Open baby-gnu opened 3 years ago
The problem appears to be data corruption on Github's side. There appear to be two labels named bug
even though bug
appears to be the primary key (or at least some sort of "unique" filter is used on that column). The issue 176
has one label with the id MDU6TGFiZWwxNjA2NzQzMTU2
and name bug
, but that label is not included in the list of "all" labels.
(ghub--graphql-vacuum
'(query
(repository
[(owner $owner String!)
(name $name String!)]
(issue [(number 176)]
(labels [(:edges t)
(:singular label id)]
id
name))
;; We cannot get information about the label by the id
;; MDU6TGFiZWwxNjA2NzQzMTU2 because the API expects the
;; name instead of the id. When we use its name "bug", then
;; we get information about another label by the same name.
(label ;; [(id "MDU6TGFiZWwxNjA2NzQzMTU2")]
[(name "bug")]
id
name)
;; This should but does not include the label with the
;; id MDU6TGFiZWwxNjA2NzQzMTU2.
(labels [(:edges t)
(:singular label id)]
id
name)))
'((owner . "saltstack-formulas")
(name . "openssh-formula"))
(lambda (data) (pp-display-expression data "*Pp*")))
(data
(repository
(issue
(labels
((id . "MDU6TGFiZWwxNjA2NzQzMTU2")
(name . "bug"))))
(label
(id . "MDU6TGFiZWw0MzI0NTg3MQ==")
(name . "bug"))
(labels
((id . "MDU6TGFiZWw0MzI0NTg3MQ==")
(name . "bug"))
((id . "MDU6TGFiZWw0MzI0NTg3Mg==")
(name . "duplicate"))
((id . "MDU6TGFiZWw0MzI0NTg3Mw==")
(name . "enhancement"))
((id . "MDU6TGFiZWw0MzI0NTg3NA==")
(name . "invalid"))
((id . "MDU6TGFiZWw0MzI0NTg3NQ==")
(name . "question"))
((id . "MDU6TGFiZWw0MzI0NTg3Ng==")
(name . "wontfix"))
((id . "MDU6TGFiZWwxMjMwMzg2OTE4")
(name . "bump:major"))
((id . "MDU6TGFiZWwxMjMwMzg3MzA0")
(name . "bump:minor"))
((id . "MDU6TGFiZWwxMjMwMzg3NjEx")
(name . "bump:patch"))
((id . "MDU6TGFiZWwxNTUzNDUwMTU2")
(name . "released"))
((id . "MDU6TGFiZWwyMDkwNzA1Mjk3")
(name . "dependencies")))))
There isn't really anything we can to but to ask Github to fix the corrupted data.
Thanks a lot for the debug @tarsius.
I found that the issue https://github.com/saltstack-formulas/openssh-formula/issues/176 label bug as an URL pointing to another repository https://github.com/saltstack-formulas/.github/labels/bug. I will see with the team if someone understand what's happening.
Thanks.
I found that the issue saltstack-formulas/openssh-formula#176 label bug as an URL pointing to another repository https://github.com/saltstack-formulas/.github/labels/bug. I will see with the team if someone understand what's happening.
In Github, you can group all your common labels under a .github
repository like the saltstack-formulas community did but each repository can have it's own labels even with conflicting names.
As an example for openssh-formula:
bug
defined in organisation .github repositorybug
defined for this repositoryCan forge
take this into account?
Regards.
Can
forge
take this into account?
Well of course it will have to.
On the other hand this sounds like the Github developers didn't fully think this through and should fix this by including the inherited labels in the list of a repository's labels. Edit: i.e. it should be a list of labels that "can be used in this repository".
Dealing with this is likely to have far reaching consequences and I likely won't find the time to do anything about it any time soon. For now I have added a variable to opt out of using labels.
git config --local forge.kludge-for-issue-294 true
Thanks, I did an update of forge and add the git config and I'm not blocked anymore.
Hello.
I'm happily using forge since several months and found an issue today with one repository:
It's working fine with other repositories like systemd-formula.
Here are few repositories wit the same error: