**** BEGIN LOGGING AT Fri Jan 12 02:59:57 2007 Jan 12 03:38:24 hey, where have all the hrw|gone ... Jan 12 03:39:13 :> Jan 12 03:42:04 slightly off topic, but... Jan 12 03:42:14 is Linux capable of doing MMC over SPI over Parport? Jan 12 03:42:17 ^^;; Jan 12 03:43:33 the only SPI I saw was on a custom board with a custom kernel driver.. so, no idea Jan 12 03:45:33 I'm hacking the newest CVS one-time-use digicam Jan 12 03:45:50 one time use digicam? cool Jan 12 03:45:58 it comes with a display? Jan 12 03:46:02 well, not one-time-use when I hack it Jan 12 03:46:06 yeah Jan 12 03:46:11 how much is it? Jan 12 03:48:27 $20 Jan 12 03:48:34 this model sucks tho Jan 12 03:48:37 wow, thats really cheap... Jan 12 03:48:38 but too late for me Jan 12 03:48:41 :) Jan 12 03:48:50 the previous model was worth >$20 Jan 12 03:48:52 not sure about this one Jan 12 04:03:30 time to go to sleep.. enough coding for tonight Jan 12 04:03:36 crap.. it's morning again Jan 12 04:24:23 03pfalcon 07org.oe.dev * reb1c9680... 10/ (1 packages/linux/linux-handhelds-2.6.inc): linux-handhelds-2.6.inc: Order list of supported machines a bit. Jan 12 04:24:27 03pfalcon 07org.oe.dev * r1cd7dafd... 10/ (3 files in 3 dirs): linux-handhelds-2.6 2.6.16: Add defconfig for h1910. Jan 12 05:10:25 03lenehan 07org.oe.documentation * r8e0bf212... 10/ (1 usermanual/chapters/recipes.xml): Jan 12 05:10:25 usermanual: Update for recipes chapter. Jan 12 05:10:25 Lots of updates for the bitbake recipes chapter I accidentally checked Jan 12 05:10:25 in as part of a different commit earlier today. This is not yet Jan 12 05:10:25 linked into the main document - a few more sections need to be filled Jan 12 05:10:25 and cleaned up before it's ready for that. Jan 12 06:27:44 03lenehan 07org.oe.documentation * rd69cc584... 10/ (1 usermanual/chapters/recipes.xml): usermanual: Fill the SRC_URI and tasks section of the bitbake recipes chapter. Jan 12 07:16:39 hi all Jan 12 07:17:24 03lenehan 07org.oe.documentation * r1c28ebcd... 10/ (1 usermanual/chapters/recipes.xml): Jan 12 07:17:24 usermanual: Update for the recipes chapter: Merge the messages and functions Jan 12 07:17:24 sections, and fill in the content describing each of the functions. Jan 12 07:27:07 03hrw 07org.oe.oz354x * r34695e98... 10/ (5 files in 4 dirs): avr-libc: dropped Jan 12 07:27:10 03hrw 07org.oe.oz354x * r19727a35... 10/ (4 files in 3 dirs): cvs: dropped 1.12.8 Jan 12 07:27:13 03hrw 07org.oe.oz354x * r6fb0a525... 10/ (3 files in 3 dirs): konqueror-embedded: apply patch from #844 to get rid of lack of -lpcre problem Jan 12 07:27:16 03hrw 07org.oe.oz354x * re0c1ef09... 10/ (8 files in 3 dirs): quilt: added 0.45 - close #1768 Jan 12 07:27:50 hi all Jan 12 07:28:03 Morning All! Jan 12 07:29:58 I would like to compile gnumeric for my hx4700, but it fails at gconf-osso configuration. I spent this week to find the problem with no result. Jan 12 07:30:58 First of all I do not know why gconf-osso is a dependency for many package (at leas gnumeric,abiworld). Jan 12 07:31:53 AFAIK gconf-osso is a maemo related stuff. Why is it even necessary? Jan 12 07:33:27 I compiled dbus* and orbit* on which gconf-osso seems to depend. Jan 12 07:33:49 I pasted the result to http://pastebin.ca/313911. Jan 12 07:34:01 Can anybody help? Jan 12 07:34:45 I start a few times from the scratch, with the same result. Jan 12 07:40:45 03lenehan 07org.oe.documentation * r7e65f37c... 10/ (1 usermanual/chapters/recipes.xml): Jan 12 07:40:45 usermanual: Update for the recipe chapter - rename the include section to Jan 12 07:40:45 require and fill in the content for the section. Jan 12 07:41:20 When I run bitbake vlc Jan 12 07:41:21 ? Jan 12 07:41:37 I have an error whitch is Jan 12 07:41:57 error : nothing provide dependence vlc Jan 12 07:47:41 it'sgood Jan 12 07:48:03 so i download source but i have a new error Jan 12 07:50:20 command for created the package ipk for vlc is : bitbake -b /.../vlc-gpe-0.8.4.bb ?? Jan 12 07:52:50 do you know the command for delete vlc package??? Jan 12 07:55:10 but in the fist time there is an error Jan 12 07:55:22 now the package don't making Jan 12 07:56:50 disaster: bitbake -b /.../vlc-gpe-0.8.4.bb -c clean Jan 12 07:57:07 disaster: Add "-c clean" to the bitbake command to delete the package Jan 12 08:02:47 kk thx Jan 12 08:04:55 when i test to making an ipk package , i have the always same error Jan 12 08:05:16 error : Task failed event exeption, abording Jan 12 08:05:27 in under there is Jan 12 08:05:35 Error .... Jan 12 08:05:49 Fztzl : autoconf execution failed Jan 12 08:05:56 any idea ? Jan 12 08:21:28 morning Jan 12 08:23:14 XorA: Hi! it seems I will not be able to use my gnumeric ever. Jan 12 08:24:00 XorA: I post some lines half an hour before. Jan 12 08:24:37 NOTE: build 200701111540: completed Jan 12 08:24:52 I just built angstrom-2007.1 image for my board without issues Jan 12 08:25:00 XorA: I compile all dbus and orbit related package maually, and still the same. Jan 12 08:28:03 XorA: If I try to ./configure with no OE (for PC) the same package (downloaded and uncompressed by OE) finished correctly. Jan 12 08:30:21 hi Jan 12 08:30:39 NOTE: package gnumeric-1.6.3: completed Jan 12 08:31:00 it's possible to put Angstrom system keyboard in azerty ? Jan 12 08:31:14 disaster: change the keymap Jan 12 08:31:51 sza2: pastebin your complete local.conf Jan 12 08:32:03 XorA: Ok. Jan 12 08:35:01 Xora, where is the fileplease ? Jan 12 08:36:49 XorA: http://pastebin.ca/313946 Jan 12 08:38:31 XorA: It takes some time to pastebin, as the system runs on my home PC, so I should login from my workplace, and needs to go through several proxies. Jan 12 08:39:41 morning Jan 12 08:41:37 xserver-kdrive is not include in angstrom distro ? Jan 12 08:42:01 sza2: http://www.angstrom-distribution.org/unstable/images/20070104/hx4700/ Jan 12 08:42:18 sza2: It seems angstrom works on hx4700 Id use that distro Jan 12 08:42:34 sza2: and most of the packages are already in the Angstrom feeds Jan 12 08:44:24 XorA: But I would like to compile myself... Jan 12 08:48:33 good morning all Jan 12 08:51:41 sza2: but try and compile Angstrom, at least we know that works for your machine Jan 12 08:54:08 hi all Jan 12 08:54:19 hey ade|desk Jan 12 08:54:48 XorA: You mean to change DISTRO = "angstrom-2007.1"? Jan 12 08:55:40 koen, xserver-kdrive is include on angstrom ? Jan 12 08:56:04 yes: http://www.angstrom-distribution.org/repo/?name=kdrive&action=search Jan 12 08:56:12 Hello, I'm having a problem Jan 12 08:56:17 http://bugs.openembedded.org/show_bug.cgi?id=1780 Jan 12 08:56:29 sza2: yes Jan 12 08:56:29 Not my bug, but I'm experiencing it Jan 12 08:56:48 koen, how i start x server ? Jan 12 08:57:11 B_Lizzard: join sza Jan 12 08:57:25 sza? Jan 12 08:57:33 sza2: also has that bug Jan 12 08:58:28 Aha Jan 12 08:58:37 I'll add my comment to that bug Jan 12 08:59:15 funnily enough the solution is mentioned in the 'error' Jan 12 08:59:23 "NOTE: multiple providers are available (osso-gnome-vfs2, gnome-vfs); Jan 12 08:59:23 NOTE: consider defining a PREFERRED_PROVIDER to match runtime Jan 12 08:59:24 gnome-vfs-plugin-file" Jan 12 08:59:41 reading hard Jan 12 09:01:14 Lemme define PREFFERED_PROVIDER, see if that fixes my woes Jan 12 09:01:39 koen, how i start x server ? please Jan 12 09:01:52 Xfbdev Jan 12 09:02:00 You need to set DISPLAY first Jan 12 09:02:10 So, "export DISPLAY=:0" Jan 12 09:02:13 for starting the server? Jan 12 09:02:17 no you don't Jan 12 09:02:25 only before starting an app :) Jan 12 09:05:53 03koen 07org.oe.dev * r13871b91... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: prefer gconf-dbus Jan 12 09:06:54 So, is it PREFERRED_PROVIDER_gconf = "gconf-dbus" Jan 12 09:06:55 ? Jan 12 09:07:13 if i shoud run vlc, imust run the server x Jan 12 09:07:38 i msut Jan 12 09:07:50 export the DISPLAY thingie Jan 12 09:07:59 You don't *have* to, but it's better Jan 12 09:08:16 And then start it like Xfbdev & vlc Jan 12 09:08:26 If you haven't got a WM Jan 12 09:08:53 XorA, B_Lizzard: I started, but it will take some time I think. Jan 12 09:09:12 Wut? Jan 12 09:10:01 export DISPLAY=:0 don't found in angstrom Jan 12 09:10:35 B_Lizzard: the error message tellss you what it wants the preferred provider for Jan 12 09:10:54 B_Lizzard: don't get distracted by my lack of coffee an recipe-dylexia :) Jan 12 09:10:59 dyslexia* Jan 12 09:11:18 and Xfbdev notfound Jan 12 09:13:03 disaster: It didn't find export? Jan 12 09:13:16 What kind of distribution doesn't have export? Jan 12 09:13:30 ho yes soory Jan 12 09:13:41 export DISPLAY=:0 it's good Jan 12 09:13:59 but Xfbdev is not found Jan 12 09:14:10 Well, what version of X did you install? Jan 12 09:14:24 Just write X, and then press tab Jan 12 09:14:35 xarg Jan 12 09:14:41 just it Jan 12 09:15:25 Capital X Jan 12 09:15:31 Did you install X? Jan 12 09:15:37 X no command Jan 12 09:16:00 Did you install X? Jan 12 09:16:12 i must install xserver-kdrive? Jan 12 09:16:20 Yeah Jan 12 09:16:25 koen say that is include in angstrom Jan 12 09:16:25 disaster: that would make sense yes :-) Jan 12 09:16:50 but when i run bitbake xserver-kdrive Jan 12 09:17:02 i have an error when in building Jan 12 09:17:22 :/ Jan 12 09:17:42 Are you on your PDA? Jan 12 09:18:18 for building the packages ? Jan 12 09:18:39 I don't really know what you want to do here Jan 12 09:18:44 my box is not an pda, just an mtherboard fo embedded system (epia) for x86 Jan 12 09:19:07 Why build xserver-kdrive if it's in Angstom? Jan 12 09:19:13 I got that Jan 12 09:19:41 yes, but how install it Jan 12 09:19:55 because xserver seems not installed Jan 12 09:20:49 On your PDA, if you have internet, do "ipkg install xserver-kdrive-fbdev" Jan 12 09:22:09 bitbake isn't needed if you want to install something Jan 12 09:22:22 Only if you want to build something Jan 12 09:22:43 But, if it already exists in the feeds, there's no reason to build it Jan 12 09:23:37 03pH5 07org.oe.dev * r114caf85... 10/ (1 packages/initscripts/initscripts-1.0/populate-volatile.sh): Jan 12 09:23:37 initscripts: fix populate-volatile.sh file and directory creation Jan 12 09:23:37 - now chmod gets called do set the correct permissions Jan 12 09:24:11 I KNOW REGULAR EXPRESSIONS Jan 12 09:24:47 disaster: if you installed angstrom epia it should also work with ipkg Jan 12 09:27:38 mourf i have an probelm Jan 12 09:27:46 the source in ftp or http Jan 12 09:28:01 because in the uit the ftp protocol is blok Jan 12 09:28:21 it don't succefull ipkg update Jan 12 09:34:39 ha i586 not existe in www.angstrom-distribution/unstable/ Jan 12 09:34:45 disaster: you can use an ftp proxy Jan 12 09:34:48 i686 is the same Jan 12 09:34:51 or no ? Jan 12 09:35:07 Hmm. I might have failed googling but is the following syntax explained somewhere: PREFERRED_VERSION_gcc-cross = "3.4.4+csl-arm-2005q3". I do understand the "normal" use of PREFERRED_VERSION, but this is not so clear. Jan 12 09:35:58 Cockroach-: PREFERRED_VERSION just has to match whatever PV is set to in the .bb you want to use. Jan 12 09:37:13 Cockroach-: wait, you are talking about the _gcc-cross override? Jan 12 09:38:24 pH5: The thing which I didn't understand was that since gcc-cross_csl-arm-2005q3.bb doesnt define version in the .bb name, I had problems when I tried to set PREFERRED_VERSION_gcc-cross earlier Jan 12 09:38:29 but I managed to google that syntax out Jan 12 09:39:37 the PV isn't set in that .bb either. Jan 12 09:40:05 Cockroach-: gcc-cross_csl-arm-2005q3 requires (imports) gcc_csl-arm-2005q3, which does set the PV Jan 12 09:40:26 mourf there is not the source ipkg for i586 and epia in the angstrom website ? Jan 12 09:40:38 source ipkg? Jan 12 09:40:41 what's that? Jan 12 09:40:59 pH5: Oh, I missed that. Jan 12 09:41:00 Thanks Jan 12 09:41:52 03andrew.paulsen 07org.oe.dev * r317ff5b9... 10/ (1 packages/xorg-lib/libx11_X11R7.1-1.0.1.bb): libx11: fix makekeys compilation, close #1447 Jan 12 09:41:55 03koen 07org.oe.dev * r1c3a6748... 10/ (1 conf/machine/epia.conf): epia: cleanup and adapt to task-base Jan 12 09:42:00 03koen 07org.oe.dev * r903c623e... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: add feed arch for epia Jan 12 09:42:18 when i run ipkg update Jan 12 09:42:50 disaster: source ipkg don't exist Jan 12 09:42:51 it try download on http://www.angstrom-distribution/unstable/i586/machine/epia/packages.bz Jan 12 09:43:29 or the folder i586 is empty Jan 12 09:43:36 but Jan 12 09:43:38 disaster: post the error message, interpreted ones are usually wrong... Jan 12 09:45:23 http/1.1 404 not found is my error message Jan 12 09:45:57 koen: you know the url to download the ipkg packages ? Jan 12 09:48:58 how i do ?please Jan 12 09:53:10 what machine i choose in http://www.angstrom-distribution.org/unstable/feed/i686/machine/ Jan 12 09:53:14 for epia ? Jan 12 09:57:39 ~change 1249 usd to eur Jan 12 09:57:44 1,249.00 United States Dollar (USD) makes 967.687 Euro (EUR) (from http://www.xe.com/) Jan 12 09:59:01 any idea ? please Jan 12 10:25:45 how i install the package vlc-gpe Jan 12 10:25:58 for angstrom Jan 12 10:26:20 in the first time i must run bitbake vlc-gpe Jan 12 10:26:37 then copy the packe ipk in my CF Jan 12 10:26:53 and run ipkg install vlc-gpe.ipk ? Jan 12 10:27:38 because when i run bitbake vlc-gpe i have error Jan 12 10:29:04 whats the error disaster Jan 12 10:43:53 hrw, koen: Now that I have an OE-patched kernel, may I be surprised at random memory corruption? Jan 12 10:44:50 not really, it is still a pxa Jan 12 10:44:54 ~lart pxa Jan 12 10:45:15 * ibot whips pxa with a wet and grimy noodle just because Jan 12 10:45:43 Any ideas what to do? CONFIG_CPU_DCACHE_DISABLE? Jan 12 10:46:06 Or will that hurt performance too much? Jan 12 10:46:51 Funny that the problem is ony present at that one place... Jan 12 10:46:54 *only Jan 12 10:50:27 how i do for instlled a package .ipk?? Jan 12 10:51:03 ipkg -i stuff.ipk Jan 12 10:51:08 (IIRC) Jan 12 10:52:42 i don't work Jan 12 10:53:45 but i download .ipk and i just installed this files Jan 12 10:54:05 but when i run ipkg install xserver-kdrive.ipkg Jan 12 10:54:13 .ipk Jan 12 10:54:42 it test to downlod by internet but this file is in local Jan 12 10:54:55 Yes, you need to use -i and not install. Jan 12 10:54:56 how i do please ? Jan 12 10:55:11 -i is not found Jan 12 10:55:20 Hm... Jan 12 10:55:33 ipkg -i xserver-kdrive.ipk Jan 12 10:56:02 Well, maybe I remembered wrong... Haven't used ipkg in a while. Jan 12 10:56:13 Run it without arguments and look at the help text it gives? Jan 12 10:56:15 ipkg install blah.ipk Jan 12 10:56:27 disaster: please read manuals at least once Jan 12 10:56:59 but ipkg install stuff.ipk try to download package by internet Jan 12 10:57:12 not install the local packages Jan 12 10:58:00 disaster: only if the version it thinks is on internet is later than the local one Jan 12 10:58:31 disaster: clear all your sources, then try it Jan 12 10:58:55 kk i test now Jan 12 11:00:33 ~lart ipkg for operating on the principal of pure luck Jan 12 11:00:34 * ibot whacks ipkg with the cluebat for operating on the principal of pure luck Jan 12 11:02:53 it strange that thje ipkg not download the package witch are present in my sourcve Jan 12 11:04:40 disaster: welcome to the craziness that is ipkg Jan 12 11:05:02 Maybe the source is broken or something? Jan 12 11:13:52 ipkg -h or seomthing like is usful Jan 12 11:14:06 I had to figure out how to downgrade a package last night Jan 12 11:15:17 BTW, is ipkg still horribly slow to start even when you give it an invalid command? Jan 12 11:16:03 the machine I have with ipkg isn't here, but maybe it isn't setup to look for stuff on the net? Jan 12 11:16:14 is there an /etc/ipk.... you can modify? Jan 12 11:16:15 From my time with OE, I recall it taking like 5-10 seconds to display the help text Jan 12 11:17:49 DataBeaver: No real delay here for ipkg -h although that is a pxa270 device with a microdrive... Jan 12 11:18:13 I think ipkg -h was fast enough... But what if you say ipkg foobar? Jan 12 11:19:08 RP: that is because ipkg has stopped parsing lists when it isn't needed Jan 12 11:20:18 RP: You're maintaining the kernel, right? Can you explain this: http://tdb.fi/~tdb/asound_trouble.txt Jan 12 11:20:50 DataBeaver: That pauses for a fraction of a second Jan 12 11:21:11 ok, the slowness is gone then. Jan 12 11:22:37 DataBeaver: I don't know what those functions do offhand Jan 12 11:23:38 snd_mask_copy does old=*mask; and snd_mask_intersect does a bitwise and (mask->bits[i]&=m.bits[i];) Jan 12 11:25:35 http://img333.imageshack.us/img333/7530/av55tb.gif Jan 12 11:26:02 DataBeaver: I don't know the code well enough to comment, sorry Jan 12 11:27:38 http://tdb.fi/~tdb/asound_trouble2.txt <-- A version without the functions manually inlined Jan 12 11:27:45 hrw|work: did you see that the zaurus has been EOL'ed? Jan 12 11:27:58 wot? Jan 12 11:28:45 koen: I know that Jan 12 11:28:52 from good source ;D Jan 12 11:29:02 we all knew that, but now it's official Jan 12 11:29:12 I guess they ran out of 8GB microdrives Jan 12 11:29:18 Err, I meant "with the functions manually inlined" a few lines back Jan 12 11:29:35 i.e. without any calls to functions othe than printk Jan 12 11:29:42 koen: I didn't know it was official... Jan 12 11:30:19 DataBeaver: You're suggesting its a compiler problem? Jan 12 11:30:21 What's this? http://pastebin.ca/314046 I have had this for several days noe. Sould it be reported as a bug or is it something i miss localy. It just happend aftre i stated with a new fresh OE dev db. Jan 12 11:30:41 ah well, our super rare Z's rise in price now :-) Jan 12 11:30:42 RP: koen hinted at cache corruption previously Jan 12 11:31:09 Although compiler problem isn't ruled out either Jan 12 11:31:12 cbrake, u there ? Jan 12 11:31:13 DataBeaver: Its not cache corruption Jan 12 11:31:16 goxboxlive: don't use an old OE checkout Jan 12 11:31:45 koen: I downloaded the mtn 032 database Jan 12 11:31:55 goxboxlive: so? Jan 12 11:31:59 And updated monotine to 0.32 Jan 12 11:32:18 goxboxlive: don't use an old OE checkout Jan 12 11:32:41 RP: k... Would an assembly dump be of any use for determining whether it's a compiler problem? Jan 12 11:32:54 mtn pull ; mtn up before complaining helps a lot Jan 12 11:33:03 I don't know enough about ARM assembly to make anything out of it. Jan 12 11:33:44 DataBeaver: What are you compiling it with? Jan 12 11:34:15 hmm, I need to bother you again. I've been building the angstrom-bootstrap-image for few times now since I've had to tweak things from here and there. Now I have the following problem: I tuned the kernel compile options, and when I do a bitbake rebuild on virtual/kernel, everything goes OK except that the tmp/rootfs/lib/modules/... does not get updated. Jan 12 11:34:26 RP: gcc version 4.1.2 20060708 (prerelease) (Debian 4.1.1-8) Jan 12 11:34:37 Rebuilding angstrom-bootstrap-image doesn't do it either. Jan 12 11:34:55 can someone here help me identify the values for a new tft , like pixelclock, vsync etc ?? i ve a pdf of the tft next to me but i dont know what to put into pxafb.c ?? Jan 12 11:34:59 DataBeaver: Just slightly off topic then :-/ Jan 12 11:35:06 and I wouldn't like to build everything from the scratch either. Jan 12 11:35:53 heh Jan 12 11:36:04 I wondered why builds were taking so long Jan 12 11:36:21 * koen increases BB_NUMBER_THREADS from 1 to 3 Jan 12 11:36:24 DataBeaver: Firstly, work out the simplest test case you can showing the problem and show me the full thing, not just a fragment of it without the function definitions Jan 12 11:37:06 Well, it's kernel code, so that might be difficult... Not sure if I can reproduce it in userspace. Jan 12 11:37:13 Can try though. Jan 12 11:38:11 Checking if userspace does it would be a good test Jan 12 11:45:32 At least replicating just that part of the code doesn't produce the problem... Jan 12 11:46:43 rob_w: whats up? Jan 12 11:47:39 DataBeaver: You replicated with the functions as defined in the kernel? Jan 12 11:48:07 cbrake, i got me target 6.5 inch tft and i now need to find its correct values like pixelclock margins etc Jan 12 11:48:17 No, I tried with just the assignment and anding first Jan 12 11:48:44 DataBeaver: Since that worked in kernel space, there is no reason to think it would work differently in userspace Jan 12 11:48:45 rob_w: is it a sharp? Jan 12 11:49:45 Well, it did produce quite different assembly output... Jan 12 11:50:47 rob_w: LCD timings -- the process I usually use is: 1) get the pxa datasheet and lcd datasheet and compare waveforms. The basics is the Hsync and Vsync have dead time at the beginning and end where no data is transfered -- often called "front porch", "back porch". Jan 12 11:51:32 The dead time for hsync is timed in pixclocks. The dead time for vsync is timed in hsync pulses. Jan 12 11:51:49 2) take your best guess -- it is very difficult to get it right the first time. Jan 12 11:52:44 front porch etc i see on my tft pdf Jan 12 11:52:48 are those the margins ? Jan 12 11:52:57 3) get a scope and start tuning the waveforms to look like the LCD datasheet. Some of the delays need to be exact. Jan 12 11:53:25 rob_w: yes, front porch is margin. Jan 12 12:19:38 RP: bitbake foo bar baz -k is broken in trunk: http://rafb.net/p/EFY2P393.html Jan 12 12:27:10 http://www.butterwizard.com/ - nice gadget Jan 12 12:28:30 koen: ok, thanks. That looks like a simple fix Jan 12 12:31:10 hmm. What's wrong with openembedded.org? Jan 12 12:31:31 hrw|work: indeed Jan 12 12:34:49 03rpurdie * r744 10bitbake/lib/bb/runqueue.py: runqueue.py: Fix prepare_runqueue parameters Jan 12 12:34:54 "The Greenphone Community SDK can be downloaded from Qtopia.net" Jan 12 12:35:11 RP: can you fix trunk<>.oz354x problem? Jan 12 12:35:37 hrw|work: Which problem is that? Jan 12 12:35:45 moment, let me dig out Jan 12 12:35:56 03koen 07org.oe.dev * rf5ae76da... 10/ (1 packages/qiv/qiv_1.9.bb): qiv: fix do_install Jan 12 12:35:59 03koen 07org.oe.dev * r6d9b1de4... 10/ (1 packages/povray/povray_3.6.1.bb): povray: fix do_configure Jan 12 12:36:32 RP: http://pastebin.ca/312372 Jan 12 12:36:39 RP: incidentally I also fixed the stuff causing the build to fail :) Jan 12 12:37:39 hrw|work: Every anonymous fuction in oz354x is going to need reindenting Jan 12 12:37:42 koen: even better :) Jan 12 12:39:18 RP: http://rafb.net/p/lgjcoU54.html Jan 12 12:40:13 koen: That was with -k? Jan 12 12:40:25 yes Jan 12 12:41:04 koen: I'll have to look at that later Jan 12 12:41:14 RP: ok, thanks Jan 12 12:47:49 03rpurdie * r745 10bitbake/lib/bb/runqueue.py: runqueue.py: Check for failed dependencies before trying to find a fnid Jan 12 12:48:06 koen: try that, I can't easily test locally... Jan 12 12:49:38 * koen tests Jan 12 12:52:05 RP: that works, thanks Jan 12 12:52:32 koen: Great, thanks for testing Jan 12 12:53:16 grrrr Jan 12 12:53:23 hi all!! Jan 12 12:53:26 how are you? Jan 12 12:53:32 why doesn't iCal show weeknumbers? Jan 12 12:53:45 * koen searches for a paper version if it Jan 12 13:17:47 hi all! Jan 12 13:22:22 * XorA wonders how mickey|zzZZzz does all that work in his sleep Jan 12 13:24:30 RP: I seem to be unable to produce that problem in userspace... I tried copypasting the exact code from the kernel. Jan 12 13:24:53 Crofton|home: re Jan 12 13:25:39 DataBeaver: I guess you'll have to try and create a minimal kernel test case then. The trick would be to minimise the amount of code so the assembler dump is easier to read Jan 12 13:26:01 i have /etc/ipkg/ Jan 12 13:26:12 where many files Jan 12 13:26:24 i have modify the source Jan 12 13:26:58 RP: I can pretty much isolate the bugging lines into about 40 or so lines of assembly Jan 12 13:27:22 because www.angstrom-distribution/unstable/feed/i586... don't exist Jan 12 13:27:35 DataBeaver: and how does it compare to the case you illustrated where it works? Jan 12 13:27:43 i have replaced by www.angstrom-distribution/unstable/feed/i686 ... Jan 12 13:27:54 ipkg update good run Jan 12 13:28:24 but no ipkg install xserver-kdrive Jan 12 13:29:17 disaster: try http://www.angstrom-distribution.org/unstable/feed/i586/ Jan 12 13:29:18 the ipk pacakage is present in http://www.angstrom-distribution.org/unstable/feed/i686/base/xserver-kdrive-fbdev_X11R7.1-1.1.0-r5_i686.ipk Jan 12 13:29:21 you can see the feeds in there Jan 12 13:30:05 lol, it's my box witch have created the folder when i try to upadte ipkg Jan 12 13:30:18 all folder in i586 are empty Jan 12 13:30:20 RP: It's somewhat similar now... With different registers and offsets thought and some differences in the instructions. Jan 12 13:33:13 I can try to write a minimal kernel testcase too... Jan 12 13:33:31 First I need to learn how to write kernel modules though. Jan 12 13:33:52 DataBeaver: The more minimal you make it, the easier it is to see the problem in the assembler Jan 12 13:36:43 http://tdb.fi/~tdb/asound_trouble3.txt <-- There's the current code, if you want to take a look while I try to make a kernel testcase Jan 12 13:36:52 koen: it's possible that an person upload the folder i586 ? Jan 12 13:42:46 ~seen AlexTE Jan 12 13:43:06 alexte was last seen on IRC in channel #openzaurus, 217d 21h 6m 53s ago, saying: 'oops :)'. Jan 12 13:54:30 Koen : It's possible?? Jan 12 14:04:17 in machine on http://www.angstrom-distribution.org/unstable/feed/i586/, the folder doesn't exist Jan 12 14:04:45 epia but next nothing folder Jan 12 14:05:05 not folder , package Jan 12 14:08:04 an idea?? Jan 12 14:13:33 hey Jan 12 14:14:01 i have delete the files in /var/lib/ipkg Jan 12 14:14:08 and run ipkg update Jan 12 14:14:17 now the package donwload and installed Jan 12 14:15:40 re Jan 12 14:18:46 hi all Jan 12 14:18:49 i must add packet to used a touchscreen dalle in serial port ? Jan 12 14:18:55 hi Jan 12 14:24:14 morning Jan 12 14:24:29 mickeyl: good morning Jan 12 14:24:39 XorA: I wonder about that myself ;) seriously, of course I'm a multiple personality and a couple of identities are working 24h Jan 12 14:25:44 koen: regarding bug 1700 (compilation of perl for Sharp), I sedded out lgcc_s as suggested. Now there is a different problem: http://rafb.net/p/BQANOa12.html CAn you advise on how to deal with this? Jan 12 14:29:31 good morning mickeyl Jan 12 14:29:35 cheer pb_ Jan 12 14:29:36 good morning florian_kc Jan 12 14:29:36 s Jan 12 14:30:44 hey pb__ Jan 12 14:33:06 Laibsch: note the line sed -i -e 's|/usr/include|${STAGING_INCDIR}|g' ext/Errno/Errno_pm.PL Jan 12 14:33:27 Laibsch: I guess you dont have errno.h in ${STAGING_INCDIR} Jan 12 14:33:37 Hello Jan 12 14:33:42 yo sirfred Jan 12 14:34:20 XorA: Your 320x240 mode capable xserver is near. ;-) Jan 12 14:34:30 sirfred: yay Jan 12 14:34:51 hi sirfred Jan 12 14:34:53 RP: Hello? Jan 12 14:35:00 RP: Hey, I was looking for you. ;) Jan 12 14:35:07 sirfred: I know ;-) Jan 12 14:35:19 RP: I would like to comment something about the w100fb and 320x240 modes. Jan 12 14:35:21 my touchscreen seems good run with cat /dev/ttyS0 Jan 12 14:35:36 RP: I would like to know your oppinion about mode changing. Jan 12 14:35:36 sirfred: sure, go on Jan 12 14:35:39 but where is the file conf of xserver-kdrive Jan 12 14:35:55 RP: OK. Well, the thing is that I'm trying to support 320x240 modes via RandR in Ximageon Jan 12 14:35:56 to specifi that /dev/ttyS0 is my mouse Jan 12 14:36:02 morning Jan 12 14:36:13 sirfred: Personally, I'd prefer to leave that to the kernel to handle through an ioctl Jan 12 14:36:28 RP: So do I, but I have some problems. Jan 12 14:36:57 RP: First, I was trusting the EXT_MEM_LOCATION and FB_LOCATION registers to know the amount of memory of the card and its mapping address. Jan 12 14:37:08 RP: But when the mode is changed to 320x240, the external mem is hidden. Jan 12 14:37:10 XorA: You are right. ../tmp/sharprom-compatible/staging/i686-linux/include/ does not include Errno_pm.PL but neither does the tmp dir for openzaurus or angstrom. Jan 12 14:37:21 RP: Is there any reason to to so? Jan 12 14:37:27 ??? Jan 12 14:37:45 RP: IIRC. 640x480 mode FB_LOCATION (intmem) EXT_MEM_LOCATION(extmem) Jan 12 14:37:46 Laibsch: errno.h not .pl Jan 12 14:37:53 320x280 moe FB_LOCATION (intmem) EXT_MEM_LOCATION(intmem) Jan 12 14:38:10 sirfred: Since we didn't use it, I had the kernel driver power it down to save power Jan 12 14:38:27 sirfred: There is no reason other than that Jan 12 14:39:04 RP: So, other than changing that register, is there any other things done with the external memory ? Is it actually powered down? Jan 12 14:39:32 sirfred: Yes, I'm fairly sure it does get powered down Jan 12 14:39:39 RP: OK. I trust you. ;-) Jan 12 14:39:59 RP: The problem is that the most memory we have, the best for offscreen based accel. Jan 12 14:40:16 RP: And another thing that came into my mind while looking at the w100fb.c code. Jan 12 14:40:42 sirfred: It sounds like its time to rethink what's getting handled by w100fb and how Jan 12 14:40:45 RP: It's supporting the w3220 card also, and I thought that card had more memory than the w100. Hasn't it ? Jan 12 14:41:08 RP: I think that it's better to have a central place to change those things (mode, and so) Jan 12 14:41:10 sirfred: The internal memory is always 384k, the external can vary Jan 12 14:41:38 RP: But IIRC, w100fb is using a fixed extmem amount, isn't it? Jan 12 14:41:43 sirfred: I agree the mode changing is the kernel drivers job. Its just a question of how userspace can tell it to leave the external memory on Jan 12 14:41:57 RP: sysfs? Jan 12 14:42:02 sirfred: A question for you, when we suspend, what should happen to video memory contents? Jan 12 14:42:20 sirfred: At the moment its preserved but I'm thinking of changing that. Jan 12 14:42:25 how i do to say at server X to start with my touchscreen (ttyS0) Jan 12 14:42:37 RP: I'm really not sure of what happens at X server core in suspend/resume cycles. Jan 12 14:42:56 RP: I know it works, but I'm not sure if the screen content is redraw and/or the offscreen memory is swapped out. Jan 12 14:43:08 But I can investigate it, of course. Jan 12 14:43:14 sirfred: kernel does chvt for suspend Jan 12 14:43:20 sirfred: FWIW, the kernel will swicth consoles before suspending Jan 12 14:43:24 sirfred: so X does same as chvt code Jan 12 14:43:41 XorA: Which one would you choose /usr/local/arm/2.95.3/arm-linux/include/linux/errno.h or /export/interim/openembedded/tmp/sharprom-compatible/work/perl-5.8.7-r21/perl-5.8.7/mint/errno.h? Jan 12 14:43:45 hrw|work: maybe :) Jan 12 14:43:58 Laibsch: /usr/local/arm/2.95.3/arm-linux/include/linux/errno.h Jan 12 14:44:06 Laibsch: as thats the one for your target Jan 12 14:44:06 Is the only way to interact with sysfs use the files? Jan 12 14:44:14 sirfred: Its something to think about as saving the memory contents is a big part of the suspend/resume delay Jan 12 14:44:21 sirfred: Most drivers don't do it Jan 12 14:44:31 So, should we need to use the ioctl API and also to open a sysfs file? Jan 12 14:44:47 RP: /sys/somewhere/w100fb/external_mem 0/1 Jan 12 14:45:05 sirfred: When your Xdriver loads it would set something in sysfs to tell w100fb to always leave the extmem on Jan 12 14:45:11 XorA: Thanks. I guess I will use another instance of sed to correct this, then. Jan 12 14:45:29 Laibsch: you should be able to generalise it Jan 12 14:45:38 RP: Is there a way to know the path to the w100fb/extmem (for example) place ? Jan 12 14:45:38 sirfred: Maybe. I'm open to alternatives Jan 12 14:46:00 XorA: ? Jan 12 14:46:09 sirfred: /sys/devices/platform/w100fb/extmem ? Jan 12 14:46:45 sirfred: its gets buried in sysfs though, doesn't it :-/ Jan 12 14:46:45 RP: Perhaps the best solution was to use something in the fb_var_screeninfo Jan 12 14:46:50 But I suppose that nothing matches this use. Jan 12 14:46:59 RP: Yes, that's the bad thing about sysfs Jan 12 14:47:06 Laibsch: move that line to do_compile_prepend() and make yourself a do_compile_prepend_sharp Jan 12 14:47:23 RP: I never know the way to locate a file. I commonly use /sysfs/bus/platform/devices/w100fb Jan 12 14:47:31 But, is that true for every machine ? Jan 12 14:47:32 XorA: I see. Thank you. Jan 12 14:48:15 sirfred: /sys/bus/platform/devices/w100fb should always work Jan 12 14:48:30 RP: So, perhaps that's the best solution. I don't see another better one. Jan 12 14:49:20 XorA: Will http://rafb.net/p/hsWbL055.html be sufficient to zap the sed stuff for Sharp? Jan 12 14:49:24 sirfred: I don't see an alternative either Jan 12 14:50:00 RP: Another thing that would be desirable is the ioctl to populate the fb_var_screeninfo.rotate field. Jan 12 14:50:12 Laibsch: no, you still need that sed for sharp, but you need to change it to ${WHATEVER_VAR_DETAILS_WHERE_MY_TOOLCHAIN_IS} Jan 12 14:50:27 sirfred: I need to find a conversation about that. Someone told me we shouldn't do that :-/ Jan 12 14:50:31 RP: So, I would be able to get from there the startup orientation of the screen, and use it to calculate the randr commands relatively to that startup orientation. Jan 12 14:50:35 RP: ok. Jan 12 14:50:41 sirfred: btw, I experimented with that kernel patch of yours. It seemed to really slow down scrolling on the device. I'm not 100% sure it was your patch, I need to confirm that Jan 12 14:50:45 RP: I can also use the DISPLAY_CNTL. Jan 12 14:51:02 XorA: Yes, I will figure this out later. But I wanted to make sure the general sed line will not be applied in the sharprom case Jan 12 14:51:03 RP: I also experienced strange behaviours sometimes. Jan 12 14:51:10 sirfred: I would like to use the rotate field if we can Jan 12 14:51:15 Laibsch: looks good then Jan 12 14:51:22 RP: To say you the truth, I'm not sure of how the HOST_DATA register works. Jan 12 14:51:38 XorA: Great. Thanks. Let's see how this turns out. Jan 12 14:51:46 sirfred: Have you considered adding an option to the kernel to load the microcode and then using that if present? Jan 12 14:51:47 RP: It's some kind of bus master, because it's very easy to hang the machine if the amount of data feeded to them is not the exact needed. Jan 12 14:51:48 Laibsch: though if you are on 32bit, that will make perl compile as x86 32bit errno.h is arm compatible Jan 12 14:52:01 RP: By the moment, I would like not to use the microcode. Jan 12 14:52:10 XorA: ? Jan 12 14:52:15 I am on 32bit. Jan 12 14:52:20 sirfred: Licence reasons? Jan 12 14:52:27 RP: I'm not sure if we are going to have an speed increment doing that, and also, license reasons, yes. Jan 12 14:52:59 RP: I never tested the arithmetic operations (scaling and so) using the ringbuffer,neither. Jan 12 14:53:25 sirfred: One day I hope its an option but I understand the concerns Jan 12 14:53:33 Laibsch: without that sed, it will look at /usr/include/errno.h and the x86 version is the same as the arm version Jan 12 14:53:44 RP: Some time ago, I thought that perhaps just register feeding would work without the microcode, so we could use the ringbuffer to feed more data. Jan 12 14:53:45 Laibsch: I added that sed originally as this is not true for x86_64 Jan 12 14:53:49 RP: But it doesn't Jan 12 14:54:00 RP: I mean, using PACKET0. Jan 12 14:54:41 RP: And what about the external memory of the better cards ? Is it always set to be 2 Mb ? Jan 12 14:55:06 Perhaps there's a way to get it from some cfg register? Jan 12 14:55:09 XorA: So I guess I do not need that sed line for Sharp ROM after all? Since I assume the prebuilt sharp stuff will not generally be 64.bit aware? Jan 12 14:55:20 sirfred: w100fb allows the machine to define how much there is Jan 12 14:56:34 RP: OK. I see, mach->mem->size Jan 12 14:56:36 Great. Jan 12 14:57:24 RP: So, resuming Jan 12 14:57:47 w100fb will provide a way to tell it not to stop the external mem, via sysfs Jan 12 14:58:08 sirfred: yes, I think thats easiest Jan 12 14:58:09 You will look for the reasons of not exposing rotate information in the fb_var_screeninfo Jan 12 14:58:19 right Jan 12 14:58:37 One question, is there a way to setup any mode with any rotation using the ioctl? Jan 12 14:58:53 RP: For example, randr = 270 degrees Jan 12 14:59:49 RP: If I understand correctly, I can set 320x240, and so the rotation will be 90 degrees. Jan 12 14:59:55 Or 240x320 and the rotation will be 0. Jan 12 15:00:12 How could I get 180 and/or 270 ? Jan 12 15:00:38 I know there's a flip sysfs file, but it would be nice not to have to touch more sysfs files, ;-) Jan 12 15:01:00 sirfred: In theory, the proper way would be the rotate field Jan 12 15:01:29 RP: ok Jan 12 15:02:49 RP: And about preserving screen information, I think that the key is if kdrive is suspending aware. Jan 12 15:03:02 RP: If that's true and is able to interact with the driver, I see no problem. Jan 12 15:03:16 sirfred: I think its aware of console switches and the kernel does a console switch before suspending Jan 12 15:03:21 Laibsch: np, fix it the right way, not the lazy way Jan 12 15:03:28 sirfred: you can test with chvt as XorA said Jan 12 15:03:28 s/np/no/ Jan 12 15:03:51 RP: I will build a debug Ximageon version, and see if I'm called . Jan 12 15:04:29 Well, I think I have no more questions. Jan 12 15:04:37 to suspend system i must add an package ? Jan 12 15:04:48 ipkg ? Jan 12 15:04:49 RP: Thanks a lot for your time. :) Jan 12 15:05:06 np. You've made some amazing progress recently :) Jan 12 15:05:44 RP: I expect to be able to implement some of the xrender accel in the future. Jan 12 15:06:16 sirfred: sounds good :) Jan 12 15:06:26 greetings to all Jan 12 15:06:30 RP: blend hooks of kdrive. It's a pitty that the w3220 seems to be different on how rotate/scale works. Jan 12 15:07:06 RP: To say the truth, all the code I've implemented in the kdrive driver was already present it libw100. Jan 12 15:07:27 sirfred: The w3220 is really looking that different? :-/ Jan 12 15:07:27 RP: It was just I was a little afraid to fight against the kdrive APIs. Jan 12 15:07:38 * RP knows the problem Jan 12 15:07:55 RP: It seems that ENG_CNTL and INC_SRC / INC_SRC2 are different. Jan 12 15:07:56 sirfred: next you'll be saying we can run compiz on Zaurus Jan 12 15:08:03 Probably INC_SRC has more bits. Jan 12 15:08:10 XorA: :) Jan 12 15:08:16 XorA: Who knows. Perhaps too less memory to do that. Jan 12 15:08:54 but only perhaps? :D Jan 12 15:08:54 hrw|work: :) Jan 12 15:09:14 At a first look, the main problem seems to be the alpha channel format. Jan 12 15:09:32 The imageon needs a 4BPP alpha channel, or 12BPP alpha channel, using separated alpha components Jan 12 15:09:38 sirfred: You'll have to give some hints to a w3220 owner and see if they can figure it out Jan 12 15:10:05 RP: Yes, I will disturb koen and mreimer as far as I have some test to feed their machines. Jan 12 15:10:08 sirfred: The alpha problem put me off trying for X accell. I couldn't work out if it would ever do anything useful Jan 12 15:10:32 RP: libw100 has some demos. Jan 12 15:11:02 RP: But the alpha plane is always 4BPP. Perhaps there's some E2_ARITHMETIC_CNTL operation to do something. Don't know. Jan 12 15:11:27 kdrive has a hook to upload something to the card, perhaps that could be hacked in some way. Jan 12 15:12:03 It's a little dirty, but if we were able to transform it while uploading, knowing the format, who knows. Jan 12 15:13:05 sirfred: I don't know the code in question too well ;-) Jan 12 15:13:47 RP: Another dirty thing I'm doing in Ximageon is assuming the display resolutions possible. Jan 12 15:13:58 RP: For RandR, the possible sizes need to be registered. Jan 12 15:14:35 RP: Perhaps I could probe all the possible modes against the framebuffer, using ioctl. Jan 12 15:14:40 But I felt lazy Jan 12 15:14:47 sirfred: I guess that will be machine specific... Jan 12 15:15:10 RP: Yes, so the good way would be to have a table of possible modes, and use the ioctl to test them. Jan 12 15:15:49 RP: Is Mardy still working on Xw100? Jan 12 15:15:58 RP: Have you heared from him lately? Jan 12 15:16:08 sirfred: I guess so, yes. Keep in mind the w100 is not the limiting factor, the lcd is Jan 12 15:16:19 sirfred: Not since I met him really Jan 12 15:16:50 RP: I think there's an ugly bug in Xw100 preventing it to use offscreen memory at all. Jan 12 15:17:31 sirfred: I'd guess your work surpasses that now? Jan 12 15:17:47 RP: Well, there's a reason I started Ximageon and didn't continue with Xw100 Jan 12 15:18:11 RP: I wanted to organize in a different way all the card information. Jan 12 15:18:20 So, I would have to change completely Xw100 Jan 12 15:18:33 Also, I have hacked kdrive to be able to use the internal and the external memory. Jan 12 15:18:45 And finally, Ximageon is (by now) not working on w3220 devices. Jan 12 15:18:48 sirfred: Is there anything Xw100 can do that Ximageon can't Jan 12 15:18:50 ? Jan 12 15:18:54 RP: w3220 Jan 12 15:19:08 So, I thought it was better to start from scratch. Jan 12 15:19:16 Yes, I can understand that Jan 12 15:19:53 Well, the thing is that kaa is using this code to calculate the pixmap pitch, and then, to know how much memory needs: Jan 12 15:19:57 pitch = (w * bpp / 8 + pKaaScr->info->pitchAlign - 1) & Jan 12 15:19:57 ~(pKaaScr->info->pitchAlign - 1); Jan 12 15:20:09 And the problem is that Xw100 is not setting pitchAlign, so is zero Jan 12 15:20:29 ouch Jan 12 15:20:30 I think that kdrive internals should never let a zero value to go on there. It should be corrected to at least 1 Jan 12 15:20:39 hi, all! Jan 12 15:20:41 So, it's perhaps a kdrive bug, better than a Xw100 one. Jan 12 15:21:08 * koen hears his name Jan 12 15:21:35 sirfred: let me reflash my hx4700 so I can do something usefull with it, like testing your new stuff :) Jan 12 15:22:08 Hi! Jan 12 15:22:11 koen: Perhaps with the changes I'll make after talking with RP, w3220 will be working. Jan 12 15:22:17 sirfred: cool Jan 12 15:22:34 koen: I think the problem is in the code I stoled from kernel, that has some w3220 specific code I didn't copy. Jan 12 15:22:35 sirfred, btw, we still have issues with w100 kernel accel with hx4700 ;-( Jan 12 15:22:48 psokolovsky: What issues? Jan 12 15:23:23 psokolovsky: As I was talking with RP now, I'm not really sure of how HOST_DATA involved operations work, I'm afraid. Jan 12 15:23:34 sirfred, hangs when outputting under "pressing" conditions, like after resume, around freq changes, etc. Jan 12 15:23:37 psokolovsky: So, it seems that sometimes it can lead to slowdowns. Jan 12 15:24:16 sirfred, ok, I'm glad there's new info on that and you're working on it Jan 12 15:24:45 psokolovsky: I was actually talking about the Xserver. I have done nothing at kernel level. :-( Jan 12 15:24:49 frequency changes sound nasty from the w100 point of view... Jan 12 15:25:16 ;-( Ok, we'll keep testing and figuring out... Jan 12 15:25:26 RP: Is there any chance that the blt routine could be called while suspending/resuming or so ? Jan 12 15:25:37 RP: I mean, from two different threads ? Jan 12 15:25:56 Or is that synchronized in some way? Jan 12 15:26:06 sirfred: there was a patch around to disable accel whilst suspending/resuming Jan 12 15:26:42 RP: Perhaps suspending is not waiting for the card to be idle. Jan 12 15:26:46 Just guessing Jan 12 15:26:51 sirfred: mreimer might have sent it to me. I've just not had a change to get to that bit of my todo list yet (but I have cleared a lot off it) Jan 12 15:27:07 RP: That's great. Jan 12 15:27:29 RP: I fear to know the length of your todo list. ;-) Jan 12 15:27:35 Well, I have to leave now. Jan 12 15:27:49 sirfred: I think the problem was we were not disabling the console so it was trying to accel ops after disabling the w100 Jan 12 15:28:09 sirfred: Its good to have caught up a bit :) Jan 12 15:28:12 RP: That seems a good source of hangs. Jan 12 15:28:33 RP: Glad to talk with you, after such a lot of time. :-) Jan 12 15:28:43 I have to leave now. See you later. Jan 12 15:29:56 hi... does anyone here knows by a chance how to check whether my kernel supports eabi ? Jan 12 15:39:48 03mickeyl 07org.oe.dev * rc5b8908a... 10/ (5 files in 2 dirs): Jan 12 15:39:48 include/preferred-gpe-versions.inc, sane-srcdates.inc: catch up with removal of gnome-vfs-dbus. Jan 12 15:39:48 NOTE: may need to adjust SRCDATE for gnome-vfs? Jan 12 15:40:43 bye all and thx for your help Jan 12 15:40:57 i go to sky on week Jan 12 15:40:59 mickeyl: gnome-vfs is now built via tarballs Jan 12 15:41:15 ok, then the entry in srcdate.inc can be removed Jan 12 15:41:29 hmm Jan 12 15:41:36 which version of gnome-vfs works w/ gtk+ 2.6.10 ? Jan 12 15:42:02 [and doesn't need bonobo] Jan 12 15:42:28 2.16.x, works with any gtk, might need a recent glib Jan 12 15:42:42 bummer, yeah Jan 12 15:42:53 which version of gnome-vfs works w/ gtk+ 2.6.10 ? Jan 12 15:42:55 ups Jan 12 15:42:57 | checking for MODULES_XML... Requested 'glib-2.0 >= 2.9.3' but version of GLib is 2.6.4 Jan 12 15:43:14 will glib 2.9.3 work w/ gtk+ 2.6.10? Jan 12 15:44:52 if not, i have a problem :) Jan 12 15:45:07 * mickeyl removes gnome-vfs for now Jan 12 15:45:36 glib is generally backwards compatible Jan 12 15:45:42 I don't imagine that upgrading it will be a problem. Jan 12 15:45:53 oh good to know, thanks. /me tries building Jan 12 15:46:55 hmm we don't have any 2.9 or 2.10 nor 2.11 Jan 12 15:47:17 i doubt whether gtk+ 2.6.10 builds against glib 2.12 Jan 12 15:47:21 i tried that once Jan 12 15:47:22 it failed Jan 12 15:47:27 hmmm Jan 12 15:48:20 mickeyl: poky-bleeding is using glib 2.12 Jan 12 15:48:37 RP: together w/ Gtk+ 2.6.10? Jan 12 15:49:02 mickeyl: sorry, with 2.10.6. My eyes played tricks Jan 12 15:49:09 *nod* Jan 12 15:51:28 any idea what i can do? Jan 12 15:51:41 should i add an older glib ? Jan 12 15:51:46 something 2.9 Jan 12 15:51:51 what was the failure with gtk+ 2.6.10 and glib 2.12? it might be easy to patch. Jan 12 15:52:13 can't remember. i'll reproduce that and drop you a note Jan 12 15:52:23 righto Jan 12 15:52:29 mickeyl: http://www.cbc.ca/technology/story/2007/01/11/coin-spy.html Jan 12 15:52:33 btw, I am (as of this morning) subscribed to the oe list again Jan 12 15:52:34 heh Jan 12 15:52:39 yay! Jan 12 15:52:43 welcome back :) Jan 12 15:53:04 chouimat: heh Jan 12 15:54:30 I guess I could subscribe to the new gpe list too, that'd be pretty exciting Jan 12 15:54:38 mickeyl: everybody is laughing in canada because of this :) Jan 12 15:54:49 pb__: ! Jan 12 15:55:01 welcome back pb__ :) Jan 12 15:55:06 pb_: absolutely. do you think you have time / motivation to do some active development again some time? Jan 12 15:55:32 mickeyl: yah, I hope so Jan 12 15:55:40 mickeyl: Ross commented "the floating object change might cause a problem. i think the only trouble will be removing any DISABLE_DEPRECATED calls in gtk" Jan 12 15:57:23 RP: i see, thanks. if I manage to come over the crucial performance problems I'm seeing atm., then I might consider going to 2.10.7 with a recent cairo and pango as well Jan 12 15:57:28 but not know Jan 12 15:57:38 s/know/now/ Jan 12 16:00:03 bbl Jan 12 16:02:50 hmm someone said my name Jan 12 16:18:21 Hello, could someone try to compile esound-gpe from actual org.oe.dev ? I'm getting some strange errors... . Jan 12 16:18:31 update repo mmp Jan 12 16:18:44 hrw|work: I already did few seconds ago Jan 12 16:19:00 Does mtn update also file timestamps? Jan 12 16:19:05 03hrw 07org.oe.dev * r454327ea... 10/ (4 files in 3 dirs): ruby: added forgotten patches Jan 12 16:19:22 (do I need to rebuild bitbake's cache) ? Jan 12 16:20:53 Hmm, I'll try bitbaking, but there was no change to esound-gpe, neither some module on which it depends... . Jan 12 16:20:57 (I believe) Jan 12 16:21:30 mmp, IIRC, it was rebuild for me yesterday, after recipe changes Jan 12 16:22:56 We'll see;) Bitbake takes a while to parse all the files... Jan 12 16:23:00 koen: is there a way to check whether the eabi patch is working? Jan 12 16:23:15 koen: i'm trying to use compiled rootfs but it seems to stuck Jan 12 16:25:16 try booting it with 'init=/bin/busybox sh' in the CMDLINE Jan 12 16:27:54 ok, thanks..! Jan 12 16:30:16 psokolovsky: It still doesn't build for me, SVN version of bitbake and most recent org.oe.dev Jan 12 16:30:36 mmp: 'bitbake -c rebuild esound-gpe' Jan 12 16:31:00 mmp, I have had trouble with that one since yesturday as well Jan 12 16:32:20 hi mccarthy Jan 12 16:32:44 koen: it still fails. It fails at fetching, scripts fetch extra/esound, and then bitbake tries to access ${S}/extra, or something like that Jan 12 16:33:06 koen: this: /build/tmp/work/armv5te-angstrom-linux-gnueabi/esound-gpe-20070109-r0/oesvn.EnhVoV/extra Jan 12 16:33:27 weird Jan 12 16:34:31 koen: i get "[42949376.550000] Failed to execute /bin/busybox sh. Attempting defaults..." Jan 12 16:34:56 cu Jan 12 16:35:01 hrw|work: bye:) Jan 12 16:36:56 koen: the older bbfile for esound-gpe fetches just module esound, the newer module extra/esound; it seems like something doesn't ignore all path elements, except the last (svn checks out directly to 'esound', not to 'extra/esound') Jan 12 16:37:34 mmp: I suspect svn behaves different across versions Jan 12 16:37:39 ~lart svn Jan 12 16:37:40 * ibot hauls svn up by the scruff of the neck and spanks him until he waddles Jan 12 16:38:08 koen: svn, version 1.4.2 (r22196) at me Jan 12 16:38:35 svn, version 1.4.0 (r21228) Jan 12 16:39:21 then, in your workdir for esound-gpe, there should be extra/esound in SVN fetch directory... Jan 12 16:40:36 ~botmail for sirfred: I have now network access to my w3220 machine, so I'm ready for tests :) Jan 12 16:45:02 mmp, koen, that's how svn always worked Jan 12 16:45:11 mmp, did mtn update recently? Jan 12 16:45:18 psokolovsky: 15 minutes ago Jan 12 16:46:18 mmp: downloads/svn/esound-gpe$ ls -l Jan 12 16:46:18 total 8 Jan 12 16:46:18 drwxr-xr-x 5 pfalcon pfalcon 4096 Jan 8 22:38 esound Jan 12 16:46:18 drwxr-xr-x 3 pfalcon pfalcon 4096 Jan 12 01:03 extra Jan 12 16:47:01 I have Jan 12 build here Jan 12 16:47:39 psokolovsky: weird, I don't have downloads/svn at all; at me, it checks out directly to the workdir. Jan 12 16:47:53 (and some tmpnam.SVN babble) Jan 12 16:48:07 mmp, impossible. Jan 12 16:48:08 err, oesvn.tmpnam one Jan 12 16:48:25 mmp, ah, sorry, you lack a svn-go-sane patch ;-) Jan 12 16:48:58 :-) So now, my svn is supposed to be insane? :) Jan 12 16:49:18 Is that patch for svn or bitbake? Jan 12 16:49:23 so, maybe it's indeed broken. koen uses 1.7 which includes that patch, and me - 1.6 with patch applied Jan 12 16:49:28 mmp, bitbake Jan 12 16:49:54 mmp, nothing is known about its sanity (actually, bitbake's sanity regarding svn ;-) ) Jan 12 16:50:08 Weird, I believe I'm using SVN trunk of bitbake.. Jan 12 16:51:47 Ahaa, but bitbake-1.6 one:) Jan 12 16:53:26 weird, even at berlios there is just 1.6 version of bitbake... Jan 12 16:53:45 ahaa, and trunk:) Jan 12 16:54:09 03pfalcon 07org.oe.dev * r42072c53... 10/ (1 conf/distro/angstrom-2007.1.conf): Jan 12 16:54:09 angstrom-2007.1.conf: Update htcuniversal's kernel version to known good, 2.6.19-hh7. Jan 12 16:54:09 * Marcin Juszkiewicz tested this version to build ok, and it is known to work. Jan 12 16:55:20 I'll give bitbake-1.7 a shot... Jan 12 16:56:18 hi pb_ Jan 12 16:57:58 03pfalcon 07org.oe.dev * ra3ec87a8... 10/ (1 conf/machine/h1910.conf): h1910.conf: Specify known good kernel version. Jan 12 16:58:37 hmm, bitbake-1.7 reports circular dependencies... . Jan 12 17:01:17 What is the difference between include and require? Jan 12 17:01:48 require errors out on missing files Jan 12 17:01:57 include doesn't Jan 12 17:02:58 koen: thanks Jan 12 17:03:02 I will use require then Jan 12 17:03:18 less surprises :) Jan 12 17:04:39 Anyone out there using parted with glibc 2.5? Jan 12 17:14:09 psokolovsky: where could I get the svn patch? I googled a bit, but have not found anything Jan 12 17:16:47 mmp, it's in OE/bitbake bugtracker, pls search there Jan 12 17:17:04 hey all Jan 12 17:17:17 psokolovsky: thanks, I thought google will find it there... Jan 12 17:28:28 hmm, should switch to the new 1.4 release of genext2fs Jan 12 17:29:17 kergoth: that supports mind control? Jan 12 17:30:39 Is it possible to specify an md5sum for a SRC_URI? Jan 12 17:30:53 apparently it has the old option to determine how many inodes will be needed again.. doesnt have the block one for good reason, too tough to be accurate without creating the fs twice Jan 12 17:34:37 Laibsch: iirc you can do ;md5sum= for each entry Jan 12 17:34:44 Laibsch: zecke knows :) Jan 12 17:34:59 koen: OK, I will try that. Jan 12 17:36:32 weird; now it doesn't fetch anything... . Will have look on it later, now I need to learn; bye:) Jan 12 17:36:37 g'day kergoth Jan 12 17:36:48 hey pb__ Jan 12 17:36:49 koen: Found a sample: packages/libvncserver/libvncserver_0.8.2.bb: Jan 12 17:37:10 iirc qte had a md5sum as well Jan 12 17:37:22 till trolltech altered the tarball upstream Jan 12 17:42:47 it especially sucks when the downloads/ contains a partial file, IARC Jan 12 17:43:47 likewise: hey Jan 12 17:43:56 likewise: Any response from coolo? Jan 12 17:45:08 Laibsch: basically, he is the sole developer on icecream. We can ask him directly for issues/help. Jan 12 17:45:25 Laibsch: I can forward his response to you. (/msg me your email addr) Jan 12 17:46:34 Did you ask him specifically about how to enable logging for the client yet? Jan 12 17:49:25 funny Jan 12 17:49:48 someone tried to explain me that 'zaurus' uses ipkg Jan 12 17:51:55 lol Jan 12 17:52:22 03kristoffer 07org.oe.dev * rfa0c91a5... 10/ (1 packages/parted/parted_1.8.2.bb): Jan 12 17:52:22 packages/parted/parted_1.8.2.bb : New parted Version Jan 12 17:52:22 * Update to newest parted. Includes many bugfixes Jan 12 17:52:22 and improvments. Jan 12 17:52:24 "how interesting! tell me more! i' Jan 12 17:52:26 He got all pissy when I tried to nuance that statement Jan 12 17:52:31 d love to hear more on this subject!" Jan 12 17:52:57 Parted just got released 20mins ago :D personal record for getting stuff into OE Jan 12 17:53:09 Kristoffer: hehe Jan 12 17:53:10 Kristoffer: :) Jan 12 17:53:24 Kristoffer: usually we track abiword pretty well Jan 12 17:53:33 :) Jan 12 17:53:34 re Jan 12 17:53:39 since the abi people upload is a few days before the release Jan 12 17:53:45 hrw: wb Jan 12 17:53:56 koen: I remember when someone added software before release ;) Jan 12 17:54:00 food.. Jan 12 17:54:02 hehe Jan 12 17:54:30 Im hoping to get this glibc 2.5 + parted + arm issue solved, or atleast get a nice backtrace (that just got implemented in 1.8.2) Jan 12 17:56:09 likewise: Did you ask him specifically about how to enable logging for the client yet? Jan 12 17:56:39 I had gparted working a while ago Jan 12 17:57:50 koen: puh, so much kiddie chit chat @ #openmoko... Jan 12 17:58:19 greentux: and the mailinglist Jan 12 18:00:15 koen: it never ceases to amuse me that half the time oe has stuff before the major distros do Jan 12 18:00:42 kergoth: mostly because we don't do QA Jan 12 18:01:36 koen: i read the ml only as digest Jan 12 18:01:56 greentux: command-click->mark as read Jan 12 18:02:00 greentux: that's what I do Jan 12 18:02:18 kergoth: that and because OE builds *and* packages Jan 12 18:02:44 true Jan 12 18:03:34 * greentux ordered today a server to provide oe (building etc) Jan 12 18:03:59 right, I should contact Gerwinin about the source mirror Jan 12 18:04:46 *lightbulb* Jan 12 18:04:59 I'll be working near gerwininin in ~2 weeks Jan 12 18:05:09 with the company car Jan 12 18:10:42 and I have to finally publish my source mirror work Jan 12 18:25:39 likewise: ping? Got your package :) Jan 12 18:25:58 Thanks a bunch Jan 12 18:26:07 Is there anything on that bios chip? Jan 12 18:26:37 NAiL: yeah, all zeroes and ones (but nothing in a particular order, that I know of) Jan 12 18:26:45 hehe, ok Jan 12 18:27:13 likewise: you men 'crypted message'? Jan 12 18:27:18 mean Jan 12 18:28:56 NAiL: no, I think that BIOS chip comes from the FlashROM manufacturer and thus contains a bunch of ones or zeroes, or both. Of course it could be an encrypted msg Jan 12 18:30:28 koen: company car? which company? :-) Jan 12 18:30:43 does openembedded.org has a car? cool :-) Jan 12 18:31:31 likewise: I think you meant to direct that last message to hrw ;) Jan 12 18:32:09 Laibsch: sorry, no I did not ask him that question, could you follow-up on that email and CC me? Jan 12 18:32:35 hrw: see my response to NAiL above, which was meant for you. Jan 12 18:32:59 likewise, do you have a preferes picopsu? Jan 12 18:33:17 likewise: sure Jan 12 18:33:18 Crofton__: The 120 Watts version, yes. Jan 12 18:33:27 ok Jan 12 18:33:29 4? Jan 12 18:33:42 for the $35 price? Jan 12 18:34:00 sure, I guess I can make koen happy with one as well :-) Jan 12 18:34:05 do you have a number, or $ limit? Jan 12 18:34:27 it sounds like I can easily bring 175 euros of "gifts" Jan 12 18:35:06 When will you know if you use the other source you mentioned? Jan 12 18:35:07 Crofton__: Well, I need one or two for myself (EFIKA and one Epia), but I'll give one to koen for his EFIKA. Jan 12 18:36:27 Crofton__: Never mind that other source (my colleagues), they are running around at the CES and then some company visits. Too busy. Also, the EUR<->USD conversion is beneficial for both of us. Jan 12 18:36:42 yeah Jan 12 18:36:50 ok Jan 12 18:37:06 I want to get one of the wide iinput range versions for myself Jan 12 18:37:27 speaking of wide.. I want wide lcd panel Jan 12 18:37:27 I'll go ahead and start paying attention to ebay then for the 120watters also Jan 12 18:37:36 Crofton__: 4 of $35 or 3 of $50 would be ok, all 120 Watts. The ebay seller looks good ($35), 117 recommendations, no negative responses. It seems a small company. Jan 12 18:37:44 ok Jan 12 18:39:04 my email addr is leonw # mailcan # com Jan 12 18:39:55 ok Jan 12 18:41:23 I've got the page book marked. I'll go ahead and do some ordering over the next few days Jan 12 18:41:38 the same guy also has an auction for the 60 I am interested in Jan 12 18:42:31 Crofton__: Thanks for any trouble and advance payment. I'm wondering if these picoPSU are suited for in-car use (as car electronics voltages can have spikes etc.) Jan 12 18:43:26 Crofton__: On the other hand, for my EFIKA-pretends-to-be-a-CD-changer-project, the CD changer would expect a neat voltage as well. Jan 12 18:44:46 I think they sell special version for cars, designed to survive cars Jan 12 18:44:53 <[cc]smart> koen: can you check OE for me if my key still has write access ? Jan 12 18:50:33 likewise, http://www.mini-box.com/M1-ATX-90w-Intelligent-Automotive-DC-DC-Power-Supply;jsessionid=ac112b1b1f430fc61f852f94428ab9cfd5cf74d95636.e3eTaxmKbNaNe34Kb38Nbh0Pahj0n6jAmljGr5XDqQLvpAe?sc=8&category=13 Jan 12 18:53:12 Crofton: thanks, but I'll look to the 12 Volts line to my CD changer first, to see if it is regulated by my car stereo head unit (this is what I expect). Jan 12 19:14:10 http://www.hrw.one.pl/2007/01/12/zaurus-line-officially-ended/ Jan 12 19:17:37 hrw: heh, good riddance. sharp's ineptitude will bother us no longer Jan 12 19:18:15 o yes Jan 12 19:19:37 XorA|gone: Are you sure /usr/include needs to be replaced and not /usr/local/include? Jan 12 19:25:35 hi Jan 12 19:26:37 at the linkstation wiki we want to import openlink distro into OE Jan 12 19:27:02 openlink is yet-another-oe-distro? Jan 12 19:27:17 timtimred: which machines you cover? Jan 12 19:27:34 atm it seems to be maintained as a big binary blob. (shock, horror, etc) Jan 12 19:27:57 au.. Jan 12 19:28:15 "Why it is worth to choose OE when there are other solutions? For me OE allows for quick start into embedded world without going through horror way of cross-compilation which is quite important in situation when we want to concentrate on out device development. After creating few simple config files we can build working system, use ready recipes to build needed tools (no need to search though internet for them as they are rather added into OE already) and tell Jan 12 19:28:23 we started (well, i started) adding it to buildroot but have been looking at OE (i think i talked with you or someone about this in #openmoko already) Jan 12 19:28:28 its my favorite quote last days Jan 12 19:28:50 initially it will be for MIPS and PPC linkstations of different varieties Jan 12 19:29:22 but now we (well, probably I) am gonna get cracking with it officially Jan 12 19:29:24 mipsel has some problems but they are solvable Jan 12 19:30:09 right. bring it on :) Jan 12 19:30:19 ppc is now supported more and more as few devels got powerpc boards recently ;) Jan 12 19:31:14 we will be supporting ARM too since they moved the LSPros to thaT i guess thinking about it Jan 12 19:31:26 timtimred: you have 2.6 kernel working on devices which you target? Jan 12 19:31:46 yes on the ppc side, and our team at the lswiki ported uboot too Jan 12 19:31:58 +kurobox Jan 12 19:32:00 people Jan 12 19:32:45 I suggest using DISTRO="angstrom-2007.1" then - it is good supported and we plan to merge all OE-based distros in it Jan 12 19:32:46 we have patches for vanilla 2.6 Jan 12 19:33:12 good Jan 12 19:33:42 i heard a rumour they might be going upstream. but i didnt take it seriously Jan 12 19:33:50 ;) Jan 12 19:34:38 likewise, ok, I will work on acquiring some of the 120w parts, they should provide loads of fun regardless Jan 12 19:35:00 Crofton: thanks! Jan 12 19:35:12 np Jan 12 19:35:28 we can exchange "gifts" at fosdem :) Jan 12 19:35:32 Crofton: At *least* I can get rid of some excessive cabling: http://www.sidebranch.com/openembedded/efika_senna.jpg Jan 12 19:35:40 will you be at the friday "beer event"? Jan 12 19:35:58 I have a similar cat issue Jan 12 19:35:59 Crofton: sure, although I've not read about it yet :-) Jan 12 19:36:30 Laibsch: yes, no cross compiled program should access /usr/* at anytime ever Jan 12 19:36:35 http://www.fosdem.org/2007/category/info Jan 12 19:36:35 Crofton: this is the picture where she decided to lie down *next to* the EFIKA. Jan 12 19:36:38 likewise: heh.. its Maestro card not MasterCard ;( Jan 12 19:36:56 hrw: check ;-) Jan 12 19:37:03 as opposed to on top .... Jan 12 19:37:37 too blurry part Jan 12 19:38:00 ok. I go back to mail Jan 12 19:39:43 hrw: so from now the zaurus devices will become more valuable... Jan 12 19:40:11 I see euro geeks have pringle'stoo Jan 12 19:42:09 greentux: for some time maybe yes. some of people will move to newer pocketpc machines which would get linux support or to openmoko phone or other ways Jan 12 19:42:29 greentux: my friend is building own ARM devboard at home Jan 12 19:42:57 http://www.balloonboard.org/ Jan 12 19:43:23 timtimred: where's the current repository for openlink stuff? http://www.kurobox.info/bbs/index.php?showtopic=750 Jan 12 19:43:56 hrw: own arm devboard. are the not enough dev boards around? Jan 12 19:43:57 Crofton: yes, it's the LART successor. Jan 12 19:44:24 theres a buildroot svn hosted offsite thats a good, ohh 50% not finished Jan 12 19:44:38 hrw: the balloonboard? looks interesting Jan 12 19:44:49 otherwise all the rest of the stuff is at http://www.linkstationwiki.net Jan 12 19:45:30 um, which appears to be down! Jan 12 19:45:50 likewise: http://forum.linkstationwiki.net/index.php?action=vthread&forum=1&topic=2108 Jan 12 19:46:07 I liked this one as well: http://www.kurobox.info/bbs/index.php?showtopic=750 Jan 12 19:46:10 okay its back up again Jan 12 19:46:24 greentux: he will base on some atmel kits Jan 12 19:46:50 "we will move from buildroot to openembedded. the main difference is that after integrating support for the bunch of different boxes we will instantly have something that we were longing for so long....IPKG-Feeds!" - that sounds nice Jan 12 19:47:05 I have hope that this year some linux based (mostly open) pda/smartphones will appear Jan 12 19:47:56 ipkg feeds... someone else can do that :) Jan 12 19:48:42 we hang out in #linkstationwiki btw Jan 12 19:49:31 timtimred: openprotium is a NAS distro for OE (it's integrated with the upstream openembedded.org sources in the monotone repository, in the .dev branch). It started with PPC target NAS devices. It may be your best starting point, or even better, merge with that project. Jan 12 19:49:58 ahha, good stuff thanks Jan 12 19:50:23 timtimred: one of it's developers (Andy) sometimes is online as 'HopsNBarley'. Jan 12 19:50:31 ballon board 2.05 (sa1100 based) is 566 EUR.... Jan 12 19:51:00 I would rather buy few h2200 devices for that cash ;D Jan 12 19:51:03 great. i'll keep a lookout. most of this weekend i'll expect to be reading docs. Jan 12 19:51:15 hrw: I have 2 sa1100 devices here. Jan 12 19:51:17 and getting started Jan 12 19:51:39 likewise: I have one - called collie Jan 12 19:51:54 hrw: I meant 'chips' (non-soldered) Jan 12 19:52:01 likewise: ah. Jan 12 19:52:06 wb koen Jan 12 19:52:46 timtimred: OE might have a steep learning curve, but please don't be put of by it (and ask here if you do not understand some parts of the docs) Jan 12 19:53:15 timtimred: to know how OE works select one of working combos for first builds. I suggest MACHINE="h2200" + DISTRO="angstrom-2007.1" - then 'bitbake angstrom-bootstrap-image' (left for night?) will show you how OE works Jan 12 19:53:29 but the sa1100 is not really a recommended chip to start new designs with. :-) Jan 12 19:54:03 cool, ive used buildroot heavily if that helps me Jan 12 19:54:34 timtimred: or use angstrom-2007.1-oabi as distro as your arm devices are 2.6.12 based so normal angstrom would not support them Jan 12 19:55:14 timtimred: it helps a bit, OE is very different in its concepts, but it very much helps if you understand what all the dependencies are. I came from buildroot too. Jan 12 19:55:29 all this is good stuff, i'll save this chat :) Jan 12 19:55:49 timtimred: btw, who chose the picture on the wiki? That seems some Dutch water channel there :-) Jan 12 19:56:04 not me!!! heh Jan 12 19:56:40 * hrw skipped buildroot Jan 12 19:56:41 Maybe koen recognized this picture? http://linkstationwiki.net/index.php?title=Category:OpenLink Jan 12 19:56:49 s/zed/zes Jan 12 19:57:08 if I would buy Zaurus half year earlier then I would get buildroot knowledge too Jan 12 19:57:09 hey likewise Jan 12 19:57:28 hey koen, recognize that piece of dutch channel on that wiki page (a few lines up)? Jan 12 19:57:36 yep Jan 12 19:57:38 my buildroot stems from gumstix and starting to import openlink for ppc into it Jan 12 19:57:40 been there Jan 12 19:58:25 timtimred: ppc uclibc or glibc? Jan 12 19:58:30 glibc Jan 12 19:58:39 timtimred: we could use some ppc people :-) Jan 12 19:58:51 it had some patches to use buildroot Jan 12 19:59:03 i cant remember if crosstool support was patched in or if it supported it Jan 12 19:59:17 it had some patches to use buildroot Jan 12 19:59:22 = glibc not buildroot Jan 12 19:59:38 timtimred: OE builds its toolchain(s) itself Jan 12 20:00:10 timtimred: glibc 2.5 from OE works on x86/arm/ppc targets Jan 12 20:00:22 likewise, i will be doing ppc first since i dont own a mipsel Jan 12 20:00:31 imo, all it really takes to grasp oe conceptually is an understanding that all of the information about building and packaging and deploying a package is just metadata, and that oe consolodates that metadata into global configuration metadata and classes of common build steps Jan 12 20:00:45 and kuro support as well, which is 99% the same hardware as a HG linkstation Jan 12 20:01:42 aye kergoth Jan 12 20:02:37 oh yeah, 2.6 for the mipsel boxes isnt ported Jan 12 20:04:06 thanks for the advice and stuff. im gonna go RTFM. Jan 12 20:04:10 :) Jan 12 20:04:15 ok, forget the openmoko channel today Jan 12 20:04:27 someone remember find option to list files 'not touched in last xyz second'? Jan 12 20:06:06 hrw: some combo of -atime and negation (!)... Jan 12 20:06:54 ok Jan 12 20:08:14 hrw: ah 'modified' you mean? http://www.grymoire.com/Unix/Find.html#uh-8 Jan 12 20:12:59 "It is difficult to find directories that have not been accessed because the find command modifies the directory's access time." Jan 12 20:13:02 fuck Jan 12 20:13:10 heh Jan 12 20:13:37 pb_: yes, I know - I'm lame when it comes to such solutions Jan 12 20:14:03 well, no, it's more that unix is lame in that respect Jan 12 20:14:05 hrw: but do you mean 'accessed' (read/write) or 'modified' (write) Jan 12 20:14:16 Why don't we have http://www.pengutronix.de/software/utelnetd/ in OE yet? Jan 12 20:14:35 * likewise on the job... Jan 12 20:14:35 likewise: read even Jan 12 20:14:38 no one usees telnet anymore? Jan 12 20:14:54 well, except for tivo Jan 12 20:14:55 for dev purposes it's ok Jan 12 20:15:00 likewise: because we have dropbear? Jan 12 20:15:16 likewise: I created source mirror for celf and need to write script which will remove archives not used in last X days Jan 12 20:15:37 heisenberg is against you hrw Jan 12 20:15:38 hrw: aha, so indeed access time... Jan 12 20:15:45 Crofton: exactly :-) Jan 12 20:16:45 likewise: in week 4 I'll be in south-east Brabant (wherever that may be) to measure how fat police agents are Jan 12 20:18:24 so I can bother you, gerwininin and the irex dudes :) Jan 12 20:18:31 koen: haha, my neighbour is one (marechaussee on a 1600cc motorcycle) Jan 12 20:18:49 koen when you will go in the USA? Jan 12 20:18:49 koen: he has gained some weight doing more paperwork lately :-) Jan 12 20:19:14 chouimat|busy: very maybe april Jan 12 20:19:26 koen lot of fat policemen there Jan 12 20:19:31 likewise: 45% of the 30k people we measured was overweight Jan 12 20:19:31 koen: for what purpose, side job? Jan 12 20:19:40 likewise: yep side job Jan 12 20:19:50 Brabant usa? Jan 12 20:19:51 likewise: http://bsst.nl/ Jan 12 20:20:01 Crofton: no, the original brabant :) Jan 12 20:20:21 Crofton: http://maps.google.com/?ie=UTF8&z=9&ll=51.44716,5.888672&spn=0.83362,2.713623&t=h&om=1 Jan 12 20:20:23 you are fascinated with body armor Jan 12 20:20:54 good evening Jan 12 20:21:02 hey tomas_f ! Jan 12 20:21:10 Crofton: it pays for my toys :) Jan 12 20:21:32 koen: you're coming over to this side of the pond? Jan 12 20:21:54 hi mreimer Jan 12 20:22:10 hi pb_. did you hear that mmp got h5xxx working again? Jan 12 20:22:38 no, I didn't hear that Jan 12 20:22:40 sounds awesome Jan 12 20:22:46 all hail mmp Jan 12 20:22:57 mreimer: I'd like to attend the CELF thing in santa clara Jan 12 20:23:02 pb_: the workaround was to not map the framebuffer (registers?) as writecombine Jan 12 20:23:16 koen: If you want, we can talk about hiring you for OpenEmbedded work, if you can spend some hours to visit Axon. Jan 12 20:23:16 mreimer: and I got invited for a conference in brazil :) Jan 12 20:23:32 koen: cool! Santa Clara's a bit of a drive from here, maybe 15 hours :-( Jan 12 20:23:39 03pfalcon 07org.oe.dev * rdd587ca4... 10/ (1 conf/machine/htcblueangel.conf): htcblueangel: Specify known good kernel version. Jan 12 20:23:43 03pfalcon 07org.oe.dev * r8e8bab6f... 10/ (1 conf/machine/rx3000.conf): rx3000.conf: Specify known good kernel version. Jan 12 20:23:58 likewise: I just have enough time to do some work before the next term starts (20070204) :( Jan 12 20:24:14 mreimer: aha. Jan 12 20:24:34 koen: ok, never mind. is more important anyways :-) Jan 12 20:24:35 pb_: interesting that it didn't break h2200, which has a similar mediaq chip Jan 12 20:24:58 XorA|gone: /usr/local/include *is* /usr! And it was really that /usr/local/include needed to replaced, rather than /usr/include. There just was no occurrence of /usr/include to replace. Jan 12 20:25:09 mreimer: http://rafb.net/p/exfHcL35.html Jan 12 20:25:26 mreimer: yeah, that is a bit odd Jan 12 20:25:32 still, great news that it's fixed Jan 12 20:25:58 I look forward to trying it out Jan 12 20:26:30 Now I am stuck at the next step: http://rafb.net/p/nRPNbT74.html "ELF file OS ABI invalid" seem to be the likely culprit but I don't know how to proceed from there. Jan 12 20:26:35 where is ${DEBIAN_MIRROR} defined? it seems to be set wrong, is resolving to http://ftp//ftp.debia.... which obviosuly doesnt work Jan 12 20:26:40 koen: those pango times look pretty good Jan 12 20:26:42 bbiab Jan 12 20:26:42 Still trying to compile perl for Sharp ROM Jan 12 20:27:13 my distro shold be set to openzaurus-unstable right? Jan 12 20:28:35 mreimer: I was looking at the raw fb speeds, 7.5MB/s vs 30MB/s Jan 12 20:28:47 koen: any time to spend at all, in the evening say? Is Nynke working/coming along? Jan 12 20:29:31 nynke is neck deep into her bachelor thesis labwork Jan 12 20:29:59 likewise: 8-17 is for bsst, the rest of the day is for me :) Jan 12 20:30:12 likewise: I hope I'll be close to eindhoven Jan 12 20:32:35 Ok, I will propagandize OE at http://www.bossaconference.org/ :D Jan 12 20:34:25 koen: cool Jan 12 20:35:20 I'm thinking about creating sweatshirt with OE logo on backs Jan 12 20:35:39 for sale at fosdem? Jan 12 20:36:02 Crofton__: to wear on conferences rather then to sell Jan 12 20:36:25 get a tatto on your forehead :) Jan 12 20:36:34 http://www.spreadshirt.net/index.3.0.html?op=producttype_detail&producttype_id=5 for example Jan 12 20:36:41 20 EUR + post Jan 12 20:39:03 Can someone tell what are the incorrect rpath arguments in this build error? http://www.pastebin.ca/314544 Jan 12 20:40:54 This seems related: http://lists.xiph.org/pipermail/flac/2006-April/000559.html Jan 12 20:41:02 hmm, libthread belongs to what package? Jan 12 20:41:09 hrw: perhaps ypu can design one and everybody can order one, if he wants Jan 12 20:41:59 greentux: I know - but I lack flash here to finish it. Jan 12 20:42:31 hrw: perhaps some other shop? Jan 12 20:42:41 greentux: thats an option too Jan 12 20:43:05 greentux: and need to check how will logo look on !white backgrounds Jan 12 20:43:54 hrw: on the website its also white background Jan 12 20:44:07 greentux: I prefer dark clothes Jan 12 20:44:28 hrw: I remember, hacker prefer dark clothes everytime ;) Jan 12 20:44:33 ;) Jan 12 20:44:55 hrw: who painted that logo? Jan 12 20:47:43 quitters Jan 12 20:48:14 perhaps we can make some shirts, some white, some black. Jan 12 20:50:17 greentux: do not remember Jan 12 20:50:28 Fetch failed: http://ftp://ftp.debian.org/debian/pool/main/p/portmap/portmap_5-25.diff.gz;patch=1 Jan 12 20:50:37 hrw: has somebody the logo in a nice file format? Jan 12 20:50:59 * greentux goes to tv for some minutes Jan 12 20:51:26 greentux: contrib/oe.svg was created by kergoth 26 months ago Jan 12 21:00:21 wow i didnt know that, gmake 3.80 got "order only" prerequisites Jan 12 21:00:32 greentux: scanline, aka micah dowty, did the oe logo Jan 12 21:01:01 order only prerequisites are prereqs where you depend only on its existance, not on its modification date Jan 12 21:01:07 which was /always/ one of my make gripes Jan 12 21:04:29 oh, very good Jan 12 21:06:46 psokolovsky: htcblueangel need tweak to angstrom config. probably other linux-handhelds 2.6 targets too as 2.6.16-hh9 is selected by default Jan 12 21:07:09 koen: the reason hx4700 is faster is that the w3220 is on a 32-bit bus rather than a 16-bit bus, and its memory bus is 200MHz rather than 100MHz. So four times faster is just about right. Jan 12 21:07:26 mreimer: I suspected that Jan 12 21:08:04 koen: any luck getting Ximageon working on hx4700? Jan 12 21:08:17 mreimer: not yet, but sirfred was hopefull Jan 12 21:08:29 mreimer: he forgot to copy the w3220 bits Jan 12 21:08:39 hrw, I added that tweak to machine config. I don't want to add anything to angstrom config w/o approval ;-). but you can build htcblueangel against angstrom now Jan 12 21:08:43 koen: I hope it's that simple Jan 12 21:08:50 * mreimer wishes he had more spare time Jan 12 21:08:54 psokolovsky: It just failed with 2.6.16-hh9 Jan 12 21:09:08 * koen should update his buildscripts Jan 12 21:09:21 hrw, yeah, wasn't any working code for BA at that time ;-) Jan 12 21:09:42 hrw, did you try with my recent updates? Jan 12 21:09:55 psokolovsky: moment Jan 12 21:10:41 hrw, actually, pls let me few more days to settle that. though if you have time, your trying it is appreciated, of course Jan 12 21:11:14 sure Jan 12 21:11:45 koen: can you do: PREFERRED_VERSION_linux-handhelds-2.6 ?= "2.6.16-hh9" Jan 12 21:12:17 koen: ? this way machines can prefer own kernel versions and we will fix versions before release Jan 12 21:13:48 NOTE: package linux-handhelds-2.6-2.6.19-hh8: started Jan 12 21:13:54 for htcblueangel Jan 12 21:13:58 hrw: sure Jan 12 21:14:08 * koen updates the repo on his laptop Jan 12 21:14:11 koen: thx Jan 12 21:15:14 first build which will test my mirror rewriterules Jan 12 21:16:10 oh dear Jan 12 21:16:21 the script kiddies have found #openmoko Jan 12 21:19:49 Want to build parted with full debugging symbols Jan 12 21:19:59 does it strip by default, since it lacks them now Jan 12 21:20:12 Kristoffer: ipkg install parted-dbg Jan 12 21:22:03 brilliant Jan 12 21:22:20 hrw: I'll add ?= for the other machines as well Jan 12 21:22:26 hrw: till we get to a release Jan 12 21:22:41 koen: who maintains the bug database? We could use the powerpc arch there. Jan 12 21:23:33 likewise: check again :) Jan 12 21:23:43 koen: I must have been blind then :-) Jan 12 21:23:57 no, I just added it ;) Jan 12 21:24:04 haha Jan 12 21:24:06 yes, it's there right at the top :-) (I knew) Jan 12 21:24:55 OpenEmbedded sets new record in customer support turn-around time. Jan 12 21:25:16 damn, should have enabled time stamps in my irc client Jan 12 21:25:17 greentux: http://openembedded.org/filebrowser/org.openembedded.dev/contrib/oe.svg Jan 12 21:25:59 what should be else on the shirt? Jan 12 21:26:00 greentux: slogans would need to be chosen carefully Jan 12 21:26:19 I proposed a slogan for fosdem, let me check. Jan 12 21:26:26 'OpenEmbedded: where Embedded meets Opensource' is one of my favorites Jan 12 21:26:27 bake it so! Jan 12 21:26:30 to be discussed, but we have not so much time... Jan 12 21:27:07 let's rename OpenEmbedded to EasyEmbedded Jan 12 21:27:23 Easy In Bedded? Jan 12 21:27:36 "bake it so".. hahah Jan 12 21:27:45 mmmh Jan 12 21:27:48 joshin: one step more: Easy in Dad Bed Jan 12 21:28:20 OpenEmbedded Jan 12 21:28:21 EmbeddedFluff Jan 12 21:28:21 Bake It So. Jan 12 21:28:39 Topic is now: www.EasyInDadBed.org - Who's your daddy now? Jan 12 21:28:52 koen: you beat me again, sigh. Time to pop a beer. Jan 12 21:29:05 "just bake it!" Jan 12 21:29:12 hmmm Jan 12 21:29:20 beer or hot chocolate + rum Jan 12 21:29:37 or hot bed Jan 12 21:29:52 or bed + hot chick Jan 12 21:30:11 "BE ODD PEED MEN" which is an anagram of... Jan 12 21:30:13 it's probably be a hot beer... Jan 12 21:30:15 OE - The hotbed for embedded Jan 12 21:30:21 hot chocolate + rum + bed + hot chick Jan 12 21:30:29 koen: not bad! Jan 12 21:30:50 OE - gets you hot chicks embedded Jan 12 21:31:18 OE where size doesnt matter Jan 12 21:31:21 :D Jan 12 21:32:02 does the openprotium guy appear here much? Jan 12 21:32:12 ~seen hopsnbarly Jan 12 21:32:23 ~seen hopsnbarley Jan 12 21:32:39 ~seen hopsnbarley Jan 12 21:32:41 i haven't seen 'hopsnbarly', koen Jan 12 21:32:43 ibot is being slow again Jan 12 21:32:46 hopsnbarley was last seen on IRC in channel #oe, 8d 54m 18s ago, saying: 'i've got a 5.x building, but had to change the recipie a bit. was hoping to bounce it off somebody.'. Jan 12 21:33:00 okay ta :) Jan 12 21:33:08 cu guys Jan 12 21:33:11 cu hrw Jan 12 21:33:37 simple comfort shirt 15,90 Jan 12 21:33:40 euros Jan 12 21:33:49 psokolovsky: NOTE: package linux-handhelds-2.6-2.6.19-hh7-r0: task do_build: completed Jan 12 21:33:57 psokolovsky: htcblueangel Jan 12 21:34:08 hrw|gone, nice! Jan 12 21:34:14 I offer code bounty for first OE chick. Jan 12 21:34:32 shit.. wrong screen Jan 12 21:34:44 psokolovsky: sorry - Ba is building still.. Jan 12 21:34:50 psokolovsky: track tinderbox please Jan 12 21:35:19 hrw|gone, ok, will start to... Jan 12 21:36:08 Hmmm, if -rpath is not supplied to the linker, how can i get a "bad -rpath option" error? Is there some set of built-in options in the linker? Jan 12 21:40:00 likewise: I believe I have icecc daemon logging in to the scheduler now: 22:45:29: login X24 protocol version: 25 [] Jan 12 21:40:00 likewise: it seems that it is necessary to set the pingable net adress for ICECC_NETNAME. Jan 12 21:40:41 Laibsch: that's new to me, are they on the same subnet? Jan 12 21:40:50 But now all of the sudden, the OE jobs do not turn up in icecc_scheduler.log anymore. They did previously, althought they where never distributed. Jan 12 21:40:55 perhaps we can collect some t-shirt wishes and distribute it @fosdem Jan 12 21:41:07 Laibsch: hmmmn. Jan 12 21:41:12 Laibsch: At least that fixed it for me. Both computers are not logged into the scheduler. Jan 12 21:41:24 Both are on 192.168.1.x Jan 12 21:41:35 No firewalls. Jan 12 21:42:35 likewise: "bad -rpath option" can be caused by things like -rpath-link as well as -rpath itself. Jan 12 21:42:56 in fact, a correctly formed -rpath option will (afaik) never yield that error. Jan 12 21:43:07 likewise: Both computers are *now* logged into the scheduler. Jan 12 21:43:17 pb_: I have no -rpath* option in the argument, see #1775. Jan 12 21:43:35 Laibsch: good. that's your starting point. Jan 12 21:44:03 likewise: except the -rpath-link that you have, right? Jan 12 21:44:18 though, admittedly, there doesn't seem to be anything obviously wrong with it. Jan 12 21:44:40 likewise: I'd say I am one step closer, but I don't know how many more steps and how difficult. Jan 12 21:44:48 pb_: see the second comment of that bug (where I call collect2 manually). Jan 12 21:45:03 Still no logs and no distribution of tasks. Jan 12 21:46:01 likewise: oh, right, you have some "-read-only-relocs" thing. Jan 12 21:46:05 that would probably cause it too Jan 12 21:46:08 Laibsch: are you sure the ICECC compiler is called at all. Jan 12 21:46:27 I wonder where that's coming from. gcc specs maybe. Jan 12 21:46:52 pb_: it occurs only in the powerpc build for angstrom, not in the arm ones. Jan 12 21:46:58 odd Jan 12 21:47:04 if you take that out, does the error go away? Jan 12 21:48:17 likewise: I made one call outside of OE for a small program "make clean;make CC=icecc -j10". That is what shows up now in http://oz.leggewie.org/wip/icecc_scheduler.log Jan 12 21:48:31 pb_: yes Jan 12 21:49:34 * likewise is googling for -read-only-relocs Jan 12 21:49:41 Could someone check this bug: http://bugs.openembedded.org/show_bug.cgi?id=1784 Jan 12 21:49:44 looks like libflac's Makefile.am is adding that flag, though in the copy I have here it seems to be marked TEMPORARILY_DISABLED Jan 12 21:49:58 if FLaC__HAS_AS__TEMPORARILY_DISABLED Jan 12 21:49:58 libFLAC_la_LIBADD = ppc/as/libFLAC-asm.la Jan 12 21:49:58 LOCAL_EXTRA_LDFLAGS = "-Wl,-read_only_relocs,warning" Jan 12 21:49:58 else Jan 12 21:50:02 [...] Jan 12 21:50:16 funky stuff Jan 12 21:51:09 B_Lizzard: hard to say much based on the information in that bug report. have you tried running abiword under gdb to find out where the abort occurs? Jan 12 21:51:25 Will do Jan 12 21:51:40 I just wanted to see if anyone has ever experienced something like that before Jan 12 21:51:47 Output is very....poor Jan 12 21:52:04 I've never heard of it, but I suspect you might be one of the first people ever to try abiword on sh3. Jan 12 21:52:19 I think italics worked before Jan 12 21:52:25 Like, 2 months ago Jan 12 21:52:33 ah right Jan 12 21:52:47 if you could isolate the point where it broke, that might be useful too Jan 12 21:53:10 If it helps, I found a post on the Knoppix forums from someone that replicated the problem Jan 12 21:53:15 Unintentionally Jan 12 21:53:35 So it's probably not arch-specific Jan 12 21:53:48 I *could* try an older version of Abiword Jan 12 21:54:23 pb_: thanks for that hint of help, would take me hours to get to this point. Jan 12 22:00:57 pb_: the ARM build for 'flac' indeed does not build with the -read-only-relocs option. Jan 12 22:06:17 likewise: I believe jobs are distributing now from OE: http://rafb.net/p/NHQI4D32.html. X24 is my non-OE, non-scheduler client computer. Jan 12 22:07:18 Laibsch: if you can build the icecc monitor (uses kde libs) you truely get an very attractive and enjoyable graphic flow of your builds being distributed. Jan 12 22:07:43 likewise: icemon? Jan 12 22:08:07 That is available as a package, but I thought it would not work with more recent versions of icecc? Jan 12 22:08:09 Laibsch: now look at the "status=" results in your scheduler log, it shows the grand total result of each build. 0=success Jan 12 22:08:34 http://oz.leggewie.org/wip/icecc_scheduler.log Jan 12 22:08:34 Laibsch: no, probably not. We should ask coolo. Jan 12 22:09:03 what supplies libdl.so.2 Jan 12 22:09:09 Trying to backtrace parted bug Jan 12 22:09:34 Nice, getting more horsepower for OE, finally! Jan 12 22:09:40 * koen envions 3 racks with 100 nodes + scheduler + big ass nfs server Jan 12 22:09:47 Laibsch: looking GOOD! Jan 12 22:10:26 Laibsch: we await your summary report of what version you used, etc. :-) Jan 12 22:11:01 version? Jan 12 22:11:20 Kristoffer: isn't that a glibc package (dynamic linker support blabla)? Jan 12 22:11:42 likewise: icecc package is from ubuntu and version 0.7.14-3 Jan 12 22:11:55 Laibsch: you use the icecc.bb from current .dev branch unmodified? Jan 12 22:12:02 typical, now why the hell am I getting those ?? in dbg Jan 12 22:12:16 Laibsch: ubuntu edgy on i686, right? Jan 12 22:12:23 yes Jan 12 22:12:33 for the distro Jan 12 22:12:52 I don't recall whether or not I made changes. Jan 12 22:13:19 Laibsch: could you do a "mtn diff classes/" while in org.openembedded.dev? Jan 12 22:13:30 likewise: You mean classes/icecc.bbclass? Jan 12 22:13:39 yes, mtn diff was what I was going to do. Jan 12 22:14:48 " mtn diff classes/icecc.bbclass " says "no changes" Jan 12 22:15:58 Laibsch: the current version used the "icecc-create-env" script, I see. I don't know where it is fetched from, though. Jan 12 22:16:54 Laibsch: ok, ./packages/icecc-create-env/icecc-create-env_0.1.bb Jan 12 22:19:16 Laibsch: just to be sure, could you "find build/tmp -regex .*icecc-create-env.*" and see if you made changes there? Jan 12 22:19:30 * Laibsch updated oe.org/icecc Jan 12 22:22:57 Laibsch: ping Jan 12 22:23:38 likewise: I have the most recent version of icecc-create-env Jan 12 22:24:03 no diff Jan 12 22:24:37 Laibsch: ok, thanks. sorry for being paranoia, I just want to make sure you aren't working on a heavily modified version that is lost when you do a rm -rf build/tmp tomorrow :-) Jan 12 22:25:12 mtn diff won't detect that Jan 12 22:26:15 koen: I did a "diff $file1 $file2" after downloading from the SRC_URI manually Jan 12 22:26:37 Laibsch: excellent. Jan 12 22:26:48 http://bugs.openembedded.org/attachment.cgi?id=1032&action=view is what I use Jan 12 22:28:41 Doesnt look like ncurses-dbg contains debugging symbols for libncurses Jan 12 22:29:18 Laibsch: Ok, and that ones matches http://www.digital-opsis.com/openembedded/icecc-create-env-0.1.tar.gz, I have just confirmed. Jan 12 22:29:36 told you ;-) Jan 12 22:30:00 Laibsch: so it seems OE head now works with your instructions on i686. I suspect some amd64 breakage though, especially on multilib machines, but that's something I can test here. Jan 12 22:30:43 Laibsch: let us know if a 2-compile setting actually gives a benefit ;-) Jan 12 22:30:54 It does. Jan 12 22:31:07 Just because my main OE machine is my router. Jan 12 22:32:19 Via C3 600 MHz, good for saving energy and better than the P1 I used as a router about a year ago which I undersped to 100 MHz. Jan 12 22:34:30 Laibsch: you sure there are no /usr/include in that file, I think maybe you read the version that had already had sed run across it Jan 12 22:35:09 XorA|gone: I think so. Jan 12 22:35:29 Laibsch: why would it look for errno.h only in /usr/local/include?? Jan 12 22:35:30 And I come to a different error now. Jan 12 22:35:40 I don't know. Jan 12 22:36:24 I will comment out the sed line now and take a look at the file. Jan 12 22:38:54 Laibsch: my $linux_errno_h = -e '/usr/include/errno.h' ? Line 118, please stop smoking the Ganja and coding Jan 12 22:39:29 :-D Jan 12 22:40:36 koen: 1784 is an age old problem, Im suprised it doesnt already have a bug Jan 12 22:41:53 anyone come across a live RD-CMP2000R Jan 12 22:54:09 03likewise 07org.oe.dev * rb23b7ec5... 10/ (1 packages/flac/flac_1.1.2.bb): flac: Fix build for powerpc. Closes #1775. Jan 12 22:58:50 did anyone find a workaround for the esound svn problem? Jan 12 23:01:18 XorA: I meant as a temporary measure to be able to answer your last question/suspicion "I think maybe you read the version that had already had sed run across it" Jan 12 23:04:03 koen: gerwinin proposed a oe dinner in ehv (but discussion may also slide off to chicks). Jan 12 23:07:39 * likewise suspects koen is under influence of choco w/ rum already... Jan 12 23:12:20 likewise: sounds good, gimme some :-) Jan 12 23:12:40 XorA: only have instant choc, no instant rum :-) Jan 12 23:15:48 likewise: was the rum I wanted Jan 12 23:25:52 03xora 07org.oe.dev * r2295b07e... 10/ (1 packages/esound/esound-gpe_20070109.bb): packages/esound/esound-gpe_20070109.bb : fix svn checkout problems. Jan 12 23:48:15 XorA|sleep: it depends on the bitbake version wether or not the checkout of esound works :( Jan 12 23:48:55 likewise: A dinner sounds good, I'll give a sign when I have the info where I'll be staying during that week Jan 12 23:49:48 koen: okido Jan 12 23:51:03 "ZO brabant" is all the info I have right now Jan 12 23:56:37 http://i97.photobucket.com/albums/l214/Ymdar/Motivator/webmunchkin.jpg Jan 13 00:02:53 hi, does any one how to change /etc/inittab Jan 13 00:03:07 i have tried editing packages/sysvinit/sysvinit/inittab Jan 13 00:03:31 but it does not seem to be the one that is getting posted to the image Jan 13 00:03:35 hmm... Jan 13 00:03:42 do i need to touch the bb file ? Jan 13 00:16:23 koen: do you know the Munchkin card game? Jan 13 00:16:53 likewise: Can you briefly explain the load number meaning in icecc? Jan 13 00:16:59 0 = not loaded Jan 13 00:17:05 1000 = overburdened? Jan 13 00:17:26 Laibsch: it's some kind of cost calculation. the lowest cost machine wins the job. Jan 13 00:17:44 it also incorporates network latency etc, but is quite heuristic. Jan 13 00:17:50 Thing is my slow Via comp is always 0 Jan 13 00:18:05 And my faster laptop is always 1000 Jan 13 00:18:18 That is strange. Jan 13 00:18:26 Laibsch: hmmm, google for "icecc load" Jan 13 00:18:36 Laibsch: or email coolo :-) Jan 13 00:19:16 nothing on google. Jan 13 00:19:28 So if anything it's gonna be coolo Jan 13 00:25:22 Laibsch: "Load of 1000 means to not schedule" Jan 13 02:14:01 03pfalcon 07org.oe.dev * rd81c0c73... 10/ (1 packages/qte/qte-2.3.10/simpad.patch): Jan 13 02:14:01 qte 2.3.10: simpad.patch contains few hunks with white-space only changes, Jan 13 02:14:01 remove them, as they complicate review and further patching. Jan 13 02:14:04 03pfalcon 07org.oe.dev * rf6e563bc... 10/ (1 packages/qte/qte-common_2.3.10.inc): Jan 13 02:14:04 qte 2.3.10: Use QT_QWS_CUSTOM consistently. Jan 13 02:14:04 * QT_QWS_CUSTOM does *not* exist in qte 2.3.10 at all. It's hello from Jan 13 02:14:05 2.3.7-or-something era. Before killing it completely, let's apply it Jan 13 02:14:07 consistently to ease looking at config. QT_QWS_CUSTOM is referenced at Jan 13 02:14:09 all *once* in entire OPIE codebase, in opie-launcher, and if its defined, Jan 13 02:14:11 there will F11 key checked. Doing so not going to kill tosa. Jan 13 02:14:13 03pfalcon 07org.oe.dev * rd74ad2ec... 10/ (1 packages/qte/qte-common_2.3.10.inc): Jan 13 02:14:15 qte 2.3.10: QT_QWS_EZX: remove bogus define. Jan 13 02:14:17 * Such define is not referenced in qte codebase or any patches. Jan 13 02:14:21 03pfalcon 07org.oe.dev * r8b68732f... 10/ (5 files in 3 dirs): (log message trimmed) Jan 13 02:14:23 qte 2.3.10: Rethink hardware platform defines. Jan 13 02:14:25 * Let there be two major platforms, Zaurus and Wince. Let there be Jan 13 02:14:27 two defines, QT_QWS_SL5XXX and QT_QWS_IPAQ, to define generic device Jan 13 02:14:29 of the corresponding platform. Jan 13 02:14:31 * Let there be also possibility to specialize for specific device of the Jan 13 02:14:33 platform. But corresponding define must be always used bu the generic platform **** ENDING LOGGING AT Sat Jan 13 02:59:57 2007