**** BEGIN LOGGING AT Sat Jun 03 02:59:56 2006 Jun 03 03:00:49 v8jlene: thanks! can I abuse your help a little? If I use ASSUME_PROVIDED += "glibc", it still is trying to build glibc Jun 03 03:01:03 must I put the whole version too? Jun 03 03:01:50 thejapa, glibc will be built regradless Jun 03 03:02:00 it is used in the cross tools Jun 03 03:02:09 regardless* Jun 03 03:02:30 emte: but what about native? Jun 03 03:02:39 Gcc and glibc are rather complicated... Jun 03 03:02:46 what about it Jun 03 03:03:09 If you look in the gcc .bb's you see that they have things like: PROVIDES = "virtual/${TARGET_PREFIX}gcc virtual/${TARGET_PREFIX}g++" Jun 03 03:03:21 OE makes the correct assumption that your target will not have the same glibc version as your host Jun 03 03:03:26 oh, I guess that in native there is no need to build glibc... it's already there, isn't it? Jun 03 03:03:27 ah Jun 03 03:03:34 no Jun 03 03:03:41 native means native to target Jun 03 03:03:45 ok Jun 03 03:03:53 i'm going crazy here. Jun 03 03:03:56 you get three stages at most Jun 03 03:04:17 intital cross, cross, and target Jun 03 03:04:29 If you have a local compiler you can assume provided for it if you want. If you have a cross-compiler you can assume provided for that as well.. etc.. Jun 03 03:04:30 it is breaking with a weird message about no rule to make target install in glibc. Jun 03 03:04:46 occasionally you can skip one or more by using qemu IF your distro supports it Jun 03 03:05:21 since qemu emulates your target Jun 03 03:05:29 Look here: http://www.openembedded.org/user-manual&dpage=ch05s05 Jun 03 03:05:56 yes, you can use any compiler you wish Jun 03 03:06:06 Note that using pre-built gcc's and glibc's is rather compilcated and since I haven't done it I can't really help much more than that... Jun 03 03:06:25 2.95 is still used for the ancients taht use the original sharp rom on thier zaurii Jun 03 03:07:14 in any case you should really let OE build them for you Jun 03 03:07:34 it will be easier to troubleshoot runtime problems Jun 03 03:07:43 i see. Jun 03 03:08:55 since you would know what options they were compiled with and the actual version Jun 03 03:09:04 as well as special patches etc Jun 03 03:10:30 i've sent a message to oe@hh.org... i've pasted it here. http://oe.pastebin.com/754887 I've been banging my head for 2 days already Jun 03 03:10:59 it's about this error, would you help me just a bit more? :) Jun 03 03:11:55 who set your cvs date? Jun 03 03:12:13 and why are you using unstable? Jun 03 03:12:34 your trying to build familiar for x86? Jun 03 03:12:38 i didn't set cvs date Jun 03 03:12:48 * emte is bewildered by why your trying to do Jun 03 03:13:10 s/why/what Jun 03 03:13:25 i thought i should build unstable, because i want to test cvs stuff. Jun 03 03:13:45 and familiar is because I have a h2200, and want to test it there later. Jun 03 03:13:58 hehe, as you can see, i'm totally new to this thing. Jun 03 03:14:08 you have i686 as your target which isnt even a valid machine Jun 03 03:14:22 i doubt unstable will compeil Jun 03 03:14:26 compile* Jun 03 03:14:50 and "native" as a macine is not valid either Jun 03 03:14:54 machine* Jun 03 03:15:38 so, i must set cvs date Jun 03 03:15:50 you need to fix your whole system Jun 03 03:16:01 MACHINE = "h2200" Jun 03 03:16:21 comment out TARGET _ARCH Jun 03 03:16:51 and i highly suggest you dont use unstable Jun 03 03:16:58 but taht is up to you Jun 03 03:17:43 TARGET_ARCH I believe has been set by setting MACHINE="native", isn't it? Jun 03 03:18:19 they are interdependent Jun 03 03:18:42 ah Jun 03 03:19:29 the thing about native was because I will try to develop things on my PC, and I wanted to run them on it before trying on h2200. Jun 03 03:20:11 Actually, it was hrw that told me to use it, but I never asked him if this was just for opie stuff. Jun 03 03:20:20 france: you around? Jun 03 03:20:43 why type of things? Jun 03 03:20:47 what* Jun 03 03:21:34 hehe, i got a weird idea about using a single backend for opie pim data and gpe pim data Jun 03 03:22:00 all in the same sqlite databases, so we can easily test both environments Jun 03 03:22:12 good luck, have you tried to talk to thier authors about that? Jun 03 03:22:20 erm. no Jun 03 03:22:44 i belive taht opie and gpe use different storage systems Jun 03 03:22:50 yep. Jun 03 03:22:53 not too sure tho Jun 03 03:23:09 opie uses sqlite 3 or xml, but sqlite3 is the preferred. Jun 03 03:23:16 gpe uses sqlite 2. Jun 03 03:23:42 it's my summer of code application Jun 03 03:24:54 actually, it was very lucky for me to be selected, i thought they wouldn't do it because i said i've never done linux development, so there's tons of things to learn Jun 03 03:25:39 i've done tons of windows dev though, but open source vs proprietary development is way different Jun 03 03:27:50 that's why i ask lots of dumb questions. :) Jun 03 03:31:32 * mithro pokes france Jun 03 03:37:09 thejapa: I haven't had a lot of luck targetting x86 - always seems to be gcc and/or glibc problems. I found that MACHINE="epia" works and that's what I currenty use. Jun 03 03:39:33 v8jlene: great Jun 03 03:39:53 i'll try it right now. What is epia? :D Jun 03 03:40:28 mini pc? Jun 03 03:41:25 thejapa: http://www.via.com.tw/en/products/mainboards/ Jun 03 03:41:43 thejapa: So yeah, the mini-ITX and non-ITX boards from via or EPIA. Jun 03 03:42:12 non-ITX? I meant nano-ITX... so they are x86 Jun 03 03:43:17 I'll try it. :) Jun 03 03:43:39 it will build i586 Jun 03 03:44:01 it has TARGET_ARCH = "i586" Jun 03 03:46:19 thanks a lot emte and v8jlene! you helped a lot more than you think :) Jun 03 03:46:30 i must go. Jun 03 03:53:21 mithro: hi Jun 03 03:53:39 france: you have email now! :) Jun 03 03:53:48 do you know where/when the OE website moved/moving Jun 03 03:54:29 mithro: they did it on 05/27/06, but did not tell me, so I know none of the details, but do a nslookup of www.openembedded.org Jun 03 03:54:47 mithro: it is some site in germany Jun 03 03:56:31 mithro: when I found out a day or so ago, I asked and kergoth said to delete everything in /home/oe as it was not needed. Jun 03 03:58:47 mithro: e-mail recieved, thanks! Jun 03 04:01:48 i must say the mailing lists have gone quite quiet Jun 03 04:02:09 and the rss from the servers as well Jun 03 06:25:59 emte: you on the OE list? Jun 03 07:16:51 good morning OE ! Jun 03 07:17:14 morning all Jun 03 08:20:36 hello world. Jun 03 09:08:14 good morning zecke Jun 03 09:08:28 moring pb Jun 03 09:08:32 +n Jun 03 09:08:35 03koen 07org.oe.dev * r37e5aa70... 10/packages/linux/ (linux-ezx/ezx6-jumbo.patch linux-ezx_2.6.16.13.bb): Jun 03 09:08:35 linux-ezx: fix patching brokenness: Jun 03 09:08:35 * OE uses quilt by default, so we can use it two times in a row Jun 03 09:08:35 * I'm too lazy to fix it properly for the .1% of OE users that do INHERIT += "patcher" Jun 03 09:08:35 * The ezx patchset is one jumbo patch now Jun 03 09:08:37 ~lart mickeyl for breaking nearly everybodies build Jun 03 09:08:37 * ibot nabs the moon and broadsides mickeyl with the sea of tranquility for breaking nearly everybodies build Jun 03 09:08:39 hi woglinde Jun 03 09:10:04 hey woglinde, zecke, pb_, koen Jun 03 09:10:15 lol Jun 03 09:10:19 hi cyril Jun 03 09:10:41 ;o) Jun 03 09:11:46 mickey|bbl: use quilt! Jun 03 09:12:31 ughs Jun 03 09:12:34 s/can/can't/ Jun 03 09:15:44 JustinP: the usual zero feedback :} Jun 03 09:16:03 france: did you turn off spam filtering on OE at handhelds.org? Jun 03 09:16:39 JustinP: feel free to check in the script under contrib/ Jun 03 09:18:56 zecke: the ipkg segfault is fixed in ipkg-0.99.163.tar.gz :) But ipkg.inc try to fetch ipkg from CVS and there is not (yet) tagged branch for this release :-/ Jun 03 09:19:07 zecke: BTW, is there any reason why not building ipkg from the tarball available here http://www.handhelds.org/download/packages/ipkg/ ? Jun 03 09:19:26 no good reason Jun 03 09:19:29 cyrilRomain: not that I would know Jun 03 09:19:31 patches welcome :) Jun 03 09:20:05 koen: I includes the patch into the existing terse.patch, but it is not a proper solution Jun 03 09:20:57 either I ask the ipkg developer to create such a tagged branch, so that we simply have to copy the ipkg*159*.bb to ipkg*163*.bb Jun 03 09:21:19 either ipkg.inc should be changed to build from http://www.handhelds.org/download/packages/ipkg/ Jun 03 09:21:31 cyrilRomain: if you feel like, you could write a test suite for ipkg Jun 03 09:25:56 zecke: why not, but I'm afraid not to have enough time to do it :-/ I work on other project than OE and although I'd like to help as much as possible, I prefer to be sure to have enough time to do such a task Jun 03 09:26:15 cyrilRomain: hehe :) Jun 03 09:30:33 zecke: for now, I perfer to participate in regression test building incremental and from-scratch Angstrom images and reporting build to Tinderbox :) Jun 03 09:31:03 zecke: BTW, shall I ask for an account for Tinderbox so that I could add comment to my builds ? Jun 03 09:31:29 cyrilRomain: you can use your bugtracker account Jun 03 09:31:38 zecke: great :) Jun 03 09:31:43 cyrilRomain: if you have no strong password... Jun 03 09:31:48 cyrilRomain: otherwise this is insecure :} Jun 03 09:32:30 as part of my uni work I might be able to hack tinderbox today Jun 03 09:33:38 zecke: I took a quick look in Tinderbox last week (I installed one on my server) Jun 03 09:34:02 zecke: BTW, did you know http://buildbot.sourceforge.net/ ? Jun 03 09:34:27 zecke: It is intended to be a python equivalent of Tinderbox Jun 03 09:34:51 zecke: I just discovered it today and don't know if it good or not Jun 03 09:35:02 cyrilRomain: I don't know, I have seen it though Jun 03 09:35:17 cyrilRomain: the buildbot servers asks the clients to start/stop building Jun 03 09:35:43 cyrilRomain: but as with OE we have many different configurations that can take different amount of time Jun 03 09:35:57 cyrilRomain: and for now we do not need bidirectional communications... Jun 03 09:55:44 zecke: about Tinderbox, I think it would be great if: Jun 03 09:55:56 zecke: 1) Tinderbox settings were outside the local.conf file. Because when you change TINDER_MACHINE of TINDER_REPORT, bitbake reparse all the bb files :-/ Jun 03 09:56:14 cyrilRomain: why would you change it ;) Jun 03 09:56:16 zecke: 2) the OEREV="unknown" in report were replaced with the DB revision _and_ a 'monotone diff' file, so that anyone would be able to reproduce the same build Jun 03 09:56:35 cyrilRomain: this is on the todo :) Jun 03 09:56:45 zecke: :) Jun 03 09:56:46 cyrilRomain: even monotone includes a diff when compiling Jun 03 09:58:47 zecke: where can I find such automatically created diff ? Jun 03 10:00:04 cyrilRomain: monotone --version will print the diff used Jun 03 10:00:21 zecke: thanks ! Jun 03 10:02:00 good morning Jun 03 10:02:25 zecke: I should be there this afternoon, if I can help for Tinderbox .... Jun 03 10:02:39 can you do perl hacking? Jun 03 10:02:51 cyrilRomain: and we could extend your python script Jun 03 10:03:09 zecke: yes (for the python script) Jun 03 10:04:05 zecke: about perl, I'm not an expert but have some developer skill, so I could try Jun 03 10:32:15 I hate perl... Jun 03 10:39:40 hi ph5 Jun 03 10:39:51 hi pb_ Jun 03 11:03:03 hi Jun 03 11:03:56 hi hrw Jun 03 11:04:04 hi hrw Jun 03 11:05:40 someone with amd64? Jun 03 11:06:54 me Jun 03 11:07:12 can you build gcc 3.4.4 with bitbake? Jun 03 11:07:29 not gcc-cross but gcc Jun 03 11:08:45 I don't know what bitbake is.. :( Jun 03 11:08:58 ok. then forget Jun 03 11:09:00 [ebuild UD] sys-devel/gcc-3.4.4-r1 [3.4.6-r1] USE="doc gcj nls -bootstrap -boundschecking -build -fortran -gtk -hardened -ip28 -ip32r10k% -multislot -nocxx -nopie -nossp -objc -vanilla" 27,037 kB Jun 03 11:16:53 hrw: yes, i think my gcc 3.4.4 builds went ok on amd64 Jun 03 11:17:09 with .dev? Jun 03 11:18:15 hrw: sorry, that's a false report; openslug does not build gcc natively. Jun 03 11:18:26 I can try right away if you wish Jun 03 11:19:51 hrw: bitbake -b org.openembedded.dev/packages/gcc/gcc_3.4.4.bb ? Jun 03 11:22:23 bitbake gcc-3.4.4 Jun 03 11:25:30 hi cliff Jun 03 11:25:42 morning hrw Jun 03 11:25:59 cbrake: looks like very morning even.. Jun 03 11:26:23 i did not get up before 8 usually :) Jun 03 11:26:43 hrw: heh, well kids change all that :-) Jun 03 11:26:51 i suppose Jun 03 11:26:54 hrw: morning is my time to get things done ... Jun 03 11:27:42 kids.. year or two and i will get own ones.. Jun 03 11:28:15 hrw: great! Its a good experience and a lot of fun :-) Jun 03 11:28:17 I'm getting old.. everyone around is getting their firstboorn :} Jun 03 11:28:27 zecke he Jun 03 11:28:31 you have 7 years Jun 03 11:28:37 hrw: force 'it' to study medicine Jun 03 11:28:43 hrw: and teach 'it' latin! Jun 03 11:29:00 woglinde: Al is doing 'bad' jokes already Jun 03 11:29:09 latin? Jun 03 11:29:19 zecke should we meet with al? Jun 03 11:29:23 woglinde: I would have to stop taking my drugs first, otherwise it would look like stitch Jun 03 11:29:37 and my wife talking to her about beeing pregnat? Jun 03 11:29:41 no Jun 03 11:29:57 better not, the influence will be bad in any case Jun 03 11:30:00 (bad for me) Jun 03 11:32:03 hrw: kicking off a gcc-3.4.4 build now ... Jun 03 11:32:31 hrw: hitting a ICE also... Jun 03 11:32:36 cbrake: I just had to do it on dev. when you came Jun 03 11:33:01 likewise: which one Jun 03 11:33:17 | /home/leon/sandbox/ixp4xx/openembedded/build/tmp/work/gcc-3.4.4-r5/gcc-3.4.4/gcc/stmt.c: In function `expand_decl': Jun 03 11:33:19 | /home/leon/sandbox/ixp4xx/openembedded/build/tmp/work/gcc-3.4.4-r5/gcc-3.4.4/gcc/stmt.c:4051: error: unrecognizable insn: Jun 03 11:33:20 | (insn 672 574 673 53 (set (reg:SI 270) Jun 03 11:33:22 | (const_int -4294958849 [0xffffffff000020ff])) -1 (insn_list:REG_DEP_ANTI 566 (nil)) Jun 03 11:33:23 | (nil)) Jun 03 11:33:25 | /home/leon/sandbox/ixp4xx/openembedded/build/tmp/work/gcc-3.4.4-r5/gcc-3.4.4/gcc/stmt.c:4051: internal compiler error: in extract_insn, at recog.c:2083 Jun 03 11:33:26 | Please submit a full bug report, Jun 03 11:33:41 sorry that should have been to pastebin.com - sorry for the spam... Jun 03 11:33:59 likewise: pastebin the compile line please Jun 03 11:34:03 0xffffffff000020ff don't look like a 32bit address Jun 03 11:34:28 doesn't... Jun 03 11:34:46 zecke: how goes getting few X2? Jun 03 11:34:47 http://pastebin.com/755318 Jun 03 11:36:10 hrw: fighting with my stinginess Jun 03 11:37:34 likewise: could you just preprocess the file? Jun 03 11:37:43 replacing -c with -E Jun 03 11:37:52 zecke: yes, hold on Jun 03 11:38:43 woglinde: I need more time and clones... Jun 03 11:38:56 woglinde: would you like to hack an ALC/FLUTE library Jun 03 11:40:03 same gcc-3.4.4 error here Jun 03 11:40:22 thx Jun 03 11:40:31 and the specs of the involved compilers could be interesting as well Jun 03 11:41:28 puh I'ms till human and don't understand the spec file Jun 03 11:48:12 and I really need a reference card for OE... :-/ Jun 03 11:49:31 zecke: http://www.inrialpes.fr/planete/people/roca/doc/draft-ietf-rmt-flute-07.txt ? Jun 03 11:51:00 zecke: hum ... a draft only of the official one that look similar ... Jun 03 11:51:16 cyrilRomain: they are RFCs Jun 03 11:51:29 cyrilRomain: yeah this FLUTE Jun 03 11:54:18 zecke: the ICE disappears when using -O2 instead of -Os. (as I expected, I have filed many of these bugs at GCC Bugzilla) Jun 03 11:54:28 hrw: see above Jun 03 11:54:44 zecke: GCC teams responds that GCC 3.4.x is no longer maintained Jun 03 11:54:51 s/teams/team Jun 03 11:55:01 likewise uh good to know Jun 03 11:55:19 likewise: the pointer size is wrong anyway Jun 03 11:55:38 likewise: it is using stuff from the host (likely) which should happen regardles of optimisation Jun 03 11:55:41 yes, but it' s good to know the wrong path is presumably in the size-optimizer. (-Os) Jun 03 11:55:53 zecke: that is scary then Jun 03 11:56:20 this is why I wanted to see a preprocessed file Jun 03 11:56:27 zecke: on it. Jun 03 11:57:07 grammar question... "I took someone to somewhere" is correct? Jun 03 11:58:16 hrw: I think yes. Jun 03 11:58:35 as in: I took Paul to Dublin? Jun 03 11:58:48 yes Jun 03 11:58:51 yes Jun 03 12:00:28 thx Jun 03 12:00:57 zecke: I think this should be it: http://oe.pastebin.com/755347 Jun 03 12:02:36 is that everythingß Jun 03 12:03:10 zecke: no, seems there is a snippet limit on pastebin Jun 03 12:03:21 /home/leon/sandbox/ixp4xx/openembedded/build/tmp/work/gcc-3.4.4-r5/gcc-3.4.4/gcc/stmt.c Jun 03 12:03:21 is not the recog.c Jun 03 12:03:34 likewise: please grep for your amd64 include directory Jun 03 12:06:50 zecke: recog.c is the source file from which my native (distro) compiler is built. I am confused... Jun 03 12:08:16 ah sorry Jun 03 12:08:18 right Jun 03 12:09:50 zecke: ok, so my native compiler (Ubuntu 6.06 LTS btw) ICEs in recog.c:2038 compiling gcc-3.4.4, source file stmt.c:4051 Jun 03 12:10:01 likewise:no Jun 03 12:10:29 zecke: no, my OE built arm-target crosscompiler does that. Jun 03 12:10:30 likewise: the cross compiler we built (arm-linux-gcc) should build gcc Jun 03 12:10:32 :-) Jun 03 12:10:36 okay Jun 03 12:10:51 so the bug is either within the cross compiler, or within the newly built one Jun 03 12:11:07 someone thinks it is okay to use 32bit addressed Jun 03 12:11:11 yes, agreed Jun 03 12:11:26 64... Jun 03 12:12:06 zecke: yes, either the sources for the to-be-built gcc go wrong, or the arm-target cross-compiler already goes wrong with wrong built-ins. Jun 03 12:13:43 zecke: I do this: grep -e '^#' stmt.i | grep -v leon Jun 03 12:14:06 zecke: all includes, but not the ones that refer local (/home/leon/sandbox/ bla bla) Jun 03 12:14:50 zecke: gives this http://oe.pastebin.com/755370 Jun 03 12:16:16 three time config.h Jun 03 12:16:46 hmm... later I fear Jun 03 12:16:52 zecke: ping. I think I am going to build the arm cross-compiler with --save-temps option, that way all preprocessed files are left in the filesystem. Jun 03 12:17:06 zecke: for later worry, you mean? Jun 03 12:17:13 I need to get some stuff done Jun 03 12:17:25 zecke: sure, me too, and the weather is too nice here Jun 03 12:17:34 hehe Jun 03 12:17:37 thanks for looking along! Jun 03 12:17:39 cu Jun 03 12:17:44 * hrw posted new post on blog Jun 03 12:17:53 * zecke needs to pick up friends (of al) at the train station Jun 03 12:18:12 likewise: I didn't know --save-temps Jun 03 12:18:29 it's a gcc option Jun 03 12:18:55 saved me a lot of hassle debugging sometimes. LOng since I have used it, I have to look it up. Jun 03 12:20:53 cu - have a nice weekend Jun 03 12:21:17 cu - still have to see davinci Jun 03 12:21:31 like|garden: book is much better Jun 03 12:23:29 hum .. 'bitbake meta-toolchain' fails here (using gcc4.1.1) http://oe.pastebin.com/755362, any ideas ? Jun 03 12:23:43 BTW, what is gcc-*-sdk for ? Jun 03 12:23:55 no :( Jun 03 12:24:14 zecke: :-/ Jun 03 12:26:10 hum ... gcc-cross-sdk_4_1.0.bb includes gcc4-build-sdk.inc, but only gcc3-build-sdk.inc exist, could it be related ? Jun 03 13:00:34 cyrilRomain: the -sdk packages are used for generating standalone installable cross toolkits Jun 03 13:00:51 grr. screensaver crashes on AMD64/Ubuntu6.06LTS/Radeon9700... Jun 03 13:01:14 pb_: i.e. one .ipk that contains gcc+binutils+glibc ? Jun 03 13:03:17 pb_: ah ok: _cross_ toolkits. thanks Jun 03 13:04:39 moin mickeyl Jun 03 13:09:04 hi Jun 03 13:11:49 hey Jun 03 13:12:53 i'd like to lower our amount of qte patches in .dev. it looks to me like we no longer need some of them, i.e. the mouse and keyboard patches that predate tslib/kernel keymap developments. Jun 03 13:13:17 any idea about these? Jun 03 13:14:05 die patches die Jun 03 13:14:16 exactly Jun 03 13:14:36 RP: I need an opinion about multicolor LEDs Jun 03 13:14:45 RP: please ping me when you're around Jun 03 13:14:50 mickeyl: they taste good, but are not healthy Jun 03 13:15:08 heh, LEDs, not LSDs Jun 03 13:15:30 wow, Shaved, Washed, Clothed I feel like a Human Being Jun 03 13:16:12 all: I have file://blabla.h in my SRC_URI, but it doesn't seem to be handled in any way. It should be copied to the S (source) dir, right? Jun 03 13:16:21 likewise: no Jun 03 13:16:24 likewise: WORKDIR Jun 03 13:17:09 ok, and patches are always applied to S. Jun 03 13:17:25 yes Jun 03 13:17:25 right Jun 03 13:17:38 you _could_ cheat with adding a directory structure to FILES Jun 03 13:17:42 but it's not recommended Jun 03 13:17:49 better copy the files manually from WORKDIR into S Jun 03 13:19:13 mickeyl: it is allowed / neat to add a directory to FILES and cp ${WORKDIR}/thatdir . ? Jun 03 13:19:31 . = ${S} during do_configure Jun 03 13:20:30 And both files and patches may either be stored in ${PN}${PV} or files/ dir along with the .bb, depending on whether they are applicable to all, or just a single version? Jun 03 13:22:15 likewise: take a look at freeNote Jun 03 13:22:22 where we do it like that Jun 03 13:22:42 we just add a directory structure into files/ and then copy it automatically by specifying it in SRC_URI Jun 03 13:23:01 dirs are allowed in SRC_URI Jun 03 13:25:07 mickeyl: thanks. Jun 03 13:30:20 mickeyl: that resulted in a Freetype only supporting Truetype fonts, 1/3 of the lib size. Jun 03 13:31:15 cool! Jun 03 13:31:19 good job Jun 03 13:31:35 which font types do you disable then? Jun 03 13:31:53 mickeyl: hold on, I'll explain a bit. Jun 03 13:33:47 mickeyl: I disabled type1 support, t42 support, winfnt, pcf, postscript and bitmap rastering Jun 03 13:34:42 mickeyl: and I disabled glyph names, mac fonts, adobe glyph lists, zlib support, lzw support, embedded bitmaps Jun 03 13:35:07 mickeyl: so, anti-aliased truetype rendering only. Also, I added the light-hinting patch. Jun 03 13:35:15 hi mickeyl Jun 03 13:36:30 hey phil Jun 03 13:37:10 likewise: excellent. that looks like a good patch to me. Jun 03 13:37:34 anyway, you should mail oe@handhelds.org to see that no one cries out because of removing a feature they rely on :) Jun 03 13:37:37 mickeyl: I do not know which distro's use which features of freetype. Jun 03 13:37:50 right, either mail or add your proposal into the bugtracker Jun 03 13:37:55 the latter may be better Jun 03 13:38:00 ok Jun 03 13:38:14 then if no one complains for - let's say a week - apply it Jun 03 13:38:29 mickeyl: I am not a OE committer myself. Jun 03 13:38:42 s/apply it/I apply it/ Jun 03 13:38:43 :D Jun 03 13:39:02 :-) Jun 03 13:40:52 Most of my stuff now sits here (mostly intrusive patches to reduce dependencies and strip package features) http://www.esrac.ele.tue.nl/~leon/linux/openembedded/ Jun 03 13:42:43 mickeyl: I think poppler needs a fullblown freetype? Jun 03 13:43:20 zecke: this is why I loosely proposed having a "feature dependency tree" in OE/bitbake. Jun 03 13:43:30 zecke: true. i guess we want to have two versions then. a reduced one and a fullblown Jun 03 13:43:42 both RREPLACEing each other Jun 03 13:44:06 mickeyl: yes, I discussed this with koen; once you start making multiple configurations of single package, it is hard to deal with dependencies. Jun 03 13:44:26 mickeyl: think of the USE flags in Gentoo. Jun 03 13:44:33 likewise: yeah, we need such thing, but it needs to be better than the USEflag Jun 03 13:44:34 hehe Jun 03 13:44:39 need to pick up al now Jun 03 13:44:46 zecke: ok, good luck Jun 03 13:45:17 I'm not late yet, so she musn't be angry :} Jun 03 13:45:44 zecke: Otherwise blame me :-) Jun 03 13:46:11 "You are late due the computer?" ... Jun 03 13:46:24 now she starts chatting she understands a little better Jun 03 13:46:32 zecke: No, "I was chatting with an unknown ' friend' " :-) Jun 03 13:46:47 cya Jun 03 13:47:57 "You are late due to your imaginary friend?" Jun 03 13:47:57 heh Jun 03 13:48:33 mickeyl: Your RREPLACE idea; I would create a limited freetype and the current freetype. Some packages depend on the limited, other on the full-blown freetype. So the two versions would have different names. Jun 03 13:48:49 freetype and freetype-ttf for example Jun 03 13:49:24 pb_: My imaginary friend returned my car too late :-) Jun 03 13:52:27 likewise: right. and both packages have to RREPLACE each other since they're providing the same library. Jun 03 13:53:01 mickeyl: freetype then RPROVIDES both freetype and freetype-ttf?? Jun 03 13:53:17 no Jun 03 13:53:22 ah right Jun 03 13:53:23 yes Jun 03 13:53:26 feature-based Jun 03 13:55:39 restarting chatzilla, as I installed a firefox extension that I want enabled... Jun 03 14:05:33 pb_: is the -sdk stuff work in progress? I am interested in trying that. Its cross-toolchain is used within OE, but also usable outside it without much hassle? Jun 03 14:06:19 likewise: it's only used outside of OE Jun 03 14:06:26 likewise: 'bitbake meta-sdk' Jun 03 14:06:50 koen: but what I compile with it, is 100% compatible with the libs inside the rootfs created by OE? Jun 03 14:07:05 it should be Jun 03 14:07:10 koen: hello btw :-) Jun 03 14:07:22 hi :) Jun 03 14:07:35 koen: you are counting days? (nights?) Jun 03 14:07:55 I'm trying to figure out how to get my luggage down to 20kg Jun 03 14:08:01 koen: haha Jun 03 14:08:26 my 80-400mm lens is 2.5kg, tripod is 2kg, etc Jun 03 14:08:46 rm 50 bottles of beer? Jun 03 14:08:56 :) Jun 03 14:09:02 likewise: rm: Permission denied. Jun 03 14:09:21 I have 4 bottles of huttenkloas for prplague :) Jun 03 14:09:44 on the way back that space is probably filled by rum Jun 03 14:10:30 NAbyss: :-) Jun 03 14:10:46 likewise: The beer's important! Jun 03 14:10:58 But, that's easily discerned from my hostmask. Jun 03 14:12:58 koen: ok replace the 50 bottles of beer with a single beertender, should save weight of glass. Get rid of the tripod, prpplague will stand in instead. Jun 03 14:13:14 :) Jun 03 14:20:44 "896 reads" on the 404 page I installed 2 days ago Jun 03 14:25:58 koen: 404? You are counting non-existant wiki pages or what? Jun 03 14:26:34 koen: how's the OE survey coming along btw? Jun 03 14:44:26 likewise: I told drupal to redirect 404s to http://www.openembedded.org/page-not-found Jun 03 14:44:44 likewise: iirc zecke and NAbyss were working something out Jun 03 14:45:36 koen: nice! :) Jun 03 14:45:49 i like drupal a lot, very easy to edit Jun 03 14:46:00 me too Jun 03 14:46:14 it was either plone or drupal :) Jun 03 14:47:23 koen: Yep, haven't heard from zecke though as yet Jun 03 14:48:01 ah, down to 17.5 kg Jun 03 14:55:12 * chouimat is back. Jun 03 14:55:23 morning Jun 03 14:55:36 hi Jun 03 15:16:44 hope this is not a delicate topic, so all links at oe.hh.org don't work anymore, right? Jun 03 15:17:05 correct Jun 03 15:17:08 ok Jun 03 15:17:12 use www.openembedded.org Jun 03 15:17:24 is there an automatic redirection ? Jun 03 15:17:43 i think of this because of site that already have link to oe Jun 03 15:17:48 www.openembedded.org has always worked Jun 03 15:18:24 ok then :) Jun 03 15:22:01 oh... lot's of bookmarks... how can i have so many... Jun 03 15:22:21 koen: about -sdk packages, you said it is only used outside OE. So to build an ARM toolchain, I should not use 'bitbake meta-toolchain' and build gcc,glibc,binutils separately ? Jun 03 15:22:55 koen: 'bitbake meta-toolchain' indeed try to build gcc-cross-sdk Jun 03 15:23:31 to build an arm toolchain to be used outside of OE, you can use meta-sdk or meta-toolchain Jun 03 15:24:16 koen: e.g. to add a toolchain to the angstrom distro, should I use meta-sdk or meta-toolchain ? Jun 03 15:27:34 JustinP, yeah i am on ~5 lists Jun 03 15:28:09 cyrilRomain: I don't really get what you want to do Jun 03 15:28:18 its just odd to see e17 more active than OE Jun 03 15:29:34 koen: I want to make an angstrom image and add an ARM toolchain on my tosa. For now, I'm doing 'bitbake gcc-4.1.1', 'bitbake binutils', 'bitbake glibc', etc... Jun 03 15:30:17 koen: I just noticed the meta-toolchain this morning and was wondering if it would build all necessary .ipk Jun 03 15:30:44 cyrilRomain: ah, *-cross-sdk is for build hosts (e.g. x86->arm) Jun 03 15:31:12 to build on the device it self you need 'bitbake gcc binutils ' Jun 03 15:31:25 koen: thank you very much :) Jun 03 15:31:55 you could create a new meta packages, like meta-native-toolchain, which would build gcc, binutils, glibc, etc Jun 03 15:33:37 koen: ... I'll build it by hand for now but keep this idea in mind ;) Jun 03 15:49:54 anyone have additions for http://www.openembedded.org/wiki/PapersAndPresentations ? Jun 03 15:50:19 please add them yourself :) Jun 03 16:06:29 hi, decided to give gcc 4 a try for building slugos, commented out all PREFERRED_VERSION lines for gcc and glibc, now it is building gcc-cross-4.0.2, but the latest one in packages is gcc-cross_4.2-20060513.bb Jun 03 16:06:45 anyone an idea why I am not getting that one and what I need to do to use this one? Jun 03 16:07:56 erm Jun 03 16:11:16 eFfeM: There was a good reason not to use gcc 4 for slugos... I can't quite remember what right now ;) Jun 03 16:12:27 eFfeM: adding a "Tested on the following slugos-variants:" wouldn't be a bad idea... Jun 03 16:13:31 (imho) Jun 03 16:13:52 NAiL: saw quite some work on gcc here in #oe for the last few days, so I thought I'd give 4.2 a try instead of just running the idle loop of my system Jun 03 16:14:01 yeah, I like that sentence Jun 03 16:14:29 koen: would it be useful if "oeandyourdistro" contained info about using a custom ebuild for qt so one can use qvfb in gentoo? Jun 03 16:15:04 anyway, still interested to give gcc4.2 a try if I only knew how Jun 03 16:15:09 eFfeM: It should have a note saying something like "If you've tested this on another slugos variant, please add it to this list" Jun 03 16:15:52 and NAiL of course if you recall why gcc4 caused a problem, feel free to let me know Jun 03 16:16:03 gcc4.2 and eabi(?) would be nice for the next release (post-3.x) Jun 03 16:16:23 NAiL: good sentence, these then would be the 1st two lines of each package related wiki page (e.g. the howto's) Jun 03 16:16:56 NAiL: since 3.x is imminent, I'd thought to give 4.2 a local try. Jun 03 16:17:03 * eFfeM likes living on the edge Jun 03 16:17:12 NAiL: you've got borzoi hardware, ya? Jun 03 16:17:25 don't know about eabi Jun 03 16:17:31 jnc: nope :( Jun 03 16:17:34 oh Jun 03 16:17:44 Can't recall what borzoi was atm Jun 03 16:17:50 btw also would like to see kexec work on slugos be Jun 03 16:17:52 but I'm pretty sure I don't have any :P Jun 03 16:17:56 eFfeM: me too Jun 03 16:17:58 i aim to ask another zaurus c3100 user what distro they run Jun 03 16:18:01 and how stable it is Jun 03 16:18:23 nail, from wikipedia: The Borzoi is an intelligent, active dog Jun 03 16:18:37 I'm hoping to get a hold of a BlueAngel later on though :-D Jun 03 16:20:25 sure i'd like to run something OE-derived :) Jun 03 16:20:49 the options aren't too productive though, last i checked Jun 03 16:27:00 eFfeM-away, NAiL: gcc 4.x are missing the armeb patches iirc Jun 03 16:27:11 ~borzoi Jun 03 16:27:13 methinks borzoi is the Sharp SL-C3100, or a dog (see http://www.dogbreedinfo.com/borzoi.htm ) Jun 03 16:27:47 eFfeM-away, NAiL: and EABI for armeb doesn't exist yet afaik Jun 03 16:32:07 eFfeM-away, did anyone tell you 4.0.2 != 4.2.x ? Jun 03 16:43:17 pH5: any idea why name is 0x0? Jun 03 16:46:55 zecke: I'm seriously considering to teach OE to populate a chroot and do builds from there Jun 03 16:47:19 zecke: but I'm not sure if it's a smart move to do that in the SoC Jun 03 16:47:22 probably not Jun 03 16:49:44 03koen 07org.oe.dev * re13ada17... 10/packages/linux/linux-ezx_2.6.16.13.bb: Jun 03 16:49:44 linux-ezx: properly fix patching troubles: Jun 03 16:49:44 * remove patches/ and .pc/ subdirs after do_ezxpatch Jun 03 16:49:44 * reinstate patching from openezx.org tarballs Jun 03 16:49:45 koen: well Jun 03 16:50:22 it would solve soooo many problems Jun 03 16:50:28 but it is a lot of work Jun 03 16:50:35 zecke: the problem seems to be further up the chain, FcObjectPtrU(pe->object) is 0x0 in FcFontRenderPrepare already for (FcPatternEltU(pat->elts)+19)->object (=44) but I have no idea what that means Jun 03 16:51:11 koen: your issue is the 'rotating' STAGING_DIR? Jun 03 16:51:26 koen: you don't need that, debian somehow packages a gcc as well Jun 03 16:51:31 koen: and is that using a chroot? Jun 03 16:52:34 zecke: my issues involve the rotating dir, relocation issues and build paths ending up in packages Jun 03 16:53:53 the first issue magically disappears when using -dev packages to populate staging Jun 03 16:54:04 so I'll attach that one first Jun 03 16:54:24 ehm Jun 03 16:54:26 attack Jun 03 16:56:24 koen: you should be able to avoid this rotating thing more easily Jun 03 16:59:35 by unbreaking glibc's do_stage Jun 03 17:12:14 zecke: 13:18 < zecke> cyrilRomain: and we could extend your python script Jun 03 17:12:17 zecke: what would you like to extend ? (it is a bash script, the python one was less straight forward) Jun 03 17:12:29 eek bash Jun 03 17:12:53 cyrilRomain: the task executor before the task executor Jun 03 17:13:00 cyrilRomain: update bitbake, oe, start the starts Jun 03 17:13:08 cyrilRomain: remove the directories after a while... Jun 03 17:16:22 zecke: I'm afraid not to get what you really want Jun 03 17:16:31 don't worry Jun 03 17:16:44 zecke: there is already an update bitbake and an update oe.db functions Jun 03 17:16:47 cyrilRomain: this script should update BitBake/OE from more than one source Jun 03 17:17:08 cyrilRomain: e.g. using the MasterMakefile from NSLu2, svn, git... Jun 03 17:17:23 cyrilRomain: automatically prune directories after x days or y builds Jun 03 17:18:14 zecke: ok, let's start with "update BitBake/OE from more than one source" Jun 03 17:18:46 we could ship a list of public mirrors in the OE tree Jun 03 17:19:17 or build poky :) Jun 03 17:19:23 or familiar from the git tree Jun 03 17:20:56 zecke: let's not mention that in here Jun 03 17:22:42 zecke: where can I find the whole list of BitBake/OE sources ? Jun 03 17:22:56 cyrilRomain: nowhere, it needs to be configurable :) Jun 03 17:23:05 cyrilRomain: you will be an executor of a configuration Jun 03 17:23:18 cyrilRomain: TinderHelper Stuff1 Stuff2 Stuff3 Stuff4 Jun 03 17:23:24 cyrilRomain: where StuffX is a configuration Jun 03 17:27:27 * chouimat is away: stuff Jun 03 18:13:22 koen: so what is the status of packaging bits? Jun 03 18:13:37 koen: you have the problem with hardcoded paths e.g. glibc Jun 03 18:13:47 that if you install it to tmp directory doesn't work right? Jun 03 18:14:21 hehe, ibot left. :) Jun 03 18:18:04 *lol* Jun 03 18:18:13 zecke: I checked and the spam filters are fine...I hope we have new spam filter next week. Jun 03 18:18:36 france: It is the first time I got spam Jun 03 18:18:43 france: and this was ~8 emails Jun 03 18:19:01 france: thanks for looking into it! and your work is really appreciated Jun 03 18:19:33 zecke: most of the time the filters work, the big problem is when a new virus comes out, it some times takes 8 hours to get into the clamav data base by the clamav people. so we tend to get bursts. Jun 03 18:19:49 zecke: same for the RBLs Jun 03 18:19:58 zecke: u are welcome Jun 03 18:25:20 zecke: it is easy to make e.g. the OE database name and path configurable, but writing a script able to update bitbake/OE from any possible repository (either a svn, a git, a darcs,...,) implies as many driver as SCM (like in bitbake the fetch/*.py scripts). hum... Jun 03 18:25:52 cyrilRom1in: this is why I labeled it task executor Jun 03 18:26:06 zecke: if we do something more generic, it should not be harder to use than specific scripts Jun 03 18:26:09 cyrilRom1in: or we could use the tinderclient.pm from freedesktop, but I don't get what it is doing :} Jun 03 18:26:26 cyrilRom1in: right. A 'configuration' would look Jun 03 18:26:38 cyrilRom1in: build oe, use monotone via wget from, use branch foo Jun 03 18:26:53 cyrilRom1in: build with this BBPATH holding conf... Jun 03 18:27:57 emte: I know 4.0.2 != 4.2.x; the issue is that bb gcc-cross starts a build with 4.0.2 while there are also 4.1 and 4.2 bb files, Isn't bb supposed to take the latest version if there is no PREFERRED ? Jun 03 18:28:18 eFfeM: we have DEFAULT_PREFERENCE and suchs as well Jun 03 18:28:26 ah, ok Jun 03 18:28:37 eFfeM: but yes in absence of any '*PREF*' the highest version should be selected Jun 03 18:32:56 koen, about the missing armeb patch: I looked into the gcc-4.2-20060513 dir and there is a patch file called armeb-fix.patch Jun 03 18:33:35 zecke: the script would provide a set of task like update bitbake/oe. Do you have a predefined list of task in mind ? Jun 03 18:33:49 no :) Jun 03 18:34:16 zecke: BTW, Bitbake is made to do task, no ? Jun 03 18:34:27 eFfeM: that's good news Jun 03 18:34:45 s/task/tasks/ Jun 03 18:34:51 zecke: no, glibc wants to compile stuff in do_stage, requiring libc-headers and stuff in staging -> broken behaviour Jun 03 18:34:59 koen, i'm trying to build it , just curious where it will strand Jun 03 18:35:30 eFfeM: 4.1.1 might be a better choice Jun 03 18:36:02 morning Jun 03 18:36:14 hey XorA Jun 03 18:36:18 hey koen Jun 03 18:36:27 koen: I worked out why abiword plugins do not work Jun 03 18:36:39 XorA: cool Jun 03 18:36:41 koen: I just need to debug the fault in configure.in/libtool.m4 Jun 03 18:37:06 koen: unfortuneately debian has conspired to render my computer useless :-( Jun 03 18:37:35 koen, ok will try 4.1.1 Jun 03 18:37:40 XorA: that's too bad Jun 03 18:38:01 koen: someone seems to have removed a large chunk of the amd64 packages from all mirrors :-( Jun 03 18:38:36 cyrilRom1in: yes it is :} Jun 03 18:40:38 koen: done. ^_^ Thanks, I forgot about contrib Jun 03 18:42:55 03justinp 07org.oe.dev * r7086df3f... 10/contrib/mtnpatch.py: Add mtnpatch.py, a script to parse and import a full monotone diff Jun 03 18:43:41 hi, is there a way to find out who is using a certain module? I have a module I need to unload but it says used by 2 without giving any name Jun 03 18:44:44 eFfeM: lsmod Jun 03 18:45:03 koen: Jun 03 18:45:03 Module Size Used by Jun 03 18:45:03 bnep 12448 2 Jun 03 18:45:14 hrm Jun 03 18:45:19 and cat /proc/modules? Jun 03 18:45:46 bnep 12448 2 - Live 0xbf286000 Jun 03 18:45:54 weird Jun 03 18:45:58 /etc/init.d/bluetooth stop I would suggest Jun 03 18:46:16 XorA: good plan Jun 03 18:46:25 and ifdown the bnep interfaces Jun 03 18:46:33 koen, I think there is indeed an app usign it nfsd has count 8 Jun 03 18:47:29 xora after the stop indeed the count drops to 0 Jun 03 18:48:53 i'm trying to get btsco to work, initially it works but after a short while I get a connect error in dmesg and no more sound, unloading and reloading sco helps a few times, but after a few times this does not work any more either and there is no streaming so I wanted to unload/reload more Jun 03 18:49:38 does your usb controller support sco transfers? Jun 03 18:49:46 and is it compiled into the kernel? Jun 03 18:50:18 the controller works under ubuntu and the sco modules are loaded Jun 03 18:50:30 i don't think there is anything that needs to be compiled in the kernel Jun 03 18:50:45 I meant if you enabled sco support for the usb chip Jun 03 18:50:50 and actually the sound playback does start, I get a second or so of sound on my headset Jun 03 18:50:56 I remember seeing something like that in the kernel config Jun 03 18:50:57 koen, let me check Jun 03 18:52:13 koen from defconfig Jun 03 18:52:13 CONFIG_BT_SCO=m Jun 03 18:52:13 CONFIG_BT_HCIUSB_SCO=y Jun 03 18:52:34 no other SCO stuff Jun 03 18:52:36 looks ok Jun 03 18:54:02 the odd thing is that I get sound for a while but then it fails and in messages I get: Jun 03 18:54:02 hci_scodata_packet: hci0 SCO packet for unknown connection handle 46 Jun 03 18:54:28 also I avoided using a usb hub Jun 03 18:55:01 I'm eager to get sound on my headset and eventually try to use it as a voip headset Jun 03 19:06:25 koen, any advise on what binutils to use with gcc 4.1.1 Jun 03 19:06:46 (I've build initial-cross, but cross fails because it has the wrong ld Jun 03 19:11:16 2.16? Jun 03 19:12:45 zecke: that is the one I am using but then configure for gcc-cross-4.1.1 gives: /home/frans/OpenSlugHead/openslug/tmp/cross/lib/gcc/armeb-linux/4.1.1/../../../../armeb-linux/bin/ld: unrecognised emulation mode: armelf_linux Jun 03 19:13:37 hm. i'm getting the old ld, might need to clean it and recompile with the new inital gcc Jun 03 19:14:50 ahah, well at least openbox is available to me, Im off to switch to XC Jun 03 19:19:23 ok rebuilding binutils-cross fixed things Jun 03 19:20:47 rebooting, back in a few minutes ... Jun 03 19:22:19 i don't get it. why i can build glibc for h2200 successfully, but i can't when it is native. hit my head in the wall if you need. Jun 03 19:23:22 gah even vim is not available :-( Jun 03 19:23:45 XorA: vim for amd64? Jun 03 19:39:48 ljp: is there a port of QtTestlib to Qt2? Jun 03 19:42:56 thejapa: debian amd64 port has gone walkies :-( Jun 03 19:48:05 * XorA tries unstable Jun 03 19:50:05 oh, i tried amd64 at the beginning of the month. i guess i have a problem with debian... i went back to gentoo Jun 03 20:14:23 a quick note on my gcc 4.1.1 adventure Jun 03 20:16:12 yes Jun 03 20:23:17 has anybody tried to use firefox with gpe on small devices? Jun 03 20:26:03 system crashed.... Jun 03 20:26:17 lamikr: small as in? Jun 03 20:26:48 will report about gcc later Jun 03 20:27:28 zecke: Well something like ipaqs... 64m of mem. I would like to test whether the flash plugin from 770 would work. Jun 03 20:27:38 lamikr: see the pdaXrom screenshots Jun 03 20:29:48 so my SD card becomes useless if i try to use linux Jun 03 20:29:50 zecke: Do you know whether the minimo has plug-in support? Jun 03 20:30:03 no, I don't know Jun 03 20:31:16 The minimo has some build problems at least in 083 branch. Patches did not apply cleanly. Jun 03 20:32:21 and the plugin support was turned off by default. Jun 03 20:56:02 anyone got a amd64 .deb for monotone 0.25? Jun 03 20:58:07 zecke: The pics in http://www.pdaxrom.org/screenshots.php?menuid=2 looked really cool. That web-page was new for me. Do you know the device they have used for screenshots? Jun 03 21:01:14 lamikr: Sharp Zaurus Jun 03 21:04:52 oh, really. Need to test the quake with h6300 on someday. Jun 03 21:05:13 lamikr: quake even worked okay in strongARMs Jun 03 21:05:27 RIP DEC Jun 03 21:06:18 zecke: How fast are the newest Zaurus models, are they near to HTC phones like universal? Jun 03 21:06:40 lamikr: I think Sharp did not buy the ~600mhz variants from Intel Jun 03 21:08:42 cr2: Ok, and I guess sharp did not have either any special graph chips. So it should work also in h6300... Jun 03 21:09:33 lamikr: w100 is used in some devices Jun 03 21:09:56 lamikr: the pxa270 based devices use the pxafb IIRC Jun 03 21:11:25 zecke: I do not have much experiences with these pxa devices, only have old ipaq3630 in addition of h6300. But it seems that intel has currently more portable Linux device models than ti/omap. Jun 03 21:11:49 So, my next device may well be pxa based. Jun 03 21:13:29 is there a compiler built in #define that gives the version of the compiler? I want to print a msg in my program specifying with which compiler it was compiled Jun 03 21:13:45 eFfeM: yes Jun 03 21:14:01 what name? Jun 03 21:14:02 eFfeM: both GNU, GNU CC and GNU LIBC have these Jun 03 21:14:04 :} Jun 03 21:14:23 I recalled there is someting but forgot the name Jun 03 21:14:29 brb Jun 03 21:15:31 lamikr: I ran firefox on a Zaurus 5500 with SA cpu :-) thats the smallest I have been with it Jun 03 21:15:46 lamikr: 240x320 screen Jun 03 21:16:21 Xora: My screen size is same :-) How much memory the zaurus 5500 has? Jun 03 21:16:27 lamikr: 64M Jun 03 21:17:04 lamikr: firefox on that device was on the edge of usable, it took ages to render pages Jun 03 21:17:24 lamikr: these days I have a 400Mhz pxa wth 640x480 screen Jun 03 21:17:27 Well, thats the same I have. I will try that if I fail to build the minimo from dev with plugin support turned on. Jun 03 21:17:49 lamikr: good luck, Ive never got minimo to build :-) Jun 03 21:18:09 eFfeM: hmm Jun 03 21:19:23 Xora: Yes, I fear that the desktop ligth firefox is now very good with 64 mb of mem... But I remember seeing it running with win95/intel 150 mhz desktop, so there are hope. Jun 03 21:19:58 lamikr: I use firefox a lot on my current PDA Jun 03 21:20:18 Xora: What PDA you now have? Jun 03 21:20:27 lamikr: Zaurus c860 Jun 03 21:21:18 XorA: That's something I havent never heard. So is zaurus still actively releasing new devices? Jun 03 21:21:45 lamikr: yes, latest is the C3200 pxa27X and a microdrive Jun 03 21:21:48 |terrier Jun 03 21:21:54 ~terrier Jun 03 21:21:54 i heard terrier is Sharp Zaurus SL-C3200, or a dog Jun 03 21:22:29 ~lart my keyboard for being funny tonight Jun 03 21:22:29 * ibot takes large quantities of Krispy Kream donuts and stuffs them one after another down my keyboard's throat until my keyboard puts on 150lbs for being funny tonight Jun 03 21:23:07 Xora: That's realy cool. I remember when I saw the first zaurus model with Linux demoed in 2001. I was really amazed and wanted to order one immediately. But then it took quite long before the first device really came to european. Jun 03 21:24:13 lamikr: it is my dream to upgrade later this year Jun 03 21:25:09 XorA: Do you know does zaurus plan to release devices with phone? Jun 03 21:25:16 eFfeM: __GNUC__ Jun 03 21:25:20 eFfeM: is the major number Jun 03 21:25:24 lamikr: that is beyond my crystal ball :-) Jun 03 21:25:43 lamikr: I plan to buy a HTC Magician (I think thats the right one) next week Jun 03 21:25:46 all: when do I put a patch in files/, or when in packagename-version/ directory? Jun 03 21:26:17 eFfeM: __GNUC_MINOR__ should have the minor number Jun 03 21:27:32 XorA: Visit in htc-linux irc, if you need more info with models. Just heard that they have managed to get the wlan driver from hx4700 to work with htc-universal. Earlier version worked them only for sending data but reading failed. Jun 03 21:28:21 lamikr: yeah, been there, I getting it as an XDA from O2, I just cant remeber exactly which model it translates to Jun 03 21:29:01 xora xdaIII Jun 03 21:29:06 i think Jun 03 21:29:22 woglinde: XDA Mini S is what I want Jun 03 21:29:51 xora oh hm Jun 03 21:30:09 XorA: In finland those HTC devices are usually not visible in shops, one needs to order them separately. Shops mostly offers devices from nokia,ericson,samsuns,philips,motorola and recently from qltek. Jun 03 21:31:40 mickeyl: hey Jun 03 21:31:43 hi Jun 03 21:31:47 RP: ping? Jun 03 21:31:56 mickeyl: I'm going to bed late tonight, but onyl due the Dallas Mavericks Jun 03 21:32:14 zecke: granted because it's an exception :) Jun 03 21:33:02 what's the recommended linux kernel interface for a vibrator? Jun 03 21:33:03 mickeyl: 'designing' some stuff now Jun 03 21:33:18 mickeyl: consider the usability of this device first Jun 03 21:33:30 Human Interface Device Jun 03 21:33:40 sounds overcomplicated to me Jun 03 21:33:46 i'm considering the LED class model atm. Jun 03 21:34:01 since it's basically a dumb output device Jun 03 21:34:14 like a led, just that it's vibration instead of photons Jun 03 21:34:22 mickeyl: have you not seen a Rampant Rabbit? Jun 03 21:34:37 mickeyl: at ROAD we had a simple platform device for that Jun 03 21:34:38 zecke: thanks for the info, I want to verify that my gcc cross is indeed 4.1.1. Jun 03 21:34:41 mickeyl: no interface :} Jun 03 21:35:16 mickeyl: and created sysfs files for it, I don't know if vibrator classifies for a 'class' Jun 03 21:35:22 i see. it would match fine to the LED if I had more than just the three levels RP has dedicated. perhaps i can convince him to add some more Jun 03 21:35:39 i just need 4 levels Jun 03 21:35:40 mickeyl: abusing the LED class? Jun 03 21:35:49 s/ab// Jun 03 21:36:05 seems just fine for a simple output device Jun 03 21:36:08 *shrug* Jun 03 21:36:16 mickeyl: I thought the idea of a class device was to give somewhat stable API to userspace as well Jun 03 21:37:02 yeah Jun 03 21:38:04 ~seen ar_ Jun 03 21:38:20 ar_ was last seen on IRC in channel #oe, 132d 7h 46m 57s ago, saying: 'any ideas?'. Jun 03 21:38:22 OT: anyone has experience with 'wix' on wine? Jun 03 21:38:35 zecke: wix? Jun 03 21:38:54 XorA: Windows XML Installer or the other way around... Jun 03 21:39:14 XorA: You create a WiX file, then use the candle and finally light and you have a msi installer Jun 03 21:39:36 zecke hm nice Jun 03 21:39:38 zecke: ah, Ive only installed msi files, never tried to make them Jun 03 21:39:53 msi with GPO is cool Jun 03 21:40:37 woglinde: GPO? Jun 03 21:40:45 grou policy order Jun 03 21:40:47 woglinde: how can I create MSI from Linux? Jun 03 21:40:51 gruppenrichtlinie Jun 03 21:40:57 zecke noway Jun 03 21:41:09 woglinde: stop swearing in here :-) Jun 03 21:42:30 xora okay :( Jun 03 21:43:05 woglinde: bloody friend :) Jun 03 21:43:09 woglinde: having inherted a broken ADS setup at work, those words hurt my ears :-) Jun 03 21:43:21 ok, managed to crosscompile hello world with gcc 4.1.1 for slugos Jun 03 21:43:56 koen, zecke, thanks for your help Jun 03 21:44:14 calling it a day for now, I'll try something serious tomorrow Jun 03 21:44:17 cu Jun 03 21:52:55 I'm watching a bad Cricket Movie and begin to start this game... Jun 03 21:55:02 heh Jun 03 21:55:03 have fun Jun 03 21:55:04 g'night Jun 03 21:55:33 night mickeyl Jun 03 21:59:06 hmm mingw doesn't build on darwin Jun 03 21:59:13 cyrilRomain|bbl: ping? Jun 03 21:59:25 zecke: pong Jun 03 21:59:56 cyrilRomain|bbl: I now have sometime for the script Jun 03 22:00:08 03mickeyl 07org.oe.dev * rbdb1cb19... 10/conf/machine/a780.conf: a780.conf: fix typo Jun 03 22:00:57 zecke: well, explain me your thought :) Jun 03 22:02:05 it might be too late Jun 03 22:02:17 :} Jun 03 22:04:00 better late than never :) Jun 03 22:04:40 cyrilRomain: well, what I need is Jun 03 22:04:56 cyrilRomain: a tool that can update data from a given/configured source Jun 03 22:05:12 cyrilRomain: do the builds Jun 03 22:05:24 cyrilRomain: and depending on a setting clean the tree :) Jun 03 22:07:08 zecke: update data from a given/configured source: we may use the fetch/*.py script, don't you think ? Jun 03 22:07:36 cyrilRomain: yes, each to be built tree could be a .bb file Jun 03 22:07:45 cyrilRomain: and we would have our own base.bbclass with the tasks Jun 03 22:08:03 cyrilRomain: fetch, unpack/update, build, eventually_clean Jun 03 22:09:13 zecke: do the builds: you mean build like (but better) 'for i in angstrom, familiar, for j in gpe-image gcc-4.1.1 do use conf j && bitbake $j' ? Jun 03 22:09:26 no Jun 03 22:09:34 bitbake CONFIG(s) Jun 03 22:09:56 you would have bitbake meta-build Jun 03 22:10:13 and each build.... would be a configuration+target Jun 03 22:15:52 zecke: you want to also generate (at least some part) of local.conf files ? Jun 03 22:16:36 cyrilRomain: no Jun 03 22:16:51 cyrilRomain: What is in my head is Jun 03 22:17:02 cyrilRomain: I have a couple trees configured Jun 03 22:17:21 conf/local.conf and this 'auto builder' is writing a conf/auto.conf Jun 03 22:17:34 or another hint for the tinderclient.bbclass Jun 03 22:17:44 e.g. clobbered/not-clobbered build Jun 03 22:18:19 zecke rhcp in a few minutes Jun 03 22:18:30 woglinde: ah 'that' RHCP Jun 03 22:19:00 woglinde: I like Dani California, but that is pretty much it Jun 03 22:20:33 zecke the metallica concert was great Jun 03 22:21:03 woglinde: it was shown on MTV as well Jun 03 22:21:16 zecke yes yes Jun 03 22:22:16 woglinde: I'm watching American Chopper now Jun 03 22:22:35 hmm, too tired to work Jun 03 22:22:56 zecke: I'm just starting to understand... yes maybe too tired to work ;) Jun 03 22:23:19 zecke that I like too Jun 03 22:23:24 ERROR: The following variable(s) were not set: TARGET_OS Jun 03 22:23:25 Please set them directly, or choose a MACHINE or DISTRO that sets them. Jun 03 22:23:30 woglinde: XXP Jun 03 22:23:33 cyrilRomain: hehe Jun 03 22:24:03 zecke I know Jun 03 22:24:16 saw it at my parents house Jun 03 22:24:46 woglinde: do you have a flash countdown on your website? Jun 03 22:25:26 zecke lol no Jun 03 22:25:37 zecke you cant have a download or it Jun 03 22:25:40 for Jun 03 22:27:06 woglinde: http://www.inf.fu-berlin.de/lehre/SS04/ALP2/vorlesung/2.1.pdf <- haha Jun 03 22:27:14 woglinde: a wrong binary search implementation Jun 03 22:27:24 zecke: what do you exactly mean by 'configuration' and 'target' ? Jun 03 22:27:27 woglinde: I think I should pass if I claim his bin search is broken Jun 03 22:27:34 cyrilRomain: dunno Jun 03 22:27:42 cyrilRomain: configuration is lik conf/local.conf Jun 03 22:27:50 cyrilRomain: and target bitbake target Jun 03 22:28:00 zecke I meant countdown Jun 03 22:28:03 damn it Jun 03 22:28:10 I am on drugs Jun 03 22:28:16 because of the concert Jun 03 22:28:20 woglinde: no swear words my bloody Jun 03 22:28:22 have to come down Jun 03 22:28:27 woglinde: too much metal in your blodd? Jun 03 22:28:37 woglinde: don't play alabama man, don't hit your wife Jun 03 22:29:44 Good afternoon all. I have started working on a project to put linux on my pocketPC smartphone. (HTC apache,PP6700). I have used Haret to boot a familiar kernel with opie but the hardware drivers need work. So I searched around on how to best do this. After setting up the crosscompile environment manually on my gentoo AMD64, I realized I didn't have a good start point for the kernel config or what files are needed in a "distribution" like Jun 03 22:31:22 I am finding that the documentation is still being written and I am stuck on bitbake bugs during package install. Jun 03 22:31:24 the error ends in "AttributeError: 'NoneType' object has no attribute 'endswith' ", Any clues? Jun 03 22:32:16 Jshanab: not much Jun 03 22:32:26 Jshanab: which version of bitbake are you using? Jun 03 22:32:40 Jshanab: it should print an environment Jun 03 22:34:06 jshanab you know the htc model porting sides? Jun 03 22:34:07 1.4.0!!!!. Coming form gentoo, it is really cool to see the recipes are just like ebuilds, but looks like I missed something that should be 1.4.2 Jun 03 22:34:33 Jshanab: well, not too much changed in this regard then Jun 03 22:34:51 Jshanab: which task of which package is failing? Jun 03 22:34:55 hey drw Jun 03 22:35:09 hi zecke Jun 03 22:35:39 any or all bitbake commands executed in the build directory ie "bitbake nano" Jun 03 22:36:07 woglinde, you were asking about sites? Jun 03 22:36:27 Jshanab: well, more specific? Is it failing directly? Jun 03 22:36:39 Jshanab: or is it parsing stuff? Jun 03 22:36:39 jshanab no I only want to know if you come form them Jun 03 22:37:29 http://wiki.xda-developers.com/index.php?pagename=ApacheProgress I found this site, but absolutly no activity or responses in monthes. zecke, let me post the output... Jun 03 22:37:41 jshanab ah okay Jun 03 22:37:59 they concentraded on the blueangel and universal Jun 03 22:39:02 http://pastebin.com/756621 Jun 03 22:40:08 I have a HTC_apache and it is a perfect oppertunity to learn this stuff and I may found a way to replace the screen with a higher resolution :-) Jun 03 22:41:51 dig this: bitbake -v -n librsvg | grep current | sed 's/NOTE: current path: //' | sed 's/\(.* -> \)*\(.*\) -> \(.*\)$/\2 --> \3/' | sed 's/\(.*\) (.*) --> \(.*\) (.*)/\1 --> \2/' Jun 03 22:41:58 PSYCO JIT not yet tested in AMD64 in gentoo, do I need it? Jun 03 22:42:04 no Jun 03 22:42:23 Jshanab: ah Jun 03 22:42:27 Jshanab: update to bitbake 1.4.2 Jun 03 22:42:35 Jshanab: and consider fixing your BBPATH Jun 03 22:42:58 BBPATH is ok, but I will try 1.4.2 Jun 03 22:43:10 Jshanab: certainly not Jun 03 22:43:46 Jshanab: this bug was first seen by lrg who had a wrong BBPATH Jun 03 22:43:53 Jshanab: and conf/bitbake.conf wasn't picked up from OE Jun 03 22:44:02 Jshanab: and then CACHE was not set at all :) Jun 03 22:47:40 does this look correct Jun 03 22:47:42 export BBPATH=/stuff/build:/stuff/org.openembedded.oz354fam083 Jun 03 22:49:09 yes :) Jun 03 22:49:42 CACHE ??? Jun 03 22:50:03 to avoid parsing again :) Jun 03 22:50:45 rhcp started Jun 03 22:50:51 woglinde: southpark as well Jun 03 22:50:53 http://www.aol.de/index.jsp?sg=Musik_special_RockinRio&cid=881679479 Jun 03 22:50:54 CCACHE ? Jun 03 22:51:14 Jshanab: no Jun 03 22:51:18 CACHE Jun 03 22:51:43 http://www.openembedded.org/repo/org.openembedded.oz354x/conf/bitbake.conf Jun 03 22:51:50 this CACHE Jun 03 22:52:01 and you can put it anywhere... /stuff/ is not a... Jun 03 22:53:24 we lost zecke Jun 03 22:53:42 ? Jun 03 22:53:54 Jshanab: http://bugs.openembedded.org/show_bug.cgi?id=895 Jun 03 22:54:01 I missed that one in the docs, thanks zecke Jun 03 22:54:25 cool! Jun 03 22:54:27 Jshanab: well, updating to 1.4.2 should fix your issue Jun 03 22:54:52 Jshanab: and this issue was uncovered due a wrong BBPATH Jun 03 22:55:33 ok, I will create an ebuild and give her a go Jun 03 23:01:06 night all Jun 03 23:01:47 zecke: well as we say in France 'la nuit porte conseil' or 'night makes ideas more clear' Jun 03 23:02:00 ;) Jun 03 23:02:03 cyrilRomain: sleep? Jun 03 23:02:12 zecke: yep, i need Jun 03 23:02:26 * zecke waits for the beging of the NBA semi final :} Jun 03 23:02:39 :) Jun 03 23:02:44 zecke provide a stream *g* Jun 03 23:02:59 woglinde: I would have picked you up... Jun 03 23:03:11 zecke no its okay Jun 03 23:03:17 I ahve the musci Jun 03 23:17:21 so guys, under what circumstances the kernel will actually "see" the sd card? i reset the ipaq and took the battery for a while Jun 03 23:17:31 after rebooting the card was mountable again Jun 03 23:18:18 from a clean start Jun 03 23:18:36 there are issues with the suspend scripts and any events associated arround Jun 03 23:19:24 mostly the fact that devices are not unmounted before they are suspended Jun 03 23:19:58 so if your device has a long power up time or the device was removed unpredictable behavoir happens Jun 03 23:20:48 subsiquent devices wont mount, total ipaq lockup, etc Jun 03 23:33:36 emte: i wasn't aware of the unmounting process everytime the ipaq goes in suspend mode Jun 03 23:35:18 emte: it was doing the job thou; i could use the device normally after resuming, but then you are right, there was a moment in which i started having very strange behavior such as spontanous reboot Jun 04 00:41:24 I have upgraded to bitbake 1.4.2 and found I have to set BBPATH at a system level, a script exporting it just isn't working for some reason. Now I get ERROR: Unable to open conf/bitbake.conf and am thinking it is the BBFILES in the local.conf Jun 04 00:56:08 It looks as if it is ignoring the BBPATH and using only conf/bitbake.conf Jun 04 00:58:09 nevermind, @#@$ old terminal window , need caffine.... Jun 04 01:08:01 Anyone know where I should set TARGET_ARCH? Jun 04 01:08:29 you shouldnt Jun 04 01:08:35 use the correct MACHINE= Jun 04 01:08:57 it will inherit TARGET_ARCH from that Jun 04 01:09:32 ok, Where is that set? LOL Jun 04 01:09:42 in local.conf Jun 04 01:13:27 thanks, i was looking at the familiar's conf file not the sample. I see it now. Jun 04 01:14:34 you are not trying to build 0.8.4 i hope Jun 04 01:14:56 that is currently in git, not monotone Jun 04 01:15:13 and it is unlikly that unstable will build at all Jun 04 01:15:38 0.8.3 Jun 04 01:16:07 I am trying to follow the Open embedded site, it is young on documentation Jun 04 01:17:51 I will have to pick something close and then tweak it Jun 04 01:18:47 what are you trying to do? Jun 04 01:20:25 I have a HTC-apache AKA sprint PPC6700. I have booted familiar on it but things don't work so some things need to be adjusted for the hardware. I am trying to get to the point where I can work on the drivers. Jun 04 01:20:52 It has a xscale270 ie armv5te Jun 04 01:21:37 I wonder if the ipaq-pxa270 would be a good choice Jun 04 01:30:33 possibly Jun 04 01:30:43 someon else i think has a HTC board ... Jun 04 01:33:10 HTC blue angel or something ? Jun 04 01:33:14 I got HTC magician there. Jun 04 01:33:58 how come the familiar conf doesn't set the machine? Jun 04 01:34:31 beciuase its a distro not a machine Jun 04 01:34:35 because* Jun 04 01:35:05 familiar covers ~ten+ series of devices Jun 04 01:37:16 ok, i am missing a few steps then. How do I create an image for my device. This needs to be a bootable image containng the kernel, the base utils and well... a distribution . What is that config for? Jun 04 01:38:00 the only config you need to edit is local.conf Jun 04 01:38:27 and in that the bare min you need to set are your paths, DISTRO, and MACHINE Jun 04 01:42:22 crap, I had followed someones instruction to copy familiar's conf to local.conf. That was wrong ? Jun 04 01:46:05 i wouldnt expect it to work Jun 04 01:46:10 what page are you reading? Jun 04 01:46:46 To tell you the truth I have forgot! Jun 04 01:47:37 emte: what about DISTRO_VERSION? Shouldn't it be set if using DISTRO="familiar"? Jun 04 01:48:09 no Jun 04 01:48:16 er your asking Jun 04 01:48:20 yeah it should Jun 04 01:48:29 that is a recent change that i forget about Jun 04 01:48:39 it used to be one string Jun 04 01:48:48 then for some reason it was changed Jun 04 01:48:55 btw, familiar-unstable breaks on gpe-calendar, koen left a but about it a few days ago. Jun 04 01:49:10 it just broke on mine, after the whoooole day building it. crap Jun 04 01:49:30 gee, koen left a bug open, not his but. Jun 04 01:49:31 ow Jun 04 01:49:41 unstable will probably fail on almost everything Jun 04 01:50:01 hehe, i know it the hard way Jun 04 01:52:39 Jshanab, you should be reading/following this: http://www.openembedded.org/wiki/GettingStarted Jun 04 01:58:06 I am, just managed to get interruppted during some build problems. I think I got it now, I am configuring local.conf. Jun 04 02:02:18 so guys, the best thing to do would be to umount all media devs manually before going into sleep mode ? Jun 04 02:03:56 CSMan, yeah thats what i do Jun 04 02:04:08 and my ipaq behaves Jun 04 02:04:18 emte: i'll make a little script for it Jun 04 02:04:21 =) Jun 04 02:04:28 just pull out the device Jun 04 02:04:34 :) Jun 04 02:07:31 Not to start a flame war, but should I use linux-uclibc or linux Jun 04 02:08:46 i would start with linux Jun 04 02:08:52 and tehn move to uclibc Jun 04 02:10:27 emte: just pull out the device ? Jun 04 02:11:47 sd, cf, pcmcia Jun 04 02:11:57 they will umount on removal Jun 04 02:12:14 unload* Jun 04 02:12:25 curiously... my h2200 has no need for umount before sleep. Jun 04 02:12:37 it is magic, i believe. and rootfs is on cf. Jun 04 02:12:46 it is booting from lab. Jun 04 02:12:49 thejapa: it's because you have 2.6.sx Jun 04 02:12:51 SO, what do I build to create the whole image? the getting started is a little vague on that point Jun 04 02:12:54 ah Jun 04 02:13:25 you should buy a pocket of your favorite sci fi writer and do bitbake gpe-image or opie-image Jun 04 02:13:38 lol Jun 04 02:13:44 if your a good reader, you may even finish the book before the build completes Jun 04 02:14:00 gpe-image would take trillions of cpu cycles on a PII machine Jun 04 02:14:37 but I did opie-image a few days ago. it didn't work though. Jun 04 02:14:51 i'm new to this stuff too. ;) Jun 04 02:15:05 I am watching stargate as i work on this, LOL. While we wait for angstrum, which do you like best GPE or OPIE, or what is the difference in less than 100 words Jun 04 02:15:45 I am gonna try to do all compiling on an AMD64 Jun 04 02:15:51 opie has more apps, but i can't find a working image for h2200. Jun 04 02:16:08 so, it leaves me with gpe. but i do have a thing for gnome stuff. Jun 04 02:23:08 This looks like I missed a step? "ERROR: Nothing provides dependency opie-image" Jun 04 02:26:20 jshanab: it should be there, which branch you are using? Jun 04 02:27:01 i mean, are you using /stuff/org... what? Jun 04 02:27:14 org.embedded.oz354fam083 Jun 04 02:28:17 is there a file org.embedded.oz354fam083/packages/meta/opie-image.bb ? Jun 04 02:29:00 yup Jun 04 02:29:53 have you set BBPATH correctly? Jun 04 02:30:19 like BBPATH=/stuff/build:/stuff/org.openembedded.oz354fam083 Jun 04 02:30:44 but....do I set DL_DIR to something? Jun 04 02:31:35 wait, BBPATH is an environment variable, not a setting in local.conf Jun 04 02:31:35 yes BBPATH is set exactly as you say Jun 04 02:31:40 ok Jun 04 02:31:59 ghidrah conf # env |grep BBPATH Jun 04 02:32:00 BBPATH=/stuff/build:/stuff/org.openembedded.oz354fam083 Jun 04 02:33:12 the DL_DIR is where bitbake will download tar.gz files to compile. you can put whatever path you want here. mine is DL_DIR = "/stuff/sources" Jun 04 02:33:17 but this is personal. Jun 04 02:33:56 in your local.conf is it BBFILES = "/stuff/org.openembedded.oz354fam083/packages/*/*.bb" ? Jun 04 02:35:44 There the bugger is...org.embedded notorg.openembedded me bad Jun 04 02:36:53 hey, fixing other ppl problems is way easier than our own, i should do it more often. :D Jun 04 02:37:38 well, I have a screen full of errors now, i guess that is progress ? Jun 04 02:38:34 emte: now after mounting as user, it won't let me dismount unless i'm root Jun 04 02:38:51 lol Jun 04 02:38:54 i know why. Jun 04 02:39:00 emte: busybox. Jun 04 02:39:01 they look like : Jun 04 02:39:02 =P Jun 04 02:39:03 ... Jun 04 02:39:04 ERROR: ("can't read from file '%s' (%s)", ('/stuff/build/tmp/staging/HTC-apache-linux/kernel/kernel-abiversion', )) while parsing /stuff/org.openembedded.oz354fam083/packages/atmelwlandriver/atmelwlandriver_3.3.5.6.bb Jun 04 02:39:06 ... Jun 04 02:40:03 jezz Jun 04 02:40:29 you cant build abi with familiar Jun 04 02:41:10 what's abi? Jun 04 02:41:16 Educate me, what is abi Jun 04 02:41:23 hehe Jun 04 02:42:55 Jshanab: Application binary interface.. in this context, it refers to EABI, afaik, which is a way of ensuring interoperability between apps compiled with different settings. Jun 04 02:43:08 Dumb question, do I need familiar? How hard is it to make my own distribution? Jun 04 02:43:12 (anyone, feel free to chip in if I'm off the mark) Jun 04 02:43:49 OMG ABI I have used that before to let old solaris code run on linux years ago Jun 04 02:44:41 Jshanab: Probably best to just use familiar.. creating a distro wouldn't be too easy. Jun 04 02:45:49 you can try building packages in org.*/packages that works Jun 04 02:46:09 but you could just use ipkg Jun 04 02:46:37 thejapa: now i have to go all the way to the server in Lima to build umount =P Jun 04 02:47:08 CSMan, I don't understand why they put a lima inside the Peru. Jun 04 02:47:12 hohoho Jun 04 02:47:15 XD Jun 04 02:47:38 very funny japinho Jun 04 02:49:00 How do I turn off abi? where is the darn kernel config, menuconfig Jun 04 02:50:43 jshanab: I don't know if abi can be turned of, and i 'm curious about how to build custom kernels too. Jun 04 02:52:49 I do it all the time in my desktop and even the crosscompiler toolchain, i just don't know the stuff to select. I am looking for a starting point. I came to the openembedded in search of the kernel .config file for familiar Jun 04 02:54:10 where in the tree is the original ping util ? Jun 04 02:54:17 I assume it is gonna be Jun 04 02:54:19 bitbake kernelheaders Jun 04 02:54:20 bitbake kernel-2.16.16 Jun 04 02:54:22 make menuconfig Jun 04 02:54:23 make clean modules modules_install bzImage Jun 04 02:54:25 etc.... Jun 04 02:54:29 compiling kernels for desktop is ok for me too. I just have no idea how to do it for h2200. Jun 04 02:54:41 or whatever handheld. Jun 04 02:55:29 it should be the same, you get a kernel-hXXXX package and the respective kernel-modules-hXXXX package Jun 04 02:55:54 emte: is that correct? Jun 04 02:56:04 It should be identical. amen CSMan Jun 04 02:56:20 no Jun 04 02:56:21 ah, but to alter the .config I believe you have no access to menuconfig Jun 04 02:56:29 its quite a bit more convoluted than that Jun 04 02:56:43 i'm not happy to be right this time. Jun 04 02:56:57 you need to remember that the distro defines rule of what to pack in the image etc Jun 04 02:58:00 The distro is just a big metapackage. The drivers to handle GPIO is the fun part Jun 04 02:58:24 i'm trying to find the original ping utility Jun 04 02:59:00 the PXA270 developer pdf is like 1300 pages! Jun 04 02:59:40 CSMan: iputils? Jun 04 02:59:46 Jshanab: That' not surprising... Jun 04 02:59:48 thejapa: good hint **** ENDING LOGGING AT Sun Jun 04 02:59:56 2006