**** BEGIN LOGGING AT Mon Apr 17 09:59:56 2006 Apr 17 10:00:30 RP: ok Apr 17 11:15:51 morning Apr 17 11:28:58 ~lart me for being too stupid to use bitbake/oe Apr 17 11:28:58 * ibot hooks into a hydrant and hoses xora down for being too stupid to use bitbake/oe Apr 17 11:40:05 koen: got an update to abiword-plugins for you Apr 17 11:41:24 XorA: cool, feel free to push Apr 17 11:46:11 03xora 07org.oe.dev * r0a660fa9... 10/packages/abiword/abiword-plugins_2.4.4.bb: Apr 17 11:46:11 abiword-plugins_2.4.4.bb : Add patch to make it build with poppler thats Apr 17 11:46:11 in the repo. Break the plugins out into a package each for better Apr 17 11:46:11 distribution. Apr 17 11:47:01 koen: there :-) Apr 17 11:47:14 cool, thanks Apr 17 11:48:53 what happened to hh.org emails today then? Apr 17 11:49:02 or is it just quiet Apr 17 11:59:53 Real: http://www.mailingmania.com/pages/ptp.php?refid=matrixz Apr 17 11:59:55 Real: http://www.mailingmania.com/pages/ptp.php?refid=matrixz Apr 17 11:59:57 Real: http://www.mailingmania.com/pages/ptp.php?refid=matrixz Apr 17 12:00:16 Fucking spammers. Apr 17 12:00:23 kick this guy Apr 17 12:00:30 heh Apr 17 12:00:45 CoreDump|home: Now that's power! Apr 17 12:01:40 XorA: it's down again Apr 17 12:02:26 XorA: don't forget to push the patch as well ;) Apr 17 12:05:16 JustinP: this behaviour is expected. I dunno if that is good or not, though Apr 17 12:06:07 ~ping RP Apr 17 12:06:08 pong RP Apr 17 12:13:33 koen: ooop Apr 17 12:14:46 erm.... trying to build snes9x-sdl-qpe, and i'm getting this error: arm-linux/bin/ld: cannot find -lqte Apr 17 12:15:19 is -lqte supposed to be something different? or what? :/ Apr 17 12:21:40 koen: fixed Apr 17 12:22:16 03xora 07org.oe.dev * r1db2f2b8... 10/packages/abiword/files/abiword-plugin-pdf-poppler.patch: abiword-plugins : add forgotten patch Apr 17 12:34:03 03rwhitby 07org.oe.dev * r2e6ce2c1... 10/packages/meta/slugos-packages.bb: slugos-packages: demoted openobex (fails to configure) Apr 17 12:34:54 koen: tried epdfview, its bad :-( needs internal files from a particular poppler version :-( Apr 17 12:35:45 epdfview is an enlightenment app? Apr 17 12:36:02 koen: no, a small pdf viewer based on poppler Apr 17 12:41:07 cedric__: pong Apr 17 12:42:26 XorA: do you have an url for that pdf viewer? Apr 17 12:42:55 koen: http://www.emma-soft.com/projects/epdfview/wiki Apr 17 12:45:53 XorA: did you try the svn version? Apr 17 12:46:25 koen: well since release was yesterday I figured it was just bolloxed :-) Apr 17 12:47:19 http://www.emma-soft.com/eblog/post/7/ Apr 17 12:47:20 koen: ah, I see, they updated to poppler0.5.1 Apr 17 12:47:28 koen: I shall do that version then Apr 17 12:49:16 :) Apr 17 12:50:28 (Sorry for this nOOb question) in GettingStarted instructions when obtaining OE, step 7: what's the difference between the repositoy's local database and the branches ? Apr 17 12:51:05 I just want to test the oz354fam083 branch, do I need a dev or dreambox repo too ? Apr 17 12:52:03 morning Apr 17 12:52:19 hey TheMasterMind1 Apr 17 12:52:20 cromain: I *think* you just need to check out org.openembedded.oz354x Apr 17 12:52:28 cromain: I'm new to this myself, tho... Apr 17 12:53:18 rm_you: so I can update the wiki with {dev,dreambox,oz354fam083,oz354x} instead of {dev,dreambox} in fact ? Apr 17 12:53:39 erm.... dunno about that Apr 17 12:55:41 cromain: yes Apr 17 12:55:59 cromain: see the monotone docs the wiki links to (venge.net/...) Apr 17 12:56:25 koen: thanks, I'll do. So then what's the difference between "monotone --db=/stuff/OE.db checkout --branch=org.openembedded.dev" and "cd /stuff/org.openembedded.dev; monotone update" ? Apr 17 12:57:40 the first creates a new checkout (reads all files from the database) the second updates the checkout to the latest revision in the database Apr 17 12:58:32 ok, if I good understand, the first one must be done once, and the second command is to keep up-to-date Apr 17 13:00:44 hrm, so does anyone know why I'm getting this error compiling snes9x-sdl-qpe? arm-linux/bin/ld: cannot find -lqte Apr 17 13:02:54 * Philippe is back (gone 13:49:28) Apr 17 13:03:03 cromain: yes Apr 17 13:03:18 hi guys Apr 17 13:03:23 cromain: and you use 'monotone pull' to update the database Apr 17 13:03:25 hey mickeyl Apr 17 13:03:31 rm_you: let me try to reproduce that... Apr 17 13:03:44 mickeyl: kk... Apr 17 13:03:47 rm_you: there have been some changes in qmake/qt handling, perhaps i broke something Apr 17 13:04:02 this is .dev, right? Apr 17 13:04:06 I noticed it isnt in the feed for 3.5.4, so... Apr 17 13:04:15 actually no, this is ox354x Apr 17 13:04:18 oh Apr 17 13:04:20 s/ox/oz/ Apr 17 13:04:30 pastebin the complete log.do_compile Apr 17 13:04:33 I can test in .dev and see if it does the same thing... Apr 17 13:04:38 yeah, that'd be good Apr 17 13:04:54 kk sec pastebinning Apr 17 13:05:39 http://pastebin.com/664907 Apr 17 13:06:28 hmm yeah, the path to the qt staging area is missing Apr 17 13:06:34 i'll have a go at that in .dev Apr 17 13:06:38 k. Apr 17 13:06:52 That would probably be why the package dissapeared from the feed for this release :P Apr 17 13:07:03 yep Apr 17 13:08:05 hi mickeyl Apr 17 13:08:10 hail RP Apr 17 13:09:33 hi zecke Apr 17 13:09:33 zecke: mornin' Apr 17 13:09:41 hey zecke Apr 17 13:10:55 mickeyl: well, if you look into that and it turns out to be a crazy simple fix, tell me so I can patch this up and get it compiling so I can install it :P Apr 17 13:11:33 ya Apr 17 13:14:28 hey Apr 17 13:14:33 morning zecke Apr 17 13:15:13 pb__: hey Apr 17 13:15:34 OT: can i get a quick recommendation for the "best" dvb-t app on linux? Apr 17 13:15:48 * mickeyl dvb-t newbie, anything for a begginer might do as well Apr 17 13:15:49 pb__: the libtool-native error was ... Apr 17 13:16:02 mickeyl: dvbnet rocks, IP over MPEG2 is just awesome? Apr 17 13:16:07 mickeyl: you want to have a VDR? Apr 17 13:16:23 zecke: perhaps later. at first trying to get that piece of hardware working by showing tv Apr 17 13:17:08 mickeyl: tzap Apr 17 13:17:13 zecke: yo Apr 17 13:17:16 mickeyl: and dvb snoop (let me look up the names) Apr 17 13:17:18 mickeyl: yo Apr 17 13:17:23 hi pb_ Apr 17 13:17:34 pb_: where should I change the osb stuff, in .dev and the branches? Apr 17 13:17:51 zecke: yes, please. I think it should be changed everywhere. Apr 17 13:19:03 pb_: I'm about to say: "Get yourself a f$!â@... oe account" Apr 17 13:19:11 zecke: drat Apr 17 13:19:18 hehe Apr 17 13:19:19 pb_: okay, I will adjust it, Apr 17 13:19:28 thanks Apr 17 13:19:49 zecke: be carefull, its maintainer is know to revert without asking Apr 17 13:20:00 koen: psst Apr 17 13:20:32 EIBA uses VFP correct. non-EIBA uses FPA? Apr 17 13:20:40 koen: I can blame the evil mind of pb_ and request france to shutdown the GPE project *muahhaaa* Apr 17 13:20:51 zecke: good thinking Apr 17 13:20:53 zecke: and I object to pointing SRC_URIs to random distro mirrors that might vanish Apr 17 13:21:08 TheMasterMind1: eabi uses vfp, non eabi is free to choose Apr 17 13:21:10 koen: that was one of the questions I have/had Apr 17 13:21:32 pb_: is that a permanent location for this source? Apr 17 13:21:47 and as we can see form the mailinglist issues hh.org is unreliable Apr 17 13:22:07 koen: NOTE: package epdfview-0.1.1+svn20060418: completed Apr 17 13:22:10 zecke: permanent enough. but, if koen insists on making a fuss, please make that change only in the oz354fam083 branch and leave it alone in .dev. Apr 17 13:22:31 zecke: in .oz354fam084 the local.conf.familiar-0.8.3 already points to there Apr 17 13:22:43 zecke: the CVS_TARBALL_STASH that is Apr 17 13:23:56 and is CVSDATE fixed as well? Apr 17 13:24:12 zecke: unless someone reverted it, it is Apr 17 13:25:15 pb_: it is your call then Apr 17 13:25:53 zecke: I never finished moving stuff from local.conf.familiar-0.8.3 to familiar-0.8.3.conf because france+gang took over the project Apr 17 13:26:14 zecke: I would prefer to have the SRC_URIs changed, but I don't have the heart to fight about it. Maybe it would be best just to leave them alone then. Apr 17 13:27:03 I guess I can easily enough work around it locally. Apr 17 13:28:10 pb_: well, I can survive possible beat ups... Apr 17 13:28:31 rm_you: 1.) inherit qmake-base Apr 17 13:28:42 2.) add -L${OE_QMAKE_LIBDIR_QT} in LDLIBS Apr 17 13:29:18 mickeyl: erm... k... I'll try to understand what that means :) Apr 17 13:29:45 rm_you: alternatively see the diff i will commit to .dev Apr 17 13:29:48 k Apr 17 13:29:59 zecke: so, anyway, what was the cause of your libtool thing? Apr 17 13:30:38 pb_: SRC_URI_append didn't work... Apr 17 13:30:49 zecke: aha Apr 17 13:32:02 r449 is my new favourite revision Apr 17 13:32:48 koen: what is trunk? Apr 17 13:32:48 sigh. who had the bright idea of swapping the word order on ARM anyway Apr 17 13:32:53 zecke: r449 :) Apr 17 13:33:46 TheMasterMind1: weitek Apr 17 13:33:54 koen: we will have a stable branch for 1.4 as well Apr 17 13:34:02 koen: so things should remain stable Apr 17 13:35:50 koen: epdfview is insanely fast Apr 17 13:36:00 XorA: cute Apr 17 13:36:14 XorA: and it's even using cairo, so it could be faster ;) Apr 17 13:36:14 koen: a nightmare to build from svn, but fast as hell Apr 17 13:37:05 heh Apr 17 13:37:28 XorA: I'm preparing to build OOo natively on my ARM boards using distcc Apr 17 13:37:29 koen: I had to manually copy a load of Makefile.* from the release to svn to build it Apr 17 13:37:52 03mickeyl 07org.oe.dev * r0b09834f... 10/packages/snes9x/ (snes9x-sdl-qpe_1.39.bb snes9x-sdl_1.39.bb): Apr 17 13:37:52 snes9x-sdl-qpe: fix linking Apr 17 13:37:52 NOTE: shouldn't sdl-config take care about that in the qpe case? Apr 17 13:37:56 'nightmare' doesn't come close to describint that Apr 17 13:39:20 awesome :P thanks Apr 17 13:40:03 hrm... should i pull of .dev grab that? Apr 17 13:40:14 s/ i / a / Apr 17 13:40:22 koen: needs some gui work, but I will keep an eye on that rpoject Apr 17 13:40:38 rm_you: yeah Apr 17 13:40:54 rm_you: you can also just take a look at the oe-commits list where the diff comes up in clear text Apr 17 13:41:09 that is.. if hh.org mail would be working :D Apr 17 13:41:35 I'm new to this whole process so.... :/ I pulled it, and i'm doing the change manually Apr 17 13:42:19 #if defined(__IEEE_LITTLE_ENDIAN) || (defined(__IEEE_BIG_ENDIAN) && defined(__IEEE_BYTES_LITTLE_ENDIAN)) Apr 17 13:42:25 is that valid? Apr 17 13:43:28 mickeyl: sdl-config shouldn't need to care about qpe: libsdl should (generally) take care of itself. Apr 17 13:43:58 though, admittedly, I have no idea what your change actually was, so it might well have been justified Apr 17 13:43:58 heh Apr 17 13:44:31 dunno. didn't look at sdl after a lot of people tweaked that Apr 17 13:44:44 it didn't find -lqpe which was explicitly given here Apr 17 13:44:47 (for whatever reason) Apr 17 13:45:02 guess i need to clean that up eventually Apr 17 13:45:52 ah right Apr 17 13:46:03 hrm... well, I apparently missed something... Apr 17 13:46:12 how many lines did you change? Apr 17 13:46:28 it does sound like fixing snes9x-sdl-qpe is the right thing then, though I don't entirely understand why there needs to be a special qpe version of snes9x-sdl at all Apr 17 13:46:38 does the regular sdl build not work with qpe? Apr 17 13:47:13 good question... Apr 17 13:47:15 pb_: .desktop files (sdl-qpe) Apr 17 13:47:35 zecke: linking with -lqpe shouldn't be needed for those, though, right? Apr 17 13:48:01 pb_: libsdl should have a SO_NEEDED which should pull the library in right? Apr 17 13:48:08 zecke: right, exactly Apr 17 13:48:17 well, DT_NEEDED to be precise, but yes Apr 17 13:50:12 ok.. so it IS fixed in .dev? Apr 17 13:50:31 i wonder if i ever fixed/changed sdl-config Apr 17 13:51:12 cya later, I need to study a bit Apr 17 13:53:38 ah, there we go, just took a couple minutes to get in there properly, pull finally updates it right :P Apr 17 13:54:59 wow I guess there's more broken than just that in oz354x :/ oh well Apr 17 13:55:12 rm_you: broken in what way? Apr 17 13:55:14 koen: another release of epdfview in last 10 mins :-) Apr 17 13:55:40 zecke: snes9x explodes in a different way now that I applied that patch Apr 17 13:57:05 http://openembedded.pastebin.com/664988 Apr 17 13:57:31 I guess I'll just have to wait for the fixed version to work its way into a real release, or switch to using .dev Apr 17 13:58:14 rm_you: fix it for us? *please* Apr 17 13:58:26 lol? Apr 17 13:58:35 rm_you: e.g. grep for main? Apr 17 13:58:47 zecke: fix snes9x in .oz354x? Apr 17 13:59:23 rm_you: most of it is compiled, it can't be too hard to get the rest running Apr 17 13:59:30 maybe Apr 17 13:59:33 I'll look at it Apr 17 14:04:58 XorA: heh Apr 17 14:07:44 * CoreDump|home still fights with kexec support Apr 17 14:08:52 mickeyl: I'm not having much luck with adding efl++ 0.1.0 to OE :( Apr 17 14:15:45 would bitbake support http proxy? Apr 17 14:19:50 http://www.netbsd.org/contrib/projects.html#qtopia wow Apr 17 14:20:17 Readon: that depends. for http/ftp it uses wget... Apr 17 14:20:21 03xora 07org.oe.dev * ra1c32f4b... 10/packages/epdfview/epdfview_0.1.2.bb: epdfview_0.1.2.bb : A new PDF viewer, very minimal based on GTK+ and poppler. Apr 17 14:20:38 Readon: see bitbake.conf and search for FETCHCOMMAND (or was it FETCH_COMMAND) Apr 17 14:22:19 koen: give that a try then Apr 17 14:23:58 How to i force bitabke to use a arm-linux-gcc-3.4.1 insted of 3.4.4 ? Apr 17 14:29:27 XorA: 28kB is pretty slick for a pdf viewer Apr 17 14:29:35 XorA: too bad poppler is 0.5MB Apr 17 14:29:41 koen: heh heh Apr 17 14:30:49 koen: commit the nonworking 0.1.1 and I'll have a look Apr 17 14:30:58 koen: a little more TLC and that would make a good PDF viewer for GPE Apr 17 14:32:04 * XorA wishes Z had 128M of ram Apr 17 14:32:54 XorA: I used to use my c700 as a display for my h5550 Apr 17 14:33:17 koen: gaffa tape them together? Apr 17 14:33:40 XorA: :) Apr 17 14:34:55 zecke: svn is also used Apr 17 14:35:02 mickeyl: ok, pushed Apr 17 14:35:47 03koen 07org.oe.dev * rb4cc2cb8... 10/packages/efl++/ (efl++-fb_0.1.0.bb efl++-x11_0.1.0.bb): efl++: add first cut at 0.1.0 Apr 17 14:46:53 please help me out here? "How to i force bitabke to use a arm-linux-gcc-3.4.1 insted of 3.4.4 " Apr 17 14:51:06 goxboxlive: force? Apr 17 14:51:31 goxboxlive: and you can't force bitbake. It is a task executor, it executes tasks (compile,unpack, fetch, patch, clean...) Apr 17 14:51:57 goxboxlive: check if OE has a gcc 3.4.1, and then set PREFERRED_VERSION (see the bitbake manual for that) for the compiler to 3.4.1 Apr 17 14:56:47 goxboxlive: or do you want to use a pre existing cross compiler? Apr 17 14:59:06 zecke: Yes i want to use my arm compiler (3.4.1) becuase my kernel refuse to bild using the 3.4.4, it will only compile with arm-linux-gcc-3.4.1 Apr 17 14:59:20 bild=build Apr 17 15:00:23 goxboxlive: oh well ;) Apr 17 15:00:29 goxboxlive: KERNEL_CCSUFFIX or such Apr 17 15:00:55 ah, awesome, my gpe-image build completed without further errors Apr 17 15:00:58 * pb_ celebrates quietly Apr 17 15:01:08 RP: how can I find out at runtime (in the kernel level) the total amount of RAM Apr 17 15:01:19 goxboxlive: KERNEL_CCSUFFIX = "-3.3.4" from the handhelds.org kernel bbfile Apr 17 15:01:35 pb_: for me dbus is not fetchable :} Apr 17 15:01:52 zecke: oh dear Apr 17 15:02:04 03xora 07org.oe.dev * r8624698b... 10/packages/enchant/enchant_1.2.5.bb: enchant_1.2.5.bb : dont package the .la files in main package. Apr 17 15:02:15 pb_: I should set another TARBALL_STASH Apr 17 15:02:23 cedric__: cat /proc/meminfo ? Apr 17 15:02:30 zecke: I'm in the familiar-0.8.4 ghetto, so my dbus is ancient. Apr 17 15:02:46 pb_: ah now f.d.o is faster Apr 17 15:02:58 pb_: yeah, I'm building from the same branch Apr 17 15:02:58 pb_: 0.23.4 Apr 17 15:03:09 oh right. yeah, 0.23.4 is what I have as well. Apr 17 15:03:17 I figured you would be on the HEAD :-} Apr 17 15:03:45 pb_: I promised to test bitbake on the branch (promised to myself) Apr 17 15:03:51 pb_: I'm about to abuse you Apr 17 15:03:53 very good Apr 17 15:03:56 (testing, not abuse) Apr 17 15:04:09 .dev seems to produce working gpe images at the moment as well which makes a nice change :) Apr 17 15:04:11 * pb_ prepares for the flood of invective Apr 17 15:04:12 RP: is there not a more elegant way? Apr 17 15:04:36 RP: that certainly is good news Apr 17 15:04:47 cedric__: You mean from within a kernel module? Apr 17 15:04:48 pb_: Have you heard about X-Extensions? Apr 17 15:04:58 RP: yep Apr 17 15:05:00 pb_: I wonder how XDamage is working (xserver side) Apr 17 15:05:12 pb_: how does it 'intercept' the boxes that are drawn? Apr 17 15:05:20 pb_: how does it work without changing the server :} Apr 17 15:05:51 zecke: it depends on support code in the server. You can't just retro-fit Damage to a system without upgrading the server. Apr 17 15:06:00 cedric__: I suspect this is specific to a pxa2xx, right? Apr 17 15:06:24 zecke: the server obviously knows which areas need to be drawn, because it is responsible for drawing them. So it basically has all the information available to it already. Apr 17 15:06:28 pb_: okay, this is what I had thougt... Apr 17 15:06:49 pb_: so it just informs the XDamage code to inform a 'clients' of this extension Apr 17 15:06:57 RP: yep, it's for cpu-pxa25x.c file Apr 17 15:07:19 Has anyone looked at the 770's xserver xdamage code? pure evil :) Apr 17 15:07:26 zecke: more or less, yeah. If you want to look at the code, it's in xserver/damageext/ Apr 17 15:07:42 RP: it really is Apr 17 15:07:47 RP: crumbs, the 770 has a specific Damage implementation? Apr 17 15:07:50 pb_: I have it open (viewcvs of cvs.netbsd.org) :} Apr 17 15:07:55 RP: like running jffs2 on a microdrive ;) Apr 17 15:08:18 pb_: Limited display bandwidth so for speed, they only update the bits that change Apr 17 15:08:21 pb_: yeah, the 770 has it at the driver level to save the little bandwidth it has Apr 17 15:08:44 koen: redraws still look slow :} Apr 17 15:08:57 RP: oh right. I thought the xserver did that on all platforms, actually. Apr 17 15:09:16 after all, no machine has infinite display bandwidth Apr 17 15:10:06 pb_: not even turing machinesà *damn* Apr 17 15:10:09 pb_: It specifically tells the kernel which bits have changed so the kernel framebuffer driver can benefit from that information Apr 17 15:10:13 Debugging it was "fun" :-/ Apr 17 15:10:24 * zecke designs a new calculation model with infinite disp. bandwidth Apr 17 15:10:29 wow Apr 17 15:10:33 what does the kernel do with that information? Apr 17 15:10:56 zecke: I guess turing machines have infinite time, so bandwidth is irrelevant Apr 17 15:11:23 iirc the kernel updates the screen from the shadow framebuffer Apr 17 15:11:50 JustinP: would you object to removing the e*20060128 files? Apr 17 15:11:51 You can learn this much from the sources that are already available. I'm not sure how much more I can say :) Apr 17 15:12:01 JustinP: i think it's enough to have the releases in there Apr 17 15:12:30 bleh Apr 17 15:12:35 ~lart broken .la files Apr 17 15:12:35 * ibot squeezes broken .la files till broken .la files turns blue like papa smurf Apr 17 15:12:35 RP: that does sound pretty crazy. I wonder why Nokia decided to do that. Apr 17 15:12:59 pb_: I hope the display is not connected over i2c :} Apr 17 15:13:12 zecke: that would indeed be unfortunate Apr 17 15:13:19 koen: about wiki update: can I remove step 4 and 5 no longer needed since you inserted download+bunzip2 of the database in step 3 ? Apr 17 15:13:27 pb_: My guess would be they got a good deal on the displays but this is pure speculation :) Apr 17 15:13:33 cromain: sure Apr 17 15:14:26 koen: ok, I'm updating the wiki (locally because I'm not a fast-english-writer) but I think this evening it'll be ok. Thanks :) Apr 17 15:15:42 RP: ah. you mean the display is structured so that the framebuffer isn't actually mappable? I suppose that would make some kind of sense. Apr 17 15:17:41 pb_: Essentially, yes, its not directly mapped Apr 17 15:18:08 funky Apr 17 15:18:41 what display controller is in the 770? Apr 17 15:18:58 http://www.washingtonpost.com/wp-dyn/content/article/2006/04/15/AR2006041500125.html Apr 17 15:19:22 A very bad review of the 770 Apr 17 15:19:44 pb_: I honestly don't know who makes it Apr 17 15:20:04 I suspect I miss some oe/bitbake subtletly here, but I changed abiword-plugins to spit out a seperate abiword-plugin-* file for each plugin, but each plugin depends on the non existant abiword-plugins, what have I missed? Apr 17 15:20:22 ah, fair enough Apr 17 15:20:33 I guess I should open up my 770 some day and have a look inside Apr 17 15:21:25 XorA: hard to say without knowing what you changed Apr 17 15:23:13 XorA: maybe some shlibs cruft? Apr 17 15:23:25 but, most likely, you failed to supply an appropriate extra_depends to your do_split_packages() invocation Apr 17 15:23:26 XorA: did an abiword-plugins ever exist? Apr 17 15:24:11 yeah, could also be stale shlibs data in your staging area. that's always been a bit of a problem with no good solution. Apr 17 15:24:43 zecke: yes it did, but Ive cleaned shlibs stuff out Apr 17 15:25:11 pb_: ah, what would be an example of an appropriate one? Apr 17 15:25:16 It could also be some cached debian.bbclass package renaming data Apr 17 15:25:38 XorA: what dependencies do you want? Apr 17 15:26:14 pb_: I dont want abiword-plugins, I want the ones shlibs adds Apr 17 15:26:25 RP: where does that data hide? Apr 17 15:26:31 and nothing else? in that case, extra_depends='' would be the thing Apr 17 15:26:47 XorA: tmp/staging/pkgmaps Apr 17 15:26:58 you always get shlibs dependencies; they don't count as extra. Apr 17 15:33:54 is a sed guru here who could give me a hand? I'm searching for a regexp that extracts numbers from strings like "jdhhd33ljddsj". I do not know how many numbers are in the string but they are all next to each other. Apr 17 15:33:58 pb_: cheers, that cured it Apr 17 15:34:07 XorA: very good Apr 17 15:35:09 CoreDump|home: Might tr not work better? Apr 17 15:35:56 * CoreDump|home reads tr's manpage Apr 17 15:38:11 CoreDump|home: Something like .*([0-9]*).* would match just a number Apr 17 15:38:30 RP: correct, but only one of them =) Apr 17 15:38:48 sed -e 's|.*(\d*).*|\1|g' Apr 17 15:38:50 I would need a generic regexp matching 1 to 3 numbers Apr 17 15:39:35 XorA: busybox sed can't handle \d :\ Apr 17 15:39:50 CoreDump|home: replace with [0-9] then Apr 17 15:40:21 and it doesn't handle the "*" =) That is my problem heh Apr 17 15:41:05 busybox sed doesn't implement *? that sounds like a grave deficiency Apr 17 15:41:06 heh Apr 17 15:41:37 I guess you could use tr -C -d "[0-9]", but busybox probably doesn't support those options either :-} Apr 17 15:42:56 in case of last resort, maybe something like sed 's/[^0-9]//g' Apr 17 15:43:03 echo "jas12yy" |sed "s/.*\(\d*\).*/\1/g" Apr 17 15:43:05 pretty gruesome though Apr 17 15:43:13 that doesn't even work on bash err Apr 17 15:43:36 CoreDump|home: The shell is probably expanding the stars Apr 17 15:43:54 yeah, it probably wouldn't. there's nothing to stop the first .* matching everything. Apr 17 15:44:06 drat Apr 17 15:44:18 you'd need to either change that to match specifically non-digits, or specify the number of digits you want. Apr 17 15:44:30 hmm Apr 17 15:44:39 looks like it :\ Apr 17 15:45:09 or make the .* minimal (.*?) Apr 17 15:45:34 I built meta-sdk, but it lacks one package I need. Can an .ipk be added, or do I need to make my own sdk-package, based on the default meta-sdk.bb ? Apr 17 15:46:28 Cockroach-: You can add it (and bump PR so it noticed the changes) Apr 17 15:47:06 CoreDump|home: try 's/[^0-9]*\([0-9]\+\).*/\1/' or some such. Apr 17 15:47:18 or, yeah, make the first .* be minimal if busybox understands that Apr 17 15:47:28 * CoreDump|home tries, thanks! Apr 17 15:47:47 RP: Ok, but instead of changing the PR I practise that BBFILE_COLLECTIONS trick with priorities Apr 17 15:49:03 * pb_ stabs ipkg Apr 17 15:49:05 not installing gpe-calendar-locale-de due to requirement for gpe-calendar Apr 17 15:49:08 that's no good Apr 17 15:54:35 I really wish hitting right shift on this lappy didnt also register as enter :-) Apr 17 15:59:33 orning Apr 17 16:14:44 mickeyl: python-core is native? Apr 17 16:14:59 does anone know of a succesful swt(java) compilation for arm? Apr 17 16:15:55 katossi: TheMasterMind1 just put a load of patches for java stuff in the bugtracker Apr 17 16:16:19 I'll take a look Apr 17 16:16:23 katossi: swt? Apr 17 16:16:39 yes, the gui toolkit Apr 17 16:17:02 used in eclipse and azureus for example Apr 17 16:17:06 I see the oe list is running again ... Apr 17 16:17:36 swing is definitely too much for our little machines Apr 17 16:18:11 katossi: swing just runs fine on a SIMpad (SA11xx) Apr 17 16:18:32 really? what vm? Apr 17 16:19:01 my experiences were honestly dissapointing Apr 17 16:20:27 TheMasterMind1: on steroids? Apr 17 16:20:38 katossi: same as the last time when we chated Apr 17 16:20:45 katossi: gcj from gcc 4.0.1 Apr 17 16:21:25 JustinP: could you consider fixing wrong DEPENDS for native packages? Apr 17 16:21:28 zecke: sorry my memory is really crap Apr 17 16:22:18 pb_: why isn't gpe embracing dogtail? Apr 17 16:22:44 zecke: but I thought the last time we talked you said only basic programs had been tested so far Apr 17 16:24:17 zecke: in fact I had never heard of dogtail until you mentioned it. Apr 17 16:24:25 but now I have, and maybe gpe will begin to embrace it Apr 17 16:24:26 heh Apr 17 16:24:37 it does sound quite cool Apr 17 16:25:13 dogtail? Apr 17 16:25:39 mithro: automated usability testing via scripts Apr 17 16:25:40 mithro: it uses ATK to control applications Apr 17 16:25:43 mithro: and more cool stuff Apr 17 16:25:55 mithro: you can write test cases with it Apr 17 16:25:58 ATK? :P Apr 17 16:26:07 accessibility toolkit Apr 17 16:26:15 mithro: How's the C parser going? Apr 17 16:26:16 ahh okay Apr 17 16:26:21 zecke: any special tricks needed to build gcj? Apr 17 16:26:35 RP: well its up to the point that it needs basic unit tests to make sure it works okay Apr 17 16:26:41 mithro: developed by SUN (IIRC), and utilized by gtk+ Apr 17 16:26:45 and then needs the ability to register new handlers Apr 17 16:26:48 koen: gcc 4.0.1 builds it for arm Apr 17 16:27:10 zecke: only if you do JAVA_arm = ",java" Apr 17 16:27:15 mithro: handlers? well this is possible already? Apr 17 16:27:29 zecke: no the functions are empty Apr 17 16:27:33 but they will be trivial to add Apr 17 16:27:38 koen: hmm Apr 17 16:27:44 mithro: nice work. I can't wait to see it in action :) Apr 17 16:27:50 i just want to make sure everything else is working right first Apr 17 16:27:54 koen: gcc4.0.1 should build gcj out of the box, unless some one reverted aeh merged it Apr 17 16:28:47 zecke: apart from the fact that 4.0.1 isn't in OE? Apr 17 16:30:19 only problem is i'm not going to have any time to work on it for the next couple of weeks, easter break is the only reason I managed to get any work on it done Apr 17 16:30:40 RP: if you want to finish it off you'll most probably find it pretty trivial ;) Apr 17 16:30:50 koen: well someone reverted my change then Apr 17 16:30:53 bastards Apr 17 16:31:08 I really hate that, people should use svn and they are still outsmarted... Apr 17 16:31:42 mithro: Last time I looked at it, I got confused but I might have a look at it again if I get a chance. I'm also using Easter to catch up on things I don't have time for :-/ Apr 17 16:32:13 Today is kernel hacking - try and clean out my inbox and pending patch pile... Apr 17 16:32:34 koen: thanks to phil gcj is working on gcc >= 4 for ARM Apr 17 16:32:37 RP: well it's a lot simplar now - you don't have to worry about any of the C code Apr 17 16:32:39 zecke: I know Apr 17 16:32:50 zecke: I want to enable it for gcc 4.1.0 Apr 17 16:33:13 mithro: The C code is the bit I understood :) Apr 17 16:33:24 he he Apr 17 16:34:08 RP: i'm using pyrex so it's pretty simple :) Apr 17 16:37:38 koen: we should start it by copying the various gcc3*.inc -> gcc4*.inc Apr 17 16:37:42 to change some defaults Apr 17 16:39:24 zecke: right Apr 17 16:39:45 zecke: right now I'll confine the change to gcc 4.1.0 Apr 17 16:40:08 RP: we sometimes do not send the build started event Apr 17 16:40:17 RP: I have no clue how this can/could happen Apr 17 16:40:38 RP: it might be, we need to register the handlers... Apr 17 16:43:42 pb_: one question to detect-stylus.c Apr 17 16:44:04 pb_: the way we use it `detect-stylus...` how likely is it that X runs? Apr 17 16:44:59 zecke: almost totally unlikely Apr 17 16:45:27 pb_: to stop the whining, I considered forking your code... Apr 17 16:45:51 zecke: that would certainly be an option Apr 17 16:45:52 pb_: Can't detect-stylus stop, after it has detected a device? Apr 17 16:45:59 yes, it does Apr 17 16:46:22 pb_: I don't get why you need to do some funky X11 Apr 17 16:46:32 pb_: specially that stylus has nothing to do with a matchbox Apr 17 16:46:41 zecke: it only does the funky X stuff if you invoke it in a different way Apr 17 16:46:48 in the case you're interested in, it does: Apr 17 16:46:50 if (flag_device) Apr 17 16:46:50 { Apr 17 16:46:50 printf ("%s\n", name); Apr 17 16:46:50 exit (0); Apr 17 16:46:50 } Apr 17 16:46:57 as soon as it figures out what the device is. Apr 17 16:47:04 CoreDump|home: hmmm...I liked it that it didn't allow me to "altboot" once I'd already altbooted....will save us so many support requests.... Apr 17 16:47:28 zecke: of course, what wrong depends? Are they depending on non-native packages? Apr 17 16:47:54 pb_: why not use two programs? Apr 17 16:48:02 zecke: no reason Apr 17 16:48:04 mickeyl: no, I wouldn't object at all, the CVS files aren't really in use any more anyway. Apr 17 16:48:06 JustinP: that is probably correct Apr 17 16:48:14 pb_: it could use a funky X extension to find out that the xserver opened a touchscren Apr 17 16:48:33 zecke: if you wanted to make an opie-safe version by evacuating the x stuff to another file and not compiling that one, that would be fine Apr 17 16:48:47 pb_: okay then I need to 'fork' it and send it to reenoo Apr 17 16:48:51 zecke: alternatively, if you wanted to make it dlopen() libx11 when it needs it, rather than linking with it directly, that would be fine too Apr 17 16:48:56 mickeyl: tested newest e17 yet? Any idea why it's locking/crashing? I think Xfbdev is segfaulting but I haven't looked for why yet....probably another dependency issue.... Apr 17 16:48:58 pb_: so you call detect stylus twice ona GPE start? Apr 17 16:49:03 zecke: right Apr 17 16:49:14 pb_: in the hope to find two styluses? Apr 17 16:49:20 :} Apr 17 16:49:22 zecke: right, we want as many styluses as possible Apr 17 16:49:23 zecke: This could be a really nasty handlers problem related to the global scope you mentioned :-/ Apr 17 16:49:45 zecke: but in fact you're right, in this day and age one can (more or less) detect the existence of a stylus by the presence of the CALIBRATE extension. Apr 17 16:49:49 RP: i can't for the life of me make kexec use my --append="kernel options". I have an ampty CMDLINE for that kernel, is there something else i need to check? Apr 17 16:50:12 RP: config file gets loaded -> INHERITS evaluated -> handlers registered Apr 17 16:50:22 CoreDump|home: How did you ensure an empty CMDLINE? Apr 17 16:50:27 RP: okay another bug to look into Apr 17 16:50:48 RP: i edited the .inc file Apr 17 16:51:12 zecke: That's fair enough and the cache changes shouldn't have changed that Apr 17 16:51:28 CMDLINE = "" Apr 17 16:51:33 CoreDump|home: It might be it shouldn't be set at all Apr 17 16:51:46 RP: we might need to do an add handlers (or whatever it is named) Apr 17 16:51:54 * zecke tries hard to return to the books... Apr 17 16:52:00 * RP has never tried kexec Apr 17 16:52:15 zecke: I'm kernel hacking atm - maybe bitbake later :) Apr 17 16:52:38 RP: I'll try that Apr 17 16:52:45 JustinP: edb-native and edje-native depend on non native versions of cpp and ncurses Apr 17 16:54:57 zecke: ah, np, will fix. Are these the releases or CVS versions (if this is in your e-mail let me know) Apr 17 16:56:41 RP: what kind of cool kernelstuff do you have working now? Apr 17 16:58:11 koen: bugfixes atm, nothing interesting :-/ Apr 17 16:58:54 koen: I want to blitz a load of one liners out of my local tree and into mainline Apr 17 16:58:54 koen: for the GettingStarted page, should we suggest to connect by default to ewi546.ewi.utnwente.nl or monotone.vanille.de (or another one) ? Apr 17 16:59:43 cromain: monotone.vanille.de Apr 17 17:00:02 koen: thanks, [ update in progres ...] Apr 17 17:03:20 JustinP: it is in the mail (version and filename) Apr 17 17:05:23 zecke: http://oe.pastebin.com/665366 Apr 17 17:05:53 zecke: yes, CVS packages. as per mickeyl's request I'm removing them (I haven't been maintaining them, which is why they have this issue :-| ) Apr 17 17:06:32 koen: OE metadata == OE database == OE tree, is it right ? Apr 17 17:07:10 OE == OE Apr 17 17:07:15 cromain: the oe metadata is stored in the db, once you extract it (checkout) you get an OE tree Apr 17 17:07:18 koen: I'm away Apr 17 17:07:30 the db alone won't do much Apr 17 17:07:44 koen: ok Apr 17 17:07:48 koen: -> RP Apr 17 17:09:24 JustinP: I'm about to leave in a second... one short question though, now that evas is modular, there's no longer much sense in distinguishing -fb and -x11 stuff, isn't it? can you confirm that libevas built with --enable-x11 doesn't directly link at x11? Apr 17 17:09:29 koen: the 'Cache' saves the last filename/data... Apr 17 17:09:40 koen: it is not very OO and it is now even out of sync Apr 17 17:10:14 the posts to enlightenment devel make it seem so Apr 17 17:11:19 emte: thanks for your help last night. I think it was the python version, i did a separate install with 2.4.3 insteadl of distribs 2.4.2 and bitbake works now. Though I also downgraded to 1.3.3 from 1.3.9. But 1.3.3 did not work with that version of python either. Apr 17 17:11:44 mickey|mexDinner: That sounds correct. I'm not *sure* of this, but I can try to confirm it.... Apr 17 17:12:23 mickey|mexDinner: of course, if we make it one build anyone who wants evas-fb will still have to build x11.... Apr 17 17:14:38 JustinP: you win some, you lose some Apr 17 17:16:50 RP: nope, #echo "CONFIG_CMDLINE=\"${CMDLINE}\"" >> ${S}/.config doesn't work either Apr 17 17:17:36 CoreDump|home: Others have had this working so perhaps ask them what they did? Apr 17 17:17:54 sure, got any names? Apr 17 17:18:03 do13 :) Apr 17 17:18:07 CoreDump|home: do13 Apr 17 17:18:13 CoreDump|home: problems with config? Try adding a newline both before and after your line Apr 17 17:18:18 ahh, will do thanks Apr 17 17:18:39 CoreDump|home: I had problems like that where I edited the defconfig and forgot a newline after the last line Apr 17 17:18:44 JustinP: problems with using a custom CMDLINE with kexec Apr 17 17:19:01 do13: ping Apr 17 17:19:08 with kexec? ah, you mean compiled into the kernel? Apr 17 17:19:37 no not compiled in but specified on the commandline with kexec Apr 17 17:19:42 CoreDump|home: did you check the final ${S}/.config? does it look right? Apr 17 17:19:51 CoreDump|home: then how does that have anything to do with .config? Apr 17 17:20:13 because you have to make sure that CMDLINE is not compiled in for that to work Apr 17 17:22:17 oh, I see Apr 17 17:22:26 mickey|mexDinner: removed Apr 17 17:22:30 zecke: fixed ;-) Apr 17 17:22:45 03justinp 07org.oe.dev * re95e1d77... 10/packages/ (25 files in 2 dirs): efl, e17: Remove CVS bbfiles for any package which has numbered releases Apr 17 17:30:55 thanks Apr 17 17:41:23 03xora 07org.oe.dev * rad9ee835... 10/packages/abiword/abiword-plugins_2.4.4.bb: abiword-plugins_2.4.4.bb : stop depending on non existant abiword-plugins Apr 17 17:43:02 afternoon Apr 17 17:43:10 hey TheMasterMind1 Apr 17 17:43:46 whats new Apr 17 17:44:15 the mailinglist is up again Apr 17 17:44:27 i don't think paul brook likes me Apr 17 17:44:29 that's good Apr 17 17:44:35 heh Apr 17 17:44:43 TheMasterMind1: what did you do? Apr 17 17:45:07 * RP is curious too :) Apr 17 17:45:20 my first email he replied to to tell me i was wrong, the second one was one line and then he ignored the third and fourth Apr 17 17:45:22 TheMasterMind1: did martin guy send you a mail yet? Apr 17 17:45:29 martin? nope Apr 17 17:45:34 what about Apr 17 17:45:55 TheMasterMind1: martin is payed by arm to integrate EABI into debian Apr 17 17:46:02 ah Apr 17 17:46:13 well i've been learning all sorts of things about ARM lately Apr 17 17:46:15 I occasionally get email asking "how the hell did you get it to work in OE' Apr 17 17:46:24 pb_: how to fix xstroke.org? Apr 17 17:46:31 the pxa is actually a BIG_ENDIAN with BYTES_LITTLE_ENDIAN Apr 17 17:46:41 when using fpa anyway Apr 17 17:46:52 but somehow he doesn't get what I (or lennert) is telling him Apr 17 17:47:01 TheMasterMind1: fpa is evil Apr 17 17:47:12 indeed Apr 17 17:47:35 TheMasterMind1: that's why vfp is so much nicer Apr 17 17:47:35 I think the kexec arm patch doesn't support --append heh Apr 17 17:47:36 i think i've finally got it (mostly) figured out now Apr 17 17:47:46 fpa? Apr 17 17:47:56 Crofton: floating point stuff Apr 17 17:48:00 ah Apr 17 17:48:12 zecke: in the medium to long term we should find somewhere else for xstroke to live, maybe at freedesktop.org Apr 17 17:50:04 pb_: tmp solution? Apr 17 17:50:11 zecke: in the short term I would like to just point the SRC_URI to another copy of the tarfile Apr 17 17:50:45 not sure if that will be practical or acceptable though Apr 17 17:53:39 pb_: is it mirrored at the fam website? Apr 17 17:54:46 zecke: yeah, but it sounds like pointing to that copy would cause a political problem. I can put a copy in my personal space on freedesktop.org though; maybe that would be the best thing. Apr 17 17:56:09 pb_: well... we just need a signed letter from george that the source will not be removed Apr 17 17:56:18 pb_: faxed and snail mailed to the Koen Trust Apr 17 17:56:22 heh Apr 17 17:56:29 I wouldn't trust that Apr 17 17:58:03 pb_: and certified by the grand parents of george Apr 17 17:58:12 zecke: he would just have a script remove it Apr 17 17:58:20 zecke: I imagine the Koen Trust blocks any communication from george anyway. Apr 17 17:58:20 "I didn't do it, it was a script" Apr 17 17:58:54 i trust george, i think we have backups as well Apr 17 17:58:55 pb_: actually no, george does not communicate at all Apr 17 17:59:32 I guess he lacks a script for that Apr 17 17:59:38 RP: yep that's it. the arm patch of kexec doesn't support --append, which makes it kinda limited Apr 17 17:59:54 zecke: okay, great. in that case, let's just use that copy. Apr 17 18:00:30 CoreDump|home: Ah. Is that a kernel limitation or the kexec tool that needs support (or both)? Apr 17 18:00:40 zecke: http://familiar.handhelds.org/source/v0.8.2/xstroke-0.6.tar.gz Apr 17 18:00:49 RP: that I can not say. Apr 17 18:01:04 CoreDump|home: It is a nasty limitation and one I'll have to address before its usable for me... Apr 17 18:01:17 koen: I trust hh.org more than xstroke.org ;) Apr 17 18:01:52 pb_: can I remove the _cvs version as well? Apr 17 18:01:57 zecke: send it to treke so he can put it on oesources (if it isn't already) Apr 17 18:02:22 zecke: yes, please. the cvs tree is obviously defunct, and I don't expect it to come back any time soon. Apr 17 18:04:43 hi mreimer Apr 17 18:04:52 hi pb_ Apr 17 18:04:58 hey mreimer Apr 17 18:05:02 hey koen Apr 17 18:05:48 am i mistaken or does glib take a long time to compile Apr 17 18:10:10 TheMasterMind1: packaging glibc takes a while as well Apr 17 18:11:57 03koen 07org.oe.dev * r928cf3c1... 10/packages/linux/handhelds-pxa-2.6_cvs.bb: handhelds-pxa 2.6 cvs: turn on EABI for eabi target, inspired by linux-openzaurus.inc Apr 17 18:12:58 TheMasterMind1: I think you're mistaken: it doesn't take _that_ long. Apr 17 18:13:13 glib is a pretty small library, it should just take a few minutes. Apr 17 18:13:37 alright Apr 17 18:13:48 * koen needs glasses Apr 17 18:18:34 JustinP: fwiw, libxml2 is using schema for some time now, so the NOTE in the php .bbs is a bit out of place now Apr 17 18:24:20 PACKAGE_ARCH_append_linux-gnueabi = "-eabi" Apr 17 18:24:23 * koen wonders Apr 17 18:28:36 ibot, botmail for reenoo: I think http://handhelds.org/~pb/sw.patch will clear up that i18n problem with rootfs construction. maybe you'd like to try it out. Apr 17 18:29:40 koen: wiki updated ! May be should suggest to checkout a specific bitbake revision for people who start sing OE ... ? Apr 17 18:30:58 koen: I already put -r419 because it was stable some day ago. What do you think about that ? Apr 17 18:31:09 cromain: I've been thinking about that, but the latest svn version is working pretty well Apr 17 18:31:28 cromain: when zecke creates the stable branch we can recomment that one Apr 17 18:32:08 koen: do you mean I remove the -r419 from the wiki and let people always getting the lastest bitbake ? Apr 17 18:32:38 cromain: yes, for the time being Apr 17 18:32:49 koen: ok Apr 17 18:33:25 How many GB do i have to have free on my harddisk for bulding 'opie-image'? Apr 17 18:33:52 koen: Hope that my update will be usefull for new OE people. Apr 17 18:35:37 koen: oh yeah, I forgot about the php builds..... Apr 17 18:36:27 JustinP: I did as well :) Apr 17 18:36:33 cromain: I'm sure they will be Apr 17 18:39:40 goxboxlive: about 4 IIRC Apr 17 18:41:23 koen: the newest bbfile doesn't have ---without-schemas, but the rest do Apr 17 18:42:03 JustinP: correct Apr 17 18:43:22 * pb_ celebrates victory over ipkg and goes back to working on apmd Apr 17 18:44:04 koen: will remove the NOTE Apr 17 18:44:31 does busybox handle mv internally or calls itself? Apr 17 18:44:50 calls itself, I think Apr 17 18:44:54 pity... Apr 17 18:44:57 busybox has very few shell builtins, or did last time I looked Apr 17 18:45:22 the CELF dudes had a patch to improve that, but (iirc) it caused some obscure problem or other when I tried to integrate it, and I never had time to debug the failure Apr 17 18:45:28 Okay, I can stand it... I need to swap my / with chroot... Apr 17 18:46:00 Probably, moving /lib/ is the worst part. Apr 17 18:48:51 03justinp 07org.oe.dev * r3486b902... 10/packages/php/php_5.0.5.bb: php_5.0.5: remove outdated NOTE Apr 17 18:57:56 koen: I saw your EABI note--very good. Have you had success getting an EABI GPE up? Apr 17 18:58:54 mreimer: I didn't try to boot it yet, but pH5 got into gpe Apr 17 18:59:07 where is ph5 now anyway... Apr 17 18:59:18 ~seen pH5 Apr 17 18:59:19 koen: good. I guess he figured out the alignment problems then? Apr 17 18:59:25 ph5 was last seen on IRC in channel #gpe, 1d 23h 37m 34s ago, saying: 'hi reenoo'. Apr 17 18:59:35 mreimer: not yet, but we boot in alignment fixup anyway Apr 17 18:59:49 03freyther 07org.oe.oz354fam083 * r1e5cafe2... 10/packages/xstroke/ (xstroke_cvs.bb xstroke_0.6.bb): Apr 17 18:59:49 packages/xstroke: Apr 17 18:59:49 xstroke.org vanished: Apr 17 18:59:49 -remove the cvs build of xstroke Apr 17 18:59:49 -change the xstroke 0.6 build to get Apr 17 18:59:49 the source from www.oesources.org Apr 17 18:59:54 03freyther 07org.oe.dev * rf7fafd00... 10/packages/xstroke/ (xstroke_cvs.bb xstroke_0.6.bb): Apr 17 18:59:54 packages/xstroke: Apr 17 18:59:55 koen: are the GPE-related EABI problems written down somewhere, so others can help? Apr 17 18:59:57 xstroke.org vanished: Apr 17 18:59:59 -remove the cvs build of xstroke Apr 17 19:00:01 -change the xstroke 0.6 build to get Apr 17 19:00:03 the source from www.oesources.org Apr 17 19:00:21 mreimer: not yet Apr 17 19:00:33 mreimer: did you try FBSD 6.1rc yet? Apr 17 19:00:48 mreimer: we're now very close to build some bootstrap, gpe and opie images for various devices for testing Apr 17 19:00:53 zecke: FreeBSD bilbo.vpop.net 6.1-PRERELEASE FreeBSD 6.1-PRERELEASE #1: Mon Mar 6 15:10:04 PST 2006 mreimer@bilbo.vpop.net:/usr/obj/usr/src/sys/BILBO i386 Apr 17 19:00:56 gpgme also disappeared....I had to change the src to a mirror Apr 17 19:01:03 not sure if it's back up now Apr 17 19:01:07 zecke: it works well for me. it's probably better than all 5.x Apr 17 19:01:22 koen: that's good to hear. I'll test h2200 Apr 17 19:01:31 koen: and maybe soon hx4700 too Apr 17 19:01:36 mreimer: hehe, I installed vmware and now have FBSD, DragonFly, OpenBSD, NetBSD and OpenSUSE installed Apr 17 19:01:48 zecke: :-) you left out qemu! Apr 17 19:01:49 I need to get a x86 CPU with VT support... Apr 17 19:02:09 zecke: get an Apple with Intel Duo--then you can have OS X too Apr 17 19:02:15 I think those have VT Apr 17 19:02:22 mreimer: right Apr 17 19:02:30 $$$ though Apr 17 19:02:32 mreimer: I want something with SATA and fast IO as well Apr 17 19:02:46 mreimer: and notebooks are normally not good at that Apr 17 19:03:02 zecke: their iMacs have Intel Duos now Apr 17 19:03:08 zecke: don't know about their minis Apr 17 19:03:14 minis as well Apr 17 19:03:14 minis do as well Apr 17 19:03:31 zecke: iMacs are SATA Apr 17 19:03:39 zecke: but probably aren't all that impressive for IO Apr 17 19:04:09 only 1.5 years left on warranty for my powerbook Apr 17 19:04:20 after that -> 7200 rpm disk replacement Apr 17 19:06:01 * zecke wanted a imac but got a barebone :} Apr 17 19:06:34 * JustinP cares not Apr 17 19:10:20 zecke: that was careless Apr 17 19:11:39 pb_: what? Apr 17 19:11:42 *confused* Apr 17 19:12:04 zecke: buying a barebone by mistake Apr 17 19:12:15 pb_: not by mistake :} Apr 17 19:12:31 I wanted something like a x86 based terminal for my desk Apr 17 19:12:38 and wrote 'iMac' on the wishlist Apr 17 19:13:20 03coredump 07org.oe.dev * refad9384... 10/packages/altboot/ (6 files in 3 dirs): altboot: Add initial support for kexec, reworked parts of the menu code Apr 17 19:14:16 zecke: aha Apr 17 19:14:41 hrm Apr 17 19:14:48 now apmd is fixed, but suspending is still slow Apr 17 19:14:58 this is no good Apr 17 19:17:09 zecke: do you know what /etc/apm/suspend.d/05ifupdown is in aid of? Apr 17 19:17:20 pb_: no Apr 17 19:17:25 pb_: usb wlan drivers? Apr 17 19:17:30 zecke: drat Apr 17 19:17:41 zecke: no, I think that's /etc/apm/suspend.d/10usbnet Apr 17 19:17:57 oh, wlan, maybe Apr 17 19:17:57 usbnet != wlan ;) Apr 17 19:18:07 sorry, moment of dyslexia there Apr 17 19:18:17 my wifi seems to work fine without it, but that's cf not usb Apr 17 19:18:18 just a speculation though Apr 17 19:18:29 I wonder if I can find out who added it and hold them to account Apr 17 19:18:39 | make[2]: *** [html-build.stamp] Fehler 1 Apr 17 19:18:39 | make[2]: Leaving directory `/home/ich/hacking/embedded/oe/builds/fam83/tmp/work/librsvg-2.6.5-r3/librsvg-2.6.5/doc' Apr 17 19:18:42 | make[1]: *** [all-recursive] Fehler 1 Apr 17 19:18:44 it seems to be the main offender causing slow suspending Apr 17 19:18:51 pb_: monotone blame Apr 17 19:19:06 good thinking Apr 17 19:20:20 ah, bad thinking Apr 17 19:20:21 monotone: unknown command 'blame' Apr 17 19:20:27 annotate Apr 17 19:20:33 pb_: annotate Apr 17 19:20:49 monotone: fatal: std::runtime_error: failure extracting schema from sqlite_master Apr 17 19:20:49 monotone: Apr 17 19:20:49 monotone: this is almost certainly a bug in monotone. Apr 17 19:21:00 drat Apr 17 19:21:18 pb_: is it locked? Apr 17 19:21:26 are you trying to do two things with monotone at once? Apr 17 19:21:37 oh yeah Apr 17 19:21:46 my pull from this morning is still "running" but appears to be stalled Apr 17 19:22:07 9596 pts/2 S+ 72:24 monotone --db=OE.db pull monotone.vanille.de org.open Apr 17 19:22:29 should I just hit ^C on it? Apr 17 19:22:55 maybe Apr 17 19:23:33 heh Apr 17 19:23:37 according to strace it seems to be waiting in select() with a 5700 second timeout Apr 17 19:24:04 maybe monotone.vanille.de has just decided to stop talking to me Apr 17 19:24:25 pb_: can't linux's top show that? Apr 17 19:24:36 zecke: no, linux sucks Apr 17 19:24:44 you must be thinking of openbsd Apr 17 19:24:53 okay, I killed it Apr 17 19:25:01 * pb_ waits patiently for monotone annotate to run Apr 17 19:25:24 heh, just one revision. that wasn't very interesting. Apr 17 19:25:27 * pb_ try monotone log instead Apr 17 19:25:35 lamikr: ping Apr 17 19:26:38 import clean BK tree at cset 1.3670 Apr 17 19:26:40 :-} Apr 17 19:26:47 pb_: bkbits.net is still online Apr 17 19:26:54 aha Apr 17 19:27:30 is it oe.bkbits.net. or openembedded? Apr 17 19:27:43 oe-devel. Apr 17 19:28:09 hm, none of those work for me Apr 17 19:28:12 hmmm Apr 17 19:28:16 same here Apr 17 19:28:19 looks like "online" is a bit of an overstatement Apr 17 19:28:28 there worked a few days ago Apr 17 19:28:32 pb_: it was online, last time I checked it Apr 17 19:28:35 how do i set ccache to speed up the compilation= Apr 17 19:28:37 but you'll never know with larry Apr 17 19:28:43 goxboxlive: install it Apr 17 19:28:50 zecke: heh, me too, but that was about a year ago Apr 17 19:29:35 i have it installed, but not in my enviorment. How do i do that? Is it just 'ccache=CCACHE'? Apr 17 19:29:54 put it in your path? Apr 17 19:30:24 maybe we should just remove network-suspend-scripts from task-bootstrap and see who complains. Apr 17 19:30:46 though I guess that probably would break usbnet, so maybe not such a good idea. Apr 17 19:30:50 pb_: h5xxx users Apr 17 19:31:14 pb_: osb stuff is not comitted, I can't build lib?svg Apr 17 19:31:15 at least the subset that uses usb wifi Apr 17 19:31:42 ah right Apr 17 19:31:51 maybe one of them would like to fix the driver instead. Apr 17 19:32:04 zecke: oh dear Apr 17 19:32:22 zecke: what's the svg problem? Apr 17 19:32:25 mreimer: does OE generate labrun scripts already? Apr 17 19:32:38 pb_: generation of the docs are failing Apr 17 19:33:00 zecke: that is unfortunate Apr 17 19:33:06 without docs we are nothing Apr 17 19:33:09 koen: no, I don't think so. We could for the jffs2 image, but otherwise we'll have to make new targets for CF and SD Apr 17 19:33:16 koen: since they take different arguments Apr 17 19:33:42 koen: I'm tempted to put the brains to do that in LAB, but don't know if that would be kosher Apr 17 19:34:43 koen: do you know what the specific problem is with the h5xxx wifi that this script solves? Apr 17 19:35:41 pb_: eons ago the drivers wouldn't survive suspend/resume Apr 17 19:35:47 http://oe.pastebin.com/665687 Apr 17 19:36:02 when the script got added they started suffering from the wlan1 .. wlan2 .. wlan3 problem Apr 17 19:36:30 zecke: crumbs, that does suck Apr 17 19:37:00 koen: okay. in that case, I think the right thing to do is remove the script, inspect the situation, and fix the original bug if there still is one. Apr 17 19:37:08 yeah Apr 17 19:37:50 pb_, thanks for working on this apmd stuff. I've been wanting to but just haven't had the time Apr 17 19:37:50 it'd be a bit unfortunate if everyone is suffering from slow suspending just because of some problem that existed ages ago and has been fixed since. Apr 17 19:37:51 usbnet is non working right now anyway Apr 17 19:38:11 * pb_ goes to look for an h5500 Apr 17 19:40:18 drat, flat battery Apr 17 19:40:22 I guess that's no surprise Apr 17 19:45:17 good nite Apr 17 19:46:10 zecke: sweet dreams Apr 17 19:50:14 koen: wifi on my h5400 seems to work okay across suspend even without that script. Apr 17 19:50:19 so I guess it's not required any more Apr 17 19:50:29 pb__: nuke it Apr 17 19:51:10 pb_: that's good news Apr 17 19:51:22 and the usbnet one seems to be doing (me) more harm than good: it causes usbf to come up when you resume even when no driver is loaded on the host, which means you end up with a default route pointing to the bit bucket. Apr 17 19:51:46 I suspect it might do that even if the usb cable isn't plugged in at all, though I'm not sure. Apr 17 19:52:11 so, overall, I kind of feel that network-suspend-scripts is toxic and should be removed. Apr 17 19:54:14 i seem to recall that the zaurus machines needed something like that, but 2.6 might have changed that Apr 17 19:54:58 yeah, that's possible. I didn't mean it should be removed from the repository altogether, just that it should be removed from familiar's task-bootstrap. Apr 17 19:55:36 if openzaurus wants or needs to keep it, that's fine Apr 17 19:56:18 damn, I'm still not in bed Apr 17 19:56:42 RP: I have a hack for the 'Summary' Apr 17 19:58:04 pb_: it seems some machines need those, which should be orthogonal to distro Apr 17 19:58:26 pb_: so only machines that needs them should have them in task-bootstrap, right? Apr 17 19:58:47 well, ultimately it's a distro decision whether to leave it out, leave it in, or make it machine specific. Apr 17 20:00:55 only the distro maintainers are in a position to make the tradeoff between not installing useless software on some machines and not causing a proliferation of packages and/or images. Apr 17 20:01:36 for example, if openzaurus (hypothetically) needed network-suspend-scripts on every machine except for one, they might decide to just ship it on all platforms to avoid needing a special task-bootstrap on that one. Apr 17 20:02:38 iirc it's only needed on collie and poodle Apr 17 20:02:59 conversely, as far as familiar is concerned, if it transpires that some or other machine wants that stuff, we would look for an alternative solution that allowed us to not ship the scripts on any machine. Apr 17 20:03:48 I know of some people running familiar on the affected zaurus machines for the debian compat Apr 17 20:03:58 which is an edge case, of course Apr 17 20:04:37 yeah. zaurus isn't a supported platform as far as familiar is concerned: those folks are welcome to run it on their zauruses, but it's up to them to provide the machine specific bits that they might need. Apr 17 20:04:56 Does familiar use busybox ? Apr 17 20:05:07 in exactly the same way as I might choose to run debian on my ipaq, but I wouldn't expect Debian to provide the tools to drive the backlight. Apr 17 20:05:10 Ecco: yes Apr 17 20:05:15 pb_, ok Apr 17 20:05:38 Is debian still maintained on ARM ? Apr 17 20:05:43 yes Apr 17 20:06:30 Ok, good. Debian uses plain coreutils, right ? Apr 17 20:06:34 RP: oh you fixed that recursive... Apr 17 20:06:53 zecke: yes, no more recursive clean Apr 17 20:08:05 Ecco: yes Apr 17 20:08:21 Ok, thx for the info Apr 17 20:08:29 Does it look doable to use debian on a Zaurus ? Apr 17 20:09:05 koen: the summary should be back as well (quite hack though) Apr 17 20:09:06 Some folks were certainly looking at that a while back. You might check the archives of the debian-handheld list. Apr 17 20:37:24 koen: I think something is ill between abiword and the plugins, they dont seem to work at all Apr 17 20:37:34 drat Apr 17 20:38:05 * koen wonders what that would be Apr 17 20:38:54 koen: some stop abiword loading at all with an symbol they cant find, which is in the main abiword exe, others just dont show up in the plugin manager Apr 17 20:40:08 XorA|gone: so they should link to abiword? Apr 17 20:41:11 koen: I think so, some such as google plugin reference XAP::getApp which is in abiword Apr 17 20:41:33 those are the ones that prevent abiword loading Apr 17 20:42:22 XorA: a missing --with-abiword= ? Apr 17 20:42:48 koen: it looks by default in ../abi/ which is right Apr 17 20:42:54 hmmm Apr 17 20:43:16 I wonder if it needs a configured abi or an onconfigured abi Apr 17 20:43:43 currently emerge abiword on getoo to see if I can spot any differences Apr 17 20:44:04 I was kinda happy with the abi/plugins split Apr 17 20:44:16 doing it in one .bb file would be challenging Apr 17 20:44:28 koen: it looks like it should be happy with them split Apr 17 20:44:44 koen: as far as I can see it only references one file from abi in plugins Apr 17 20:44:54 should be ok Apr 17 20:45:15 it very irritating Apr 17 20:45:18 try mailing tf Apr 17 20:45:32 hopefully he has the solution Apr 17 20:45:40 or can create one :) Apr 17 20:46:03 i never relalized how many countries there actually are until i did bitbake dropbear :( hehe Apr 17 20:46:53 tyz: 217 locales last I heard.... Apr 17 20:47:29 ahhh hehe Apr 17 20:52:11 the PXA255 doesn't have fpu, right ? Apr 17 20:52:21 CSMan: correct Apr 17 20:54:23 what do they mean by "built-in multimedia capabilities" ? just a smooth GUI? Apr 17 20:54:39 marketing speach Apr 17 20:55:07 the GUI is pretty snappy, but i don't know about video playback Apr 17 20:55:14 even audio playback Apr 17 20:56:07 gets chunky when trying to do more things Apr 17 20:58:27 'night all Apr 17 20:59:01 nite koen Apr 17 21:00:44 CoreDump|home: so i guess these ipaqs are not really meant for video playback Apr 17 21:00:54 nope Apr 17 21:00:55 03coredump 07org.oe.oz354x * r292fee4e... 10/packages/altboot/ (altboot_1.0.6-rc4.bb altboot_1.0.6.bb altboot_1.0.7-rc1.bb): altboot: mv 1.0.6-rc4 1.0.6, added 1.0.7-rc1 with kexec support and changes all over the place. Apr 17 21:01:00 03coredump 07org.oe.dev * r38593741... 10/packages/altboot/ (6 files in 3 dirs): altboot: mv 1.0.6-rc4 1.0.6, added 1.0.7-rc1 with kexec support and changes all over the place. Apr 17 21:01:41 CSMan: my Akita has a PXA270 und fullscreen video still sucks Apr 17 21:03:53 * XorA is quite happy with mpeg4 video on pxa255 Apr 17 21:04:02 CoreDump|home: so we can't blame linux for that, nor gcc when trying to optimize Apr 17 21:04:05 24fps at 320x240 no problems at all Apr 17 21:04:21 XorA: using vlc ? Apr 17 21:04:29 using mplayer optimised for am Apr 17 21:04:31 arm Apr 17 21:04:36 your own build ? Apr 17 21:04:45 mplayer-atty in OE Apr 17 21:04:58 i'll grab it and try Apr 17 21:05:00 =) Apr 17 21:05:01 bvdd? Apr 17 21:05:09 should work on all pxa devices, on w100 chips get added bonus of overlay Apr 17 21:05:45 CoreDump|home: only if someone with a bvdd capable device fixes it, I have no access to one Apr 17 21:06:36 I tried -atty, didn't notince difference to normal mplayer w/ kernel 2.6 Apr 17 21:07:23 is 2.6 better ? Apr 17 21:07:39 oh yeah =) Apr 17 21:07:43 oh yeah Apr 17 21:07:45 hmmm... Apr 17 21:07:56 ~praise RP Apr 17 21:07:58 All hail RP! Apr 17 21:08:39 ~praise RP Apr 17 21:08:41 All hail RP! Apr 17 21:08:48 ~lart Embeddix Apr 17 21:08:48 * ibot dumps 42 tons of dirt, manure, and fish heads on Embeddix Apr 17 21:09:10 this damn remote driver is driving me crazy Apr 17 21:09:31 XorA: what's the link to the snapshot dir ? Apr 17 21:09:44 mine is out of date Apr 17 21:09:46 I'm thinking it's a problem with timing and the "state machine" that ebmeddix implemented.... Apr 17 21:10:40 o_O why did they removed 0.8.3 from the download section ? Apr 17 21:11:12 *remove Apr 17 21:12:51 CSMan: snapshot dir? Apr 17 21:13:27 heheh, i think there used to be one... Apr 17 21:13:54 it's familiar's thou Apr 17 21:14:19 http://familiar.handhelds.org/snapshots/v0.8.3/ Apr 17 21:15:58 XorA: i think i have to update everything, should i just go to familiar and get the latest image ? Apr 17 21:16:01 kcachegrind kicks so much ass. Apr 17 21:16:05 * kergoth` had no idea Apr 17 21:17:55 k, i got the feed Apr 17 21:18:08 JustinP: what happened to 0.8.3 ? Apr 17 21:19:45 hmm... hostap in dev image doesn't spport encryption? Apr 17 21:19:48 anybody know of a workaround for the localedef oom problem? Apr 17 21:19:57 mreimer, swap Apr 17 21:20:04 poly check dmesg, a module is missing Apr 17 21:20:05 thanks polyonymous Apr 17 21:20:37 CoreDump|home, ieee80211_crypt? Apr 17 21:20:49 no, something with "ar4" Apr 17 21:21:06 hmm... how come it's missing,I wonder... lemme check Apr 17 21:22:29 ahh... load module hostap_crypt_wep... Apr 17 21:22:51 And I do not have one... Apr 17 21:23:08 I had a different problem then ;) Apr 17 21:23:17 kergoth: what sample data are you feeding kcachegrind? Apr 17 21:23:28 but somewhat similar... I wonder how come it's missing and I wonder if I had this module at all, then.. Apr 17 21:23:33 i've wanted to feed it oprofile sample data, but conversion tools last i looked were sorely lacking Apr 17 21:23:44 03mickeyl 07org.oe.dev * rb7f6e8d6... 10/conf/distro/preferred-e-versions.inc: preferred-e-versions: give people a chance to chose over fb/x11 Apr 17 21:23:48 03mickeyl 07org.oe.dev * rf0201c27... 10/conf/bitbake.conf: bitbake.conf: fix E_CVS entry and make E_URI appear here as well Apr 17 21:23:52 03mickeyl 07org.oe.dev * r1224054f... 10/site/arm-linux-gnueabi: site/arm-linux-gnueabi: add AC entry for size_off_t Apr 17 21:23:57 03mickeyl 07org.oe.dev * rdb84028d... 10/classes/e.bbclass: e.bbclass: don't duplicate E_CVS and E_URI Apr 17 21:24:01 03mickeyl 07org.oe.dev * r1a6db0f8... 10/packages/efl/ (6 files): efl: edje and ecore no longer need imlib2 (yay!) Apr 17 21:27:30 CSMan: I'm an OZ person, I know nothing about familiar Apr 17 21:27:48 play safe =P Apr 17 21:28:18 XorA: did you bitbaked mplayer-atty? Apr 17 21:28:33 and g-ether doesn't work exactly like Ecco described... Apr 17 21:29:21 mickey|mexDinner: thanks, I appreciate it Apr 17 21:29:27 :) Apr 17 21:29:52 JustinP: fixing efl++ and noticed that on the way Apr 17 21:31:06 ^_^ Apr 17 21:31:33 mickeyl: would appreciate some help testing....I have no idea why it's crashing Apr 17 21:31:45 mickeyl: I'll try an strace tonight, but I don't know how much it will help... Apr 17 21:32:49 polyonymous, what do you mean ? Apr 17 21:32:55 mickeyl: uh-oh Apr 17 21:33:05 mickeyl: monotone: warning: discarding revision cert packet 1224054fda74d44a07c933940064334aa0ca4218 with unmet dependencies Apr 17 21:33:08 etc Apr 17 21:33:34 nothing to worry about. monotone got slightly confused, repull and it will disappear Apr 17 21:33:35 Ecco, linux doesn't see it Apr 17 21:33:49 mickeyl: yeah, was just about to say nm, it's working now ;-) Apr 17 21:33:53 * mickeyl hugs monotone safety net Apr 17 21:33:59 g_ether works just fine here Apr 17 21:34:09 is 0.8.4-rc1 running the 2.6 kernel ? Apr 17 21:34:13 CoreDump|home, do you have kernel from .dev? Apr 17 21:34:17 nope Apr 17 21:34:28 ok Apr 17 21:34:34 It worked fine with OZ image for me as well Apr 17 21:34:40 CSMan: that wasn't for you heh Apr 17 21:34:48 I've just flashed this .dev kernel... Apr 17 21:34:53 ah i c Apr 17 21:35:12 And it's wrong timing, as I have a plane early in the morning... Apr 17 21:35:12 CoreDump|home: k. do you know if it has it? Apr 17 21:35:22 no clue, I'm not a fam guy Apr 17 21:35:50 Ah, g-ether is working now for no apparent readon Apr 17 21:35:53 reason Apr 17 21:35:57 =) Apr 17 21:37:51 not sure what can I do about wlan, though... Apr 17 21:41:34 CSMan: yes, I maintain it :-) Apr 17 21:41:41 afternoon Apr 17 21:42:07 polyonymous, yeah, that's what I said Apr 17 21:42:20 ahhh, ok, I re-read your sentence Apr 17 21:42:30 "it does not work" "like Ecco said" Apr 17 21:42:37 Yes, exaqctly. Apr 17 21:42:41 I read "it does not work like Ecco said" Apr 17 21:42:47 sorry :-) Apr 17 21:42:55 Yeah, and suddently it works with no appearant reason Apr 17 21:42:57 :))) Apr 17 21:43:02 damn Apr 17 21:43:09 however, it looks like in CoreDump|home's distro that it is reliable Apr 17 21:43:15 and when I rolled out my own kernel so did it Apr 17 21:43:20 and in both cases Apr 17 21:43:27 Well, .oz branch kernel seems to be. Apr 17 21:43:28 usbcore is either a module or not in the kernel at all Apr 17 21:43:33 I think I'll have to reflash old kernel Apr 17 21:43:43 wheras it was hardcoded within the OZ 3.5.4.1a2 kernel Apr 17 21:45:19 XorA: you built it with GUI ? Apr 17 21:45:38 oh well, I'll fall back to the old image then. Apr 17 21:46:23 CSMan: no, but I think amd are Apr 17 21:47:28 it's building qt o_O Apr 17 21:49:12 CSMan: choose libsdl-x11 Apr 17 21:53:58 XorA: may I ask how ? Apr 17 21:58:19 fell back to oz kernel, works fine. Will look deeper into it when I get time... Apr 17 21:58:32 see ya, folks Apr 17 21:58:42 cya polyonymous Apr 17 22:01:43 CSMan: PREFERRED_PROVIDER_virtual/libsdl = "libsdl-x11" in local.conf Apr 17 22:03:42 not something that I could've guessed =P Apr 17 22:03:44 thanks Apr 17 22:03:56 g'night Apr 17 22:05:50 03mickeyl 07org.oe.dev * r1d4c1927... 10/ (classes/qt4x11.bbclass packages/qmake/qmake2-native_2.00a.bb): qmake class, qmake-native: for qt4, we store mkspecs below $QTDIR Apr 17 22:05:54 03mickeyl 07org.oe.dev * rba1ea41f... 10/packages/efl++/ (5 files in 2 dirs): efl++: make 0.1.0 build, remove obsolete 0.0.5 Apr 17 22:23:41 * Philippe is away: visual contact - melancholic dreams Apr 17 23:09:41 * chouimat is away: food Apr 17 23:57:15 Just a warning, the bug tracker might be down for a while tonight Apr 18 00:02:23 hey ggilbert, long time no see Apr 18 00:04:37 yo mithro Apr 18 00:04:55 sup? Apr 18 00:52:53 03raymond 07org.oe.dev * r7d5b3ff4... 10/packages/autofs/ (autofs-4.1.4/Makefile.rules-cross.patch autofs_4.1.4.bb): autofs - Makefile was not in fact using the cross compiler. Apr 18 01:15:50 03raymond 07org.oe.dev * r4ad5fe71... 10/packages/beecrypt/ (beecrypt-native_3.1.0.bb beecrypt_3.1.0.bb): beecrypt: initial add Apr 18 01:15:54 03raymond 07org.oe.dev * rb913760f... 10/packages/bonnie/bonnie++_1.03a.bb: bonnie: initial add Apr 18 01:15:58 03raymond 07org.oe.dev * r2d0d27dc... 10/packages/cdrtools/cdrtools-native_2.01.bb: cdrtools-native: initial add Apr 18 01:16:02 03raymond 07org.oe.dev * rc5a6eaac... 10/packages/check/check_0.9.2.bb: check: Add staging of library and includes Apr 18 01:25:05 03raymond 07org.oe.dev * rfc24b7fc... 10/packages/cpuburn/ (cpuburn/gcc_CC.patch cpuburn_1.4.bb): cpuburn: initial add Apr 18 01:25:09 03raymond 07org.oe.dev * r49b08a06... 10/packages/ctags/ctags-native_5.5.4.bb: ctags-native: initial add Apr 18 01:25:13 03raymond 07org.oe.dev * r67f9c7bd... 10/packages/curl/ (curl-native_7.15.1.bb curl_7.15.1.bb): curl: upgrade to 7.15.1 Apr 18 01:25:17 03raymond 07org.oe.dev * rf33c5827... 10/packages/dosfstools/ (10 files in 3 dirs): Apr 18 01:25:17 dosfstools: Compile fixes. Add functionality for creating MSDOS image Apr 18 01:25:17 files from OE. Apr 18 01:25:22 03raymond 07org.oe.dev * r619ae4c3... 10/packages/fontconfig/ (fontconfig-native_2.2.95.bb fontconfig_2.2.95.bb): Apr 18 01:25:22 fontconfig: Work around archive created with patched files read-only. Apr 18 01:25:22 * Fontconfig-native should use freetype-config-native if it exists; not freetype-config which is for non-native packages. Apr 18 01:25:26 03raymond 07org.oe.dev * r20384ee2... 10/packages/geode-drivers/ (5 files): geode-drivers: GeodeGX/5535 GeodeLX/5536 kernel modules. Apr 18 01:25:30 03raymond 07org.oe.dev * r9cdd0b65... 10/packages/glibc/ (4 files in 3 dirs): Apr 18 01:25:31 glibc: 5090_all_stubs-rule-fix.patch allows building from very deep Apr 18 01:25:33 directories Apr 18 01:33:32 03raymond 07org.oe.dev * r154a6527... 10/packages/gpm/ (gpm-1.20.1/init gpm_1.20.1.bb): gpm: Add gpm init script Apr 18 01:33:36 03raymond 07org.oe.dev * rce8ac120... 10/packages/hibernate-script/ (4 files in 2 dirs): hibernate-script: initial add Apr 18 01:33:40 03raymond 07org.oe.dev * rf97087a6... 10/packages/ica/ica-bin_8.0.bb: ica: initial add Apr 18 01:33:44 03raymond 07org.oe.dev * r04607d1a... 10/packages/kbd/kbd_1.12.bb: kbd: initial add Apr 18 01:33:48 03raymond 07org.oe.dev * re9584869... 10/packages/libaio/ (4 files in 2 dirs): libaio: initial add. Patched for MIPS. Apr 18 01:33:53 03raymond 07org.oe.dev * rc00221d7... 10/packages/linux/ (linux-geodelx-2.6.11/defconfig linux-geodelx_2.6.11.bb): linux: GeodeGX/5535 GeodeLX/5536 kernel and config. Apr 18 01:37:33 03raymond 07org.oe.dev * r4564762c... 10/packages/mtools/ (3 files in 2 dirs): mtools: initial add Apr 18 01:37:37 03raymond 07org.oe.dev * rfc7fc8e1... 10/packages/openssh/ (4 files in 3 dirs): openssh: Add ssh_config Apr 18 01:37:41 03raymond 07org.oe.dev * r8d3fa722... 10/packages/rsync/ (files/rsyncd.conf rsync_2.6.7.bb): rsync: Add rsyncd.conf Apr 18 01:37:45 03raymond 07org.oe.dev * rb9ca9fd7... 10/packages/syslinux/ (3 files in 2 dirs): syslinux: Upgrade to 3.11 Apr 18 01:40:21 hi, anybody knows what's this means? Apr 18 01:40:31 PROVIDES = "virtual/libc ${@['virtual/${TARGET_PREFIX}libc-for-gcc', '']['nptl' in '${GLIBC_ADDONS}']}" Apr 18 02:03:14 evening Apr 18 02:05:53 wolverine, it means that package provides it Apr 18 02:11:07 thanks emte Apr 18 02:11:20 I want build glibc Apr 18 02:11:37 which enabled nptl Apr 18 02:11:54 and that is bad? Apr 18 02:12:23 the provides flag has nothing to do with compiling Apr 18 02:12:27 "bitbake glibc" can't find this provides Apr 18 02:12:35 its used mainly in the ipkg files Apr 18 02:13:07 you might want to look at the packages dir and choose a usable metafile Apr 18 02:13:09 yes, I have glibc_2.3.3+cvs???.bb, and I defined glibc_2.3.5.bb which enabled nptl Apr 18 02:13:29 or a valid distro/machine that has a glibc version defined Apr 18 02:15:10 machine is i586 Apr 18 02:15:12 :) Apr 18 02:17:46 DEBUG: providers for virtual/i586-linux-libc-for-gcc are: ['glibc', 'glibc-intermediate'] Apr 18 02:17:47 NOTE: preferred version 2.3.5 of glibc not available Apr 18 02:17:47 NOTE: preferred version 2.3.5 of glibc-intermediate not available Apr 18 02:17:47 DEBUG: providers for virtual/i586-linux-gcc are: ['gcc-cross'] Apr 18 02:17:47 DEBUG: selecting /home/jiff/projects/thirdparty/openembedded/org.openembedded.dev/packages/gcc/gcc-cross_3.3.6.bb as PREFERRED_VERSION 3.3.6 of package gcc-cross Apr 18 02:17:49 ERROR: /home/jiff/projects/thirdparty/openembedded/org.openembedded.dev/packages/gcc/gcc-cross_3.3.6.bb depends on itself (eventually) Apr 18 02:21:31 wolverine, what distro are yuo using? Apr 18 02:22:08 using/building Apr 18 02:33:26 * chouimat is away: Zzzz Apr 18 02:34:10 night Apr 18 02:35:00 * JustinP wonders if this EABI thing really works Apr 18 02:35:11 * JustinP builds Apr 18 02:40:04 a distro defined by myself Apr 18 02:40:59 JustinP: what eabi stuff? Apr 18 02:42:59 TheMasterMind1: heh, you haven't been around long, huh? Apr 18 02:47:43 nope Apr 18 02:47:45 just wlaked in Apr 18 02:47:46 walked even Apr 18 02:52:57 emete, what this means ${@['virtual/${TARGET_PREFIX}libc-for-gcc', '']['nptl' in '${GLIBC_ADDONS}']} Apr 18 02:53:27 what's result if nptl defined? Apr 18 02:53:46 wolverine: '' Apr 18 02:54:23 Is it a python or shell syntax? Apr 18 02:55:06 python Apr 18 02:56:09 I see, it means if nptl not defined, it will returns ${TARGET_PREFIX}/libc-for-gcc Apr 18 02:56:15 thanks Apr 18 02:57:46 but why disalbed nptl in glibc? Apr 18 02:58:07 I built a glibc 2.3.4 with gcc 3.3.4, for i586 Apr 18 02:58:31 It has problem with thread Apr 18 02:58:37 * JustinP knows not Apr 18 04:13:56 03mwester 07org.oe.dev * ree282fd5... 10/packages/linux/ (2 files in 2 dirs): unslung-kernel: patch ppp-mppe.c to remove errant floating point computation Apr 18 04:57:10 wolverine, because you probably never removed all the patches that are in OE for emebedded devices that do not support it Apr 18 04:57:43 you should really look at/use an x86 based distro Apr 18 05:24:19 hiya ... i know i am in the wrong channel .. but can someone point me into the right direction , i have a question with regards to visual C++ Apr 18 05:24:38 i want to know how to creat setup files for my project Apr 18 05:24:55 try windows? Apr 18 05:25:32 i would guess that it may have some type of "wizard" tho Apr 18 05:25:39 thanks Apr 18 05:25:44 every hunk of junk in windows seems to Apr 18 05:26:07 Mamelodi: IIRC, they bundle installshield express with VC++, take a look on the install cds Apr 18 05:26:19 sweet Apr 18 05:26:28 thanks alot Apr 18 05:26:46 have a job to do. an i am noob when it comes to visual anything Apr 18 05:27:06 Mamelodi: Failing that, look into 'wix' on MSDN.. seriously good shit. Apr 18 05:27:23 kk Apr 18 05:27:27 will do Apr 18 05:28:03 wtf does that have to do with OE? :P Apr 18 05:28:06 Mamelodi: Most of my work's webapps, but for my Win32 development, I use Delphi and WiX for the installer Apr 18 05:28:16 mithro: Eh, vaguely related to software development =P Apr 18 05:28:37 lol Apr 18 05:28:42 i was lost Apr 18 05:28:57 got nothing to do with OE really Apr 18 05:29:12 but i seemed to have picked the right channel Apr 18 05:29:31 no, all the mean people are asleep Apr 18 05:29:37 cept me Apr 18 05:29:42 7:40 here Apr 18 05:29:45 in the morning Apr 18 05:29:47 where i am lol Apr 18 05:29:53 1539+1000 here.. Apr 18 05:30:23 thanks for the help Apr 18 05:30:26 gtg Apr 18 05:56:09 i'm awake Apr 18 06:54:40 hey Apr 18 06:54:56 I'm a physics student now (2nd subject) Apr 18 07:00:41 good morning all Apr 18 07:00:47 K'morn. Apr 18 07:00:55 s/^K/g/ Apr 18 07:01:24 * Necronom does a vmstat 1 on the 3100, and sees that it is pull 11k interrupts/sec. Apr 18 07:01:44 Hrm. English *is* my first language. That must mean it is time for bed. Apr 18 07:03:24 bonjour Apr 18 07:03:34 pb_: koen hey Apr 18 07:03:53 koen: I love how everything that runs on X is branded with GPE Apr 18 07:04:09 koen: I wonder what happens if I port and improve Opie PIM... Apr 18 07:04:18 koen: it will get called GPE Opie PIM? :} Apr 18 07:06:57 zecke: probably OpenGPE PIM Apr 18 07:10:13 GPE evolution, GPE nautilius, GPE epiphany... :} Apr 18 07:11:05 GPE x.org? Apr 18 07:11:06 =P Apr 18 07:11:33 "No, but you have to attend at all Tuesdays (or make us believe that you worked at other times during the week) and present your work at the end of the semester." Apr 18 07:11:43 relief... No Exam, Free Attendance Apr 18 07:14:55 ~hail AMD Apr 18 07:14:56 * ibot bows down to AMD and chants, "I'M NOT WORTHY!!" Apr 18 07:15:20 mr Danks was quite busy last night pushing changes Apr 18 07:22:56 hi Apr 18 07:23:06 good morning, is OE temporary broken? Apr 18 07:23:38 ERROR: Openembedded's config sanity checker detected a potential misconfiguration. Apr 18 07:24:05 NOTE: Handling BitBake files: / (2608/3253) [80 %]ERROR: Error Method already seen: qte_arch in' qte-common_2.3.10.inc' now in 'qte_4.0.0-snapshot.bb' Apr 18 07:24:53 the 'method already seen' is harmless Apr 18 07:25:05 the sanity checker ones deserves some attention Apr 18 07:25:43 this is not the only method, thats seen before, there are many more Apr 18 07:25:57 regarding gcc and glibc Apr 18 07:26:50 what makes me ask, if its broken, is that i did a fresh pull and co Apr 18 07:26:52 I get 20-30 warnings like that Apr 18 07:27:03 you can ignore them Apr 18 07:27:11 Please set TARGET_OS directly, or choose a MACHINE or DISTRO that does so. Apr 18 07:27:18 zecke: could you silence that warnings? Apr 18 07:27:40 the TARGET_OS error is something to worry about Apr 18 07:28:06 i set tosa and openzaurus Apr 18 07:29:42 yo zecke Apr 18 07:29:50 koen: do you fix them (the warnings)? Apr 18 07:30:04 koen: bad things could happen if the definitions of these functions are different Apr 18 07:30:13 zecke: no, most of the warnings are bogus Apr 18 07:30:28 koen: no, none of the warnings are bogus Apr 18 07:30:35 zecke: especially if gcc_{1,2,4,5,6,7} all include gcc.inc Apr 18 07:30:41 koen: if the function is shared between two bb files -> to the .inc file Apr 18 07:30:50 koen: yeah, but this function is defined in each .bb file Apr 18 07:30:59 koen: and not in the .bbclass, .inc or similiar Apr 18 07:31:19 but it's only executed per .bb, right? Apr 18 07:31:27 koen: no we have a global method scope :} Apr 18 07:31:33 (always had) Apr 18 07:31:55 koen: I will silence these warnings for the release though Apr 18 07:34:46 also i get NOTE: exceptions.SyntaxError:EOL while scanning single-quoted string (line 1) while evaluating: Apr 18 07:34:46 ${@base_set_filespath Apr 18 07:35:09 sws: which bitbake version? Apr 18 07:35:09 regarding angstrom and familiar Apr 18 07:35:13 450 Apr 18 07:35:21 did a svn up this morning Apr 18 07:35:28 sws: do you have a bogus DISTRO set? Apr 18 07:35:55 i copied over openzaurus-unstable to openzaurus-myunstable Apr 18 07:37:14 oh, sorry, that file cant exist, i did rm -Rf OE before the fresh pull ;) Apr 18 07:38:46 sws: sadly we can't have a better parsing report ;) Apr 18 07:38:55 k, errors regarding DISTRO are gone, the already seen Method persists, i will ignore them Apr 18 07:41:23 morning Apr 18 07:41:26 morning Apr 18 07:41:32 hi Apr 18 07:41:35 just as i'm heading to bed Apr 18 07:41:35 hey XorA & TheMasterMind1 Apr 18 07:41:56 * XorA goes to upgrade bitbake to see zecke's latest magic Apr 18 07:42:20 i should probably post the rest of my patches to bugzilla soon Apr 18 07:42:26 maybe tomorrow.. Apr 18 07:42:27 XorA: Don't forget RP! Apr 18 07:42:32 anyway, g'night Apr 18 07:42:36 TheMasterMind1: cya Apr 18 07:42:42 and RP of course Apr 18 07:42:49 'night TheMasterMind1 Apr 18 07:42:49 ~hail zecke RP Apr 18 07:42:50 * ibot bows down to zecke RP and chants, "I'M NOT WORTHY!!" Apr 18 07:43:07 ah time to get food Apr 18 07:43:28 yipee I have a second subject now ( just a couple of years late...) Apr 18 07:55:08 is it possible to get a lesser font for the console on poodle? i try to play around with a kernel 2.6 bootstrap-image Apr 18 07:55:50 sws: compile another font into the kernel? Apr 18 08:07:27 wierd, reading that washington post review of the 770 and his main gripe appears to be that USA networks dont allow bluetooth dialup, hardly nokias fault :-) Apr 18 08:10:51 XorA: and it lacks a keyboard Apr 18 08:11:35 XorA: or use the TT metric: 'it doesn't run on a 5000d, so it sucks' Apr 18 08:11:44 koen: heh heh Apr 18 08:12:29 did someone address the tosa usb to pc connection problem with kernel 2.6? even my poodle works with OE default kernel 2.6 Apr 18 08:12:30 zecke: you mean a global namespace for python functions? I think it would be better to split the namespaces up rather than trying to enforce uniqueness across all .bb files. Apr 18 08:13:47 pb_: well this way one does not need to recompile the functions of the bbclasses over and over again Apr 18 08:14:09 pb_: one could make it better though Apr 18 08:14:26 pb_: compile methods from .inc and .bbclass only once Apr 18 08:14:33 zecke: true. I suppose the ideal thing would be to have an explicit "global" flag that you could set on methods that you want to be shared. Apr 18 08:14:37 pb_: always compile methods from .bb Apr 18 08:14:42 right, exactly Apr 18 08:15:07 hmm still hungry Apr 18 08:16:39 zecke: not that I'm particularly opposed to uniquifying the names of the existing functions, but it seems like a bit of a maintenance overhead if you want to ensure that this is always the case, particularly in the presence of collections. Apr 18 08:17:19 hrm, I guess I should go to work Apr 18 08:17:37 pb_: good point (collections) Apr 18 08:17:50 pb_: the goal was to make awarenes and then silent the warning Apr 18 08:18:03 zecke: bitbake is now so fast, the text doesnt actually get time to draw over my ADSL link :-) Apr 18 08:18:51 XorA: and 1.5 will get even faster Apr 18 08:18:59 XorA: 1.4 might as well Apr 18 08:19:12 RP: Threaded builds shouldn't be too hard Apr 18 08:20:19 mickey|zzZZzz: will python's threading support extend? Apr 18 08:20:28 mickey|zzZZzz: e.g. use monitors instead of locks? Apr 18 08:20:46 mickey|zzZZzz: with Dual Core CPUs, Threading should get the norm Apr 18 08:22:09 with a threaded bitbake we could ask CosmicPenguin for an 8way opteron buildbox :) Apr 18 08:22:34 zecke: yeah, that'll be interesting. of course, Intel have had hyperthreading for years, and that doesn't seem to have promoted a huge amount of threading, but still. Apr 18 08:23:10 koen: and if we get a niagara we could host many threaded builds Apr 18 08:23:10 doing real threads with hyperthreading can actually degrade performance Apr 18 08:23:13 hey schurig Apr 18 08:23:24 koen: hi! Apr 18 08:23:27 zecke: hi 2u2 Apr 18 08:23:38 pb_: Do you know if the Pentium D chips have HT as well? Apr 18 08:23:41 morning all Apr 18 08:23:42 schurig: hey, how are you? Apr 18 08:23:54 schurig: RP added the csl gcc3, so we should have 'real' EABI support now Apr 18 08:24:03 zecke: intel has abandoned HT Apr 18 08:24:16 * zecke had hoped for four cpus Apr 18 08:24:50 zecke: I don't think they do Apr 18 08:25:16 koen: fine Apr 18 08:25:42 zecke: it sounds like Pentium EE does still have HT though Apr 18 08:25:48 koen: however, for one of my targets (I currently support 3) I don't have time for Kernel 2.6 in the fore-seeable future, so I cannot use a newer toolchain anyway Apr 18 08:26:00 pb_: yeah but no multicore? Apr 18 08:26:07 schurig: that's too bad Apr 18 08:26:17 koen: except there is a patch to enable kernel 2.4 to work with NTPL ... or I patch glibc2.4 to work with linuxthreads Apr 18 08:26:18 pb_: I plan to spend ~600Euro for the complete system not only for a CPU Apr 18 08:26:21 zecke: Writing a multi threaded bitbake is what I might look at next Apr 18 08:26:22 zecke: oh, I thought there were dual core EEs Apr 18 08:26:33 zecke: heh, right Apr 18 08:26:49 zecke: a mac mini? Apr 18 08:26:58 zecke: My Pentium D 2.8 doesn't have HT, just dualcore. Apr 18 08:27:16 RP: we should have a pipeline of to be build packages Apr 18 08:27:29 Good morning Apr 18 08:27:35 hey Bernardo Apr 18 08:27:40 RP: the only corner case I can imagine now is that we have selected X to proivde A Apr 18 08:27:47 RP: but if X fails, we go to Y Apr 18 08:28:01 RP: we would have to update all deps on X or such... Apr 18 08:28:02 zecke: The only concern is I might write the threading in C ;-) Apr 18 08:28:18 RP: rewrite of Bitbake? Apr 18 08:28:24 RP: well use Python Threads Apr 18 08:28:40 zecke: Not a total rewrite, no Apr 18 08:29:57 RP: no reason to use C there Apr 18 08:32:19 RP: okay then I will leave threading to you, and go to improve the testing Apr 18 08:33:24 RP: OT: Hacking QtE to support BGR sounds doable. The term is 'hacking' Apr 18 08:35:00 RP: what blockers do we have for BitBake 1.4? Apr 18 08:37:43 zecke: You could hack the kernel and qemu too :-} Apr 18 08:38:05 RP: QtE is to punish ;) Apr 18 08:38:18 RP: I won't make the kernel suffer more from my contributions Apr 18 08:38:24 QtE does deserve hacking more, I admit :) Apr 18 08:39:11 zecke: parsing the .bb's resulted in lots of errors this morning Apr 18 08:39:31 ArjanS: hehe, which kind of errors? Apr 18 08:39:39 ArjanS: about methods being redefined? Apr 18 08:39:48 yeah, exactly :) Apr 18 08:41:01 but I must say, both creating and especially parsing the cache have got some great speedups! Apr 18 08:41:15 parsing takes less then a second now! Apr 18 08:44:31 ArjanS: 'creating' it could be minutes faster :} Apr 18 08:45:05 RP: google python cow dict... Apr 18 08:45:13 RP: dict = { 'cow' : 1 } Apr 18 08:45:15 ... Apr 18 08:45:26 ArjanS: Wow.. Apr 18 08:45:32 right, really going to work now Apr 18 08:45:33 * pb_ bbiab Apr 18 08:52:10 oops away for a week and a bit, 223 revs , wow Apr 18 08:52:20 morning all Apr 18 08:52:32 hi Apr 18 08:53:35 hi do13_ Apr 18 08:55:10 hey do13_ Apr 18 08:56:09 hi zecke, koen, Bernardo Apr 18 08:59:07 hi do13_, I have a problem connecting tosa openzaurus kernel 2.6 using usb to my pc, module g_ether loads ok but tosa is not recognized when put in the cradle, is it a known problem? Apr 18 09:01:02 sws: Hmm. I haven't used udc since month. I'll check this again. Thx Apr 18 09:01:03 zecke: doing a 'monotone revert' on a file bitbake doesn't pick it up Apr 18 09:01:27 zecke: it insist on build gcc-4.1.0-r1, while the .bb is back to r0- Apr 18 09:01:28 koen: what is the mtime? Apr 18 09:02:01 hi do13_ Apr 18 09:02:15 zecke: -rw-r--r-- 1 koen koen 507 Apr 18 11:11 gcc_4.1.0.bb Apr 18 09:02:54 Hey Richard Apr 18 09:03:17 zecke: shouldn't it use something content based to spot changes? Apr 18 09:03:27 koen: well your filesystem sucks, use SUN UFS Apr 18 09:03:38 koen: it uses mtime to check if it needs to reparse the file Apr 18 09:04:34 koen: reading content can be expensive... :-) Apr 18 09:04:38 zecke: What am I looking for in google? Apr 18 09:05:45 RP: I wanted to search for a ready to use COW dict Apr 18 09:06:07 RP: and entered the terms, all you find is a introduction on how to put 'cow' (the string) into a dict Apr 18 09:06:47 koen: I hope to release bitbake soon, so I can hack on the tinderbox Apr 18 09:07:10 zecke: Right. Kind of disappointing :-/ Apr 18 09:07:17 njs: do you have an idea on how to find deleted files? Apr 18 09:08:10 zecke: Traditional COW would mean you copied the whole dictonary when any write was made to it of course. I still prefer the idea of parent child tracking and grafting the dicts Apr 18 09:08:24 zecke: what do you mean? Apr 18 09:08:50 njs: I want something like the 'Attic' in viewcvs Apr 18 09:09:08 zecke: err, hrm Apr 18 09:09:10 njs: I know there once was a file that was named f*o.b* and I wonder how I can find it using monotone Apr 18 09:09:17 compiling tosa openzaurus kernel 2.6 bootstrap-image: | Patch patches/usb-owner.patch does not apply (enforce with -f) Apr 18 09:09:17 NOTE: Task failed: /home/sander/openembedded/build/tmp/work/tosa-linux/wlan-ng-modules-0.2.3+svn20060418-r0/temp/log.do_patchcmd.9363 Apr 18 09:09:33 RP: hehe that is our 'COW' Apr 18 09:09:43 RP: we have the configuration Apr 18 09:09:49 zecke: the stupid way is: monotone log | egrep 'f.*o\.b' :-) Apr 18 09:10:00 RP: and set it as parent to our new dict Apr 18 09:10:03 | patching file src/prism2/driver/prism2_usb.c Apr 18 09:10:03 | Hunk #1 FAILED at 313. Apr 18 09:10:17 RP: now we lookup a key, we search in our local dict, then the parents Apr 18 09:10:33 RP: to enter a key, we look if we need to shadow the dict (COW) Apr 18 09:10:44 RP: we copy the 'value' of the key, into our dict Apr 18 09:10:45 sws: It seems this has been fixed upstream, try to remove the patch Apr 18 09:10:50 RP: and then do the manipulation Apr 18 09:10:54 I guess the smart way is: http://thread.gmane.org/gmane.comp.version-control.monotone.devel/6315/focus=6315 Apr 18 09:11:04 RP: so we have COW on a per key basis -> your parenting? Apr 18 09:11:04 but that's blue sky, not anything you can use now Apr 18 09:11:19 zecke: Yes, we're talking about the same thing Apr 18 09:12:36 RP: the only attention (which is luckily true for OE/BitBake) Apr 18 09:12:43 RP: we do not alter keys of our parents Apr 18 09:13:10 do13_ : removing the patch worked Apr 18 09:13:16 zecke: I guess that's a nice simplification :) Apr 18 09:14:08 zecke: Is everything based on the configuration data dict? Apr 18 09:14:14 TodoLost: Release bitbake, Implement a bonsai in python for svn and monotone, hack tinderbox Apr 18 09:14:25 RP: each parsed bbfile Apr 18 09:14:40 zecke: is the "o" there a typo or intentional? I can't tell :-) Apr 18 09:15:03 zecke: and configuration never changes once we start parsing files? Apr 18 09:15:25 RP: Well, I hope that Apr 18 09:15:45 njs: sadly not (typos in general) Apr 18 09:15:54 lol Apr 18 09:16:25 zecke: do you do this manually? Apr 18 09:16:47 mithro: COW? yes, do you know a better way? Apr 18 09:17:08 hey mithro Apr 18 09:17:22 you can do it via abusing classes if you wanted - dunno if it would be much faster Apr 18 09:17:24 OT: is packaging of glibc faster? Apr 18 09:17:43 (but it is an important case in python so it should be pretty fast) Apr 18 09:18:00 mithro: what 'classes' Apr 18 09:18:48 class A: pass Apr 18 09:19:03 how can one abuse classes? Apr 18 09:19:11 okay Apr 18 09:19:31 class A: p = "A" Apr 18 09:19:37 class B(A): pass Apr 18 09:19:41 mithro: we have a class with __getitem__, __setitem__ 'emulating' most parts of a dict Apr 18 09:19:50 it is an Adapter to dicts Apr 18 09:20:03 B.p == "A" Apr 18 09:20:10 B.p = "C" Apr 18 09:20:17 A.p == "A" Apr 18 09:20:23 I'm going to make a wild guess that just copying dicts is cheaper than overriding __magic__ methods Apr 18 09:20:34 B.p == "C" Apr 18 09:20:50 because the former is just some tuned C, while the latter requires python function calls, and even _dispatching_ those is crazy expensive :-) Apr 18 09:21:05 mithro: that's just syntactic sugar for dict assignment, though :-) Apr 18 09:21:12 njs: well, the main idea was to save memory Apr 18 09:21:26 zecke: oh, umm, sure, right, never mind then :-) Apr 18 09:21:30 njs: this is why you only copy 'shadowed' keys from the parent to the local one Apr 18 09:21:42 copy.copy (not deepcopy) Apr 18 09:21:58 njs: while keeping a compatible semantic to the old {} Apr 18 09:22:05 zecke: see how classes are copy on write? Apr 18 09:22:39 it would be intresting to see if its faster Apr 18 09:22:41 oh, duh, that's true, classes do do fallback when looking stuff up Apr 18 09:23:01 mithro: interesting Apr 18 09:23:12 however it's always copy on write Apr 18 09:23:16 mithro: and is lookup of A.p faster than dict[p]? Apr 18 09:23:46 ie even if you assign it "A" it's still assigned Apr 18 09:24:15 zecke: i think A.p is equal to dict[p] - if not using magic methods Apr 18 09:24:34 nice idea, i will try it Apr 18 09:24:58 i know it has continually gotten faster Apr 18 09:25:05 as it's fundemental part of python Apr 18 09:25:19 zecke: have fun reading the metaclass articles :-) Apr 18 09:25:47 i can help with the metaclass stuff if you need it Apr 18 09:26:01 libtpproto-py uses metaclasses quite alot Apr 18 09:26:05 anyway of to dinner Apr 18 09:26:06 bblr Apr 18 09:29:42 njs: OT: how do you generate these gcov websites for monotone? Apr 18 09:34:23 zecke: using the scripts it mentions at the top :-) Apr 18 09:34:48 "Code is available via netsync venge.net, branch org.vorpus.monotone-cov." Apr 18 09:35:10 it's all a bit shoestrings-and-duct-tape, but seems to work well enough :-) Apr 18 09:36:27 hi Apr 18 09:36:30 hey mickeyl Apr 18 09:37:19 (including various crazy stuff to adapt it to funky things about monotone, like it knows how to combine gcov data from multiple binaries (our unit tests produce a separate binary from our black box tests), and it knows how to recognize unit tests in the source and ignore them wrt coverage...) Apr 18 09:37:49 hail mickeyl Apr 18 09:37:51 njs: tape is everything one needs ;) Apr 18 09:37:57 mickeyl: hey Apr 18 09:38:10 hey mickeyl Apr 18 09:38:27 03mickeyl 07org.oe.dev * r84ac7904... 10/packages/lirc/ (9 files): lirc many versions: drop maintainership and remove obsolete versions Apr 18 09:38:30 hey mickeyl Apr 18 09:38:31 03mickeyl 07org.oe.dev * r6b727804... 10/packages/efl/etk_20060418.bb: add etk, a nice widget set based on the enlightenment foundation libraries Apr 18 09:38:35 03mickeyl 07org.oe.dev * r4306a260... 10/packages/poppler/ (poppler_20060418.bb poppler_0.5.1.bb): poppler: add a cvs version, because epdf needs some of the most recent features which are not in a release yet Apr 18 09:38:39 03mickeyl 07org.oe.dev * r6f801065... 10/packages/efl/etk_20060418.bb: etk: stage additional headers Apr 18 09:38:43 03mickeyl 07org.oe.dev * r794ba5c6... 10/packages/efl/epdf_20060418.bb: add epdf, a PDF rendering library based on poppler and the EFL Apr 18 09:39:09 mickeyl: does the ABI of poppler_cvs change? Apr 18 09:39:35 mickeyl: don't forget to push the x11 fix as well Apr 18 09:40:05 XorA: i don't think so Apr 18 09:40:06 mickeyl: where are the screenshots of epdf and epdfviewer? Apr 18 09:40:08 koen: yep Apr 18 09:40:26 zecke: patience is a virtue :D Apr 18 09:40:46 there is an epdfviewer to cause confusion with epdfview I added yesterday Apr 18 09:40:51 mickeyl: patience is what youth doesn't offer Apr 18 09:41:05 zecke: http://emma-soft.com/projects/epdfview/chrome/site/screenshots/version-0_1_1.png Apr 18 09:41:06 XorA: screenshots as well ;) Apr 18 09:41:16 zecke: indeed. you show your matureness by being able to realize that! Apr 18 09:41:23 s/matureness/maturity/ Apr 18 09:41:51 cya later, I need to pack my stuff, to visit the r-room Apr 18 09:42:26 koen: thats a different program is it not Apr 18 09:42:59 XorA: it's a screenshot :) Apr 18 09:43:25 koen: I was assuming epdf for E != epdfview that I added Apr 18 09:45:10 03mickeyl 07org.oe.dev * r61d534dc... 10/packages/xlibs/ (2 files in 2 dirs): diet-x11 x11r7-1.0.0 tell the truth about thy utf8 support (fixes #778) Apr 18 09:47:57 do13_: how may I debug that pxa2xx_udc thing? are there debug messages in runtime? Apr 18 09:49:04 koen: oh well, compiling plugins against compiled abiword doesnt help :-( **** ENDING LOGGING AT Tue Apr 18 09:59:57 2006