notfound 1118618 6.4.3-5
found 1118618 6.1.5-1~exp2
fixed 1118618 6.4.3-1
tags 1118618 + trixie pending
thanks

On Thu, 15 Oct 2026 23:04:01 +0200 Paul Gevers <[email protected]> wrote:

> This all doesn't feel like it should block migration of the version in
> unstable. Also, it doesn't look like the difference between the version
> in unstable and the version in testing is all that big. Is this bug even
> tagged with the right version?

This bug is a mess because (1) it was reported against the wrong package and (2) I was not familiar enough with the bug reporting system to know what the metadata should be, and (3) the package was in the middle of a transition. The binary package containing this bug, libamdhip64-5, had already started a transition to libamdhip64-6. The libamdhip64-6 on unstable at the time this bug was filed already fixed the problem. I didn't know how to handle a bug on a binary package that could not be fixed on unstable/testing because there was no associated source package anymore. [1]

The libamdhip64-5 containing this bug has been removed from the unstable / testing archives and now only exists on Trixie / Bookworm. There's no rocr-runtime 6.4 (or later) package on Trixie / Bookworm, so the bug is irrelevant there unless we do backports. Still, I've opened a bug [2] with a trixie-pu to fix it anyway, so I guess I'll tag this +trixie pending.

Since libamdhip64-5 does not exist on unstable/testing anymore and the problem was fixed in libamdhip64-6 on unstable, I'm going to close this bug as fixed.

Sincerely,
Cory Bloor

[1]: There were also a lot of other things going on at the time in the ggml dependency tree. I'd been pressured to  "keep everything on the same version" and I applied that as "change everything all at once" which lead to many transition problems.
[2]: https://bugs.debian.org/1152417

Reply via email to