Rewriting history

[patent saffron]

ok starting a thread

Staged the rewritten history to a branch upstream: https://github.com/dagger/dagger/tree/shykes/fix-committed-binary

[vestal badge]

LGTM 👍

$ git diff main..upstream/shykes/fix-committed-binary
diff --git a/dagger b/dagger
deleted file mode 100755
index ce84325fd..000000000
Binary files a/dagger and /dev/null differ

[patent saffron]

temporary disabled force push protection on main

force push done

Looks ok

[vestal badge]

LGTM!

[patent saffron]

re-enabling branch restrictions

restrictions re-enabled ✅

Ok I think we're done

[vestal badge]

hm, i merged https://github.com/dagger/dagger/pull/9133, and noticed that this included a change that still had the dagger binary in it?

so now it's in main again 😱 (sorry that's my bad)

cc @shrewd geode - i made a new branch i can force-push to main https://github.com/dagger/dagger/compare/main...jedevc:dagger:jedevc/fix-committed-binary

[vestal badge]

not sure, why does github think devin co-authored that pr?

[patent saffron]

came here to ask the sane thing 🤔

[patent saffron]

@vestal badge @shrewd geode is there still a pending history rewrite?

what's the next step to close this?

[vestal badge]

we've rewritten history on main and rebased every pr that had the bad commits in them, so we shouldn't see this

however, we still need to keep in mind that anyone who force-pushes their pr might accidentally re-introduce it - so we still need to review carefully, and keep an eye out for reintroducing it

[patent saffron]

How hard would it be to have a check for large binaries being committed? It seems like it would kill 2 birds with one stone: easily catch regressions in this particular case; and also avoid future instances of a contributor (AI or human) slipping a binary into the history

[vestal badge]

probably not too much work - but depends on how we do the check - there may be legitimate reasons to commit a binary for tests etc (though i don't think we do that today)