**** BEGIN LOGGING AT Sun Mar 19 10:59:57 2006 Mar 19 11:01:20 _law_: libglademm builds again in .dev Mar 19 11:01:40 hrm Mar 19 11:01:43 ~lart openssl Mar 19 11:01:43 * ibot readies the nuke launcher and fires some rounds at openssl Mar 19 11:02:36 <_law_> koen: nice Mar 19 11:10:34 koen: it seems cpumask_t is undefined in linux/interrupt.h, partly because we don't ship linux/cpumask.h ... Mar 19 11:11:04 morning zecke Mar 19 11:12:17 hey zecke Mar 19 11:19:39 good morning folks Mar 19 11:20:00 hey chris144 Mar 19 11:20:05 hi koen Mar 19 11:21:30 koen: are you familiar with x86_64 psyco problems? Mar 19 11:21:52 chris144: no, is a 32/32 system on the opteron Mar 19 11:21:59 koen: ok Mar 19 11:22:32 koen: seems like I cannot get psyco to work on my AMD x86_64 Mar 19 11:25:28 koen: just removing the #include from linux/rtc.h does the trick for atd. Mar 19 11:28:26 linux/interrupts.h, linux/irq.h and linux/cpufreq.h look like they shouldn't be in the llh package anyway. Mar 19 11:40:56 pH5: so lets remove them :) Mar 19 11:51:15 03koen 07org.oe.dev * r39aa6de6... 10/packages/gtkmm/gtkmm_2.8.3.bb: Mar 19 11:51:16 gtkmm 2.8.3: fixes Mar 19 11:51:20 * put more headers in -dev packages Mar 19 11:51:22 * install -d the gtkmm dir in staging before trying to copy file into it Mar 19 11:51:30 03rpurdie 07org.oe.dev * rb63cec57... 10/packages/quilt/ (quilt-native_0.39.bb quilt_0.37.bb quilt_0.39.bb): quilt: Fix SRC_URI for older packages Mar 19 11:51:38 03rpurdie 07org.oe.dev * r8ebef355... 10/packages/zaurusd/zaurusd_svn.bb: zaurusd: Correct DEPENDS Mar 19 11:51:45 03rpurdie 07org.oe.dev * rf722b6b3... 10/conf/machine/qemuarm.conf: qemuarm.conf: Use arm926ejs optimisation (and hence armv5te), pcmciautils as the pcmcia manager and correct SERIAL_CONSOLE Mar 19 11:51:51 03rpurdie 07org.oe.dev * r8ff4c967... 10/conf/machine/include/zaurus-clamshell-2.6.conf: zaurus-clamshell-2.6.conf: Add zaurusd to default images. Mar 19 11:52:01 03rpurdie 07org.oe.dev * rc2935576... 10/conf/machine/include/ (zaurus-clamshell-2.4.conf zaurus-clamshell.conf): zaurus-clamshell.conf: Make chkhinge 2.4 only Mar 19 12:02:33 pH5: do you still have a problem with socket()? Mar 19 12:34:45 03rpurdie 07org.oe.oz354x * r8ee7ea5b... 10/packages/quilt/ (quilt-native_0.39.bb quilt_0.37.bb quilt_0.39.bb): quilt: Fix SRC_URI for older packages Mar 19 12:34:51 03rpurdie 07org.oe.oz354x * ra422b8c8... 10/packages/linux/ (14 files in 2 dirs): Add linux-openzaurus-2.6.15 from .dev Mar 19 12:34:55 03rpurdie 07org.oe.oz354x * r362dc63c... 10/packages/zaurusd/zaurusd_svn.bb: Add zaurusd from .dev Mar 19 12:35:00 03rpurdie 07org.oe.oz354x * rea068ebd... 10/packages/hostap/ (files/hostap_cs.conf-upstream hostap-conf_1.0.bb): Add hostap-conf from .dev Mar 19 12:35:07 03rpurdie 07org.oe.oz354x * r0789e572... 10/packages/alsa/ (4 files): Add ALSA 1.0.10 from .dev Mar 19 12:35:13 03rpurdie 07org.oe.oz354x * r99ea5c48... 10/packages/tslib/tslib_cvs.bb: tslib: Add inherit pkgconfig (fix from .dev) Mar 19 12:35:19 03rpurdie 07org.oe.oz354x * r70121f46... 10/packages/base-files/ (3 files in 3 dirs): base-files: Updates from .dev Mar 19 12:35:33 03rpurdie 07org.oe.oz354x * r063098c0... 10/conf/machine/ (zaurus-clamshell-2.4.conf zaurus-clamshell.conf): zaurus-clamshell: Make chkhinge 2.4 only Mar 19 12:40:57 * mithro pokes RP Mar 19 12:41:01 you alive? Mar 19 12:51:37 ping koen Mar 19 13:00:40 heh Mar 19 13:00:47 259 mb of debian updates Mar 19 13:01:13 I should update more often Mar 19 13:01:28 hi koen Mar 19 13:01:43 hey gremlin[it] Mar 19 13:02:39 just a couple of info about ep93xx and oe ... to build u set MACHINE = ep93xx and what for DISTRO ? Mar 19 13:02:43 on local.conf Mar 19 13:02:55 distro doesn't really matter Mar 19 13:03:36 mhh ok ... i have familiar-unsatble btw ... Mar 19 13:04:28 koen: I'm just booting 2.6.16-rc6 - looks like the socket problem is still there :-/ Mar 19 13:04:30 "udevd[614]: init_udevd_socket: error getting socket: Function not implemented" Mar 19 13:04:48 anyone know rpurdie's postal address? Mar 19 13:06:47 to go over my problem with libc maybe i have to 'downgrade' binutils ... setting another version is binutils-cross_VERSION = _2.15.94.0.1 .. right ? Mar 19 13:10:13 mithro: yes Mar 19 13:11:08 That was to say I was alive. I do also know my postal address though! :) Mar 19 13:14:58 RP: could you message me your postal address, I have a present for you :P Mar 19 13:22:34 hi Mar 19 13:22:42 hi hrw Mar 19 13:22:43 RP: I see starting work on 354x ;) Mar 19 13:23:12 hrw: I have a build running. Its not looking like too many changes are needed :) Mar 19 13:23:48 hrw: Locally, I just have an updated zaurus-clamshell-2.6.conf and I see to specify a different CVSDATE for zaurusd Mar 19 13:24:29 RP: create zaurusd_fixeddate.bb for this Mar 19 13:25:19 RP: I want to avoid any cvsdate changes Mar 19 13:25:36 I'm not sure I agree with all these fixeddata.bb files, especially given bitbake's lack of parsing speed... Mar 19 13:25:48 hrw: I'm just setting CVSDATE_zaurusd = "xx" Mar 19 13:25:50 ok Mar 19 13:26:26 hrw: Should I make a copy of openzaurus-3.5.4.conf for 3.5.4.1? I presume I set DISTRO_VERSION to something like 3.5.4.1-alpha1? Mar 19 13:27:35 RP: or do openzaurus-3.5.4.1.conf with 'include conf/distro/openzaurus-3.5.4.conf' Mar 19 13:28:20 ok, that sounds better again. Mar 19 13:29:09 users still does not complain about 3.5.4 ;) Mar 19 13:29:36 Long may it continue :) Mar 19 13:29:37 RP: 3.5.4.1.conf can set DISTRO_VERSION to 3.5.4.1-alpha1 etc Mar 19 13:29:50 hrw: agreed. Mar 19 13:30:25 hi kergoth Mar 19 13:30:37 RP: we can do 3.5.4.1-test images when first ones will work - there are many users wanting to test Mar 19 13:31:01 * hrw want working 2.6.15 too ;) Mar 19 13:31:29 hrw: Yes. The alpha releases will just be to get the ball rolling. I'm hopeful to have them build by tomorrow and then we can see where things stand :) Mar 19 13:31:57 morning mickeyl Mar 19 13:32:19 hi mickeyl Mar 19 13:33:36 good morning folks! Mar 19 13:33:53 koen: any idea about: | /oe/c7x0/tmp/staging/arm-none-linux-gnueabi/include/asm/arch/irqs.h:11:28: error: asm/mach-types.h: file not found Mar 19 13:33:54 ? Mar 19 13:34:02 bails out while compiling opie-alarm Mar 19 13:36:00 That file is built by the kernel based on the current defconfig. I suspect the linux-headers package your using is broken in some way Mar 19 13:36:52 RP: ah right. Mar 19 13:36:55 * mickeyl looks to pH5 and koen Mar 19 13:37:18 mickeyl: touch the asm/mach-types.h file as a nasty hack that might just work ;-) Mar 19 13:37:28 heh, good idea Mar 19 13:37:29 * mickeyl trying Mar 19 13:37:37 jnc: ping? Mar 19 13:38:28 hmm another error now Mar 19 13:38:39 | /oe/c7x0/tmp/staging/arm-none-linux-gnueabi/include/linux/interrupt.h:36: error: expected specifier-qualifier-list before 'cpumask_t' Mar 19 13:38:39 | /oe/c7x0/tmp/staging/arm-none-linux-gnueabi/include/linux/interrupt.h:66: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'cli' Mar 19 13:38:39 | /oe/c7x0/tmp/staging/arm-none-linux-gnueabi/include/linux/interrupt.h:70: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'sti' Mar 19 13:38:39 | /oe/c7x0/tmp/staging/arm-none-linux-gnueabi/include/linux/interrupt.h:74: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'save_flags' Mar 19 13:38:42 | /oe/c7x0/tmp/staging/arm-none-linux-gnueabi/include/linux/interrupt.h:79: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'restore_flags' Mar 19 13:38:45 | /oe/c7x0/tmp/staging/arm-none-linux-gnueabi/include/linux/interrupt.h:84: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'save_and_cli' Mar 19 13:38:48 | /oe/c7x0/tmp/staging/arm-none-linux-gnueabi/include/linux/interrupt.h:157: error: expected specifier-qualifier-list before 'atomic_t' Mar 19 13:38:51 oops, that was longer than expected, sorry Mar 19 13:39:11 Should we tell you about pastebin? ;-) Mar 19 13:39:15 mickeyl: there are some files in the l-l-h package that shouldn't be there at all Mar 19 13:39:17 heh Mar 19 13:39:44 mickeyl: in this case, I think just removing the #include from linux/rtc.h should do the trick Mar 19 13:40:38 pH5: didnt you (or kone) just added it today? Mar 19 13:41:01 s/kone/koen Mar 19 13:41:22 pH5. ok, trying that Mar 19 13:42:00 NOTE: package opie-alarm-1.2.1+cvs-20060319: completed Mar 19 13:42:03 thanks pH5 and RP Mar 19 13:42:15 * mickeyl continues his first angstrom opie-image Mar 19 13:42:37 I'm really a bit uncomfortable with that package, because I went around the headers 'fixing' things without a deeper understanding. Mar 19 13:43:14 pH5: Perhaps see what gets done to the normal l-l-h headers and copy that? Mar 19 13:45:28 RP: I had llh 2.6.12.0, my headers and vanilla 2.6.16-rc6 in meld next to each other. But looking at the vanilla 2.6.12 / llh 2.6.12.0 diff might be a better idea even... Mar 19 13:49:14 cvs [checkout aborted]: connect to anoncvs.handhelds.org(141.2.14.64):2401 failed: Connection timed out Mar 19 13:49:18 * mickeyl kicks anoncvs.handhelds.org Mar 19 13:53:34 anyone know the default port for syncml? Mar 19 13:56:02 hail mickeyl Mar 19 13:56:04 wb zecke Mar 19 13:56:27 cheers pb_ Mar 19 13:56:35 morning zecke Mar 19 13:56:54 morning pb__, zecke Mar 19 13:56:55 hi Mar 19 13:57:05 zecke: yeah, that two months passed quick ;) seriously, i couldn't leave it and had to look at it... when i found it was a one liner, i could hesitate to check it in :D Mar 19 13:57:15 couldn't even Mar 19 13:57:19 zecke: heh, finished soldering my chip. looks kind of ugly, but it seems to work. Mar 19 13:57:35 pb__: congrats :) Mar 19 13:57:59 mickeyl: may we call you Dr. Lauer now? Mar 19 13:58:07 zecke: heh, no :/ Mar 19 13:58:10 :( Mar 19 13:58:14 mickeyl: doh Mar 19 13:58:17 i guess the time warp was only in my mind Mar 19 13:58:30 * zecke notes to himself a PhD is not a one liner. Mar 19 13:58:52 yeah, fixing qt is a one liner Mar 19 13:58:53 hrm, i can't seem to get my new phone to use syncml Mar 19 13:58:58 not finishing your thesis :) Mar 19 13:59:04 bbiab, cleaning dishes Mar 19 13:59:24 mickeyl: you should get a dog for that Mar 19 14:01:51 mickeyl: I have removed the Dust from the SIMpad and it works perfectly with a 2.4er kernel Mar 19 14:02:17 mickeyl: just these badly written Games hardcoding the resolution to 240x320 Mar 19 14:03:25 zecke: but you can fit few of them on screen... Mar 19 14:03:49 right. multiplayer! Mar 19 14:03:53 zecke; did'nt you fix some of 'em? Mar 19 14:04:46 pb__: normal man have two hands so equipped with two styli.. can play atleast two at once Mar 19 14:08:50 * koen finished upgrading mythtv to 0.19 Mar 19 14:08:58 zecke: now that we have UCB/MCP support in the kernel it shouldn't be hard to finish 2.6 for SIMpad or is it? Mar 19 14:09:08 koen: while you're there... could you upgrade tetex and lyx? :) Mar 19 14:09:09 "konqueror-embedded: "No rule to make target `-lpcre' needed by `konqueror'. Stop. " Any known fix? Mar 19 14:09:16 eeek, that again Mar 19 14:10:10 mickeyl: in OE? Mar 19 14:10:29 ya Mar 19 14:11:13 RP: have a look at the OE mailing list archive, this question was answered there. I don't know the fix because it never occured to me but it was installing something on your host Mar 19 14:11:22 IIRC Mar 19 14:11:41 Just googling. Looks like it needs libpcre-dev Mar 19 14:11:50 ah right Mar 19 14:11:55 hmm Mar 19 14:11:59 could you add a hint into the .bb file? Mar 19 14:13:42 mickeyl: Will test and add the hint if found to work Mar 19 14:13:50 thanks Mar 19 14:14:37 OT: anyone of you know whether GSM has been adopted as an ISO standard and if so, which # ? Mar 19 14:16:54 mickeyl: building lyx right now Mar 19 14:17:04 cool. Mar 19 14:17:24 heh Mar 19 14:17:33 somewhere lyx requires openssl Mar 19 14:17:48 and openssl doesn't know about arm-eabi Mar 19 14:17:59 ah bummer Mar 19 14:18:54 wow Mar 19 14:19:17 OE integrates the openssl buildsystem in to openssl.inc Mar 19 14:19:23 * koen won't touch that Mar 19 14:19:30 arg Mar 19 14:29:20 hi all Mar 19 14:29:26 hi dirk Mar 19 14:29:39 hey Marcin Mar 19 14:29:40 hey do13 Mar 19 14:29:49 hey koen Mar 19 14:30:21 03mickeyl 07org.oe.dev * r1d509672... 10/packages/perl/perl.inc: perl: fix SRC_URI to use ftp, the http structure has been moved Mar 19 14:31:46 pH5: do you think you can add a few 'rm linux/interupts.h' type statement to the llh.bb? Mar 19 14:32:41 hi Dirk Mar 19 14:33:04 hi Richard Mar 19 14:33:14 hrw: will 3.5.4.1 also have updated images for c7x0? Mar 19 14:33:24 koen: It will Mar 19 14:33:46 RP: I suspect it needs the matchbox 1.0 bits as well Mar 19 14:34:28 koen: What was added that it needs? Mar 19 14:34:54 RP: you need it to unbreak the fonts on high dpi devices Mar 19 14:35:10 RP: I think that we can update to 1.0 - 1.0 looks better then 0.9+svn Mar 19 14:35:14 koen: The dpi changes made the branch in the end? Mar 19 14:36:03 is so, matchbox 1.0 is advisable Mar 19 14:36:46 RP: enabling xsettings is good to have, regardless if we lie about dpi :) Mar 19 14:39:11 RP: I added tosa to the zaurusd package. could you add this to your svn http://www.do13.de/oz/temp/zaurusd-tosa.patch Mar 19 14:39:44 mickeyl: GSM are many ETSI standards Mar 19 14:39:53 mickeyl: which one in particular are you interested in? Mar 19 14:41:38 do13: I'll get that added, thanks. Looks good. Mar 19 14:42:04 RP: thx Mar 19 14:42:56 does zaurusd has some doc how it works? Mar 19 14:43:32 hrw: Not yet... Mar 19 14:43:52 OT: How much is a clamshell device nowadays? Mar 19 14:44:08 zecke: about E250 Mar 19 14:44:28 zecke: c1000 from pricejapan.com Mar 19 14:45:52 koen: it says E320 Mar 19 14:46:04 zecke: 322 euro including shipping Mar 19 14:47:12 koen: hmm, I will wait for my ROAD device then :} Mar 19 14:47:18 hrw: Basically zaurusd is a lot of shell script black magic Mar 19 14:47:26 zecke: thanks, I've just wondered what to cite when I briefly talk about gsm and umts. Since there are so many I just cited the specification area of the 3gpp group now. Mar 19 14:47:31 zecke: when? :D Mar 19 14:47:56 * zecke looks innocent Mar 19 14:47:59 I don't know Mar 19 14:47:59 hey dirk Mar 19 14:48:04 mmh Mar 19 14:48:07 you promised march :D Mar 19 14:48:08 hey mickeyl Mar 19 14:48:10 march 2006 even Mar 19 14:48:23 zecke: kick those marketing folks Mar 19 14:48:33 mickeyl: and you have seen it in February :) Mar 19 14:48:43 fair enough :) Mar 19 14:48:58 even played with it in februari :) Mar 19 14:49:21 mickeyl: I'm about to get me a palm :} Mar 19 14:49:23 I need PIM Mar 19 14:49:36 heh Mar 19 14:49:41 good choice Mar 19 14:49:45 they're cheap nowadays Mar 19 14:49:47 cheap and functional Mar 19 14:49:56 get one without a color display Mar 19 14:49:59 they last _pretty_ long Mar 19 14:50:03 mickeyl: I will steal one from steve :} Mar 19 14:50:05 get m105 ;) Mar 19 14:50:18 zecke: Do more work on opie :) Mar 19 14:50:21 one of the old Sony ones Mar 19 14:50:35 do13: hehe, I wish I could :} Mar 19 14:50:47 zecke: Sony Clie rulz Mar 19 14:51:03 do13: I'm a bit in the air now. I would like to see Qtopia4 Mar 19 14:51:26 but sadly TT marketing doesn't know what they want Mar 19 14:51:33 zecke: qtopia4 is incompatible, listen to lorn and stay with qt/e 2 for the next 10 years Mar 19 14:51:37 *cough* Mar 19 14:51:47 :) Mar 19 14:52:06 koen: Actually, I think 5+ years of QtE2 are more realistacally Mar 19 14:52:17 koen: no one at TT would have expected that when writing Qt2 :} Mar 19 14:52:41 i guess the embedded team didn't expect how slow v4 is :/ Mar 19 14:52:48 do13: zaurusd svn is updated, thanks Mar 19 14:52:58 RP. thx Mar 19 14:56:27 hmm what task should I schedule now.. SIMpad kernel hacking, Doc writing, BitBake C Parser, or iSync Qtopia/Opie sync Mar 19 14:56:58 C parser Mar 19 14:57:25 C parser Mar 19 14:57:37 thirded ;-) Mar 19 14:58:07 don't you want to get your SIMpad running, or sync PIM with OSX? Mar 19 14:58:32 no simpad, no osx here and I do not read docs Mar 19 14:58:38 C parser Mar 19 14:58:54 if I want to sync my pim with osx a opensync->isync conduit would help a lot more than some blechtopia thing Mar 19 14:59:27 koen: a generic one? Mar 19 14:59:40 koen: well I would start with Qtopia/Opie to understand the API Mar 19 14:59:43 OT: what do you think about new layout on my website? http://www.hrw.one.pl/ Mar 19 14:59:51 zecke: doc writing :-} Mar 19 14:59:52 koen: I have though of implementing syncML on the client as well Mar 19 15:00:05 looks nice Mar 19 15:00:19 hrw: looks nicer :) Mar 19 15:00:31 mickeyl: okay. I have Python Questions Mar 19 15:00:41 * mickeyl runs Mar 19 15:00:42 mickeyl: If I implement the C parser and the data module in C Mar 19 15:00:58 mickeyl: is there a clean way where I can avoid the Py calling overhead? Mar 19 15:01:05 sooner or later I will implement that theme on OZ.org Mar 19 15:01:58 zecke: which calling overhead exactly? You need to run inline python functions, you can't avoid that. Mar 19 15:02:56 zecke: implementing the parser and data module in C will give a HUGE performance boost. The couple of python functions necessary to talk to the C extension module won't be significant. Mar 19 15:03:05 you can safely ignore them for now :) Mar 19 15:03:06 mickeyl: from the parser I will do setData this goes C->PyC->Python->C-PyC Mar 19 15:03:47 * zecke feels like moving this big stone up the Performance Hill Mar 19 15:04:44 that doesn't sound right. you need a shortcut to set data from C to C, especially considering that the C module is implemented in C as well Mar 19 15:04:49 parser module, even Mar 19 15:05:13 mickeyl: right, is there some shortcut? Mar 19 15:05:24 i think so. you should have a look at how sip is doing its internal magic. Mar 19 15:05:25 mickeyl: we have created that 'data' abstraction Mar 19 15:06:27 * zecke takes a break before starting the work :} Mar 19 15:06:54 To save me going totally mad, what's wrong with the code "if (version.LooseVersion(bb.__version__) < version.LooseVersion(minversion)):" Mar 19 15:07:21 It results in: File "/usr/lib/python2.3/distutils/version.py", line 296, in __cmp__ Mar 19 15:07:21 return cmp(self.version, other.version) Mar 19 15:07:21 AttributeError: LooseVersion instance has no attribute 'version' Mar 19 15:08:11 heh Mar 19 15:08:17 don't call it version Mar 19 15:08:22 version is a distutils module Mar 19 15:08:44 or is it that what you want to talk to= Mar 19 15:08:45 ? Mar 19 15:09:04 I'm trying to use that module Mar 19 15:10:00 argh Mar 19 15:10:09 you're not mad Mar 19 15:10:17 looks likes a bug in that module Mar 19 15:10:26 >>> x=distutils.version.LooseVersion() Mar 19 15:10:26 >>> y=distutils.version.LooseVersion() Mar 19 15:10:26 >>> x > y Mar 19 15:10:26 Traceback (most recent call last): Mar 19 15:10:26 File "", line 1, in ? Mar 19 15:10:26 File "/usr/lib/python2.4/distutils/version.py", line 296, in __cmp__ Mar 19 15:10:28 return cmp(self.version, other.version) Mar 19 15:10:30 AttributeError: LooseVersion instance has no attribute 'version' Mar 19 15:11:02 dir(x) Mar 19 15:11:29 RP: did you parse a version yet? Mar 19 15:11:40 self.version gets set by parse Mar 19 15:11:49 if you compare prior to that, the attribute is not set yet Mar 19 15:11:56 (from a first glance at the module) Mar 19 15:13:30 mickeyl: No, I didn't parse a version so its looks like I'm abusing the module... Mar 19 15:13:48 yeah Mar 19 15:14:19 to be honest, it shouldn't bail out like that though. in my own python modules, I set all attributes to a sane default value in the constructor Mar 19 15:16:42 I'd swear I had this working but in my big cleanup, I've lost the final version of the sanity class file :-/ Mar 19 15:18:08 mickeyl: the cmp function should have a sane decorator Mar 19 15:18:15 mickeyl: that could check this invariant Mar 19 15:20:22 ya. major portions of the Python STL are still from the 1.5 days and that shows. Mar 19 15:20:31 s/STL/SL/ Mar 19 15:20:39 habits... Mar 19 15:22:31 morning Mar 19 15:22:51 hey chouimat, how are things? Mar 19 15:24:59 mickeyl: good but a little slow Mar 19 15:25:11 | configure: error: "MySQL needs a off_t type." Mar 19 15:27:32 Offhand does anyone know if/how opie can be set to handle a framebuffer with BGR instead of RGB pixel ordering? Mar 19 15:28:21 pH5: didn't you talk about adding a cjk input method to OE a while ago? Mar 19 15:30:35 * pb__ heads off to write some marketing literature Mar 19 15:30:36 bleh Mar 19 15:30:49 koen: other things stole all my time, but yes, basically the missing tasks are finding a working HWR engine for kana that is not written in C++ and getting im-ja or similar to work. Mar 19 15:30:51 pb__: poor soul Mar 19 15:31:03 zecke: haven't got much soul left Mar 19 15:31:22 pH5: what makes kana HWR in C++ bad? Mar 19 15:31:26 it seems I am 0wn3d by the sales and marketing department Mar 19 15:32:05 zecke: I don't speak C++ well, which makes ripping out and integrating into GPE yudit's recognition engine very time consuming. Mar 19 15:32:38 pH5: leira (who entered and left #oe just now) was talking about getting scim to work in OE Mar 19 15:33:06 pb__: My commiserations :-( Mar 19 15:33:35 * RP is thankful OH doesn't have a sales and marketing department... Mar 19 15:33:43 time to food something Mar 19 15:33:44 (yet) Mar 19 15:33:57 * hrw had a talk with client at Friday... Mar 19 15:33:59 zecke: yes, yet... :-/ Mar 19 15:36:03 * zecke needs food as well Mar 19 15:38:56 ~lart opensync.org Mar 19 15:38:56 * ibot gets a hotmal account and SPAMs opensync.org Mar 19 15:39:12 why can't they put the files in a location wget can actually understand? Mar 19 15:42:55 koen: I was under the impression that scim is a little heavy-weight, but that is interesting in any case. Mar 19 15:45:28 mickeyl: I'm still looking for a way into PyC Mar 19 15:45:39 mickeyl: should I start by updating setup.py to build the C++ stuff Mar 19 15:45:48 mickeyl: or should I start compiling it by hand? Mar 19 15:49:42 i recommend starting from scractch Mar 19 15:50:34 mickeyl: compile by hand? Mar 19 15:50:58 yep. distutils integration is just a Fingerübung Mar 19 15:52:50 mickeyl: How do you like the new pythin.org site? Mar 19 15:54:05 i like it better than the old one Mar 19 15:54:23 it looks more professional for sure Mar 19 15:54:30 whatever that may be :) Mar 19 15:56:44 folks, I was just experimenting: 'python do_whatever_prepend' works, but 'python do_whatever_append' does not. anybody familiar with that? is there a reason why that is so? thanks Mar 19 15:57:37 zecke: I did have a semi working C data module at one point. At the time I then got very confused about bitbake internals and gave up on it. GIven what I know now, I should probably look at it again :) Mar 19 15:58:15 RP: I feel like the dawrfs from Southpark Mar 19 15:58:29 1. C++ Parser 2. ??? 3. Success/Profit Mar 19 15:58:56 zecke: I know the feeling all too well... Mar 19 16:01:00 lol Mar 19 16:01:18 mickeyl: from tkappinit.c /* XXX I've heard rumors that this doesn't work */ Mar 19 16:04:34 heh, cute Mar 19 16:05:37 chris144: sounds like a bug Mar 19 16:06:26 pb_: you think so? should I put it in bugzilla? Mar 19 16:06:37 chris144: Make sure the function you're appending to didn't end in something like a "return" first though Mar 19 16:07:44 chris144: I guess you should. Please attach a testcase showing an _append and a _prepend, together with the resulting data in each case. Mar 19 16:08:10 RP: no there was no return... Mar 19 16:08:34 pb_: ok, I'll reproduce and put it in bugzilla. Mar 19 16:13:00 *cry* Mar 19 16:13:35 lart BSD flex Mar 19 16:14:11 zecke: If I remember rightly, even old versions of gnu flex misbehave with the C parser Mar 19 16:14:32 RP: well mine doesn't even know %option reentrant stack Mar 19 16:14:49 and I think I need that for inherit,include,require Mar 19 16:14:55 Well, 3.5.4.1-alpha1 sort of worked. Got as far as the touchscreen calibration and then failed :) Mar 19 16:17:45 Why would bb.__version__ be getting returned as None? Mar 19 16:18:24 RP: opie or gpe? Mar 19 16:18:31 koen: opie Mar 19 16:18:50 gpe is building Mar 19 16:18:59 that can probably be explained with hardcoding event somewhere Mar 19 16:19:11 or does zaurusd fix that for opie nowadays? Mar 19 16:20:11 koen: Its a broken udev Mar 19 16:20:40 0.84 is broken as well? Mar 19 16:20:48 koen: It wasn't using 0.84 Mar 19 16:20:57 koen: zaurusd will fix that eventually but the zaurii input device numbers seem quite fixed - I've never seen them change Mar 19 16:21:23 lucky you Mar 19 16:21:26 I agree its a race though :) Mar 19 16:22:08 let's see how that magical udev rule will handle that Mar 19 16:22:59 koen: Quite easily - create /dev/input/touschsreen0 for touchscreens and /dev/input/keyboard0 for keyboards Mar 19 16:28:29 Yay. 3.5.4.1 opie works on spitz including alsa sound nearly out the box :) Mar 19 16:28:53 well, it does work out the box now I fixed udev... Mar 19 16:31:40 This is beautiful :) Mar 19 16:31:45 cool Mar 19 16:33:51 excellent! good job Mar 19 16:34:57 At the moment its using pcmcia-cs. I'm feeling tempted to leave it alone for 3.5.4.1 given its all working... Mar 19 16:35:44 next weekend I will be offline Mar 19 16:37:51 RP: what about test images during week? Mar 19 16:38:29 OT: would parsing 3222 files in 30 seconds be an improvement? Mar 19 16:38:35 hrw: I think that's quite realistic given the way this spitz opie image feels Mar 19 16:38:42 zecke: it is Mar 19 16:38:43 zecke: yes Mar 19 16:39:49 posted info on oesf forums Mar 19 16:44:07 zecke: mmh.. Yeah! Mar 19 16:45:30 koen|food: looks like people are still fetchiing OZ Mar 19 16:45:37 http://ewi546.ewi.utwente.nl/rrd/graph/traffic_eth0_604800.png is nice Mar 19 16:48:45 chris? Mar 19 16:49:57 hrw: 5MBit/s peak. Nice :) Mar 19 16:51:21 RP: OZ effect Mar 19 16:51:34 03rpurdie 07org.oe.oz354x * rcac0aae6... 10/packages/udev/udev_084.bb: udev 0.84: Promote to default Mar 19 16:51:40 03rpurdie 07org.oe.oz354x * r0c581dde... 10/packages/hostap/ (5 files): hostap-modules/hostap-utils: Create virtual/kernel-hostap so hostap-modules can be either external, a kernel module or built into the kernel Mar 19 16:51:44 03rpurdie 07org.oe.oz354x * r14b9fce0... 10/packages/base-files/ (3 files in 3 dirs): base-files: Update spitz and borzoi fstab for usbdevfs -> usbfs change. Mar 19 16:51:52 03rpurdie 07org.oe.oz354x * rf991efa0... 10/conf/machine/zaurus-clamshell-2.6.conf: zaurus-clamshell-2.6: Updates for 2.6.15 kernel and other changes from .dev Mar 19 16:52:04 hrw: lots of people are doing wget -r Mar 19 16:52:05 03rpurdie 07org.oe.oz354x * r761983bf... 10/conf/distro/openzaurus-3.5.4.1.conf: Add openzaurus-3.5.4.1 DISTRO Mar 19 16:52:16 pcmciautils going in or not RP? Mar 19 16:52:43 ade|desk: Can't make my mind up... Mar 19 16:53:20 go on you will, you will , you will Mar 19 16:55:18 koen: is GPEsyncd working well? Mar 19 16:55:31 zecke: it is Mar 19 16:55:45 koen: did you have data loss yet? Mar 19 16:55:58 zecke: sure, lots of them in the past Mar 19 16:56:18 hmm Mar 19 16:56:23 but generally no important stuff is in one place Mar 19 16:58:13 go Richard! Mar 19 16:58:18 :D Mar 19 16:58:34 * mickeyl happy to see 3.5.4.1 so early Mar 19 16:58:44 how much RCs will 3.5.4.1 have? Mar 19 16:59:08 as far as we need. i'd hope just 1 Mar 19 16:59:44 s/far/much/ Mar 19 17:00:56 any ideas what will be a reason for 3.5.4.2? other then tosa/2.6 Mar 19 17:02:05 hrw: pong Mar 19 17:02:20 jnc: you do wifi work in OE Mar 19 17:02:25 yep Mar 19 17:02:38 jnc: could you test WPA in oz 3.5.4 and write howto about it? Mar 19 17:02:51 * jnc reads irclog Mar 19 17:03:49 * zecke pulls hairs :} Mar 19 17:04:03 NAbyss: reboot, i would guess. that is what i did anyways. probably not needed Mar 19 17:04:18 jnc: Yeah, I did that.. card still seems to work :) Mar 19 17:04:23 hrw, jnc: Testing/fixing wpa in 3.5.4.1 if needed would also be appreciated Mar 19 17:04:29 Just need to test it with an AP.. will give it a burl at work Mar 19 17:04:41 But, for now.. last minute assignment work calls Mar 19 17:04:42 RP: if it works in 3.5.4 then it also works in 3.5.4.1 Mar 19 17:04:57 hrw: True Mar 19 17:05:00 RP: kernels have WE19, wpa-supplicant does not change Mar 19 17:05:41 jnc: if something needs to be updated then give me sign Mar 19 17:06:55 * ade|desk finally finishes watching koen's presentation at fosdem, shame the sound is so faint Mar 19 17:06:56 03rpurdie 07org.oe.dev * rcb0dad15... 10/packages/udev/udev_084.bb: udev 084: Make default Mar 19 17:07:15 03rpurdie 07org.oe.dev * r6c946bf3... 10/packages/base-files/ (3 files in 3 dirs): base-files: Update spitz and borzoi fstab for usbdevfs -> usbfs change. Mar 19 17:07:25 ade|desk: thx for reminding... I have Liam presentation to watch Mar 19 17:07:29 03rpurdie 07org.oe.dev * rb8b3a6ad... 10/packages/hostap/ (8 files): hostap-modules/hostap-utils: Cleanup DEPENDS/(R)PROVIDERS - virtual/kernel-hostap is a runtime entity and allows the hostap modules to be external kernel modules, internal kernel modules or built into the kernel. Mar 19 17:07:56 RP: that hostap thing has a bug in bugzilla - can you put info there? Mar 19 17:08:19 hrw: This might even fix that bug Mar 19 17:08:26 yep Mar 19 17:09:16 hrw: 3.5.4 is no-go, i have c3000 and c3100 hardware; will do it for 3.5.4.1 if you like? Mar 19 17:10:55 RP: is it time to look at making cxx00 suspend/resume work in 2.6 linux kernel? Mar 19 17:11:10 jnc: ok Mar 19 17:11:33 * hrw want sharp hc-6000n Mar 19 17:11:38 jnc: Try the 3.5.4.1 branch - I've just been testing that on spitz and opie works at least Mar 19 17:11:54 RP: okay, i can do that Mar 19 17:12:04 jnc: I've got a little more time available now so perhaps we should look at that, yes Mar 19 17:12:16 :) Mar 19 17:12:22 glad. i'm selling my c3000 this month Mar 19 17:12:36 argh time to make food Mar 19 17:13:21 * ade|desk mumbles something about pcmciautils at RP ;) Mar 19 17:13:30 the good news is that may be enough time to fix the c3000 suspend/resume, and also i will be receiving an sl-5500 Mar 19 17:13:35 more hardware to test :) Mar 19 17:14:37 jnc: The trouble is you need me to find enough time to port that horrible sharp i2c power code into 2.6 and its likely to be painful. In the meantime, 2.6.16 will be released and I'll have a two week deadline to get a ton of zaurus patches merged into mainline :-/ Mar 19 17:16:33 it would be better to get patches in mainline ;) Mar 19 17:17:02 i agree Mar 19 17:17:13 worry about horrible code later Mar 19 17:17:28 We'll see how my time stacks up... Mar 19 17:17:42 jnc: Does suspend/resume work happily on your 3100? Mar 19 17:17:53 haven't tried it, that one runs SharpROM Mar 19 17:18:26 RP: at OH you work as kernel expert? Mar 19 17:18:44 jnc: Given its the same code, it would be interesting to know Mar 19 17:18:54 hrw: Amongst other things, yes Mar 19 17:19:33 ok. time to clean home a bit - cu tomorrow Mar 19 17:20:07 i agree, but i am using that device every day. i know this sounds silly, but until the c3000 suspend/resume is fixed, i'm not about to trash the working SharpROM setup on the c3100 Mar 19 17:20:10 :/ Mar 19 17:20:57 I understand. I'd like to be able to use my device every day too :-/ Mar 19 17:22:12 the applications i use include spreadsheet (hancomsheet), and a text editor. the rest is not very useful. i can switch when suspend/resume is working, gnumeric runs reliably, and i can access the device using only keystrokes. Mar 19 17:22:34 those are not your problem to deal with :) Mar 19 17:23:20 it's just that i need that stuff personally Mar 19 17:24:05 RP: can you identify all the bits of code that are present in the sharprom for cxx00 that are not present in the oz kernel? Mar 19 17:24:27 like, do you have a mental picture of all those bits of code, or are some of them question marks? Mar 19 17:24:41 jnc: I have a very good image of that Mar 19 17:24:46 RP: If you're talking about suspend/resume invoked by the power button, it appears to work for the 3100 running 2.6.15. Mar 19 17:25:23 NAbyss: It appears to work for most people except jnc on his C3000 :-/ Mar 19 17:25:43 RP: Could the userspace environment make a difference? I'm running OPIE here Mar 19 17:26:18 NAbyss: We tracked it down to the kernel in jnc's case - its the specific hardware and kernel setup :-/ Mar 19 17:26:22 Ah.. Mar 19 17:29:34 RP: are you the guy from the hx2000 port Mar 19 17:29:44 Harvy: I am Mar 19 17:29:56 03rpurdie 07org.oe.dev * r35362094... 10/site/arm-linux: Add dpkg configure hints Mar 19 17:30:00 03rpurdie 07org.oe.dev * ra5ca4ea8... 10/packages/dvbtools/ (dvbtraffic.bb sctzap_cvs.bb): dvbtools: Change CVSDATE references -> SRCDATE Mar 19 17:30:04 03rpurdie 07org.oe.dev * r522ddd4c... 10/packages/konqueror/ (2 files): konqueror-embedded: Add information about pcre workaround to .bb file. Mar 19 17:30:46 RP: alsa trying to test .deb package with OE? Mar 19 17:31:18 RP: just a quick question then. Are the patches in monotone? Mar 19 17:31:42 Harvy: The patches are on the website but there is a .bb file in OE which tells you how to apply them properly Mar 19 17:32:09 koen: Its a fix from a while back that got lost in my dev trees. It removes the need for your dpkg autofoo.patch Mar 19 17:33:02 do you know which .bb file? Mar 19 17:33:30 Harvy: packages/linux/linux-openzaurus_2.6.15.bb Mar 19 17:33:49 Harvy: But I suspect they're a bit broken against 2.6.15 Mar 19 17:42:29 3.5.4.1 GPE sort of works too. We need to do something about the fonts though Mar 19 17:42:47 RP: unbreak matchbox ;) Mar 19 17:42:53 The autorotate is also "interesting" Mar 19 17:43:19 Its making assumptions from poky that aren't valid in GPE :-/ Mar 19 17:43:23 RP: ah in the .dev branch; any thing in the .ozxxxxx branch? I am asuming that RPSRC is an env to you web site...? Mar 19 17:43:37 RP: using libgtkinput in portrait mode would be a nice touch Mar 19 17:44:32 Harvy: An older version of the .bb file should be in the .oz/fam branch. RPSRC is a variable for my site, yes Mar 19 17:44:37 koen: patches welcome :) Mar 19 17:45:00 RP: is mallum still alergic for patches that contain 'gpe'? Mar 19 17:45:36 koen: yes Mar 19 17:46:15 koen: which bits of matchbox do I need to update? Mar 19 17:46:28 RP: panel and wm Mar 19 17:46:44 RP: cheers wil try again for my hx2100 once emerge world done for my build system.. Any other useful hints Mar 19 17:47:28 RP: OE has a patch to provide maemo style keyboard popups in entry fields, but the matchbox end of the patch uses 'gpe' in function names Mar 19 17:49:51 koen: That's probably enough to get it ignored :-/ Mar 19 17:50:11 RP: D'oh :( Mar 19 17:50:37 koen: Is there a mb-panel 1.0 in OE? Mar 19 17:51:21 RP: no, since I don't have the skill to forward port the patches Mar 19 17:51:36 RP: and I expected the bits to enter poky within seconds Mar 19 17:51:39 03hrw 07org.oe.oz354x * r994453b4... 10/packages/keymap-locales/ (10 files in 7 dirs): Mar 19 17:51:43 added keymap-locales from .dev Mar 19 17:51:45 also added RDEPENDS="keymaps" as we have it separated in OZ Mar 19 17:52:05 RP: I stopped doing duplicate work after looking what is in OH svn, but not in OE Mar 19 17:52:24 s/OH?poky/ Mar 19 17:53:16 koen: I'm not sure there is that much in poky svn that's different. It could be I missed a set of commits though Mar 19 17:53:41 I do my best to keep things in sync :-/ Mar 19 17:53:49 RP: opensync for example, horribly duplicated, and parts are missing in OE Mar 19 17:54:20 I'm not sure that was ever finished which is why it wasn't merged Mar 19 17:54:25 it's frustating to add package foo to OE and see it appearing in poky a few weeks later, but slightly different Mar 19 17:55:04 koen: I can understand that :-/ Mar 19 17:55:24 You have no idea the balancing act I have to do... Mar 19 17:56:08 I can't see why cwiiis can to OE->poky instead of doing stuff only in poky Mar 19 17:56:18 s/can/can't/ Mar 19 17:56:27 one word. monotone Mar 19 17:57:24 alias svn=monotone in .bashrc Mar 19 17:57:32 svn add foo, svn commit foo Mar 19 17:57:35 and it works! Mar 19 17:57:43 apart from the time delay Mar 19 17:58:26 on my slow cable link monotone is faster as svn, because it saves the network roundtrip Mar 19 17:58:35 (crappy upstream svn server) Mar 19 17:59:22 I can't see how it's 'faster' to create .bbs for poky yourself that are already in upstream OE Mar 19 17:59:34 but if that's the excuse used, I'll ask no more Mar 19 18:00:03 I nearly always get asked to merge packages from OE (or check if they exist) before anything gets added Mar 19 18:00:24 Except by Cwiiis, as you've noted Mar 19 18:01:13 I have a merge run to do, which is complicated by all the xlib changes Mar 19 18:02:10 NIH Mar 19 18:02:34 NIH? Mar 19 18:02:43 Not Invented Here Mar 19 18:03:54 IMO NIH is the reason the hh.org folks dislike OE Mar 19 18:04:26 Quite likely. Its not within their control Mar 19 18:05:23 I agree with the xlib changes, they just pose me a slight challenge with poky as if I screw up, it could have nasty consequences. Mar 19 18:10:45 On balance, I'd say poky has done OH more favours than it has harm which is the mina thing Mar 19 18:11:01 the next commit message will make me incredibly popular with some people Mar 19 18:12:48 03koen 07org.oe.dev * r61427ee8... 10/conf/distro/angstrom-2006.9.conf: Mar 19 18:12:50 angstrom-2006.9: specify extra machine specific installs Mar 19 18:12:52 * See? It can be done without polluting conf/machine/ Mar 19 18:14:15 Could you not use += instead of append? Mar 19 18:14:47 I've seen bitbake blowup with foo_something += Mar 19 18:14:51 but you have a point Mar 19 18:14:55 * koen adds a leading space Mar 19 18:15:13 If bitbake blows up can we fix bitbake then? Mar 19 18:15:26 There is no bug report about this... Mar 19 18:16:02 RP: you make koen file a wildcard bitbake bug Mar 19 18:16:46 RP: I vaguely remember being told that _append should be used Mar 19 18:17:03 koen: You're right. Just ignore me ;-) Mar 19 18:17:39 03koen 07org.oe.dev * r833150c6... 10/conf/distro/angstrom-2006.9.conf: angstrom-2006.9: _append needs a leading space Mar 19 18:19:59 zecke: I'm going to add sanity.bbclass to OE. We do need to think about providing a better mechanism to hook it into OE though (a hook that gets called once at the right time before doing any work with bb files) Mar 19 18:20:39 RP: the handlers can need some work Mar 19 18:20:55 zecke: Agreed Mar 19 18:25:24 btw gang, i'm targetting angstrom for new work on gcc4/amd64 buildhost Mar 19 18:25:34 thanks to koen and everyone for the EABI toolchain Mar 19 18:31:04 03rpurdie 07org.oe.dev * r00da6c55... 10/ (classes/sanity.bbclass conf/sanity.conf conf/bitbake.conf): Mar 19 18:31:06 Add sanity.conf/sanity.bbclass. This checks for common user misconfigurations Mar 19 18:31:08 and gives helpful information about fixing them. The checks can be expanded over Mar 19 18:31:10 time. It can be disabled by following the slightly cryptic instructions in Mar 19 18:31:12 sanity.conf. Mar 19 18:43:12 zecke: http://pastebin.com/611043 Mar 19 18:43:34 It seems bitbake is trying to tell me something, but I don't know what Mar 19 18:48:50 hah Mar 19 18:49:14 * RP -> food Mar 19 18:53:26 later all Mar 19 18:57:19 koen|away: that is encrypted use hexdump to see the message ;) Mar 19 19:00:49 RP: when you're back from food, what's the reason C3000 has the init-reboot=poweroff, init-poweroff=freeze mapping/ Mar 19 19:00:51 ? Mar 19 19:01:30 i think nabyss confirmed that for C3100 too Mar 19 19:03:14 i.e. trying to reboot means the device powers off, trying to power off the device gets into a stuck state Mar 19 19:06:02 hum.... angstrom bootstrap-image doesn't really go anywhere on spitz Mar 19 19:07:26 i get up to "Freeing init memory: 80K" and then nothing happens Mar 19 19:13:00 gosh, my zaurus charger for the c3100, it makes this irritating high-pitched noise Mar 19 20:03:27 jnc: The reason is that we use the standard pxa reboot handler which does a watchdog reset. The bootloader doesn't always seem to get on with that Mar 19 20:04:01 jnc: I suspect it hangs as we leave some register corrupt. SOmetimes it does power one again, sometimes it doesn't Mar 19 20:04:37 ah okay Mar 19 20:04:41 thanks for that info Mar 19 20:05:06 after the rootfs is mounted,... hmm Mar 19 20:05:08 is my init corrupt? Mar 19 20:05:10 * jnc pokes Mar 19 20:05:33 i don' Mar 19 20:05:45 i mean, i don't see where it's trying to run init here, on the angstrom build Mar 19 20:08:11 RP: on my C3000, that behavior has been pretty consistent. i don't know what you mean by sometimes it does power on again Mar 19 20:08:16 i've never seen that happen ;) Mar 19 20:09:32 jnc: Is this a reboot on the commandline or via sysrq? Mar 19 20:10:32 commandline Mar 19 20:10:39 haven't ventured the sysrq stuff Mar 19 20:10:47 I usually use sysrq as its quicker :) Mar 19 20:11:03 :) Mar 19 20:11:19 jnc: I've been looking at the Sharp code and I have a theory about your suspend/resume problem. There is some assembler code that's poking arount the memory refresh register buried in the sharp code Mar 19 20:11:47 that makes sense AFAIK Mar 19 20:21:48 RP: you know what's wild? with a broken init, suspend/resume seems to work Mar 19 20:21:59 i think, not 100% sure yet Mar 19 20:22:23 i'm not able to make it work in this configuration, so i don't have anything to test against Mar 19 20:22:30 It shouldn't make any difference... Mar 19 20:22:36 right Mar 19 20:23:22 is "require sanity.conf" in bitbake.conf supposed to be "include sanity.conf" ? Mar 19 20:23:49 it's like include, but it fails if sanity.conf is not able to be included; afaik Mar 19 20:24:13 well it's failing for me, yet I have all the sanity files Mar 19 20:24:20 jacques: No, its not Mar 19 20:24:45 I get: ERROR: Unable to parse conf/bitbake.conf (conf/bitbake.conf:397: unparsed line: 'conf/sanity.conf') Mar 19 20:24:46 jacques: Does your bitbake setup work? Mar 19 20:25:00 jacques: Is this an old version of bitbake? Mar 19 20:25:05 sigh Mar 19 20:25:08 here we go Mar 19 20:25:20 when you guys say "old" you mean "older than yesterday" Mar 19 20:25:36 old as in last version Mar 19 20:25:38 :) Mar 19 20:25:42 rwhitby just updated bitbake in the nslu2 stuff like yesterday Mar 19 20:26:13 jacques: no, I mean older than when the "require" directive was added to bitbake which was "quite a while ago" but I can't remember exactly when Mar 19 20:26:29 the topic could probably use a date stamp to indicate when bitbake was last updated ... Mar 19 20:27:07 sanity is overrated - commenting out the "require" line seemd to work Mar 19 20:27:19 lol Mar 19 20:27:36 jacques: A bitbake from yesterday would know about the require keyword. Your error says your bitbake doesn't. Do you have multiple bitbake's installed? Mar 19 20:28:14 RP has a very excellent point Mar 19 20:28:21 unfortunately bitbake doesn'y give a version number when it runs (or I don't know how to get it to do that) Mar 19 20:28:22 are you using bitbake from an SVN checkout? Mar 19 20:28:44 '--version' works Mar 19 20:28:52 how can I find out which version of bitbake is being run? Mar 19 20:28:53 that's pretty much standard for commandline tools Mar 19 20:28:57 hi jacques Mar 19 20:29:01 jnc@baker:~/opensource/openembedded/org.openembedded.dev/packages/linux$ bitbake --version Mar 19 20:29:04 I don't have bitbake installed - the nslu2 stuff is all self-contained Mar 19 20:29:04 BitBake Build Tool Core version 1.3.3.4, bitbake version 1.3.3.2 Mar 19 20:29:41 meaning each of my nslu2 mtn dirs has its own itbake installed Mar 19 20:30:07 ok, so what does bitbake --version show? Mar 19 20:30:17 checking Mar 19 20:31:20 sigh Mar 19 20:31:34 this is a pain - all interlocking shell scripts to setup the env Mar 19 20:32:33 and now whenever I source the setup-env script my shell exits Mar 19 20:32:37 didn't used to do that Mar 19 20:33:54 simply put, I don't know how to run bitbake anymore the way they've setup the scripts Mar 19 20:34:03 I just type "make" in the proper dir Mar 19 20:34:23 jacques: Can you find the bitbake binary? Mar 19 20:34:57 "binary" in a rather loose sense that is :) Mar 19 20:35:05 I can find the bitbake dir in the mtn dir, yes Mar 19 20:35:20 but since python pat hetc isn't set correctly I can't just type ./bitbake -version Mar 19 20:35:49 bitbake:__version__ = "1.3.2" Mar 19 20:36:27 out of interest, why is bitbake.conf using "require" for sanity.conf? That file doesn't define any variables that are necessary for correct operation, does it? Mar 19 20:36:44 maybe rwhitby didn't get around to updating bitbake after all? Mar 19 20:36:53 is 1.3.2 too old? Mar 19 20:36:59 pb_: The idea was to give an error if your bitbake setup wasn't correct Mar 19 20:37:17 jacques: Its too old - much older than yesterday Mar 19 20:37:40 RP, OK thanks, I guess I was operating on incorrect information - sorry :-( Mar 19 20:38:37 rwhitby ok, I'm going to update bitbake in our monotone repo Mar 19 20:38:37 rwhitby BTW, that's done by "make import-bitbake ; make push-bitbake" Mar 19 20:38:37 RP: oh, right. "wasn't correct" in the sense of too old? Mar 19 20:38:51 that was yesterday - not sure what happened Mar 19 20:38:59 I guess we should lean on Coremaster Zecke to make bitbake export a variable with its actual version number to facilitate that kind of check. Mar 19 20:39:39 jacques: heh, what happens if you do "make import-bitbake" yourself? Mar 19 20:39:43 pb_: It already does and sanity.bbclass does check it Mar 19 20:40:03 RP: ah. then why is the "require"... er, required. Mar 19 20:40:08 pb_, let me try that Mar 19 20:40:19 pb_: "Wasn't correct" was to catch the case where it didn't find a sanity.conf file for whatever reason (broken paths) Mar 19 20:40:58 pb_, it tries to do: svn co svn://svn.berlios.de/bitbake/trunk/bitbake Mar 19 20:41:36 RP: ah. I guess that's laudable, but if it has the effect of changing a useful error into a cryptic one then I guess it's a net loss. Mar 19 20:41:43 jacques: that is indeed the location of bitbake Mar 19 20:43:11 pb_: Its one cyptic error to watch for rather than several more obscure ones so its not all bad... Mar 19 20:43:37 pb_, now I got the 1.3.3.2 version - strange Mar 19 20:44:53 * jacques ponders the wisdom of updating bitbake while a build is ongoing. Mar 19 20:45:05 ah well, what's the worst that could happen Mar 19 20:45:24 ruh-roh - someone heard his name being used in vain Mar 19 20:45:40 heh Mar 19 20:45:41 pb_, why not just make it check for a new version on intial cache build? Mar 19 20:45:50 emte: er? Mar 19 20:46:17 just making it a bit smarter than returning a version number Mar 19 20:46:29 howdy rwhitby - I thought you were asleep otherwise I would have discussed this bitbake issue with you first Mar 19 20:46:56 jacques: I was Mar 19 20:47:05 it was just coincidental that I logged in now Mar 19 20:47:18 * rwhitby reads logs Mar 19 20:47:46 you would only want to update it before the intial meta cache as bad things might happen if it were updated part way through building an image Mar 19 20:47:53 rwhitby, basically I had an issue which was because of old bitbake and I was saying you had just updated bitbake yesterday Mar 19 20:48:26 but the mtn checkout I just did had the old bitbake (1.3.2 vs. 1.3.3.2) Mar 19 20:48:44 it would be really good if bitbake was kept in the same monotone database as OE, then they would always be in sync. Mar 19 20:48:46 * emte isnt sure that helps make any more sense or not ... Mar 19 20:49:16 rwhitby, then what happens to the projects taht use bitbake and not OE? Mar 19 20:49:44 that topic has been discussed on the mailing list as well Mar 19 20:50:03 are there actually any projects in that category? Mar 19 20:50:11 emte: not replace the SVN, but someone keeps it in sync in monotone (like I do in the org.nslu2-linux branch). that someone being someone who knows better than I when it needs to be updated in monotone from svn Mar 19 20:50:14 supposedly Mar 19 20:50:25 rwhitby: For reference, I pushed sanity.conf into OE which performs some basic sanity checks of a user's setup. It might ot might not like the NSLU's standard setup although it doesn't check anything too controvershal Mar 19 20:50:26 rwhitby: that would be a good idea Mar 19 20:50:31 you would have to ask kergoth for sure i'd think Mar 19 20:51:10 rwhitby: One thing sanity.bbclass does check is the bitbake version number which is at least something... Mar 19 20:51:33 RP: that's great - who keeps the version number up to what is required? Mar 19 20:51:54 why not jsut make it try to update from svn on initial cache Mar 19 20:52:06 if its already updated it will use very little bw Mar 19 20:52:10 (cause that person would be an ideal candidate to run "make import-bitbake update-bitbake" in the monotone dir :-) Mar 19 20:52:33 the only problem you would have is if someone "installed" bitbake Mar 19 20:52:50 then your goimg to have to sudo the update Mar 19 20:52:52 rwhitby: If anyone adds something to OE that depends on new bitbake features, the number gets bumped. Since few people acutally change bitbake (just myself and zecke atm), this should be maintainable for now Mar 19 20:53:48 RP: excellent. if you happen to remember to ping me (or anyone else on the nslu2-linux core team) in this channel when you do that, that would be icing on the cake. Mar 19 20:54:35 rwhitby: ok, I'll keep that in mind. Mar 19 20:54:50 emte: we have forbidden anyone in nslu2-linux from "insalling" bitbake. It just causes more trouble than what it's worth if it's installed. Mar 19 20:55:02 s/insalling/installing/ Mar 19 20:55:32 rwhitby, i agree Mar 19 20:55:37 but it is an option Mar 19 20:57:12 * RP -> back later Mar 19 20:57:46 rwhitby: sure Mar 19 21:09:34 03mickeyl 07org.oe.dev * r4a54c232... 10/packages/openssl/ (9 files in 4 dirs): Mar 19 21:09:38 openssl 0.9.7e make it work with gnueabi-arm Mar 19 21:09:40 drop older versions Mar 19 21:11:50 hm, is there nothing nicer than this "if test "x$D" != "x"; then" to ensure that pre/post inst/rm scripts are executed on the device? Mar 19 21:19:36 tmbinc: no Mar 19 21:20:51 s/test/[/ Mar 19 21:21:08 heh, that isn't nicer Mar 19 21:21:30 ;) Mar 19 21:21:41 yes because you need to add the ] Mar 19 21:41:53 hmm, my build failed on gcc-cross Mar 19 21:41:59 haven't seen that lately Mar 19 21:43:23 gosh that parser is voodoo Mar 19 21:45:38 hmm Mar 19 21:45:39 NOTE: preferred version 2.4 of glibc not available Mar 19 21:45:50 which is pretty odd, because one bitbake run before it just built it Mar 19 21:46:03 mickey|dinner: stupid bitbake Mar 19 21:46:13 mickey@allanon:/local/pkg/oe/c7x0$ ls -l tmp/work/armv5te-linux-gnueabi/ |grep glibc Mar 19 21:46:13 drwxr-xr-x 10 mickey mickey 528 Mär 18 20:53 glibc-2.4-r2 Mar 19 21:46:13 drwxrwxr-x 9 mickey mickey 480 Mär 19 03:35 glibc-2.4-r3 Mar 19 21:46:43 mickey|dinner: could you find out, if it is in the already failed list? Mar 19 21:46:50 where is that list? Mar 19 21:47:13 mickey|dinner: bin/bitbake Mar 19 21:47:30 zecke: the c parser? Mar 19 21:47:47 pb_: no the parse_py, in the lack of a conformance test suite Mar 19 21:47:57 ah Mar 19 21:47:59 pb_: I need to make it compatible to the existing one :} Mar 19 21:48:18 heh Mar 19 21:48:20 pb_: did you write it? Mar 19 21:48:27 no, that was kergoth Mar 19 21:48:47 pb_: too bad, that would have been an explanation where your sould ended up Mar 19 21:49:09 yeah, that is too bad Mar 19 21:49:56 I gave up on writing that marketing document after about 1200 words. I guess I will try again in the morning. Mar 19 21:50:23 maybe then I will be gripped by the marketing fever Mar 19 21:51:30 that sounds dangerous Mar 19 21:52:31 I dunno, it sounds kind of fun in a way Mar 19 22:01:14 <[g2]> pb_ what are you pitching ? Mar 19 22:01:42 just some internet service Mar 19 22:01:52 er, internet services thing Mar 19 22:02:31 all rather tedious Mar 19 22:03:25 <[g2]> so how does one setup a local fetch repo ? Mar 19 22:04:25 <[g2]> and 2) is there some way to extract a copy of just the used meta data files from the OE tree Mar 19 22:04:34 zecke: hmm, ok, it was the sanity.conf i just downloaded that broke the build Mar 19 22:04:47 zecke: anyway, the error message still persists, but the build goes on Mar 19 22:05:00 need to investigate eventually Mar 19 22:05:10 mickey|dinner: that would be nice Mar 19 22:05:19 mickey|dinner: did you pass the sanity test? Mar 19 22:05:29 no, i had ASSUME_PROVIDED in my site.conf :D Mar 19 22:05:31 =, even Mar 19 22:05:36 or should I do mickeyl.setVarFlag('MIND', 'insane', True) Mar 19 22:05:45 heh Mar 19 22:16:21 hmm Mar 19 22:16:23 n1 = *(((unsigned int *) src1)++); Mar 19 22:16:31 error: invalid lvalue Mar 19 22:16:41 how to rephrase that for gcc4 ? Mar 19 22:16:58 error: invalid lvalue in increment Mar 19 22:16:59 actually Mar 19 22:17:14 mickey|dinner: tmp var Mar 19 22:17:51 mickey|dinner: or cast src1 to unsigned int* in advance Mar 19 22:30:10 03mickeyl 07org.oe.dev * r171521f7... 10/packages/libmodplug/libmodplug_0.7.bb: libmodplug: use oe_libinstall, not cp Mar 19 22:30:14 03mickeyl 07org.oe.dev * rcaf6f396... 10/packages/libxine/ (4 files in 3 dirs): libxine-fb 1.0: apply gcc4 patch Mar 19 22:45:12 hmm Mar 19 22:45:30 a NoOP parser takes 51 seconds Mar 19 22:58:31 vs. 13 minutes Mar 19 23:00:10 zecke: regarding bug 558, the easiest way to identify the offending code is to set /proc/sys/alignment to raise SIGBUS, run opie-opieplayer under the debugger, and look to see where it crashes. Mar 19 23:00:51 pb_: did it move to sys/? Mar 19 23:01:06 pb_: you assume that gdb, c++ and arm are playing together Mar 19 23:01:24 er, no, I just typed it wrong Mar 19 23:01:27 it is in cpu Mar 19 23:01:50 zecke: is that a dangerous assumption to make? Mar 19 23:01:55 pb_: I think I need to try myself, and revive my remote gdb setup Mar 19 23:02:12 obviously c++ equals danger, but still Mar 19 23:02:13 pb_: most of the time, it is not working :) Mar 19 23:02:18 ah Mar 19 23:02:21 that is unfortunate Mar 19 23:02:29 pb_: most of the time not even catchsegv can catch a segfault Mar 19 23:03:03 zecke: catchsegv usually works less reliably than gdb for me, even on plain C code Mar 19 23:03:04 pb_: we C++ programmers are real maam's, we want tools! Mar 19 23:04:15 so it seems Mar 19 23:04:48 pb_: I had hoped someone else would fix that alignment trap :} Mar 19 23:05:14 heh Mar 19 23:07:50 pb_: did you write buildDegraph? Mar 19 23:08:06 possibly Mar 19 23:08:10 I don't remember for sure Mar 19 23:08:44 1 10.089 10.089 10.127 10.127 Mar 19 23:11:44 http://oe.pastebin.com/611491 Mar 19 23:13:18 yeah, it sucks Mar 19 23:23:23 * pb_ go to sleep now Mar 19 23:23:24 night night Mar 19 23:23:32 n8 Mar 19 23:23:42 good thing, every file is only parsed once Mar 19 23:24:01 'night pb_ Mar 19 23:30:06 Hmm. Boot opie with a CF card inserted after a fresh install without the screen calibrated. Game Over :-/ Mar 19 23:31:18 RP: until next reboot? Mar 19 23:32:35 zecke: Until you reboot with the card removed, yes Mar 19 23:41:04 Sound is also tempramental on c7x0 for 3.5.4.1. Lets hope the next ASoC updates fix that :) Mar 19 23:41:53 how do I start building for 3.5.4.1? do I have to switch from .oz354fam083? Mar 19 23:42:14 Zero_Chaos: Switch to the oz354x branch Mar 19 23:42:51 RP: I don't have one of those right now.... is it just .oz354x? Mar 19 23:43:22 Zero_Chaos: It is. Make sure monotone is set to pull that branch (monotone list vars should give a clue) Mar 19 23:44:03 RP: I think I can do it, thanks. Mar 20 00:25:11 good nite Mar 20 00:25:27 RP: I have the C++ code now in place Mar 20 00:25:39 just no way to call it from python and no way back into python Mar 20 00:25:45 I will work on that tomrrow Mar 20 00:26:07 RP: I'm still wondering how I would call a C++ data implementation directly Mar 20 00:26:31 and I can only think of storing a 'pointer' as an attribute Mar 20 00:29:57 zecke: You mean call the C++ from the C++ ? Mar 20 00:31:54 zecke: I'll be interested to see your code. Perhaps I can find some time to help with it... Mar 20 00:33:35 anyhow, bed for me too. 'night all Mar 20 00:43:53 RP: yes Mar 20 00:44:08 RP: I have a PyObject that I suspect to be a C++ implementation Mar 20 00:44:27 RP: I want to avoid creating a parameter list and invoke python calls.. Mar 20 00:45:13 I move to bed now as well Mar 20 02:17:05 n1 = *( ++( unsigned int *) src1) ); Mar 20 02:17:09 that isn't valid IMO Mar 20 02:17:12 never was. Mar 20 02:17:32 oh wait i missed a paren Mar 20 02:17:56 n1 = *( ( (unsigned int *) src1 ) ++ ); Mar 20 02:19:12 that should probably be n1 = *(( ++( (unsigned int*)( src1 ) ) )); Mar 20 02:19:59 which is too clever. shoot that programmer. Mar 20 02:26:34 i found a bug in the Angstrom... i think Mar 20 02:57:06 actually it's probably more likely an bug with linux-openzaurus, and the way it handles oz hardware Mar 20 06:29:32 hmm, doesn't linux support a1200? Mar 20 06:30:12 hmm, the tf card and the pda can't be accessed via gentoo here... Mar 20 06:30:17 anyone here would give a suggestions? Mar 20 07:09:28 what is a tf card? :o Mar 20 07:09:53 Uh.. tinyflash? Mar 20 07:15:05 morning all Mar 20 07:23:26 hi RP Mar 20 07:23:32 * Bernardo is off to work Mar 20 07:42:22 hi all Mar 20 07:46:40 good morning Mar 20 07:52:39 d0tslash: tf is a card which is a bit like SD, but it's much smaller... Mar 20 07:53:08 by the way, where to get the source for motorola a1200? Mar 20 07:56:08 hmm, Is there any info about developing on smartphone for motorola? Mar 20 07:59:49 Yo ! Mar 20 08:00:55 03hrw 07org.oe.oz354x * rae75d4d1... 10/packages/rsync/ (rsync_2.6.5.bb rsync_2.6.6.bb rsync_2.6.7.bb): rsync: update to 2.6.7 Mar 20 08:01:00 03hrw 07org.oe.dev * r52ebc371... 10/packages/rsync/ (rsync_2.6.5.bb rsync_2.6.6.bb rsync_2.6.7.bb): rsync: update to 2.6.7 Mar 20 08:16:16 BTW, tf stands for transflash. Mar 20 08:20:25 morning Mar 20 08:20:34 morning hrw|work Mar 20 08:21:05 NOTE: multiple providers are available (glibc, glibc-intermediate); Mar 20 08:21:06 NOTE: consider defining PREFERRED_PROVIDER_virtual/arm-linux-libc-for-gcc Mar 20 08:22:01 hrw|work: I can provide some 3.5.4.1 images for c7x0. The opie image has some sound problems I'm hoping the new ASoC code will fix and it gets confused if you have a unrecognised CF card inserted before you calibrate the screen. Apart from that its looking good though Mar 20 08:22:21 hi, anybody can let me know how to deal with multiple providers? Mar 20 08:22:25 RP: good to know Mar 20 08:22:31 minipanda: this one you can ignore Mar 20 08:22:34 minipanda: Just ignore that Mar 20 08:22:47 RP: that info about card is longstanding opie bug Mar 20 08:23:00 hrw|work: I thought it might be :-/ Mar 20 08:23:51 hrw|work: RP ok .. but i need a kernel image for my OMAP OSK5912 board Mar 20 08:24:26 so maybe i need 'bitbake linux-omap1' works and generate an deplyable image somewhere Mar 20 08:24:36 minipanda: Those notes are not stopping you though and you can safely ignore them Mar 20 08:25:54 * RP -> back shortly Mar 20 08:29:25 2.6.16 released Mar 20 08:31:45 moin Mar 20 08:31:57 hi zecke Mar 20 08:33:56 hrw|work: I won't get to 30 seconds parsing Mar 20 08:34:10 hrw|work: a NoOP parser takes around 50 seconds on this ibook Mar 20 08:34:20 zecke: you got <1minute which is fscking awesome Mar 20 08:35:05 hrw|work: 10 seconds from this 50 where spend in buildDepgraph Mar 20 08:35:41 BB>> build linux-omap1-2.6.12-rc2 Mar 20 08:35:42 SHELL: Building linux-omap1-2.6.12-rc2 Mar 20 08:35:42 BB>> Mar 20 08:43:48 where ca I find oe_libinstall syntax? Mar 20 08:46:20 mndctrl: base.bbclass :} Mar 20 08:47:05 ahh ;) tnx... Mar 20 08:51:53 morning Mar 20 08:56:19 hi XorA Mar 20 09:01:51 2.6.16. I guess I need to push some patches now... :) Mar 20 09:02:13 you need Mar 20 09:03:14 RP: hi Mar 20 09:03:58 2.9mb changelog... :} Mar 20 09:04:27 2.9mb!? Mar 20 09:06:44 hi zecke Mar 20 09:07:13 I wonder if the LED subsystem makes it in... Mar 20 09:09:40 RP: BBFILES := "ftp://ftp.handhelds.org/file/file.bb" is that desirable? Mar 20 09:10:04 RP: current parser could do that, I'm not going to implement it Mar 20 09:10:29 http://sam.rpsys.net/openzaurus/3.5.4.1-alpha1/ for anyone interested Mar 20 09:10:51 zecke: I think we have bigger things to worry about Mar 20 09:11:08 RP: well the current code can do that :} Mar 20 09:11:21 zecke: I can forgive you for breaking that for a speed improvement ;-) Mar 20 09:11:28 zecke: first get cparser working then add missing things Mar 20 09:17:56 hrw|work: pure voodoo Mar 20 09:28:35 hi OE-users ! Mar 20 09:28:48 hi ;) Mar 20 09:29:24 I still try to build opie-image and it fails with following deps: Mar 20 09:29:27 | ERROR: Cannot satisfy the following dependencies for task-opie-base: Mar 20 09:29:28 | ttf-dejavu-sans ttf-dejavu-sans-mono Mar 20 09:29:46 grma: ttf-dejavu was not built? Mar 20 09:30:06 i will try Mar 20 09:30:14 looks like Mar 20 09:31:25 a other tings: i have problems with Tremor, but i come back with that later, lets try ttf-dejavu Mar 20 09:31:33 thanks Mar 20 09:31:41 hmm.. starting from DejaVu 2.4 I will have to replace some rules it it Mar 20 09:32:21 grma: which branch? which version of bitbake? Mar 20 09:32:27 I need some help creating a .bb file for my shard library. Please check out: http://hanafjedle.net/modules.php?name=Forums&file=viewtopic&p=37#37 Mar 20 09:32:36 anon posting should be allowed... Mar 20 09:32:47 mndctrl: forums? Mar 20 09:33:10 yeah, I've posted my current .bb file etc... Mar 20 09:33:32 I will not put data into forums :} Mar 20 09:33:41 bah! Mar 20 09:33:42 ;) Mar 20 09:33:59 mndctrl: lack of 'inherit autotools'? Mar 20 09:34:21 ok? hmm... Mar 20 09:34:29 mndctrl: are you sure the library is in the same dir? Mar 20 09:35:07 mndctrl: take a look at the work directory of your library Mar 20 09:35:21 mndctrl: I suspect you need to set S Mar 20 09:35:48 actually its not... the libgsmc-1.0 dir bitbake creates and cd's into, contains nothing... Mar 20 09:36:05 I've tried to put stuff in there, but no luck... Mar 20 09:36:21 zecke: ok, I'll try a S, but my .tar file contains no dirs Mar 20 09:36:29 03pH5 07org.oe.dev * r0cd761de... 10/packages/linux-libc-headers/ (2 files in 2 dirs): linux-libc-headers-2.6.15.99: remove #include from linux/rtc.h Mar 20 09:36:36 mndctrl: then set S = "${WORKDIR}" Mar 20 09:36:43 zecke: ok... Mar 20 09:36:43 mndctrl: I was not able to unzip your data Mar 20 09:37:03 but what do I need in the .tar package? only .c files? Mar 20 09:37:34 mndctrl: the source, the build files (Makefile) Mar 20 09:37:48 zecke: no, the URI dir is broken, I've only put it in dev.openembedded.org/packages Mar 20 09:38:55 but what do I need in the .tar package? only .c files?? Mar 20 09:39:29 mndctrl: there is no right answer to your question Mar 20 09:39:45 okthen ;) Mar 20 09:39:59 mndctrl: it could be enought to have .c files and a Makefile Mar 20 09:40:24 mndctrl: OE is not a make/autotools/cmake/any* replacement Mar 20 09:40:49 ok... Mar 20 09:41:22 it still fails, but I've got some stuff to work on ;) tnx guys... Mar 20 09:42:00 mndctrl: do you have a Makefile at all? Mar 20 09:42:28 nope ;) thats what I needed to work on... lol Mar 20 09:46:29 mndctrl: a small hint. do not use 'gcc', 'g++' Mar 20 09:46:40 mndctrl: use $(CC) $(CXX) in the Makefile Mar 20 09:47:43 ok, tnx ;9 Mar 20 09:47:44 ;) Mar 20 09:51:43 on bitbake opie-mediaplayer2 it has problems when checking out Tremor from svn (http://pastebin.com/612074), i have seen this on other svn servers because of svn-version problems between server version 1.0 and my client 1.2, can someone tell me a way how to solve this without needing svn, or downloading the bz2 file somewhere else ? Mar 20 09:52:51 grma: manual checkout works? Mar 20 09:54:52 RP, i downloaded last qemu snapshot , how to be sure that your patches are be added ? ( which files ? ) Mar 20 09:55:31 zecke: will try Mar 20 09:55:45 zecke: but i don't think so Mar 20 09:56:13 Genesis: Check hw/pl110.c exists Mar 20 09:58:14 oki Mar 20 09:58:34 it's right Mar 20 09:58:53 i compile snapshot from http://www.dad-answers.com/qemu/ Mar 20 09:59:51 Personally, I'd just check it out from CVS or better still let OE do it since we're on this channel but ... Mar 20 10:00:25 i don't know how to do Mar 20 10:00:58 OE can compile qemu automaticaly ?! Mar 20 10:01:49 or the community build it and post it somewhere ? Mar 20 10:02:12 OE can compile qemu and also build images that run under it Mar 20 10:02:21 zecke: Tremor svn co http://svn.xiph.org/trunk/Tremor --> Bad request Mar 20 10:05:17 zecke: I have seen this on other svn servers too, with some svn servers I have no problem with others I get always Bad request .. :-( Mar 20 10:06:08 I dowloaded Tremor now with wget from http://svn.xiph.org/trunk/Tremor, but how can i tell OE that it is there ? Mar 20 10:11:03 grma: are you firewalled? Mar 20 10:15:17 zecke: yup, but with other svn servers i have no problem ? Mar 20 10:15:22 good morning Mar 20 10:18:36 grma: do they use http, https or svn protocol? Mar 20 10:18:40 good morning all Mar 20 10:19:08 grma: works here Mar 20 10:19:15 chris144: regarding your 'bitbake' bug Mar 20 10:19:24 chris144: I fear this is not a bug Mar 20 10:19:30 zecke: ok explain Mar 20 10:19:36 chris144: it should warn though Mar 20 10:19:54 chris144: I think it won't be possible to mix functions with python,shell and lua Mar 20 10:20:13 chris144: the _append and _prepend are just textual Mar 20 10:20:27 chris144: and the 'flag' is set to be a function of kind.. Mar 20 10:20:38 zecke: i do http, what dose you svn --version say, my svn cleint version is 1.2 .. Mar 20 10:21:04 zecke: well, the append ist not 'really' essential; prepend works nicely. So i thought... Mar 20 10:22:26 zecke: we should document it somehow. Where should it go, in the wiki? Mar 20 10:25:02 Genesis: http://www.rpsys.net/openzaurus/qemu/ Mar 20 10:25:40 and anyone else interested in running qemu images :) Mar 20 10:28:16 :o) Mar 20 10:28:17 thx Mar 20 10:30:06 re Mar 20 10:30:28 grma: I have version 1.2.3 r15833 Mar 20 10:30:43 Genesis: It runs a gpe-image from .dev as of yesterday but only gets as far as touchscreen calibration which is stops dead Mar 20 10:30:44 chris144: bitbake manual Mar 20 10:30:44 RP, : why not write it in the OE wiki ? Mar 20 10:31:00 chris144: I will have to look in the function handling Mar 20 10:31:12 chris144: but it smells like this: On prepend the 'function' flag gets set Mar 20 10:31:28 chris144: but execution will fail anyway because you mix python and shell? Mar 20 10:31:35 03hrw 07org.oe.oz354x * r711ea366... 10/packages/meta/bootstrap-image.bb: bootstrap-images: fix feed urls for OpenZaurus Mar 20 10:31:51 for information , my target is a i.mx21 Mar 20 10:32:20 Genesis: It will probably end up there. This was just something quick and simple I could point people at - it also applies to something called poky which is in fact where these changes came from Mar 20 10:32:36 yeah Mar 20 10:32:39 i bookmark that Mar 20 10:32:49 http://www.rpsys.net/openzaurus/temp/qemuarm-poky1.png Mar 20 10:33:01 last week Mar 20 10:33:02 zecke: thanks than i have a firewall problem :-( Mar 20 10:38:47 Genesis: For qemu purposes you're stuck with an ARM Integrator/CP Mar 20 10:41:14 zecke: sorry, I was on the phone Mar 20 10:42:38 RP, : i'm a noob in embedded linux , i'll get what is integrator/CP , i saw that in qemu examples Mar 20 10:43:10 for the moment i try to set up my first OE environment Mar 20 10:43:22 ( perharps i should begin by that ) Mar 20 10:43:32 03pH5 07org.oe.dev * r599ce309... 10/packages/linux-libc-headers/ (2 files in 2 dirs): linux-libc-headers-2.6.15.99: fix linux/videodev2.h Mar 20 10:44:01 zecke: http://pastebin.ca/46324 is the file I was working on (it wors as it is now) Mar 20 10:45:05 zecke: if I replace python tarball_do_unpack with some do_patch_prepend stuff it works ! Mar 20 10:46:09 chris144: I lack the time now :} Mar 20 10:46:59 chris144: why tarball_do_unpack? Mar 20 10:47:05 zecke: no problem, we can take it up sometime, it is in bugzialla Mar 20 10:47:30 chris144: do_unpack is a Shell Function already Mar 20 10:47:49 chris144: I need to check/consider if the flag should be changed Mar 20 10:48:30 zecke: the .bb I was working on unpacks in two steps, 1st rpm2cpio, 2nd the resulting tarball with oe_unpack_file Mar 20 10:48:49 zecke: but as I said, no hurry... Mar 20 10:49:14 chris144: The only we might see is bitbake not updating the function flag (from shell->python) Mar 20 10:49:29 chris144: I need to findout if that is wanted :} Mar 20 10:49:35 but now I need to hack on RMI Mar 20 10:49:50 zecke: ok no prob, see ya Mar 20 10:49:54 :-) Mar 20 10:50:17 zecke: good morning Mar 20 10:50:18 mickey|zzZZzz: could you add dirks key to the database? I'll be without reliable net till wednesday Mar 20 10:50:54 koen|work: do you know bootstrap-image to be functional under angstrom-2006.9? Mar 20 10:51:05 * chris144 looks to make a nice gfxboot screen for his embedded PC104 board now Mar 20 10:51:15 jnc: ph5 has booted it on his hx4700 Mar 20 10:51:31 koen|work: okay. am having trouble with that on spitz Mar 20 10:51:38 init never seems to be called Mar 20 10:53:41 what does it mean if the kernel mounts a rootfs, and then frees 80K of init memory Mar 20 10:54:41 jnc: did you build your kernel with eabi support? Mar 20 10:54:51 good question Mar 20 10:54:55 right now only the ep93xx kernel has teh eabi bits in OE Mar 20 10:54:59 ah Mar 20 10:55:00 then no Mar 20 10:55:09 i built with linux-openzaurus kernel Mar 20 10:55:18 is it more code, or a config option? Mar 20 10:55:32 * rwhitby notices 2.6.16 is released ... and starts an nslu2-linux build of it ... Mar 20 10:55:32 both Mar 20 10:55:37 zoh Mar 20 10:55:44 jnc: 2.6.16 will have it Mar 20 10:55:54 okay :) Mar 20 10:56:00 lunchtime Mar 20 10:56:10 linux-openzaurus-2.6.16 might land this evening if we're lucky Mar 20 10:56:38 RP: did your leds system get in? Mar 20 10:56:43 RP: cool, just in time for me to flash a Angstrom image on c7x0 tomorrow Mar 20 10:57:21 rwhitby: Its in -mm so it might make 2.6.17 if we're lucky Mar 20 10:57:50 excellent - looking forward to it and the RTC system becoming official so we can have it in our nslu2 debian kernel Mar 20 10:58:03 * RP also hopes for the rtc subsystem and ASoC Mar 20 10:58:15 ASoC ? Mar 20 10:58:24 ah, dwery Mar 20 10:58:31 no, that's AZ. Mar 20 10:59:10 ASoC is an ALSA subsystem for system on chip style sound support Mar 20 10:59:35 We're using it on the Zaurii as its ideally suited **** ENDING LOGGING AT Mon Mar 20 10:59:56 2006