|
From: Michael B. <mb...@de...> - 2010-03-12 00:36:27
|
Hi, On Thu, Mar 11, 2010 at 06:28:19PM -0500, Chris Frey wrote: > On Fri, Mar 12, 2010 at 12:17:28AM +0100, Michael Banck wrote: > > Does anybody else see another way forward? > > This is really unfortunate. An external dependency is killing 0.22 > when we've put real effort into making it possible to run 0.22 and > 0.40 side by side. Well, making it possible to run both and actually supporting the stable version are two different things. If you look at the trac list for 0.22, you see loads of tickets which look fatal. AFAICT, there were two commits on the 0.2x branch in the last 2-3 years. Maybe it is my fault, but I stopped fowarding (or even looking) at 0.22 bugs in Debian for a while now, assuming nobody else cares, either. > The only way to properly kill 0.22 is by making 0.40 better. And if > all those plugins have to go, 0.40 is not yet better. I agree. That is why I also mention delayed releases - I have the feeling that *some* plugins sometimes get ported after a release, but they're always catching up with the new changes. > Remember Fedora's error. I have no problem with 0.40 being in testing > and unstable. I do have a problem with 0.22 dying before a proper > replacement is ready. Well, 0.22 will live on in Debian stable, so probably supported until late 2011. > What is missing in libsyncml that 0.22 can't work with it? I'm afraid > I'm not up to speed on the syncml side of things. I believe the story is: libsyncml-0.4.7 (I think) and above use libsoup-2.4, while the others use libsoup-2.2. The syncml plugin for opensync-0.22 only works with libsyncml-0.4.6 (I think) and lower. I really don't want to point fingers, and I take the blame for the wrong upload, but in the end, I would have just as well removed opensync-0.22 for the next stable release even if 0.3x would still be in experimental. Michael |