Split package versions in testing (due to libplasma CI failure)

[Colin S]

Hi,

Thanks for your quick response!

On 20/04/2026 12:59, Hefee wrote:
> Well we won't touch the plasma packages till the migration is done. because
> the CI failures do not stop the transition. The blocking bit is
> plasma5support, that needs to end for 5 days in unstable. In two days Plasma
> 6.6 will hopefully migrate to testing. Than we can fix those issues.

Oh, thank you for clarifying. I got confused after seeing no clear 
reason why the transition didn’t happen and just decided to ignore the 
“not blocking” part.

So plasma5support wait is the reason why auto-libplasma transition is 
“100%” but not finished? Or is libplasma transition stuck by the 
indirect boost1.90 collision mentioned on the transition page?

I like to understand enough so I can answer my own questions in future, 
so I appreciate your feedback. But I can also just wait to see what 
happens after plasma5support ages out too if this is too much. :-)

>> 3. Is it possible to do anything with the packaging/migration to avoid
>> this half-migrated state for the major point releases in future?
> We hope yes ;) With 6.6 we created a package to bundle all plasma packages
> together, so it is not possible anymore to mix different plasma versions. This
> bundle package is named plasma-version and all plasma packages depend on
> plasma-version-base-6.6.
> With a newer Plasma they will depend on 6.7 and so on with that apt is forced
> to remove all old packages, because  plasma-version-base-6.6 conflicts against
> plasma-version-base-Y.X. But this bundling was introduced together with Plasma
> 6.6 packages, so it does not help for the current transition from 6.5 -> 6.6.
> But hopefully it is the last half transition to testing.

Nice! This is great! :-)

Best regards,