**** BEGIN LOGGING AT Wed Mar 07 02:59:57 2007 Mar 07 04:18:09 03pfalcon 07org.oe.dev * rb7646efb... 10/ (5 files in 2 dirs): linux-handhelds-2.6 2.6.19: 2.6.19 moved to branch, trunk is 2.6.20 now. Mar 07 04:18:13 03pfalcon 07org.oe.dev * r190b6a69... 10/ (26 files in 13 dirs): Mar 07 04:18:13 linux-handhelds-2.6: Add 2.6.20-hh1 and 2.6.20-hh2. Mar 07 04:18:13 * defconfigs are copies of .19's for now. Mar 07 04:57:23 Crofton|home: What are you still doing up???? Mar 07 05:48:06 http://sourceware.org/ml/libc-alpha/2004-09/msg00166.html Mar 07 05:48:23 is this problem solved? Mar 07 06:13:31 hi, can anyone help me with a sharprom dev question? Rather... how to get the compiling to work properly? Mar 07 06:18:06 hi, can anyone help me with a sharprom dev question? Rather... how to get the compiling to work properly? Mar 07 06:23:39 Post what's not working on pastebin Mar 07 06:23:48 maybe it's not a sharprom specific issue? Mar 07 06:26:55 While I'm here.. has anyone had any trouble with bootsplash-native? It's make file has broken LDFLAGS elements.. points to $(PREFIX)/$(LIB) which ends up being staging/i686-linux instead of staging/i686-${distroname}-linux on mine Mar 07 06:51:29 ah of course Mar 07 06:51:37 bootsplash-native should depend on jpeg-native Mar 07 06:52:38 is this the case in a more recent checkout of org.openembedded.dev? Mar 07 07:08:13 what's the standard diff format people expect for patches submitted to oe bugzilla? Mar 07 07:10:13 bah I'll just use mtn in all it's slowness :0 Mar 07 07:10:32 i wish oe mtn allow to insert patch file freely Mar 07 07:11:07 jselect: diff -Nru (less the r if not recursive) Mar 07 07:11:43 thanks NAiL, I'm submitting the bug now with mtn diff ${FILE} but I'll use diff later as my mtn db is all messed up Mar 07 07:12:03 http://sources.redhat.com/ml/libc-alpha/2005-09/msg00017.html Mar 07 07:12:22 this need to apply into current linux-header Mar 07 07:12:32 fo oe Mar 07 07:12:38 /s/fo/of/ Mar 07 07:13:33 alrighty. Submitted bug 1947.. my first bug for oe :) Mar 07 07:13:46 and of course the patch to fix it. Mar 07 07:23:57 <_law_> has someone OE running with monotone 0.33 ? Mar 07 07:31:06 morning Mar 07 07:31:14 _law_: I installed 0.33 yesterday Mar 07 07:31:59 jselect: diff -Nrup even ;) Mar 07 07:32:24 ~lart uboot-gta01 Mar 07 07:32:24 * ibot does a little 'dpkg -P uboot-gta01' action Mar 07 07:32:36 <_law_> hrw, any problems? Mar 07 07:34:20 _law_: none yet but I just did pull/merge/update once Mar 07 07:34:51 <_law_> hrw, ok thanks Mar 07 08:38:40 hrw: libpython doesn't get packaged for x86 :( Mar 07 08:38:49 ;( Mar 07 08:59:36 03koen 07org.oe.dev * rdf564f45... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: add feed arch for the neo1973 Mar 07 09:03:13 koen: uboot-gta01 need oabi still? Mar 07 09:03:38 hrw: haven't rebuilt uboot since 2 weeks ago Mar 07 09:03:50 http://youtube.com/watch?v=nuzRsXKm_NQ - matchbox in bios Mar 07 09:03:52 mkimage turns out to the arch independant... Mar 07 09:08:55 RP: Is jffs2 lzo compression in vanilla yet? Mar 07 09:09:32 XorA: good morning! Mar 07 09:09:32 morning Mar 07 09:09:39 koen: so how is the Neo then? Mar 07 09:09:54 XorA: bad blocks == jffs2 screwage Mar 07 09:10:11 root@fic-gta01:/etc/ipkg$ ipkg list_installed | awk '{print $1}' | xargs ipkg -force-reinstall install Mar 07 09:10:16 that's running now Mar 07 09:10:38 XorA: EABI is running great, though :) Mar 07 09:11:04 yay Mar 07 09:11:12 I thought the bad blocks problem was fixed Mar 07 09:11:16 koen: have you reinstalled the rootfs with DFU? Mar 07 09:11:23 LaF0rge: yes Mar 07 09:11:28 koen: I'm just adding something to the errata now Mar 07 09:11:30 while pressing AUX every 10 seconds :) Mar 07 09:11:38 koen: We're not erasing the remainder of the partition Mar 07 09:11:46 koen: so either you need to pad the image to the full rootfs part size Mar 07 09:11:58 koen: or you need to manualy 'nand erase rootfs' before starting the DFU Mar 07 09:12:32 LaF0rge: I suspected something like that Mar 07 09:12:37 ahah my Neo is in Edin, Ill probably pick it up from depot tomorrow then Mar 07 09:12:42 LaF0rge: good to know it isn't badblocks Mar 07 09:16:51 LaF0rge: what's the secret for turning on bluetooth? I can't see the power mgmnt entry for bluetooth in /sys/... Mar 07 09:18:45 03koen 07org.oe.dev * r11605062... 10/ (3 files in 3 dirs): xserver-kdrive-common: add GTA01 support Mar 07 09:19:12 rwhitby: modprobe gta01-pm-bt ? Mar 07 09:19:41 koen: can you rename uboot/uboot* to uboot/u-boot* like it is with other ones? Mar 07 09:19:56 hrw: I'm about to leave for the lab Mar 07 09:20:51 ok Mar 07 09:20:53 LaF0rge: it seemed to be loaded, but I've lost connectivity with the device now (it's at work, and I'm at home, and rebooting the host connected to it didn't work), so I'll look again tomorrow. Mar 07 09:35:59 NOTE: package u-boot-native-gta01-1.2.0+svn20070307-r3: task do_rebuild: completed Mar 07 09:40:14 Is it possible to tell bitbake to go all the way with every recipe, even if the dependency is only on do_stage? I'm trying to build with $WORKDIR on a 1G partition, so I'd need rm_work to be run right after every build, and not in the end. Mar 07 09:41:04 pH5: use bitbake 1.6 or set BB_NUMBER_OF_THREADS to 0 or 1 Mar 07 09:42:14 hrw: thanks! Mar 07 09:43:32 I'm not sure does second way help Mar 07 09:45:21 is Tim Chick here? Mar 07 09:45:40 hrw: it's worth a try. I don't want to go back to 1.6 Mar 07 09:47:03 NOTE: package linux-gta01-2.6.20-moko8-r1: task do_deploy: completed Mar 07 09:47:40 without oabi toolchain ;D Mar 07 09:49:47 03hrw 07org.oe.dev * r3e81bb40... 10/ (1 packages/uboot/uboot-gta01_svn.bb): uboot-gta01: fix parse error Mar 07 09:49:52 03hrw 07org.oe.dev * r6ab3973f... 10/ (1 packages/uboot/u-boot-mkimage-gta01_svn.bb): Mar 07 09:49:52 u-boot-mkimage-gta01: build uboot-mkimage needed for Neo1973 kernel Mar 07 09:49:52 - same tool is built with uboot-gta01_svn.bb recipe but Neo1973 u-boot Mar 07 09:49:52 require OABI toolchain Mar 07 09:49:55 03hrw 07org.oe.dev * r821beca3... 10/ (3 files in 2 dirs): u-boot-mkimage-gta01: renamed to u-boot-mkimage-gta01-native Mar 07 09:50:02 03hrw 07org.oe.dev * r387d69e6... 10/ (1 packages/linux/linux-gta01_2.6.20.bb): linux-gta01: depend on u-boot-mkimage-gta01-native so kernel can be built with EABI Mar 07 10:10:06 hrw: thanks for updating uboot Mar 07 10:11:34 koen|lab: np. it was faster then building oabi toolchain Mar 07 10:11:54 morning all Mar 07 10:12:12 http://ewi546.ewi.utwente.nl/tinderbox/showbuilds.pl?tree=Angstrom is my build Mar 07 10:17:13 hrw: but, uboot-utils aims for the same Mar 07 10:17:30 so if your mkimage thing also works for !arm, we could make it general Mar 07 10:18:41 koen|lab: feel free Mar 07 10:22:45 there's no ftdi stuff in OE already is there? Mar 07 10:22:50 http://www.intra2net.com/de/produkte/opensource/ftdi/ Mar 07 10:22:58 (for http://wiki.openmoko.org/wiki/Debug_Board#Usage_Instructions) Mar 07 10:32:33 good morning Mar 07 10:36:15 * CoreDump|afk can't believe he didn't switch to svn ealier Mar 07 10:36:31 CoreDump|afk: from cvs? Mar 07 10:37:01 hrw: yeah Mar 07 10:37:26 ;) Mar 07 10:37:30 =) Mar 07 10:37:34 CoreDump|afk: anyway I prefer monotone over svn Mar 07 10:37:41 hrw: heh Mar 07 10:37:49 * CoreDump|afk waits for the DHL delivery guy Mar 07 10:38:50 hi chrisromain Mar 07 10:38:58 hi all Mar 07 10:39:04 hi florian Mar 07 10:40:05 ho Mar 07 10:40:12 hi zecke Mar 07 10:40:50 hi zecke Mar 07 10:40:59 hrw: BB_NUMBER_THREADS didn't help, I'll have to look for something else. Mar 07 10:41:15 pH5: ask zecke or RP then Mar 07 10:41:35 pH5: do you plan to update x11 to r7.2? Mar 07 10:45:33 hrw: not until may Mar 07 10:46:52 are there "perfect example" recipes that one should copy when starting new library and application recipes? Mar 07 10:47:59 I usually start from empty file Mar 07 10:48:27 rwhitby: there is contrib/oe-stylize.py and http://openembedded.org/wiki/StyleGuide to help you with that Mar 07 10:49:03 pH5: yeah, I know about those. I was hoping someone would say "I wrote this recipe most recently, and it is perfect" for both a library and an application. Mar 07 10:49:41 rwhitby: 'inherit lib_package' Mar 07 10:50:42 koen|lab: thx. No complete recipes from you that are completely perfect? Mar 07 10:50:44 rwhitby: that should take care of packaging, staging, fixing .la files, etc Mar 07 10:50:54 rwhitby: not that I can think of right now Mar 07 10:51:30 rwhitby: people are too modest :) Mar 07 10:51:45 pH5: except when someone else makes a stylistic mistake ... Mar 07 10:52:54 guess I'll just travel the usual "write it, get yelled at, fix it" route ... Mar 07 10:56:02 should this package (http://www.nongnu.org/confuse/) go under packages/confuse/.. or under packages/ftdi/ (which is where I'm putting ftdi_eeprom and libftdi)? ftdi_eeprom depends on it, but it's not in OE already. Mar 07 10:56:30 confuse/ Mar 07 10:57:07 and the bb file should be libconfuse.bb or confuse.bb ? Mar 07 10:57:25 (the source tarball is confuse.tar.gz, but it's called libconfuse ...) Mar 07 10:58:30 * rwhitby thinks packages/libconfuse/libconfuse_${PV}.bb Mar 07 10:58:36 rwhitby: normally confuse.bb then debian renaming will make the output libconfuse* Mar 07 10:59:10 XorA: that doesn't seem to be followed very well with existing packages (e.g. libftdi/libftdi.bb ...) Mar 07 10:59:33 rwhitby: it seems to be how OE used to do things, obviously a clash there Mar 07 10:59:52 so it should be packages/confuse/confuse_2.5.bb then Mar 07 11:00:05 packages/jpeg/jpeg certainly follows that pattern Mar 07 11:00:52 rwhitby: I generally try to use the tarball name. libftdi has the lib prefix in the tarball name (as do the xorg libs, for example). Mar 07 11:01:24 pH5: you put it better than me Mar 07 11:05:14 "try to avoid having to set S =" Mar 07 11:07:11 koen|lab: say hi to charles eugene from me :-) Mar 07 11:11:21 oxo: who? where? Mar 07 11:15:17 koen|lab: I mean mister van Eijckeren Mar 07 11:15:57 oxo: done :) Mar 07 11:16:56 koen|lab: no complains about sepukku.bbclass yet? Mar 07 11:18:01 zecke: haven't tried it yet Mar 07 11:28:42 robtaylor: didn't you have a rtp library somewhere? Mar 07 11:29:08 NAiL: Not yet, no Mar 07 11:29:10 morning all Mar 07 11:29:17 hey RP Mar 07 11:30:52 sed -n 's/^#define \([^_][a-zA-Z0-9_]*\).*/\1/p ; \ Mar 07 11:30:52 s/^#define \(_[^_A-Z][a-zA-Z0-9_]*\).*/\1/p' | \ Mar 07 11:30:52 sort -u > tmp-macro_list Mar 07 11:30:52 sed: -e expression #1, char 88: unterminated address regex Mar 07 11:30:52 :0: fatal error: when writing output to : Broken pipe Mar 07 11:30:52 compilation terminated. Mar 07 11:30:54 sorry Mar 07 11:31:13 hi RP Mar 07 11:31:21 can anyone tell me what is wrong with this? Mar 07 11:31:21 hey RP Mar 07 11:31:23 Packaged contents of confuse-dev into /home/moko/build/tmp/deploy/ipk/armv4t/confuse-dev_2.5-r0_armv4t.i Mar 07 11:31:24 pk Mar 07 11:31:35 should that be libconfuse-dev, or is that correct? Mar 07 11:32:13 rwhitby: you got confuse_ipk or libconfuse? Mar 07 11:32:24 neither. Mar 07 11:32:48 gues lib_package didn't work automagically for me. Mar 07 11:33:32 rwhitby: did you install more than just a library? what are the names of the other packages? Mar 07 11:34:00 rwhitby: I think it's correct. we generally don't split -dev and -dbg packages, so they just keep the .bb name. Mar 07 11:34:16 recipe: http://pastebin.ca/384842 Mar 07 11:34:18 zecke: are -dev packages renamed debian-style for single libraries? Mar 07 11:34:30 pH5: good question :) Mar 07 11:34:43 output: http://pastebin.ca/384845 Mar 07 11:34:43 rwhitby: 'inherit autotools lib_package' Mar 07 11:35:23 exclude-tests is temporary, cause I can't be bothered how to fix the automake stuff which fails to find ../src includes. Mar 07 11:36:11 hrw, RP: What does the "keyboard" feature mean? Any kind of keyboard or a full keyboard? Mar 07 11:36:54 florian: any keyboard which ned keymap setting Mar 07 11:37:19 florian: so if your device has working keymap in kernel then you can but do not have to use this feature Mar 07 11:37:22 hrw: ok, so i can't use it to decide if i want a virtual keyboard or not. Mar 07 11:37:47 florian: yep. it cover hardware stuff Mar 07 11:38:08 I mean 'base system' stuff not x11/opie/gpe ones Mar 07 11:38:48 hrw: sure, but if it indicated a full featured keyboard i could use it to decide this. Mar 07 11:39:19 so when does this debian renaming kick in? Mar 07 11:39:25 florian: collie has keyboard but do not have it in features because 2.4.18 gives working keys and we want to have as small as possible in rootfs Mar 07 11:39:44 for the record sed need -e Mar 07 11:39:44 hrw: ah ok Mar 07 11:39:51 lib_package.bbclass assumes that ${PN} is the name you want for the package, which in this case is 'confuse', not 'libconfuse'. Mar 07 11:39:52 hrw: is that documented somehow? Mar 07 11:39:59 XorA|dentist: enjoy ;) Mar 07 11:40:01 florian: nope rather Mar 07 11:41:44 03pH5 07org.oe.dev * rbefbf790... 10/ (1 packages/gnome/goffice_0.3.2.bb): goffice: fix QA issue, add all debug files to -dbg package. Mar 07 11:44:16 koen|lab: is there a template for 'lib_staging.bbclass' functionality? Mar 07 11:47:54 and sed -e 'foo bar' Mar 07 11:48:08 hi all Mar 07 11:48:12 hi, i try to compile gtk+2.10.6 for armv4 but it fails on building "cups". Is there any "no-cups" patch or somethig like ? Mar 07 11:48:47 lkm`: it fails.. how it fails? Mar 07 11:50:15 hrw: Compiling phpcups.c... | CROSS COMPILE Badness: /usr/include in INCLUDEPATH: /usr/include/php | cc1: internal compiler error: in add_path, at c-incpath.c:362 Mar 07 11:51:04 thx Mar 07 11:51:13 lkm`: change the bbfile and disable GtkPrinting Mar 07 11:51:25 zecke: bug is in cups Mar 07 11:52:14 hrw: right, he asked to disable cups though and this looks like a dep of GtkPrinting Mar 07 11:52:34 I like to get some feedback for http://bugs.openembedded.org/show_bug.cgi?id=1946 Mar 07 11:52:55 zecke: That's fbgrab-viewer-native. Standalone viewer Mar 07 11:53:10 -native? Mar 07 11:53:14 lkm`: can you edit packages/cups/cups_1.2.7.bb and add "--without-php" into EXTRA_OECONF? Mar 07 11:53:32 i shell try .. Mar 07 11:54:39 zecke: The viewer is always for the host poc Mar 07 11:54:40 pc Mar 07 11:55:19 stefan_schmidt: You assume the presence of X11 (which is okay as I don't want you to force to add X11-native) Mar 07 11:56:07 stefan_schmidt: but don't you want to have fbgrab-viewer running on your Neo19730? or any other device (e.g. your dreambox)? Mar 07 11:57:46 stefan_schmidt: the question is, do you want to setup a OE environment to build your 'native' packages, most likely you don't want that for something as simple as fbgrab-viewer. What about adding a -native and non-native version? ;) Mar 07 11:58:31 zecke: Sounds good to me. I'll take a look. Mar 07 11:58:59 stefan_schmidt: in that case remove the PROVIDES line, otherwise you will have to hunt an error Mar 07 11:59:33 zecke: ah, thnaks Mar 07 12:01:21 lkm`: does that change help? Mar 07 12:05:18 hrw: yes, but now gtk+ failed. *** Can't link to Pango. Pango is required to build Mar 07 12:05:51 pango should be already built.. Mar 07 12:06:26 yes it is .. i try to rebuild it Mar 07 12:07:02 hrw: we have seen that yesterday, It fails with pangocairo library missing... Mar 07 12:07:42 zecke|away: i also, buth with other gtk+. For that reason i choose the exact version Mar 07 12:08:39 lkm`: which machine/distro combo you use? Mar 07 12:08:47 i check and really there is no libpangocairo as GTK+ expected. Mar 07 12:08:56 final confuse_2.5.bb: http://pastebin.ca/384869 <- comments? Mar 07 12:09:34 rwhitby: looks nice Mar 07 12:09:40 rwhitby: remove 'confuse' from description Mar 07 12:09:42 lkm`: I think gtk2.6 is the most tested version. Which version do you build, which distribution? Mar 07 12:09:45 hrw: i made some distro confics. Except task-base i want Minimo Mar 07 12:09:52 "Library for parsing configuration files." is enough Mar 07 12:09:54 rwhitby: http://www.intra2net.com/de/produkte/opensource/ftdi/TGZ/${P}.tar.gz Mar 07 12:10:14 lkm`: better use angstrom-2007.1 ;D Mar 07 12:10:48 zecke|away: i try to make some build that contains graphical web browser (Minimo) for arm920T Mar 07 12:11:07 koen: ${P} then doesn't work for the -native .bb Mar 07 12:11:41 confuse-${PV} then Mar 07 12:11:48 yep, done that. Mar 07 12:12:47 I need to put 'S = "${WORKDIR}/confuse-${PV}"' in too for -native. Mar 07 12:13:16 03hrw 07org.oe.dev * r33540f0b... 10/ (1 packages/cups/cups_1.2.7.bb): cups: we do not want web interface so do not check for PHP, Perl, Python, Java Mar 07 12:14:06 hrw: angstrom ? Mar 07 12:14:35 confuse-native_2.5.bb: http://pastebin.ca/384873 <- comments? Mar 07 12:15:25 lkm`: angstrom Mar 07 12:15:39 lkm`: this is distro which has to unify all our distros Mar 07 12:16:04 rwhitby: use files/ dir instead of confuse-2.5/ one Mar 07 12:16:14 rwhitby: so you will not have to set FILESPATH Mar 07 12:16:34 hrw: ah, yes. I did actually use files, but still copied that in :-) Mar 07 12:19:47 koen: openmoko-image is hard target Mar 07 12:19:52 s/hard/big Mar 07 12:19:58 indeed Mar 07 12:22:27 hrw: ok, i shell try it is some time . Mar 07 12:22:52 lkm`: it is maintained and buildable. I'm now building for armv4t device Mar 07 12:23:24 hrw: i just wanted to build some browser .. so it is not a problem to choose another version of gtk+ Mar 07 12:24:06 03rwhitby 07org.oe.dev * rd0cd7aea... 10/ (6 files in 4 dirs): confuse: Library for parsing configuration files. Mar 07 12:24:16 which version is used in angstrom ? Mar 07 12:24:32 there: *that* recipe is now perfect :-) Mar 07 12:25:09 lkm`: cd de bu open t stam Mar 07 12:25:11 ops Mar 07 12:25:12 cd de bu open t stam Mar 07 12:26:00 lkm`: gtk 2.10.9 glib 2.12.9 pango 1.16.0 atk 1.10.3 Mar 07 12:27:55 03nail 07org.oe.dev * r6839c5ad... 10/ (1 conf/machine/turbostation.conf): turbostation: Add a kernel-feed for the turbostation Mar 07 12:28:51 NAiL: feeds are distro thing not machine Mar 07 12:28:54 NAiL: that doesn't belong in conf/machine Mar 07 12:29:32 NAiL: turbostation + umbaumba distro will do not work.. Mar 07 12:29:56 ... Mar 07 12:30:37 do I really have to make sure my machine builds on all distros? Mar 07 12:31:43 oh well Mar 07 12:31:54 NAiL: it doesn't belong there Mar 07 12:31:59 I'll clone off openprotium and make yet another distro so I can add my fucking feed Mar 07 12:33:44 * rwhitby is an idiot. confuse is already packages by openmoko. Mar 07 12:34:16 rwhitby: it is in oe already? Mar 07 12:34:23 it will be soon Mar 07 12:34:50 rwhitby: s/is/was/ Mar 07 12:35:16 it's in om svn, but will be in OE soon Mar 07 12:35:38 according to CIA, it is in OE already Mar 07 12:35:52 f'ing Xilinx Mar 07 12:36:48 koen: becuase I just committed it 10 minutes ago? but totally independent creation of the existing package in OM svn Mar 07 12:36:49 03nail 07org.oe.dev * r405d9f5a... 10/ (1 conf/machine/turbostation.conf): turbostation: Remove feed. Mar 07 12:36:53 03nail 07org.oe.dev * r39b23b4b... 10/ (1 conf/distro/openturbostation.conf): openturbostation: Initial commit of distro Mar 07 12:37:53 koen: what's the difference between "inherit autotools binconfig pkgconfig" and "inherit autotools lib_package" ? Mar 07 12:38:00 re Mar 07 12:38:08 koen: thx :> Mar 07 12:38:37 rwhitby: binconfig is used for -config style packages Mar 07 12:38:44 e.g. cups-config to get the location of the headers Mar 07 12:40:29 so I should merge those two lines then Mar 07 12:43:09 hmm - builds and packages fine without lib_package Mar 07 12:45:27 hmm - actually, it doesn't. inherit lib_package causes the emitted ipks to be libconfuse ... Mar 07 12:46:16 right Mar 07 12:46:28 and it seperates the helper binaries into confuse-bin Mar 07 12:50:22 03rwhitby 07org.oe.dev * r1b7c658e... 10/ (4 files in 3 dirs): confuse: Imported and merged recipe from OpenMoko Mar 07 12:51:07 * rwhitby goes to bed before he makes even more work for himself for no reason ... Mar 07 12:52:20 hrw: GTK+ 2.10.9 also gives the same error: *** Can't link to Pango. Pango is required to build Mar 07 12:53:23 hrw: but actually i have builded pango-1.16.0-r0 already Mar 07 12:59:18 * rwhitby wonders why the openmoko ftdi-eeprom-native.bb does not leave something in i686-linux/bin ... Mar 07 12:59:59 rod: because it has the wrong staging dir? (non-native)? Mar 07 13:03:08 heh Mar 07 13:03:16 turkish scriptkiddies have low standards Mar 07 13:03:27 "HACKED BY UZMAN" in a wiki Mar 07 13:03:39 :D Mar 07 13:03:55 koen: ..that's pathetic Mar 07 13:04:30 http://openzaurus.berlios.de/Users <- the page in question Mar 07 13:04:53 LaF0rge: would autotools_stage_all in do_stage() fix that? Mar 07 13:06:56 LaF0rge: my goal is to follow the debug board wiki instructions, but using all tools built from OE and OM repos. Mar 07 13:07:45 rod: I am not the OE guy here, i have no idea :) Mar 07 13:08:50 LaF0rge: neither am I yet :-) I'll try some things out and submit a patch Mar 07 13:09:09 rod: we need some do_stage Mar 07 13:09:11 let me try Mar 07 13:09:27 rod: reference the openocd-native recipe Mar 07 13:10:12 oh, it has that. but NATIVE is missing Mar 07 13:10:31 inherit native, yes. Mar 07 13:10:35 STAGING_BINDIR_NATIVE Mar 07 13:10:52 LaF0rge: I'll pastebin something soon Mar 07 13:10:55 * NAiL kicks CIA-3 Mar 07 13:10:55 ow Mar 07 13:11:40 - install -m 0755 ftdi_eeprom/ftdi_eeprom ${STAGING_BINDIR} Mar 07 13:11:40 + install -m 0755 ftdi_eeprom/ftdi_eeprom ${STAGING_BINDIR_NATIVE} Mar 07 13:12:26 LaF0rge: scratch that, I need to sleep. Mar 07 13:12:41 doesn't "inherit native" do that stuff? Mar 07 13:13:02 rwhitby: don't think so Mar 07 13:13:20 we're explicitly asking it to put a file to some directory, and giving the wrong one Mar 07 13:13:49 I thought the combination of "inherit native" and "autotools_stage_all" instead of the hand-crafted do_stage() would do it? Mar 07 13:16:25 * rwhitby notices the checking in #openmoko, and goes to bed :-) Mar 07 13:17:21 morning Mar 07 13:26:42 koen: | /home/hrw/devel/build/openmoko/tmp/work/armv4t-angstrom-linux-gnueabi/gcc-4.1.1-r10/gcc-4.1.1/gcc/real.h:78: error: size of array 'test_real_width' is negative Mar 07 13:26:51 koen: did you get this? Mar 07 13:27:01 not that I know off Mar 07 13:54:23 greetings, I'm trying to create a bb recipe with libboost-serialization1.33.1 as a runtime dependency, so I set RDEPENDS = "libboost-serialization1.33.1", but when I build it I get this error: ERROR: Nothing provides runtime dependency libboost-serialization1.33.1 Mar 07 13:55:13 I'm guessing this is because there isn't a .bb file called libboost-serialization1.33.1_1331.bb, only a package that gets created when boost is built Mar 07 13:55:41 how are dependencies referenced that aren't .bb files? Mar 07 13:56:02 xuumbi: why do you need to set that as dependency? Mar 07 13:56:14 xuumbi: OE will pick it up automagically with the shlib code Mar 07 14:02:14 koen, how old is the )# db snapshot? Mar 07 14:02:38 koen: hmmm, I thought I had to set it Mar 07 14:03:06 koen: I'll give it a try, thanks Mar 07 14:03:55 lkm`: I had to delete tmp and rebuild everything to make pango visible to other packages last week. I think it was due to a problem with the new ipk layout, but I'm not sure. I tried rebuilding pango several times first, which worked fine, but other programs still couldn't find it. Mar 07 14:04:50 apaulsen: it has do to with pkgconfig fixes, not with the deploy rework Mar 07 14:04:53 zecke|away: yep Mar 07 14:05:24 koen|away: ok, I won't badmouth the deploy rework :) Mar 07 14:06:11 zecke|away: i also know quite a bit about rtp, so ping me when you get back with what you want to do Mar 07 14:08:45 robtaylor: where can I find the sourcecode? Is it within farsight? Mar 07 14:09:08 zecke|away: well, what exactly do you want to do? Mar 07 14:09:27 zecke|away: plain rtp, all you need is gstreamer Mar 07 14:09:50 zecke|away: for nat traversal google-talk style you need farsight Mar 07 14:09:51 robtaylor: I wanted to see if I can use any preexisting class to create NTP Times Mar 07 14:10:14 zecke|away: rtcp? there's not much support for that anywhere at the moment Mar 07 14:10:48 robtaylor: yes, we wrote our own class now but thanks for listening Mar 07 14:10:48 zecke|away: why do you want NTP times? Mar 07 14:11:17 robtaylor: we patch and merge various RTP/RTCP sources on the fly Mar 07 14:11:31 he zecke Mar 07 14:11:32 zecke|away: who's 'we' ? Mar 07 14:12:08 robtaylor: the company that gives me office space (www.gmit-gmbh.de) Mar 07 14:12:28 zecke|away: ah, closed source stuff then? Mar 07 14:12:32 robtaylor: For DVB-H we patch various RTP bits to make ALL the clients and players happy Mar 07 14:13:29 robtaylor: not everything, but yes this is more than just a fun/spare time project Mar 07 14:14:09 zecke|away: hmm, yeah, looks like its something comparable with flumotion Mar 07 14:14:44 robtaylor: Hmm, I think it is not but let me reread on flumotion Mar 07 14:15:09 what is the best distro for mipsel platform? Mar 07 14:15:36 crink: embedded? Mar 07 14:15:50 yes Mar 07 14:15:58 hm Mar 07 14:16:03 angstorem? Mar 07 14:16:07 au1200 series Mar 07 14:16:12 amd -> rmi Mar 07 14:16:23 but dont know about mips status Mar 07 14:16:31 yeah, i'd say angstrom, though its usually tested on arm, of course Mar 07 14:16:40 robtaylor: it is different, we try to solve a different problem :) Mar 07 14:16:56 dreambox ist mips but the repositrie differs a lot from oe-dev Mar 07 14:17:08 zecke|away: ah :) using gstreamer? or custom media stuff? Mar 07 14:17:09 and dreambox is a settopbox Mar 07 14:17:34 cron k you can try angstroem and provide patches for mips kernel and toolchain stuff it not works out of box Mar 07 14:17:44 crink I mean Mar 07 14:17:45 sorry Mar 07 14:17:48 * crink had tried angstrom, having tried opendreambox Mar 07 14:18:05 dev, dreambox Mar 07 14:18:14 crink nothing works? Mar 07 14:18:43 imho, od is better than angstrom, but i can't sure Mar 07 14:18:58 lots of time wasted Mar 07 14:19:15 crink: you need to be clear on exactly what you want Mar 07 14:19:33 crink: a gpe platform? just a basic platform? Mar 07 14:19:48 opie? Mar 07 14:19:48 robtaylor: We just patch RTP no matter where it is from (helix,vlc,darwin, Insert Product Name here). We add artificial delays, allow redunancy, hot swapping of streams, fixing sequence numbers and session ids to make dvb-h terminals happy etc... Mar 07 14:19:48 just basic platform Mar 07 14:20:09 zecke|away: i think you use patch in a different way to me ;) Mar 07 14:20:14 mipsel linux console with nptl support Mar 07 14:20:39 robtaylor: probably, and sorry for that then Mar 07 14:20:50 zecke|away: i was actually asking about your media layer, but i guess from your answer its a closed custom jobbie Mar 07 14:21:12 crink than angstroem and bootstrap-image Mar 07 14:21:14 hey crink, I had some troubles with mipsel lately, in the end I created an own distro and matched the glibc and gcc versions that came with the vendor; it was without nptl though Mar 07 14:21:27 and if it not works, we mast patch Mar 07 14:21:29 zecke|away: which probably makes sense as you're not trying to do anything too hard Mar 07 14:21:31 must Mar 07 14:21:45 woglinde: thx i will give it a try. Mar 07 14:21:49 03rpurdie * r788 10/ (2 files in 2 dirs): runqueue.py: Fix parameters to get_user_idstring() Mar 07 14:22:03 robtaylor: media layer in which sense? (e.g. we don't alter payloads of the RTP packets) Mar 07 14:22:07 crink maybee jin can provide you patches Mar 07 14:22:19 if he is kindly Mar 07 14:22:46 zecke|away: yeah, thats what i meant. i had the impression histvserver was muxing more than just rtp streams as input Mar 07 14:23:00 crink but how familiar are you with building toolchains and stuff? Mar 07 14:23:03 zecke|away: but anyhow, for NTP code, you can look at jrtplib Mar 07 14:23:15 woglinde: very handy with them Mar 07 14:23:26 zecke|away: thats a very capable rtp library Mar 07 14:23:55 i know there is Gx-linux but it is not free Mar 07 14:25:00 robtaylor: thanks Mar 07 14:25:12 crink: you could always give debian a try ;) Mar 07 14:25:15 crink but ENABLE_BINARY_LOCALE_GENERATION = "0" in your local.conf so qemu istn build and glibc buildinh is faster Mar 07 14:25:29 only 3-4 hours Mar 07 14:26:53 good Mar 07 14:26:56 thx Mar 07 14:27:13 args but = put Mar 07 14:31:07 does anyone know what happened to the openalchemy packaging effort? Mar 07 14:31:14 are there some recipes available somewhere? Mar 07 14:31:46 I had a start online somewhere Mar 07 14:32:01 koen: I would be interested in continuing/contributing to that Mar 07 14:32:04 then I looked at their makefiles and concluded crosscompiling won't ever work Mar 07 14:32:24 koen: why didn't you report to openalchemy-general? Mar 07 14:32:33 koek: they will fix it, just need to tell it Mar 07 14:33:16 http://www.openembedded.org/~koen/openalchemy.tar.bz2 Mar 07 14:34:58 LaF0rge: ooh, i like the inplace handwriting recognition Mar 07 14:35:10 rob: ? Mar 07 14:35:40 rob: alchemy doesn't even try to do any recognition. just save everything as graphics Mar 07 14:35:51 1-bit graphics at lo-res are not all that big Mar 07 14:35:55 LaF0rge: ah, i see Mar 07 14:36:23 LaF0rge: not sure how that fits in a an Subject: line then Mar 07 14:36:54 (http://openalchemy.org/index.php/Image:Email2.jpg) Mar 07 14:38:42 * oxo wants to package console-data, putting the fonts in seperate packages Mar 07 14:39:02 is there a way to do that without (explicitly) listen all (200+) fonts in PACKAGES? Mar 07 14:39:07 btw, i was having some interesting discussions with Ole Reinhardt at FOSDEM about writing a modern handwriting recogniser Mar 07 14:39:07 rob: haven't used that email app so far Mar 07 14:39:36 oxo: python, do_split_packages Mar 07 14:40:09 oxo: see http://www.openembedded.org/repo/org.openembedded.dev/packages/gnome/goffice_0.3.2.bb Mar 07 14:41:01 thank you Mar 07 14:41:30 I think I can do that Mar 07 14:42:26 robtaylor: mallum has his own xstroke thingy as well Mar 07 14:43:45 koen: yeah, all the existing stuff is very basic in its recognition Mar 07 14:44:28 koen: ive been looking at a combined time delay NN and hidden markov model based scheme Mar 07 14:45:00 it should in theory be able to recognise cursive words and printed text Mar 07 14:45:38 koen: I've passed on this info Mar 07 14:45:49 koen: but we could have saved two months in latency if it was reported earlier Mar 07 14:49:02 koen: I'm not blaming you, but if you run into such an issue ever again, please ping them or me. thanks! Mar 07 14:50:49 LaF0rge: ehm... you lost me Mar 07 14:50:55 LaF0rge: what issue? Mar 07 14:51:13 ah, openalchemy Mar 07 14:51:27 * koen is concentrating on too many things at once Mar 07 14:52:03 koen: yes, openalchemy Mar 07 14:52:12 keon: yes, I know that feeling very well Mar 07 14:52:13 * chouimat dcc a cloning machine to koen :) Mar 07 14:52:45 oh dear Mar 07 14:52:52 and I need to vote today as well Mar 07 14:53:39 voting machines? Mar 07 14:53:47 not yet running OE ? ;) Mar 07 14:53:56 I suspect machines Mar 07 14:54:32 * koen conjures up Crofton__ to do some van Eck phreaking Mar 07 14:54:53 we still have an old chap with a hole punch for voting in UK Mar 07 14:55:34 and he normally smells of tea and digestive biscuits Mar 07 14:56:01 * chouimat prefers the pen and paper Mar 07 15:00:18 Hi! Mar 07 15:01:08 What is Debian/OE way to control what /etc/init.d/ scripts actually run on a given runlevel? Mar 07 15:01:22 update-rc.d exists for it Mar 07 15:02:35 hrw, why update-rc.d run for each script installed in init.d/ ? what if I want to install service, but leave its startup to user? Mar 07 15:03:13 And after all, what is Debian analog of RH chkconfig util? Mar 07 15:04:50 hrw: ^ ? Mar 07 15:05:16 psokolovsky: then do not use update-rc.d and let user start it by hand? Mar 07 15:05:24 and I do not know what chkconfig is Mar 07 15:07:16 hrw, RH has special syntax for init.d/ scripts (2nd or 3rd line), which lists recommended priority and runlevels for service. chkconfig then can easily activate/deactivate service for automatic startup Mar 07 15:07:56 hrw, well, if I want to remove that update-rc.d call for some service - does it need gross RFC, or will be ok to discuss here? Mar 07 15:08:22 probably depends on service Mar 07 15:08:26 I have to go now Mar 07 15:08:51 regarding http://bugs.openembedded.org/show_bug.cgi?id=1532 Mar 07 15:09:34 I dont understand, how can i install libidl-dev in host? Mar 07 15:21:21 hi all Mar 07 16:39:13 03nail 07org.oe.dev * r783d972a... 10/ (1 packages/images/openturbostation-image.bb): openturbostation: Add image generation Mar 07 16:39:17 03nail 07org.oe.dev * rc67676d2... 10/ (10 files in 4 dirs): turbostation: initscripts Mar 07 16:39:24 03nail 07org.oe.dev * r8f07b698... 10/ (13 files in 4 dirs): turbostation: More init-stuff Mar 07 16:39:35 03nail 07org.oe.dev * r90400ad9... 10/ (1 packages/meta/openturbostation-packages.bb): openturbostation: Add packages Mar 07 16:39:42 03koen 07org.oe.dev * r1ed4ea88... 10/ (1 packages/matchbox2/matchbox-panel-2_svn.bb): mb-panel2: rconflict with mb-panel-1 Mar 07 16:39:48 03nail 07org.oe.dev * rc68d955a... 10/ (14 files in 3 dirs): openturbostation: init-stuff Mar 07 16:39:57 03pH5 07org.oe.dev * r2a83a3cf... 10/ (1 packages/linux/linux-handhelds-2.6.inc): Mar 07 16:39:57 linux-handhelds-2.6: do_deploy after do_install and before do_package. Mar 07 16:39:57 - This makes sure that bitbake doesn't try to deploy after rm_work. Mar 07 16:40:07 03florian 07org.oe.dev * rbc33dfdd... 10/ (1 packages/xserver-common/xserver-common_1.15.bb): xserver-common: Add 1.15 Mar 07 16:40:17 pH5: empty do_rm_work() in recipe is simpler ;) Mar 07 16:40:24 03nail 07org.oe.dev * r4628d4ec... 10/ (1 packages/initscripts/initscripts-openturbostation_1.0.bb): openturbostation-initscripts: Fix erroneous patch Mar 07 16:40:36 03pH5 07org.oe.dev * r81b1167c... 10/ (1 packages/gnumeric/gnumeric_1.6.3.bb): gnumeric: fix QA issue, package all debug files in -dbg package Mar 07 16:40:42 03koen 07org.oe.dev * rb0902151... 10/ (1 packages/tasks/task-openmoko.bb): task-openmoko: ship matchbox panel plugins and openmoko panel plugins, install a .desktop for rxvt Mar 07 16:40:54 thar she blows :P Mar 07 16:41:50 03florian 07org.oe.dev * rf08623e9... 10/ (1 packages/tasks/task-gpephone.bb): gpephone-image: Include libgtkinput in all images. Mar 07 16:42:43 how does OE figure out which files to put in the staging dir during do_stage? Mar 07 16:43:04 it doesn't Mar 07 16:43:30 autotools_stage_all just does 'make install DESTDIR=staging' and other recipes use oe_libinstall by hand Mar 07 16:43:47 oh, that explains ;) Mar 07 16:46:52 pH5: thanks for looking into these issues. Do you find a lot of false positives? Mar 07 16:52:35 are I am right that MACHINE_EXTRA_RRECOMMENDS is for machine specific modules and they are loaded automatical at bootup? and BOOTSTRAP_EXTRA_RDEPENDS modules are loaded on demand via udev? Mar 07 16:53:59 mr_nice: no Mar 07 16:54:21 mr_nice: whether or not a module is loaded automatically depends on the module Mar 07 16:54:24 zecke|away: no false positives, but I only looked out for the debug packaging so far because those errors stop the build. Mar 07 16:54:25 MACHINE_EXTRA_RRECOMMENDS is for machine specific modules installed by default Mar 07 16:54:36 BOOTSTRAP_EXTRA_RDEPENDS shouldn't exist in OE anymore Mar 07 16:54:40 pH5: ah hehe Mar 07 16:54:56 pH5: I think that case is even complete and sound Mar 07 16:55:33 koen: so MACHINE_EXTRA_RECOMMENDS is what I want if I want to add pcmcia and vfat support? Mar 07 16:56:32 mr_nice: not really Mar 07 16:56:39 vfat isn't needed to boot your machine Mar 07 16:57:09 koen: hm, I want to include vfat support so that I can mount a cf card. where do I need to add the module? Mar 07 16:57:24 in your distro or local.conf I guess Mar 07 16:58:34 koen: so just copy the angstrom-2007.1-oabi.conf to angstrom-2007.1-oabi-simpad.conf and add which bitbake variable? Mar 07 17:00:31 no Mar 07 17:00:44 why does everyone want to copy files for bogus reasons? Mar 07 17:01:07 koen: because I don't understand bitbake Mar 07 17:02:38 koen: shuld I just overwrite the angstrom-2007.1-oabi.conf? It should be availabel for everyone so local.conf is the wrong file or? Mar 07 17:02:56 no Mar 07 17:05:06 koen: I want to change oe that way that someone else can easily create a image for his simpad - with vfat support included. Mar 07 17:05:31 why not work with the angstrom people? Mar 07 17:05:53 like "I'd like to ship vfat by default, what do you think?" Mar 07 17:06:04 and they might respond with "great idea!" Mar 07 17:06:20 (notice how that conversation lacks 'simpad') Mar 07 17:07:24 koen: hm, so first ask there and then change the file or ask for some one else to change it? I think that no one will have time for simpad stuff or? Mar 07 17:08:08 in this case I already committed it :) Mar 07 17:08:57 koen: hehe, great thx! Mar 07 17:09:36 koen: openalchemy folks are starting to cleanup now Mar 07 17:09:37 03koen 07org.oe.dev * r18946765... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: recommond kernel-module-vfat so people can mount their favourite fat formatted cards and disk by default Mar 07 17:09:42 LaF0rge: cool Mar 07 17:09:53 LaF0rge: I hope to have some time off in brazil to look at it Mar 07 17:10:05 koen: I'm giving the bb recipes a revamp, will put them up somewhere public as soon as something builds Mar 07 17:43:09 03nail 07org.oe.dev * rd53ee743... 10/ (1 conf/machine/turbostation.conf): turbostation: Fix two variables Mar 07 17:43:17 03nail 07org.oe.dev * rdb3f92ea... 10/ (1 packages/images/openturbostation-image.bb): openturbostation: Fixes to build image Mar 07 17:43:28 03nail 07org.oe.dev * rd9f359eb... 10/ (3 files in 2 dirs): openturbostation-init: Rename .bb Mar 07 17:43:44 03nail 07org.oe.dev * r0c12cf61... 10/ (1 conf/machine/turbostation.conf): openturbostation: Fix misspelling Mar 07 18:05:09 Imagine a bitbake where it tries to run every possible task in parallel... Mar 07 18:05:21 * RP wonders how the machine is still running... Mar 07 18:07:34 RP: that sounds good :-) Mar 07 18:09:33 RP: take a look at the debian QA slides and the parallel computer Mar 07 18:09:54 RP: he claims to recompile debian in 7 hours on this cluster, 7 hours for OpenOffice :) Mar 07 18:10:49 zecke|away: It would be interesting to let bitbake mt loose on something like that :) Mar 07 18:13:09 * LaF0rge would already be interested to see bitbake run loose on a single niagara box Mar 07 18:13:11 yes, I can we could start building our own cluster Mar 07 18:13:15 The OE foundation should buy a big-ass cluster just to test that Mar 07 18:13:31 LaF0rge: hmm, niagara *yummie* Mar 07 18:13:45 zecke: I want one of these as OM buildhost :) Mar 07 18:14:24 LaF0rge: I would be happy with a couple of POWER cpu's as well Mar 07 18:14:44 LaF0rge: talking about the buildhost, I'm about to stop paid work for today. How do you want to proceed with autobuilds Mar 07 18:16:13 RP, zecke|away: Would it somehow be possible to emulate the old bitbake behaviour where every dependency is built up to do_build instead of just running the needed tasks? Mar 07 18:17:01 I'm trying to build in a 1GB ramfs, but it doesn't work because bitbake won't run the rm_work tasks before the disk runs full. Mar 07 18:17:08 zecke: i can give you an [root] account on the machine that would in the end run those builds (for now), if that is what you have in mind ;) Mar 07 18:17:34 pH5: We need a get_next_task() function in runqueue.py Mar 07 18:18:08 a better get_next_task()... Mar 07 18:18:35 LaF0rge: status is unchanged: I need to update my autobuild script and someway to put a status message on a website (but I think the autobuilder has apache running anyway) Mar 07 18:19:24 zecke: well, you can just do that on one of our [v]hosts. you get root to it and do whatever you think is required Mar 07 18:19:39 we e.g. have buildhost-old.openmoko.org which can be used for this Mar 07 18:19:58 or even buildhost-test.openmoko.org as well ;) Mar 07 18:20:23 cu Mar 07 18:20:35 LaF0rge: okay, I will talk to you about this probably at ~20:30 (I should be home at that time) Mar 07 18:20:59 zecke: ok, just send your ssh key via email to laforge@openmoko.org Mar 07 18:23:00 RP: how? to (optionally) build a package all the way, some "get_next_task" would have to return the next task of the current provider, while currently get_next_task is used to fetch the next task out of the "buildable" runqueue. aren't those completely different things? Mar 07 18:24:16 pH5: It could give preference to the next buildable task with the same provider as the one last built. Or give priority to the rm_work task. Lots of options Mar 07 18:24:44 pH5: It was never intended to be as simplistic as it currently is, it just does that because that happens to work Mar 07 18:31:23 RP: Do you have any thoughts on why linux-rp-2.6.20 won't mount a rootfs on /dev/mmcblk0p1 when a 2.6.17 build will? I have been trying to test a 2.6.20 build for openzarus-3.5.5 and even with a 15s root delay I get no joy Mar 07 18:32:08 RP : unfortunately I only can see the last screen worth of error messages as I still lack a serial cable :( Mar 07 18:32:34 RP: Ah, thanks! I (begin to) see. Mar 07 18:32:37 hvontres|poodle: No real idea I'm afraid, I'd have to test Mar 07 18:33:42 RP: ok. I'll try to rebuild one with my angstrom setup to see if it works any different. Maybe it's some wierd gcc 3.4 vs 4.1.1 thing Mar 07 18:34:33 RP: BTW, bb 1.7.999 is working very nicely. Is seems the parser got a bit quicker. Mar 07 18:34:40 LaF0rge: and we will need some sort of dummy user for the bugzilla. Some mail account where the mails ends up in /dev/null Mar 07 18:35:10 hvontres|poodle: Could be, I loose track of what was changed and where :) Mar 07 18:35:18 zecke|away: just create a new hotmail account :) Mar 07 18:52:25 re Mar 07 18:54:45 hvontres|poodle: we created a gmx account for the previous attempt and it got blocked because it looked like mailbombing :( Mar 07 18:55:17 koen: let us hope we have less build errors Mar 07 18:56:27 zecke|away: :D Mar 07 18:56:31 hm? Mar 07 18:56:41 why do you have to use gmx/hotmail/... accounts at all? Mar 07 18:57:04 laziness Mar 07 18:57:05 LaF0rge: we are poor students and don't own hardware with direct net access Mar 07 18:57:26 koen: I'm not too lazy to add a MX record, well I am Mar 07 18:57:39 zecke: ah, but you could have said that to anyone who has Mar 07 18:58:06 zecke: or maybe I'm just too much into the 'wrong' kidn of crowd were the average number of hosted machine per person is somewhat unusually high. Mar 07 18:58:27 zecke: you want an account just for the autobuild tests? Mar 07 18:58:29 LaF0rge: In the past we have experienced with tinderbox1, tinderbox2, tinderbox3 and tinderbox1,2 used email to send build logs Mar 07 18:58:46 LaF0rge: for now I would say yes Mar 07 18:59:11 zecke: Ok, that account should be more a role than something linked to your name, rihgt? Mar 07 19:00:00 LaF0rge: whatever suits you, I will certainly watch that it continues to build stuff Mar 07 19:00:33 LaF0rge: something like build-reports@host.com Mar 07 19:01:17 koen: I'm creating autobuild@imap.openmoko.org which will get an autobuild-reports@openmoko.org alias that should be used Mar 07 19:01:47 now if I would find that cyradm password ;) Mar 07 19:05:42 damn, I still haven't packaged modest Mar 07 19:10:27 koen: don't worry, i intended to do the same... Mar 07 19:11:59 florian: not packaging modest? ;) Mar 07 19:13:17 koen: not yet Mar 07 19:15:14 ciao Mar 07 19:29:36 hi all ! Mar 07 19:29:49 hi gremlin Mar 07 19:30:00 hi woglinde Mar 07 19:35:03 someone know t2-project.org ... seem to be somehitng similar to oe ... :S Mar 07 19:35:22 hm Mar 07 19:36:15 hm in the german article it said that the get support from atmel Mar 07 19:36:29 yes ... Mar 07 19:36:38 but let me see Mar 07 19:37:04 is the news i start reading and find this T2-project ... Mar 07 19:37:30 hm they use c and lua Mar 07 19:37:34 for building Mar 07 19:37:50 yes so seem ... Mar 07 19:38:19 uh they main developer are based in berlin like me Mar 07 19:38:27 whit 'similar to OE' I mean a tool to cross build target images ... Mar 07 19:38:57 ah Mar 07 19:38:59 okay Mar 07 19:39:02 heh. "Due the ever-changing nature of OE and the requirement to obtain many seperate components as well as the need to install the build tool bitmake into the system makes OE way harder to use than T2." Mar 07 19:39:02 its rocklinux Mar 07 19:39:03 hehe Mar 07 19:39:11 The T2 SDE (formerly ROCK Linux) Mar 07 19:39:38 I know them... they had a booth next to the GPE booth at Cebit two years ago... Mar 07 19:40:02 pH5 hahahah really an hard task :) Mar 07 19:40:43 I used it for a contract ... it's somewhat ok ... but not as flexible as OE or gentoo ebuild ... still have a buildroot like feels about it Mar 07 19:41:14 hehe Mar 07 19:41:20 they have no opie Mar 07 19:41:47 woglinde that could be a plus :P Mar 07 19:41:55 NO! Mar 07 19:42:11 * chouimat kicks gremlin[it] ... just for fun ...gtk should die!!! Mar 07 19:42:45 HAHAHA :) Mar 07 19:42:58 no qt-embbed Mar 07 19:43:09 gremlin[it]: I'm serious for the gtk part :) Mar 07 19:44:02 hey!!!! do you know what ?? they did a remake of the Teenage mutant ninja turtle movie yay!!!! I'm so happy that I think I will be sick Mar 07 19:44:16 oehm Mar 07 19:44:29 :) Mar 07 19:44:52 yeah, I'm only 19 but seeing that made me feel old Mar 07 19:45:01 http://www.wellingtongrey.net/miscellanea/archive/2007-01-15%20--%20science%20vs%20faith.html Mar 07 19:45:13 I'm thinking, "back in my days, we had real actors who went through the trouble of putting on real suits..." Mar 07 19:45:27 * chouimat is currently having fun with http://edu.kde.org/kturtle/ Mar 07 19:46:13 i just thought of something, how did they make all of those facial expressions? Mar 07 19:46:19 koen hehe Mar 07 19:46:26 obv. the guys inside couldn't do all that Mar 07 19:47:22 ahahahah koen :) Mar 07 19:49:57 hey guy i got an error with db4 install step ... Mar 07 19:50:17 exactly db-4.3.29-r7 task do_install Mar 07 19:50:25 | rmdir: `/media/hd1/oe/ipaq/tmp/work/arm-oabi-angstrom-linux/db-4.3.29-r7/image//usr/share/doc': Directory not empty Mar 07 19:52:22 woglinde: do you have some time for me? Mar 07 19:52:28 mr_nice yes Mar 07 19:52:48 gremlin yes see that too Mar 07 19:52:53 woglinde: how did you add network support for you simpad? Mar 07 19:53:02 mr_nice? Mar 07 19:53:14 I have pcmcia ne2000 card Mar 07 19:53:23 that works Mar 07 19:53:34 not good but it works better than usb Mar 07 19:53:50 woglinde: did you add some modules to you images? Mar 07 19:54:26 mr_nice the standard pcmcia network modules are activated in the kernel config Mar 07 19:54:42 mr_nice the pcmcia wlan cards too Mar 07 19:55:16 hyh, strange relation .. Mar 07 19:55:28 woglinde: hm, I will check it a second. thx Mar 07 19:55:56 woglinde: did you checked if the simpad wakes up? Mar 07 19:56:07 to build minimo you need glib2 and libIDL installed on host machine Mar 07 19:56:08 mr_nice no Mar 07 19:56:11 will do later Mar 07 19:57:07 woglinde: I have handed in your simpad.conf as a bugreport to oe. I hope that is ok Mar 07 19:58:03 this is directed at any/everyone: I have a magnetic card reader with an Atmel AT89S52 (an 8052 fam. micropr), and I want to yank the hex image of the program code off of the chip. I've drawn up an entire schematic, read through dozens of datasheets, and so far it looks like the best I can do is solder some 30-gauge wire onto a handful of pins, hook it up to the USB port, and write some... Mar 07 19:58:05 ...software to control the read timing. Is there a better way? Mar 07 19:59:31 mr_nice its okay Mar 07 20:03:28 woglinde: do you think it is better to exclude apm from the default config if it wont wake up? Mar 07 20:03:57 mr_nice no Mar 07 20:04:00 better fix apm Mar 07 20:04:13 woglinde: hehe Mar 07 20:06:04 re Mar 07 20:06:19 hi hrw Mar 07 20:06:51 koen, LaF0rge: do we have dfu-util native in OE or openmoko? Mar 07 20:08:35 ok. found dfu-util in openmoko Mar 07 20:11:13 clear Mar 07 20:12:52 is there any way to run rootfs image in some emulator, qemu for example ? Mar 07 20:13:05 arm compiled Mar 07 20:13:38 lkm: build for qemuarm and read RP page about qemu Mar 07 20:16:06 hi, all! Mar 07 20:16:29 how could I add not-tree files to mtn diff? Mar 07 20:16:46 this is actually to support new device. Mar 07 20:18:02 mtn add file Mar 07 20:21:44 cu Mar 07 20:23:22 hrw that was short Mar 07 20:31:44 Good night Mar 07 20:49:52 hrw|gone: I haven't merged the dfu stuff yet since my OE build machines are out of reach for the neo Mar 07 20:52:34 koen: Hello. Mar 07 20:53:41 hey sirfred Mar 07 20:53:44 koen: Do you know if the kdrive tslib input driver is oe specific or belongs to the upstream version? Mar 07 20:54:04 sirfred: iirc it is in xorg 7.2 Mar 07 20:54:17 sirfred: ph5 or pb_ should know the details Mar 07 20:54:53 koen: I've got a problem with randr resolution change I'm not sure the worse way to handle it. Mar 07 20:55:05 I mean the less worse Mar 07 20:56:38 When I change the screen to a different mode, coordinates sent by tslib don't change. This is expected, as calibration is unique, and calculated for the original mode Mar 07 20:57:58 Kdrive have a transformation matrix for mouse coordinates, but as it's made of integers, I cannot reduce them to fit a smaller mode. Mar 07 20:58:53 One solution will be to tell tslib to translate the coordinates, but doing that from the video driver doesn't look trivial (for me), other than with dirty static values. Mar 07 20:59:19 Any idea? Mar 07 20:59:47 Other than this, randr with resolution change is working on my Husky. Mar 07 21:00:53 hmmm, I have a bitbake recipe that runs a build and copies binaries into ${D}${libdir}/, but for each binary it seems to be creating a .debug folder with a copy of the binary in it Mar 07 21:01:22 why would bitbake do that? Mar 07 21:02:21 I have other bitbake recipes that run builds and copy files over that don't have this problem, could it be something unique to ${libdir} and binaries? Mar 07 21:04:57 wait, they're not the same binaries... the binaries in .debug are significantly bigger, they must contain debugging info... Mar 07 21:05:14 still, doesn't make sense to me that they would be included Mar 07 21:07:13 re Mar 07 21:07:17 koen: I built it Mar 07 21:07:37 koen: to flash neo I need to connect debugboard or not? Mar 07 21:17:06 jo zecke Mar 07 21:17:22 zecke you are changing providers fast Mar 07 21:18:56 woglinde: I'm in SteveLAN now Mar 07 21:19:25 woglinde: steve helps to search for jtag ports on the greenphone Mar 07 21:19:39 damnit now my simpad screen isnt right Mar 07 21:19:47 dont get it Mar 07 21:19:54 must be hardware issue Mar 07 21:20:05 g'night Mar 07 21:20:10 nite gremlin Mar 07 21:20:49 woglinde: did the screen ever work for you? Mar 07 21:22:16 zecke yes Mar 07 21:22:23 I told it Mar 07 21:22:24 hm Mar 07 21:22:30 last week it works Mar 07 21:22:37 today it dont works right Mar 07 21:22:54 maybee it must run out of power Mar 07 21:23:55 okay it dont wake up from suspen Mar 07 21:23:56 d Mar 07 21:25:06 hyh, minimo + task-base has size ~ 29MB Mar 07 21:27:02 zecke hm we have to fix this too Mar 07 21:27:04 pcmcia: request for exclusive IRQ could not be fulfilled. Mar 07 21:27:04 pcmcia: the driver needs updating to supported shared IRQ lines. Mar 07 21:45:05 update for anyone who might be interested in the problem I'm having, apparently OE creates unstripped versions of binaries in .debug folders, which is fine, I just don't see why they would be included in a package without being explicitly referenced Mar 07 21:50:06 regarding that .. what is the difference between release and debug in DISTRO_TYPE ? Mar 07 21:55:45 lkm: not sure, I'd like to know myself Mar 07 21:57:24 xuumbi: in my conf files it is set to debug. But as i see binaries in rootfs are already striped Mar 07 21:59:09 any hint on how to find the JTAG pins without removing the PXA? Mar 07 22:00:31 but i'm confused that rootfs with minimo only is 30 mb ... Mar 07 22:03:58 zecke: I am guessing they "forgot" to mark them on the slikscreen? Mar 07 22:06:42 03nail 07org.oe.dev * r0982ef3e... 10/ (1 packages/meta/openturbostation-packages.bb): openturbostation: Fix spelling Mar 07 22:06:46 03nail 07org.oe.dev * r90048af3... 10/ (1 conf/distro/openturbostation.conf): openturbostation: Split feed into packages and kernel Mar 07 22:06:51 03nail 07org.oe.dev * re90fceca... 10/ (3 files in 3 dirs): linux-turbostation: Update defconfig with 2 bajillion modules Mar 07 22:06:55 03nail 07org.oe.dev * r5f1842dd... 10/ (1 conf/distro/openturbostation.conf): openturbostation: Add missing + Mar 07 22:06:59 03koen 07org.oe.dev * r47d87166... 10/ (1 packages/gtkhtml/gtkhtml-3.8_3.8.2.bb): gtkhtml: add 3.8.2, has cross-compile badness Mar 07 22:07:56 yes Mar 07 22:07:59 phoneme runs Mar 07 22:13:36 woglinde: on your simpad? Mar 07 22:15:02 hvontres yes Mar 07 22:15:10 cant test qt support Mar 07 22:15:24 because of my displayproblem Mar 07 22:17:57 I discoverd a nice site effect Mar 07 22:18:05 if date is set under 1970 Mar 07 22:18:17 woglinde: use vnc to export the fb Mar 07 22:18:34 the jvm cant start bescause it hat problems with gettimeofdate Mar 07 22:18:36 woglinde: but hey, congratulation for compiling :) Mar 07 22:22:26 uh *CONGRATULATIONS: test Test completed with 411 tests passed and 0 failures Mar 07 22:29:26 woglinde: so do you have a .bb file or did you do the build "by hand" ? Mar 07 22:29:47 hvontres by hand for now Mar 07 22:30:39 woglinde: do you have some instructions somewere? I might try to see if I can get QT working on my Zarus Mar 07 22:31:03 hvontres hm Mar 07 22:31:14 which toolchain do you have? Mar 07 22:32:20 woglinde: Currently, I have a 3.4 toolchain for both .dev and .oz354 and a 4.1.1 EABI toolchain for .dev at home Mar 07 22:32:37 hm I see Mar 07 22:32:41 qt with mt? Mar 07 22:33:04 you need to get the -mr1 release Mar 07 22:33:07 woglinde: I think the EABI one is with, but I can always rebuild Mar 07 22:33:16 mr2 hasnt arm or qt support for now Mar 07 22:33:56 woglinde: hmmm. I think I did manage to get mr1 to build once, but the QT stuff wasn't working too well... Mar 07 22:34:04 hm mom I will write a short description on a pastebin site Mar 07 22:46:49 honvtres -> http://pastebin.ca/385589 Mar 07 22:46:56 args Mar 07 22:47:02 hvontres Mar 07 22:47:49 hi Mar 07 22:47:55 hi coredump Mar 07 22:49:43 koen: is there a difference between DISTRO OpenMoko and DISTRO Angstrom currently? Mar 07 22:49:43 Good night. Mar 07 22:49:50 n8 sirfred Mar 07 23:03:23 when i'm trying to build openmoko image i get an error from qemu-native because it doesn't work with gcc 4.x Mar 07 23:03:37 apt-get install gcc-3.4 Mar 07 23:03:48 or Mar 07 23:04:35 03rpurdie 07org.oe.dev * ra90e7f0a... 10/ (1 classes/tinderclient.bbclass): tinderclient.bbclass: Add hacky workaround for bitbake 1.8.x (pending rewrite in 1.9) Mar 07 23:04:41 add ENABLE_BINARY_LOCALE_GENERATION = "0" Mar 07 23:04:49 if you dont want all LOCALES build Mar 07 23:05:05 that prevents the build of qemu too Mar 07 23:05:11 if you don't want any locales Mar 07 23:05:19 he c is default Mar 07 23:05:29 hi rp Mar 07 23:05:33 hi woglinde Mar 07 23:06:24 rp did you read the mail about bitbake 1.8 and quilt? Mar 07 23:06:46 woglinde: I did, thanks. Its a known issue I'd kind of fogitten about :/ Mar 07 23:07:23 he bitbake 1.8 is really an improvement Mar 07 23:07:27 thanks Mar 07 23:08:08 woglinde: np and its not all my work :) Mar 07 23:08:37 Basically nothing in the metadata says do_patch tasks depend on quilt-native populate_staging... Mar 07 23:34:19 * NAiL has trouble cross-compiling gcc on an x86_64 host Mar 07 23:40:35 !obug 1951 Mar 07 23:40:37 duh Mar 07 23:40:41 !oebug 1951 Mar 07 23:40:43 * * Bug 1951, Status: NEW, Created: 2007-03-07 15:45 Mar 07 23:40:44 * * nail(AT)nslu2-linux.org: gcc fails to build on x86_64 Mar 07 23:40:45 * * http://bugs.openembedded.org/show_bug.cgi?id=1951 Mar 07 23:48:15 woglinde: Got it. Thanks Mar 07 23:49:46 woglinde: you are still awake? Mar 07 23:55:33 yes Mar 07 23:55:45 now building phonme from svn Mar 07 23:56:29 zecke: Would you say it would be right to redirect python functions to a logfile, the same way we do with shell ones? Mar 08 00:13:48 sometimes angstroem suckz why the hell I need pyhton-native for gtk+ Mar 08 00:19:13 woglinde: My guess would be to build the python bindings? Mar 08 00:19:52 nabyss that cant be done by the normal system pyhton? Mar 08 00:20:37 woglinde: IIRC, the aim of OE in general is to have as few external dependencies as possible.. Mar 08 00:20:37 now I remeber why I like opie more than gpe Mar 08 00:20:48 nabyss hm okay Mar 08 00:21:04 woglinde: That, and using the system python can give version hell when debugging.. i.e. bugs aren't isolated to a given mtn checkout Mar 08 00:35:13 !oebug 1947 Mar 08 00:35:14 * * Bug 1947, Status: NEW, Created: 2007-03-06 23:16 Mar 08 00:35:15 * * jakers(AT)voyagergaming.com: Bootsplash-native cannot find jpeg Mar 08 00:35:16 * * http://bugs.openembedded.org/show_bug.cgi?id=1947 Mar 08 00:36:14 Hmm redboot seems fairly promising. I'm after a boot loader that can boot squashfs directly with a kernel inside.. any hints guys? Mar 08 00:36:40 uboot with patches Mar 08 00:38:08 woglinde: I've got that on my list for research from a few days ago. Thanks :) Mar 08 00:38:41 good night Mar 08 01:05:33 03coredump2 07org.oe.dev * rce10d15c... 10/ (2 files in 2 dirs): xserver-kdrive-common: Unb0rk Xserver Mar 08 01:15:57 hmm uboot doesn't seem to have a "generic" boad config to work with. Looking at using either uboot or redboot on a custom x86 based board. Mar 08 01:59:33 For the record, this bug can be resolved, http://bugs.openembedded.org/show_bug.cgi?id=1791 Mar 08 02:00:04 http://rafb.net/p/eao6W930.html Mar 08 02:00:38 bootstrap-image for mipsel with nptl support Mar 08 02:02:39 i have the file http://www.razamicroelectronics.com/products_alchemy/ Mar 08 02:03:34 click "Register/Access Embedded Developer Site (alchemy customers only)" on the site Mar 08 02:03:57 insert your email address, then you can get the file. **** ENDING LOGGING AT Thu Mar 08 02:59:57 2007