IRC log of #maemo-meeting for Thursday, 2010-04-22

*** dneary has quit IRC01:03
*** dneary has joined #maemo-meeting04:17
*** dneary has quit IRC04:54
*** X-Fade has joined #maemo-meeting09:45
*** achipa has joined #maemo-meeting09:51
X-Fadehi all10:00
achipahi10:00
*** sandman_ has joined #maemo-meeting10:00
X-FadeIt seems we are a bit light today :)10:00
sandman_good morning everybody10:01
achipalet's wait a few more minutes ;)10:01
X-Fadehi sandman_10:01
Stskeeps(what meeting's scheduled?)10:03
X-FadeWe're going to talk about Qt and Qt experimental packages in Extras.10:04
achipa"discuss the future of Qt10:04
achipalibraries in Extras-devel."10:04
achipawell, try to, anyway :)10:04
Stskeepsah, i'll listen in then10:04
sandman_good morning everybody (got a date with the dentist), but from what we talked yesterday, it seems that much was already discussed via email10:06
sandman_s/good morning everbody/ btw: harryf won't show up today/10:06
X-Fadesandman_: It seems that you are the only one then ;)10:07
sandman_from Qt? yes10:07
X-FadeI feel we need to discuss more, especially about officially supported versions and experimenta in Extras.10:08
X-FadeEspecially what happens when an experimental version becomes stable/supported.10:08
*** harryF has quit IRC10:08
sandman_our proposal would be to either publish 4.7 snapshots as qt4-experimental-*  to extras-devel, or to publish them in our own repo on Qt servers10:09
X-FadeThe issue isn't really where they are published, but that a plan needs to be made about transitioning and preventing a mess.10:10
sandman_we somehow need to provide devs with 4.7 snapshots until PR1.3 (when 4.7 will officially replace 4.6 in the firmware)10:11
X-FadeAnd will you push things like qtm linked against PR1.2 or against a snapshot?10:12
sandman_qtm is a snapshot right now - qtm will become an official, stable part of Qt together with 4.710:13
X-FadeOk, so from then on it will be published either in firmware or in nokia-apps.10:13
sandman_I don't know about the plans, since that isn't done by us here in Munich, but by the Brisbane team10:13
*** harryF has joined #maemo-meeting10:13
achipasandman_: there was a notion from Kate that there will be a stable version for 4.6, in Extras10:13
X-FadeProblem is that once PR1.3 is there we then need to rip it out again.10:14
sandman_from Qt's side there will be no offiial QtM release for 4.6 (at least that *I* know of)10:14
X-FadeThere doesn't seem to be a lot of thought going into the consequences of doing these tricks in extras.10:15
sandman_our (Qt's) plan should be to publish Qt and QtM 4.7 snapshots (either extras-devel or another repo) once PR1.2 is out10:16
sandman_as soon as PR1.3 is around the corner we move 4.7 into the firmware10:16
X-FadeWe need to see if we can prevent dual installs of Qt as much as possible though.10:17
sandman_*we* don't want to move to extras10:17
sandman_snapshots -> e-d / stable -> firmware10:18
X-FadeNo, me neither ;)10:18
sandman_whatever Kate and F-N are doing is out of our hands though :)10:18
X-FadeJust need to make developers aware that using qt4-maemo5 is really the experimental part.10:18
sandman_yes - that name is way too generic10:19
sandman_we'll change that10:19
achipaplus, don't forget your bindings, whatever you do, we need to follow10:19
X-FadeThey expect this to be the qt library to use and then get bitten.10:19
achipa(pyside, pyqt et al)10:19
sandman_do we really need bindings for snapshots?10:19
* sandman_ is a hardcore C++ coder :P10:19
X-FadeI feel bindings should really track stable.10:19
sandman_yes10:20
X-FadeAnd if they want to track experimental, then they need to be renamed as such?10:20
achipaerr10:20
achipathis is in complete contradiction with what the PySide team does10:20
X-FadeSo there is a clear divide between official 'will not break' and experimental?10:21
sandman_I think this comes from the fact that the old official Qt 4.5 on the device was not really usable10:21
sandman_so it made more sense to track the unstable 4.610:21
sandman_I guess everybody should adapt to the new situation where the official 4.6 is to be preferred over an unsupported snpshot10:22
X-FadeSo how about removing qt4-maemo5 after PR1.2 release?10:22
achipasandman_: qtm and qml, there's the reason this will happen again (and harryF agreed on this  :) )10:22
sandman_achipa: then you need a pyside-experimental package to go along with the qt4-experimentl packages10:23
sandman_(and don't complain if we break things on a daily basis :) )10:24
sandman_X-Fade: we will remove them and replace them with qt4-experimental 4.7 packages10:24
X-Fadesandman_: Well that is where you need me.10:24
X-FadeYou guys think too generic :)10:25
achipado you think it would be a bad idea to have a qt4-experimental METApackage ?10:25
sandman_rm -rf qt4-maemo5-* ? ;)10:25
achipaand then we could ha depend in on semi-stable devel snapshots ?10:25
X-Fadesandman_: Not as simple as that, but in the end that is what happens then yes.10:26
achipa(this is what essentially Qt does with TP, beta1, beta2, rc, etc)10:26
sandman_achipa: qt4-experimental-dev does that already (more or less)10:26
achipasandman_: but that does not shield us from bleeding edge breakages10:27
sandman_achipa: we won't follow the exact Qt 4.7 release schedule, since we have our separate repo and need to merge and test first10:27
achipaaha, ok10:27
sandman_achipa: that's why it's called experimental :) ... really you can't expect something 100% stable from an unreleased product10:28
achipa(the problem is that the chain can get long - snapshot of qt, snapshot of the bindings, snapshot of qtm, it's enough to have an oops in just one of them and dozens of apps break)10:28
X-FadeThat is why I think we should stress that experimental is just that.10:28
X-FadeNot inteneded to be use by anybody below the advanced hacker.10:29
X-Fade*used10:29
sandman_achipa: once again - we think it's a good idea to provide binary snapshots to make other dev's lives easier, so they do not have to compile Qt themselves10:29
*** sbradley has joined #maemo-meeting10:29
X-Fadehi sbradley.10:29
*** kate_ has joined #maemo-meeting10:29
sandman_sandman_: whatever you (or Pyside or pyqt) will do with it is something we can't and won't control10:29
sbradleyhi, have I missed much?10:30
sandman_sbradley: hi - just half an hour :)10:30
kate_Hi, i just joined, sorry being late, dentist appointment took longer than expected10:30
achipahi10:31
X-Fadesbradley: After PR1.2 release, removal of qt4-maemo5 from extras-devel10:31
X-Fadeqt4-maemo5 to be renamed to qt4-experimental10:31
sandman_sbradley: are there plans from Brisbane's side to publish QtM packages for 4.6 ?10:31
sbradleysandman_: I believe we'll just be using the "official" Qt releases for the QtM dependencies10:32
sbradleyso 4.6 until 4.7 becomes the official supported version10:32
X-Fadesbradley: And publish them through official repos? Or extras-devel?10:33
sbradleyextras-devel but promoted to extras10:33
X-Fadesbradley: Are these official Nokia supported versions?10:33
sandman_sbradley: will those be binary compatible (4.6 and 4.7 QtM) ?10:33
sbradleyX-Fade: yes, we wont be publishing unstable snapshots (at least not initially). just the official QtM releases10:34
X-Fadesbradley: Then they should not go to extras.10:34
X-Fadesbradley: Then they should go to Nokia-apps.10:34
X-FadeSo they are available for everybody with the proper firmware and available from Nokia.10:35
X-FadeOtherwise once QtM gets in PRX.X, there will be a conflict.10:35
sandman_sbradley: X-Fade: BIC handling needs to be clear though - otherwise we get in a huge mess with PR1.310:36
sbradleysandman_: I believe so (I'll need to check); it's just dynamically linked to Qt so it shouldn't matter whether it was built against 4.6 or 4.710:36
X-FadeWe had that mess with Qt4.5.3 in extras and I really want to prevent that.10:36
X-FadeOfficially supported libs should come from the official source.10:37
X-FadeAnd extras is not that.10:37
kate_For QTM, i expect there should be unstable versions as early as possible with qt4-experimental10:37
sandman_sbradley: the thing is not the Qt dependency, but the the API of the QtM libraries (read: can a app compiled 4.6 QtM run against a 4.7 QtM version)10:37
sandman_if NOT, we need to name the packages differently10:37
sbradleysandman_: yeah, there's only one version of QtM (1.0), it's not related to the Qt versions10:37
sandman_sbradley: ok - good :)10:38
sandman_so I'm with X-Fade on this one - move them to nokia-apps as soon as these packages are stable and final10:39
kate_There will be a lot of development for QtM after current release, as example IM and Presence are still coming to next versions10:39
X-Fadepackage domains will otherwise really bite us later.10:39
sbradleywhat's the process for getting stuff into nokia-apps?10:39
sandman_kate_: then put those in e-d and make them depend on the upcoming 4.7 snapshots (qt4-experimental-*)10:39
kate_About official QtM, earliest Maemo version where it will be official part will be PR1.310:39
kate_Before PR1.3 QtM must be extras ?10:40
X-FadeI'm not against having experimental qtm in extras-devel as long as it is named qtm-experimental and linked against qt4-experimental ;)10:40
sandman_we should really only have 4.6 stable and 4.7 snapshot packages after PR1.210:40
sandman_not a mix of 4.6-stable, 4.6-snapshot and 4.7-snapshot10:41
sbradleysandman_: I believe 4.6 snapshots will be stopped once 4.7 is out10:41
X-FadeYes, so qt4-maemo5 will be removed.10:42
sbradleyand agreed, we need clear "experimental" type naming. it's a bit confusing at the moment10:42
X-FadeThat leaves us with qt4 in Pr1.2 (4.6.2) and qt4-experimental in -devel (4.7-snap)10:42
kate_When Qt4.6 and Qt4.7 are binary compatible, is there any reason make QtM linked Qt4.7 or should we handle issue with something LD_LIBRARY_PATH etc10:42
sandman_X-Fade: Qt is decided - the question is now QtM10:43
X-Fadekate_: Problem is more on the debian package side (dependencies)10:43
kate_If we do so, one version of QtM should work with both Qt4.6 and 4.710:43
kate_The debian dependencies can be for Qt4.6 but if you would like use Qt4.7-experimental, set LD_LIBRARY_PATH10:44
sbradleyif the 4.7 packages have a different name we could just use an OR dependency, so 4.6 or 4.7 work10:44
sandman_sbradley: i guess QtM 4.6 will be released when Qt 4.7 is released, right?10:44
kate_I think that QtM release will be very soon, for PR1.210:44
sandman_sbradley: kate_: Qt 4.6 and 4.7 will be binary compatible and will both be called libqt4-*10:44
achipasbradley: btw that was how I destroyed extras with PyQt :)10:44
sbradleyyeah the first (non-beta) QtM release is happening soon10:44
X-FadeThen let's see if we can get that QtM release in nokia-apps.10:45
kate_And first official QtM will be for 4.6 and it will live in extras ?10:45
kate_Then when we got PR1.3 it will be part of the system10:45
sandman_kate_: no - in nokia-apps10:45
sandman_it's provided and supported by Nokia after all10:46
sandman_let me summarize a bit:10:48
sandman_1) remove qt4-maemo5 (4.6) after PR1.210:48
sandman_2) upload Qt 4.7 snapshots as qt4-experimental to e-d afterwards10:48
sandman_3) as soon as 4.6 QtM is released, get those packages to nokia-apps10:49
sandman_4) remove 4.6 QtM from extras devel10:49
sandman_5) maybe upload new QtM packages to e-d, but only for qt4-experimental (4.7) and with 'experimental'in the pacakge name10:50
sandman_6) PROFIT :)10:50
sbradleyhehe10:50
achipaX-Fade: would it be more good or more harm if we had a sort of provides: qt4-maemo5 in the qt4-experimental packages ?10:50
X-FadeI like that list.10:50
achipa(to help the transition)10:50
sbradleyyeah, sounds good10:51
kate_And QML will be only with Qt4.7-experimental10:51
sbradleywhat's the process for getting stuff into nokia-apps?10:51
sandman_achipa: makes sense10:51
sandman_kate_: yes10:51
X-Fadesbradley: I'll have to find out who is responsible for that10:51
achipamight I add a 7) recommend bindings to follow the same nomenclature10:52
sandman_achipa: you might ;) /me always forgets about binding10:52
achipa(so that we don't have a 'pyside' package that depends on experimental, as that defeats the whole point of clear naming)10:52
*** tekojo has joined #maemo-meeting10:52
sandman_achipa: someone needs to forward that to the pyside/pyqt maintainers though10:53
X-Fadetekojo: Do you know who is responsible for nokia-apps?10:53
tekojoThe repo?10:53
X-Fadeyep10:53
achipasandman_: I *am* the pyqt maintainer for Maemo ;)10:53
tekojoI can find out10:53
sandman_achipa: i can never remember if you're the pyqt or pyside maintainer, so I made the statement generic :P10:54
achipasandman_: will forward the conclusions to PySide folks :)10:54
achipa(though the binding maintainer pains are bigger as we can't get away with a simple ldconfig/path trick :( :) )10:55
achipaone more question, rather mundane, but one I got a LOT of requests for/about10:56
achipaQMaemo5HomescreenAdaptor10:57
achipa(for Maemo desktop widgets)10:57
sandman_what's wrong with it?10:57
achipait's currently in e-d, but not as a lib, but as an included example class10:57
achipais it going to be included in the QMaemo5 module at some point ?10:57
sandman_achipa: it won't be included, as it is not a supported part of Qt - it's a creative friday project from Ralf, that we decided to polish up and publish for everyone as an OS project on gitorious10:58
achipathe problem is twofold - one, people include it and forget about it - so we can't provide any updates should they be necessary - and that can mean breakage10:59
sandman_achipa: I'm not sure about the lib/example packagin - I need to talk to harryF and Ralf first10:59
sandman_achipa: let's talk about that on #qt-maemo this (european) afternoon11:00
achipasandman_: okay11:00
X-Fadesbradley: I will follow up on the nokia-apps part.11:01
sandman_so - is everybody ok with my summary?11:01
sbradleyX-Fade: thanks11:01
sbradleysandman_: yes11:01
kate_OK11:01
X-Fadesandman_: yes from my part.11:01
achipasandman_: yes11:01
tekojoto get into nokia-apps the packages need to go to the internal integration queue11:01
sandman_ok - then I guess we can get back to coding :)11:02
kate_Remember that QtM has some extra dependencies, libtpsession and modest-plugin that lives in extras-devel11:02
X-Fadekate_: Those would need to move then too.11:03
sandman_QtM stable for 4.6 or experimental QtM for 4.7?11:03
kate_Yes, i just wanted remind11:03
kate_QtM stable, they are adaptation layers needed for messaguing11:03
sbradleyactually yeah, there's also those bearer dependencies that I believe we don't want to officially support11:04
achipaX-Fade: you okay with the temporary Provides: libqt4-maemo5 thing ?11:04
X-Fadeachipa: If that really helps?11:04
sbradleylibconnsettings, libconninet and libicd-network-wlan-dev11:04
achipaX-Fade: well at least we don't insta-break EVERYTHING qt related in extras-devel11:05
achipaX-Fade: as qt4.7 is supposed to be binary compatible and all11:05
* sandman_ is off - I have another meeting11:05
X-FadeThanks all!11:06
* achipa thanks & off11:06
sbradleythanks11:07
*** tekojo has left #maemo-meeting11:14
*** achipa has quit IRC12:56
*** kate_ has quit IRC13:04
*** kate_ has joined #maemo-meeting13:06
*** GNUton-BNC has joined #maemo-meeting13:30
*** dneary has joined #maemo-meeting19:06
*** dneary has quit IRC19:50
*** dneary has joined #maemo-meeting19:51
*** VDVsx has joined #maemo-meeting20:04
*** dneary has quit IRC21:50
*** dneary has joined #maemo-meeting21:52
*** dneary has quit IRC21:57
*** VDVsx has quit IRC23:33

Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!