**** BEGIN LOGGING AT Sun Aug 06 02:59:57 2006 Aug 06 03:03:39 03justinp 07org.oe.dev * r9f0b841c... 10/ (4 files in 3 dirs): e-wm: set menu autoscroll options to better defaults for touchscreens Aug 06 03:03:45 03justinp 07org.oe.dev * r8d6a9f8d... 10/ (3 files in 2 dirs): emacs: add --without-sound to fix building problems, fix FILES Aug 06 06:50:22 hi all ... Aug 06 06:54:22 i got an error doing 'mtn update' : mtn: misuse: workspace required but not found Aug 06 07:27:38 gremlin[it]: perhaps you need to do a new checkout after updating? Aug 06 07:28:51 ouch ... i hope no :( ... u sure JustinP Aug 06 07:28:51 ? Aug 06 07:33:26 gremlin[it]: or migrate your workspace....not sure if you can, though Aug 06 07:35:21 i try to take a look to monotone man page but i didn't see 'workspace migration' ... only database ... so ok re-check ... sigh Aug 06 07:39:25 gremlin[it]: if you just upgraded from pre-0.26 to post-0.26, then yeah, you have to check out any workspaces again Aug 06 07:39:57 thank njs :) Aug 06 07:41:17 (there just wasn't any way to provide migration tools, things changed too much0 Aug 06 09:03:17 03koen 07org.oe.dev * r73b98eb0... 10/ (1 packages/libwpd/libwpd_0.8.6.bb): libwpd: update to 0.8.6 Aug 06 09:19:29 03koen 07org.oe.dev * r866db4bc... 10/ (1 packages/angstrom/task-angstrom.bb): task-angstrom: ship procps till the busybox ps gets debugged and include dbus-glib till pH5's patch gets applied Aug 06 09:19:42 hey zecke Aug 06 09:25:32 he koen and zecke Aug 06 09:25:39 hey woglinde Aug 06 09:25:55 the libz staging is some sort of broken .. all packages miss functions like infalte and such .. any hints ? Aug 06 09:26:16 rob_w hm Aug 06 09:26:25 rebuild from scratch Aug 06 09:26:34 look where libz.so stays Aug 06 09:26:51 koen, ok Aug 06 09:39:47 someone removed SO_NEEDED libz.so? Aug 06 09:40:05 | tmp-dive_1.s:89:invalid character '@' in first operand Aug 06 09:51:52 gmm gmp update for OSX Aug 06 09:52:27 gmp is soo stupid Aug 06 09:52:41 with-arch=pentium-m --with-tune=prescot Aug 06 09:52:46 default apple GCC Aug 06 09:53:04 and gmp compiles itself with -march=i386 -mtune=i386 Aug 06 09:56:07 03koen 07org.oe.packaged-staging * r80e98fa7... 10/ (1 classes/packaged-staging.bbclass): classes/packaged-staging.bbclass: fix case where CROSS_DIR is being created for the first time Aug 06 10:08:39 03koen 07org.oe.dev * r2ec54635... 10/ (1 packages/e17/entice_20060416.bb): entice: add 20060416 snapshot which has some ipc fixes Aug 06 10:12:02 * zecke considers adding make check runs to our -native packages Aug 06 10:12:47 unbelievable, autoconf fails the testcase... Aug 06 10:14:10 hm Aug 06 10:14:17 struggeling with bibtex Aug 06 10:14:19 *sigh* Aug 06 10:14:27 just call it once more Aug 06 10:19:20 Quick question, before the jornada7xx arch was called jornada720. We changed this because it was misleading (jornada 7xx is 710,720,728). But I've found some references in opie to jornada720 arch, would this cause any problems? Aug 06 10:20:03 Ken|JLime: maybe some minor ones Aug 06 10:20:14 Ken|JLime: how did your touchscreen problem progress? Aug 06 10:20:48 else if (d->m_model == Model_Jornada_720 ) Aug 06 10:20:53 never got it to work Aug 06 10:21:02 started thinking that it might be due to something like this Aug 06 10:21:02 Ken|JLime: tslib utils work? Aug 06 10:21:17 Ken|JLime: no, this code is from odevice.cpp Aug 06 10:21:18 havent tried that yet :) thx for reminding me Aug 06 10:21:34 Ken|JLime: it is only for controling suspend, hardware buttons, backlight Aug 06 10:21:48 oki, was going to add hp6xx code to that so backlight works Aug 06 10:22:21 do I need to define Model_Jornada_6xx somewhere? Aug 06 10:22:34 odevice.h Aug 06 10:22:40 oki Aug 06 10:22:48 we should have an Jornada 'family' in odevice.h Aug 06 10:23:01 Ken|JLime: but ask Chicken (msg him) Aug 06 10:23:15 Ill make this as a patch against OE for now, but will send the patch to the opie team at same time Aug 06 10:23:19 will do Aug 06 10:23:57 sending stuff upstream is important Aug 06 10:24:42 yeah know, otherwise the 6xx will never get properly supportd Aug 06 10:37:46 man playing a patch monkey is not funny Aug 06 10:38:02 he Aug 06 10:38:16 own your own werksstudent Aug 06 10:38:17 haha Aug 06 10:38:49 woglinde: patching gcc 3.3.3,3.3.4,3.4.3,3.4.4,4.0.0,4.0.2,4.1.0,4.1.1,4.1.2,4.1-snapshot,4.2-snapshot Aug 06 10:39:18 hm Aug 06 10:39:43 why we dont step 3.3.3 and 3.3.4 to 3.3.6 and only maintain one 3.3 gcc Aug 06 10:40:09 dunno, we want to maintain only a kernel gcc :} Aug 06 10:48:20 oh Makefile.in of gcc3.3 is better than 3.4 Aug 06 10:48:34 but 4.x is somehwat better again Aug 06 11:02:56 03kristoffer 07org.oe.dev * rb4d676ca... 10/ (8 files in 7 dirs): (log message trimmed) Aug 06 11:02:56 base-files/base-files/jornada720/ : Moved to jornada7xx Aug 06 11:02:56 * moved so its mapped against proper ARCH Aug 06 11:02:56 initscripts/initscripts-1.0/jornada7xx/devices : Changed to better devices Aug 06 11:02:56 * changed /dev/input/ts0 to correct value Aug 06 11:02:56 linux/linux-jlime-jornada*/defconfig : Minor changes Aug 06 11:02:58 * EVDEV support for 7xx Aug 06 11:03:02 03kristoffer 07org.oe.dev * r235bdc3a... 10/ (4 files in 3 dirs): (log message trimmed) Aug 06 11:03:04 Adding Jornada Support for backlight,power... in Libopie. Aug 06 11:03:06 packages/libopie/libopie2/odevice_jornada.cpp-fixup.patch : Aug 06 11:03:08 * addition of jornada6xx with correct links Aug 06 11:03:10 * change of jornada720 -> jornada7xx to follow archs Aug 06 11:03:12 packages/libopie/libopie2/odevice.h-jornada-fixup.patch : Aug 06 11:03:14 * Add jornada6xx Aug 06 11:04:00 zecke could you look at the patches and say if they look okey? Aug 06 11:04:22 I will put that task on the stack Aug 06 11:04:55 Ken|JLime: could you please make your commits more granular? Aug 06 11:06:22 koen yeah, will do. Know I've sucked at it in the past Aug 06 11:14:46 hmmm Aug 06 11:14:59 the entice bug seems not to be in entice itself Aug 06 11:16:14 let's wait a month and play update bingo again :) Aug 06 11:17:23 I hate gcc getting rebuilt everytime theres a new patch Aug 06 11:19:00 heh? Aug 06 11:19:13 it gets rebuild everytime the PR is increased Aug 06 11:19:50 For me it rebuilds everytime a new patch is implemented Aug 06 11:20:23 and that is not true Aug 06 11:20:29 see next pushes Aug 06 11:20:51 I hadnt updated in awhile, saw some cse patch of some sort Aug 06 11:21:02 and effe updated the PR Aug 06 11:21:08 aha Aug 06 11:21:10 doh Aug 06 11:21:14 well logical then atleast Aug 06 11:21:17 to force rebuilding gcc_X.X.X.bb on the native autobuilder Aug 06 11:22:02 Oki, well shouldnt be any rebuilds more then. Seeing as initial, cross is both at 4.1.1 now Aug 06 11:22:05 which is what I use Aug 06 11:33:43 zecke: which next commits ? Aug 06 11:34:20 koen: my gcc patches etc... Aug 06 11:34:32 mtn: 27,697 | 24 | 9,141 Aug 06 11:34:40 only 859 revs left :) Aug 06 11:35:02 koen: i386-darwin/bin/ipkg-build:63: large_uid_files=`find . -uid +99 || true` Aug 06 11:35:08 hmm that line is GNUism Aug 06 11:35:10 dunno why Aug 06 11:36:15 find: -uid: unknown expression primary Aug 06 11:38:33 -user uname Aug 06 11:38:34 True if the file belongs to the user uname. If uname is numeric Aug 06 11:38:34 and there is no such user name, then uname is treated as a user Aug 06 11:38:34 ID. Aug 06 11:39:41 and what about '+' Aug 06 11:40:19 "All primaries which take a numeric argument allow the number to be pre- Aug 06 11:40:19 ceded by a plus sign (``+'') or a minus sign (``-''). A preceding plus Aug 06 11:40:19 sign means ``more than n'', a preceding minus sign means ``less than n'' Aug 06 11:40:20 and neither means ``exactly n''." Aug 06 11:40:41 but -user +1 return nothing for me Aug 06 11:41:15 NOTE: Applying patch 'fixup-aeabi-syscalls.patch' Aug 06 11:41:15 NOTE: Applying patch 'file' Aug 06 11:41:19 hmm I suck Aug 06 11:43:46 zecke: ship gnu-find-native with OE? Aug 06 11:44:07 koen: GNU getopt-native Aug 06 11:44:12 koen: GNU sed -native Aug 06 11:45:02 koen: let us use -user Aug 06 11:45:10 koen: or use python instead of os.system Aug 06 11:45:58 hey zecke, koen Aug 06 11:46:04 sups? Aug 06 11:46:42 busy playing patch monkey Aug 06 11:46:47 hey mithro Aug 06 11:48:13 zecke: actually, the '+' doesn't work as expected Aug 06 11:49:03 powerbook-wlan:/Users koen$ find . -user +0 Aug 06 11:49:04 find: -user: +0: no such user Aug 06 11:49:04 powerbook-wlan:/Users koen$ find . -user 0 Aug 06 11:49:04 ./.localized Aug 06 11:49:08 sigh 3% of the words for the paper done Aug 06 11:49:21 woglinde: nice Aug 06 11:49:28 woglinde: earn yourself a pizza Aug 06 11:49:31 nope Aug 06 11:49:32 by getting to 60% Aug 06 11:49:37 we will have barbecue Aug 06 11:49:53 *sigh* Aug 06 11:50:14 ? Aug 06 11:50:19 patch monkey? Aug 06 11:50:40 so the find in osx suckz? Aug 06 11:50:43 woglinde: no. I don't get it: You could have a pizza "fuer lau" and decide on burning stuff Aug 06 11:51:19 woglinde: this only means BSD is posix compliant but not GNU compliant Aug 06 11:51:49 hm gnu find works as expected Aug 06 11:52:06 GNU find works like GNU find Aug 06 11:52:19 wow, a miracle? Aug 06 11:52:20 I always get annoyed by the BSD utils being so minimalistic Aug 06 11:52:21 ;) Aug 06 11:52:33 zecke yes Aug 06 11:52:38 marvelous Aug 06 11:52:41 koen: they force you to use '|' Aug 06 11:52:46 and insisting on the options ot go first Aug 06 11:53:00 zecke: '|' is pretty expenive on ARM Aug 06 11:53:12 so portable scripts consume too much power Aug 06 11:53:18 koen hm because it opens a new subshell? Aug 06 11:53:28 woglinde: yes Aug 06 11:53:35 -m4: script.4s: 1: Cannot open foo: No such file or directory Aug 06 11:53:35 +m4: script.4s: 1: cannot open `foo': No such file or directory autom4te: m4 failed with exit status: 1? Aug 06 11:53:38 3. tools.at:90: 3. autom4te cache (tools.at:90): FAILED (tools.at:106) Aug 06 11:55:20 reading m4 makes my eyes bleed Aug 06 11:55:42 koen: once you are blind, you get used to it Aug 06 11:58:03 woglinde: http://www.microsoft.com/windowsvista/features/foreveryone/performance.mspx Aug 06 11:58:12 woglinde: "Windows Vista introduces a new power state called Sleep" Aug 06 11:58:21 now Windows has two power states Aug 06 11:58:23 Crash and Sleep Aug 06 11:59:18 before sleeping everything is saved to RAM Aug 06 11:59:36 Q: normally windows applications fit into a cacheline? Aug 06 12:01:33 heh Aug 06 12:01:44 maybe if you have 80MB cache Aug 06 12:01:52 not 128k like the xscales Aug 06 12:02:09 I wished the netbook had a bigger cache Aug 06 12:02:18 and windows gets faster by swapping to USB memory sticks Aug 06 12:02:39 the funny thing is that you can't disable swap Aug 06 12:02:58 well, you can but XP will become nonfunctional Aug 06 12:03:32 jikes Aug 06 12:03:42 xscaled only have 64kb cache Aug 06 12:03:45 hihi Aug 06 12:05:18 what a buzzword Aug 06 12:05:20 superfetch Aug 06 12:05:48 Windows Vista will rock! Aug 06 12:05:54 hehe Aug 06 12:06:00 on amd64 500000 ghz Aug 06 12:06:41 Quad Core AMD64 Aug 06 12:06:45 10.0000+ Aug 06 12:07:00 and two Windows Vista PCI Express Accelerator cards Aug 06 12:07:17 superfetch suckz Aug 06 12:09:42 http://www.microsoft.com/windowsvista/features/foreveryone/networking.mspx Aug 06 12:09:51 so vista will spam the net Aug 06 12:10:01 to find out if links are active or not Aug 06 12:10:03 this suckz Aug 06 12:10:19 'spam' Aug 06 12:10:31 well I use rendezvous and I don't think my net is spammed Aug 06 12:10:47 as a user I find it convient to easily access our laserjet printers Aug 06 12:11:36 woglinde: wow, Windows Vista will tune itself to receive more data Aug 06 12:12:17 woglinde: wow, the internet must have been designed on Windows Vista Aug 06 12:17:15 zecke: been pondering adding rendezvous to thousandparsec :) Aug 06 12:18:09 zecke so it means vista can lower the mtu if it regonize that it is behind ip-nat? Aug 06 12:18:18 woglinde: yes! Aug 06 12:18:29 woglinde: and it might even has one congestion algorithm implemented Aug 06 12:18:47 That wont work without implementing enough bugs Aug 06 12:18:55 woglinde: and they had to reimplement their networking stack completely to add this Aug 06 12:19:34 maybee the took the openbsd stack Aug 06 12:21:55 woglinde: or they licensed OSX Aug 06 12:22:08 hm? Aug 06 12:22:10 woglinde: now they look into ways to uglify it Aug 06 12:22:19 i tought it was bsd licenes Aug 06 12:41:13 03koen 07org.oe.dev * r63fa6491... 10/ (1 packages/libcontactsdb/libcontactsdb_0.5.bb): libcontactsdb: add 0.5 Aug 06 12:41:19 03koen 07org.oe.dev * r59db0e08... 10/ (1 packages/gpe-contacts/gpe-contacts_0.46.bb): gpe-contacts: add 0.46 Aug 06 12:42:01 so i copy .libs dir from tmp/work/armv5te/zlib/ into tmp/staging/armv5te/libs/.libs Aug 06 12:42:28 no Aug 06 12:43:13 but freetype then compile ?! Aug 06 12:43:30 rob_w: well Aug 06 12:43:39 rob_w: 1. look at the problem, identify it Aug 06 12:43:46 rob_w: 2. understand the issue Aug 06 12:43:47 .. but thnen fontconfig breaks again missing the functions Aug 06 12:43:49 rob_w: 3. fix it Aug 06 12:44:00 ok ok Aug 06 12:44:19 brb Aug 06 12:44:25 aeh I tought koen already fixed this Aug 06 12:44:49 afaik a clean rebuild solves it Aug 06 12:44:55 rob_w you are not the first one who enter this problem Aug 06 12:45:11 koen hm didnt you told this rob_w before? Aug 06 12:45:42 woglinde: yes, but in my experience people don't always follow advice Aug 06 12:45:50 woglinde: and I can be wrong :) Aug 06 12:46:04 hm Aug 06 12:50:10 wow I have sticked the finger into GMP's arse Aug 06 12:51:25 http://spooky-possum.org/cgi-bin/pyblosxom.cgi/premature.html Aug 06 12:51:26 hehe Aug 06 12:53:40 woglinde: Yes GNU tools are nice! Aug 06 12:54:02 he rms is nice Aug 06 12:54:06 so are GNU tools Aug 06 12:54:27 woglinde: well. RMS is passionate, has one goal and lives up to his standards Aug 06 12:58:22 woglinde: do you want to read his biography Aug 06 13:04:25 hm i have Aug 06 13:04:29 aeh oh right Aug 06 13:04:34 you have thsi book Aug 06 13:04:43 later Aug 06 13:05:49 Hi all, if I've created a .bb package for a SW (x11vnc) which does just compile it (no stage, no install) - does it make sense to submitt it to OE? Aug 06 13:07:04 koen@bitbake:~/OE/monotone/org.openembedded.dev/packages/libvncserver$ ls Aug 06 13:07:04 libvncserver_0.6.bb libvncserver_0.7.1.bb x11vnc_0.7.1.bb Aug 06 13:07:09 it's already in OE Aug 06 13:08:32 created a .bb for libvncserver_0.8.2 (to be used for fbvncserver-ipaq) & x11vnc_0.8.3 (to perform tests @ arm arch for the author) Aug 06 13:11:24 add them to the bugtracker: bugs.openembedded.org Aug 06 13:11:52 thanks - I'll do so Aug 06 13:17:04 freetype building for anyone? Aug 06 13:17:09 2.1.10 Aug 06 13:17:18 sh3-linux-gcc: /work/dev/donkey/build_6xx/tmp/staging/sh3-linux/lib/.libs/libz.a: No such file or directory Aug 06 13:17:18 | make: *** [/work/dev/donkey/build_6xx/tmp/work/sh3-linux/freetype-2.1.10-r1/freetype-2.1.10/objs/libfreetype.la] Error 1 Aug 06 13:18:07 isnt libz.a part of gzip? Aug 06 13:18:13 no Aug 06 13:18:21 libz.a is static library of zlib Aug 06 13:18:23 zlib Aug 06 13:18:34 smells like broken .la Aug 06 13:18:43 any suggestions? Aug 06 13:18:47 grep Aug 06 13:19:45 libz.a and libz.la is in my staging dir Aug 06 13:19:58 but not in /lib/.libs Aug 06 13:20:03 but inside /lib instead Aug 06 13:20:12 well grep the la files Aug 06 13:20:15 fot libz Aug 06 13:20:48 how you mean? Aug 06 13:21:55 bitbake libz and grep for .la? Aug 06 13:24:03 03freyther 07org.oe.dev * r67526c5a... 10/ (1 packages/autoconf/autoconf-2.59/sizeof_types.patch): Aug 06 13:24:03 packages/autoconf/autoconf: Update SIZE_OF check. Use the new $SIZE check only on cross compiling, try to use arm-linux-size from binutils Aug 06 13:24:03 Guard the new and old size_of check using if $cross_compiling test Aug 06 13:24:03 return true otherwise you use the old code. Aug 06 13:24:03 Also backport the AC_CHECK_TARGET_TOOL method from autocon-2.60 and Aug 06 13:24:06 make use of it when finding a size tool. Aug 06 13:24:09 1. identify the issue Aug 06 13:24:12 2. understand it Aug 06 13:24:14 3. fix it Aug 06 13:24:15 03freyther 07org.oe.dev * r2396228f... 10/ (22 files in 13 dirs): (log message trimmed) Aug 06 13:24:15 packages/gcc: For targets (libstdc++,java,mudflap...) use xgcc -E as CPP Aug 06 13:24:15 GCC compiles the xgcc cross compiler and will compile target Aug 06 13:24:15 libraries e.g. libstdc++. It uses autoconf to configure the target Aug 06 13:24:17 libraries which uses /usr/bin/cpp to determine the presence of Aug 06 13:24:21 host includes. Aug 06 13:24:22 By using xgcc we will not risk to include /usr/include to find standard Aug 06 13:24:27 03freyther 07org.oe.dev * r8ff7dd43... 10/ (4 files in 4 dirs): Aug 06 13:24:28 packages/glibc: Add patch, patch glibc2.4 to use $target-readelf Aug 06 13:24:31 Use $target-readelf instead of hardcoding readelf which is not Aug 06 13:24:32 present on non GNU, non binutils system. Aug 06 13:24:33 zecke:) thx Aug 06 13:24:34 03freyther 07org.oe.dev * re0115984... 10/ (1 packages/gmp/gmp-native_4.1.2.bb): Aug 06 13:24:36 packages/gmp/gmp-native: OSX/Darwin/gcc hack. Force gmp to not used assembly versions Aug 06 13:24:41 Set none as CPU to force GMP to not use 'optimized' assembly which Aug 06 13:24:43 does not compile with apple gcc (MacIntel). Aug 06 13:24:45 Messing with TARGET_SYS etc. messes with CFLAGS, PATH as well. We do Aug 06 13:24:46 not need to care for the CFLAGS for now, but we need to mess with the Aug 06 13:24:49 PATH. Aug 06 13:25:23 Ken|JLime: .libs is a custom of sucky libtools Aug 06 13:25:38 Ken|JLime: and for whatever reasons it wants to find libraries in there Aug 06 13:25:42 re Aug 06 13:25:45 Ken|JLime: this can have many reasons Aug 06 13:26:03 Ken|JLime: grep is the best target tool to find the reason Aug 06 13:26:06 re rob_w Aug 06 13:26:07 koen, but i did update and rebuild after i dumped tmp/ Aug 06 13:27:01 Ken|JLime: I'm currently at compiling gcc-cross-initial Aug 06 13:27:08 Ken|JLime: it will take some time until I'm there... Aug 06 13:27:45 ken|jlime where the libz.a coulnt bee find? Aug 06 13:28:41 freetype-2.1.10-r1 is the first to complain about missing libz.la Aug 06 13:28:53 libz.a i meant Aug 06 13:31:37 woglinde it wanted to find it under staging /lib/.libs/ but it was to be found under /lib Aug 06 13:32:21 why do you guys to refues to use grep? Aug 06 13:32:28 damn Aug 06 13:32:33 why do you guys refuse to use grep? Aug 06 13:33:18 zecke i moved the libz into staging and grepped the mentioned funtions out if them correctly but yet the apps "ignore" the libz library`s Aug 06 13:33:44 hm but I dont get why freetype needs staticle libz Aug 06 13:33:58 rob_w: that is "sticking something into someones arse"-tactic Aug 06 13:34:05 rob_w: all I ask for is analaysis... Aug 06 13:34:20 analyze the issues, at the end of the day we are engineers Aug 06 13:35:02 nm tmp/staging/arm-linux/lib/libz.a | grep inflate inflate.o: Aug 06 13:35:02 0000035c T inflate Aug 06 13:35:31 rob_w and the symbol isnt in libz.so? Aug 06 13:35:33 /work/dev/donkey/build_6xx/tmp/staging/sh3-linux/lib/.libs/libz.a: No such file Aug 06 13:35:37 thats what me makes wonder Aug 06 13:35:48 the question everyone should himself Aug 06 13:35:57 I lib/.libs/ a sensible path Aug 06 13:36:07 is lib/.libs/a a sensible path Aug 06 13:36:09 Answer: No Aug 06 13:36:15 .libs never should go into staging Aug 06 13:36:17 i think Aug 06 13:37:03 Why does freetype think libz is in that path? Aug 06 13:37:08 Answer: I don't know! Aug 06 13:37:21 study the configure.log Aug 06 13:37:23 So to find an answer to the question is to look at config.log Aug 06 13:37:28 hm I should start build Aug 06 13:37:30 something Aug 06 13:37:31 use grep on *.la files Aug 06 13:37:37 this is no rocket science as well Aug 06 13:37:41 alternative b) is Aug 06 13:37:59 ..ask zecke Aug 06 13:38:07 send your disks+computers to the BMCO Forum at the AttilaStr. and I can do the post mortem for you Aug 06 13:38:13 haha Aug 06 13:38:47 use a free version of grep, e.g. GNU grep Aug 06 13:38:52 grep for libz in .la files Aug 06 13:38:52 lol Aug 06 13:39:05 :) Aug 06 13:40:14 hm do I need a new bitbake? Aug 06 13:40:48 woglinde: why? Aug 06 13:40:56 ${@base_read_file('/devel/arm/tmp/${distro}/staging/arm-linux/kernel/kernel-abiversion')} Aug 06 13:41:00 NOTE: exceptions.SyntaxError:EOL while scanning single-quoted string (line 1) while evaluating: ${SHLIBS}kernel (${KERNEL_VERSION}) Aug 06 13:41:01 LDFLAGS_append = " -L. -lz" Aug 06 13:41:01 Aug 06 13:41:16 in zlib_1.2.3.bb Aug 06 13:41:16 rob_w: that is not what I haved asked for Aug 06 13:41:25 rob_w: either do what you are asked to or be quite Aug 06 13:41:41 i am learning , sorry Aug 06 13:41:47 randomly tossing garbage will piss me off Aug 06 13:42:30 rob_w: once again Aug 06 13:42:35 1. identify the problem Aug 06 13:42:38 2. analyze it Aug 06 13:42:39 3. fix it Aug 06 13:42:44 at which point are you? Aug 06 13:42:55 1. freetype looks for libs/.libs/libz.a Aug 06 13:43:07 why does it even look there Aug 06 13:43:18 we came to the conclusion it is not smart to look there Aug 06 13:43:22 ok Aug 06 13:43:26 16:03 < rob_w> LDFLAGS_append = " -L. -lz" Aug 06 13:43:36 why should -L. or -lz lead to lib/.libs/? Aug 06 13:43:47 the gimp is actually pretty fast on ARM Aug 06 13:43:53 (if you have enough ram) Aug 06 13:44:07 koen: hehe, I just wanted to ask that Aug 06 13:44:28 koen: on my first Laptop (200mhz mmx Pentium) I used to open many, many, many pictures with gimp... Aug 06 13:44:38 locate jpg | xargs gimp Aug 06 13:44:43 that's what I used to do Aug 06 13:44:52 rob_w: so how do you want to continue? Aug 06 13:45:02 "one day I got bored and opened 20000 images in the gimp" Aug 06 13:45:15 ok i will clean freetype and rebuild to see the errror once more .. ok ? Aug 06 13:45:32 rob_w: you are starting to piss me off Aug 06 13:45:45 rob_w: why should the issue vanish now? Aug 06 13:45:48 hey dude , thats ok i ll go for myself now Aug 06 13:46:00 why do you refuse to do basic analysis? Aug 06 13:46:10 i just want to see hte error once more ok, so i can do your steps !! Aug 06 13:46:26 find . -name '*.la' -exec grep libs '{}' \; Aug 06 13:46:59 replace '.' with tmp/staging... Aug 06 13:47:17 rob_w: you can grep now as well... Aug 06 13:48:33 zecke, ok so this now shows me what ? Aug 06 13:49:13 rob_w: Well Aug 06 13:49:28 rob_w: as this is a developer channel, I will treat you like a developer Aug 06 13:49:37 We use a compiler to compile software Aug 06 13:49:47 these software use buildsystems to build the software Aug 06 13:49:58 GNU Autoconf/Automake allows to write portable software Aug 06 13:50:13 GNU libtool allows to create, use and load libraries in a portable manner Aug 06 13:50:27 now you configure freetype, which uses GNU auto-tools to configure itself? Aug 06 13:50:35 it tries to find the zlib? Aug 06 13:50:36 * rob_w follows Aug 06 13:50:43 or it uses zlib indirectly Aug 06 13:50:53 this is the first question we need to answert Aug 06 13:51:03 GNU autoconf writes a config.log Aug 06 13:51:18 a protocol which describes the tests, results and actions it has done on configuring Aug 06 13:51:23 sure Aug 06 13:51:24 it can be used to analyse issues Aug 06 13:51:36 so findout if it mentions the zlib somehow? Aug 06 13:52:18 lemme check Aug 06 13:53:42 hecking for gzsetparams in -lz... yes Aug 06 13:53:42 checking zlib.h usability... yes Aug 06 13:53:42 checking zlib.h presence... yes Aug 06 13:53:42 checking for zlib.h... yes Aug 06 13:54:08 that is not config.log but anyway... Aug 06 13:54:56 i dont find a config.log inside the freetype .. Aug 06 13:55:47 "In aller Freundschaft" Aug 06 13:55:58 if someone tells you to open config.log and it is not present Aug 06 13:56:06 don't go for any other random file Aug 06 13:56:26 but i had a feeling to look into temp/log_do_configure .. Aug 06 13:56:29 also Aug 06 13:56:33 hehe find ./ -name "*-.log" Aug 06 13:56:37 ups Aug 06 13:56:39 hehe find ./ -name "*.log" Aug 06 13:56:54 log_do_configugre != config.loh Aug 06 13:56:55 log_do_configugre != config.log Aug 06 13:57:04 simple speech Aug 06 13:57:29 woglinde, that is new to me , sorry Aug 06 13:57:48 config.log is what configure produced Aug 06 13:58:00 log_do_configure is what oe produce Aug 06 13:58:07 d Aug 06 13:58:12 rob_w: check for config.log within the freetype directory Aug 06 13:58:15 i though those correspondent in some manner Aug 06 13:58:39 run.do_configure and log.do_configure correspondent Aug 06 13:58:54 config.log is created by the invocation of 'configure' Aug 06 13:59:03 yes because to see what configure says, normaly its on terminal and then gone Aug 06 13:59:27 hmm the freetype bbfiles looks fishy :} Aug 06 13:59:50 let me see Aug 06 14:00:15 hm no maintainer Aug 06 14:01:22 there is a freetype-2.1.10/builds/unix/config.log Aug 06 14:01:43 that is the one Aug 06 14:01:56 as ./configure is executed from within builds/unix Aug 06 14:02:06 i see Aug 06 14:04:03 To summarize Aug 06 14:04:12 We know that freetype checks for zlib and finds it Aug 06 14:04:25 hm no no Aug 06 14:04:39 looking at config.log will tell us where it finds it, and how it finds it Aug 06 14:04:43 there was problems with the normal-system libz Aug 06 14:05:07 so koen switched to the libz which is shipped with freetype Aug 06 14:05:16 thats the first configure.patch Aug 06 14:05:52 args Aug 06 14:09:55 koen are you around? Aug 06 14:10:11 I'm around Aug 06 14:11:42 koen what was excalty the problem with freetype an normal libz? Aug 06 14:12:14 the .libs thing Aug 06 14:12:24 so someone switched it to use the internal libz Aug 06 14:13:08 aeh? Aug 06 14:13:39 woglinde, but the configure patch looks to me as it doesnt change anything ,, it was like that before Aug 06 14:13:49 I thought it was something at runtime Aug 06 14:14:03 woglinde: one would have to specify --without-zlib to use the internal one Aug 06 14:14:14 woglinde: which we don't want to do for security concerns Aug 06 14:14:19 zecke yes Aug 06 14:14:21 koen@bitbake:~/OE/monotone/org.openembedded.dev/packages$ grep zlib freetype/freetype_2.2.1.bb Aug 06 14:14:22 EXTRA_OECONF = "--without-zlib" Aug 06 14:14:57 hmm 2.1.10 freetype looks more nice Aug 06 14:15:07 rob_w which version auf freetype you are building? Aug 06 14:15:20 koen: did you write this funky do_configure? Aug 06 14:15:21 right , oe tryed 2.1.10 Aug 06 14:15:30 DEFAULT_PREFERENCE = "-1" Aug 06 14:15:31 m Aug 06 14:15:31 but 2.2.1 builds just fine Aug 06 14:15:32 hm Aug 06 14:15:40 koen why you dont bumped? Aug 06 14:15:45 zecke: what do_configure? Aug 06 14:15:52 wihin freetype Aug 06 14:16:09 LIBTOOL="${HOST_SYS}-libtool" Aug 06 14:16:23 zecke: nope Aug 06 14:16:33 zecke: maybe kergoth/mallum/pb Aug 06 14:16:43 or it was inherit from your OZ buildroot thingy Aug 06 14:17:42 hm Aug 06 14:21:25 koen: BTW: it looks like we can start building on OSX Aug 06 14:22:25 zecke, that sounds cool Aug 06 14:23:18 http://rafb.net/paste/results/PIu8lu51.html Aug 06 14:24:50 rob_w try EXTRA_OECONF = "--without-zlib" in freetype 2.1.10 Aug 06 14:24:53 .bb Aug 06 14:28:11 woglinde, looks good Aug 06 14:29:36 Included konqueror to opie-image getting "konqueror: /usr/lib/libstdc++.so.6: no version information available (required by konqueror) Aug 06 14:29:57 hmm Aug 06 14:30:42 zecke, please dont take this personal, i know that things can be annoying when teaching dummy`s Aug 06 14:31:20 rob_w: hehe, well :) Aug 06 14:32:10 zecke, but i bet you have been annoying to some when u where at your starts , right ? Aug 06 14:32:38 rob_w: I have annyoed tronical (Simon Hausmann) Aug 06 14:32:50 kenj hm Aug 06 14:33:09 why the heck configure takes the host libstdc Aug 06 14:33:09 When I start konqueror I should clarify Aug 06 14:33:27 Ken|JLime: can this be a mismatch of libstdc++s? Aug 06 14:33:29 When I start on my jornada I mean, the compilation works fine. Aug 06 14:33:43 rob_w: I think woglinde has taken over this support request :) Aug 06 14:33:50 he Aug 06 14:33:58 just building glibc Aug 06 14:33:58 Ken|JLime: do you have installed another libstdc++? Aug 06 14:34:00 ath the mom Aug 06 14:34:07 :D Aug 06 14:34:08 that can last Aug 06 14:34:19 Ken|JLime: e.g. try copying the libstdc++ from staging and see how it works Aug 06 14:34:49 koen: did you add gpe-theme-clearlooks? Aug 06 14:35:04 Could this be due to different libstdc++ (staging vs jornada one) Aug 06 14:35:29 zecke: yes, why? Aug 06 14:35:40 koen: why do you write cp -d Aug 06 14:35:44 koen: and not cp -P Aug 06 14:35:49 zecke: because I suck Aug 06 14:36:02 koen: you even sit in front of a BSD system Aug 06 14:36:36 rob_w: could you pastebin your config.log please? Aug 06 14:36:59 zecke, now that i use --without-zlib ? Aug 06 14:37:24 no, the old one :) Aug 06 14:37:39 ouch Aug 06 14:37:44 hehe, don't worry Aug 06 14:38:12 i can do it for you once more .. stay tuned Aug 06 14:38:25 rob_w: I will do it myself (once there) Aug 06 14:38:32 ok thanks Aug 06 14:38:46 *food* Aug 06 14:38:58 yes Aug 06 14:39:04 steaks are grilled here Aug 06 14:39:06 woglinde: al's friend is too drunk to get pizza :( Aug 06 14:39:18 zecke lol Aug 06 14:39:18 i am already at libsm Aug 06 14:46:01 argh first it builds pango-1.10.2-r0 then it builds pango-1.13.2-r0 Aug 06 14:46:12 lol Aug 06 14:46:16 that is not good Aug 06 14:46:20 did pango-1.10 fail? Aug 06 14:46:29 no both went fine Aug 06 14:46:50 at least it did .13 after .10 ;-) Aug 06 14:51:18 is there any fair coming up where oe people attend ( in germany ) ? Aug 06 14:53:00 OEDEM Aug 06 14:53:13 I will have a 'booth' at the CCC Aug 06 14:53:16 cool Aug 06 14:53:20 at the end of december Aug 06 14:53:34 thats always berlin , right ? Aug 06 14:53:46 yes Aug 06 14:59:56 where was this picture archive from fosdem or such ? Aug 06 15:01:00 ahh NOTE: package gtk+-2.8.9: completed Aug 06 15:01:41 rob_w: there are some pictures at http://dominion.kabel.utwente.nl/koen/cms/index.php/v/Koen/confs/fosdem/?q=gallery Aug 06 15:02:36 thats the one .. Aug 06 15:04:14 i always love to check back some faces ;-) Aug 06 15:05:08 hi all ... I've a problem building qte-mt-2.3.10-r3: Aug 06 15:05:08 NOTE: package qte-mt-2.3.10-r3: task do_package: started Aug 06 15:05:08 NOTE: Couldn't find shared library provider for libgcc_s.so.1 Aug 06 15:05:45 qte-mt builds; but a subsequence built of zudoku fails Aug 06 15:06:00 because it can not link with -lstc++ (I guess) Aug 06 15:08:19 any idea what's wrong? Aug 06 15:13:31 pgfeller: well ~pastebin your error... Aug 06 15:15:50 zecke: ... I'm new to OE & IRC, soory: what is/means ~pastebin? Aug 06 15:16:51 ~pastebin Aug 06 15:16:54 methinks pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.com/ (BROKEN AND SUCKING NUTS), or http://pastebin.ca, or http://channels.debian.net/paste, or for #oe use http://oe.pastebin.com, or http://bzflag.pastebin.ca/, or http://paste.lisp.org/ for the lisp/scheme nuts Aug 06 15:20:06 thanks zecke (& ibot) ... I'll use http://pastebin.com/ - as oe. timed out. Aug 06 15:27:55 zecke: it seems as pastebin.com is under havy load (timeouts) ... is channels.debian.net/paste OK as well? Aug 06 15:28:04 sure Aug 06 15:28:08 or rafb.net/paste :) Aug 06 15:28:09 try pastebin.ca Aug 06 15:28:35 here it is: http://channels.debian.net/paste/3359 Aug 06 15:30:10 pgfeller: paste some more stuff from log.do_compile Aug 06 15:30:21 pgfeller: or did you see -frtti or -fno-rtti? Aug 06 15:32:10 full log.do_compile @ http://channels.debian.net/paste/3360 ... Aug 06 15:33:04 pgfeller: well... Aug 06 15:33:07 zecke: a grep for rtti returned no result Aug 06 15:33:17 dunno then Aug 06 15:33:24 it could be -lsupc++ Aug 06 15:33:33 it sounds more like mixing -frtti and -fno-rtti Aug 06 15:33:41 pgfeller: watchout for mickeyl in this channel Aug 06 15:33:45 he might know more Aug 06 15:35:16 zecke: thanks ... added/logged the problem as [Bug 1279] in Bugzilla & 'll try to catch mickeyl. Aug 06 16:01:04 koen: what kind of input would you be interested in for your packaged staging branch? Aug 06 16:02:48 03freyther 07org.oe.dev * r4d0a5260... 10/ (25 files in 16 dirs): Aug 06 16:02:48 packages/*/*.bb: Stop GNUism on cp. -d -> -P, -a -> pPR... Aug 06 16:02:48 Fix GNUism on the usage of cp Aug 06 16:02:48 -d as --no-derefence -> -P Aug 06 16:02:48 --deference -> -L Aug 06 16:02:48 -r -> -R Aug 06 16:02:50 -a same as -dpR -> -PpR Aug 06 16:03:33 zecke: what you'd like to see happen to give me a positive review at the final evaluation Aug 06 16:05:49 zecke: btw, I'm going to try OE on osx now Aug 06 16:07:02 koen: Well. As we knew from the beginning we would not finish everything Aug 06 16:07:15 koen: I'm interested most in decent documentation Aug 06 16:07:37 koen: What I want at the end is Having a 3-4 pages document describing the approach Aug 06 16:07:59 koen: a bit info on the current implementation, the issues/limitations of this approach Aug 06 16:08:14 koen: and most important a perspective on how to proceed from here Aug 06 16:08:55 e.g. I have thought you have used a PACKAGE_FUNC Aug 06 16:09:27 koen: on OSX uicmoc-native fails as the mac is not known as a host Aug 06 16:09:37 koen: and ipkg is failing Aug 06 16:12:54 let me put it this way: Aug 06 16:12:57 mtn: bytes in | bytes out | certs in | revs in Aug 06 16:12:57 mtn: 535.1 k | 443.4 k | 36/1528 | 10/382 Aug 06 16:13:08 hrw|gone: http://ewi546.ewi.utwente.nl/tinderbox/showbuilds.pl?tree=OpenZaurus-354x Aug 06 16:13:16 gpe/icon-list.h gone missing Aug 06 16:13:36 sounds like a too old libgpewidget Aug 06 16:13:55 http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=78&logfile=20060806073133.log Aug 06 16:14:00 it just turned up Aug 06 16:14:21 koen: macintel's are nice, the battery lifetime is too low :( Aug 06 16:15:26 hmm perl-native sucks Aug 06 16:16:00 zecke: where is the tinderbox source located on ewi? Aug 06 16:16:11 $HOME Aug 06 16:16:12 /home/tinderbox? Aug 06 16:16:17 hehe Aug 06 16:16:24 well there is a SVN directory Aug 06 16:18:16 pvanhoof: hey Aug 06 16:18:24 pvanhoof: I like reading on tinymail :) Aug 06 16:18:28 hey zecke Aug 06 16:18:34 thanks, np etc :) Aug 06 16:19:41 I was wondering .. how can I get tinymail more involved with openembedded and visa versa, zecke ? Aug 06 16:19:50 I think you are on the core team of oe, right? Aug 06 16:20:18 'core' is relative Aug 06 16:20:19 I will probably release something in a eta of < 2 months > 1 months Aug 06 16:20:35 I'm in my api rush atm Aug 06 16:20:47 pvanhoof: like I said, put up a tarball ;) Aug 06 16:20:51 making sure things are exactly the way I want it to be, getting reference counting right Aug 06 16:21:15 koen, yeah but. After carefully looking at the api changes that I really want .. I will really need a month to get it right Aug 06 16:21:42 the web stuff (people who want to use tinymail for web applications) is something that is definitely going to be post 1.0 Aug 06 16:22:08 but reference counting and renaming types is something that is going to be pre 1.0 Aug 06 16:22:31 I really don't want to confuse developers with reference counting problems when they would switch from a 1.0 to a 2.0 api :-\ Aug 06 16:22:46 pvanhoof: being a C++/Qt guy myself, I think cooperating with the GPE guys is the best thing to do Aug 06 16:22:57 pvanhoof: it will sink into OE anyway :) Aug 06 16:23:06 yep, I'm of course working together with florian and nils on this Aug 06 16:23:24 pvanhoof: and once a Qt4 based Opie lifts off, I will heavily consider tinymail Aug 06 16:23:25 well, florian and nils did some tinymail/modest work for Nokia Aug 06 16:23:49 zecke, well .. if qt comes with glib mainloop integration, it would actually be very possible to implement a libtinymailui-qt Aug 06 16:24:07 I don't think that gtypeinterface and gobject are a problem/clash for qt Aug 06 16:24:20 and I would be interested in something like that Aug 06 16:24:41 absolutley. Currently opie uses libetpan.sf.net and as things stagnates considering tinymail is currently not an option Aug 06 16:24:52 ok Aug 06 16:25:09 a libtinymail-etpan would also be possible, as a replacement for libtinymail-camel Aug 06 16:25:11 pvanhoof: doesn't qt4 already have a glib mainloop integration? Aug 06 16:25:11 pvanhoof: for the OE integration, all I/we can currently offer is doing automatic builds and minor unit testing on tinymail Aug 06 16:25:26 koen: Qt4.2 will have Aug 06 16:25:37 zecke, there's gunit unit tests in libtinymail-test Aug 06 16:25:49 zecke, but will devices have support for this integration? Aug 06 16:25:57 I was told it was an optional integration Aug 06 16:26:00 not a default Aug 06 16:26:19 zecke: I'm trying to get the tinderbox use the OE css files Aug 06 16:26:27 well, the thing is .. I can redo the parts that use gmainloop easily Aug 06 16:26:36 koen: use the sheriff admin interface Aug 06 16:26:41 but it would become qt and gtk specific Aug 06 16:26:46 which is something I prefer to avoid Aug 06 16:26:59 9.3G ../tinder.log/ Aug 06 16:27:00 wow Aug 06 16:27:12 koen: well, run the cron job Aug 06 16:27:17 actually, the gmainloop parts are already in the gtk specific implementation (libtinymailui-gtk) Aug 06 16:27:26 hmm, so it's not really a problem Aug 06 16:27:40 ok, so .. a libtinymailui-qt is perfectly possible then :) Aug 06 16:27:53 a qt/kde developer just has to lear gobject :) Aug 06 16:27:55 learn Aug 06 16:28:32 zecke, I have been thinking about doing some parts myself, to get other people started Aug 06 16:28:52 I would join, if I weren't too busy :} Aug 06 16:28:59 the thing is, I would do that. If I know people would join. the reason is because I can't maintain it Aug 06 16:29:09 for the same reason you wont join :) Aug 06 16:29:21 busy, schedule, other plans for tinymail, etc Aug 06 16:29:54 so I would certainly help with the initial stuff, getting the qt/kde developer up and running. But then the maintenance would be in his/her hands Aug 06 16:30:03 pvanhoof: sadly Trolltech is currently too busy reinventing the wheel themselves :( Aug 06 16:30:08 I'm even prepared to learn a significant amount of qt API Aug 06 16:30:26 zecke: could you send me the tinderbox login info? Aug 06 16:30:32 oh ic, they are planning to also make some sort of mail api? Aug 06 16:30:33 (I forgot) Aug 06 16:30:52 koen: what is the address you use at the bugzilla? Aug 06 16:31:05 koen@dominion.kabel.utwente.nl Aug 06 16:31:11 pvanhoof: They have this Qtopia product Aug 06 16:31:19 I'll get a @pobox.com address soon(ish) Aug 06 16:31:37 pvanhoof: and they are good at implementing (inferior) solutions themselves for this product Aug 06 16:31:49 zecke, well tell them I (the maintainer of tinymail) is interested in cooperating Aug 06 16:32:11 let is compete on the things on which we should compete :). Not on email frameworks Aug 06 16:32:16 but well, I can't force them :) Aug 06 16:32:24 s/is/us Aug 06 16:32:37 pvanhoof: that is what we/I'm/Opie devs are struggling with Aug 06 16:33:39 well, tinymail is particulary interested in commercial activities around it. I'm not dirty of it Aug 06 16:33:53 so if they want to sell a Qtopia with some tinymail integration: great Aug 06 16:34:09 koen: you should be able to login into the tinderbox Aug 06 16:34:24 koen: it will connect to bugs.oe.org to authenticate Aug 06 16:34:37 they have something like tinymail in qtopia? Aug 06 16:34:50 pvanhoof: They have sadly developed the Anti-Pattern to not listen to input Aug 06 16:35:13 right well, then that will kill them anyway Aug 06 16:35:17 nothing to worry about then :) Aug 06 16:36:18 zecke: : Insufficient privileges to sheriff tree (need sheriff tree)! at Tinderbox3/DB.pm line 41. Aug 06 16:36:19 pvanhoof: they have an mail application and a small Interface to open MBoxes Aug 06 16:36:50 so nothing much Aug 06 16:37:09 well, tell them I'm very interested Aug 06 16:37:20 remember, they don't listen :) Aug 06 16:37:35 if they listen they will say they develop something better already :} Aug 06 16:37:38 but that I can't maintain it forever. However, I'm interested in helping doing the initial stuff Aug 06 16:38:04 I will try, for Opie using Qt4 we will defintaley consider/use tinymail Aug 06 16:38:24 well, if they answer that :), I would like to see the valgrind that shows how they use less mb than tinymail for an equally large folder Aug 06 16:38:48 and foe OpenEmbedded integration we can do basic QA (e.g. build stuff regulary) Aug 06 16:38:53 pvanhoof: that is easy! Aug 06 16:39:13 well it of course depends on how you show the headers Aug 06 16:39:15 pvanhoof: just throw away every item you have read, you will need infinite amount of time, but less memory Aug 06 16:39:30 if you show 5 headers per button click Aug 06 16:39:50 which is however not what people want :) Aug 06 16:40:58 anyway, let them know I'm not an anti qt guy that wants to compete with whatever they have Aug 06 16:41:04 competing on that level doesn't work Aug 06 16:41:26 I'm rather interested in what they would need for a libtinymailui-qt Aug 06 16:41:38 or you guys Aug 06 16:42:10 and if they want to totally qt-ify that and call it kinymail .. sure guys :) Aug 06 16:42:27 lol Aug 06 16:43:31 them being c++ will also like tinymails design Aug 06 16:43:49 it's very much designed like how you would do a java or .net framework Aug 06 16:43:57 very much not like how you would do a C framework :) Aug 06 16:44:36 it starts with: everything has an interface (or in C++: everything has a abstract class where every method is abstract) Aug 06 16:44:59 and of course uses a lot of those fancy design patterns Aug 06 16:45:35 but me using that word depends on you liking or not liking people who use buzzwords :). Anyway, tinymail uses it. It doesn't buzzword about it Aug 06 16:46:27 and so, they can make a c++ class that implements the tinymail interfaces Aug 06 16:46:42 pvanhoof: one question. You tried to use XP when creating tinymail Aug 06 16:46:49 pvanhoof: who was your customer on site? Aug 06 16:46:58 well, XP not really strictly Aug 06 16:47:06 pvanhoof: which aspect of XP was the hardest to obey? Aug 06 16:47:10 if possible, I tried using XP principles Aug 06 16:47:23 well impossible was standup meetings :) Aug 06 16:47:33 I can standup and meet with myself, but .. Aug 06 16:47:51 creating unit tests before implementing didn't always happen Aug 06 16:47:51 pvanhoof: which was the hardest thing to obey/make extreme Aug 06 16:48:02 pvanhoof: which was the most easies principle to use? Aug 06 16:48:21 adding documentation while implementing was the most easy part Aug 06 16:48:48 but it's not like, easy if you don't have a team :) Aug 06 16:49:03 XP is really oriented for teams. so I wouldn't say tinymail was developed using XP Aug 06 16:49:26 easy isn't the right word. Rather: possible Aug 06 16:49:45 * zecke was just curious beacuase a lot people toss the word XP Aug 06 16:49:50 I just try to get as much of one iteration done each time: {design, implement, test, document } Aug 06 16:50:28 The principle I have picked is. Try to write code that can be tested Aug 06 16:50:33 yep Aug 06 16:50:41 but for example release often, which is a core point of XP Aug 06 16:50:44 I didn't do that Aug 06 16:51:07 because well, if I release early .. it will force me to stick to the API that I already once released Aug 06 16:51:15 yet that API is not ready Aug 06 16:51:39 I want the api to be flexible and adaptible .. but closable Aug 06 16:51:51 so strategy in stead of inheritance and stuff like that Aug 06 16:52:05 not easy to get that right Aug 06 16:54:20 anyway, *now* most of the important code as unit tests Aug 06 16:54:24 has Aug 06 16:54:49 and now that people are joining and contributing, I'm going to implement rules like: Aug 06 16:54:59 give me a unit test with your implementation Aug 06 16:55:13 or, your implementation will need to pass these unit tests, else it will simply not go in Aug 06 16:56:05 so, it's a pragmatic agile project :) Aug 06 16:57:08 gtg now. I have this one in autojoin. SO I guess cya guys Aug 06 16:57:13 girlfriend :) Aug 06 16:57:20 later! Aug 06 17:14:49 who is the guy responsible for cpan.bbclass? Aug 06 17:26:54 zecke, NOTE: build 200608061947: completed Aug 06 17:26:55 Aug 06 17:27:12 so beside the zlib thingi all went fine .. Aug 06 17:31:01 koen: | NOTE: make -j 2 LIBTOOL=arm-linux-libtool -e MAKEFLAGS= libz.so.1.2.3 libz.a Aug 06 17:31:01 | make: *** No rule to make target `libz.so.1.2.3'. Stop. Aug 06 17:33:06 zecke: weird Aug 06 17:33:13 that is OSX Aug 06 17:33:31 | Building shared library libz.1.2.3.dylib with arm-linux-gcc -march=armv5te -mtune=xscale. Aug 06 17:33:34 | Checking for unis Aug 06 17:33:36 lol Aug 06 17:33:42 yeah let us build a .dylib Aug 06 17:39:31 *moron* Aug 06 17:44:13 why isn't zlib using autoconf... Aug 06 17:48:16 zecke ask that pppd too Aug 06 17:48:24 hehe Aug 06 17:48:30 woglinde: for pppd, I know the reaon Aug 06 17:48:35 once it took me only one day to convert pppd to auto tools Aug 06 17:48:51 woglinde: but zlib uses copy and pasted auto tests Aug 06 17:49:09 woglinde: but it is doing uname -a to determine the target Aug 06 17:49:28 wtf? Aug 06 17:49:38 /devel/arm/tmp/openzaurus/work/gcc-cross-4.1.1-r6/temp/run.do_install.5371: line 586: -arm-linux-strip: command not found Aug 06 17:49:55 * zecke looks innocent Aug 06 17:50:39 ah no, I can't be quilty for that Aug 06 17:50:44 zecke: I'll modify InitialValues.pm later Aug 06 17:51:06 ah Aug 06 17:51:11 you have found it! Aug 06 17:51:23 hm? Aug 06 17:51:27 my problem? Aug 06 17:51:42 woglinde: no, koen edited css of the tinderbox Aug 06 17:51:54 woglinde: and InitValues is used by tbox3 when you create a new build Aug 06 17:52:11 woglinde: will you implement the bonsai monitors? Aug 06 17:52:30 woglinde: check where this leading '-' comes from Aug 06 17:52:46 woglinde: at first I thought my AC_CHECK_TARGET_TOOL was guilty Aug 06 17:52:54 woglinde: but it can't be quilty... :) Aug 06 17:53:27 hm Aug 06 17:53:40 he mickeyl Aug 06 17:53:41 mickeyl: hey! Aug 06 17:53:44 cheers guys Aug 06 17:54:00 ??? Aug 06 17:54:03 lol Aug 06 17:54:04 hey mickeyl Aug 06 17:55:27 *cough* Aug 06 17:56:26 pgfeller_: I can't reproduce your problem. Which g++ are you using? Aug 06 17:56:35 * koen reads http://zecke.blogspot.com/2006/08/gnu-toolchain-hacking.html Aug 06 17:56:44 djeez Aug 06 17:56:48 python is slow on osx Aug 06 17:56:59 eek Aug 06 17:57:05 dare you reading... Aug 06 17:57:13 koen: fink miscompiled its own copy of python2.3 Aug 06 17:57:22 mickeyl: decorators are awesome in py2.4 Aug 06 17:57:38 mickeyl: anyway to add them 'automatically' at runtime Aug 06 17:57:39 zecke: http://planet.linuxtogo.org/ :) Aug 06 17:57:41 haha Aug 06 17:57:47 good I catched my own fault Aug 06 17:58:00 zecke: yeah. It's my #2 big fun feature since 2.0 Aug 06 17:58:13 my personal #1 was generators Aug 06 17:58:27 * woglinde wonders why nobody stumble over it Aug 06 17:58:34 mickeyl: how to add tracing to python? Aug 06 17:58:47 mickeyl: add decorators to each method... Aug 06 17:59:06 mickeyl: I have written an @stamped decorator simulating bitbake in a set of custom scripts Aug 06 18:04:40 jesus Aug 06 18:04:46 I am so dull Aug 06 18:04:55 its only local changed Aug 06 18:06:24 any hints on jtagging a sl4 simpad? Aug 06 18:06:34 where is mr_nice when needed? Aug 06 18:09:52 hgnr Aug 06 18:09:58 koen? Aug 06 18:10:01 3 recipes/s Aug 06 18:10:07 stupid python Aug 06 18:10:19 and i thought that was a hint on jtagging Aug 06 18:11:00 hmm I think I will order a pizza... Aug 06 18:19:05 linux works on loox 420? Aug 06 18:19:37 zecke, koen - as an osx addict i would like to volunteer to help test oe on mac Aug 06 18:19:51 poushag: no testing needed ATM Aug 06 18:19:57 poushag: too much stuff breaks Aug 06 18:20:02 i would be starting from scratch - can you write up something on the wiki pls? Aug 06 18:20:10 oh too bad Aug 06 18:20:16 :( Aug 06 18:21:52 are you at least going to write it up (with the errors too) on the wiki so others can collaborate? Aug 06 18:24:59 not yet Aug 06 18:25:12 jtag> cable ppdev /dev/parport0 DLC5 Aug 06 18:25:13 Initializing Xilinx DLC5 JTAG Parallel Cable III on ppdev port /dev/parport0 Aug 06 18:25:13 Error: Cable initialization failed! Aug 06 18:25:14 once ready, it will not be different to the stuff being availavle Aug 06 18:25:33 what could be wrong? Aug 06 18:25:48 contacts not right, no power, wrong io, wrong interrupt? Aug 06 18:29:39 so you wont write it up anywhere that those of us mac addicts can keep up? (other than these irc chats) Aug 06 18:30:27 poushag: again, when it is done. There is nothing mac specific Aug 06 18:30:47 poushag: following the normal instructions will lead to success Aug 06 18:31:25 zecke: which macbook are you owning? Aug 06 18:31:37 mickeyl owning: None Aug 06 18:31:56 mickeyl having access to: black macbook Aug 06 18:32:13 mickeyl with 2gb of RAM, 80gb harddrive, 2ghz core duo Aug 06 18:32:48 hey guys Aug 06 18:32:59 id really appreciate some help here, im almost diying Aug 06 18:33:18 does the parport have to be in a special mode (ecp+eed at the moment) in order for jtag to work? Aug 06 18:33:45 and do multiple entries of parport0 in /proc/iports conflict? Aug 06 18:33:50 well the getting started wiki about building oe on mac has a totally empty page - so i set up oe on a linux vm (on pc) instead Aug 06 18:34:33 poushag: to tell you the trueth. This 'port' is at a stage where: Aug 06 18:34:44 poushag: the people able to help will find their way through anyway Aug 06 18:34:54 poushag: the people not able to help will only cost time to support Aug 06 18:35:47 zecke: im very sorry to beg and intervene, could you just spend 10seconds of your time to a lonely unhappy soul, without its toy? Aug 06 18:36:10 x29a: sure, but I don't have an answer Aug 06 18:36:29 are you not into jtagging alldaylong? Aug 06 18:36:52 not frequently :) Aug 06 18:36:54 ok i will be patient - sry for being too excited - i guess im enjoying oe and osx so much separately i cant wait to combine the two! :D Aug 06 18:36:57 x29a can you read german? Aug 06 18:37:05 woglinde: yeah Aug 06 18:37:14 even write and speak Aug 06 18:37:28 maybe in this forum you have luck Aug 06 18:37:30 http://palmtopmagazin.de/board/tablett/ Aug 06 18:37:44 poushag: it is as easy as installing: XCode (gcc), GNU sed, GNU getopt, SVN, wget Aug 06 18:37:50 its searchable Aug 06 18:37:52 zecke: how long can you operate it when it's fully charged? Aug 06 18:37:54 poushag: and then the normal getting started Aug 06 18:38:07 mickeyl: when compiling and both cores are used Aug 06 18:38:11 1 1/2 hours Aug 06 18:38:35 heh Aug 06 18:38:49 mickeyl: I'm not sure how long it will take on a normal load Aug 06 18:38:54 woglinde: yeah, im into it, some hits on "jtag" Aug 06 18:38:59 im so disappointed Aug 06 18:39:00 I can see that intel saves more power as the ppc chips Aug 06 18:39:02 x29a I know Aug 06 18:39:07 Aug 06 18:39:32 koen: I hate intel :} Aug 06 18:42:44 i already installed xcode some time ago - as for the other items i guess i can go hunt them down myself - but what you said zecke is exactly what the wiki should say - i suppose i will fill in the wiki myself then - a good wiki writeup will encourage more osxers to try it - rather than giving the impression that nobody has done it yet Aug 06 18:43:24 poushag: this is Free Software. We do not people trying it. We need people doing stuff Aug 06 18:43:44 poushag: but feel free to write the Wiki site Aug 06 18:43:51 poushag: and ask me if you have questions... Aug 06 18:43:53 fwiw, there should be such a page already in the wiki Aug 06 18:44:00 see the bottom of GettingStarted Aug 06 18:44:21 k Aug 06 18:44:52 koen there is - its empty - but not for long (as i quote zecke on it) Aug 06 18:47:45 re gremlin Aug 06 18:51:47 hi woglinde Aug 06 18:59:09 ~lart fontconfig Aug 06 18:59:10 * ibot DoSes fontconfig Aug 06 19:00:33 YEAH Aug 06 19:00:34 it works! Aug 06 19:00:50 x29a fine :) Aug 06 19:00:51 woglinde: entering parallel adress directly did the trick Aug 06 19:05:25 koen: fontconfig compiles and links fc-case Aug 06 19:06:02 zecke: and? Aug 06 19:06:17 koen: using CXXFLAGS and LDFLAGS from the target compiler Aug 06 19:06:25 koen: and we already build fontconfig-native Aug 06 19:14:16 koen, this user man page refs mtn 0.25: http://www.openembedded.org/user-manual&dpage=ch03s02 Aug 06 19:14:55 you want me to file a oe bug on it? Aug 06 19:15:34 (i dont see a link for editing it) Aug 06 19:15:46 poushag: org.openembedded.documentation checkout Aug 06 19:17:16 hm why needs opie-image gtk Aug 06 19:19:02 *sigh* I hope angstroem is soon enough simpad ready Aug 06 19:19:30 03koen 07org.oe.documentation * r09a345a2... 10/ (1 usermanual/usermanual.xml): usermanual/usermanual.xml: update to match monotone migration Aug 06 19:20:08 hehe Aug 06 19:20:17 * koen hugs sed Aug 06 19:20:20 hehe Aug 06 19:20:23 poushag: You are free to help us to improve documentation Aug 06 19:20:23 sed rockz Aug 06 19:20:26 koen: GNU sed? Aug 06 19:20:30 lol Aug 06 19:20:38 gsed Aug 06 19:21:01 powerbook-wlan:~/Projects/OpenEmbedded/org.openembedded.documentation/usermanual koen$ sed --version Aug 06 19:21:02 GNU sed version 4.1.4 Aug 06 19:21:02 gsed,ggrep,gfind,gcp,gmake,gdiff,gpatch Aug 06 19:21:18 gawk Aug 06 19:21:26 koen: have you ever used FileVault? Aug 06 19:21:48 zecke: I tried to , but OSX said I didn't have enought diskspace Aug 06 19:22:02 i have been doing so on the wiki already - the docs will get some attention from me later i suppose Aug 06 19:22:09 $HOME >37GB Aug 06 19:22:11 koen: ah Aug 06 19:22:23 koen: it needs du -h + GBs to create the DMG Aug 06 19:24:14 zecke: It would be need if we could hookup 'make xhtml ; cp xhtml/* /var/www/oe/usermanual/' into ciabot Aug 06 19:25:08 s/need/neat/ Aug 06 19:25:43 03freyther 07org.oe.dev * r9e7caafb... 10/ (5 files in 3 dirs): (log message trimmed) Aug 06 19:25:43 clases/cpan.bbclass,packages/perl: add EXTRA_CPANFLAGS, fix native Aug 06 19:25:43 Add EXTRA_CPANFLAGS to cpan.bbclass, similiar to EXTRA_OECONF Aug 06 19:25:43 and EXTRA_OEMAKE Aug 06 19:25:43 Use this in libxml-parser-perl-native to point to the expat Aug 06 19:25:44 include and libraries Aug 06 19:25:46 perl-native inherits native already. No need to use BUILD_CC, Aug 06 19:25:50 03freyther 07org.oe.dev * rc730cb69... 10/ (4 files in 3 dirs): Aug 06 19:25:52 packaes/zlib: Do not use uname to determine the library ending Aug 06 19:25:54 Do not use uname to decide if the library ends with dylib Aug 06 19:25:56 or so. And where to put the Version numbers... Aug 06 19:25:58 For darwin unset LDSHARED as we do not have gcc -shared to Aug 06 19:26:00 zlib should use autofoo to decide such things... Aug 06 19:26:28 hmm /me bitches keithp to fix fontconfig :} Aug 06 19:27:20 zecke, which version of Xcode are you using for oe? Aug 06 19:27:37 gcc version 4.0.1 (Apple Computer, Inc. build 5341) Aug 06 19:31:42 hmm Aug 06 19:31:45 gcc version 4.0.0 20041026 (Apple Computer, Inc. build 4061) Aug 06 19:34:32 uda1380 support changed alot from .12 to .16 .. Aug 06 19:35:38 well i thought i installed xcode before - but i actually didnt yet - i will pull the gcc pkg from my tiger dvd (whichever version that one is) Aug 06 19:36:41 unless you think for some reason thatone is too old Aug 06 19:40:10 zecke: powerbook-wlan:~/Projects/OpenEmbedded/build/tmp/angstrom/staging koen$ ls Aug 06 19:40:10 arm-angstrom-linux-gnueabi powerpc-darwin Aug 06 19:40:11 okay, more OSX stuff next week Aug 06 19:40:36 koen: next step will be universal toolchains Aug 06 19:41:01 that'd be cute Aug 06 19:41:21 koen: well it is as simple as -arch ppc -arch i386 Aug 06 19:41:33 koen: but I wonder what autotools will do... Aug 06 19:44:29 hm okay Aug 06 19:44:36 now I trigger the libz error too Aug 06 19:45:57 koen: Do you think writing this documentation and pushing your stuff is feasible? Aug 06 19:46:17 zecke: ? Aug 06 19:46:20 you've lost me Aug 06 19:47:03 koen: for your packaged staging stuff Aug 06 19:47:25 zecke: docs yes, what do you mean with 'pushing stuff'? Aug 06 19:47:28 koen: I have written that I would like to see documentation, documentations Aug 06 19:47:40 koen: oh okay Aug 06 19:47:49 koen: I asked if you used PACKAGE_FUNC Aug 06 19:47:56 koen: and you said you were pulling... Aug 06 19:48:07 zecke: pulling on OSX :) Aug 06 19:48:13 I thought you would push it afterwards Aug 06 19:48:23 I synced :) Aug 06 19:48:32 yes, I used a PACKAGE_FUNC Aug 06 19:48:39 but? Aug 06 19:49:43 s/used/am using/ Aug 06 19:49:45 http://www.openembedded.org/repo/org.openembedded.packaged-staging/classes/package_ipk.bbclass Aug 06 19:50:31 haha now libz compiles Aug 06 19:50:35 aeh freetype Aug 06 19:50:38 with libz Aug 06 19:50:42 will check it in Aug 06 19:51:02 koen: but that is ipkg's one? Aug 06 19:51:10 koen: you can do PACKAGE_FUNCS += my_stuff Aug 06 19:51:29 zecke: you mean inside packaged-staging.bbclass? Aug 06 19:51:56 yes Aug 06 19:52:28 and documentation on what you have done would be nice (well we have the revs...) Aug 06 19:52:38 noted Aug 06 19:54:23 heh Aug 06 19:54:37 OE on osx tries to build ncurses without a toolchain Aug 06 19:54:50 elite Aug 06 19:55:17 not here... Aug 06 19:58:57 go cia-bot Aug 06 19:58:59 zecke, did you use fink to install gnu stuff (sed, etc)? Aug 06 19:59:08 admitingly yes Aug 06 19:59:48 sudo fink install quilt :) Aug 06 20:00:40 03woglinde 07org.oe.dev * r672032a9... 10/ (3 files in 2 dirs): Aug 06 20:00:40 packages/freetype: fix build without internal libz Aug 06 20:00:40 Some user reports that building freetype without the included Aug 06 20:00:40 libz breaks there build. They are right. Justify the remaining Aug 06 20:00:40 freetype.bb like freetype_2.2.1.bb Aug 06 20:01:15 justify? Aug 06 20:01:23 he Aug 06 20:01:23 I do not like the freetype files Aug 06 20:01:26 they look broken Aug 06 20:01:31 it works Aug 06 20:01:34 They should use the system zlib Aug 06 20:01:41 and not a broken, insecure copy Aug 06 20:01:51 they do now Aug 06 20:02:05 :) Aug 06 20:02:12 thanks koen Aug 06 20:02:26 hmm, I need to order a pizza... Aug 06 20:02:30 hm Aug 06 20:02:33 its to late Aug 06 20:02:39 koen: Aug 06 20:02:40 NOTE: package ncurses-5.4: completed Aug 06 20:02:40 NOTE: build 200608062217: completed Aug 06 20:02:40 Build statistics: Aug 06 20:03:27 so do you have to run x11 to build oe on osx? Aug 06 20:03:28 koen: do you have an EABI toolchain on OSX? Aug 06 20:03:45 poushag: no Aug 06 20:03:54 k thx Aug 06 20:06:16 | checking for xml2-config... /sw/bin/xml2-config Aug 06 20:06:18 argh Aug 06 20:06:59 hehe Aug 06 20:07:33 zecke we have to play lotto Aug 06 20:07:36 woglinde: the good thing about the OSX stuff is. The stuff that fails to compile with uclicb are likely to fail with a BSD host as well Aug 06 20:07:52 with luck we win enough money to work all the time on oe Aug 06 20:08:02 zecke: time to expand the no host includes patch? Aug 06 20:08:14 koen: it has /sw/include in it already Aug 06 20:09:09 koen: I should make these paths configurable... somehow... Aug 06 20:11:57 is there an alternative to using fink to install the gnu stuff? like darwinports or something? Aug 06 20:12:21 poushag compile by hand? Aug 06 20:12:44 hehehe - yeah i guess it will be fink then Aug 06 20:12:50 poushag: sure. darwinports should work as well Aug 06 20:13:03 woglinde, are you an osx user too? Aug 06 20:13:03 poushag: dports didn't work on MaxIntel Aug 06 20:13:27 mines a G4 PPC dual 1GHz Aug 06 20:13:37 dual? Aug 06 20:13:56 yup 2 processors Aug 06 20:14:07 poushag: bastard! Aug 06 20:14:09 it was almost top of the line when i got it Aug 06 20:14:12 * zecke needs a quad core g5 Aug 06 20:14:55 http://www.apple.com/xserve/ Aug 06 20:15:03 the nextone up from mine was 2x1.25ghz w slightly better gfx (128mb cache i think) Aug 06 20:15:23 but i dont regret my choice - had to save a little $ Aug 06 20:16:24 i want to drop by my old university and see the 1100 node g5 supercomputer cluster (at Virginia Tech) - maybe you guys heard of it Aug 06 20:16:26 http://images.apple.com/xserve/images/index_rack_010604.gif Aug 06 20:16:27 nice Aug 06 20:17:22 i believe it was announced on slashdot when completed - do you all recall reading about it? Aug 06 20:17:58 they started with regular g5s then the following year upgraded to g5 xserves (still 1100 nodes) Aug 06 20:17:59 I don't read /. often Aug 06 20:18:42 i wanted to buy one of the retired g5s - but apple was reselling them for too high a price (for me at the time) although a little cheaper than new ones of course Aug 06 20:21:42 mickey|tv: tcl includes/detects my host... Aug 06 20:22:12 nite Aug 06 21:07:54 03freyther 07org.oe.dev * r37507395... 10/ (3 files in 3 dirs): (log message trimmed) Aug 06 21:07:54 packages/quilt: Fix build of quilt 0.45 Aug 06 21:07:54 The aclocal.m4 is an automatically generated file, created by aclocal.m4. Aug 06 21:07:54 The OpenEmbedded project automatially removes this automatically generated Aug 06 21:07:54 file to regenerate it using aclocal. Aug 06 21:07:56 Readd the content from aclocal.m4 as acinclude.m4 which will be seen by Aug 06 21:07:58 the same set of programs... Aug 06 23:30:37 grr, cant get a successful build. terrible luck with those lately Aug 07 01:03:50 | checking host system type... arm-unknown-linux-gnu Aug 07 01:03:50 | checking sysdep dirs... configure: error: The arm is not supported. Aug 07 01:03:50 | FATAL: oe_runconf failed Aug 07 01:03:50 NOTE: Task failed: /home/kergoth/code/oe/build/tmp/work/armv5te-linux/glibc-2.3.5+cvs20050627-r8/temp/log.do_configure.15185 Aug 07 01:03:57 * kergoth boggles Aug 07 02:21:35 * kergoth sighs Aug 07 02:37:47 * ThalanMel|oot greets Aug 07 02:37:51 hi Aug 07 02:41:01 I hope this is the right chan, I've been trying to upgrade from PalmOS 4.5 to v5, but nobody seems to provide actual downloads or anything. Then I found OpenEmbedded. Sounds interesting. Aug 07 02:41:56 OE's for linux on PDAs, not PalmOS Aug 07 02:44:19 Ah woe Aug 07 02:44:39 I thought I read something about linux on palm devices **** ENDING LOGGING AT Mon Aug 07 02:59:56 2006