On Tue, Sep 2, 2025 at 12:11 PM Andre Vehreschild <ve...@gmx.de> wrote: > > On Tue, 2 Sep 2025 11:00:41 +0100 > Jonathan Wakely <jwakely....@gmail.com> wrote: > > > On Tue, 2 Sept 2025 at 10:47, Andre Vehreschild via Gcc <gcc@gcc.gnu.org> > > wrote: > > > > > > Er, well, I missed to tell what I did to the branch: > > > > > > - rebased to current master, > > > - reverted a squashed commit, and > > > - applied 8 separate commits. > > > > > > So nothing spectacular. Nothing I would expect anything to chew on for so > > > long. > > > > It looks like you're pushing nearly one thousand new commits: > > > > $ git log --oneline master..c0159d2c27e3b9b849fee8ce8d47a92806646888 | wc -l > > 953 > > These 953 changes came because of a git pull on the branch. At least that is > when they occurred. The rebase only added threehundertsomething. > > > That's going to take far too long. > > > > Could you use forge.sourceware.org for branches like that? > > I'll rather abandon using branches on gcc's git when those basic operations > take that long and block others. > > > There's no reason for the server hooks to process this work if it's > > just in-development work being done on a branch which you plan to > > rebase often anyway, so it could be hosted inthe forge instead of the > > main git repo (but is rebasing a public devel/xxx branch a good idea > > anyway?) > > Philosophy! > > There are thousand ways to use git. So I'll go back to 90's and just publish > patches on the list. Not being able to use the most core operations on > branches > is demotivating to me.
I think we need to sit down and see why this takes so long. It shouldn't. Telling people to not do it works, but it's hardly the correct thing to do here ... I've been using devel/ branches just fine, but I usually force-push new state that I've locally rebased ontop of trunk, so it's always just a small set of local changes over a git revision known. Richard. > - Andre > > > > > > > > > > > - Andre > > > > > > On Tue, 2 Sep 2025 11:39:18 +0200 > > > Andre Vehreschild <ve...@gmx.de> wrote: > > > > > > > Hi Jakub, > > > > > > > > I am pushing to gfortran-test. The process seems to be finished, but is > > > > not returning (sorry it's in German): > > > > > > > > Objekte aufzählen: 11679, fertig. > > > > Zähle Objekte: 100% (11679/11679), fertig. > > > > Delta-Kompression verwendet bis zu 28 Threads. > > > > Komprimiere Objekte: 100% (2995/2995), fertig. > > > > Schreibe Objekte: 100% (9904/9904), 10.15 MiB | 998.00 KiB/s, fertig. > > > > Gesamt 9904 (Delta 8217), Wiederverwendet 7591 (Delta 6861), Paket > > > > wiederverwendet 0 (von 0) remote: Löse Unterschiede auf: 100% > > > > (8217/8217), > > > > abgeschlossen mit 1517 lokalen Objekten. > > > > > > > > And here it hangs. I don't know what the server is doing, but my side is > > > > doing nothing. No load on CPU or network. Shall I kill the git process > > > > here? > > > > > > > > Sorry for the inconvenience. I was waiting for the server to finish > > > > whatever it is doing. > > > > > > > > - Andre > > > > > > > > > > > > On Tue, 2 Sep 2025 11:26:54 +0200 > > > > Jakub Jelinek <ja...@redhat.com> wrote: > > > > > > > > > On Tue, Sep 02, 2025 at 11:04:45AM +0200, Richard Biener via Gcc > > > > > wrote: > > > > > > > > > > > On Tue, Sep 2, 2025 at 11:01 AM Rainer Orth via Gcc > > > > > > <gcc@gcc.gnu.org> > > > > > > wrote: > > > > > > > > > > > > > > I'm currently trying to push to the repo for an hour, but only get > > > > > > > > > > > > > > remote: > > > > > > > --------------------------------------------------------------------- > > > > > > > remote: -- Another user is currently pushing changes to this > > > > > > > repository. -- remote: -- Please try again in another minute or > > > > > > > two. -- remote: > > > > > > > --------------------------------------------------------------------- > > > > > > > remote: error: hook declined to update refs/heads/master To > > > > > > > git+ssh://gcc.gnu.org/git/gcc.git ! [remote rejected] > > > > > > > master -> master (hook declined) error: failed to push some refs > > > > > > > to > > > > > > > 'git+ssh://gcc.gnu.org/git/gcc.git' > > > > > > > > > > > > > > Another commit a few minutes before worked just fine. Is there > > > > > > > someone doing an excessive commit or has some lock been left > > > > > > > behind? > > > > > > > > > > > > On IRC Jakub said > > > > > > > > > > > > 10:05 < jakub> locked by user vehre at 2025-09-02 07:34:09.557509 > > > > > > (pid = 3710103) > > > > > > 10:05 < jakub> 3710103 ? R 27:15 python hooks/update.py > > > > > > refs/heads/devel/gfortran-test > > > > > > cb4b73da237153871fb840a3a31a79354933a8bb > > > > > > c0159d2c27e3b9b849fee8ce8d47a92806646888 > > > > > > 10:05 < jakub> so, it isn't stale, it is actively doing something, > > > > > > but it takes more than 30 minutes already > > > > > > > > > > It is now already more than 90 minutes (and the process taking almost > > > > > 14GB of memory). > > > > > Andre, what are you trying to do, can it be killed so that others can > > > > > access the repository? If what you're trying to do needs to be really > > > > > done, can it be done in smaller steps, such that it won't lock > > > > > everybody out for more than a few minutes at a time? > > > > > > > > > > Jakub > > > > > > > > > > > > > > > > > > > > > > -- > > > Andre Vehreschild * Email: vehre ad gmx dot de > > > -- > Andre Vehreschild * Email: vehre ad gmx dot de
