On 2016-01-09 05:43:57, Gianfranco Costamagna wrote: [...]
>>Furthermore, it's very likely that borg 1.0 gets released before >>stretch, so all those arguments will be completely irrelevant because >>borg *will* be API stable. > > this would fix the issue once for all To be fair, while this will fix the issue for a while, we shouldn't attach a sense of eternity to this resolution. :) At some point, 1.0 will stop being supported in borg, and stretch will stop being supported in Debian. If anyone here is still running lenny boxes, you know what I am talking about. With all our talk of Debian being really stable for a really long time, while true, it doesn't mean "forever". I had an eye-opening experience when I looked at the Solaris release schedules, which are maintained for more than a decade. Solaris, however, does not ship stuff like PHP and Ruby which are changing constantly... Anyways, the point is, we will always have to deal with those issues. The solution, upstream, is to provide standalone binary packages to workaround that problem. Maybe, if we *really* have such a problem i n the future, we can provide versioned packages. For example, the autoconf suite does exactly that: the Debian archive has automake1.5, automake1.6 and so one. It uses the alternatives system to support multiple installs in parallel. That is annoying, but it works. >>This, in short, is a non-isse right now. If 0.28 was in stable and >>0.29 in testing, this would be a bug, but then the fix would be a >>backport, not a removal from stable (which you can do, if you are >>really stuck, anyways). >> >>Now, can i open this bug about backporting? :) > > > nope. Backport has been asked so long time ago, and I have personally > backported > one reverse dependency in jessie-backports great! > (setuptools-scm > https://packages.qa.debian.org/s/setuptools-scm.html > ) that backport is now out of date, btw. :) > but when I asked zigo about python-msgpack he told me "I'll backport it soon" > and in fact I remember seeing the package in jessie-backports new queue. > > For some obscure reasons the package got rejected, and I'm still investigating > about this. > > You can ask all the backports you want, but for sure we need to have the > build dependencies > prior to do them. prior to backporting, sure, but i'm talking about opening a bug report for the backport, where we can track stuff exactly *like* that. so if you don't mind, i'll create that backporting issue and make it depend on the msgpack backport. :p i'll be happy to do those backports myself. of course if zigo is taking care of it, i won't step on his toes, but it is common practice for the backporter to be a different person than the package maintainer... > BTW a little bit off-topic, but the package 0.29 will be in xenial (unless > you provide 1.0 in the next > month or so). > > While I can agree about backporting stuff in Debian, this usually will become > a trouble in debian-based > distros, where usually nobody actively maintain leaf packages (well, I'm a > recent MOTU Ubuntu Developer, > so I'll probably be able to backport it). > I don't want, neither Debian wants to break Derivative distros just because > of API changes. honestly, that is Ubuntu's problem, not ours. and while our decisions do affect Ubuntu, they profit from our packaging work for free, so they can hardly complain about things being out of sync on that other side. they are the ones who are responsible to do those syncs, not us. and yes, if we fix those issues in Debian (e.g. through backports), that *will* provide a clear path for derivatives as well. > So please, don't talk about this issue like a Debian specific one, it affect > also other distros too. meh. let's say it's a "debian family problem" and leave it at that if you want. :p but i don't think the debian borg maintainers should take the fall for the 6 month rolling schedule of Ubuntu. that would be totally unfair. xenial will be a snapshot of debian testing/unstable, as any other ubuntu release. if Ubuntu people do not want borg 0.29, they can remove it on their end... > I agree about what Danny said, we can downgrade this bug, but it will happen > on Ubuntu/Mint/Whatever > if we don't fix it in some way. > > Thanks for the clarification, I think we have ~20 days to downgrade it, and > probably it will be downgraded. > > Even if we don't downgrade it in 20 days, it will migrate as soon as we do > it, so really not an issue > (except for people who won't be able to install the package of course). sure. i just happened to look at backporting borg and figured it would be nonsensical if borg was removed from testing before that. :p > This bug is the sign that we really care about borg userbase, not the > opposite, and I'm pretty sure we will > make a even better package after this discussion (and some debian/NEWS files > :p ) sure, thanks for that! a. -- People arbitrarily, or as a matter of taste, assigning numerical values to non-numerical things. And then they pretend that they haven't just made the numbers up, which they have. Economics is like astrology in that sense, except that economics serves to justify the current power structure, and so it has a lot of fervent believers among the powerful. - Kim Stanley Robinson, Red Mars
