`gix` and others no longer keep the `CHANGELOG.md` uptodate

[Byron]

Discussed in #2362

^{Originally posted by roy-work January 10, 2026}
The one in gix/CHANGELOG.md ends at v0.74.1, but the current version is v0.77.0.

Just looking to see what the breaking changes are.

[Byron]

My latest comment reproduced here, CC @EliahKagan .

---

Thanks for explaining the situation! Should there be an issue for this behavior of cargo-smart-release? Or did this happen due to factors outside it?

I don't think it was outside influence, as it did just run through without issues.

Another thing I noticed, and wasn't sure if was intentional, is that the changelog entry was written to the GitHub release but not to the committed file. That seems bad to be, though I haven't looked into whether it's happened before or if it might somehow be intentional.

This is a new problem for me, and it's expected to write changelogs to files and then add them to a commit.
Fortunately, that issue will fix itself the next time the changelogs are generated, after all, they always regenerate from the entire history.

cargo changelog --write gix already does the trick, but I noticed that the tags are out of order. Try it yourself and you know what I mean. Maybe it's us misusing jiff? It says that v0.74.1 is 1 day older than the tag before, 0.77.0, which is months in the future.

I don't know if this is due to some regression in cargo-smart-release or something else.

It feels like a regression, but it's probably also related to some input data that it doesn't manage well. After all, CSR didn't really change, despite dependency updates maybe.

---

Steps towards a fix

I am trying https://github.com/copilot/tasks/pull/PR_kwDOKKB1fM68c-_i?session_id=8aee6be6-c3c9-43c8-b194-9e9afc1d6565 .

[EliahKagan]

I believe GitoxideLabs/cargo-smart-release#97 has fixed this, based on the changes and comment discussion there and the observed effect in gix/CHANGELOG.md. If the fix is found to be incomplete, then this issue could be reopened or a new issue created.

CC @roy-work -- and thanks for reporting this in #2362!

[Byron]

I think it's fine to keep the issue closed, but I also wanted to warn that I don't think the fix would necessarily affect the seeming lack of writing changelogs at all.
But we can watch this more closely for the release in 11 days.

[EliahKagan]

But we can watch this more closely for the release in 11 days.

Oh. If you expect his may recur, then I was too hasty in closing it. Sorry about that!