**** BEGIN LOGGING AT Sat Dec 02 02:59:57 2006 Dec 02 03:29:04 hi all Dec 02 06:02:51 gee mtn takes a while to update:( Dec 02 06:43:15 morning koen :) Dec 02 07:25:39 morning all Dec 02 07:27:31 moriing Ifaistos Dec 02 07:50:53 hvontres|poodle : Hi ! seems people are still sleeping :) Dec 02 08:43:02 morning all Dec 02 08:44:57 morning RP, morning all Dec 02 08:46:14 hi cyrilRomain Dec 02 08:48:53 RP: did you get my mail from 2006.11.12 ? I sent you patches for linux-openzaurus 2.6.17 and 2.6.18 (just an update to make patch apply, similarly to the one you did on 2.6.18+git) ? Dec 02 08:49:27 cyrilRomain: I did although I'm afraid its got lost in my inbox. I realised that yesterday and will put that right today, sorry Dec 02 08:49:41 RP: I saw linux-openzaurus was renamed to linux-rp, so maybe I can send you an updated version Dec 02 08:50:08 cyrilRomain: Don't worry, I can deal with that :) Dec 02 08:50:17 RP: thanks :) Dec 02 08:50:24 good morning all Dec 02 08:50:29 hey koen|gprs Dec 02 08:51:41 morning koen|gprs Dec 02 08:53:00 koen|gprs: Do I have an account edit the OE website? Dec 02 08:53:16 RP: you should have Dec 02 08:53:35 * koen|gprs checks Dec 02 08:54:05 koen|gprs: I noticed only one of the email lists is listed Dec 02 08:54:58 your user name is "Richard Purdie" Dec 02 09:10:30 koen|gprs: thanks Dec 02 09:19:32 The contacts is updated. We should add the NSLU2 people there too. We need a webmaster :-/ Dec 02 09:24:18 morning all. Dec 02 09:27:56 hi Ifaistos Dec 02 09:42:53 03pfalcon 07org.oe.dev * rfeb1b692... 10/ (1 packages/linux/linux-handhelds-2.6_2.6.16-hh5.bb): Dec 02 09:42:53 linux-handhelds-2.6 2.6.16-hh5: Remove obsolete version. Dec 02 09:42:53 * Please use 2.6.16-hh7 instead (-hh8 switched to common buttonmap). Dec 02 09:45:43 hi fooks Dec 02 10:01:12 giel: ping Dec 02 10:16:51 Anybody willing to push bug 1624? Dec 02 10:37:40 likewise: yeah? Dec 02 10:53:51 hey likewise Dec 02 10:54:47 Laibsch: Can you reproduce 1624? If so, can you work out why -L/usr/lib appears in the commandline? Dec 02 10:55:55 RP: Yes, as I mentioned in comment #2, I can confirm the issue. Dec 02 10:56:28 Laibsch: I've replied to the bug - we need to find why -L/usr/lib is there Dec 02 10:58:33 giel: ping Dec 02 11:01:17 RP, Laibsch: maybe libpng_1.2.12 should define the do_compile the same way libpng_1.2.8 does, instead of using autotools ... ? Dec 02 11:02:11 with ZLIBLIB=${STAGING_LIBDIR} Dec 02 11:02:30 cyrilRomain: Or at least have the OEMAKE_append line... Dec 02 11:03:08 1.2.12 doesn't stage correctly for me either. I suspect I have some local changes interfering though :-/ Dec 02 11:05:07 Like STAGING_LIBDIR missing from my bitbake.conf :} Dec 02 11:06:06 hm maybe solving 1624 would even solve 1344 Dec 02 11:07:53 cyrilRomain: Does prboom build against libpng 1.2.8? Dec 02 11:08:35 RP: against zlib Dec 02 11:11:11 cyrilRomain: It builds against libpng which needs zlib Dec 02 11:11:23 RP: it seems that recent versions of prboom has been removed, so maybe should the 1344 mark as WONTFIX, or one should bump version of prboom Dec 02 11:11:48 RP: right Dec 02 11:12:52 Why were recent versions removed? Dec 02 11:13:50 RP: dunno. Dec 02 11:21:18 cyrilRomain|bbl: I find any evidence that verison of prboom was ever in OE Dec 02 11:21:24 I can't find Dec 02 11:22:48 koen|tdose____: Is prboom 2.4.3 some unpushed changeset on ewi? Dec 02 11:24:14 RP: could you please push this small fix in OE http://pastebin.ca/264412 ? This solves opie* stuff compilation with sharprom-compatible. I have posted on OE dev mailing list, but atm no-one pushed it in OE. Dec 02 11:24:22 RP: prboom? Dec 02 11:24:25 RP: no idea Dec 02 11:24:34 hi koen Dec 02 11:24:42 RP: last time I tried to compile doom I used pb's version Dec 02 11:24:45 hey nicolasfr Dec 02 11:25:05 ~2 years ago Dec 02 11:25:34 koen|tdose: I'm just wondering about #1344 Dec 02 11:25:35 koen: about fetchmail bug, you wrote this comment for the bug: sharprom-compatible.conf:FULL_OPTIMIZATION = "-fexpensive-optimizations Dec 02 11:25:36 -fomit-frame-pointer -O2". Dec 02 11:26:04 yes, that's what's in the config file now Dec 02 11:26:13 question is: how to use "-fexpensive-optimizations -fomit-frame-pointer -O1" only for fetchmail? Dec 02 11:26:37 you probable want FULL_OPTIMIZATION_sharprom-compatible = "-fexpensive-optimizations -fomit-frame-pointer -O1" in fetchmail Dec 02 11:26:45 probably* Dec 02 11:27:06 ok! Dec 02 11:27:29 What version of make are you guys using ? it seems there are some issues with 3.81 (u-boot_1.1.4 has one of them) Dec 02 11:29:21 03rpurdie 07org.oe.dev * r1931d326... 10/ (1 classes/palmtop.bbclass): palmtop.bbclass: Fix sharprom override (from nicolasfs #oe) Dec 02 11:29:24 another question: when building stuff for sharprom-compatible I absolutely do no want libqpe and libqte to be built/packaged even if other recipes depend on them. Is it possible to do that with a directive in sharprom-compatible.conf ? Dec 02 11:29:29 RP: thanks a lot Dec 02 11:29:54 nicolasfr: ASSUME_PROVIDED += "qte packages" Dec 02 11:30:56 RP: I have tried that but with no luck. Maybe I made a typo... Dec 02 11:31:24 nicolasfr: Note the +=. Make sure you don't use ASSUME_PROVIDED = anywhere Dec 02 11:31:56 RP: ok Dec 02 11:36:16 re Dec 02 11:39:46 http://pastebin.ca/264423 fixes bug 1625 (fetchmail). Anyone can merge? Dec 02 11:41:44 hi Jin^eLD Dec 02 11:42:12 hey Dec 02 11:42:58 nicolasfr: Can you test adding FULL_OPTIMIZATION_pn-fetchmail = "-fexpensive-optimizations -fomit-frame-pointer -O1" to sharprom-compatible.conf please? Dec 02 11:43:51 03nicolasfr 07org.oe.dev * rbf07ce72... 10/ (1 packages/fetchmail/fetchmail_6.2.5.bb): fetchmail: fix optimizations flags for sharprom compatible distro Dec 02 11:49:50 RP: strange because I usually make sure I don't file bug about locally modified or added package. Moreover I don't use prboom. Dec 02 11:50:21 RP: "FULL_OPTIMIZATION_pn-fetchmail = "-fexpensive-optimizations -fomit-frame-pointer -O1" to sharprom-compatible.conf" works fine Dec 02 11:50:22 Nevermind, when 1624 will be corrected, I'll either WONTFIX 1344 or try to bump prboom to 2.4.7 Dec 02 11:54:10 s/WONTFIX/INVALID/ Dec 02 11:58:54 it seems bug 1624 fixed itself. Dec 02 12:01:00 03koen 07org.oe.dev * r7db6bb94... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom 2007.1: remove spurious hyphen from TARGET_PREFIX Dec 02 12:12:39 qte-common-2.3.10.bb has "CXXFLAGS += "-fno-visibility-inlines-hidden" which breaks for sharprom-compatible. Is there a way to do something like CXXFLAGS_sharprom-compatible -= "-fno-visibility-inlines-hidden" (this does not work of course, but the idea is there ;) ) Dec 02 12:16:39 nicolasfr: Have you put the information in http://pastebin.ca/264412 into the BTS? Dec 02 12:17:22 Laibsch: RP has just merged it Dec 02 12:17:55 Laibsch: but I don't remember if it was put in BTS Dec 02 12:19:22 Laibsch: can you try to update OE and bitbake libpng 1.2.12. I'd like to know if 1624 is fixed or not Dec 02 12:21:09 * cyrilRomain would like to test too, but still in NOTE: package glibc-2.5-r4: task do_compile: started Dec 02 12:22:17 cyrilRomain: I think 1624 only affected sharprom-compatible distro Dec 02 12:22:33 nicolasfr: ok Dec 02 12:22:59 Any idea why the dir name for u-boot is uboot (without the dash) but the packet is u-boot ? Dec 02 12:25:37 Oh, bug weekend? Dec 02 12:25:40 Nice! Dec 02 12:25:51 nicolasfr: I will pull and verify. Dec 02 12:26:02 nicolasfr: Did you receive my PM? Dec 02 12:27:00 03koen 07org.oe.dev * r85c514dd... 10/ (1 packages/glibc/glibc-initial_2.5.bb): glibc initial: add 2.5, needed for powerpc Dec 02 12:33:04 * koen|tdose stabs glibc Dec 02 12:33:07 likewise: yes? Dec 02 12:33:12 Is http://permalink.gmane.org/gmane.comp.handhelds.openembedded/14027 still an issue? Dec 02 12:33:57 * koen|tdose starts heading to the logfs talk Dec 02 12:34:09 giel: just checking if you are visiting tdose today? Dec 02 12:35:11 koen|tdose: Are you a professional attendee of conferences now? ;-) Dec 02 12:35:27 All these conferences that nobody ever heard of Dec 02 12:37:35 03koen 07org.oe.dev * r54848140... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom 2007.1: some more tweaks for powerpc Dec 02 12:39:09 koen : Didi you build glibc for efika ? Dec 02 12:39:40 Laibsch: yes I received your PM, but it seems my answers do not get through (gaim bug maybe) Dec 02 12:39:46 what is the PARALLEL_INSTALL_MODULES = "1" thing? Dec 02 12:44:22 morning Dec 02 12:44:48 hi Dec 02 12:44:57 Jin^eLD: its for machines which can run 2.4/2.6 kernels - add -2.4/-2.6 into kernel modules packages names Dec 02 12:45:06 hi Cliff Dec 02 12:45:19 hrw: Hi Dec 02 12:45:43 hrw: ah, I see, thx Dec 02 12:45:48 hrw: hello Dec 02 12:46:43 how goes bugdays? Dec 02 12:47:34 phone.. Dec 02 12:49:18 my youngest brother got new phone. Dec 02 12:49:46 so now my 3 old cellphones are not used. Dec 02 12:50:06 hrw: You are contributing to e-waste :) Dec 02 12:50:20 hrw: http://bugs.openembedded.org/buglist.cgi?cmdtype=runnamed&namedcmd=LastWeekFixed Dec 02 12:50:25 hrw: 12 bugs and counting Dec 02 12:50:45 Laibsch: 'namedcmd'is your own search Dec 02 12:51:02 yes, you are right Dec 02 12:51:06 Just a moment Dec 02 12:51:58 hrw: how is the new build machine working out? Dec 02 12:52:01 http://tinyurl.com/y7s3ju Dec 02 12:52:32 cbrake: its nice to have such power under finges Dec 02 12:52:57 Laibsch: thx. saved Dec 02 12:53:27 hrw: you running a 64 bit OS? Dec 02 12:53:35 cbrake: 64bit debian sid Dec 02 12:54:06 OE qemu fails to run, cant build 2.4-crapix kernels - otherwise all is working Dec 02 12:54:18 hrw: great Dec 02 12:54:41 REFERRED_PROVIDER_task-bootstrap is no longer needed, right? Dec 02 12:54:56 Jin^eLD: MACHINE_TASK_PROVIDER = "task-base" is better Dec 02 12:55:16 we need to force machine maintainers to switch to task-base Dec 02 12:55:25 task-bootstrap is scheduled for removal Dec 02 12:55:51 I know.. that's why I asked :) thanks Dec 02 12:57:52 cbrake: I do not like one thing in amd64 platform - x86 compatibility. Dec 02 12:59:00 hrw: de nada. Oh what a feeling. To give back to one of the head honchos ;-) Dec 02 12:59:02 cbrake: in bios there are still things like 'A20 handling' from xt times or '15-16M memory hole' for xt/at ISA machines etc Dec 02 12:59:22 Laibsch: ke? Dec 02 12:59:33 ke? Dec 02 12:59:43 que I meant. Dec 02 13:00:17 I meant it was nice to not always be sucking blood out of you and others. but to be able to give back. Dec 02 13:00:31 ah Dec 02 13:00:49 head honcho = one of the real master Dec 02 13:01:19 Laibsch: http://bugs.openembedded.org/show_bug.cgi?id=1568 is your toy... Dec 02 13:01:53 Laibsch: its rtti/nortti thing Dec 02 13:02:37 I am afraid this toy is overstretching my capabilities. Dec 02 13:02:40 * hrw want bitbake gui... Dec 02 13:11:29 hrw: yeah, its amazing that x86 compatibility still exists! Dec 02 13:13:47 lack of 64bit drivers for 64bit mswin.. Dec 02 13:14:24 hrw: re rtti. Maybe you can give a few more keywords, and maybe just maybe I have a chance at taking a stab at the problem? Dec 02 13:14:41 Laibsch: -no-rtti gcc Dec 02 13:15:13 * Laibsch reads http://publib.boulder.ibm.com/infocenter/macxhelp/v6v81/index.jsp?topic=/com.ibm.vacpp6m.doc/compiler/ref/ruoptrtt.htm Dec 02 13:19:32 hrw : Do you know if gpe-image or opie-image build for x86 ? Dec 02 13:21:31 Ifaistos: I built xfce-image (not yet tested) but this is with my xfce recipes which need testing before push Dec 02 13:21:58 Ifaistos: koen built some images for my progear/angstrom - so gpe-image is probably one of them Dec 02 13:28:29 hrw : i started building it in background. see what comes out... Dec 02 13:29:45 hmm.. we have very old php... Dec 02 13:30:08 1751 hrw 25 0 293m 289m 3608 R 31.9 30.8 0:34.70 jc1 Dec 02 13:30:08 3763 hrw 25 0 181m 176m 3420 R 32.6 18.8 0:09.98 jc1 Dec 02 13:30:08 3999 hrw 25 0 119m 114m 3472 R 29.9 12.2 0:06.10 jc1 Dec 02 13:30:13 ~lart java Dec 02 13:30:14 * ibot accelerates a free AOL cd to 50,000 rpm and lets java feel it Dec 02 13:32:16 I wonder how people were able to develop java apps without few GB of RAM in devboxes Dec 02 13:36:37 I wonder how companys like motarola get the brilliant idea that java will make a great embeddeed phone OS. Dec 02 13:37:34 bah 'omg that name is taken!' >_< Dec 02 13:53:55 hi florian Dec 02 13:55:44 ok. time to go shopping Dec 02 13:55:45 cu all Dec 02 13:57:37 hi all Dec 02 14:30:12 hmm, why is it saying "NOTE: preferred version 2.3.5+cvs20050627 of glibc-intermediate not available" Dec 02 14:30:17 it takes 2.5 then Dec 02 14:30:45 I added this PREFERRED_PROVIDER_virtual/arm-linux-libc-for-gcc = "glibc-intermediate" because it complained that it was not set Dec 02 14:30:53 and after that it complains about 2.3.5 not being available.. Dec 02 14:30:58 any ideas what this is all about? Dec 02 15:09:31 zecke: usermanual.xml in trunk has SRCDATE mentioned for SVN, so I guess it's just about regenerating the manual :> Dec 02 15:15:39 Jin^eLD: glibc-intermediate is only used for NPTL -- which 2.3.5 does not support -- not exactly sure on this. Dec 02 15:17:40 oh, so I might not need the intermediate setting at all Dec 02 15:18:04 Jin^eLD: right -- you need to set the provider to not specify glibc-intermediate Dec 02 15:18:20 ok then I think I know where my mistake is Dec 02 15:18:29 I set it there PREFERRED_PROVIDER_virtual/arm-linux-libc-for-gcc Dec 02 15:18:38 Jin^eLD: my build partition is offline right now as I'm resiszing it, or I would check. Dec 02 15:18:51 Jin^eLD: right Dec 02 15:18:52 thanks anyway.. I will look at other distros Dec 02 15:19:38 hey planet oe Dec 02 15:20:08 hi Dec 02 15:20:17 btw, are blackfin targets supported by OE? Dec 02 15:20:24 openssl refuses to build on i386 netvista Dec 02 15:20:28 * cbrake notes e2fsck -f takes a long time. Dec 02 15:20:54 it does mention that it trys to configure for linux-i386 but then compile breaks immidiatly Dec 02 15:22:37 the funny part is that it even doesnt skip over when i fake a tmp/stamps/openssl*.do_compile Dec 02 15:23:50 rob_w : it needs a patch as there is no i386 as a target in openssl Dec 02 15:24:24 hmm i could raise that string into 586 or something Dec 02 15:24:57 Jin^eLD: Blackfin needs uclinux (no mmu) Dec 02 15:26:11 Ifaistos, but why doens it skip after i touched a tmp/stamps/openssl*.do_compile Dec 02 15:26:45 Ifaistos: I thought OE has some uclib stuff? Dec 02 15:26:49 03pfalcon 07org.oe.dev * rb79002cd... 10/ (1 conf/machine/rx3000.conf): Dec 02 15:26:49 rx3000: Machine definition for HP iPaq rx3000 series. Dec 02 15:26:49 * Newer iPaq series, rx3715, rx3415, rx3115 models. Dec 02 15:26:49 * Hardware-wise, similar to h4000, but uses Samsung CPU. Models differ in CPU Dec 02 15:26:49 speed/flash size. (It appears WiFi/BT is present in all models.) Dec 02 15:26:49 * Got good porting progress in HH.org, port appears to be essentially complete. Dec 02 15:26:53 * Closes #1587. Dec 02 15:26:55 03pfalcon 07org.oe.dev * r3221131e... 10/ (5 files in 5 dirs): linux-handhelds-2.6: Add draft defconfig for rx3000. Dec 02 15:26:59 03pfalcon 07org.oe.dev * r20f06569... 10/ (1 packages/linux/linux-handhelds-2.6_cvs.bb): Dec 02 15:27:00 linux-handhelds-2.6_cvs: DEFAULT_PREFERENCE = "-1" Dec 02 15:27:02 * How it came to be otherwise? %) Dec 02 15:27:38 rob_w: have a look at this http://owmnr.digital-opsis.com/project/browser/trunk/org.openembedded.dev/packages/openssl/openssl-0.9.7e/openssl-i386.patch Dec 02 15:27:56 great Dec 02 15:28:33 Jin^eLD: uClinux and uClibc are two different things :) Dec 02 15:28:38 oh.. Dec 02 15:29:01 so OE does not support/has nothing to do with uClinux? Dec 02 15:29:46 Jin^eLD: Not for the moment Dec 02 15:29:51 :( Dec 02 15:31:09 but u can use ulibc afaik != Dec 02 15:31:55 well, I asked because I may get my hands on some BlackFin hardware Dec 02 15:32:00 and I of course would prefer using OE Dec 02 15:32:05 Ifaistos, ur patch looks good Dec 02 15:32:48 i s anyone maintaining familiar distro in oe.dev ? Or is allmost dead? Dec 02 15:34:03 goxboxlive, good question Dec 02 15:34:39 rob_w do you know if familar is taking advantages of task-base? Dec 02 15:35:46 i would think so ... never checked task-base yet Dec 02 15:35:52 I am trying to build a image now, because of lack of support of opie in angstrom (wich works kind of great btw). Just some few issues again. Dec 02 15:36:22 are u stuck ? Dec 02 15:36:56 stucked with buidling famailiar? Dec 02 15:39:01 goxboxlive, but you built opie for angstrom all the time. what's up? Dec 02 15:42:46 psokolovsky_ i am missing getkey in busybox, and there are some opie definitions i am using in my local.conf that should belong to angstrom.conf. I mean if we should choose to have a main distro on the HTC devices we should be able to build a image right out of OE, and not using local hacks. Dec 02 15:42:50 nicolasfr: Is bug 1625 fixed for you? Still fails for me. Dec 02 15:43:48 goxboxlive, then, please submit any changes you have in you local.conf as ticket to OE bugs, and feel free to add me to cc: right away. Dec 02 15:44:35 goxboxlive, as this weekend id bugdays, I exactly want to catch Koen and ask to readd OPIE building support back into angstrom.conf. You see, he's not there now ;-) Dec 02 15:45:17 goxboxlive, we need to work togother to get gradual and sustained support for OPIE in live distro, random switching here and there hardly help. Dec 02 15:47:38 psokolovsky_ I am agree with you, but dont you think i have asked to get 'getkey' in to angstrom. The answer i allways get is "Opie is not a part of an Angstrom distro" so i cant help. Dec 02 15:48:19 rob_w: you got a umpc? Dec 02 15:48:36 goxboxlive, and now you want to maintain own distro? ;-D also, where's bug for thsi getkey matter? I have know idea what's it all about. Dec 02 15:49:09 psokolovsky_: mickeyl added a getkey applet to busybox Dec 02 15:49:29 psokolovsky_: it might not be patched into the current busybox Dec 02 15:49:39 psokolovsky_: and we should force mickeyl to send this thing upstream Dec 02 15:49:59 psokolovsky_ / zecke it's not beeing patched, but the patch is in the directory. Dec 02 15:49:59 zecke, cool ;-) is it unavoidably needed for opie? Dec 02 15:50:24 psokolovsky_: I think we used to for this "5,4,3,2,1,0 Opie starts now" interrupt Dec 02 15:50:33 no it's not, but you have to change the opie code to get rid of the failure messages. Dec 02 15:50:38 psokolovsky_: I have no idea if we still have this in /etc/init.d/opie Dec 02 15:50:39 goxboxlive, so, you're the requestor, please provide bug # where it is recorded. we'll go further from there. Dec 02 15:51:18 zecke, psst Dec 02 15:51:23 goxboxlive, then let's get rid of it and failure messages too. I'm waiting for bug with description of the matter ;-) Dec 02 15:52:00 psokolovsky_ ok, i can place a bug about it. Dec 02 15:52:22 rob_w: that is a piece of shit device, destroy all your units! Dec 02 15:52:22 goxboxlive, please do, and describe the issue in detail pls. Dec 02 15:52:26 I'll handle it Dec 02 15:52:34 yes Dec 02 15:52:37 zecke, u dont know which i ve ! Dec 02 15:52:37 sorry, goota go, ask for few hours Dec 02 15:52:46 zecke, never Dec 02 15:53:20 rob_w: This device always made me quit my job... Dec 02 15:53:39 only after 1 1/2 days exposed to windows on this device I was about to quit Dec 02 15:53:50 zecke, do you the " we will rule the world " factor ur the pure technical parts ?? Dec 02 15:54:15 s/do you / do you hate Dec 02 15:54:29 rob_w: How they compressed the windows feeling and squeezed it into a overheated device :) Dec 02 15:54:41 rob_w: the hardware and software plainly sucks :) Dec 02 15:55:01 zecke, thats we WE will show them different ,, that not the form-factor was bad , but the OS Dec 02 15:55:16 rob_w: Well the USB Device is bad as well :) Dec 02 15:55:18 my kids love it ;-) Dec 02 15:55:22 rob_w: It is not USB 2.0 Dec 02 15:55:38 hmm let me check on mine but i though it was 2.0 Dec 02 15:55:50 * rob_w grabs for it Dec 02 15:55:58 rob_w: then windows can't handle High Speed devices Dec 02 15:56:28 pskolovsky_ It is allready reported as a bug. bug 1474 It's beeing worked on AFAIK see. Dec 02 15:59:03 zecke, which device are you talking actually? Dec 02 15:59:16 rob_w: samsung q1 Dec 02 15:59:35 at least this is the device I threw out of window (no kidding) Dec 02 16:01:24 hmm Dec 02 16:01:32 :) Dec 02 16:01:52 rob_w: the touchscreen in 1024x768 was not linear at the edges and about 10cm off :) Dec 02 16:01:55 the thing is battery will kill you anytime Dec 02 16:02:11 wow 10cm is quite a too big tolerance Dec 02 16:02:24 rob_w: and I was not able too hook usb keyboard/mice in between the USB dongle Dec 02 16:02:33 mine runs kinda my own touchscreen driver and is acurate Dec 02 16:02:47 rob_w: because this device is not USB2.0 compliant, or the drivers suck Dec 02 16:03:05 rob_w: well in the native resolution it is almost accurate Dec 02 16:03:15 rob_w: if you switch resolutions you lost Dec 02 16:05:39 great.. Dec 02 16:05:54 :) Dec 02 16:06:21 my next toy I'm exposed to is a T830 from fujitsu siemens Dec 02 16:06:55 zecke, that looks more then a pda Dec 02 16:07:13 UMTS equipped device Dec 02 16:07:43 hi there Dec 02 16:07:43 okay, now I need to beat to finish my assignment Dec 02 16:08:19 I'm porting a small gtk+ app to collie target. Dec 02 16:08:49 and the problem I have is how to make FileChooserDialog fit into 320x240 geom. Dec 02 16:09:57 The dialog is too wide because the left pane that holds a list of shortcuts has Add/Remove buttons. Dec 02 16:10:10 The width of these buttons eats most of the space. Dec 02 16:10:22 Any ideas? Dec 02 16:12:51 vlg, either check the containers you used the the propberty`s of the widgets inside of those containers .. Dec 02 16:13:01 or get to fight the gtkrc theme Dec 02 16:14:52 rob_w: I hoped that there would be a gpe widget set of common dialogs that would provide same functionality as stock dialogs, but adjusted to the 320x240. But libgpewidgets shuned dealing with this thorny issue. Dec 02 16:16:32 for file choosing things , check gpe-filemanager , it does implement this functionlity also and successfull Dec 02 16:17:10 rob_w: thanks - I'll do that. Dec 02 16:22:35 vlad? Dec 02 16:23:21 vlg: Are you vlad? Welcome to #oe Dec 02 16:24:26 Laibsch: yes I am - thanks. Dec 02 16:25:03 zecke: Is http://permalink.gmane.org/gmane.comp.handhelds.openembedded/14027 still an issue? Dec 02 16:25:36 vlg: Glad to see you here. Getting granule to compile from CVS? Dec 02 16:26:03 We have bug weekend. Everybody makes an effort to make OE even better. Dec 02 16:27:20 Laibsch: yes and no luck so far ... Dec 02 16:29:46 Laibsch: are you Rolf? Dec 02 16:30:04 Yes Dec 02 16:30:51 I will take a look at what is going wrong with granule in a little while. Dec 02 16:31:43 As I wrote you in a mail I believe the file structure is different from what OE expects. Compilation itself seems to work. Installation step fails with no target to make install. Dec 02 16:32:39 Laibsch: what I discovered is that it never does do_fetch, do_unpack, do_configure, and do_compile. Instead, bitbake jumps right into do_install. That's why it fails - there is nothing to install. Dec 02 16:33:54 vlg: No, that is because you completed these steps earlier. Dec 02 16:34:01 bitbake remembers that. Dec 02 16:34:21 If you want to go through the steps again, do "bitbake -c rebuild granule" Dec 02 16:35:19 Laibsch: OK - that was what I was looking for - clearing its cache - I'll try that in a moment and see if that's the reason. Dec 02 16:36:20 Of course, that only rebuilds the package in question. Dec 02 16:36:26 Which is a good thing. Dec 02 16:37:03 You can also start over completely by rm -Rf tmp/ But that will likely set you back a couple of hours worth of compilation time ;-) Dec 02 16:42:17 Laibsch: I don't trust sourceforge CVS - last night I checked in a bunch of files in, and today, after running 'bitbake -c ...', the thing tells me that there is no updates in 'src' directory. Perhaps, it is a problem on Sourcforge's end. Dec 02 16:42:46 What do you suggest? Dec 02 16:43:25 OE checks out CVS every day, changed or not. Unless you pin it to a certain date. Dec 02 16:44:49 I'll bite the bullet, clean up tmp and start from scratch. Dec 02 16:50:12 Hello Dec 02 16:50:20 I had a question to ask Dec 02 16:50:57 I have been trying to build sdl-perl, and I keep getting this message during compilation: Dec 02 16:50:58 CROSS COMPILE Badness: /usr/include in INCLUDEPATH: /usr/include/SDL Dec 02 16:51:21 I searched in google, and it seems to be a safety mechanism of a sorts Dec 02 16:51:36 I tried to disable it in bitbake.conf, but nothing changed Dec 02 16:51:42 It still comes up Dec 02 16:52:04 B_Lizzard: fix sdl-perl :) Dec 02 16:52:10 no escaping Dec 02 16:52:41 I get a similar error message with galculator Dec 02 16:52:53 Aye, fix sdl-perl, but where to start? Dec 02 16:53:31 As far as I know, the right place to put "include" is /usr/include Dec 02 16:54:06 B_Lizzard: sdl-perl is likely to call /usr/bin/sdl-config due a broken configure script Dec 02 16:54:17 B_Lizzard: a bug has already been posted for that: http://bugs.openembedded.org/show_bug.cgi?id=1066 Dec 02 16:54:18 B_Lizzard: you can shoot the ignorant developer and then patch it Dec 02 16:54:20 anyone know how to disable the cursor in Xfbdev? Dec 02 16:54:35 zecke: I'll look into it Dec 02 16:55:01 cbrake: the best way to do it is have libxcursor use a transparent cursor theme Dec 02 16:55:20 B_Lizzard: thx Dec 02 16:55:21 the transparent cursor package exists in OE Dec 02 16:57:32 zecke : Is the icecream class working ? Dec 02 17:00:46 Ifaistos: with the patch from the bug tracker it should Dec 02 17:00:59 I'm away now :) Dec 02 17:02:44 Hee, icecream class. Dec 02 17:02:46 .... Dec 02 17:02:53 shit, now i want icecream. Dec 02 17:03:48 anybody has a good linux compatible UPS to recommend? Dec 02 17:06:27 APC does stuff that works with linux Dec 02 17:06:35 VoodooZ: I have Back-UPS CS 350 APC which works well with linux - it is usb-controlled and shuts down system gracefully when battery runs out of juice. Dec 02 17:11:30 vlcool. What software do you use for it? What distro? Do you find it enough for a workstation in terms VA rating? Dec 02 17:21:16 vlg: do you use your distro's built-in support or a third-party software like NUT? Dec 02 17:26:18 Laibsch: 1625 is fixed. Can you look at the fetchmail recipe to see if it has a line reading: FULL_OPTIMIZATION_sharprom-compatible = "-fexpensive-optimizations -fomit-frame-pointer -O1" Dec 02 17:27:17 Yes, it does. I pulled and checked before I commented. Dec 02 17:42:36 03Ondrej 07org.oe.dev * rd81c2457... 10/ (1 packages/mythtv/mythtv_0.18.1.bb): mythtv: fix RDEPENDS - close #254 Dec 02 17:45:36 hi Dec 02 17:45:44 ~lart OE for lack of stable branch Dec 02 17:45:45 * ibot beats OE to within 2.54cm of his life for lack of stable branch Dec 02 17:46:09 I build glibc second time now... Dec 02 17:49:53 gcc-cross is broken.. Dec 02 17:52:34 Laibsch: which toolchain do you use? gcc-2.95 from hh.org ? Dec 02 17:53:41 good time Dec 02 17:56:34 * RP minght finally have a few minutes to spend in front of the computer... Dec 02 18:07:34 hrw|afk : did you gcc-cross 4.1.1 failed on the x86_64 machine ? Dec 02 18:08:17 Ifaistos: it fails on java Dec 02 18:09:12 hrw|afk : strange i am building angstrom for progear and it works ok Dec 02 18:09:36 I now added --without-x into recipe and rebuilding Dec 02 18:09:36 hrw|afk : its building gdb now Dec 02 18:10:08 bbl - boston beaked beans (~lart online dictionary) waits for me Dec 02 18:34:17 what is everyone's recommendation for a bootloader if I'm doing a ground up Xscale design? U-boot? Dec 02 18:34:41 Gerrath_: If it does what you need Dec 02 18:35:29 well I need to boot the kernel and be able to easily download a new kernel. A splash screen would be nice bot not required so my requirements are very simple. Dec 02 18:37:45 Gerrath_: "easily download a new kernel" is the key bit. Define easy. If its something u-boot supports, you should be fine, if not... Dec 02 18:37:48 03rpurdie 07org.oe.dev * ra817d73e... 10/ (3 files in 3 dirs): fetchmail: Move optimisation tweak to distro config (so it applies to all versions) Dec 02 18:39:32 RP, I need to be able to flash a new kernel to the board in manufacturing. U-boot does support it, it seems like U-boot is perfect for what I need but I also have sensed some hesitation from people to recommend it and just thought I'm either missing something or there is something better out there :-) Dec 02 18:41:30 Gerrath_: Field upgrades are probably more difficult Dec 02 18:41:42 RP: Hi! Dec 02 18:42:04 RP, that is not a concern.. I'm in an FDA environment and pretty much field upgrades are a big no-no! Dec 02 18:42:13 RP: wanted to ask, is constraint for led triggers to not be modules is so fatal? ;-) Dec 02 18:42:17 Gerrath_: no problem then :) Dec 02 18:42:19 hi psokolovsky_ Dec 02 18:42:47 psokolovsky_: The triggers themselves can be modules, the trigger core can't Dec 02 18:43:27 RP: So it's ok for our machine-specific triggers to be tristate, right? Dec 02 18:43:57 psokolovsky_: Sure, its just the trigger core that requires to be compiled in Dec 02 18:44:09 (and the trigger core is tiny) Dec 02 18:45:05 RP: Nice, thanks! Anton had a bit of concern about it. Too bad Kconfig is not advanced enought to support cleanly any not compiled in/module/builtin for triggers. ;-) Dec 02 18:45:06 ~lart gcc-cross/java Dec 02 18:45:07 * ibot --purges gcc-cross/java Dec 02 18:45:22 last hour is only gcc-cross... Dec 02 18:47:01 psokolovsky_: Kconfig should be able to support it... :-/ Dec 02 18:48:32 * RP -> back later Dec 02 18:48:34 RP: I meant, it couldn't comfortably. It clearly lacks RECOMMENDS/SUGGESTS support ;-). It would be nice if it pulled with main machine, all its triggers by default, but allowed user to still deselect them. Well, nevermind ;-) Dec 02 18:49:26 uf. looks like java part of gcc is built Dec 02 18:49:33 NOTE: package gcc-cross-4.1.1-r8: task do_install: started Dec 02 18:49:36 :) Dec 02 18:49:51 1h on amd64... Dec 02 18:54:43 hrw, will go into OE? with ARM crosscompile? Dec 02 18:55:45 psokolovsky_: ? Dec 02 18:55:57 hrw, gcc... java... Dec 02 18:56:09 psokolovsky_: I had problem with gcc-cross amd64->x86 Dec 02 18:56:41 ah. sorry. I thought that you work in making gcc support compilation java into ARM binaries... Dec 02 18:57:01 psokolovsky_: I hate java Dec 02 18:57:13 hrw, lol Dec 02 18:57:19 psokolovsky_: gcj works on ARM :) Dec 02 18:57:33 psokolovsky_: even freeswing and gcj work on a simpad Dec 02 18:57:34 for me java is like newest game - good excuse to force you to buy more ram, faster cpu.. Dec 02 18:57:36 hrw, what about C#? ;-E what do you think should be done to not remove mono? Dec 02 18:57:50 psokolovsky_: update to 1.2 Dec 02 18:58:05 psokolovsky_: bitbake gcc-cross took 800M ram Dec 02 18:58:17 zecke|away, well, good news. I mean, if they are fresh? do you know somebody tried that lately? Dec 02 18:58:21 just for 3 jc1 Dec 02 18:58:31 heh Dec 02 18:58:37 psokolovsky_: lately not, bu pb_ fixed upstream gcc :) Dec 02 18:59:16 zecke|away, I wish I had time to deal with high-level stuff as java and c# instead of kernel ;-D Dec 02 18:59:32 psokolovsky_: use this java kernel Dec 02 18:59:40 lol Dec 02 19:00:12 psokolovsky_: http://bugs.openembedded.org/show_bug.cgi?id=1630 is the problem Dec 02 19:00:13 psokolovsky_: http://www.jnode.org/ Dec 02 19:00:48 yeah, I heard about jnode, javaos, etc ;-) Dec 02 19:01:46 psokolovsky_: or use the smalltalk vm which can be booted from grub Dec 02 19:02:04 zecke|away, ok, ok, I got idea, thanks ;-)) Dec 02 19:02:14 hehe Dec 02 19:03:28 zecke|away : Except inheriting the icecc.bbclass do i need to set anything else ? Dec 02 19:04:41 zecke|away : i get the same error as the one mentioned in the bugtracker Dec 02 19:08:08 damn :( Dec 02 19:08:08 Ifaistos: *sorry* Dec 02 19:08:09 Ifaistos: I think you can bitch me in 45 minutes to fix it Dec 02 19:08:47 zecke|away : Thanks :) I found some time to work on the compile cluster... Dec 02 19:09:12 psokolovsky_: any issues with Opie, Opie in OE. My offer to 'inform' you still holds Dec 02 19:10:02 RP, is U-boot the BL you would recommend or is there another on? Dec 02 19:10:11 rp, s/on/one Dec 02 19:10:17 Gerrath_: use u-boot Dec 02 19:10:27 avoid redboot Dec 02 19:10:41 hrw, thanks. redboot is the redhat BL, correct? Dec 02 19:11:03 Gerrath_: do not know who wrote it. many people complain about it Dec 02 19:12:09 Crofton: can you provide pair of user/pass in ossie/* recipes? it is still unfetchable Dec 02 19:12:51 zecke|away, So far, I'd like to convince Koen to re-add support for building OPIE images to Angstrom. I also fixes support for buttons on pocketpc devices. I'm still slow on it... But sure, thanks for help, I will have many chances to address for it, I'm sure! Dec 02 19:13:39 bbl Dec 02 19:14:43 hrw|tv, good timing Dec 02 19:17:13 I just finidhed something, so I will set up a lame account for that Dec 02 19:19:14 thx Dec 02 19:19:34 can you remind me of the user/pass syntax for svn fetcher? Dec 02 19:25:11 svn://user:pass@host? Dec 02 19:25:27 Crofton: it has same syntax as cvs one Dec 02 19:25:42 heh, I don't know that one either :) Dec 02 19:25:47 I made the account :) Dec 02 19:26:08 the fetcher code wasn't helpful Dec 02 19:32:17 zecke|away : Got it working !!! Dec 02 19:39:48 hrw|tv, testing, bunch of stuff is rebuilding though Dec 02 19:55:37 Ifaistos: oh, I' still struggling with my assignment Dec 02 19:56:46 zecke|away : NP. Tommorow is also good :) Dec 02 19:56:57 Ifaistos: attach a patch :) Dec 02 19:57:45 zecke|away : No patch. It just needs to set the ICECC_PATH. If you don;t do that you get the error in the bugtracker Dec 02 19:58:30 zecke|away : The only strange thing while building the toolchain is that is using only the local host... Dec 02 19:59:33 zecke|away, you set up mozilla tinderbox? Dec 02 19:59:55 Crofton|oe: the tbox3 yes Dec 02 20:00:13 tbox3? Dec 02 20:00:58 Crofton: there is tbox which is used on tinderbox.mozilla.org Dec 02 20:01:07 yes Dec 02 20:01:11 and tinderbox2 Dec 02 20:01:18 Crofton: there is tbox2 which is advertised on the mozilla pages but shit Dec 02 20:01:24 Crofton: then there is tbox3 Dec 02 20:01:25 yes Dec 02 20:01:31 where is tbox3? Dec 02 20:01:42 Crofton: which got copied by the OSDL people, landed in freedesktop.org cvs and got hacked a bit Dec 02 20:01:48 and is used on ewi Dec 02 20:02:04 basically, I need something that can rebuild our sw and report build faliurs Dec 02 20:02:15 and in the future test program failures Dec 02 20:02:29 Crofton: check the tbox stuff from the freedesktop.org tbox Dec 02 20:02:43 Crofton: in future use my own django/python based tinderbox Dec 02 20:03:41 so many tinderboxs ... Dec 02 20:03:52 <^^Sokol^^> hello, i need help about Simpad ad Opie 0.8.4 ... It says : booting boot/zImage from partition >root< / find_inode failed for name=zImage / Failed to find inode ... etc... Dec 02 20:03:54 :) Dec 02 20:04:00 Crofton: mine will be the best one Dec 02 20:04:09 of course ... Dec 02 20:04:30 Crofton: but for your project use the one from f.d.o, I only patched the default password for the admin screen away Dec 02 20:04:42 Crofton: and used the sql schemta files from John Keiser to initialize the db Dec 02 20:05:04 ~google zecke selfish Dec 02 20:08:24 nicolasfr: The one that mickeyl advised to use. Dec 02 20:14:23 03mwester 07org.oe.dev * r7a7f97e7... 10/ (1 packages/linux/ixp4xx-kernel_2.6.19.bb): ixp4xx-kernel - bump svn version to add patch for dsmg600 default command line. Dec 02 20:18:10 zecke|away : Is it normal that the remote clients don't get any compile jobs while the toolchain is building ? Dec 02 20:18:15 Laibsch: that is? Dec 02 20:18:46 | checking for scrollkeeper-config... no Dec 02 20:18:46 | configure: error: Couldn't find scrollkeeper-config. Please install the scrollkeeper package Dec 02 20:18:50 I love that... Dec 02 20:19:12 Ifaistos: I have a blacklist of modules Dec 02 20:19:15 gpe-image is NOT buildable Dec 02 20:19:36 Ifaistos: I think the toolchain is part of this Dec 02 20:19:47 hm, don't remember. How do I find out. Dec 02 20:19:53 nicolasfr: hm, don't remember. How do I find out. Dec 02 20:20:22 Ifaistos: Are you using icecc successfully? Dec 02 20:21:35 laibsch : It seems so. Currently it sends all build requests for the toolchain compile to the local host. Dec 02 20:22:14 laibsch : waiting to see what will hapen when cross-compiling kicks in Dec 02 20:22:26 Cool. Dec 02 20:23:13 Did you have to do anything special? I remember that so far it was not so easy: bug 1485 Dec 02 20:25:46 laibsch : The error shows up if you have not setup ICECC_PATH in your local.conf Dec 02 20:31:27 cu Dec 02 20:33:25 Well, I set it via an environment variable : export ICECC_PATH=`which icecc` Dec 02 20:33:32 I guess that should be OK. Dec 02 20:34:20 Ifaistos: Does it absolutely have to be in local.conf? Dec 02 20:49:15 What package(s) should I install for getting GPE alongside Opie? The FAQ is fairly outdated. Dec 02 20:49:27 task-gpe-base Dec 02 20:56:38 XorA|gon1: Thank you. Should that be gpe-task-base? Dec 02 20:56:48 I have that installed but GPE does not start. Dec 02 20:57:04 Not even when I call "/etc/init.d/gpe-dm start" Dec 02 20:57:18 Nothing happens. And I cannot change to chvt 2. Dec 02 20:57:31 But the Z is idle'ing along. Dec 02 20:59:00 Laibsch : Not sure. I just put it in the local.conf Dec 02 20:59:19 XorA|gon1: Do I need to specify exactly what X server to use? Which one would that be for the collie? The FAQ says something along those lines. Dec 02 20:59:41 Ifaistos: Are you seeing results now? Dec 02 21:00:20 Laibsch : No cross-compile yet... building glibc-intermediate Dec 02 21:00:54 On more than one machine? Dec 02 21:01:24 Laibsch : I think it will not work... My host is AMD64 and my clients are PIII and P4. The toolchain will be for AMD64 so it will not run on them... Dec 02 21:02:19 Ifaistos: So are the clients seeing load at the moment or not? Dec 02 21:02:25 Laibsch: I think xserver should be pulled in automatically Dec 02 21:02:28 I never got it to distribute. Dec 02 21:02:34 it is xserver-kdrive-fbdev Dec 02 21:02:43 Laibsch : No. just the local host Dec 02 21:02:54 XorA|gon1: I do have xserver-common, but not xserver-kdrive-fb. Dec 02 21:03:26 RDEPENDS_gpe-base-depends := "\ diet-x11 \ virtual/xserver" Dec 02 21:03:32 so should be an RDEPEND Dec 02 21:12:15 ~prod koen|tdose Dec 02 21:12:24 * ibot zaps koen|tdose with a high voltage cattle prod Dec 02 21:12:34 heheheheheh Dec 02 21:12:41 hey XorA|gon1 Dec 02 21:12:56 hey koen|train Dec 02 21:12:59 XorA|gon1: Well, that seems to be a bug in oz354x which I filed about 30 minutes ago under bug 1633. Furthermore, that dependency on diet-x11 is declared for meta-gpe but not for gpe-task-base. Dec 02 21:13:21 Laibsch: oh, it might be different in oz354x Dec 02 21:13:39 It seems like it. I am sure hrw|gone will fix it soon. Dec 02 21:14:01 I tried to use precompiled packages to test GPE-only stuff. Dec 02 21:14:24 So far I had Opie only since we do not yet have Japanese input for GPE. Dec 02 21:15:09 XorA|gon1: Would it be diet-x11 or virtual/xserver that pulls in the machine-specific xserver? Dec 02 21:15:40 Laibsch: virtual/xserver would pull any any server that provides virtual/xserver Dec 02 21:16:09 So that is the machine-specific one that I need to specify explicitly? Dec 02 21:16:54 Laibsch: poossibly, its been a long time since i played with pulling in X by ipkg Dec 02 21:27:52 virtual/xserver specifies which one to build, ${XSERVER} specifies which one to install Dec 02 21:29:15 koen|train: ${XSERVER} is new for Angstrom is it not? Dec 02 21:29:29 nope Dec 02 21:29:45 it has existed since kdrive is in the oe tree Dec 02 21:29:59 since kdrive builds multiple servers Dec 02 21:30:14 hmm.. in my distro.. my usb harddisk is recognized but somehow mount says "No such device", allthourh fdisk -l lists everything properly; any ideas what I may be missing? do I need to configure busybox with some special mount? it's a vfat drive Dec 02 21:36:18 ok.. the kernel modules were missing :> now... I know there are those DISTRO_FEATURES which map to lists of kernel modules in task-base, right? Dec 02 21:36:22 is your filesystem compiled into the kernel? Dec 02 21:36:58 how would I properly create custom features? add them to task base or list modules somewhere in my distro.conf? Dec 02 21:37:10 koen|train: got a module... just installed it and got further Dec 02 21:39:05 resolving all dependencies by hand now... which leads back to my question :> Dec 02 21:41:39 the missing dependencies are a bug with objdump iirc Dec 02 21:42:06 and iirc DISTRO_EXTRA_DEPENDS can hold extra stuff you want to install Dec 02 21:42:43 but tiredness + beer + too much greek food + train may introduce errors in my statements ;) Dec 02 21:42:52 :)) Dec 02 21:43:04 no, you are right about DISTRO_EXTRA_DEPENDS, I use that to specify my custom packages Dec 02 21:43:18 but I thought that the kernel modules were handled in a different way Dec 02 21:43:30 because I saw this features thing and lists of modules that match the feature in task-base Dec 02 21:43:41 kernel modules your machine really, really needs go into MACHINE_EXTRA_DEPENDS Dec 02 21:44:04 to mymachine.conf you mean? Dec 02 21:44:06 since you might have a sucky bootloader that limits the kernel to 1MB and you need extra modules in the rootfs Dec 02 21:44:56 btw, what is better - to compile stuff into the kernel or to use modules? Dec 02 21:45:04 hmm, I have redboot on that system, not sure what it can handle Dec 02 21:46:46 I prefer to have stuff in the kernel, but people like RP prefer to have it as modules so the can test different versions Dec 02 21:47:06 is there any impact in terms of performance/memory consumption? Dec 02 21:47:11 any difference I mean Dec 02 21:47:26 not any noticable AFAIK Dec 02 21:47:52 Compiled in modules can't be unloaded so always use resources. If you always need then that's fine... Dec 02 21:48:35 RP: can they be unloaded automatically if not used? I mean - is there something in place for that, or would I have to resort to custom rmmod scripts? Dec 02 21:49:05 Jin^eLD: There are the standard kernel mechanisms for dealing with that Dec 02 21:49:31 RP: at least on PC I do not remember seeing that modules were unloaded automatically Dec 02 21:49:45 is some configuration setting required for that or so? Dec 02 21:50:14 Jin^eLD: Offhand I don't know although I think modprobe and friends have some feature that does this... Dec 02 21:51:46 i c Dec 02 21:56:18 OE really rocks :) I think I finally figured out a setup to switch between platforms and the distro is also looking good Dec 02 21:56:33 cool :) Dec 02 21:57:07 :) Dec 02 21:57:31 * RP wishes more people could see that :) Dec 02 21:57:49 I was surprised how easy it was to create a patched custom kernel Dec 02 21:58:08 Jin^eLD: That was where I came into OE :) Dec 02 21:58:16 :) Dec 02 21:58:41 I think the way is still difficult, there are tons of options and settings and everything, and documentation.conf misses quite a few things Dec 02 21:58:58 I don't think I would get to where I am without the #oe channel support Dec 02 21:59:10 Jin^eLD: I think we'd all agree our documentation needs improvement... Dec 02 21:59:50 :> Dec 02 22:02:21 btw, I thought there are logs from this channel online, could not find them Dec 02 22:02:31 ~logs Dec 02 22:02:41 hmm Dec 02 22:02:42 apt/ibot/jbot/purl all log to http://ibot.rikers.org// where channelname is html encoded ie: %23debian | lines that start with a space are not shown | some channels have stats at http://ibot.rikers.org/stats/.html.gz, or updated "nightly" Dec 02 22:02:44 aah Dec 02 22:02:48 thanks Dec 02 22:03:04 want to look up something I asked before ;) Dec 02 22:04:00 icecream trial was semi-sucesfull... Dec 02 22:04:19 got the bbclass to run Dec 02 22:04:29 but Dec 02 22:04:51 it seems all the machines in the cluster have to be the same arch Dec 02 22:05:12 Ifaistos: I talked to likewise today, and he said he'd look at icecc.bbclass in the near future as well Dec 02 22:05:27 Jin^eLD: I have a script to get all the oe logs if you want: http://www.rafb.net/paste/results/wGpOdd39.html Dec 02 22:05:28 icecc must be of same version (or really close versions) Dec 02 22:06:01 cyrilRomain: that's cool, thank you Dec 02 22:06:11 koen|train : I think we have an issues if the machines in the compile farm are of different architecture Dec 02 22:07:39 koen|train : I'll have a look at it tommorow. Dec 02 22:08:09 I tried to build glibc for powerpc today, and it failed Dec 02 22:08:15 I know why it failed, but I'd have to sit down to come up with a solution Dec 02 22:09:08 koen|train : which glibc ? i managed to build 2.3.5_svn with gcc 4.1.1 Dec 02 22:09:29 2.5 Dec 02 22:10:04 it fails on building glibc-initial, because it passes -some-funky-ppc-option to x86 gcc Dec 02 22:10:51 koen|train : i used Distro generic and it builds ok. Dec 02 22:11:00 koen|train : task-base Dec 02 22:11:04 glibc 2.5? Dec 02 22:11:51 koen|train : no. "2.3.5+cvs20050627" Dec 02 22:13:27 bye guys. Dec 02 22:45:13 hmm, I still can't quite figure this one thing out.. for a distro + multimachine; I get ipk packages like armv5te - those would match all compatible arm machines Dec 02 22:45:21 and then I have packages with mymachine.ipk in the name Dec 02 22:45:33 so those are special for the machine Dec 02 22:45:37 but all packages land in deploy/ipk Dec 02 22:45:42 can I finegrain it Dec 02 22:46:31 so "*armv5te.ipk" get in their own directory, "*mymachine1.ipk" go to another directory, and "*mymachine2.ipk" go to yet another directory? Dec 02 22:55:03 Jin^eLD: I think it could be done modifying classes/package_ipk.bbclass Dec 02 22:55:36 so no way to do this via some .conf file? Dec 02 22:55:38 Jin^eLD: but classes/packaged-staging.bbclass uses the DEPLOY_DIR_IPK as well, so it is maybe not so trivial to change Dec 02 22:56:03 Jin^eLD: no AFAIK Dec 02 22:56:29 mmhmm Dec 02 22:58:04 Jin^eLD: if you use multiple local.conf file Dec 02 22:58:18 you can try defining DEPLOY_DIR_IPK in each one Dec 02 22:58:26 hmm true, indeed Dec 02 22:58:28 I could do that Dec 02 22:58:35 because I do use multiple local.conf files, good idea Dec 02 22:58:36 thanks Dec 02 23:04:32 Jin^eLD: hmm yes I think it will work. Dec 02 23:42:37 Anyone have a moment to help and OE-newb? It's a an OE-build question, but it involves familiar/GPE, so this might be the wrong place... Dec 02 23:44:39 gneuf: I am an OE-new myself but I could try, I had nothing to do with familiar/GPE so chances to get useful info from me may be slim Dec 02 23:44:41 :) Dec 02 23:45:19 Using the latest bitbake, sources pulled from the handhelds repository... Dec 02 23:45:53 bitbake gpe-image fails because it can't find the glibc-localedata-i18n. Dec 02 23:46:29 Which, of course, exists in tmp after it gets built during the build... Dec 02 23:46:52 I know it's and "inherit"/"ipkg" mistake on m,y part... Dec 02 23:46:56 uhm Dec 02 23:47:16 But so far (4 days), I can't pinpoint it in any faqs or bug posts... Dec 02 23:47:23 gneuf: what is the version of bitbake do you use ? Dec 02 23:47:52 gneuf: I guess I am too newb for this... Dec 02 23:48:19 Initially, the 1.3 from the handhelds repo, but then 1.6.3 Dec 02 23:48:28 Same error both times. Dec 02 23:48:54 I had to patch ipkg.py for the "*control" bug on debian... Dec 02 23:49:23 ok, 1.6.3 is the right one to use (don't try to use the lastest trunk version!) Dec 02 23:49:34 Bypass qemu and use the local one (although I now have patches that make qemu-native compile) Dec 02 23:49:40 On debian Dec 02 23:49:55 LAtest trunk still expolding on me.... Dec 02 23:50:33 So, I also tried: full cleans, full cache deletions, full tmp removal.... Dec 02 23:51:27 PREFERRED_PROVIDER for glibc - I specified the full CVS name of "glibc-2205....."" Dec 02 23:51:53 And today I'm trying interactive building and debugging to see if I can figure out Dec 02 23:52:29 exaclty why the path to the file desn't seem to get inherited when meta-gpe-images is constructed in work/ dir Dec 02 23:52:44 No luck so far... Dec 02 23:54:53 cool, running GaPaE and Opie at the same time Dec 02 23:55:10 GPE Dec 02 23:56:52 Neat trick. Dec 02 23:57:01 Indeed Dec 02 23:57:22 And I found out that way that gaim did install an icon into the menu. Dec 02 23:57:31 Just not for opie ;-) Dec 03 00:03:30 Anyone have a moment to help and OE-newb? It's a an OE-build question, but it involves familiar/GPE, so this might be the wrong place.. Dec 03 00:06:41 is anyone familiar with redboot? Dec 03 00:08:11 Familiar, but not really up to date Dec 03 00:08:12 I figure out how to write the kernel image to flash Dec 03 00:08:30 doing a "go address" does not really work, just prints some crap on the screen Dec 03 00:08:36 bye Dec 03 00:08:38 bye Dec 03 00:08:52 I guess I need some offset? Dec 03 00:09:25 gneuf: Don't meta-ask. Just ask Dec 03 00:13:53 gpe-image fails with "Cannot find package glibc-localedata-i18n" Dec 03 00:15:32 Can you bitbake glibc? Dec 03 00:16:19 Yes - after editing the ipkg.py tar commands, which have "*control" in them instead of "./control". Dec 03 00:16:29 The binary locales get built. Dec 03 00:16:43 And so does glibc-localedata-i18n Dec 03 00:17:07 I alos found OE bug #1432. Dec 03 00:17:10 ...also... Dec 03 00:17:25 BUt I couldn't find out what was done to fix it. Dec 03 00:18:17 So far, I'm thinking that the .ipk doesn't get built... Dec 03 00:18:22 Or built in the wrong place. Dec 03 00:21:07 Jin^eLD: that crap is probably GDB debug protocol Dec 03 00:23:10 huh? Dec 03 00:24:34 I think I just killed my board :) Dec 03 00:24:49 fis erase did not turn out as I expected ;) Dec 03 00:30:52 hey rwhitby - almost done! what a grind... Dec 03 00:47:29 wtf is it with tcl scripters and (im)propper indetion damnit Dec 03 01:12:28 gneuf: Seems like you are experiencing bug 1432. Nothing seems to have been done yet to fix it. Dec 03 01:15:52 HopsNBarley: cool Dec 03 02:24:11 . Dec 03 02:31:33 does any one know how to back up my copy of pocket pc before I install OE? Dec 03 02:40:13 zezom: Use Bootblaster to save the .gz images of the RTOM before proceding... Dec 03 02:40:21 .........................................ROM... Dec 03 02:40:41 On an iPaq Dec 03 02:40:43 :) Dec 03 02:41:00 gneuf, thanks.. :) Dec 03 02:41:22 You have read, of course, http://www.handhelds.org/feeds/bootldr/install.html#savingwince Dec 03 02:41:53 no I havn't I did try looking for a how to a few days back though Dec 03 02:42:08 thanks again Dec 03 02:42:10 Which model? Dec 03 02:42:27 rx4540 Dec 03 02:42:41 Newly, experimentally supported. Dec 03 02:43:02 I believe you should be looking for the "Ramdisk Rescue" method for loading it if it's supported... Dec 03 02:43:03 it's all good, I'll just mess with it for a while if I need to Dec 03 02:43:23 And my ROM backup instructions may thus be incorrect... Dec 03 02:44:08 Try Dec 03 02:44:10 http://www.handhelds.org/moin/moin.cgi/RamdiskRescue Dec 03 02:44:11 ok, I'll make sure I'll ready through the whole page before I start any way Dec 03 02:46:23 gneuf, do you know if the wireless is supported on the rx4540? it's just that it has a G card in it instead of the B cards that most have Dec 03 02:47:39 Not sure. Dec 03 02:48:17 that's cool Dec 03 02:48:26 I'm just keeping my fingers crossed :) Dec 03 02:49:26 You'd have to check if the chip has changed Dec 03 02:49:45 And if there's a module that supports it. Dec 03 02:50:34 I don't even know if there's a port yet - I've only seen the 2200, 4000, and 4700 supported so far... Dec 03 02:50:43 is their a "bootable business card" or PDA's ? Dec 03 02:51:03 PDA's Dec 03 02:51:24 yeah me too I'm using the hx4700 profile atm Dec 03 02:51:27 Ramdisk Rescue is a dual-boot CF process. Dec 03 02:51:46 You can reset to WinCE, but you have to restore the RAM (your stuff). Dec 03 02:51:49 but can it do things like lspci and all that? Dec 03 02:51:59 Yah. Dec 03 02:52:04 nice Dec 03 02:52:20 You could theoretically install a complete Debain system once it's up... Dec 03 02:52:34 With a big PCMCIA disk or flash... Dec 03 02:52:51 THen you'd get real bash, full apt-get, etc... Dec 03 02:53:03 I have 1gig of internal and a 2 gig sdcard Dec 03 02:53:14 Instead of the standard monolithic shell-and-utils. Dec 03 02:53:30 If the SD is supported, then you could but a lot 'o Debian on it... :) Dec 03 02:53:37 ...put... Dec 03 02:53:46 cool Dec 03 02:53:52 brb Dec 03 02:53:56 np **** ENDING LOGGING AT Sun Dec 03 02:59:57 2006