Mar 11, 2021 at 12:52 PM Dima Pasechnik <dim...@gmail.com> >>> wrote: >>> > > >>> > > On Thu, Mar 11, 2021 at 10:11 AM Dima Pasechnik <dim...@gmail.com> >>> wrote: >>> > > > >>> > > > On Wed, Mar 10, 2021 at 4:00 PM E. Madison Bray < >>> erik....@gmail.com> wrote: >>> > > > > >>> > > > > On Tue, Jan 12, 2021 at 11:33 PM tobia...@gmx.de < >>> tobia...@gmx.de> wrote: >>> > > > > > >>> > > > > > >>> > > > > > For what's worth, + 1 for migrating to github. >>> > > > > > >>> > > > > > The interface is cleaner, it has many more features and >>> integrations, and is more active which could attract more contributions. >>> There are a few scripts/tools that allow to migrate from trac to github. >>> But most of them are unmaintained for a few years already, so I'm not sure >>> if they still work (which should be taken as a sign that one should migrate >>> sooner than later). >>> > > > > >>> > > > > In 2019 Julian Rüth and I, with the help of some others, already >>> put >>> > > > > in some effort to set up an organization for SageMath on GitLab: >>> > > > > https://gitlab.com/sagemath >>> > > > > >>> > > > > Between GitHub and GitLab, we felt that the latter would be more >>> > > > > acceptable to the broader Sage community. We also implemented a >>> bot >>> > > > > that can mirror GitLab merge requests as Trac tickets, though >>> it's >>> > > > > been in need of troubleshooting for a while. >>> > > > > >>> > > > > This was also done before the advent of GitHub Actions, and the >>> > > > > ability to provide custom CI runners for GitLab Pipelines seemed >>> > > > > advantageous, since we could maintain our own fleet of runners, >>> be it >>> > > > > on Sage developers' personal machines (if they are generous >>> enough to >>> > > > > host them) or any conceivable constellation of cloud computing >>> > > > > platforms. >>> > > > > >>> > > > > In practice this has gained little traction, in part due to lack >>> of >>> > > > > advertising. The GitLab Runner solution also proved a bit >>> troublesome >>> > > > > to maintain, as it required some constant attention to make sure >>> there >>> > > > > were always working runners available. I tried to keep that up >>> for a >>> > > > > while myself, but have had other obligations. >>> > > > >>> > > > I think it should be mentioned that GitLab has an analog of GitHub >>> Actions, >>> > > > and the difference is that its runners may be self-hosted, or >>> provided >>> > > > by GitLab. >>> > > > E.g. see >>> https://gitlab.com/sagemath/dev/trac/-/pipelines/266731297 >>> > > >>> > > I just tried to switch to a "community" runner, and got an error >>> which >>> > > is probably >>> > > obvious to people versed in Docker: >>> > > >>> > > https://gitlab.com/sagemath/dev/trac/-/jobs/1089520433 >>> > >>> > I think it might be because the Docker builds have been otherwise not >>> > working for a while (due to lack of reliable runners). So a more >>> > recent "build-from-clean" job is needed. These jobs are run when >>> > develop/master are updated as well as on tags. Whereas >>> > "built-from-latest" is run on branches for tickets. It tries to build >>> > the branch on top of the "latest" Docker image e.g. for develop. But >>> > the last one that built successfully is too old, and so trying to make >>> > the diff between that ticket and the version of develop it's based on >>> > fails. Hence the message: >>> > >>> > "Could not find commit fbca269f627bf6a8bc6f0a611ed7e26260ebc994 in >>> > your local Git history. Please merge in the latest built develop >>> > branch to fix this: git fetch trac && git merge >>> > fbca269f627bf6a8bc6f0a611ed7e26260ebc994" >>> > >>> > But for the automated CI that's not a very useful message... >>> > >>> > I know Matthias has done some impressive things to get around GitHub >>> > Actions' time limit on jobs by breaking the build up into "stages" >>> > that can be split across multiple jobs. I see no reason that couldn't >>> > work with GitLab as well. >>> > >>> > But it would still be better to have our own fleet of runners--they >>> > would be faster, and we could test on more different custom hardware >>> > configurations. The problem is that this is at a minimum a part-time >>> > job... >>> >>> Well looks like I need to correct the record a bit. Perhaps I've been >>> a bit too sanguine about the state of the GitLab builds. In fact, the >>> latest develop commit, 9.3beta8, built quite successfully: >>> https://gitlab.com/sagemath/sage/-/pipelines/266734885 >>> >>> And it ran on one of the fleet of runners I've been maintaining here >>> at Paris-Saclay, which I haven't touched in months. So I guess it's >>> still working after all ^^; Ever since I set this up I had been >>> having a problem with runners randomly erroring out, and not being >>> deleted correctly when they do. I have tried many times to fix it to >>> no avail, and I kind of gave up for a while. I assumed eventually >>> this caused things to grind to a halt, but apparently not. >>> >>> Knowing that it's still working at least somewhat gives me motivation >>> to try again to investigate the problem with the erroring runners and >>> see if it can't be fixed. Maybe an upgrade of the gitlab-runner >>> controller is in order... >>> >> -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/bd753905-52b7-47b8-9b41-0cd5048361e1n%40googlegroups.com.
