Re: [BUG] b4: assertion error with `b4 send`

Konstantin Ryabitsev <[email protected]> writes:

> On Thu, Jun 04, 2026 at 10:23:39PM +0200, Andreas Hindborg wrote:
>> Hi,
>> 
>> I got this assertion error when sending patches today:
>> 
>> ---
>> Sent 11 messages
>> Tagging sent/20250305-unique-ref-29fcd675f9e9-v17
>> Recording series message-id in cover letter tracking
>> Created new revision v18
>> Updating cover letter with templated changelog entries.
>> Invoking git-filter-repo to update the cover letter.
>> New history written in 0.06 seconds...
>
> [...]
>
>>   File "/nix/store/xl7fx5x8g52g03npg1dpjzvaiijpv98z-python3.13-git-filter-repo-2.47.0/lib/python3.13/site-packages/git_filter_repo.py", line 4767, in _record_metadata
>>     self._compute_metadata(metadata_dir, orig_refs)
>>     ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^
>>   File "/nix/store/xl7fx5x8g52g03npg1dpjzvaiijpv98z-python3.13-git-filter-repo-2.47.0/lib/python3.13/site-packages/git_filter_repo.py", line 4712, in _compute_metadata
>>     assert(usoa == intermediate) # old wasn't pruned => usoa == intermediate
>
> Looks like this tripped something in git-filter-repo, so not something we
> control anyway.
>
>> It looks like everything is OK and b4 set up for the next spin, cover
>> letter looks right.
>
> Yes, looks like it was during some kind of cleanup.
>
>> I observed this on nixos first with b4 0.14.3 via nixpkgs 25.05, then
>> upgraded to 0.15.2 via nixpkgs unstable.
>
> We actually got rid of git-filter-repo in the latest master, so I'm just going
> to blame this on gremlins and blame the alignment of the celestial spheres. :)

Ok, cool, I'm going to pull down that and see what happens.


Best regards,
Andreas Hindborg