**** BEGIN LOGGING AT Fri Mar 29 02:59:59 2013 Mar 29 08:35:07 good morning Mar 29 10:01:50 gtk-demo doesnt show arabic or chinses text. Do I need to add font packages ? Mar 29 10:10:45 jkroon: I'm not sure, but have you installed the appropriate -locale-* packages? Mar 29 10:11:18 jkroon: setting IMAGE_LINGUAS appropriately should ensure that all of these get installed Mar 29 10:13:28 bluelightning, ah Mar 29 10:13:43 bluelightning, ill try that, thanks Mar 29 10:21:00 ... but where do I find what the corresponding fields should be for arabic or chinese.. I know "sv-se" is for swedish, and "ja-jp" is japanese I guess.. Mar 29 10:25:21 nm, found it Mar 29 10:41:41 Is using "ROOTFS_POSTPROCESS_COMMAND += 'install_linguas;'" deprecated ? Mar 29 10:49:24 jkroon: that's automatic now Mar 29 10:49:35 ok Mar 29 10:49:47 (assuming IMAGE_LINGUAS is set that is) Mar 29 10:52:23 cool Mar 29 11:12:43 STOP flooding! Mar 29 11:18:36 Hi Mar 29 11:20:13 Here a small question. In some .inc I have: Mar 29 11:20:24 INITSCRIPT_PACKAGES = "foo bar ohoh" Mar 29 11:20:33 What would be the best/short way to remove "ohoh" package from list and get Mar 29 11:20:38 INITSCRIPT_PACKAGES = "foo bar" Mar 29 11:42:15 I cant seem to include any ttf font supports Arabic :-( For japanese I found ttf-wqy-zenhei, but for instance ttf-droid will not get included properly Mar 29 12:54:21 matt3535: you can use oe_filter_out... "git grep" for examples Mar 29 12:54:33 matt3535: however sometimes it's just better to set the value Mar 29 12:56:14 matt3535: with the possible exception of unicode fonts I guess we may not have support for arabic Mar 29 12:56:38 matt3535: if you can find the font itself elsewhere it would not be difficult to add a recipe for it though Mar 29 12:59:16 * jkroon is missing the pango-module-arabic-lang package in the image.. Mar 29 12:59:48 I guess I need the pango shapers to start with.. Mar 29 13:17:45 bluelightning, yeah, gtk-demo on my Fedora 18 box renders the arabic text.. Mar 29 13:18:30 but how would I know which font is being used for the text rendering.. Mar 29 13:20:14 that is a good question... I'm not sure to be honest Mar 29 13:26:33 jkroon: strace shoud tell you that Mar 29 13:31:15 mckoan, wont strace give me all the fonts gtk-demo wants to load ? Mar 29 13:32:33 jkroon: maybe, I don't know the code of gtk-demo Mar 29 13:34:41 jkroon: you could try creating an hello world app in Fedora and investigate how does it work, then port it on OE Mar 29 13:36:03 yeah Mar 29 13:37:41 well, I did a "rpm -qfi /usr/share/fonts/* | grep Arabic", it gives me one rpm, paktype-naqsh-fonts, which has a .ttf file in it. Ill try shipping it with the image Mar 29 13:40:47 can I add something to the SRC_URI based on the call convention? like, SRC_URI_callconvention-hard += "hardfp.patch" ? Mar 29 13:41:00 does this work? Mar 29 13:44:29 yes! it worked! Mar 29 13:44:35 * jkroon dances Mar 29 13:47:01 jkroon: تهنئة Mar 29 13:47:43 mckoan, im swedish, i dont understand a word of arabic.. but i customers want it :-) Mar 29 13:47:59 s/i customers/our customers Mar 29 13:48:27 jkroon: :-D Mar 29 13:48:29 @bluelightning: thanks this is exactly what I want ! Mar 29 13:48:54 jkroon: I know just a few words Mar 29 13:49:04 mckoan, what did that one mean then ? Mar 29 13:49:16 jkroon: congratulations ;-) Mar 29 13:49:28 mckoan, aha :-D Mar 29 13:50:51 now I just need a font for Hebrew.. Mar 29 13:55:29 HI All Mar 29 13:55:39 i am getting this error MojDbTextUtils.cpp:38: undefined reference to `u_errorName_48' Mar 29 13:55:42 karan: hi Mar 29 13:55:48 what might be the error Mar 29 13:56:10 i am trying for it Mar 29 13:56:16 karan: maybe you should try to explain what you are doing Mar 29 13:56:23 same is compiling in my normal make system Mar 29 13:56:34 when i build in oe-core its giving this error Mar 29 13:57:23 same make file in oe core not working Mar 29 13:57:31 this is CMake module Mar 29 14:01:02 please can any one help Mar 29 14:09:36 karan first you do not tell us what recipe this is Mar 29 14:10:05 karan second it is a missing symbol at link time Mar 29 14:11:00 okok Mar 29 15:21:00 ever since I started developing with OE, I'd have a real use for a 16-core machine with 64 GB RAM and an enterprise-grade SSD. Mar 29 15:21:13 -j 16 would be awesome :) Mar 29 15:24:32 * florian knows this feeling Mar 29 15:26:45 anyone know of a way to test the "sync" command? I'm writing from my embedded Linux (booted from NAND flash) to a microSD card and it seems like "sync" doesn't always guarantee that everything has been written to my microSD card. Mar 29 15:26:59 it can't Mar 29 15:27:17 the sd card also has a controller that writes to flash in the card :) Mar 29 15:27:32 is there anyway that *I* can know? Mar 29 15:27:59 do I have to use the driver for the card to check some register for "all bytes written" or something like that? Mar 29 15:28:15 it's the card itself, not the driver Mar 29 15:28:35 I can't be the first person to encounter this problem, though Mar 29 15:28:50 you are not Mar 29 15:28:57 perhaps a low-level "flush" command that the driver can send to the card? Mar 29 15:29:09 how do others deal with this? Mar 29 15:30:00 My manufacturing facility is powering on devices for the first time before final assembly for a smoke test. The device is creating initial files, but they don't get fully written before the device is powered off. Mar 29 15:30:33 the worst part is that the writes are minuscule in size. Probably 10-20 files all totaling less than 10kiB Mar 29 15:30:50 ah, so you have a known batch of exactly one card? Mar 29 15:32:08 we buy the same model of card in batches, but we've bought more than one batch Mar 29 15:32:15 but they all seem to suffer the same Mar 29 15:32:48 and of course, it's kind of random since it depends on the timing inside the card Mar 29 15:33:06 it seems to be caching writes… (makes sense now that I think of it) Mar 29 15:33:27 yes, but if you have a known set of card-types you can better accomodate the card firmware behaviour Mar 29 15:34:01 e.g. find out if you can send the card a eject, power-off, etc command that makes it flush gracefully Mar 29 15:35:04 thank you. What should I be googling for to find these commands? or should I start with the driver and see what kind of commands it knows how to send? Mar 29 15:35:20 sorry, no idea Mar 29 15:35:34 but one thing to check would be the sd-card spec Mar 29 15:35:41 thx. of course, testing this will be a pain Mar 29 15:35:45 good idea, Mar 29 15:35:46 . Mar 29 16:42:17 moin Mar 29 17:53:01 i just read about windrive CGP linux based on yocto ? so is tehre a meta-layer from windriver somewhere ? Mar 29 17:53:11 s/windrive/windriver Mar 29 20:17:45 master seems to rerun a lot of tests now Mar 29 20:43:48 anybody got a workable technique to enforce some test discipline? Mar 29 20:44:18 we keep saying we're not going to do it again, but every time it gets a little worse Mar 29 20:44:51 the whole gotta-get-these-features-in-at-the-last-minute thing... Mar 29 20:45:18 which ends up squeezing the time (and life) out of test Mar 29 20:45:28 just wondering... Mar 29 21:03:57 mr_science: Time based releases help a lot with that. Mar 29 21:27:04 meaning what exactly? Mar 29 21:28:28 got a pointer URL or something? Mar 29 21:35:13 nm, if it's what i think it is, then it doesn't fit well with our "promised" features... Mar 29 21:47:13 broonie: but you're right, that's probably what we *should* be doing... **** ENDING LOGGING AT Sat Mar 30 02:59:59 2013