hey "haha github uptime lol" and all that but this is serious: some pull request merges were completely mangled yesterday and the repo may not be in the state that its maintainers expect. If your project processed any updates yesterday, please double-check https://www.githubstatus.com/incidents/zsg1lk7w13cf

#github

Incident with Pull Requests

GitHub's Status Page - Incident with Pull Requests.

this is secondary but I don’t love how they completely changed the “Resolved” description within the last hour while leaving the timestamp at 13 hours ago
@0xabad1dea That's a yikes for sure
@0xabad1dea They have to pretend it was resolved quickly 🙃

@luxliquida to be clear, it already said it was resolved since yesterday. What changed is the textual description of the resolution event, without being indicated.

(the change was roughly from "We will provide instructions on how to mitigate as soon as possible" to "those affected have been contacted". this isn't some shocking coverup, just bad practice to overwrite and keep the old timestamp)

@0xabad1dea Oh, I figured as much. But I doubt it's an isolated bad practice, and it speaks to a larger culture of "make sure the dates are good for PR"
@0xabad1dea @luxliquida anybody have an example of a wrong merge in the wild"? I wonder if it's subtly wrong or blindingly wrong...
@0xabad1dea wow. that's goddamn nasty
@0xabad1dea At this point I see the last non-Flavor-Ade drinker left at Github trying to send a message "look, please just stop using this service, it's just me and the botlickers and I'm retiring next week".
@0xabad1dea So every project has to run git fsck --full or something?
As if we didn't have enough reasons to reconsider our reliance on GitHub! Personally, I just use it as a backup to my personal server (along with Codeberg of course) precisely because of these shenanigans.
@0xabad1dea Looks like my employer made a great choice when, a few months back, they decided to make GitHub the new default as a development platform, moving away from a GitLab service that is hosted internally, run by a dedicated team, and has been generally awesome for the last decade.