Pushing a tag doesn't update in Packagist

[jaitaiwan]

Environment:

• Bitbucket
• Bitbucket Packagist POST Hook

Replication steps:

1. Create a new commit.
2. Push commit to repo.
3. Realise one forgot to tag the release.
4. Tag release and push it to the repo.

Symptoms:

• Package does not auto-update to latest release likely due to #200 (Specific Comment)
• Force update does not result in latest release appearing

[stof]

@jaitaiwan can you link to the package being affected ?

[jaitaiwan]

No worries: centricwebestate/provender is the main one. It seems to happen sporadically, sometimes the force update works, other times it's totally useless.

[stof]

AFAICT, all tags appear on Packagist for this repo

[jaitaiwan]

Yes that's because I've since done a push of changes AND tags in the one go for a new release.

[stof]

Does it also affect force updates, or only the update from the hook ?

If it only affects the hook, the question is whether Bitbucket triggers the hook when pushing only a tag.

[jaitaiwan]

Yes it effects force updates.

Did you read the full post? I mentioned that under the "Symptoms" heading. I also referenced a comment in which I confirm that bitbucket fails to push adequate data on non-commit pushes...

[Seldaek]

I don't think it's due to that bitbucket issue, as the only information we need from the hook data is the repository URL. If it affects force updates, then it might just be a delay in the bitbucket API propagating new tags, or network failure of some sort.

[jaitaiwan]

I can't say for certain, but my hooks into SemaphoreCI seem nearly instant.
There is the random occasion where the hook isn't run, but when it is, its
nearly instant.

On Sun, 31 May 2015 at 22:10 Jordi Boggiano notifications@github.com
wrote:

I don't think it's due to that bitbucket issue, as the only information we
need from the hook data is the repository URL. If it affects force updates,
then it might just be a delay in the bitbucket API propagating new tags, or
network failure of some sort.

—
Reply to this email directly or view it on GitHub
#500 (comment).