My thoughts... this is just a draft, add your thoughts/extra information or if you just agree, and we''ll try to get a final mail. I would be nice if we hear from modax before sending anything too (he is in vac). On Sun, Aug 16, 2009 at 02:41:27PM +0200, Marc Brockschmidt wrote:> Heya, > > As announced on dda [RT1], we want to get an impression when releasing > Squeeze is feasible. We have proposed a (quite ambitious) freeze in December > 2009, and some developers have noted that their planned changes wouldn''t be > possible in this time frame. So, to find out when releasing would work for > most people, it would be great if you could answer the following questions: >I will answer your including also Qt 4 plans, that as you know is a quite important part for KDE 4 :) There is also a bunch of widely used kde apps that do not belong to KDE core, but they are leaf apps.> * Which major upstream releases of KDE are expected in the next two > years? Which of those are material for Debian stable, which might be a bit > flaky? >We currently have KDE 4.3.0 in unstable and we will have at least a couple of point releases before this year ends. KDE 4.4 is expected about January 2010. KDE 4.5 is expected about July 2010. It is not a good idea to do more speculation about dates after this point because in KDE they are also testing new release management methods... About flaky releases, usually 0 point releases have some bugs and 1 point releases are highly preferred to ship in a Debian stable release. We currently have Qt 4.5.2 in the archive and if it even could get another point release, the development and expectations are on Qt 4.6. There is not public release schedule for Qt 4.6, but worse case scenario, I think there will be at least a pre-release by december. Qt 4.5 was released in March 2009, I really expect Qt 4.6 being release before Qt 4.5 is one year old. Whether KDE 4.4 will need Qt 4.6 or not, is still unknown, it depends a lot of Qt 4.6 release date.> * How much time do you usually need from a new upstream release of KDE > to a stable Debian package in unstable? >Major releases: We usually are able to have the packages ready by the day of the release since we always get the final tarballs between 5-7 days before the release. In almost all the major releases, we need to put some stuff throught NEW and sometimes this is faster, sometimes it is not so fast.> * How many "big" transitions will the upcoming changes cause? When should those > happen? Can we do something to make them easier? >Nothing I can think of now, we already have done most of the complicated stuff when moving from kde 3 to kde 4. Something it would be very nice to do for Squeeze is killing more kde 3 apps. Specially, some of the kde 3 apps that are not ported and have somehow a replacement in kde 4. This is being doing slowly [0]. Anyway, with the current dates you are proposing, I am almost sure kdelibs(4c2a) from kde 3 will have to be kept for Squeeze (plus Qt 3, of course) and it will be hard because they are quite dead upstream... What could be removed is arts (and we would love to), but it needs some work on it. In summary: We would like a Squeeze release with KDE 4.4.0 at least (4.4.1 if possible) and Qt 4.6. For us december work quite *bad*, if you freeze in december we''ll have some kde 4.3.x and from january we''ll be using 4.4, if the freeze last a few months (*), by July we will be using 4.6 and it will be quite hard to care about bugs in KDE 4.3.x, that will be from 2 releases ago.... (*) I am sorry but it is not crazy thinking about this happening... [0] http://wiki.debian.org/kdelibs4c2aRemoval
I mostly agree with Ana''s comments, but I think I have a couple of additions. I can''t be certain but I would bet that KDE 4.4 will not depend on Qt4.6 unless there is at least a release candiate of 4.6 at the time of KDE soft feature freeze (actually that would have to be very soon, it would probably be hard to remove 4.6 dependencies). Personally I like the idea of going for a very late patch release of a KDE release, e.g. the latest 4.3 because 4.4 will very likely introduce a lot of new stuff. Speaking as a KDEPIM developer, 4.4 is the release we are currently targeting or addressbook and calendar porting to Akonadi [1]. This is very nice for us as developers, however there are still issues for certain types of deployment, mainly because non of the embedded style database backends (SQLite, MySQL/Embedded) qualify for being our (Akonadi''s) default yet. Cheers, Kevin [1] Some functionality might even depend on Nepomuk -- Kevin Krammer, KDE developer, xdg-utils developer KDE user support, developer mentoring -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: This is a digitally signed message part. URL: <http://lists.alioth.debian.org/pipermail/pkg-kde-talk/attachments/20090817/3fa0d091/attachment.pgp>
KDE follows a 6 months cycle release. Qt 4.6 should be released by the end of 2009 or early 2010. imho, some people will try to push KDE 4.4 dependency on Qt 4.6. for Debian stable, I think releasing with Qt 4.6.0/KDE 4.4.0 is a very bad idea. bugfix/patch release is preferred and according to the timeline, it will be Qt 4.5.x/ KDE 4.3.x. basically, we have have the materials (Qt 4.5.2/KDE 4.3.0 and soon KDE 4.3.1). Now, we should fix bugs... In summary: I agree with Ana''s comments.
Hello, On sekmadienis 16 Rugpj?tis 2009 22:27:12 Ana Guerrero wrote:> > As announced on dda [RT1], we want to get an impression when releasing > > Squeeze is feasible. We have proposed a (quite ambitious) freeze in > > December 2009, and some developers have noted that their planned changes > > wouldn''t be possible in this time frame. So, to find out when releasing > > would work for most people, it would be great if you could answer the > > following questions:Nobody from Debian Qt/KDE team has expected (or dreamt of) the freeze to happen so early hence the announcement was quite shocking. If we freeze in December 2009, squeeze will be stuck with KDE 4.3.{3,4} and Qt 4.5.3. Is this good or bad? Well, I don''t see much to clap for here: 1) *Very* bad timing for Qt release. As it was pointed out, Debian will freeze with Qt 4.5 (basically EOLed upstream) and Qt 4.6 right around the corner. Given Qt major releases are rather slow (every ~8-10 months) and Debian stable actually releases with Qt 4.5 in late Q2 of 2010, that would be only 2-4 months until Qt 4.7. So who is going care to about Qt 4.5 then? Nor upstream, nor users and it is a big question if we will. 2) Regarding KDE itself, actually 4.3.{3,4} is not the worst candidate for Debian stable from the KDE 4.3 series, probably the best. But is 4.3 series a good candidate for Debian stable? Frankly, quality of KDE 4.3.0 suggests KDE 4 still needs some work on stability front: memory leaks, kwin cpu usage, io usage and plasma-desktop crashes / bugs. Some of this might be Xorg bugs as well. What is more, if we push Qt 4.6 into stable somehow, KDE 4.3 probably won''t work well with it and given that the most probable candidate for flackiness is plasma, nobody should expect fixes for this from upstream. Hence the whole picture of KDE 4.3 desktop still looks lacking. 3) Some important features are still missing like usable network manager integration, root administration in systemsettings (and overall situation with policykit), completely broken nepomum/soprano. Not to mention that some important third party KDE applications (e.g. k3b) are not ported to KDE 4 yet or existing ports are lacking in features. Only time can resolve these problems. 4) unstable has been open for too short period of time. Remember, KDE 4 is a fresh thing in unstable/testing, 9 months of limited (and otherwise restricted by overall size of Qt/KDE and testing transitions since KDE links whole Debian together) testing is not enough. Not to mention that neither of us expected to release with KDE 4.3 and didn''t give (nor will give now being late to the party) to it as much care as it probably needed to.> > * Which major upstream releases of KDE are expected in the next two > > years? Which of those are material for Debian stable, which might be a > > bit flaky? > > We currently have KDE 4.3.0 in unstable and we will have at least a couple > of point releases before this year ends. > > KDE 4.4 is expected about January 2010. > KDE 4.5 is expected about July 2010. > > It is not a good idea to do more speculation about dates after this point > because in KDE they are also testing new release management methods... > > About flaky releases, usually 0 point releases have some bugs and 1 point > releases are highly preferred to ship in a Debian stable release.Yes, Debian stable should not ship KDE or Qt .0 releases, .1 in the worst case, .2 or .3 would be perfect.> > * How much time do you usually need from a new upstream release of KDE > > to a stable Debian package in unstable? > > Major releases: > We usually are able to have the packages ready by the day of the release > since we always get the final tarballs between 5-7 days before the release. > In almost all the major releases, we need to put some stuff throught NEW > and sometimes this is faster, sometimes it is not so fast.Yes, "stable Debian package" does not take long but it is as stable as packaged upstream KDE is which is a big variable in early point releases.> > * How many "big" transitions will the upcoming changes cause? When should > > those happen? Can we do something to make them easier? > > Nothing I can think of now, we already have done most of the complicated > stuff when moving from kde 3 to kde 4.I agree, nothing very complicated remains.> Something it would be very nice to do for Squeeze is killing more kde 3 > apps. Specially, some of the kde 3 apps that are not ported and have > somehow a replacement in kde 4. This is being doing slowly [0]. Anyway, > with the current dates you are proposing, I am almost sure kdelibs(4c2a) > from kde 3 will have to be kept for Squeeze (plus Qt 3, of course) and it > will be hard because they are quite dead upstream...Squeeze will and should ship KDE 3 libs / Qt 3. But the goal to get rid of as much KDE 3 / Qt 3 apps as possible is still important.> What could be removed is arts (and we would love to), but it needs some > work on it.I don''t understand why nobody considers splitting artsd off (and do not pull it in by default) and forgetting about it. Who wants to transition kdelibs4c2a and rip users off from sound in their KDE 3 applications?> In summary: > We would like a Squeeze release with KDE 4.4.0 at least (4.4.1 if possible) > and Qt 4.6. For us december work quite *bad*, if you freeze in december > we''ll have some kde 4.3.x and from january we''ll be using 4.4, if the > freeze last a few months (*), by July we will be using 4.6 and it will be > quite hard to care about bugs in KDE 4.3.x, that will be from 2 releases > ago....In my opinion, Squeeze should aim at >= KDE 4.4.2, >= Qt 4.6.1. That would be around the end of Q1 2010 / beginning of Q2 2010. If freezing in late summer / early autumn, aim at the latest Qt 4.6.x and >= KDE 4.5.1. -- Modestas Vainius <modestas at vainius.eu> -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part. URL: <http://lists.alioth.debian.org/pipermail/pkg-kde-talk/attachments/20090821/f1af2781/attachment.pgp>
It''s worth mentionning Qt 4.5.x doesn''t support GCC 4.4 [1]. Qt 4.6 will add GCC 4.4 support. [1] Qt breaks strict aliasing rules resulting in memory corruption and miscompilation.
Hi In our response, I think we should really avoid mentioning at all a possible freeze in december 2009, as this is not what we are asked about. We are asked about what we would like to see in a release and when we can get that. So:> * Which major upstream releases of KDE are expected in the next two > years? Which of those are material for Debian stable, which might be a bit > flaky?KDE is currently releasing new feature versions in january and july, but they are also discussing their release management methods and plans. We would expect KDE 4.4.0 in end of January 2010 and KDE 4.5.0 in end of July 2010. Each month after this, there is new bugfix releases of the latest version, at least until .3, and maybe a bit longer for .4. I don''t think they have gotten to a .5 point release yet. the .0 release has with 4.1.0 and 4.2.0 been a bit flaky. 4.3.0 has been a bit better in this regard, but we would still like to ship something like a .2 release. The release schedules of Qt is also relevant. Qt4.6 is expected to arrive around christmas, and has some new nice features, but from a release point of view, it is the first release that is expected to build correctly with gcc 4.4.> * How much time do you usually need from a new upstream release of KDE > to a stable Debian package in unstable?KDE releases the tarballs to packagers one week before release, so packagers have time to prepare the packaging. We usually get something ready (with the 4.x releases) for release day, sometimes with some bugs, sometimes not. These bugs are usually weeded out during a week or two after this. For point releases, which is only bugfix releases, we usually get the packaging fully right at first upload, which happens around release day.> * How many "big" transitions will the upcoming changes cause? When shouldthose> happen? Can we do something to make them easierA removal-transition of arts would be nice, but requires a package name change of kdelibs4c2a, as kdelibs4c2a changes ABI without arts. Of each new upstream release (4.x), the kde4libs source package will probably have to break: kdebase-workspace-bin ( << 4.x), which generates a small mess, but nothing major is expected. A .la file removal transition should be considered carefully regarding to apps using kde3 kdelibs. If we could plan the Debian release cycle completely after what would be good for KDE, squeeze will freeze shortly after KDE 4.5.0 is uploaded to Debian (July 2010) and the .1 and .2 and .3 releases will be allowed into squeeze. Squeeze will release in November or December 2010. This will give us a whole lot of interesting features and also some of the tricky kmail bugs that only can be solved by a almost full rewrite will have happened, and KDE will have the possibility to take advantage of the new things in Qt4.6, and we will reach Qt4.6.2 or something like that on the Qt side. -- I''m not able to insert a 3D provider from Windows, how does it work? From the folder inside iMacOS NT you must download the command prompt for receiving the program of the file to a RAM mailer over a ADSL terminale.