Re: .92 tarball deadline and hard code freeze

On Fri, Feb 28, 2020 at 10:25 am, Mart Raudsepp <[email protected]> wrote:
> This sounds more like a "problem" with the distribution, not that such
> releases shouldn't happen.
> If a minor release upgrade prompts for a full system reboot, then 
> maybe
> that update shouldn't be packaged up on such distros - it shouldn't
> mean the translations shouldn't get out the door in general in a 
> tagged
> release.

Fedora will continue to take all GNOME updates. Our update script isn't 
smart enough to conditionally decide whether a release has interesting 
changes or not like a human would do.

> Or they could be gated to all get pushed down all at once, so
> the reboot is made for many .4 packages, some of which have bug fixes
> too.

That's a good point, but it's certainly also nice to have fewer total 
packages to update. :)

(Of course, don't hesitate to release updates when you actually have 
bugfixes.)

> Also so late as .4 often seems to be more like a "oh, I haven't done
> any work on this module at all this cycle, so there's only translation
> updates - let me get at least these out for users to benefit from the
> translators work, but I'll not bump the minor as there have been no
> code changes" kind of deal too, so distros prompting reboots can
> include them in their next cycles updates, while others know to take
> them earlier.

Yeah, that's a good point. I have a module in this situation, GNOME 
Sudoku, with only translation updates since 3.34.1. On the one hand, 
since there are no changes since 3.34, there's really no need to do a 
3.36: I could just call the next release 3.34.2 so that 3.34 users 
benefit from the translation changes.

On the other hand, it's probably better to avoid the extra 3.34 update, 
so I think I'll call it 3.36.0? Hadn't really given it much thought 
until now.

Michael