Created
June 27, 2012 17:46
-
-
Save robxu9/3005641 to your computer and use it in GitHub Desktop.
27 June 2012 Trinity Meeting
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
[11:00] <@[wcm]> == everything below here will be logged (11:00 AM EST) == | |
[11:00] <MutantTurkey> morning folks | |
[11:01] <@Xu_R> o/ | |
[11:01] <MutantTurkey> I see it's the usual packed house >_> | |
[11:01] <@Xu_R> <_< more people would be nice... | |
[11:01] == schreiberstein [~schreiber@p57BCE34B.dip.t-dialin.net] has joined #trinity-desktop-meeting | |
[11:02] == SlavekB [~slavek@mail.axis.cz] has joined #trinity-desktop-meeting | |
[11:02] == samelian [~same@89.114.234.9] has joined #trinity-desktop-meeting | |
[11:02] <MutantTurkey> Xu_R: just gotta remind em | |
[11:02] <eliddell> Heh, and here I thought his comment had conjured them up. ;) | |
[11:02] <@Xu_R> I bumped the email yesterday :/ hopefully people show up :P | |
[11:04] <samelian> Xu_R: next time please use GMT time :) | |
[11:04] <samelian> I can't remeber your timezone :) | |
[11:04] <MutantTurkey> really should use UTC right? | |
[11:04] <@Xu_R> samelian: yup, sorry, was a spur of the moment email... it's 1600 GMT ;P (too late) | |
[11:04] <MutantTurkey> start at 11:10 ok? | |
[11:05] <samelian> well, is very good hour for me | |
[11:05] <Z_God> yeah same here, had to look up it was UTC-5 (added it to the etherpad then) | |
[11:05] <samelian> there is 18:05 | |
[11:05] <@Xu_R> MutantTurkey: kk | |
[11:05] <@Xu_R> ah, ok. I tried to find a good time for everyone | |
[11:05] == kb9vqf [~kb9vqf@edge.pearsoncomputing.net] has joined #trinity-desktop-meeting | |
[11:05] == mode/#trinity-desktop-meeting [+o kb9vqf] by ChanServ | |
[11:06] <@kb9vqf> Hello all | |
[11:06] <@Xu_R> kb9vqf: long time no see :) | |
[11:06] <@kb9vqf> :) | |
[11:06] <MutantTurkey> the white wizard approaches | |
[11:06] <schreiberstein> ^^ | |
[11:06] <@Xu_R> indeed, he does | |
[11:06] <@kb9vqf> I figured I should check in on this little thing I am in charge of ;-) | |
[11:07] <@Xu_R> ;) | |
[11:07] <MutantTurkey> good turnout - though we are missing our prime bug tester Darrell | |
[11:07] <@Xu_R> and David as well | |
[11:07] == Xu_R changed the topic of #trinity-desktop-meeting to: http://trinity.etherpad.trinitydesktop.org/48 | |
[11:07] <@kb9vqf> speaking of bugs, has everyone seen the (old) news on the TDE front page? | |
[11:08] <MutantTurkey> he's posted to the ArchLinux list looking for package hosting - hopefully that means he is almost complete | |
[11:08] <MutantTurkey> kb9vqf: did you just post that? | |
[11:08] <@Xu_R> The etherpad is in the topic | |
[11:08] <@kb9vqf> nope, look at the date | |
[11:08] <MutantTurkey> seems that I missed that update | |
[11:08] <@Xu_R> saw that already :P | |
[11:09] <MutantTurkey> If i do say so, the website works pretty damned nice | |
[11:09] <@Xu_R> indeed it does. | |
[11:09] <@kb9vqf> Darrell also recently updated the About page | |
[11:10] <MutantTurkey> do we have an FAQ yet? | |
[11:10] <@kb9vqf> Only on the Etherpad IIRC | |
[11:10] <MutantTurkey> my bugzilla-foo is reporting that over 213 bugs have been marked resolved since the release of 3.5.13 | |
[11:11] <@Xu_R> ok, I'm starting the meeting right now then. Remember, we have no meetbot, so make sure to mark important items on the etherpad | |
[11:11] <@kb9vqf> Bugs are a major issue still | |
[11:11] <@Xu_R> and I'll grab the logs from webchat ([wcm]) | |
[11:11] <@kb9vqf> but yes, good progress has been made | |
[11:12] <@kb9vqf> The reason I referenced the TDE home page is for the link in that news item--it brings up a report on bug status for all of TDE | |
[11:12] <@Xu_R> I'm going to go straight down the etherpad list. Stop me if there needs to be any discussion. | |
[11:12] <MutantTurkey> kb9vqf: I think regular updates are important on the homepage, to keep people clued' in | |
[11:12] <@Xu_R> ah, I see the link. *notes on etherpad* | |
[11:13] <@Xu_R> (that's a long link >.>) | |
[11:13] <@Xu_R> Ok. First topic: Qt->TQt conversion. | |
[11:14] <@Xu_R> See if there are any 3rd party packages impacted by this change (see relevant mailing list thread) | |
[11:14] <@Xu_R> Bringing some apps into the tree because of TQt3? | |
[11:15] <@Xu_R> Brainstorming how to identify and handle apps that need Qt3 | |
[11:15] <@Xu_R> Any comments so far? | |
[11:15] <MutantTurkey> is this even relevant? as long as everyone has Qt3 installed for those 3rd party applications - there shouldn't be a problem right? | |
[11:16] <@kb9vqf> Right | |
[11:16] <Z_God> I wonder how non-TQt Qt3 applications look and integrate in TQt-Trinity | |
[11:16] <MutantTurkey> they should be the same | |
[11:16] <MutantTurkey> afaik? | |
[11:16] <@kb9vqf> AFAIK they would only have native Qt3 themes available | |
[11:17] <eliddell> Thing we might have to watch for is API break between last Nokia QT3 and 3.3.8d | |
[11:17] <MutantTurkey> but it would be the same even if you linked it against TQt3 | |
[11:17] <@kb9vqf> The major deal breaker here seems to be QCad; when was the last release of QCad for Qt3? | |
[11:17] <@kb9vqf> i.e. was it already abandoned? | |
[11:17] <MutantTurkey> last release was 2012, but I think that is Qt4 | |
[11:17] <Z_God> as for Qt3 apps, I still use Twinkle here | |
[11:18] <@kb9vqf> In that case someone should contact them and see if they have dropped all Qt3 support | |
[11:18] <@kb9vqf> Personally I expect the third-party Qt3 app issue to go away in the next few years | |
[11:18] <eliddell> Listed release for qcad in kde-sunset is 2.0.5 | |
[11:18] <MutantTurkey> kb9vqf: because no-one will maintain it? | |
[11:18] <@kb9vqf> pretty much | |
[11:19] <@kb9vqf> or that there will be an equivalent Qt4 app available by that time | |
[11:19] <@kb9vqf> I initiated TQt3 to solve major problems with keeping Qt3 and Qt4 on the same system | |
[11:19] <@kb9vqf> so far that has worked very well | |
[11:19] <@kb9vqf> IMHO more effort should go into polishing the Qt3-->Qt4 theme engine | |
[11:20] <@kb9vqf> that way Qt4 apps will look and act native on TDE | |
[11:20] <@kb9vqf> (as much as possible anyway!) | |
[11:21] <Z_God> otherwise they could just be compiled against TQt I suppose | |
[11:21] <@kb9vqf> There is some porting required (essentially just renaming headers and object names) | |
[11:22] <Z_God> is there a guide for that? | |
[11:22] <@kb9vqf> other than that, yes, they could just be built against TQt3 | |
[11:22] <Z_God> Qt3->TQt3 porting guide or smth like that? | |
[11:22] <@kb9vqf> There are some scripts in GIT, but they are probably out of date | |
[11:22] <Z_God> there should be a webpage with step by step info imo | |
[11:22] <@Xu_R> it should essentially be putting t in front of most names asnd headers, right? | |
[11:22] <@kb9vqf> The guidelines are: All QObject, QWidget, (Q objects) need to become TQObject, TQWidget, etc. | |
[11:23] <@kb9vqf> tqt.h should be included from the command line for gcc | |
[11:23] <@kb9vqf> and yes, static qt variables are now tqt_* instead of qt_* | |
[11:23] <@kb9vqf> Fairly simple, but tedious to get 100% right | |
[11:23] <@kb9vqf> I do agree with Z_God, a web page should be put up with this information | |
[11:24] <@kb9vqf> Much of it can be extracted from tqt.h | |
[11:24] <@Xu_R> I've noted that on the Etherpad for now. It needs to be moved to the wiki. | |
[11:24] <@kb9vqf> yes | |
[11:26] <@Xu_R> ok. anything else before I keep moving down the list? | |
[11:26] <@kb9vqf> any questions about TQt3 | |
[11:26] <@kb9vqf> ? | |
[11:27] <@kb9vqf> if not go ahead to the next topic :) | |
[11:28] <@Xu_R> ok... Discuss whether an install can be made with a base OS and TDE without any underlying apps pulling in Qt4.. | |
[11:28] <@Xu_R> ^ we may want to skip this for now and come back for it later... | |
[11:28] <@kb9vqf> Xu_R: It is of course theoretically possible | |
[11:28] <@kb9vqf> the LiveCDs would do this for example | |
[11:29] <@Xu_R> It's all packaging in the end :) | |
[11:29] <@kb9vqf> right | |
[11:29] <@kb9vqf> the quality of the packaging, specifically | |
[11:29] <@Xu_R> yup. | |
[11:29] <@Xu_R> ok. 3.5.13.1 progress (Slavek's backporting work :D) | |
[11:30] <@Xu_R> SlavekB: Status update on how your PPA is going? | |
[11:30] <SlavekB> Ok, I'm ready. | |
[11:30] <@Xu_R> Go ahead. :) | |
[11:31] <SlavekB> Most of the patches I have already incorporated into the v3.5.13-sru branch. | |
[11:32] <@Xu_R> great! :D | |
[11:32] <SlavekB> Still working on the latest packages for which are the "Fix inadvertent" TQ "changes" or "GCC 4.7 fixes." | |
[11:33] <MutantTurkey> is this going to be a real release? | |
[11:33] <@kb9vqf> That is my intention | |
[11:33] <MutantTurkey> ok awesome | |
[11:33] <@kb9vqf> I'm just waiting for all the bugs to be worked out :) | |
[11:33] <@Xu_R> http://trinity.etherpad.trinitydesktop.org/16 etherpad for the SRU | |
[11:33] <MutantTurkey> SlavekB: what bugs are we waiting on? | |
[11:33] <Z_God> are there bugs? | |
[11:33] <@Xu_R> SlavekB: are there still problems, even on gcc 4.7.1? | |
[11:33] <@kb9vqf> Every now and again I hear reports of regressions from the mailing lists | |
[11:34] <Z_God> but that's about the nightlies mostly I suppose | |
[11:34] <SlavekB> I have no where to test 4.7.x | |
[11:34] <@kb9vqf> 3.5.13 won't build on 4.7.1 | |
[11:34] <@kb9vqf> or any gcc >= 4.7.x | |
[11:34] <SlavekB> So, incorporate patches. :) | |
[11:34] <Z_God> or use an older compiler for the release | |
[11:34] <@kb9vqf> Z_God is right | |
[11:34] <@Xu_R> oic. | |
[11:34] <@Xu_R> yes, then force an older compiler for now. | |
[11:34] <@kb9vqf> too much patching and we risk introducing bugs again | |
[11:35] <@kb9vqf> R14 is where the build problems are being addressed | |
[11:35] <@Xu_R> ok | |
[11:35] <SlavekB> In any case - all the packages I have also tested on Precise. | |
[11:35] <Z_God> will there be mention of this release as R13.1 or smth somewhere as well or just 3.5.13.1? | |
[11:35] <MutantTurkey> will 3.5.13.1 have a precise release? | |
[11:35] <MutantTurkey> because people have been asking | |
[11:35] <SlavekB> Yes, I plan to prepare it. | |
[11:36] * kb9vqf doesn't know how well that will work, as 3.5.13 won't build on Precise due to many system library changes | |
[11:36] <@kb9vqf> But you can definitely try :-) | |
[11:36] <Z_God> 3.5.13 oneiric installs with some hacks on precise | |
[11:36] <MutantTurkey> 6 months is too short of a cycle. | |
[11:37] <@kb9vqf> My thoughts exacrl | |
[11:37] <@kb9vqf> *exactly | |
[11:37] <@Xu_R> kb9vqf: I haven't kept up with what Ubuntu's done; what system library changes? :/ | |
[11:37] <MutantTurkey> instead of focusing on improving and stabilizing the current release they focus on the next one... | |
[11:37] <SlavekB> All packages mentioned on SRU etherpad I have already tested on Precise. | |
[11:37] <@kb9vqf> Good! | |
[11:37] <@kb9vqf> I had run into problems with the stock 3.5.13 builds on Precise | |
[11:37] <@kb9vqf> looks like you fixed them | |
[11:37] <Z_God> what would be needed for a live-cd with precise & trinity? | |
[11:37] <SlavekB> I have a few fixes for build-deps | |
[11:38] <Z_God> it would be the new LTS ubuntu with trinity, I'd really like to see that | |
[11:38] <@kb9vqf> I have an automated CD builder mostly developed here, but I am waiting for the newest LibreOffice release with TDE support (3.6) | |
[11:38] <MutantTurkey> me too - i intend to upgrade when support is available | |
[11:38] <@kb9vqf> MutantTurkey: If you look at our past releases we seem to have settled on a ~1 year release cycle | |
[11:38] <MutantTurkey> kb9vqf: that is in it's 2nd beta correct? | |
[11:38] <MutantTurkey> kb9vqf: that seems more sane | |
[11:38] <Z_God> kb9vqf: great! that'll be fantastic | |
[11:39] <@kb9vqf> Honestly I haven't kept up to date with the release schedule of LO, but when 3.6 hits Debian Wheezy I can build a LiveCD. | |
[11:39] <MutantTurkey> kb9vqf: we want to support 12.04 for as long as it's supported right? | |
[11:39] <@kb9vqf> Yes | |
[11:39] <MutantTurkey> though as long as slackware is around we will be supporting stuff in the dark ages :p | |
[11:39] <@kb9vqf> Note that we still support Lucid | |
[11:39] <@kb9vqf> and it is many years old now | |
[11:39] <MutantTurkey> right | |
[11:39] <Z_God> but still supported | |
[11:39] <Z_God> I still install lucid with trinity now on new boxes... | |
[11:40] <@kb9vqf> I am seeing very little incentive to upgrade past Natty due to GTK 3.0 | |
[11:40] <MutantTurkey> what is the trouble with that? | |
[11:40] <@kb9vqf> Ugly as all get out | |
[11:40] <MutantTurkey> I mean other than the obvious pita of dealing with it. | |
[11:40] <@kb9vqf> yes, that | |
[11:40] <@kb9vqf> :-) | |
[11:40] <MutantTurkey> do we do a gtk3 theme engine and picker? | |
[11:41] <@kb9vqf> No one has developed either of those for any toolkit | |
[11:41] <MutantTurkey> the problem is we need a better way to do this than preloading kgtk.so | |
[11:41] <@kb9vqf> i.e. Qt4 doesn't have it either | |
[11:41] <MutantTurkey> kde seems to have just written native gtk themes that match the KDE4 default theme. | |
[11:41] <Z_God> I think the focus should be on Qt4 theming for the R14 release | |
[11:41] <@kb9vqf> GTK3.0 has something like three working themes and one of those is Oxygen | |
[11:41] <@kb9vqf> Z_God is correct | |
[11:41] <Z_God> and GTK3 maybe for a later release | |
[11:41] <@kb9vqf> Dunno...GTK3 is a mess | |
[11:42] <MutantTurkey> yep | |
[11:42] <@kb9vqf> I won't do it, but if someone wanted to I would welcome the effort ;-) | |
[11:42] <MutantTurkey> kb9vqf: is it worth mentioning qt5? | |
[11:42] <@kb9vqf> I know about Qt5 | |
[11:42] <@Xu_R> I think we can deal with GTK3 later. Qt (any version) is what we should stay with. | |
[11:42] <MutantTurkey> i know about it too... do we have any course of action for it? | |
[11:42] <@Xu_R> Qt5... I should hope it's not as much a change as Qt3->Qt4 was. | |
[11:43] <@kb9vqf> If Nokia goes fully to QML (i.e. dumps widgets entirely) then TQt3, Qt4, and GTK 2/3 will be the only plaforms for developing native desktop applications | |
[11:43] <@kb9vqf> Scary, huh? | |
[11:43] <Z_God> is Qt still done by Nokia? | |
[11:43] * kb9vqf refuses to program complex desktop apps in....javascript | |
[11:43] <Z_God> I think they sold parts and sacked loads of people | |
[11:43] <MutantTurkey> I totally agreee wit hthat. | |
[11:44] <@Xu_R> kb9vqf: I still find QML + JS ridiculous... but what can you do >.< | |
[11:44] <MutantTurkey> with 'that | |
[11:44] <@kb9vqf> Not use it? | |
[11:44] <@kb9vqf> :) | |
[11:44] <@Xu_R> true ;) | |
[11:44] <@kb9vqf> Z_God brings up a good point; Qt is now its own project loosely tied to Nokia | |
[11:44] <MutantTurkey> kb9vqf: you can still hard code it though right? | |
[11:44] <@kb9vqf> As of now, yes | |
[11:44] <Z_God> I don't know if more people could communicate with Qt devs from here | |
[11:44] <@kb9vqf> Qt6 may not | |
[11:44] <Z_God> I added some on G+ they seem friendly | |
[11:44] <MutantTurkey> Z_God: well looks can be misleading | |
[11:45] <Z_God> so far they were also polite when I pointed out a problem | |
[11:45] <@kb9vqf> As long as they keep QWidget and friends we can theoretically use Qt4 for smaller applications and theme it | |
[11:45] <Z_God> although minor | |
[11:45] <Z_God> problem^ | |
[11:45] <MutantTurkey> and is there any course of action for network-manager? (sorry while were are talkin about deprecated items) | |
[11:45] <MutantTurkey> nm seems to have gone the way of the crapper with its quality | |
[11:45] <@kb9vqf> If they go fully to QML then we will have at least three incompatible toolkits with different theming :-/ | |
[11:45] <@Xu_R> :/ | |
[11:45] <@kb9vqf> nm stinks right now | |
[11:46] <MutantTurkey> wicd works well! | |
[11:46] <Z_God> nm-applet works fine here, it's just GTK3 | |
[11:46] <@kb9vqf> nm-applet won't even show its tray icon unless you run Unity | |
[11:46] <samelian> MutantTurkey: i plan to make a NM client from scratch | |
[11:46] <MutantTurkey> kb9vqf: yeah what the heck is that? | |
[11:46] <MutantTurkey> samelian: really? | |
[11:46] <samelian> when I will done with polkit | |
[11:46] <samelian> yes | |
[11:46] <MutantTurkey> okers | |
[11:46] <MutantTurkey> okay' | |
[11:46] <@kb9vqf> cool. Will you use a UI similar to knetworkmanager? | |
[11:47] <@kb9vqf> (many users here loved the old knetworkmanager UI) | |
[11:47] <samelian> not sure yet, I found the KNM a little unfriendly | |
[11:47] <samelian> well, we can discuss this | |
[11:47] <@kb9vqf> Its power was in the two-clicks-to-connect idea | |
[11:47] <Z_God> kb9vqf: I wish I could track Qt5 development, but I really do not have enough technical knowledge, it would be good if there would be more interaction though from the Trinity side | |
[11:47] <@kb9vqf> The wizards were not so good | |
[11:47] <samelian> I want to made something more controllable | |
[11:48] <MutantTurkey> something with good vpn support would be nice | |
[11:48] <samelian> sure, I intent to implement completely support | |
[11:48] <@kb9vqf> Z_God: I am still a bit unhappy with Qt >= 4; last I checked they were STILL shipping bugs in the latest Qt4 release. At this point release that is inexcusable. | |
[11:48] <samelian> I will follow closely the NM APO | |
[11:48] <samelian> API | |
[11:48] <@kb9vqf> be warned it will probably change again... | |
[11:49] <Z_God> kb9vqf: are the bugs reported in their tracker or are the kind of considered features? | |
[11:49] <Z_God> they* | |
[11:49] <@kb9vqf> I was waiting to do something similar until NM hit 1.0 | |
[11:49] <@kb9vqf> but if you want to do it please go ahead :-) | |
[11:49] <Z_God> the Qt devs hang out on IRC, they don't mind discussing things it seems | |
[11:50] <@kb9vqf> Z_God: I haven't used Qt4 in some time, but the bugs I am referring to are in their bugtracker, as those developers affected by them complain loudly about them ;-) | |
[11:50] <@kb9vqf> I don't think I was clear though | |
[11:50] <Z_God> kb9vqf: ok, but do you think you could somehow steer Qt5 development a bit or at least somehow research could be put into it? | |
[11:50] <@kb9vqf> The major problem I have with Qt4 is that it releases very, very quickly | |
[11:51] <MutantTurkey> kb9vqf: what is the status of with Qt4 and trinity? | |
[11:51] <Z_God> it seems it'll be important to have good knowledge on Qt5 in any case either because it will somehow be useful or to be able to explain why not | |
[11:51] <@kb9vqf> Z_God: Possibly. Honestly, I don't have much time, but if they at least keep the widget class framework Qt5 will still be usable | |
[11:51] <@Xu_R> we have theming between TQt3 and Qt4 now, correct? | |
[11:51] <@kb9vqf> Xu_R: This ties in with my previous comment about Qt4's fast release cycle | |
[11:52] <@kb9vqf> Theming worked on Qt 4.7, but gives foolishly large widgets on Qt 4.8 | |
[11:52] <Z_God> kb9vqf: yeah, I understand time is the issue, I will see of I can ask them about this in particular | |
[11:52] <@kb9vqf> I haven't had a chance to see what changed from 4.7 to 4.8 ye | |
[11:52] <@kb9vqf> *yet | |
[11:52] <@kb9vqf> Z_God: I know they are really trying to push QML as the Next Best Thing | |
[11:52] <Z_God> I think Qt5 will be relatively similar to Qt4 regarding applications and porting source | |
[11:53] <@kb9vqf> Yes, they stated that | |
[11:53] <Z_God> they mainly want to change the back-end I understand to be able to use newer GPUs properly | |
[11:53] <@kb9vqf> the question is, if you are not using QML, will you still get any attention in the bugtracker? | |
[11:53] <@Xu_R> :/ | |
[11:53] <@kb9vqf> we saw this with KDE4 | |
[11:53] <Z_God> I don't know how it will compare to Qt3 regarding speed on such systems for example | |
[11:54] <@kb9vqf> Z_God: It will probably be faster, but remember that the toolkit was trimmed down to do this | |
[11:54] <Z_God> I can imagine GPU acceleration may be critical for certain things on newer devices | |
[11:54] <MutantTurkey> it won't be faster | |
[11:54] <@kb9vqf> I personally get nervous about relying on a GPU | |
[11:54] <samelian> kb9vqf: I don't understand why you insist to mix Qt3 with Qt4 | |
[11:54] <MutantTurkey> why do you need a crazy GPU do draw basic widgets? | |
[11:54] <MutantTurkey> samelian: hasn't that been answered to death? | |
[11:55] <@Xu_R> o_o GPU for that.. that doesn't make sense at all. | |
[11:55] <samelian> yeah, forget it :) | |
[11:55] <@kb9vqf> samelian: Like it or not Qt4 is here to stay on Linux. We need to do what we can to not step all over each others toes | |
[11:55] <Z_God> MutantTurkey: because your CPU sucks, eat more battery, etc. | |
[11:55] <samelian> well, there are a lot gui toolkits, but we don't care about they | |
[11:55] <@kb9vqf> Very few people have any idea of how to make a GPU though. | |
[11:55] <samelian> Qt4 is not the successor of Qt3, is something completely different | |
[11:56] <@kb9vqf> It seems to be going against the FOSS mindset somewhat to rely on a black box that almost no one can recreate | |
[11:56] <@kb9vqf> just my $0.02 | |
[11:56] * kb9vqf notes that Linux runs on fully open CPU hardware | |
[11:56] * kb9vqf also notes that there is no open GPU anywhere | |
[11:57] <@Xu_R> :/ | |
[11:57] <@kb9vqf> Seems we have taken nVidia's generosity a bit too far... :-) | |
[11:57] <Z_God> makes sense, it would be good to have articles on this on the trinity site | |
[11:57] <MutantTurkey> we do have open intel drivers and such | |
[11:57] <@kb9vqf> Drivers, yes. Hardware, no | |
[11:57] <MutantTurkey> right | |
[11:57] <@kb9vqf> Show me even a basic GPU in an FPGA that can run Qt4/5 and I'll shut up :-) | |
[11:58] <@kb9vqf> s/GPU/open GPU/ | |
[11:58] <@kb9vqf> Anyway, I think we are fairly off topic at this point | |
[11:58] <Z_God> yes, this could make Trinity a critical part of a free system | |
[11:58] <samelian> instead to waste our time with Qt4, would be much useful to add glib event loop support to Qt3 | |
[11:58] <@kb9vqf> glib? why? | |
[11:59] <samelian> for example, my last conflict with creator of polkit :) | |
[11:59] <@kb9vqf> do you have a patch for it? | |
[11:59] <@kb9vqf> if so I would consider adding it | |
[11:59] <@kb9vqf> I assume that this only allows compatibility with glib, and does not make Qt3 *rely on* glib? | |
[11:59] <samelian> nope, I this moment I'm trying to write a polkit client using pure dbus calls | |
[11:59] <samelian> but is very hard | |
[11:59] <samelian> because is poor documented | |
[11:59] == SlavekB [~slavek@mail.axis.cz] has quit [Read error: Connection reset by peer] | |
[11:59] <@kb9vqf> right | |
[12:00] * kb9vqf hates DBUS | |
[12:00] <samelian> and David won't to support this effort | |
[12:00] <@kb9vqf> the idea was OK, but the implementation stinks | |
[12:00] <samelian> also, qt3 dbus binding is really sux | |
[12:00] * kb9vqf knows | |
[12:00] <samelian> a lot small but sever bugs | |
[12:00] <@kb9vqf> are they in our bug tracker? | |
[12:00] <@kb9vqf> I can't fix what I don't know about :-) | |
[12:01] <samelian> nope, I made patches in my kde3 fork | |
[12:01] == SlavekB [~slavek@mail.axis.cz] has joined #trinity-desktop-meeting | |
[12:01] <@kb9vqf> we should incorporate those patches | |
[12:01] <@kb9vqf> can you send a link? | |
[12:01] <samelian> sure, I will prepare a list | |
[12:01] <@kb9vqf> thanks! | |
[12:02] <samelian> some bugs affect my udisks2 backend, and a new discovered one affect polkit client | |
[12:02] <@kb9vqf> samelian: So this doesn't get lost in the development process, can you file a BLOCKER bug on the bug tracker with the patches attached? | |
[12:02] <samelian> well, I think is not a blocker actually, because trinity do not use udisks2 | |
[12:03] <@kb9vqf> No, but I am using Blocker to tag things that block the next release (R14), and I want the bugs fixed | |
[12:03] <samelian> maybe will be a blocker when my polkit agent will be integrated into trinity | |
[12:03] <samelian> aha, ok then | |
[12:03] <@kb9vqf> Interestingly enough we seem to be working on some of the same things, but with two completely different approaches | |
[12:03] <@kb9vqf> this is a good thinkg | |
[12:03] <@kb9vqf> *thing | |
[12:03] <samelian> of course :) | |
[12:04] <@Xu_R> open approaches are good | |
[12:04] <samelian> as I said on ML, is better two solutions than none | |
[12:04] <@kb9vqf> right | |
[12:05] <SlavekB> Can I have some notes for 3.5.13.1? | |
[12:05] <SlavekB> (I've been away a while...) | |
[12:05] <samelian> my intent is to modernize kde3 to support latest technologies | |
[12:06] <@Xu_R> SlavekB: you don't have to ask, just state outright what you want to do with 3.5.13.1 :) | |
[12:06] <samelian> I think with systemd we will be able to control the OS directly from KDE3/Trinity | |
[12:06] <samelian> like system services, etc | |
[12:06] <MutantTurkey> sigh systemd sucks | |
[12:06] <MutantTurkey> not all systems use it | |
[12:06] <samelian> well... | |
[12:06] <@kb9vqf> samelian: This is my intent as well, but I am approaching it from a much lower level | |
[12:06] <@Xu_R> samelian: how do you interact with systemd? dbus? | |
[12:07] <MutantTurkey> trinity cannot depend on something that only a few systems use | |
[12:07] <samelian> Xu_R: I do not interact with systemd yet | |
[12:07] <MutantTurkey> good old initscripts have worked for almost 40 years.... and suddendly they're not enough? | |
[12:07] <@Xu_R> samelian: ah, ok. | |
[12:07] <SlavekB> A couple of errors to solution: | |
[12:07] <SlavekB> MutantTurkey yesterday reported a problem in Oneiric | |
[12:07] <SlavekB> kdm_greet hangs on reboot/shutdown from kdm - see bug 1068 | |
[12:07] <@kb9vqf> MutantTurkey: Well, upstart has decreased boot times and increased reliability, so I think it was a good idea | |
[12:07] <samelian> MutantTurkey: the idea behind systemd is to parallelize the startup | |
[12:08] <samelian> i think systemd support the old init scripts too | |
[12:08] <@kb9vqf> The major issue I have with many of these newer technologies is that the API is almost constantly changing, and also that they do not provide C bindings, instead exposing undocumented DBUS interfaces | |
[12:09] <samelian> well, I like dbus actually | |
[12:09] <@Xu_R> SlavekB: I marked that on the Etherpad :) | |
[12:09] <samelian> because is toolkit agnostic | |
[12:09] <SlavekB> language files in kaffeine - see bug 858 => perhaps related to this: Switch application language does not work at all | |
[12:09] <samelian> and permit asyncrounous operations | |
[12:09] <@kb9vqf> Well, I don't care for it personally | |
[12:09] <@kb9vqf> too bulky | |
[12:09] <@Xu_R> noted that bug. | |
[12:09] <@kb9vqf> every time I sit down and say "I want to use DBUS in this application" I regret it several days later | |
[12:09] <SlavekB> big problems with kgtk-qt3 - with LD_PRELOAD in startkde | |
[12:10] <samelian> is a little to hard to handle dbus errors, but it's ok so far | |
[12:10] <SlavekB> Xu_R: Thanks, I see it :) | |
[12:10] <@Xu_R> just wondering... is that kgtk hack bascially dead with gtk3 now? | |
[12:10] <@kb9vqf> yes | |
[12:10] <Z_God> SlavekB: maybe it would be good to disable the GTK intregration stuff is this prevents issues with for instance precise | |
[12:10] <MutantTurkey> completely | |
[12:10] <Z_God> better something ugly than broken | |
[12:11] <MutantTurkey> yeah | |
[12:11] <MutantTurkey> users can still pick regular gtk themes | |
[12:11] <MutantTurkey> so | |
[12:11] <MutantTurkey> not ap roble mthere | |
[12:11] <@kb9vqf> I would suggest detecting gtk == 2 and only enabling integration for that version | |
[12:11] <MutantTurkey> not a problem there. | |
[12:12] <eliddell> What would detection do to a system with both GTK2 and GTK3? | |
[12:12] <@kb9vqf> OK, detect GTK3 and disable integration | |
[12:13] <SlavekB> I would like to blow LD_PRELOAD from starttde / startkde. | |
[12:13] <samelian> kb9vqf: trinity's kdm is patched for consolekit support? | |
[12:13] <@kb9vqf> samelian: No idea | |
[12:13] <samelian> aha | |
[12:13] <samelian> I found a patch in fedora 9 | |
[12:13] <@kb9vqf> My point is that some folks like myself will stay on GTK 2 systems purely for the integation | |
[12:13] <samelian> and seems ok | |
[12:13] <@kb9vqf> what does this do for TDE? | |
[12:14] <samelian> for example, will working polkit agent | |
[12:14] <@kb9vqf> OK, this is a good object lesson | |
[12:14] <@kb9vqf> Nice new technology, right? Just add support and get new features? | |
[12:14] <@kb9vqf> Oops, deprecated: http://www.google.com/url?sa=t&rct=j&q=consolekit&source=web&cd=1&ved=0CFMQFjAA&url=http%3A%2F%2Fwww.freedesktop.org%2Fwiki%2FSoftware%2FConsoleKit%2F&ei=UzHrT-vrHaby2QW0-r2gAQ&usg=AFQjCNH3CKzsvsrlmkkojWygFRbErtLkXQ&cad=rja | |
[12:15] <@kb9vqf> Sorry, www.freedesktop.org/wiki/Software/ConsoleKit/ | |
[12:15] <samelian> yes, I know | |
[12:15] <@Xu_R> consolekit was deprecated and replaced with... systemd, right? >_> | |
[12:15] <@kb9vqf> Linux is only hurting itself this way | |
[12:15] <samelian> but I think the same API will be implemented in systemd | |
[12:15] <eliddell> And systemd is never going to be universal. | |
[12:15] <SlavekB> it would be good to have a package pinentry-qt / pinentry-tqt. | |
[12:15] <@kb9vqf> By the time someone implements support for something it is deprecated and thrown out | |
[12:16] <@kb9vqf> /rant | |
[12:16] <MutantTurkey> sigh it is true | |
[12:16] <@Xu_R> It does seem that way... >.> | |
[12:16] <samelian> well, we need to live with this :) | |
[12:16] <@Xu_R> /sigh | |
[12:16] <eliddell> Alas... | |
[12:16] <MutantTurkey> so the solution is to stop dealing with abstract crap | |
[12:16] <@kb9vqf> samelian: But who pays for it? | |
[12:16] <MutantTurkey> kernel doesn't change so easily | |
[12:16] <@kb9vqf> Software development is not free | |
[12:16] <samelian> developers, of course :) | |
[12:16] <MutantTurkey> deal as directly as possible | |
[12:17] <@kb9vqf> MutantTurkey is right IMHO | |
[12:17] <@kb9vqf> that's why I took the approach I did regarding the HAL replacement | |
[12:17] <MutantTurkey> and the thing is - every one of these tools uses the kernel in the end | |
[12:17] <MutantTurkey> other than dbus lol | |
[12:17] <@kb9vqf> and I will probably do something similar for network management | |
[12:17] <MutantTurkey> if we can do directly - do directly | |
[12:18] <MutantTurkey> don't do easy, do right! | |
[12:18] <samelian> but what about authorizations? | |
[12:18] <@kb9vqf> I have also found tools that directly access procfs to be much faster than those using DBUS | |
[12:18] <SlavekB> For Precise is available only pinentry-gtk and pinentry-qt4 - neither is a good choice to Trinity. | |
[12:18] <@kb9vqf> samelian: I would say to use the standard Linux permissions | |
[12:18] <samelian> these days is not enough, IMHO | |
[12:18] <@kb9vqf> if anything the run-as-root master daemon model introduces security holes | |
[12:18] <samelian> we need more granular policies | |
[12:19] <@kb9vqf> yes, but that should be done at the kernel level | |
[12:19] <@kb9vqf> not in userspace | |
[12:19] <@kb9vqf> e.g. extended ACLs in procfs | |
[12:19] <MutantTurkey> this i what I have been saying | |
[12:19] <@kb9vqf> anyway, as I said before we are working on two different approaches--time will tell which one is better/more maintainable | |
[12:20] <samelian> sure | |
[12:20] <@kb9vqf> it may be that both work just fine :-) | |
[12:20] <MutantTurkey> which one gets deprecated faster... | |
[12:20] <@kb9vqf> basically :-) | |
[12:20] <@Xu_R> ^ | |
[12:20] <@kb9vqf> but that is winning the lotto with Linux these days | |
[12:20] <@kb9vqf> Torvalds could have a bad cup of coffee and get rid of procfs tomorrow | |
[12:20] <@kb9vqf> who knows ;-) | |
[12:20] <samelian> haha | |
[12:21] <SlavekB> One more note for 3.5.13.1: how about patch "Fix mimetype detection is Incorrect, some files appear as Unknown"? | |
[12:21] <@kb9vqf> I think the general consensus was that increasing the buffer size helps somewhat, but TDE should really use the mime detection library from the system | |
[12:22] <@kb9vqf> That has not yet been done | |
[12:22] <SlavekB> And wait for it, or incorporate this patch into the 3.5.13? | |
[12:23] <SlavekB> Or leave it as it is? | |
[12:23] <@kb9vqf> Leave as-is | |
[12:23] <SlavekB> Ok. | |
[12:23] <@kb9vqf> I don't know how much of a rewrite is neeed to use the system library | |
[12:24] <SlavekB> Also, I do not know. | |
[12:24] <samelian> well, seems that this meeting do not follow the topics :) | |
[12:24] <@kb9vqf> not at all | |
[12:24] <@Xu_R> XD | |
[12:24] <@kb9vqf> let's get back on track... :-) | |
[12:24] * kb9vqf will be back shortly | |
[12:24] <@Xu_R> we're still kinda on topic, tbh | |
[12:25] <@Xu_R> 3.5.13.1... last thoughts anyone? | |
[12:25] <MutantTurkey> hip hip! | |
[12:25] <MutantTurkey> hooray | |
[12:25] <Z_God> mention R13.1 maybe | |
[12:25] <@Xu_R> but wouldn't we have to code that into Trinity? | |
[12:25] <Z_God> to clear certain confusion in advance for the next release | |
[12:25] <@kb9vqf> No renaming 3.5.13.x to R13 in this SRU | |
[12:25] <@kb9vqf> That will introduce problems | |
[12:25] <Z_God> I mean just a mention in the notes | |
[12:26] <SlavekB> What to do: Switch application language does not work at all? | |
[12:26] <@kb9vqf> OK :-) | |
[12:26] <@kb9vqf> Bug number? | |
[12:26] <Z_God> or maybe R13.0.1 | |
[12:26] <Z_God> I don't know | |
[12:26] <SlavekB> Sorry, I must just write it :( | |
[12:26] <@kb9vqf> Z_God: How about "3.5.13.1 [R13.0.1 in the new TDE version scheme]" | |
[12:27] <@Xu_R> *R13.1 | |
[12:27] <Z_God> kb9vqf: yep, exactly | |
[12:27] <@kb9vqf> Yes, Xu_R is right | |
[12:27] <Z_God> but maybe a decision needs to be made which digit is reserved for what in the version number | |
[12:27] <@kb9vqf> this is a major point release | |
[12:27] <Z_God> ok | |
[12:27] <SlavekB> Previously, we have agreed on naming 3.5.13.1 | |
[12:27] <@kb9vqf> the 'z' field in the Rx.y.z scheme is a patch level, implying no ABI breakage | |
[12:28] <@kb9vqf> SlavekB: Yes | |
[12:28] <@kb9vqf> That is what we will use | |
[12:28] <@Xu_R> officially, it is still 3.5.13.1, as this release still uses the old version system. | |
[12:28] <@kb9vqf> Z_God is just saying that we should mention what the equivalent version is in the new scheme | |
[12:28] <@kb9vqf> (in the release notes) | |
[12:28] <@Xu_R> ah, ok. that would be good as a transition for when we annouce R14. | |
[12:29] <SlavekB> In the new scheme I would see it as R13.0.1. | |
[12:29] <@kb9vqf> Why? | |
[12:29] <@kb9vqf> Is the ABI 100% compatible with 3.5.13? | |
[12:29] <SlavekB> Yes. | |
[12:29] <@kb9vqf> Oh, ok | |
[12:30] <@kb9vqf> then yes, R13.0.1 is correct | |
[12:30] <MutantTurkey> R14 R15 R16 each one represents what was to represent a single point release from 3.5.X | |
[12:30] <MutantTurkey> so 3.5.14 would be R14 and R15 would be 3.5.15 | |
[12:30] <MutantTurkey> so wouldn't just R13.1 imply a minor version bump? | |
[12:30] <@kb9vqf> Right, major revisions of the software. That implies heavy ABI and API changes | |
[12:30] <MutantTurkey> isn't that what an SRU is/ | |
[12:30] <MutantTurkey> ? | |
[12:30] <MutantTurkey> a minor version | |
[12:30] <@kb9vqf> I don't know | |
[12:31] <@kb9vqf> we need to set some rules on this I guess | |
[12:31] <SlavekB> I understand the "sru" as a "patch" version. | |
[12:31] <@Xu_R> is it just MAJOR.MINOR or is it MAJOR.MINOR.PATCH? | |
[12:31] <@kb9vqf> The latter | |
[12:31] <@Xu_R> then it's 13.0.1 | |
[12:31] <@kb9vqf> SlavekB: You have not introduced any new functionality, correct? | |
[12:32] <SlavekB> Correct. | |
[12:32] <@kb9vqf> Then 13.0.1 would seem to be most appropriate | |
[12:32] <@Xu_R> then it is indeed a patch. | |
[12:32] <SlavekB> If anything, the little things. | |
[12:32] <@kb9vqf> Honestly I don't see us incrementing the Minor version very often | |
[12:32] <@kb9vqf> but I am leaving it there in case we end up not releasing a major verison for an extended period of time | |
[12:32] <@Xu_R> ok | |
[12:33] <@kb9vqf> the Minor version would be bumped if the ABI breaks or if new functionality was added | |
[12:33] <Z_God> there have been patch releases in KDE3 history with minor features added | |
[12:33] <SlavekB> So for the R14 I would prepare R14.0.1 ;) | |
[12:33] <@kb9vqf> the Major version gets bumped only when significant development work has gone into TDE, the API, ABI have changed, and new features have been added | |
[12:33] <@kb9vqf> SlavekB: Correct | |
[12:34] <MutantTurkey> someone write that down... | |
[12:34] <@Xu_R> noted. | |
[12:34] <Z_God> they were fully backwards & forwards compatible though, I think it happened in the 3.5 serie | |
[12:34] <@Xu_R> on etherpad. | |
[12:34] <@kb9vqf> We decided not to restrict the project with that kind of compatibility | |
[12:34] <SlavekB> MutantTurkey: What about the issue that you reported yesterday? | |
[12:34] <MutantTurkey> kdesktop is not functioning properly. | |
[12:34] * kb9vqf will be back a bit later | |
[12:34] <MutantTurkey> i have not tested it in a new user though | |
[12:35] <MutantTurkey> kb9vqf: you should drop by more often... | |
[12:35] <Z_God> kb9vqf: just saying that because you wrote "or if new functionality was added" | |
[12:35] <@Xu_R> ok. 3.5.13.1 discussion finished? We'll be moving onto R14 discussion. | |
[12:35] <@kb9vqf> Z_God: Well, bumping the minor version indicates that the ABI might be broken, it does not guarantee it | |
[12:35] <SlavekB> Well - I'll try to create there a test machine with Oneiric. | |
[12:36] <Z_God> kb9vqf: ok :) | |
[12:36] <SlavekB> Still perhaps: How do I name the package? | |
[12:37] <SlavekB> I still hold the original name 3.5.13 with my ax_number_ | |
[12:37] <@Xu_R> SlavekB: will that upgrade cleanly to original name with 3.5.13.1? | |
[12:38] <SlavekB> original names contain svn revision number | |
[12:39] <SlavekB> For example: kdebase 3.5.13-0debian9+r1261450+pr19~squeeze | |
[12:39] <@Xu_R> oh... :/ yikes... what's higher than that in debian version logic? | |
[12:39] <SlavekB> now they have: 3.5.13-0debian9+r1261450+pr19~squeeze+ax16 | |
[12:39] <@Xu_R> w/o the revision... maybe include the git has somehow? | |
[12:40] * Xu_R has no suggestion - he thinks debian versions are really weird | |
[12:41] <@kb9vqf> SlavekB: Try 3.5.13-1debian* | |
[12:41] <SlavekB> So I wanted to know whether the new name can be simply 3.5.13.1-buildnumber~squezee | |
[12:41] <@kb9vqf> Introducing the point might cause a problem | |
[12:41] <SlavekB> kb9vqf: 3.5.13. >1 | |
[12:41] * kb9vqf seems to remember a way that you could test the "greater than" logic from the command line | |
[12:42] <@kb9vqf> If I remember correctly, for this to work '.' would need to be greather than '-' | |
[12:43] <SlavekB> I'll try it. | |
[12:43] <@kb9vqf> I think this is the command: dpkg --compare-versions "version1" gt "version2" | |
[12:44] <SlavekB> And what about those: r1261450+pr19 ? | |
[12:45] <@kb9vqf> you will need to incrememnt the 3.5.13-x portion | |
[12:46] <@kb9vqf> so 3.5.13-1-ubuntu-4-foobar would become 3.5.13-2-ubuntu-0-newstring | |
[12:46] <@kb9vqf> anyway, we should probably continue with the meeting topics | |
[12:46] <@kb9vqf> Xu_R: Next topic? | |
[12:46] <@Xu_R> R14 Release Goals/Progress (Going Forward) | |
[12:47] <@kb9vqf> Bugs, bugs, bugs | |
[12:47] <@Xu_R> gcc 4.7.1 seems to have fixed many bugs we had with 4.7.0 | |
[12:47] <@kb9vqf> I want to see most of the Critical bugs and all of the Blocker bugs fixed before R14 is released | |
[12:47] <@Xu_R> we need a list of packages that are building good post-rename | |
[12:47] <@kb9vqf> AFAIK all are building | |
[12:47] <@Xu_R> ok | |
[12:47] <@kb9vqf> the build failures on the list are unrelated to the rename | |
[12:48] <@kb9vqf> I have successfully built everything in the nightly builds | |
[12:48] <@kb9vqf> so we're close, but the bugtracker does need some attention still | |
[12:48] <@Xu_R> outstanding images/branding? | |
[12:48] <@kb9vqf> Darrell would know | |
[12:48] <@kb9vqf> that won't block release though | |
[12:48] <@Xu_R> that won't. | |
[12:49] <@kb9vqf> well, actually it might, but I think most of the problems have been fixed | |
[12:49] <@kb9vqf> besides, changes to images can be committed at the eleventh hour | |
[12:49] <@kb9vqf> i.e. they won't break builds | |
[12:49] <eliddell> We hope. ;) | |
[12:49] <@kb9vqf> hehe | |
[12:49] <@kb9vqf> If they do something is VERY wrong | |
[12:49] <@Xu_R> categorization of menu, svg artifact rendering strangeness | |
[12:50] <@Xu_R> Upstream Chanages with TDE impact? (aka, udev, hal, udisks2, etc) | |
[12:50] <@kb9vqf> I added some notes | |
[12:50] <eliddell> I need a copy of the SVG file that was demonstrating the artefact to trace that further. | |
[12:50] <@kb9vqf> Basically, we know HAL is on its last legs now | |
[12:50] <@kb9vqf> udev should exist with the same API (i.e. minimal impact) | |
[12:50] <@kb9vqf> udisks2 we don't rely on, but samelian will be affected by this | |
[12:51] <@Xu_R> ok | |
[12:51] <@kb9vqf> A priority should be to ship a hal-free TDE with R14.0 | |
[12:51] <SlavekB> By the way, package koffice-i18n-trinity is missing. And so k3b-i18n-trinity. | |
[12:51] <@kb9vqf> in the nightly builds? I know :-) | |
[12:51] <@kb9vqf> We NEED people to test the HAL-free versions of TDE from GIT | |
[12:51] <SlavekB> No - have never been. | |
[12:52] <@kb9vqf> oh, ok | |
[12:52] <@kb9vqf> SlavekB: File a Blocker bug report on this please, with the location of the latest i18n files for those applications | |
[12:53] <SlavekB> kb9vqf: Ok | |
[12:53] <@kb9vqf> Because of recently reported rapid HAL degredation we will need to ship a HAL-free TDE with R14.0 | |
[12:53] <@kb9vqf> This puts some pressure on the new code in GIT | |
[12:53] <@kb9vqf> and it needs to be tested/fixed | |
[12:54] <@Xu_R> which brings us to the next topic... how are we going to do a testing period? | |
[12:54] <@kb9vqf> One month minimum | |
[12:54] <@kb9vqf> It will be announced | |
[12:54] <@Xu_R> will there be a freeze and a rally for bugs? | |
[12:54] <@kb9vqf> Rally for bugs first, then freeze for testing | |
[12:54] <@Xu_R> ok | |
[12:54] <@kb9vqf> At least I think that makes sense :-) | |
[12:55] <@kb9vqf> Would anyone be willing to have a bug day soon? | |
[12:55] <@kb9vqf> I want to see those critical bugs squashed | |
[12:55] <@kb9vqf> (at least a majority of them) | |
[12:55] * kb9vqf listens to the crickets chirping in the channel | |
[12:55] <@Xu_R> I've got a TDE VM that I can test with. | |
[12:55] <@Xu_R> :P | |
[12:56] * Xu_R chirps | |
[12:56] <@kb9vqf> anyone else? | |
[12:56] <@kb9vqf> TDE can't release until those bugs are dealt with, and it will take some time for me to do so my myself | |
[12:56] <@Xu_R> :/ | |
[12:57] <@Xu_R> we should also try to get more people on IRC | |
[12:57] <@Xu_R> (I mean, look at our meeting right now >.>) | |
[12:57] <@kb9vqf> I count 24 idling on the main #trinity-desktop channel right now | |
[12:57] <@Xu_R> hm... and usually not that many active. | |
[12:58] <@kb9vqf> OK, bugs will have to be handled by me | |
[12:58] <@kb9vqf> Release may be pushed to Winter 2012, but that's not such a bad thing I guess | |
[12:58] <@kb9vqf> (I am factoring in a sizable Q/A period) | |
[12:58] <@Xu_R> ok | |
[12:59] * kb9vqf notes that means Quality Assurance, not Questions and Answers ;-) | |
[12:59] <@Xu_R> I know ;) | |
[12:59] <@kb9vqf> Next topic I guess | |
[13:00] <schreiberstein> I don't want to disturb your talk, but I can do some beta testing on wheezy if you like. Maybe it'll help to find some bugs or to test if everything works correctly without hal. | |
[13:00] <@kb9vqf> schreiberstein: That would be helpful | |
[13:00] <@Xu_R> ok... last topic... final thoughts, future milesotnes, kudos, etc? | |
[13:01] <@kb9vqf> schreiberstein: You would need to compile tdebase with TDE HW library support enabled, can you do this? | |
[13:01] * Xu_R has another meeting now, so he will not be as active on this now | |
[13:01] <@kb9vqf> schreiberstein: Also remember to install kpowersave-nohal-trinity instead of kpowersave-trinity ;-) | |
[13:02] <@kb9vqf> Regarding the last topic, we have already hit some milestones | |
[13:02] <SlavekB> Due to planned release of TDE and Wheezy I did not intend to prepare 3.5.15.1 for Wheezy. | |
[13:02] <@kb9vqf> Good idea | |
[13:02] <@kb9vqf> We now have a base toolkit (TQt3) that doesn't collide with Qt4 or future Qt versions | |
[13:02] <schreiberstein> kb9vqf: i never compiled tee or kde3 + i never built packages on debian. But i know about the build process and i know something about linux from scratch. | |
[13:02] <schreiberstein> *tde | |
[13:02] <@kb9vqf> schreiberstein: All you have to do is pull the tdebase sources with 'apt-get source tdebase-trinity' | |
[13:03] <@kb9vqf> then modify debian/rules with the required CMake flag | |
[13:03] <@kb9vqf> then 'dpkg-buildpackage -rfakeroot' | |
[13:03] <@kb9vqf> you will see a bunch of new .deb files when the build is done | |
[13:03] <@kb9vqf> install those :-) | |
[13:03] <@kb9vqf> oh, and before you do this, you may want to 'apt-get build-dep tdebase-trinity' | |
[13:04] <schreiberstein> ah ok. so the hal-free trinity is in nightly build deb-src? | |
[13:04] <@kb9vqf> The flag is -DWITH_TDEHWLIB="ON" | |
[13:04] <@kb9vqf> schreiberstein: Yes | |
[13:04] <schreiberstein> nice.. ok I'll try | |
[13:04] <@kb9vqf> be aware that it won't work unless your user has r/w access to certain files in procfs | |
[13:05] <@kb9vqf> so if it fails, try running it with tdesudo for testing | |
[13:05] <@kb9vqf> I haven't figured out the best way to set the permissions in procfs yet | |
[13:06] <schreiberstein> couldn't one simply do a chmod -R 777 /proc ? or is it different from /dev in that way? | |
[13:06] <@kb9vqf> You probably could, but that would be a Bad Idea | |
[13:06] <SlavekB> very bad idea | |
[13:06] <@kb9vqf> Unless of course you want random people snooping through your machine's RAM ;-) | |
[13:06] <@kb9vqf> You would want to set permissions for the specific device nodes that kpowersave needs | |
[13:06] <schreiberstein> it's a test machine anyway ;-) | |
[13:07] <@kb9vqf> Things like the backlight | |
[13:07] <@kb9vqf> and the CPU governers | |
[13:07] <samelian> to handle permissions maybe is better idea to use suided helpers | |
[13:07] <@kb9vqf> *governors | |
[13:07] <@kb9vqf> samelian: Possibly--I had considered suiding the entire kpowersave application | |
[13:07] <samelian> hmm, I think this is bad idea | |
[13:07] <@kb9vqf> but much better would be to use groups | |
[13:08] <@kb9vqf> samelian: Me too ;-) | |
[13:08] <@kb9vqf> schreiberstein: For testing just run kpowersave with tdesudo | |
[13:08] <@kb9vqf> We'll keep working on a good way to grant kpowersave access to the backlight and such | |
[13:08] <samelian> my idea is to write some daemons controlled via dbus | |
[13:09] <@kb9vqf> Obviously this got pushed up due to HAL failing | |
[13:09] <samelian> and the permissions will be handled by polkit | |
[13:09] <@kb9vqf> samelian: I would rather keep access controls in the kernel | |
[13:09] <schreiberstein> shall i do this on a real machine or inside a vm? because of the powermanagement | |
[13:09] <@kb9vqf> schreiberstein: If you have a laptop use that | |
[13:09] <@kb9vqf> Otherwise we also need testers for the removable device support | |
[13:10] <@kb9vqf> A real machine would be best, just so that I can get a good cross section of real-world hardware tested | |
[13:10] <schreiberstein> i don't have one.. but my 2nd pc is an amd machinee featuring an athlon x2 6000. | |
[13:10] <@kb9vqf> samelian: Correct me if I'm wrong, but can't udev set permissions for procfs on device plug? | |
[13:11] <samelian> not sure, I think it is possible | |
[13:11] <@kb9vqf> I would envision there being base system groups as there are already (audio, video, etc.), and the administrator simply granting access to the hardware by adding users to those groups | |
[13:11] <@kb9vqf> It would also prevent some random user from, for example, suspending the machine. | |
[13:11] <samelian> this is pretty sux, because you need to logout/login | |
[13:12] <samelian> to apply new permissions, I mean | |
[13:13] <@kb9vqf> Well, in my mind it's better than having a daemon suid root that can introduce new attack vectors | |
[13:13] <@kb9vqf> In my experience as well (with HAL) that daemon introduces a single point of failure | |
[13:14] <@kb9vqf> if it goes down, everything that relied on it in all sessions goes down as well | |
[13:14] <samelian> I think this is not a big issue | |
[13:15] <@kb9vqf> Besides, there are ways around having to log in/out | |
[13:15] <@kb9vqf> http://superuser.com/questions/134406/linux-refreshing-groups-without-having-to-re-login | |
[13:15] <@kb9vqf> samelian: In my experience it became a big issue | |
[13:15] <samelian> if udisks2 is killer for example, systemd will respawn it | |
[13:15] <samelian> killed* | |
[13:15] <@kb9vqf> What if it locks up? | |
[13:15] <samelian> unlikely | |
[13:15] <@kb9vqf> hehe | |
[13:15] <@kb9vqf> I'd argue with that | |
[13:16] <@kb9vqf> we should probably just continue parallel development in this area | |
[13:16] <@kb9vqf> the users can decide which method they would rather use | |
[13:16] <samelian> yes, to check the possibilities | |
[13:17] <@kb9vqf> I just remember many embarrassing moments when HAL froze on multi-user systems | |
[13:17] <@kb9vqf> Not fun | |
[13:17] <samelian> the HAL is a giant :) | |
[13:18] <@kb9vqf> Or when Network Manager failed to properly handle multiple users | |
[13:18] <@kb9vqf> Personally I think the root daemon is a hack | |
[13:18] <@kb9vqf> and that the real problem should be fixed in the kernel | |
[13:18] <@kb9vqf> e.g. add real-time permissions updating | |
[13:19] <@kb9vqf> Xu_R: I think this pretty much concludes the meeting | |
[13:19] <@kb9vqf> Thank you all for coming | |
[13:19] <@Xu_R> it does. | |
[13:19] <@Xu_R> Thanks all! | |
[13:20] == samelian [~same@89.114.234.9] has left #trinity-desktop-meeting ["Leaving"] | |
[13:20] == Xu_R changed the topic of #trinity-desktop-meeting to: done - http://trinity.etherpad.trinitydesktop.org/48 | |
[13:21] == kb9vqf [~kb9vqf@edge.pearsoncomputing.net] has left #trinity-desktop-meeting ["Konversation terminated!"] | |
[13:21] == SlavekB [~slavek@mail.axis.cz] has left #trinity-desktop-meeting ["Kopete 0.12.5 : http://kopete.kde.org"] | |
[13:22] == eliddell [~eliddell@d24-235-195-164.home1.cgocable.net] has left #trinity-desktop-meeting [] |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment