On Mon, Jan 11, 2010 at 8:06 PM, Piotr Drąg <piotrdrag(a)gmail.com> wrote:
W dniu 11.01.2010 20:40, Miloš Komarčević pisze:
> Don't think stopping commits to branches globally is a good idea. Why
> would one be disallowed from making changes to an older branch in
> order to correct and/or improve translations of a still supported
> program version?
>
The point here is that anaconda is special. Older branches are not
really used anymore after creation. So any work pushed to them will be
wasted time.
Yes and no. As David points out, someone could decide to make a
respin/remix pulling in all the updates. Localized respins/remixes
particularly come to mind, where anaconda translations are of great
interest.
> Also, e.g. system-config-printer translations have in the past
always
> been done on a branch, even during the development cycle (i.e. before
> rawhide freezes).
>
s-c-printer uses different development model than anaconda. We can work
on master branch (which will be F13) until F13 is released, and then
master is being switched to to-be-F14, and we are working on that next
release.
s-c-printer translations were on 1.1.x branch in Tx all throughout (or
most of) the F12 development cycle, not master. Similarly for F11.
Therefore, I don't think it's safe to assume the maintainer will not
create a branch before rawhide gets frozen for R13.
My point is that Tx should not stop people from committing to
branches, but people should certainly be clear about what they are and
should be doing (warnings, TQSG, etc.).
Miloš