**** BEGIN LOGGING AT Sun Apr 09 09:59:56 2006 Apr 09 10:04:42 w should probably add an override like _highdpi or somthing like that Apr 09 10:05:11 CoreDump|home: isn't the dpi tied to a machine on handhelds? Apr 09 10:05:40 we used to have BIG_SCREEN and SMALL_SCREEN which was massively unsuccessfull Apr 09 10:05:47 hmm Apr 09 10:05:58 Good day Apr 09 10:06:05 morning Cockroach- Apr 09 10:06:13 let's wait for RP to reappear so he can say "machine classes" Apr 09 10:06:15 hey Cockroach- Apr 09 10:07:59 koen: problem is that once you add a new machine, you'll have to make sure to add it to all the patches needed for that machine. Would there be a way to say MACHINE xyz has HIGHDPI, a single SRC_URI_append_HIGHDPI=blahh would cover all current and future machines Apr 09 10:08:18 indeed Apr 09 10:10:25 It seems pretty undesirable that you need different files for machines with different display resolutions. It'd be much better to patch the software so that it respects some X property or something and would then work on any display. Apr 09 10:11:04 In fact, some mixture of Xft.dpi and the actual screen dpi probably conveys the right information already. Apr 09 10:12:31 well, i wasn't talking about patching software "x" to do "y" on HIGHDPI, but more selecting propper default preferences for HIGHDPI / LOWDPI. Apr 09 10:12:55 like shipping large icons or not Apr 09 10:12:58 like fontsize and so on Apr 09 10:13:01 exactly Apr 09 10:13:24 right, and my point is that you ought to be able to select the proper fontsize automatically, without needing different builds of the package. Apr 09 10:13:48 for icons, the best thing would be to put the two sets of icons in different packages and install one or the other or both depending on the MACHINE class. Apr 09 10:14:37 pb_: you can't patch each and every application to auto-detect the font-size. That's what preferences are there for in the first place. Apr 09 10:15:10 CoreDump|home: almost all applications should already be taking their fontsize from a central preferences registry anyway. Apr 09 10:15:24 indeed Apr 09 10:15:44 if there are any that don't, then you need to patch those few to either use a suitable fixed size, or respect what gconf (for example) says. Apr 09 10:15:55 i agree completly Apr 09 10:16:06 okay, so why are you asking for this extra override? Apr 09 10:16:36 however, a setting which works great in 100DPI might be unusable in 200 DPI. You have to preconfigure the centry preferences somehow Apr 09 10:16:47 central Apr 09 10:17:07 that should be done by a script (or a small helper program) that runs at login time. Apr 09 10:17:37 before login time in the case of gpe-login Apr 09 10:18:01 for fonts, the right thing to do is almost certainly to leave the application-visible font size alone and just twiddle Xft.dpi to get the result you want. That is already a per-display setting, so you can do it at session startup time with no problems. Apr 09 10:18:48 a wrong Xft.dpi caused the highly annoying "200 DPI" bug on GPE Apr 09 10:19:20 and even now, a 100DPI font-size is rather large on 200 Apr 09 10:19:27 sure, it would do Apr 09 10:19:40 I think you probably want to be setting Xft.dpi to something around 140 on a 200dpi screen. Apr 09 10:20:20 as far as I know, nobody has written any logic to do that yet. Apr 09 10:20:48 well, someone hardwired it to 100 in thepast heh Apr 09 10:21:13 indeed, because 100 is the value that works correctly (or close to correctly) for the majority of displays. Apr 09 10:21:32 no argument there.Font-size on 100DPI is perfect Apr 09 10:21:45 the 200dpi Zaurus screens are kind of an extreme case, and do indeed want slightly bigger fonts. Apr 09 10:22:03 smaller when run at 200DPI ;) Apr 09 10:22:17 ooooh Apr 09 10:22:18 http://www.eetimes.com/news/latest/showArticle.jhtml?articleID=179101800 Apr 09 10:24:31 CoreDump|home: anyway, I recommend you create a new script in Xinit.d (or Xsession.d) which inspects the current display configuration and installs an appropriate Xft.dpi property. I think that should be all you need to make the fonts work correctly for everybody. I strongly recommend that you don't start trying to make xserver-common (or gpe-session-scripts, or whatever contains that property) be a machine specific package. Apr 09 10:24:56 bbiab, heading over to the office for a while Apr 09 10:25:28 well, you got a point there. A single adjustment would be better than patches all over the place Apr 09 10:28:21 dang, ran out of space on my Z Apr 09 10:30:35 i wonder if applications would be better installed inside a loopback filesystem stored on the FAT32 drive, for cxx00 Apr 09 10:31:52 RP: danboid was asking about you, w/re to pxa270 acceleration Apr 09 10:32:19 jnc: I made a reply - if he asked further questions, I did see them as I wasn't on irc Apr 09 10:32:27 didn't Apr 09 10:32:56 ohh okay Apr 09 10:33:04 yeah i think it's straight Apr 09 10:33:44 i'm trying out oz 3.5.4.1test on c3000, it suspends and resumes the first time okay up to 20min i think Apr 09 10:34:00 after that it's just weird :/ Apr 09 10:34:48 :-/ Apr 09 10:35:24 i really wish this worked, i'm going to write the WPA howto and make that work nice Apr 09 10:41:26 oh boy Apr 09 10:42:06 one character too much in local.conf and have to do a long lasting build again :) Apr 09 10:42:11 if someone wants to buy a c3000.... Apr 09 10:42:19 suspend doesn't yet work with openzaurus Apr 09 10:42:30 but i'll give a fair price on it Apr 09 10:45:57 hmpf, 'gst-launch sinesrc ! osssink' segfaults on familiar x86-build running in chroot-jail when exiting. How discouraging :) Apr 09 12:32:13 03coredump 07org.oe.oz354x * r6a41892e... 10/packages/xserver-common/ (4 files in 2 dirs): xserver-common: Adjust Xft.dpi with a script at runtime Apr 09 12:33:06 morning Apr 09 12:34:08 hi mickeyl Apr 09 12:34:18 hi mickeyl Apr 09 12:35:13 hi zecke Apr 09 12:35:29 hey Apr 09 12:35:43 X guy: xstroke.org is gone, please update SRC_URI Apr 09 12:36:09 koen: my goal is, to not require any mails Apr 09 12:36:44 morning zecke Apr 09 12:38:13 gosh I love reading mails from Eric Shattow... Apr 09 12:39:03 jnc is right here. what did he do this time? ;) Apr 09 12:39:27 mickeyl: his monotone 0.26 mail... Apr 09 12:40:00 mickeyl: I hate it, when I feel like being treated like an idiot Apr 09 12:40:03 ah. well it's a bit on the command'y site, but i think it's just ment as an announcement for those of us who don't monitor the mt progress Apr 09 12:40:10 mickeyl: and all of his mails are written in that tone ;) Apr 09 12:41:50 * RP -> back later Apr 09 12:41:59 e-mail is just an imperfect communication medium... Apr 09 12:42:40 * zecke builds monotone now Apr 09 12:44:56 koen: looking at the feedback of the 'other SCMs' Apr 09 12:45:08 koen: nobody beside me is allowed to cast their vote :} Apr 09 12:45:40 What's good about monotone, anyway? :) Apr 09 12:45:55 polyonymous: it sounds elite and mathematical ;) Apr 09 12:45:58 you need to ask the question another way round Apr 09 12:46:01 polyonymous: seriously, it is very secure Apr 09 12:46:09 I see. Apr 09 12:46:13 pick an alternative and convince us, what's better Apr 09 12:46:25 polyonymous: it is disturbed and distributed Apr 09 12:46:35 polyonymous: it allows to have easy forks Apr 09 12:46:48 mickeyl, it's not that I object using monotone, I was really trying to understand - I see you fight a lot of its limitations and was wondering what are the benefits. Apr 09 12:47:08 the benefits are that nothing else comes close to our preferred style of work Apr 09 12:47:21 which has been heavily influenced by bitkeeper Apr 09 12:47:32 I see. Apr 09 12:47:43 mickeyl: at 'work' I have huge issues with svk... something started to be wrong Apr 09 12:47:46 Well, thank you for clarification. Apr 09 12:47:56 np Apr 09 12:48:23 * polyonymous is way too used to cvs and svn. Apr 09 12:48:27 zecke: svn (and its relatives) and me are still no friends. can't really comment why, but i don't like it Apr 09 12:48:51 mickeyl: I think it has something to do with being close to PhD Apr 09 12:49:08 zecke: heh, that's possible Apr 09 12:49:11 mickeyl: e.g. Felix von Hundelshausen - prefers CVS for the exact reasons Apr 09 12:49:29 mickeyl, you're very insightfull with this "still" word. You never know what happens next -- once I couldn't imagine why people use vim and now I can't stand other editors :) Apr 09 12:49:43 polyonymous: hehe Apr 09 12:49:44 polyonymous: no kidding :) Apr 09 12:50:05 That's why I was wondering about mt benefits instead of just saying it sucks :) Apr 09 12:51:23 seems like libxine's ogg demuxer is totally screwed up Apr 09 12:51:24 i like its distributed nature and that (despite some problems we faced because of its bugs) you toss your changesets in a maelstrom and magically they end up being merged with all those from others. you can kill the mt process between additions, merges, etc. and it won't corrupt your data Apr 09 12:52:05 the mt guys should hire me. I have a lucky hand in uncovering the most ugly of all bugs =) Apr 09 12:52:12 mickeyl, aha. that sounds appealing, although you have have some trust in computers. I do not have it, prefer to do without magic :) Apr 09 12:52:12 heh, that's true Apr 09 12:52:37 CoreDump|home: I ask you to toy with monotone 0.26 Apr 09 12:52:48 i guess i still have stom trust, yes. otherwise i would refuse to board airplanes and stuff Apr 09 12:52:54 s/stom/some/ Apr 09 12:52:55 njs: can one incrementally roserify, or replay changes? Apr 09 12:53:16 mickeyl: I have still no respone on my mail to qtopia-interest Apr 09 12:53:33 mickeyl, that's true. But I would definitely refuse to board fully automated aircraft with no human pilot :) Apr 09 12:53:44 zecke: doesn't really surprises me. i guess they really don't know. or they don't care ;) Apr 09 12:54:39 hi mickeyl Apr 09 12:54:40 even qt-embedded-interest is more or less deserted Apr 09 12:54:42 hey pb_ Apr 09 12:55:47 mickeyl: time to either learn Gtk-- or FLTK Apr 09 12:56:10 gtkmm is quite simple. Apr 09 12:56:35 ~pester Nokia for the Hildon-- bindings Apr 09 12:56:42 Nokia for the Hildon-- bindings: Are we there yet? .. Are we there yet? .. Are we there yet? see http://beverlys.net/LJ/BuggingYou.swf Apr 09 12:56:46 zecke: au contraire, time to embrace qt under x Apr 09 12:56:59 pb__: *nod* Apr 09 12:59:05 No, libxine tremor support is totally senseless, who would expect it to work... Apr 09 12:59:54 pb_: Qt is multi platform we embrace X all the time Apr 09 13:00:13 zecke: excellent Apr 09 13:00:17 pb_: time to create a freemobile.org and adapt freedesktop.org standards Apr 09 13:00:26 right, it is always time for that Apr 09 13:02:03 reenoo|afk: try handhelds.org/~pb/locale-depends.patch Apr 09 13:03:41 hey mickeyl & zecke Apr 09 13:04:48 pb_: what is your opinion on monotone 0.26? Apr 09 13:05:04 pb_: should we ask for an SCM election or declare monotone the whiner? Apr 09 13:05:18 mickeyl: http://www.linuxjournal.com/article/8213 Apr 09 13:05:34 zecke: monotone certainly is a bit of a whiner, but I guess one could live with that Apr 09 13:06:12 zecke: monotone 0.26 certainly sounds like an improvement, though I was somewhat dismayed by this recent instance of "bad revisions" cropping up again Apr 09 13:06:42 pb_: hehe, I smelled that fuel for the fire... Apr 09 13:07:21 pb_: what 'he' did was - as I guess - perfectly legal but just dumb... Apr 09 13:07:32 koen: hi. that's a cool article! thasnk Apr 09 13:07:34 thanks, even Apr 09 13:08:12 zecke: reverting to bitbake r412 fixed the PACKAGE_ARCH problems I was seeing Apr 09 13:08:14 zecke: yeah, indeed. I guess the thing that bothers me is that these legal-but-dumb things have the capability to get the repository into a state that is, apparently, quite difficult to recover from. Apr 09 13:08:45 that is actually not true Apr 09 13:08:58 koen: please correct me then Apr 09 13:08:58 we just choose this option as the least invasive for users Apr 09 13:09:11 zecke: and, more specifically, that monotone's distributed nature seems to bring more trouble than benefits. Apr 09 13:09:40 zecke: it doesn't matter how much I (or njs) explain, the trolls will call it corruption anyway Apr 09 13:09:43 pb_: right, o-hand.com ISV is copying monotone to svn, svn to monotone :( Apr 09 13:10:01 koen: right :( Apr 09 13:10:57 zecke: we could have done a merge, but that would have required hours of cpu time on the users end Apr 09 13:11:25 (verifying the complete oz354x+oz354fam083 branch and throwing it away) Apr 09 13:13:39 zecke: I've forwarded you part of the email conversation between RP, njs and me about those revs Apr 09 13:13:49 mickeyl: wait for the Qt4 scenegraph, it will beat the crap out of evas ;) Apr 09 13:14:23 koen: someone comitted from a .dev co to oz354x branch? right? Apr 09 13:14:34 zecke: the other way around Apr 09 13:20:47 zecke: what "he" did was a simple typo when commiting, and "he" assumed that a "mt disapprove" would fix the bad commit. And "he" didn't even commit the whole tree, but one single directory. Apr 09 13:22:35 ;) Apr 09 13:25:27 zecke: http://dominion.kabel.utwente.nl/koen/pda/files/screwage.tiff Apr 09 13:25:54 zecke: monotone-viz from fink works quite well on OSX Apr 09 13:26:08 although I wished fink would start using the native OSX version of gtk Apr 09 13:43:31 * chouimat is back. Apr 09 13:43:46 morning Apr 09 13:44:24 koen: how 'native' is Gtk+ native? Apr 09 13:44:45 koen: does it do menu bars? accelerated double buffering? where is it getting the fonts from? Apr 09 13:47:45 zecke: it's getting more native every day Apr 09 13:48:04 koen: can you do native Gtk# apps already? Apr 09 13:48:19 no idea Apr 09 13:48:25 I haven't used c# in a while Apr 09 13:49:00 having c# working on osx would be cute, as I wouldn't have to use XP at work Apr 09 13:49:36 koen: I will have to do some reading (learning) today, but I hope to look into the BitBake issue Apr 09 13:52:22 hmm Apr 09 13:52:50 in .dev i can't seem to convince bitbake to chose linux-openzaurus as a provider for hostap-modules. no matter what i specify it will always attempt to build hostap-modules Apr 09 13:53:06 * mickeyl rm -rf OE Apr 09 13:53:11 * mickeyl new checkout Apr 09 13:59:34 RP: I just checked in a small fix for linux-oz 2.6.16, could you check if I got it right? http://ewi546.ewi.utwente.nl/tmp/viewmtn/revision.psp?id=e6a2ccb39d3b001fdcc1071ba1860678a1fe8647 Apr 09 14:00:00 mickeyl: yeah, I have it since yesterday as well... :( Apr 09 14:00:13 mickeyl: we discussed one fix and implemented another non-fix Apr 09 14:00:24 mickeyl: smells like a bitbake bug though Apr 09 14:00:24 ok, thanks for confirming it's not my installation Apr 09 14:00:47 koen: i don't think this is correct. IIRC we want to unite all machines under defconfig-cxx00 Apr 09 14:00:53 probably RP rather forgot to checkin that file Apr 09 14:00:56 RP ? Apr 09 14:01:13 * mickeyl flashing oz 3.5.4.1 alpha1 (gpe-image) Apr 09 14:02:19 mickeyl: http://ewi546.ewi.utwente.nl/tmp/viewmtn/revision.psp?id=7c80dc8d4f2c7059b470a4f07a2db7b06d4c07f4 Apr 09 14:02:32 "Rename file packages/linux/linux-openzaurus-2.6.16/defconfig-cxx00 as packages/linux/linux-openzaurus-2.6.16/defconfig-spitz" Apr 09 14:03:00 that's why I thought my fix is correct Apr 09 14:03:07 oops Apr 09 14:03:10 hmm odd Apr 09 14:03:12 mickeyl: using -vvv I see hostap-utils -> hostap-modules (hostap-modules) Apr 09 14:03:18 but double checking can't hurt when it is about kernels Apr 09 14:03:29 mickeyl: but it doesn't print a 'consider defining PREFERRED_PROVIDER' Apr 09 14:03:39 mickeyl: it seems not to see the other one at all Apr 09 14:03:41 zecke: yeah, that's kind of odd Apr 09 14:04:09 to be honest, i don't understand all the RPROVIDES -> PROVIDES or rather RDEPENDS->DEPENDS magic not at all. is that documented anywhere? Apr 09 14:04:27 documented? Apr 09 14:04:32 magic? Apr 09 14:04:32 what's that? Apr 09 14:04:44 mickeyl: merlin, gandalf, you are all into magic it seems Apr 09 14:05:57 heh, that's true Apr 09 14:06:00 your forgot Apr 09 14:06:03 allanon.vanille.de Apr 09 14:06:10 allanon is a famous magician as well Apr 09 14:06:30 mickeyl: oh, what about copperfield or this dutch magican? Apr 09 14:06:40 mickeyl: I just know avahlon Apr 09 14:06:46 heh those new guys are just playing tricks Apr 09 14:06:49 mickeyl: I wouldn't mind to live their Apr 09 14:07:44 03koen 07org.oe.dev * re6a2ccb3... 10/packages/linux/linux-openzaurus_2.6.16.bb: linux-openzaurus 2.6.16: change defconfig-cxx00 to defconfig-spitz as the former doesn't exist Apr 09 14:13:58 mickeyl: here's a commit especially for you :) Apr 09 14:14:32 * mickeyl plays a drumroll Apr 09 14:14:48 koen: I have a big merge and some changes for you as well Apr 09 14:15:24 03koen 07org.oe.dev * r184b27ce... 10/packages/efl++/efl++.inc: efl++: fix SRC_URI Apr 09 14:16:02 koen: ah yeah! Apr 09 14:16:13 what a reminder... Apr 09 14:17:00 * mickeyl needs to cleanup that stuff and publish the cvs Apr 09 14:17:02 or svn Apr 09 14:17:04 or mt Apr 09 14:17:06 :D Apr 09 14:17:15 mickeyl: I also have .bbs for 0.1.0 Apr 09 14:17:27 but I have to build them first Apr 09 14:17:33 cool. thanks Apr 09 14:30:14 pb_: thanks for the patch. will give it a try later today Apr 09 14:30:36 geat Apr 09 14:30:37 great, too Apr 09 14:34:48 let us see what I broke Apr 09 14:36:37 koen: http://www.flickr.com/photos/natfriedman/33566213/ <- lol Apr 09 14:37:12 heh Apr 09 14:38:33 03coredump 07org.oe.oz354x * r4f74234b... 10/packages/xserver-common/ (files/calibrate_zaurusd.patch xserver-common_1.8.bb): xserver-common: Restart zaurusd after calibration. Fixes non-working softkeys on first boot. Apr 09 14:38:38 03coredump 07org.oe.oz354x * rd7ad139f... 10/packages/matchbox-panel/matchbox-panel_0.9.3.bb: matchbox-panel: DEFAULT_PREFERENCE = "-1" for buggy version 0.9.3 Apr 09 14:38:42 03coredump 07org.oe.oz354x * r0a1e8311... 10/conf/distro/preferred-gpe-versions-2.7.inc: preferred-gpe-versions-2.7.inc: Revert to matchbox-panel 0.9.2 Apr 09 14:38:46 03coredump 07org.oe.oz354x * r0578d254... 10/packages/xserver-common/ (files/setDPI.sh xserver-common_1.8.bb): xserver-common: Adjust rxvt font-size as well. Apr 09 14:38:50 03freyther 07org.oe.dev * r21066969... 10/packages/ (nonworking/pptp-linux pptp-linux): Apr 09 14:38:50 packages/pptp-linux: Apr 09 14:38:50 Move away from nonworking as we have used it several Apr 09 14:38:50 times on our SIMpad already. Apr 09 14:38:55 03freyther 07org.oe.dev * r3d4d2c8a... 10/ (conf/build/Power Macintosh-darwin.conf site/powerpc-darwin): Apr 09 14:38:55 OS X OE builds: Apr 09 14:38:57 Use the power of OE to redefine some variables if we Apr 09 14:38:58 are on a PowePC darwin Operating System. We define Apr 09 14:39:01 values to be more friendly to autotools Apr 09 14:39:03 03freyther 07org.oe.dev * rbec705b0... 10/packages/flex/ (files/flex-lvalue.diff flex.inc): Apr 09 14:39:05 packages/flex/flex(-native): Apr 09 14:39:07 Apply patch from the Debian patch tracker for *BSD Apr 09 14:39:09 libc based Operating systems Apr 09 14:39:13 03freyther 07org.oe.dev * r6591fa99... 10/classes/ (cross.bbclass native.bbclass): Apr 09 14:39:15 OS X: core and native.bbclass: Apr 09 14:39:17 -We do not have rpath on Darwin, do not use it Apr 09 14:39:19 for the native builds Apr 09 14:39:23 03freyther 07org.oe.dev * rdfc96c23... 10/packages/quilt/ (files/non-gnu.patch quilt-native_0.42.bb quilt_0.42.bb): Apr 09 14:39:25 packages/quilt: Apr 09 14:39:28 Be buildable on non GNU systems. This patch was mainly Apr 09 14:39:30 taken from the quilt mailing lists. Apr 09 14:39:31 03freyther 07org.oe.dev * rb40b9fb8... 10/packages/patcher/patcher_20040913.bb: Apr 09 14:39:33 packages/patcher: Apr 09 14:39:36 Remove language construct no longer being used/available Apr 09 14:40:08 I'm merging the multiple heads now Apr 09 14:40:45 oooh Apr 09 14:41:04 getting closer to live demos on my powerbook Apr 09 14:41:10 " language construct no longer being used" ? Apr 09 14:41:35 ah, nm Apr 09 14:41:38 _delete Apr 09 14:41:40 yep Apr 09 14:41:54 patcher was the only user, and it was 'wrong' Apr 09 14:42:17 yeah Apr 09 14:42:26 though sometimes i would find that construct pretty helpful Apr 09 14:43:26 mickeyl: I didn't understand the python code -> saw nobody used it -> removed the python code Apr 09 14:43:43 wasn't there a SL-Cxxxx patch for button-settings? Apr 09 14:57:19 koen: I'm slowly replacing fink with OE :} Apr 09 14:57:24 koen: e.g. to build a newer flex Apr 09 14:57:25 :) Apr 09 15:11:26 I have not been able to compile fbreader it is missing a .h file Apr 09 15:28:28 ~lart SuSE Apr 09 15:28:28 * ibot takes a big bite out of SuSE's jugular vein Apr 09 15:28:43 mickeyl: | make[1]: Entering directory `/data/build/koen/OE/build/tmp/angstrom/work/armv5te-linux-gnueabi/efl++-x11-0.1.0-r0/efl++-0.1.0/src'| make[1]: *** No rule to make target `common.pro', needed by `Makefile'. Stop. Apr 09 15:29:20 'File not found on Server' ... which file... idiots Apr 09 15:35:29 koen: common.pro there? Apr 09 15:38:00 yes Apr 09 15:39:44 a, bit not in src/ Apr 09 15:40:24 s/bit/but/ Apr 09 15:50:54 i'll take a look Apr 09 15:51:39 bbl, family diner because of my wife's birthday Apr 09 15:52:56 Looks like someone should prioritize things in his life... ;) Apr 09 16:06:52 zecke: http://thread.gmane.org/gmane.comp.handhelds.openzaurus.devel/762/focus=762 Apr 09 16:08:20 koen: hmm, make zecke flithy rich so he can stop studying and OSX... Apr 09 16:10:23 hi all Apr 09 16:10:43 zecke: I think the new bitbake doesn't like the python stuff in angstrom.conf Apr 09 16:10:46 hey medaglia Apr 09 16:10:52 anyone knows if I need xorg to compile opie? Apr 09 16:11:16 opie doesn't use X Apr 09 16:11:23 koen, no? Apr 09 16:11:25 uhm... Apr 09 16:11:36 So... all I need is qt-embedded and opie? Apr 09 16:44:25 koen: what did you try to build - to get the PACKAGE_ARCH bug Apr 09 17:02:29 zecke: gpe-image, c7x0, angstrom-2006.9 Apr 09 17:12:55 http://www.datenfreihafen.org/~stefan/EZX/photos/a780-kernel-messages.jpg <- butt ugly Apr 09 17:13:33 hah Apr 09 17:13:49 mvista Apr 09 17:14:42 koen: what is the state of PPC integration in OE? Apr 09 17:14:56 koen: I know about the dreambox, I wonder about other boards :} Apr 09 17:15:00 zecke: I have no idea Apr 09 17:15:17 greg72 did some work, no idea if his patches were integrated Apr 09 17:15:31 but Ifaistos and hopsnbarley are working on it now Apr 09 17:22:04 lol Apr 09 17:22:04 http://www.inf.ufsc.br/~alebm/ezx/ Apr 09 17:22:25 mickey|family: you have been forked to the motorola phone Apr 09 17:23:16 ppatching makefiles... smart Apr 09 17:24:33 trouble after trouble, grr Apr 09 17:24:47 ? Apr 09 17:24:48 some gst-plugins do not seem to work when built for x86 Apr 09 17:25:04 or it's the environment Apr 09 17:25:07 oh :( Apr 09 17:25:17 sinesrc -> osssink works Apr 09 17:25:39 but gst-register fails when trying to add vorbis or mad plugins Apr 09 17:30:39 Cockroach-: I started using gst as of yesterdax - so I'm no help at all Apr 09 17:31:29 :) Apr 09 17:32:08 It's been a while since I last played with gstreamer Apr 09 17:32:45 koen: That change is 50% right :) Apr 09 17:32:48 giving appropriate flags to gst-inspect does tell me the plugins fail, but google turns up nothing Apr 09 17:33:10 and since this environment is like Frankensteins monster, I cannot bug gstreamer-folk too much Apr 09 17:33:22 Have we a way to handle SRC_URIs that are tarballls within tarballs? Apr 09 17:33:43 RP: no Apr 09 17:33:52 RP: one could hack up unpack Apr 09 17:33:52 just wondering if gstreamer/gst-plugins are patched in OE so, that they do not work in x86 anymore Apr 09 17:34:04 RP: base.bbclass... Apr 09 17:34:54 RP: did you ever test IrDa (pxa_ficp) on a Cxxxx? Apr 09 17:35:16 RP: I fear STUART and PXA_ficp conflict with each other Apr 09 17:35:39 zecke: I never did, no... Apr 09 17:35:54 zecke: Nobody's ever reported a bug mind... Apr 09 17:36:10 RP: alwin did (to me) some weeks ago Apr 09 17:36:22 RP: but I just realized it might be an issue somewhere else Apr 09 17:36:44 zecke: He saw it on the cxxxx? Apr 09 17:36:58 oh C7x0 Apr 09 17:37:03 *sorry+ Apr 09 17:37:12 What's the difference between linux-gnueabi and eabi ? Apr 09 17:37:24 RP: see csl faq Apr 09 17:37:32 RP: EABI is a family of ABIs Apr 09 17:37:32 zecke: Its been confirmed working on c7x0 in the past Apr 09 17:39:27 :o Apr 09 17:39:37 well Isn't this just the day of days Apr 09 17:40:02 zecke: The FAQ doesn't explain the difference I'm looking at :-/ Apr 09 17:41:14 Luck, open source software with sources and finally understanding what I'm seeing and I'm able to get it to work Apr 09 17:41:49 RP: :( Apr 09 17:42:07 I just lacked another plugin before those others started to work. This constant mood switching between frustration and relief starts to get to me :) Apr 09 17:42:21 RP: try the Debian Wiki? :} Apr 09 17:42:40 zecke: I'll just guess I want the linux version :} Apr 09 17:43:53 i need a appointment at the dentist... Apr 09 17:45:53 zecke: :-( Apr 09 17:46:10 they are in a miserable state... Apr 09 17:46:38 one could think I'm frong england (what a wonderful stereotype) Apr 09 17:46:48 frong/from... Apr 09 17:47:40 I've seen too much of my dentist mainly due to major orthodontal work a few years ago :-/ Apr 09 17:48:36 my pain killers + desease + coca cola... Apr 09 17:48:43 the suckers are breaking apart :} Apr 09 17:49:01 My other problem is I grind them :-/ Apr 09 17:49:16 I think I do that at night... Apr 09 17:49:25 and in the morning bits are missing... Apr 09 17:49:28 * RP puts the pen down he was chewing. Must stop doing that... Apr 09 17:50:31 If you do it at night, they can give you gum shield like things which stop you wearing the teeth down Apr 09 17:52:03 I do it during the day which resulted in my jaw muscles becoming overstressed. I was unable to eat for several days :-/ Apr 09 17:52:14 * RP puts the pen down again :} Apr 09 17:53:00 RP: :( Apr 09 17:54:01 zecke: It does make you stop doing it though so it wasn't all bad - I'm a lot better than I was :) Apr 09 17:59:51 * RP -> food Apr 09 18:01:02 03rpurdie 07org.oe.dev * rd2b8bc17... 10/packages/linux/ (linux-openzaurus_2.6.15.bb linux-openzaurus_2.6.16.bb): linux-oz-2.6: Correct the defconfig SRC_URIs to match recent changes Apr 09 18:01:04 RP: linuxeabi and linux-gnueabi are just different ways to tell gcc and autofoo that you're using eabi (AFAIK) Apr 09 18:07:30 03koen 07org.oe.dev * r745a9ea4... 10/packages/gpe-windowlist/gpe-windowlist_cvs.bb: gpe-windowlist: inherit gpe to ship the icons as well Apr 09 18:11:53 god Apr 09 18:12:00 so many clueless people Apr 09 18:12:08 RP: hehe Apr 09 18:12:30 koen: I figured it is a lot harder to get a clue, than to be clueless Apr 09 18:13:31 "If monotone has yet to be chosen as the SCM for OE, Apr 09 18:13:31 why has it been forced upon the community even when early Apr 09 18:13:32 attempts were made to have OE parallel in other SCM?" Apr 09 18:13:46 I don't seem to remember people testing other scms Apr 09 18:14:22 emte: hey :) Apr 09 18:14:38 emte: I don't think you have tested svn, darcs, git or mercurial :) Apr 09 18:15:40 koen: this big bad company behind OE has forced everyone to use monotone Apr 09 18:15:48 koen: I stumbled into the gstreamer & ogg plugin problem and google came up with irclog from last august where you seemed to have the same problem (invalid ogg stream serial). How was that solved? Apr 09 18:16:53 Cockroach-: you'd have to ask reenoo for the details Apr 09 18:17:02 Cockroach-: iirc it was a problem with spider or playbin Apr 09 18:17:12 it should be all solved with gst 0.10 Apr 09 18:17:30 it manifests when using gst-launch from command line Apr 09 18:17:44 so I don't think it's spider thing Apr 09 18:18:01 on the log there was mentioned this tremor fixed point ogg library Apr 09 18:21:16 zecke: is that the same company that paid for our porches? Apr 09 18:21:18 zecke: the mt 0.26 thing was going to be asked eventually, and i'm not bossing people around. Apr 09 18:21:50 it's just a prompt for discussion, nothing personal y'know Apr 09 18:23:02 the point is that we already were discussing it Apr 09 18:23:56 i missed this. where? Apr 09 18:24:03 as you might have noticed people are very picky about forcing time schedules onto them Apr 09 18:25:30 jnc: we were (are) discussing it right here in #oe Apr 09 18:26:40 (although I don't know for the past few days, since I was away from internet) Apr 09 18:31:14 koen: and my ultra expensives workstation... Apr 09 18:31:18 koen: Sun niagra... Apr 09 18:31:43 koen: i was looking on the mailing list for a discussion. not everyone is in the same timezone :/ Apr 09 18:31:55 koen: what where we talking about Apr 09 18:33:12 zecke: that some people are clueless and that we got filthy rich by 'forcing monotone onto the community' Apr 09 18:33:21 jnc: that's why logs and scrollback exist Apr 09 18:33:46 koen: ah right. I still wait for the stuff njs promised me when we stay with monotone Apr 09 18:34:32 koen: I still wait until people recognize monotone is only a set of shell scripts wrapped around rsync and rcs Apr 09 18:34:36 o? Apr 09 18:34:43 I already got my new macbook pro Apr 09 18:34:52 bastard! Apr 09 18:35:00 ;) Apr 09 18:35:05 koen: Stefan Eilers' macbook pro also returned from repair Apr 09 18:35:09 * koen still only has an old powerbook Apr 09 18:35:26 for the logs: zecke and I are being sarcastic Apr 09 18:35:37 Apr 09 18:37:13 oh well Apr 09 18:37:16 time to bug people Apr 09 18:37:19 reenoo: you around? Apr 09 18:37:53 Cockroach-: sorta Apr 09 18:38:11 reenoo: blame koen, he said to ask from you :) Apr 09 18:38:48 reenoo: Do you remember the problem with gstreamer & ogg plugin? "invalid ogg stream serial no" Apr 09 18:39:19 koen: when was this, i will look in my scroll back for that day Apr 09 18:39:41 last week or so Apr 09 18:39:45 when trying (for example) 'gst-launch filesrc location=/foo/bar/1.ogg ! oggdemux ! vorbisdec ! osssink' Apr 09 18:39:46 oh Apr 09 18:39:53 Cockroach-: not sure. missing plugin, incorrect pipeline, or general breakage Apr 09 18:40:02 plugin is there Apr 09 18:41:22 i don't have the log from a week ago. Apr 09 18:41:38 ~logs Apr 09 18:41:40 apt/ibot/jbot/purl all log to http://ibot.rikers.org// where channelname is html encoded ie: %23debian | lines that start with a space are not shown | some channels have stats at http://ibot.rikers.org/stats/.html.gz, or updated "nightly" Apr 09 18:42:03 that should work :) Apr 09 18:42:06 Cockroach-: is that on an ARM? you won't have any joy with those plugins without an FPU Apr 09 18:42:20 reenoo: no, I'm running x86 at the moment Apr 09 18:42:24 thats why I wonder Apr 09 18:42:28 koen: I would like to upgrade to monotone 0.26, monotone is using it itself, and there is no other project like OE Apr 09 18:42:37 koen: what will break, will only break for us Apr 09 18:43:08 reenoo: I've read something about that tremor fixed point library, but is that only in 0.10.x series Apr 09 18:43:12 zecke: I've heard about some problems with rosterifying Apr 09 18:43:22 instead of this 0.8.9 I'm using at the moment. Apr 09 18:43:22 koen: can you remember what date this discussion was? Apr 09 18:43:34 zecke: and we need to figure out how to rosterify 3rd party branches Apr 09 18:43:50 koen: did you look at the wiki site? Apr 09 18:43:58 which wiki site? Apr 09 18:44:22 koen: the one refered to in the UPDATE document Apr 09 18:44:22 i could certainly pull all the files and zgrep, it would be a waste of bandwidth Apr 09 18:44:50 zecke: I have Apr 09 18:45:11 koen: isn't it talking about one strategy on how to convert private branches Apr 09 18:45:41 zecke: the other problem is that we need to install a new viewmtn and port our patches Apr 09 18:46:00 okay Apr 09 18:46:00 zecke: yeah it has a way, njs and I were discussing that a few weeks ago Apr 09 18:46:22 Cockroach-: I'm afraid I'd have to look up the exact pipeline. best bet is to install all plugins and use playbin. assuming you don't have any space constraints. Apr 09 18:47:47 Ok. At the moment there is no problem with space, but on the final platform it's a different story Apr 09 18:48:35 you can check which plugins are used with lsof once you have a working pipeline Apr 09 18:49:02 but, if the 0.8.9 doesn't have tremor based ogg plugin nor the 3630 has the power to decode ogg with libogg, then I cannot use ogg at all Apr 09 18:49:23 Unless I'm able to compile 0.10.x Apr 09 18:49:41 well, on the h3630 you probably want to use the ivorbis plugin Apr 09 18:50:49 zecke: I drafted up a plan for the change a while ago, but emte killed my motivation Apr 09 18:52:27 ah, the plugin name was ivorbis Apr 09 18:52:37 but the problem seems to be oggdemux Apr 09 18:55:15 lol SuSE sucks too much... Apr 09 18:55:53 zecke: doesn't anything with RPM suck? Apr 09 18:56:19 koen: First they enable a firewall, then the Browser Share methods claims, well it might not work due enable Firewall... Apr 09 18:56:55 no default update servers.. Apr 09 18:58:11 teach OE to emit debs and use it as a desktop distro Apr 09 19:02:04 Cockroach-: i'm using ivorbis (i.e. tremor) with gstreamer 0.10.something in a oe-based distro, what's the problem with 0.10? Apr 09 19:02:16 don't use oggdemux - just ivorbis (iirc) Apr 09 19:02:58 right, ivorbis does its own demuxing Apr 09 19:06:55 emte: ping Apr 09 19:19:30 * chouimat is away: stuff Apr 09 19:43:01 re Apr 09 19:52:18 * RP returns Apr 09 19:52:38 RP: wb Apr 09 19:52:50 hmmm, a newline of thinking Apr 09 19:53:06 koen: Why would there be two different tarballs from code sourcery, one twice the size of the othe with the difference being one called EABI and the other called Linux EABI Apr 09 19:53:10 zecke: I've figured out how to defeat ewi's firewall Apr 09 19:53:13 03tmbinc 07org.oe.dreambox * r8dd02b17... 10/conf/distro/opendreambox-1.2.conf: opendreambox-1.2: use swig-native version 1.3.27 Apr 09 19:53:17 03tmbinc 07org.oe.dreambox * re7264c45... 10/packages/tuxbox/libdvbsi++.bb: libdvbsi++: update to 20060329 Apr 09 19:53:21 03tmbinc 07org.oe.dreambox * r98e139d5... 10/packages/swig/ (3 files in 2 dirs): swig,swig-native: add version 1.3.29 Apr 09 19:53:26 03tmbinc 07org.oe.dreambox * r931d9c28... 10/packages/python/ (6 files in 2 dirs): python-2.4.3: import package from .dev branch Apr 09 19:53:29 RP: comninatorial brainfuck? Apr 09 19:53:30 03tmbinc 07org.oe.dreambox * r4ebeb4e1... 10/packages/netkit-base/ (netkit-base-0.17/gcc4_buildfix.patch netkit-base_0.17.bb): netkit-base-0.17: add gcc4 buildfix Apr 09 19:53:37 (to quote Jeff Waugh) Apr 09 19:53:38 03tmbinc 07org.oe.dreambox * r65a7370d... 10/packages/linux/ (3 files in 2 dirs): linux-dm7025: upgrade to kernel 2.6.12.6, add nptl support Apr 09 19:53:50 tmbinc: ping Apr 09 19:53:55 03tmbinc 07org.oe.dreambox * r2786f076... 10/packages/linux/linux-dm7020.bb: linux-dm7020: use other mirror Apr 09 19:53:59 tmbinc: do we get some updates in .dev as well? :D Apr 09 19:54:00 03tmbinc 07org.oe.dreambox * r70b54c98... 10/packages/linux-libc-headers/ (3 files in 2 dirs): linux-libc-header-2.6.12.0: add package, add mips nptl fixes Apr 09 19:54:03 combinatorial* Apr 09 19:54:03 03tmbinc 07org.oe.dreambox * r7b67edbc... 10/packages/gstreamer/gstreamer_0.10.4.bb: gstreamer-0.10.4: fix metadata Apr 09 19:54:08 03tmbinc 07org.oe.dreambox * re66f25d8... 10/packages/binutils/ (7 files in 2 dirs): binutils-2.16.1: add package Apr 09 19:54:13 03tmbinc 07org.oe.dreambox * r2d1f29f2... 10/packages/glibc/ (5 files in 3 dirs): glibc-2.3.5+cvs20051107: add mips nptl fixes Apr 09 19:54:16 03tmbinc 07org.oe.dreambox * rbad2a103... 10/packages/libxml/libxml2_2.6.22.bb: libxml2: use different source Apr 09 19:54:21 03tmbinc 07org.oe.dreambox * r27394325... 10/packages/gcc/ (6 files in 2 dirs): gcc-4.1.0: add package Apr 09 19:54:23 03tmbinc 07org.oe.dreambox * r7fdc3b5c... 10/packages/gcc/ (21 files in 2 dirs): gcc-3.4.5: add package Apr 09 19:54:27 03tmbinc 07org.oe.dreambox * r2103e7aa... 10/packages/glibc/ (glibc_2.3.2+cvs20040726.bb glibc_2.3.5+cvs20050627.bb): glibc-2.3.2: fix cvs breakage by removing ports Apr 09 19:54:28 go tmbinc, go Apr 09 19:54:32 03tmbinc 07org.oe.dreambox * r4ba45f02... 10/packages/meta/dreambox-image.bb: dreambox-image: remove enigma2-plugin-systemplugins-oldsoftwareupdate Apr 09 19:54:34 * mickeyl plays cheerleader Apr 09 19:54:35 03tmbinc 07org.oe.dreambox * r9adbeb82... 10/packages/gcc/gcc-cross_4.1.0.bb: gcc-cross-4.1.0: add package Apr 09 19:54:40 03tmbinc 07org.oe.dreambox * rec5e41ee... 10/conf/machine/dm7025.conf: machine/dm7025: use gcc-3.4.5 as default Apr 09 19:54:42 we have an O Apr 09 19:54:43 gimme a t Apr 09 19:54:43 03tmbinc 07org.oe.dreambox * r61e94c14... 10/packages/autoconf/ (autoconf-2.59/sizeof_types.patch autoconf_2.59.bb): autoconf-2.59: add sizeof patch Apr 09 19:54:45 gimme a b Apr 09 19:54:48 we have an E Apr 09 19:54:48 03tmbinc 07org.oe.dreambox * r1f0a2a79... 10/packages/gstreamer/gst-plugins.inc: gst-plugins: fix metadata Apr 09 19:54:51 03tmbinc 07org.oe.dreambox * r169a4081... 10/packages/dreambox/dreambox-dvb-modules.bb: dreambox-dvb-modules: provide modules compile with gcc-4.1 Apr 09 19:54:51 OE OE oE Apr 09 19:54:54 heh Apr 09 19:54:56 03tmbinc 07org.oe.dreambox * rb51d76c1... 10/packages/enigma2/enigma2.bb: enigma2: add gstreamer as depends Apr 09 19:54:59 03tmbinc 07org.oe.dreambox * r3e5eb042... 10/packages/glibc/ (3 files in 2 dirs): glibc-2.3.5: add mips rlimit rtprio fix Apr 09 19:55:04 03tmbinc 07org.oe.dreambox * r76ecb4ef... 10/packages/strace/ (strace-4.5.14/arm-eabi.patch strace_4.5.14.bb): strace-4.5.14: import from .dev branch Apr 09 19:55:15 holy cow, that guy was busy Apr 09 19:55:49 probably just stored up his commits Apr 09 19:55:54 mickeyl: most of the stuff is already in .dev (or imported from .dev) Apr 09 19:56:04 tmbinc: swig 1.3.29 isn't :D Apr 09 19:56:15 and the rest are mips specific nptl fixes, where i'm not sure if they don't break anything else :) Apr 09 19:56:15 27 even Apr 09 19:56:20 yeah ok Apr 09 19:56:29 just a reminder, no hurries Apr 09 19:56:34 question is if you want the newer version, it emits even bigger binaries ;) Apr 09 19:57:09 but yeah, you're right, i should take the time to reduce the differences between the both trees :/ Apr 09 19:57:20 yes, that'd be good in general Apr 09 20:03:38 tmbinc: well I'm toying around with in oz354fam083 branch, so building gstreamer there produces 0.8.9. but thanks for the ivorbis info, have to test that tomorrow after work. Apr 09 20:03:53 ...after a dull 8 hours of poking the 770 :P Apr 09 20:04:37 * koen awaits maemo 2.0 Apr 09 20:05:06 how is the oe 770 state btw? is there anything useful to build in OE (from a non developer's point of view)? Apr 09 20:05:27 not really Apr 09 20:05:49 What I'm learned, big companies with paranoid lawyers should stay away from open source ;) Apr 09 20:05:53 someone should integrate the latest maemo svn bits Apr 09 20:06:00 they are too afraid to do anything Apr 09 20:06:20 that + matchboxkeyboard should give you something enduser useable Apr 09 20:06:32 koen: I'd be prepared to try and help sort out OE and the 770 Apr 09 20:06:51 koen: I'm regularly using poky images on the 770 Apr 09 20:06:51 and add a kernel so we don't have to depend on the initfs Apr 09 20:07:06 hey hrw Apr 09 20:07:10 * hrw for a moment (as usual in weekend) Apr 09 20:07:17 * Cockroach- sets mode +hibernate Apr 09 20:07:20 cheers hrw Apr 09 20:07:20 If we can get a deb fetcher, I have a kernel .bb we can use Apr 09 20:07:25 RP: having OE build maemo bits would be more beneficial in the short term Apr 09 20:07:25 hi hrw Apr 09 20:07:32 mickeyl: if card is working with orinoco_cs and hostap_cs then patch for hostap_cs is good thing? Apr 09 20:07:56 hrw: if it's a prism-based card, yes. if it's a hermes based card, no Apr 09 20:07:57 koen: I'd like to see maemo support in OE Apr 09 20:08:00 more info: with 2.6.16+pcmciautils my card == orinoco instead of hostap - I have a patch for it Apr 09 20:08:10 RP: i'd like to see both maemo and OE stock images on 770 Apr 09 20:08:18 RP: after my thesis, i'm ready to help doing that as well Apr 09 20:08:30 stock being gpe and e of course :) Apr 09 20:08:30 RP: I'd like maemo not fsck !maemo stuff during world build Apr 09 20:08:37 mickeyl: Agreed, I'd like to see both Apr 09 20:09:01 RP: the kernel does have a tar.bz2, right? Apr 09 20:09:10 it had when I last downloaded it :) Apr 09 20:09:19 Poky is basically the same as .dev as far as the 770 is concerned as we just run an xserver + mb on it Apr 09 20:09:28 koen: I could only ever find a .deb Apr 09 20:09:57 RP: could you add such an image to .dev if you have a chance? I'd like to have a lightweight X based image just w/ a window manager Apr 09 20:09:59 koen: Point me at a .tar.* and I'll put a .bb into OE Apr 09 20:10:11 time to send patch to LKML (first time in a live) Apr 09 20:10:15 hrw: woot! Apr 09 20:10:22 * mickeyl bbi1h Apr 09 20:10:31 hostap_cs.c | 3 +++ Apr 09 20:10:33 kernel-source-2.6.12.3_2.6.12.3-osso14.tar.gz Apr 09 20:11:06 I do not have to be subscribed to lkml to post? Apr 09 20:11:45 mickeyl: We'd need to refactor some of the gpe assumptions but I think this would be a good thing. Next time I feel like having lots of people complain at me, I'll give it a go :) Apr 09 20:11:57 RP: http://ewi546.ewi.utwente.nl/OE/source/kernel-source-2.6.12.3_2.6.12.3-osso14.tar.gz in about ~25 minutes Apr 09 20:12:10 Apart from that, it would be very simple to add the openedhand task package and image into OE which would give you a lightweight x image Apr 09 20:12:30 koen: Where on maemo did you find that? Apr 09 20:12:36 stage.maemo.org Apr 09 20:12:37 RP: pcmciautils 010 segfault on my c760/2.6.16 Apr 09 20:12:45 RP: it came and went Apr 09 20:12:51 I just made a backup in time Apr 09 20:13:46 kernel-source-2.6.12.3_2.6.12.3-osso14.tar.gz 8% 3868KB 27.9KB/s 25:40 ETA Apr 09 20:13:49 * koen taps foot Apr 09 20:13:49 koen: Ah, right. I wonder if we're supposed to have that :) Apr 09 20:14:03 GPL, bla bla bla Apr 09 20:14:07 ~lart pcmciautils author for /usr/bin/ln Apr 09 20:14:08 * ibot raises middle finger to pcmciautils author for /usr/bin/ln Apr 09 20:15:55 RP: 013 segfault too ;( Apr 09 20:16:00 hrw: What did you do to make it segfault? Apr 09 20:16:14 hrw: Its never segfaulted for me :-/ Apr 09 20:16:35 http://pastebin.com/650133 - strace Apr 09 20:17:26 mmap2 (NULL, xxx) doesn't look healthy Apr 09 20:18:34 indeed, xxx would be a bad value for the length Apr 09 20:23:58 cu Apr 09 20:24:41 koen: why did you want to defeat it? Apr 09 20:25:23 defeat what? Apr 09 20:26:08 koen: ewi's firewall Apr 09 20:26:26 zecke: monotone 0.26's new port number Apr 09 20:30:39 pb_: Was it you who added the csl toolchain originally? Apr 09 20:30:47 yeah, I think so Apr 09 20:32:34 pb_: Its a shame we can't seem to find a stable CVS date for it :-/ Apr 09 20:32:55 pb_: I have a version with the latest cvs date which builds fine but the images crash... Apr 09 20:33:06 that is unfortunate. how do they crash? Apr 09 20:33:33 RP: that's a step up from the previous csl toolchain, that one didn't even compile lots of stuff Apr 09 20:34:29 pb_: Init segfaults (although some of the bootup process proceeds and eventually crashes) Apr 09 20:35:29 oh dear Apr 09 20:35:48 I guess someone would have to debug the init crash. At least init is pretty simple, so that might not be too hard. Apr 09 20:36:14 pb_: I also had a version that causes oopses in the kernel (starting in load_elf_binary) Apr 09 20:37:33 pb_: My current plan is to try the 2005q3 release. If that works, would you object to the addition of the most recent cvs .bb files and the removal of the older ones (which as koen mentioned, didn't stand up to testing) Apr 09 20:39:33 RP: if the 2005q3 release is what you're testing, it would seem better to add .bb files for that rather than the most recent cvs code. Apr 09 20:39:49 but in principle, no, I wouldn't object to you doing that. Apr 09 20:39:55 pb_: I have both :) Apr 09 20:40:04 Like I said, recent cvs doesn't work Apr 09 20:40:28 right. there doesn't seem much point in adding it if it doesn't work. Apr 09 20:40:55 So just having 2005q3 might be the best way forward? Apr 09 20:41:00 If it works :) Apr 09 20:41:25 right Apr 09 20:41:32 RP: it should work as nokia seems be using it Apr 09 20:41:47 koen: I'm using that reasoning :) Apr 09 20:41:52 RP: do you use the csl glibc as well? Apr 09 20:42:11 koen: I'm using the glibc from the 4.0 build Apr 09 20:42:45 from a csl tarball or the one from OE? Apr 09 20:42:52 The one already in OE Apr 09 20:42:57 ah Apr 09 20:43:02 that would be a problem Apr 09 20:43:08 This is broken with 3.4? Apr 09 20:43:21 pbrook hinted at the fact that the csl glibc is heavily patched Apr 09 20:44:00 But doesn't the OE glibc work or has that been gving problems? Apr 09 20:45:10 it works, but if you go with csl compiler you need aq csl glibc as well Apr 09 20:45:14 (at least for eabi) Apr 09 20:46:42 could someone give a an explaination what eabi is good for? Apr 09 20:46:49 * CoreDump|home has no idea Apr 09 20:46:56 CoreDump|home: one abi for arm Apr 09 20:47:11 instead of the softfloat/hardfloat/something else we have now Apr 09 20:47:11 koen: I'm not sure that entirely makes sense. It was probably once the case but if glibc works for gcc 4.0 eabi, surely it should work for gcc-csl 3.4? Apr 09 20:47:25 heh Apr 09 20:47:25 ahh i c Apr 09 20:47:33 zecke: which stuff did I promise you? Apr 09 20:48:02 RP: csl EABI is not 'upstream' eabi Apr 09 20:48:20 RP: csl tends to 'forget' to apply patches upstream Apr 09 20:48:45 "o, we forgot to enable thumb interwork" Apr 09 20:48:58 "if you want eabi use our toolchain" Apr 09 20:49:03 etc, etc Apr 09 20:50:03 koen: I realise its not complete. If there is enough for it to work for gcc 4.0, I don't see why it shouldn't work for gcc 3.3.4-csl though. Unless there is a flaw in the logic I'm not seeing Apr 09 20:50:20 I might just add glibc as well at this rate :-/ Apr 09 20:50:50 csl makes money off their own toolchain, not off upstream Apr 09 20:52:36 their glibc has loads of fixes to work with eabi which aren't really fit for upstream according to paul Apr 09 20:53:05 (think embeddix/lineo, but less evil) Apr 09 20:54:02 IIRC it's generally because they have a contract with someone to provide X, not provide-an-upstream-that-has-X-included Apr 09 20:54:21 (esp. since you can't exactly guarantee a customer that you _can_ get the patches they need into upstream) Apr 09 20:54:35 that sums it up pretty good Apr 09 20:55:09 RP: try asking pbrook or drow in #debian-arm specific question about your problems, there's a good chance they can give you patches Apr 09 20:55:21 njs: yeah, exactly. also, in this case, because csl's EABI customers don't want to be tied to the upstream gcc release cycle. Apr 09 20:55:32 RP: that's how pH5 and I got the syscall patch for glibc Apr 09 20:55:41 njs: hey Apr 09 20:55:43 koen: I'll keep that in mind, thanks :) Apr 09 20:56:14 koen: The glibc in OE does work to some extent though? Apr 09 20:56:23 pb_: yah Apr 09 20:56:33 RP: it gets into X with a gpe-image Apr 09 20:56:38 anybody know anything about the development of konq/E ? it would seem to not have been worked on since 2003, is that correct? It mentions on the homepage that it takes code from the most up to date konweror? is that the case? is the build in opie taking advantage of code more recent than 2003 or not? Apr 09 20:56:51 koen: right. That's more than enough for me :) Apr 09 20:56:55 dan2003: ?? Apr 09 20:57:11 dan2003: well konqE for Qt2 was last updated years ago Apr 09 20:57:27 dan2003: konqE for Qt3 was made to compile with any version of KDE >3.0 Apr 09 20:57:37 dan2003: and some people used it on setup boxes etc... Apr 09 20:57:54 dan2003: now the ROAD company (www.road-gmbh.de) contracted someone to do a backport to QtE Apr 09 20:58:02 dan2003: to QtE2.3.12 Apr 09 20:58:27 dan2003: but from your writings I can't parse your question Apr 09 20:58:35 dan2003: there are too many unknown variables Apr 09 20:58:38 zecke: I pushed stuff to hg and git as well :) Apr 09 20:58:46 dan2003: way too updates Apr 09 20:58:47 ahh, so we may see an update yet.. Apr 09 20:58:53 koen: how did you get an account? Apr 09 20:59:18 for git? I asked RP Apr 09 20:59:29 pb_: I will do some GPE work (or at least wear a GPE T-Shirt at the next fair) if you cast your vote for monotone Apr 09 20:59:31 I gave accounts to several people... Apr 09 20:59:36 although I might have hallucinations about hg Apr 09 20:59:55 koen: I gave away accounts for darcs and hg, only florian has one Apr 09 21:00:11 ah yes, I remember Apr 09 21:00:28 hg didn't get updated so I lost interest Apr 09 21:00:57 zecke: I am conflicted between monotone and git, so your offer might tip the balance Apr 09 21:01:01 koen: he? it was updated like any other system Apr 09 21:01:23 zecke: the tree at alwins-world didn't Apr 09 21:01:33 hmm Apr 09 21:01:40 you mentioned some problem with tailor a while ago Apr 09 21:03:02 RP: http://ewi546.ewi.utwente.nl/OE/source/kernel-source-2.6.12.3_2.6.12.3-osso14.tar.gz is ready to be fetched Apr 09 21:03:18 osso14... Apr 09 21:04:17 pb_: (looking at backlog) actually, I'm not _entirely_ sure why OE uses monotone either, since the project doesn't seem to take advantage of the more advanced capabilities much :-) Apr 09 21:04:20 someone should get nokia to release their 2.6.16 version Apr 09 21:04:44 well, renames, maybe, there seem to be a lot of file and dir renames, and git/hg have no idea how to deal with those, really. Apr 09 21:04:48 njs: about a year ago it was about the only system around Apr 09 21:05:06 njs: we need a lot of copy and rename Apr 09 21:05:16 koen: right, that too... Apr 09 21:05:25 re Apr 09 21:05:25 njs: to update to new upstream releases, we copy file_old_version file_new_version.bb Apr 09 21:05:33 hrw|tv: hey! Apr 09 21:05:36 git couldn't even import the OE repo.... Apr 09 21:06:04 zecke: qt/e r31/32 is working nice for people - I will give them for all users tomorrow Apr 09 21:06:36 hrw|tv: what is your timeframe for OZ3.5.4.1? Apr 09 21:06:36 koen: ok, thanks - I'll sort out a .bb for that Apr 09 21:06:46 hrw|tv: I want to include the CapsLock fixes as well Apr 09 21:06:47 RP: cool, mucho gracias Apr 09 21:06:47 koen: What makes you say they have a 2.6.16 version? Apr 09 21:07:06 RP: EABI, although they could just have patched 2.6.ancient Apr 09 21:07:23 zecke: 3.5.4.1.. 3weeks max (to release it before 1may) Apr 09 21:07:48 users will shout when we will give them alpha2 ;) Apr 09 21:07:48 and unplug internet for some weeks ;) Apr 09 21:07:57 hrw|tv: I will be at a funeral from Tuesday to Thursday, so I could have an r33 by next week Apr 09 21:08:15 zecke: no problem from my side Apr 09 21:08:41 koen: That's true Apr 09 21:09:12 hrw|tv: Are the gpe issues fixed? Apr 09 21:09:29 RP: CoreDump|home will know Apr 09 21:09:38 they are Apr 09 21:09:52 So alpha2 can roll? Apr 09 21:09:55 CoreDump|home: so we can release alpha2? Apr 09 21:09:59 in my book yes Apr 09 21:10:17 ok - will do fresh build then Apr 09 21:10:33 * RP will also set a build away Apr 09 21:11:06 I should really sync that kernel size fix into oz354x Apr 09 21:11:31 NOTE: selecting already-staged gcc-cross-initial to satisfy gcc-cross-initial; upgrading from 3.4.4+csl-arm-2005q3-r0 to 4.1.0-r1 Apr 09 21:11:43 Except is isn't already staged :-/ Apr 09 21:12:01 :( Apr 09 21:12:29 hrw|tv: njs is working on a cherrypicker Apr 09 21:12:41 koen: great Apr 09 21:12:56 njs: and now we kind of like you ;) Apr 09 21:13:24 for sentimental reasons I would feel bad to move to git - but luckily this is not the biggest and not the only reason Apr 09 21:14:15 I must admit some developer support from monotone has softened my feelings a lot as well :) Apr 09 21:14:58 koen: http://www.youtube.com/watch?v=KkAtRfA1UXc Apr 09 21:16:04 cute Apr 09 21:16:25 didn't Robert Love port powerball to use the thinkpads' sensors? Apr 09 21:16:52 hi, all!!! Apr 09 21:16:56 hey slapin Apr 09 21:17:16 Do anyone have any success with QEMU running ARM EABI binaries? Apr 09 21:17:17 although accellerometers still give me a feeling of death-by-matlab Apr 09 21:17:33 slapin: not yet Apr 09 21:18:35 RP: I wonder what would happen if we'd use vanilla gcc 4.1 + glibc Apr 09 21:18:40 glibc-csl* Apr 09 21:18:51 NOTE: Checking Stamp: /usr/oe/build4/tmp/stamps/armv5te-linux/gcc-cross-initial-3.4.4+csl-arm-2005q3-r0.do_populate_staging Apr 09 21:18:51 NOTE: selecting already-staged gcc-cross-initial to satisfy gcc-cross-initial; upgrading from 3.4.4+csl-arm-2005q3-r0 to 4.1.0-r1 Apr 09 21:19:21 Something is very wrong with bitbake :-/ Apr 09 21:19:51 koen: Its an interesting idea :) Apr 09 21:33:12 cu Apr 09 21:34:49 RP: what is bitbake actually doing wrong? Apr 09 21:35:44 nothing in the output you pasted there looks obviously incorrect. Apr 09 21:36:05 pb_: Its building 4.1.0 when I want it to build 3.4.4 Apr 09 21:36:20 pb_: 4.1,9 isn't staged, contry to what what message would suggest Apr 09 21:36:29 or 4.1.0 :) Apr 09 21:36:43 it doesn't think 4.1.0 is staged. the message says that 3.4.4+csl-arm-2005q3-r0 is staged. Apr 09 21:37:11 (but that 4.1.0 is the preferred version, so it plans to build that one) Apr 09 21:37:47 pb_: It could be argued that message is ambigious Apr 09 21:38:13 pb_: But you are correct, it does thing 4.1.0 is the preferred version so I need to work out why it thinks that (I' Apr 09 21:38:19 m debugging that now) Apr 09 21:38:31 RP: what distro? Apr 09 21:38:54 I changed the ?= in angstrom to = a while ago Apr 09 21:39:10 koen: Grrr :-/ Apr 09 21:39:43 koen: why? Apr 09 21:39:55 preferred-e-versions.inc for example Apr 09 21:40:15 koen: Could you fix preferred-e-versions instead? Apr 09 21:40:39 RP: that's my plan, but I need to talk to JustinP first Apr 09 21:40:58 and I think I saw some other stuff trying to poke at toolchain stuff as well Apr 09 21:41:30 http://ewi546.ewi.utwente.nl/tmp/viewmtn/revision.psp?id=ba0fa1662b1852ad4ed93d9e5cbbb939b90908a1 Apr 09 21:41:51 koen: Personally, I'd nuke it as its blatently out of place :-/ Apr 09 21:41:52 and I got one too many reports from "it doesn't work" Apr 09 21:42:07 RP: yeah, that was my thought as well Apr 09 21:42:23 RP: but I try to communicate first instead of reverting blindly Apr 09 21:42:25 I missed that checkin. I was pleased a distro was doing something sensible for a change - I have noticed this and was using it :-/ Apr 09 21:42:47 koen: Yes, its a good idea to mention it but it does need fixing Apr 09 21:43:19 I would be very happy to change it back to ?= as soon as other parts in OE stop breaking it Apr 09 21:43:22 *very* happy Apr 09 21:43:56 (as angstrom is suppossed to be the poster child for "doing it right") Apr 09 21:44:00 koen: You have my blessing to fix OE to stop breaking it... Apr 09 21:47:27 wow I already have 1.200 out of 6000 rosterified Apr 09 21:52:21 zecke: congratulations Apr 09 21:52:57 RP: I've sent a mail to oe@ about it Apr 09 21:53:12 RP: and managed to fit a conf/machine overhaul into the same mail as well ;) Apr 09 21:53:22 pb_: hehe, I once again too verbose Apr 09 22:06:40 I love the PXA270 overlay Apr 09 22:07:10 zecke: ? Apr 09 22:07:13 flicker free graphics and transparency... Apr 09 22:07:25 RP: I currently play a video and have Opie on top of it Apr 09 22:07:37 zecke: ah :) Apr 09 22:11:03 the future of graphics is promosing Apr 09 22:18:05 zecke: with kernel 2.6? Apr 09 22:19:01 koen: dunno which kernel Apr 09 22:19:19 zecke: opie on directfb? Apr 09 22:19:27 koen: no, sadly not Apr 09 22:19:42 koen: a for loop to toggle transparency Apr 09 22:19:52 *u32 = *u32 | 0x80008000; /* MSB is transparency Bit (two pixels per loop) */ Apr 09 22:19:59 heh Apr 09 22:20:00 well after setting up the layers Apr 09 22:20:20 transparent until repaint ;) Apr 09 22:20:50 I did something like that in BASIC Apr 09 22:21:16 once I got a real compiler it had an acceptable speed Apr 09 22:21:19 koen: do you know how usable keithp's windowing system wasß (twinß) Apr 09 22:21:33 I haven't really looked into twin Apr 09 22:21:53 since all teh 'embedded' stuff I have is >200MHz with loads of ram Apr 09 22:22:07 hmm... I should get a X hacking job Apr 09 22:22:39 * koen should go to sleep Apr 09 22:22:42 hmm there's a thought, Macintosh Colour Classic as a target platform Apr 09 22:22:56 koen: hehe, that is a good option as well Apr 09 22:23:00 another 3 internet free days to come Apr 09 22:23:03 'night all Apr 09 22:23:09 koen: in Germany? Apr 09 22:23:21 zecke: no, amsterdam (work) Apr 09 22:23:30 koen|sleep: where do you work now? Apr 09 22:24:00 still measuring police officers for their new bulletproof vests Apr 09 22:24:23 ah good, just switched offices then Apr 09 22:24:40 yeah we measure it on site Apr 09 22:24:47 so we tour the country Apr 09 22:24:51 koen|sleep: can you shoot on police officers as well? Apr 09 22:24:52 anywayz Apr 09 22:24:56 heh Apr 09 22:25:09 no, I can't Apr 09 22:25:12 ZZZZZZZZZZZZZ Apr 09 22:25:34 koen|sleep: well, then your job sucks! Apr 09 22:25:36 koen|sleep: have fun! Apr 09 22:32:49 03tmbinc 07org.oe.dreambox * r70274558... 10/packages/linux/linux-dm7025.bb: linux-dm7025: add experimental dvb multipid patch Apr 09 22:32:53 03tmbinc 07org.oe.dreambox * r8509a841... 10/packages/enigma2/enigma2.bb: enigma2: update Apr 09 22:33:22 tmbinc: multipid? Adding more than one PID? Apr 09 22:35:13 zecke: something like "dvr0" just multiple times Apr 09 22:35:25 so you can record more than one service from one transponder Apr 09 22:35:40 based on the - yet uncommented - RFC on the linux-dvb ml Apr 09 22:36:01 I should subscribe to this list... Apr 09 22:36:18 do you know Patrick Boetcher, and have you heard of him lately? Apr 09 22:36:36 well, the patch really ugly at the moment, but we need to betatest an image with this patch :) Apr 09 22:36:55 customers are crying because they can't do that at the moment (record more than one program at once on one transponder) Apr 09 22:37:05 I'm mainly using dvbnet, so adding more than one PID, would allow me to add more IPs ;) Apr 09 22:37:37 hm no it only works for PES (aka. TS/PES) filters Apr 09 22:37:46 but can't you just create multiple interfaces? Apr 09 22:38:08 or write a userspace replacement for dvbnet Apr 09 22:38:11 using a tun device Apr 09 22:39:22 yes these are the alternatives, I'm just slowly getting to DVB and LinuxDVB... Apr 09 22:39:23 then you could use section filters, poll() them (or use threads), or even filter on 0x2000 (to get the whole transponder), and do some intelligent section-reassembling (probably not so trivial if you don't want to waste memory), then you have "everything" ;) Apr 09 22:40:01 interesting idea, this way I could capture everything coming from my modulator :} Apr 09 22:40:24 I need more time :( Apr 09 22:41:18 what hardware setup do you have? Apr 09 22:41:28 on the TX side? Apr 09 22:41:39 03tmbinc 07org.oe.dreambox * ra5f4f0c4... 10/packages/dreambox/dreambox-dvb-modules.bb: dreambox-dvb-modules/dm7025: update to 20060410 Apr 09 22:41:57 tmbinc: That is all black magic to me Apr 09 22:42:17 tmbinc: we have a 'data planet' IPE, and a green cable coming into my room Apr 09 22:42:27 tmbinc: which I connect to Input A of my modulator :} Apr 09 22:42:40 tmbinc: you see I'm a complete expert on DVB ;) Apr 09 22:44:25 hehe Apr 09 22:45:01 tmbinc: I only need to care/understand more of this stuff when IP multicast fails Apr 09 22:45:44 zecke: In python, if you have a dictonary, can you have a function trigger if something tries to access a key that doesn't exist in said dictonary? Apr 09 22:46:18 RP: yo can implement a __getitem__ and not throw KeyError Apr 09 22:46:30 RP: you would create an Adapter (when talking about patterns) Apr 09 22:47:05 zecke: ok, I just wondered if it was possible. I have some ideas about rewriting the cache handling... Apr 09 22:47:23 RP: ah, I can imagine your idea Apr 09 22:47:37 RP: quite clever Apr 09 22:47:40 zecke: self.pkgdata ? Apr 09 22:47:48 no ;) Apr 09 22:48:03 collect the information and save the variables that were hit Apr 09 22:48:03 What were you thinking i was thinking? :) Apr 09 22:48:26 this solves my biggest headaches (saving the values unexpanded) Apr 09 22:48:48 pleased to be of service :) Apr 09 22:49:20 so what is your idea, some kind of delayed loading? Apr 09 22:49:28 but how is it saving the wolrd? Apr 09 22:49:45 zecke: I'm wondering if we can lose pkgdata if we have a second smaller cache of the data we mostly use Apr 09 22:50:00 pkgdata would then just get filled on demand Apr 09 22:50:14 RP: pkgdata, is currently used for the Provider -> filename map? Apr 09 22:50:21 RP: or does it store more? Apr 09 22:50:36 It stores a the metadata for each .bb file Apr 09 22:51:51 RP: ah okay, the proxy pattern Apr 09 22:52:02 hmm... yeah. more or less we only need to cache the bits which influence depedency generation. it gets a bit tough to predict whether python functions add things to dependency handling or not though Apr 09 22:52:09 * mickey|bbl on the way to bed Apr 09 22:52:11 g'night Apr 09 22:52:18 mickey|bbl: you can#t sleep now Apr 09 22:52:22 heh Apr 09 22:52:25 mickey|zzZZzz: opening a can full of worms, and... Apr 09 22:52:28 :D Apr 09 22:52:30 mickey|bbl: Once we've scanned a file, we can cache the resulting data Apr 09 22:52:33 'night mickey|zzZZzz Apr 09 22:53:26 RP: hg is not too bad, there are things I don't understand though ;) Apr 09 22:53:36 RP: and keithp claims metadata of hg will change soon Apr 09 22:54:01 zecke: I think it has too many issues for my liking... Apr 09 22:54:29 RP: darcs, is a patch tracker, completely useless for our workflow Apr 09 22:54:46 RP: hg could match, but monotone is doing fine as well Apr 09 22:54:57 RP: having the njs bot in this channel is of great help Apr 09 22:55:17 RP: how often do you have to do svn cleanup on your poky tree? Apr 09 22:55:36 zecke: I've no idea what that does :} Apr 09 22:55:55 zecke: Agreed, as I said, it comes down to git, monotone or at the outside, svk and monotone looks to be winning atm Apr 09 22:56:04 RP: I do svn diff, CTRL+C because it takes so long and then need to do svn cleanup to remove lock files Apr 09 22:56:12 Personally, I like git but I accept that its nasty for people to use Apr 09 22:56:40 RP: I did some merges with git, but once again I don't remember how to do it Apr 09 22:56:48 zecke: I tend not to use ctrl+C and delete lock files manually if its a problem :} Apr 09 22:57:12 RP: and I hate its' GNUism - which keithp claims will be fixed soon Apr 09 22:58:04 zecke: There is a future in git Apr 09 22:58:11 zecke: I also think monotone has a future though Apr 09 23:01:43 njs addressed all of my concerns, recently. Apr 09 23:02:11 the thing that might benefit from discussion is some kind of reccommendation or policy on how devs ought to use branches Apr 09 23:02:35 call me crazy, but there's a lot of duplicated efforts atm Apr 09 23:03:34 how do i check someone elses changes before they commit? i have to track down their patches in a bugzilla report, read the mailing list, search IRC logs, and so on. then i have to apply those patches myself, and based on that information, i say simply "yes" or "no" Apr 09 23:04:35 it would be easier if a guideline was suggested on how people make branches for pending changes, so that code can be reviewed quickly Apr 09 23:06:32 jnc: The current patch review process is via patches, not brancges Apr 09 23:07:01 do you think monotone is well suited for that? Apr 09 23:07:21 versus say, other SCMs Apr 09 23:07:24 jnc: No more or less so than most other SCMs Apr 09 23:09:35 RP: okay pkgdata is utilised when the build is going Apr 09 23:09:54 RP: can we save a lot of time from that? I think not, the compile time should outweight that Apr 09 23:10:17 zecke: Picture this scenario: Apr 09 23:10:29 * zecke pushes the record button Apr 09 23:11:20 When we parse the .bb files first time around, we save a subset of variables (DEPENDS, RDEPENDS, PREFERRED_PROVIDER*) etc. to a secondary cache Apr 09 23:11:35 yes Apr 09 23:11:46 We also save the parsed .bb file to the cache as we do at the moment Apr 09 23:12:41 When we get run with a valid cache, we don't load the parsed .bb file from the main cache but just access the secondary cache and its subset of variables Apr 09 23:13:00 Until we actually build a .bb file in which case we load that single .bb file from the main cache Apr 09 23:13:18 okay this matches 100% on what we have talked about and agreed on :) Apr 09 23:13:32 Good :) Apr 09 23:13:49 I'm just a little happier now I've looked at what bitbake is currently doing Apr 09 23:14:00 RP: monotone is, in some theories at least, _supposed_ to be better suited for patch review/workflow stuff than other SCMs... Apr 09 23:14:04 Its actually more straightforward to do this than I thought Apr 09 23:14:29 njs: That's good to know :) Apr 09 23:14:46 RP: There are two gotchas though Apr 09 23:15:01 zecke: go on... Apr 09 23:15:15 RP: hehe, need to get something done... Apr 09 23:15:29 RP: first, the BitBake parser should not call update_data Apr 09 23:15:37 RP: we should cache the raw data first Apr 09 23:15:49 RP: then update_data Apr 09 23:15:53 zecke: Yes, I was looking at that PREFERRED_PROVIDER business I added... Apr 09 23:16:16 RP: and the second one is dependency tracking of Vars Apr 09 23:16:24 RP: DEPEND depnds on FILE Apr 09 23:16:25 Ah, I understand what you mean and agree Apr 09 23:16:30 i did a google search, but did not come up with a definition for what "rosterify" is or what it does Apr 09 23:16:39 RP: and a python method - from bitbake Apr 09 23:16:47 RP: base_* methods don't need to be cached Apr 09 23:16:56 RP: but autotools methods as well Apr 09 23:17:18 zecke: Tracking variables down to the file level is going to be tricky Apr 09 23:18:06 RP: hehe, we could just start with some know values Apr 09 23:19:49 zecke: You start to wonder if we should track it down to the line in the file :) Apr 09 23:19:59 zecke: aka bitbake2 :) Apr 09 23:20:13 (or bitbake-ng) Apr 09 23:20:26 hehe Apr 09 23:20:30 jnc: rosterify is the migration command from 0.25 to 0.26 Apr 09 23:20:38 Hmm. I should sleep Apr 09 23:20:45 ah. it is not so much a generic term? Apr 09 23:20:49 A repository hook failed: 'pre-commit' hook failed with error output: Apr 09 23:20:52 yeah svk... Apr 09 23:22:01 jnc: not really Apr 09 23:22:14 jnc: well, it's the generic term for that migration, maybe :-) Apr 09 23:23:17 mmm someone yelled? Apr 09 23:23:25 * emte reads history Apr 09 23:23:31 emte: yeah, are you the nigel who mailed oe? Apr 09 23:23:55 yes i am me Apr 09 23:24:16 what did i say this time? Apr 09 23:24:18 emte: well, if you have hints on how to make the decision making more transparent Apr 09 23:24:27 emte: I would welcome them ;) Apr 09 23:24:41 decision re? Apr 09 23:25:13 emte: about the usage of OE, and forcing it on the community Apr 09 23:26:32 usage of oe ... Apr 09 23:26:42 emte: usage of monotone... Apr 09 23:26:46 sorry it is way too late Apr 09 23:26:52 hmm i mentioned something about monotone .. Apr 09 23:26:56 ah Apr 09 23:27:24 yeah i should have rephrased what i ment in my original post Apr 09 23:27:44 emte: well, I ask myself what you meant by forcing Apr 09 23:27:45 i was not aware taht any of the test repos still existed Apr 09 23:28:05 emte: they should not be used for real work, but they are all online Apr 09 23:28:22 reading the oe.org site it pretty boldly states that we use monotone and evein in the Getting Started there is no other otions Apr 09 23:28:26 options* Apr 09 23:28:46 i half agree Apr 09 23:28:48 emte: oh well, these are test SCMs, not an option Apr 09 23:29:17 if they shouldnt be used for real work why would most people use them? Apr 09 23:29:30 emte: to test other SCMs, to replace monotone Apr 09 23:29:39 but for a valid evaluation they need to be used Apr 09 23:29:44 emte: it is one thing using svn for three files Apr 09 23:29:57 yeah i know Apr 09 23:29:58 emte: sure, but for evaluation and not for work Apr 09 23:30:08 emte: so they don't belong to Getting Started Apr 09 23:30:14 they are nothing official Apr 09 23:30:22 it was more the suprise statement taht monotone hadnt been chosen yet that suprised me Apr 09 23:30:24 I simply cherry pick from .dev to branch Apr 09 23:30:34 monotone 0.26 Apr 09 23:30:53 BitKeeper -> monotone -> COMPLAINS, COMPLAINS, COMPLAINS -< ??? Apr 09 23:30:58 but it did remind me taht i need to update hg to 0.8.1 Apr 09 23:31:06 where ? hopefully will be monotone 0.26 Apr 09 23:31:15 but could be barzar or darcs (just kidding) Apr 09 23:31:43 emte: we 'decided' on monotone after BitKeeper, we could decide on something else Apr 09 23:31:47 just to see what the new gui tool is like Apr 09 23:31:54 so this is why we have 'Win Phil back' Apr 09 23:33:13 i know that Phil was maintaining a parallel repo, but liek i said i figured votes had been cast a while ago Apr 09 23:33:20 like* Apr 09 23:34:12 emte: Phil did not maintain a parallel repo Apr 09 23:34:46 no? i thought he was using the same one as kergoth ... Apr 09 23:34:57 kergoth did not maintain a parallel repo Apr 09 23:35:04 what are you smoking? ;) Apr 09 23:35:53 then i imagined the tailor fixes kergoth submitted when he was doing the monotone->svk stuff? Apr 09 23:36:49 then again i confuse myself often .. Apr 09 23:36:50 emte: when did he do that? I might miss something Apr 09 23:37:07 emte: look at the date where we switched to monotone Apr 09 23:37:09 few months ago Apr 09 23:37:14 emte: look at the date of Win Phil back Apr 09 23:37:30 yeah i was Apr 09 23:37:32 emte: no, kergoth -at that time- was not involved in 'Win Phil back' Apr 09 23:37:36 goes upto april 6 Apr 09 23:37:58 started in jan Apr 09 23:38:56 emte: look at the Wiki pages Apr 09 23:39:07 emte: no service is maintained by phil or kergoth Apr 09 23:41:57 no there isnt Apr 09 23:42:43 'night all Apr 09 23:43:18 RP: cya Apr 09 23:44:17 Aug 18 13:09:35 kergoth emte: it amuses me that i'm losing track of all my fixes to tailor because i'm not source controlling them. Apr 09 23:45:04 emte: well, where does it say "I'm offering an official svn/svk service for OE?" Apr 09 23:45:18 emte: it is just saying he worked on it Apr 09 23:45:21 it doesnt, it never was Apr 09 23:45:45 emte: so there was never an officual alternative to monotone Apr 09 23:45:56 emte: so no need to change Getting Started, or to offer an alternative Apr 09 23:46:42 correct, unless you want peopel to use the testing repos Apr 09 23:46:54 which we dont Apr 09 23:47:04 right Apr 09 23:47:18 I'm going to bed now Apr 09 23:47:29 yeah i need to go to the store for food Apr 09 23:49:09 Aug 26 07:38:46 CosmicPenguin emte: did you shut down your hg repo? Apr 09 23:49:12 whoops Apr 09 23:49:25 never told him i did ... Apr 09 23:49:38 ohwell i am sure he figured it out ... Apr 10 00:13:10 hello Apr 10 00:13:30 did anybody here run monotone on a 32mb ram box? Apr 10 00:23:53 sounds painful Apr 10 00:33:00 http://pastebin.ca/48815 can someone tell me how to fix this? seems like it cannot find pcap.h.... Apr 10 00:36:11 ljp: indeed Apr 10 00:36:21 ljp: wanted to know if its feasible or not Apr 10 00:36:45 with swap :) Apr 10 00:39:45 http://pastebin.ca/48815 can someone tell me how to fix this? seems like it cannot find pcap.h.... Apr 10 01:00:31 do any of the kernels in OE use pxamci / support SD cards / send hotplug events for card insert/removal Apr 10 01:03:15 /home/zaurus/src/build/tmp/cross/lib/gcc/arm-linux/3.4.4/../../../../arm-linux/bin/ld: cannot find -lpcap Apr 10 01:03:18 WTF? Apr 10 01:03:21 ummm... Apr 10 01:05:08 depends on libpcap apparently Apr 10 01:05:17 bitbake libpcap Apr 10 01:05:34 TheMasterMind1: I have done that... Apr 10 01:10:27 anyone tell me when bitbake won't link to pcap properly, this is really annoying.... Apr 10 01:10:38 is the library in the staging area Apr 10 01:10:58 TheMasterMind1: it's in build/tmp/work/libpcap- Apr 10 01:11:08 TheMasterMind1: where else should it be? Apr 10 01:15:14 Zero_Chaos: I seem to have a copy of libpcap.so in tmp/staging/arm-linux/lib Apr 10 01:15:27 johnX: I'll check Apr 10 01:16:22 johnX: yup, got that too Apr 10 01:17:09 so what are you trying to build that's refusing to link to pcap? Apr 10 01:17:25 johnX: rawfakeap Apr 10 01:17:35 johnX: it's my .bb Apr 10 01:20:00 unset LDFLAGS perhaps Apr 10 01:20:19 TheMasterMind1: where and why? Apr 10 01:21:11 TheMasterMind1: it's not set in conf/local.conf...... Apr 10 01:21:17 I think he means "did you set the LDFLAGS properly in the Makefile?" Apr 10 01:21:34 johnX: -lpcap it proper I think... what else do I need? Apr 10 01:22:00 to point it at the directory where the arm-linux libs are... Apr 10 01:22:41 oe_runmake LDFLAGS="${LDFLAGS}" Apr 10 01:23:14 johnX: how do I do that? Apr 10 01:23:27 TheMasterMind1: in the .bb? Apr 10 01:23:29 actually, you should do it the way TheMasterMind1 has it Apr 10 01:24:16 yes, in the .bb Apr 10 01:25:27 johnX: where? (it errored) Apr 10 01:26:03 you should look at another .bb that's made to compile a program that uses Makfiles Apr 10 01:27:03 johnX: yeah, I have done that, and this isn't my first .bb Apr 10 01:28:59 then you should know what oe_runmake is Apr 10 01:29:12 function oe_compile() { oe_runmake LDFLAGS="${LDFLAGS}" } Apr 10 01:29:14 in your bb Apr 10 01:29:15 TheMasterMind1: didn't say I was very good.... Apr 10 01:29:25 TheMasterMind1: thanks Apr 10 01:31:06 TheMasterMind1: well, I added the line, and it parses, but still ld cannot find -lpcap Apr 10 01:31:28 quilt-native-0.42 isn't building on my FC4 machine. 0.39 builds fine Do I submit bug? Apr 10 01:31:58 Zero_Chaos: if you put your .bb on pastebin I'll take a look at it Apr 10 01:32:36 johnX: k, one sec Apr 10 01:34:03 johnX: http://pastebin.ca/48819 Apr 10 01:34:36 johnX: the patch is just to remove CC = gcc Apr 10 01:42:38 johnX: any ideas? Apr 10 01:43:03 I got it to compile, but I'm not sure if what I did makes any sense, so I'm going to try and run it on my Z Apr 10 01:43:24 btw, /etc/ap_manuf is already shipped in kismet Apr 10 01:43:55 johnX: you won't be able to run it without modified drivers, and ap_manuf is acctually different than the one with kismet... but thanks for the reminder, I'll move/rename it Apr 10 01:44:15 johnX: well, you can run it, but it won't *work* Apr 10 01:44:33 well, it runs and prints the help message Apr 10 01:44:43 what I did was set CFLAGS the same way you set LDFLAGS Apr 10 01:44:54 johnX: sweet Apr 10 01:44:59 also, your patch to the makefile isn't necessary with CFLAGS set Apr 10 01:44:59 johnX: I'll try it Apr 10 01:47:46 johnX: so compile should have oe_runmake LDFLAGS="${LDFLAGS}" CFLAGS="${CFLAGS}" and no patch? Apr 10 01:47:59 That's what I have Apr 10 01:48:04 hmmm Apr 10 01:48:21 no luck? Apr 10 01:48:57 johnX: FATAL: arm-linux-strip /home/zaurus/src/build/tmp/work/rfakeap-0.2-r0/install/rfakeap/./usr/sbin/rfakeap: command failed Apr 10 01:49:00 NOTE: Couldn't find shared library provider for libpcap.so.0 Apr 10 01:49:01 ? Apr 10 01:49:46 did you clean the work directory and recompile? Apr 10 01:49:51 and there is something wrong, because it runs on my computer... which it shouldn't Apr 10 01:49:53 johnX: yeah Apr 10 01:51:47 it worked for me Apr 10 01:51:53 johnX: maybe time to rm -rf tmp/ and start over.... sometime too many changes in the branch Apr 10 01:53:06 Zero_Chaos: it runs on your computer? does file call it an ARM executable or i386/amd64? Apr 10 01:55:20 rfakeap: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), for GNU/Linux 2.6.9, dynamically linked (uses shared libs), for GNU/Linux 2.6.9, not stripped Apr 10 01:56:58 yeah, I don't think deleting tmp will help Apr 10 01:57:13 johnX: okay, well, I'm all for suggestions. Apr 10 01:57:33 johnX: I know that gcc refers to my host machine's gcc, not the cross compiler.... Apr 10 01:57:52 bitbake -b /path/to/your/file.bb -c clean Apr 10 01:58:06 johnX: and if Makefile forces CC = gcc it's not going to work Apr 10 01:58:10 TheMasterMind1: k Apr 10 01:58:30 oe_runmake linux CC="${CC}" CXX="${CXX}" CFLAGS="${CFLAGS}" CXXFLAGS="${CXXFLAGS}" LD="${LD}" LDFLAGS="${LDFLAGS}" Apr 10 01:58:44 heh Apr 10 01:58:44 err, get rid of the linux in there Apr 10 01:59:00 and don't apply your patch Apr 10 01:59:06 k Apr 10 02:00:19 TheMasterMind1: that worked, but I still get this: Apr 10 02:00:19 FATAL: arm-linux-strip /home/zaurus/src/build/tmp/work/rfakeap-0.2-r0/install/rfakeap/./usr/sbin/rfakeap: command failed Apr 10 02:00:23 NOTE: Couldn't find shared library provider for libpcap.so.0 Apr 10 02:00:26 TheMasterMind1: ideas? Apr 10 02:04:07 I find it interesting that it worked on mine... Apr 10 02:04:38 johnX: when you type "gcc", does it call arm-gcc or i686-gcc? Apr 10 02:05:38 from a bash prompt it calls 386 gcc... Apr 10 02:05:47 actually I've been looking for why the gcc bit worked on mine Apr 10 02:06:00 johnX: yeah, that is what I wondered. Apr 10 02:06:04 see if staging/arm-linux/lib has libpcap.so.0 Apr 10 02:06:29 mine doesn't have the libpcap.so.0, just libpcap.so and libpcap.so.0.9.3 Apr 10 02:06:52 johnX: as a matter of fact, mine is the same as yours, not libpcap.so.0 only the others Apr 10 02:07:00 johnX: eh, it should still work. Apr 10 02:10:40 hmm... Apr 10 02:10:54 n/m, it still didn't work, I still have a i686 executable Apr 10 02:12:28 in log.do_compile what is showing up as the compiler ? Apr 10 02:13:07 gcc Apr 10 02:13:14 hence the problem Apr 10 02:16:39 try not inheriting autotools Apr 10 02:16:53 um... okay Apr 10 02:17:52 johnX: if I do that, and apply my patch, it works properly... I wonder... WTF? Apr 10 02:18:02 don't apply your patch either, btw Apr 10 02:18:06 it's not needed Apr 10 02:18:11 johnX: but it seems to work so... here is not not complaining Apr 10 02:18:18 johnX: I'll try without the patch now Apr 10 02:18:44 I removed the inherit autotools line as soon as I got your .bb without thinking about it Apr 10 02:19:12 johnX: w00t Apr 10 02:19:17 johnX: thanks man Apr 10 02:19:22 TheMasterMind1: thanks for your help as well Apr 10 02:19:25 no prob Apr 10 02:19:32 * Zero_Chaos will go back and improve his previous .bb files... Apr 10 02:19:55 ah, and thanks to TheMasterMind1 I'll kill off a lot of unnecessary patches in my local .bb's as well too Apr 10 02:20:19 johnX: yeah, I was hoping there was a better way to do that... Apr 10 02:20:36 TheMasterMind1: you know how to force CFLAGS without patching Makefile? Apr 10 02:20:49 TheMasterMind1: for instance if I want -03 Apr 10 02:23:18 yes by setting it in the bb file Apr 10 02:24:12 TheMasterMind1: so I can just add it to that big string you sent me? or, how? Apr 10 02:24:18 * Zero_Chaos is easily confused Apr 10 02:25:14 at first I was thinking CFLAGS="${CFLAGS} -O3" but if -O2 is already set, how will gcc react? Apr 10 02:26:03 johnX: I have no idea who wins in the fight between -O -O1 -O2 -O3 Apr 10 02:26:17 johnX: probably will just -Os and say "fsck ya'" Apr 10 02:27:43 johnX: maybe just check the CFLAGS and add them all, changing what you want.... Apr 10 02:27:55 johnX: CFLAGS="-O3" Apr 10 02:29:25 johnX: well, that didn't work, there is no -O3 even in there when I check log.do_compile Apr 10 02:33:27 TheMasterMind1: how do I set -O3 in the .bb file; and does it override everything else if I do? Apr 10 02:40:44 oe_runmake CFLAGS="-O3 ${CFLAGS}" Apr 10 02:40:57 TheMasterMind1: thanks Apr 10 02:41:23 TheMasterMind1: will that override -O2, etc? Apr 10 02:42:54 TheMasterMind1: from log.do_compile: Apr 10 02:42:56 arm-linux-gcc -march=armv4 -mtune=xscale -I/home/zaurus/src/build/tmp/staging/arm-linux/include -fexpensive-optimizations -fomit-frame-pointer -frename-registers -O2 -L/home/zaurus/src/build/tmp/staging/arm-linux/lib -Wl,-rpath-link,/home/zaurus/src/build/tmp/staging/arm-linux/lib -Wl,-O1 rfakeap.c -o rfakeap Apr 10 02:43:06 TheMasterMind1: no -O3 at all.... Apr 10 02:43:12 TheMasterMind1: any ideas? Apr 10 02:47:58 um depends on the makefile and configure script, dunno. its tricky, you can grep CFLAGS */*.bb and see how some of the other packages achieve overriding CFLAGS Apr 10 02:48:27 TheMasterMind1: I'll check it out, thanks Apr 10 02:56:58 EXTRA_OEMAKE="-O3" Apr 10 02:57:13 TheMasterMind1: Rock, I'll try it Apr 10 02:57:18 TheMasterMind1: where? Apr 10 02:57:20 :-) Apr 10 02:58:27 in the bb.. and it should be EXTRA_OEMAKE = "CFLAGS='-O3 ${CFLAGS}'" i think Apr 10 02:59:34 or the -O3 after the ${CFLAGS} since i think it uses the last -O argument Apr 10 03:00:04 are optimisation flags even valid with arm ? Apr 10 03:00:18 TheMasterMind1: testing Apr 10 03:00:31 tho i guess it would be discarded if they werent Apr 10 03:01:07 emte: why wouldn't they be? gcc is gcc Apr 10 03:01:30 TheMasterMind1: ROCK! that worked Apr 10 03:01:39 * Zero_Chaos falls to his knees and worships TheMasterMind1 Apr 10 03:01:56 i dont know if anyone spent the time to do the work Apr 10 03:01:57 TheMasterMind1: thanks for the help man, I'll never have to make a stupid patch again... Apr 10 03:02:08 np Apr 10 03:02:32 i use grep on packages/*/*.bb religiously, there's a lot ot learn from the bbs already out there Apr 10 03:02:33 i know the codesorcery arm stuff isnt in gcc until 4.x series Apr 10 03:02:52 which is all the treading stuff Apr 10 03:02:57 threading* Apr 10 03:03:37 emte: heh, I don't even run gcc 4 on my host yet Apr 10 03:03:43 emte: and I'm a gentoo user... Apr 10 03:03:44 according to the gcc changelog anyway Apr 10 03:03:57 poor you Apr 10 03:04:05 emte: I know.. lazy acctually Apr 10 03:04:23 you must ... with all that free time gentoo gives you :P Apr 10 03:04:56 yeah i know you ment gcc4.x Apr 10 03:06:31 anyway i am under the impression taht all the arm optimisation stuff is built into the tune stuff Apr 10 03:07:00 emte: I'm going sleepy now, so don't feel bad if I don't respond :-) Apr 10 03:07:04 night all Apr 10 03:07:08 lol Apr 10 03:07:11 night Apr 10 03:07:52 -mtune=xscale Apr 10 05:29:05 I built the qte package on the org.openzaurus.354x branch, but I noticed that the resulting library has a different file size than that in the OZ 3.5.4.1-alpha1 release Apr 10 05:29:18 and when I tried to use it, it seems to cause Opie to crash Apr 10 05:29:34 You might have a different toolchain? Apr 10 05:30:11 doesn't oe use it's own toolchain? Apr 10 05:30:36 Yes, but what I'm thinking is you may have a dfiferent version of the toolchain, or different settings Apr 10 05:30:52 ah, you might be right Apr 10 05:31:00 i think I do have another toolchain in the path Apr 10 05:31:06 ok, i'll try removing that Apr 10 06:51:44 why do people keep removing the "note: download oe.db" remarks from gettingstarted? Apr 10 06:51:48 * koen reverts Apr 10 06:54:23 hello Apr 10 07:14:11 hey Genesis Apr 10 07:16:39 koen: Because they oppose monotone, and try to subvert proper usage.. Apr 10 07:16:46 That's my guess at least. Apr 10 07:21:10 maybe I should add a note saying "We know we are saying it three times, but saying it one time isn't enough" Apr 10 07:31:23 morning Apr 10 07:31:24 ~ewi Apr 10 07:31:25 hmm... ewi is ewi546.ewi.utwente.nl - main backup server for OE monotone. Hosts database snapshot: http://ewi546.ewi.utwente.nl/OE/OE.db.bz2 Apr 10 07:31:31 hey hrw|work Apr 10 07:31:43 koen: db snapshots exist still? Apr 10 07:31:49 they should Apr 10 07:32:36 ok fetching Apr 10 07:32:44 had problem with fetching it today Apr 10 07:36:15 approx. 2 hours ago? Apr 10 07:36:27 rather 1-1.5 Apr 10 07:36:57 the cronjob ran 2 hours and 10 minutes ago Apr 10 07:38:36 hmm. the problem with releasing 3.5.4.1-alpha2 is that first I have to provide all updates to upgrades feed.. Apr 10 07:44:51 morning Apr 10 07:46:38 hi XorA Apr 10 07:46:45 booting from SD is nice Apr 10 07:47:44 hrw|work: can you ipkg install firefox/gimp reliable when root is on SD? Apr 10 07:48:47 hrw|work: my ipkg always enters D state when I try that :-( Apr 10 07:50:47 someone needs to show Negroponte OE :-) Apr 10 07:53:39 XorA: I can Apr 10 07:53:54 btw - firefox 1.5.0.1/.dev works? Apr 10 07:54:49 hey XorA & mallum Apr 10 07:54:51 bbl Apr 10 07:54:58 hey koen|away Apr 10 07:57:46 bah, my .dev build still tries to build hostap-modules :-( Apr 10 07:58:14 hrw|work: no Apr 10 07:58:55 hrw|work: it builds, runs, but never opens a window Apr 10 07:59:10 morning all Apr 10 07:59:29 hey RP Apr 10 08:01:19 good morning Apr 10 08:05:03 hi Bernardo RP Apr 10 08:05:04 XorA: ;( Apr 10 08:05:37 hrw|work: koen told me same bug inflicts recent minimo as well, on the rare occasions we can build it Apr 10 08:21:01 hrw|work: did you ever get all your SD cards working reliably? Apr 10 08:21:13 without timeout and detection errors Apr 10 08:21:57 TheMasterMind1: nope Apr 10 08:22:19 TheMasterMind1: I got new card from user and that one just works Apr 10 08:22:31 weird Apr 10 08:23:57 the code that enables SD card usage with pxamci - is that implemented model-specifically or in the general driver.. do you know? Apr 10 08:25:58 sd<>mmc is in mmc core iirc Apr 10 08:26:36 ok thats what it looked like from the patches i looked through, but i wasn't sure Apr 10 09:03:39 hey Apr 10 09:03:47 hail zecke Apr 10 09:03:54 hi zecke Apr 10 09:04:17 ave phil Apr 10 09:04:51 gosh ccache and precompiled headers are playing tricks on each other Apr 10 09:21:47 meeting... Apr 10 09:59:49 who broke the c7x0 keymap? **** ENDING LOGGING AT Mon Apr 10 09:59:56 2006