GitLab seems quite poorly

[Ben Gamari]

Ben Gamari <ben at smart-cactus.org> writes:

> Matthew Pickering <matthewtpickering at gmail.com> writes:
>
>> I don't know what has changed since yesterday but pushing new branches
>> to GitLab is broken.
>>
>> The two most recent branches pushed to origin 404 when you click on the link.
>>
>> https://gitlab.haskell.org/ghc/ghc/branches
>>
>> https://gitlab.haskell.org/ghc/ghc/tree/wip/mp/15646
>>
>> I also pushed some commits to a fork and the MR which they appear on
>> didn't update at all. Perhaps a symptom of the same problem.
>>
>> Also a recent MR by Ben complains of the source branch for the MR not
>> existing.  https://gitlab.haskell.org/ghc/ghc/merge_requests/625
>>
>> The summary is that GitLab seems unusable at the moment. Did you
>> fiddle with anything recently Ben?
>>
> Yes, I am aware of these issues have been working in the background to
> resolve or report them. In short, I updated to 11.9.0 yesterday. This
> seems to have introduced a bug in the handling of branches whose names
> contain forward slashes. Unfortunately it hasn't been easy to nail down
> precisely what broke.
>
> I have tried rolling back to 11.8.2 but sadly the issues seem to persist.
>
For what it's worth, I managed to come up with a patch that appears to
avoid the issue. It appears to be due to caching within GitLab, although
I can't say I understand precisely how it is going wrong. I have
reported the regression as gitlab-ce#59403 [1].

Cheers,

- Ben


[1] https://gitlab.com/gitlab-org/gitlab-ce/issues/59403
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 487 bytes
Desc: not available
URL: <http://mail.haskell.org/pipermail/ghc-devs/attachments/20190323/2ca26977/attachment.sig>