[TYPO3-core] Change procedure revisited

Ernesto Baschny [cron IT] ernst at cron-it.de
Tue Nov 22 10:11:21 CET 2011


Markus Klein schrieb am 22.11.2011 03:56:
>> -----Original Message-----
>>
>> However this back-port can be done by anyone who cares to do it, but must
>> not absolutely be the committer as far as I understand the situation. This is...
> 
> And this is probably the biggest problem about the whole procedure. There's no responsible person for doing the backports.
> Anyone is usually no one.

Not true.

> As somebody recently pointed out: "We may have projects running and I will not be able to commit anything to previous branches. Next I will forget about it and we will not get those patches to the core. And what if I fixed 10 bugs in the master? Do I have to monitor them all? That will never work :("
> 
> So please agree on a way to handle this, asap. Thanks.

Whoever scratches it's itch about it takes the responsibility.

We cannot demand anything from the volunteer work of anyone. We could
say "the release managers should do it", but I just don't have the time.
The "friendly ghost of the week", might be another candidate, but there
is enough "other work" for him/her to do. Even the "commiter" might not
be tagged with this responsibility (because it usually also requires
testing, verifying *again*). Else we get stuck in a situation where
noone dares to merge anything to master, because he would then be
declared responsible for something he doesn't want to do.

So it depends on the patch, sometimes it's a matter of minutes and you
(the commiter) can just "do it" (cherry-pick and merge), sometimes it
involves testing in older branches, and you (the commiter) might not
feel like doing it (at least right away).

In my eyes this is ok, valid and not stopping our workflow. As soon as
the interest is there by anyone in backporting a specific bug, he can
just "jump in" and do it.

The script and overview of merges mentioned by Francois is there to help
us spot stuff that we still *might* backport.

It's not that we have *tons* of bugfixes missing a backport, since you
imply in your statement which I responded with "Not true". I feel that
it is going really well, even with the informal established routine of
"whoever has the time, does the backport":

http://www.typo3-anbieter.de/typo3-merges/

Is there any fix you are missing in 4.5? Why don't you go ahead and do
the backport? If nothing is bothering you, it's not a problem. If its
bothering your customer, he might pay you to do the backport.

But since the question is popping up every couple of weeks, we might
want to write something about it, at least if we agree on something.

Cheers,
Ernesto


More information about the TYPO3-team-core mailing list