**** BEGIN LOGGING AT Fri Oct 07 02:59:56 2005 Oct 07 03:50:32 hi guys, do you have this problem with gst-plugins-0.8.7-r3 too? http://pastebin.com/386042 Oct 07 04:22:49 Arjan: may want to check whether the source tarball has been extracted correctly Oct 07 04:56:56 Hey. Has anyone here with an ARM based device successfully used snes9x version 1.43? Oct 07 04:57:12 There's a bb for it, it should work under GPE Oct 07 04:57:28 im trying to port it to SDL, its working on the desktop but my pDA just sits there Oct 07 04:57:41 does anyone know if the 1.43 version works in GPE on an arm device? Oct 07 04:57:45 or can anyone test? Oct 07 05:22:19 03florian 07org.oe.dev * r209a6256... 10/: gpe-ownerinfo: Add new release. Oct 07 05:30:06 03rw 07org.oe.dev * r634dd1df... 10/: bitbake.conf: add FREEBSD_MIRROR Oct 07 05:30:10 03rw 07org.oe.dev * rcaf25ee1... 10/: rpm2cpio: fix SRC_URI (uses newly introduced FREEBSD_MIRROR). fixes Bug #338 Oct 07 05:42:55 hmm, glibc-2.3.5 doesn't seem to build Oct 07 05:42:57 checking for suffix of object files... configure: error: cannot compute suffix of object files: cannot compile Oct 07 05:43:33 MP0: which platform? Oct 07 05:43:45 openomap Oct 07 05:44:03 my oe repository is dated 2005-10-07T07:49:42 Oct 07 05:44:05 well how old is your machine.conf? Oct 07 05:44:08 hehe Oct 07 05:44:19 that sounds like the march or mtune is set wrongly Oct 07 05:45:41 the omap5912osk.conf seems to use tune-arm926ejs.conf Oct 07 05:46:04 -march=armv5te -mtune=arm926ej-s Oct 07 05:46:52 now look at the config.log of glibc Oct 07 05:47:09 it will say it does not know arm926ej-s (don't ask me what the correct tune would be) Oct 07 05:51:01 zecke: :\ I'm porting over snes9x 1.43 WIP to SDL so it can run under opie Oct 07 05:51:09 i wrote a backend that works on my desktop Oct 07 05:51:17 however on my PDA, it starts, displays the framerate and all Oct 07 05:51:25 but sits there at a black screen...not frozen Oct 07 05:51:26 but not working Oct 07 05:51:37 i'm baffled :| Oct 07 05:51:48 I'm going to assume ej-s is a typo and try ejs Oct 07 05:54:02 hmm... 03koen 07 * r1.3540 10openembedded/conf/machine/tune-arm926ejs.conf: tune-arm926ejs.conf: fix typo in mtune parameter Oct 07 05:54:14 ... strange Oct 07 06:07:27 ... yeah, it's arm926ejs: http://www.ip97.com/gcc/ARM-Options.html Oct 07 06:07:50 koen keeps complaining about having to remove the hyphen... I take it that the hyphen keeps creeping back into oe Oct 07 06:08:23 crap Oct 07 06:30:12 mreimer: ping Oct 07 06:48:00 03koen 07org.oe.dev * r3b01989f... 10/: packages/gpe-login/: add gpe-login 0.83 Oct 07 06:49:26 Good Morning! Oct 07 06:49:52 Has anyone else had a problem with compiling an image, and then most of the network programs not work (like atftpd, appWeb, etc...) Oct 07 06:50:07 the instant you try to run it, I get a 'segmentation fault' Oct 07 06:50:59 what sucks is - the image I had before this (which was huge & had the debug info compiled in) worked fine. I then removed the 'tmp' directory & told it not to build with the debug stuff, and now I have this problem Oct 07 06:51:23 install catchsegv and see where it segfaults Oct 07 06:55:00 thanks koen - will do, brb Oct 07 07:14:43 morning Oct 07 07:15:44 03koen 07org.oe.dev * r9eacd7bb... 10/: packages/btscanner/btscanner_1.0.bb: fix compile with recent bluez headers, patch adapted from https://www.rocklinux.net/submaster/data//2005/06/0909132213173.patch Oct 07 07:15:48 03koen 07org.oe.dev * rd158643d... 10/: packages/btscanner/btscanner-1.0/0909132213173.patch: don't forget to commit the patch too..... Oct 07 07:16:23 running 'catchsegv atftpd' still gives me the 'Segmentation Fault' immediately, and I'm not seeing any other information. Am I using it incorrectly? Oct 07 07:28:40 I have a little question about .bb files Oct 07 07:29:14 yes? Oct 07 07:29:18 can anyone help me? Oct 07 07:29:29 probably not Oct 07 07:29:41 ok Oct 07 07:29:50 so what can I do? Oct 07 07:30:05 Pigi: hey, could you send me the 'FreeBSD' patch? Oct 07 07:30:12 dmunicio: ask, and hope someone will answer Oct 07 07:30:14 Ciao all. Oct 07 07:30:16 you can... ask a question instead of asking if we can help.. Oct 07 07:30:25 zecke, sure. Oct 07 07:30:51 pb_, can I /msg you ? Oct 07 07:30:58 hey Pigi Oct 07 07:30:59 ok, I want to include a bitmap in my rootfs in /usr/share/pixmaps for example Oct 07 07:31:15 hey koen ! Oct 07 07:32:27 hey koen Oct 07 07:32:28 dmunicio: rootfs or a 'package'? Oct 07 07:33:34 so I create a .bb file similar to gpe-terminal.bb in which I only want to copy this file to the rootfs Oct 07 07:33:42 I noticed on some chatlogs you were complaining about arm926-ej-s Oct 07 07:33:42 zecke, check mail ! Oct 07 07:33:54 Pigi: I can't :( Oct 07 07:34:05 (in reference to configuring glibc) Oct 07 07:34:17 the current oe repository has it still mispelt Oct 07 07:34:34 I fixed it in mine, but is there any way to get it fixed in the oe repository? Oct 07 07:34:46 zecke, how am I supposed to send you the patch ? Oct 07 07:35:07 zecke, or simply you can't ATM ? Oct 07 07:35:39 ATM I can't Oct 07 07:36:33 ok, np Oct 07 07:38:47 pb_ disregard my previous msg :) Oct 07 07:42:07 Hello everybody Oct 07 07:43:51 I've setup the (amazing) OE/BB build system and I have successfully built some packages for openzaurus on C7x0. Oct 07 07:44:42 However when I try to build any -image (opie, gpe, e) compilation fails on module-init-tools-cross-3.1-r3 Oct 07 07:45:41 dmunicio: well creating the rootfs is a packe itself Oct 07 07:45:52 dmunicio: it will install available packages Oct 07 07:46:07 ~pastebin Oct 07 07:46:09 i heard pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca/ Oct 07 07:46:15 Specifically, on insmod.static I get the message "/usr/bin/ld: cannot find -lc" Oct 07 07:46:16 prasinos: pastebin your error please Oct 07 07:46:33 prasinos: oh then you need to install it Oct 07 07:46:49 prasinos: (glibc static) Oct 07 07:47:37 http://pastebin.com/386214 < should I retry? Oct 07 07:47:48 zecke: you mean I have to install glibc-static on the machine I am doing the compilation. (I'll try now...) Oct 07 07:47:59 prasinos: right Oct 07 07:48:09 anyone else have any suggestions on my immediate 'Segmentation fault' whenever I run a few servers (like appWeb, atftpd)? Oct 07 07:48:41 dwildes: dmesg, check the alignment trap, and do as koen said, install catchsegfault and strace Oct 07 07:49:58 zecke: a dmesg gives me no info on the trouble, a 'catchsegv ' still gives me the segfault, and with no other information, dump file or anything Oct 07 07:50:32 haven't installed strace though yet, I'll run it and see what happens. I'm looking through a strings output now Oct 07 07:53:03 dwildes: if you can not run a full blown image, you might simply want to avoid striping of the binaries? Oct 07 07:53:14 dwildes: and you can allow debugging on a per package level Oct 07 07:53:15 zecke: That wat it! It needed glibc-static-devel while I had only glibc-devel installed. Thanks a lot. Oct 07 07:54:30 prasinos: well next thing is to find out why it needs it Oct 07 07:54:44 prasinos: or at least find an appropriate place in the wiki to add the knowledge to Oct 07 07:56:14 zecke: when I make bootstrap-image... al the ipkg are installed, aren't they? Oct 07 07:56:35 i mean all the ipkg Oct 07 07:56:51 I did have another image built earlier that had all of the debug turned on in the local.conf & that image worked okay. Now I was wanting to make my image smaller by commenting out the debug statements. I then removed the 'tmp' directory, and recompiled from scratch. After I done that, I'm getting the segfaults. (if that helps any) Oct 07 07:58:15 later all, guys Oct 07 08:18:22 anyone willing to help me figure out this strace output? :-) Oct 07 08:18:53 I got the image compiled with strace now & finally getting some 'segmentation fault' info Oct 07 08:21:19 URL for the strace output: http://pastebin.ca/24851 Oct 07 08:23:27 * MP0 looks Oct 07 08:25:55 I'm not sure if this is normal but it's having a bit o difficulty with nonexisting libraries Oct 07 08:26:45 it seems like it got killed right after it attached its signal handlers and checked the kernel version. Oct 07 08:43:35 can anyone help me Oct 07 08:44:39 I think the problem is missing libraries, but I don't know how to solve it. Oct 07 08:45:13 if you look at your strace, it tries to open some libraries and fails Oct 07 08:46:00 MP0 - yeah, I'm digging through now to see where I can install those libraries. Hopefully that'll solve this. :) Oct 07 08:49:55 dmunicio: You still there? You need to edit the meta file which creates the image you're trying to build and add your package name to it Oct 07 08:55:00 anybody succesfull creating any akita image? Oct 07 08:55:33 I compiled several packages, but I' having problems with images Oct 07 08:55:42 I don't know if it's my local.conf Oct 07 08:57:42 MP0: apparently those two files (ld.so.preload & ld.so.nohwcap) do not have any 'real' relation to the issue. They are more/less config files for the LD libraries Oct 07 08:59:06 so, I don't believe they are causing the segfault, it must be the line #43 in the post Oct 07 09:00:45 lardman|busy: have you read my post? Oct 07 09:01:10 dmunicio00: since I replied to you? Oct 07 09:01:30 yes Oct 07 09:01:45 dmunicio00: I've not seen anything Oct 07 09:03:13 you know what - every app that is 'segfaulting' end with that last line: Oct 07 09:03:17 _sysctl({{CTL_KERN, KERN_VERSION, 0, 20d49, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}, 2, 0xbffffa94, 30, (nil), 0}) = 0 Oct 07 09:05:22 And I don't have the executable 'sysctl' in the image. Doesn't that suppose to be included within glibc? Oct 07 09:07:26 dmunicio00: If you want to add your package to the bootstrap image, then look at packages/meta/bootstrap-image.bb which points to task-bootstrap.bb. Edit the last one and add the name of your package to it, then it will be installed. Oct 07 09:09:56 is it normal getting fetch failed due to servers down ? Oct 07 09:14:11 is anyone else running an x86 image of OE? Oct 07 09:14:54 indeed Oct 07 09:15:14 cool - do you have 'atftpd' or 'appWeb' installed? Oct 07 09:16:36 oh, no Oct 07 09:16:50 I subscribe literally to the E part of OE Oct 07 09:18:36 i said: are you sure? I thought that all the ipk files are installed Oct 07 09:19:27 CosmicPenguin: thanks anyway Oct 07 09:19:35 sorry Oct 07 09:20:08 np - I'm sure it'll get figured out eventually *hahah* Oct 07 09:31:52 dmunicio00: Yes, I am sure, if all of the ipkg files which are present in tmp/deploy/ipk were installed my flash images would be rather large Oct 07 09:44:40 bye Oct 07 09:45:13 mickey|resting: monotone.vanille.de is timing out again...:-( Oct 07 10:11:23 zecke: pong Oct 07 10:44:48 where does unconfigured.sh come from? I don't see it in rootfs **** BEGIN LOGGING AT Fri Oct 07 11:24:33 2005 Oct 07 11:45:20 Why is packages/tslib/tslib/tslib.sh empty? Oct 07 11:46:59 | checking for main in -lm... configure: error: Link tests are not allowed after GCC_NO_EXECUTABLES. Oct 07 11:53:35 the configure script for gcc-cross-4.0.0-r1 appears to be broken Oct 07 11:53:48 it complains about Oct 07 11:53:49 | checking for main in -lm... configure: error: Link tests are not allowed after GCC_NO_EXECUTABLES. Oct 07 11:54:04 The accepted solution appears to be either to remove the check from the configure script Oct 07 11:54:06 or Oct 07 11:54:19 set an env variable to skip the check Oct 07 11:56:01 namely ac_cv_search_strerror="none required" Oct 07 11:56:01 export ac_cv_search_strerror Oct 07 11:56:34 before I done an update this morning (the last one was about 3weeks ago), I was able to compile 'mysql', and now I can't compile it. :-( I'm wondering if this is where my other 'segment faults' are a potential result from? Oct 07 11:56:38 it ends with: Oct 07 11:56:43 default.c:397: internal compiler error: in get_indirect_ref_operands, at tree-ssa-operands.c:1449 Oct 07 11:57:42 I'm using gcc='4.0.1', glibc='2.3.2' for machine='x86' Oct 07 11:58:48 * MP0 doesn't know Oct 07 11:58:58 MP0: what platform are you compiling for? Oct 07 11:59:30 my gcc-cross-4.0.0-r1 packages compiled at least Oct 07 11:59:52 just checked under tmp/deploy/ipk Oct 07 12:00:17 openomap Oct 07 12:00:46 I'm going to try doing bitbake again and hope I get a better result Oct 07 12:02:21 <---- has not once been able to build a bootstrap-image Oct 07 12:03:07 what machine spec are you using? I can try a compile over here to see if I get same/similar results Oct 07 12:03:48 BB_VERSION = "1.3.2"OE_REVISION = "694a5fa155d742b0e06d8723ba2a338fac5c716e"TARGET_ARCH = "arm"TARGET_OS = "linux"MACHINE = "omap5912osk"DISTRO = "openomap"TARGET_FPU = "soft" Oct 07 12:04:27 I had to fix a mispelling in one of the machine .conf files Oct 07 12:05:09 in Oct 07 12:05:10 org.openembedded.dev/conf/machi Oct 07 12:05:12 ne/tune-arm926ejs.conf Oct 07 12:05:47 -mtune should be -mtune=arm926ejs as opposed to arm920ej-s Oct 07 12:06:19 hmm, just tried to bitbake again... same problem Oct 07 12:07:00 basically the problem is well described in http://www.nabble.com/crosstool-and-uClibc-t120686.html Oct 07 12:07:05 so hopefully I can fix up the files Oct 07 12:07:09 * MP0 crosses fingers Oct 07 12:07:11 okay - I'm configured for your platform, about to compile Oct 07 12:07:45 how long did you compile before getting errors? Oct 07 12:08:14 uhm, hours Oct 07 12:08:18 I'm on a 400MHz machine Oct 07 12:08:35 bootstrap-image anyways Oct 07 12:08:56 ah - on a 1.8ghz AMD, so it might be an hour or so before I'll have info for ya Oct 07 12:09:34 * MP0 mouth waters Oct 07 12:10:02 * MP0 thinks of all the RAM Oct 07 12:10:14 I've got 2gigs RAM :) Oct 07 12:10:45 You realize that after gcc 4 builds on your computer... Oct 07 12:10:53 ... I'm going to have to kill you and take your RAM Oct 07 12:11:02 but your compile dies in gcc-cross right? Oct 07 12:11:05 * dwildes laughs Oct 07 12:11:20 yeah Oct 07 12:11:59 well - it's off, I'll let ya know what happens Oct 07 12:12:17 k Oct 07 12:18:39 meanwhile, I've made OE think it hasn't touched gcc-4 and try to d/l it again. Oct 07 12:18:51 we'll see how a new download from a different source goes Oct 07 12:21:32 sounds good! Oct 07 12:41:11 so anyone going to end up coming to the Linux Kongress? Oct 07 12:42:32 If there's gonna be a Linux Senate, I'll get 12 old men. Oct 07 12:46:33 * koen watches how OE succesfully checks out opie-pcmciaapplet Oct 07 12:46:39 why am I not surprised by that? Oct 07 12:47:43 Low expectations? Oct 07 12:48:25 I guess Oct 07 12:52:14 * MP0 cheers on bitbake Oct 07 12:52:19 go little bitbake, go! Oct 07 12:55:41 nogo, | checking for main in -lm... configure: error: Link tests are not allowed after GCC_NO_EXECUTABLES. Oct 07 12:55:54 stupid gcc-cross-4.0.0 Oct 07 12:56:48 MP0 - I'm downloading 'gcc-cross-initial-3.4.4+csl-arm-20050603-r0' now. I'm only on a 256k link, so I still have about 15min left :( Oct 07 12:57:00 :-( Oct 07 12:57:24 but, so far so good for bitbake bootstrap-image Oct 07 12:57:37 dwildes: you don't want to use the csl compiler Oct 07 12:57:56 dwildes: http://bugs.openembedded.org/show_bug.cgi?id=204 Oct 07 12:58:51 that's what bitbake bootstrap-image is downloading Oct 07 12:59:42 koen: I'm just testing a build for MP0 to see if it compiles in my environment (gcc-cross-4.0.0) Oct 07 13:01:00 gcc-4.0.x where x < 2 miscompiles code Oct 07 13:01:03 don't use it Oct 07 13:01:36 reenoo - you may have solved my problem with an earlier issue :D Oct 07 13:02:08 the openembedded repository I Oct 07 13:02:08 I was compiling with gcc-4.0.1 on x86, and the packages did compile, but immediately segfault when ran Oct 07 13:02:16 'm using is from 2005-10-07T07:49:42 Oct 07 13:02:25 and it's using the csl compiler and 4.0.0 Oct 07 13:02:33 (4.0.0 fails to configure) Oct 07 13:03:09 You're saying we shouldn't use them... how do we get openembedded to use something else? Oct 07 13:03:12 but what is strange is - the packages DID work a few weeks ago Oct 07 13:03:31 not until did I update this morning, & recompile did I start getting the segfaults Oct 07 13:03:32 MP0: look at conf/distro/familiar-0.8.3.conf Oct 07 13:03:40 should we use GCC=4.0.2 ? Oct 07 13:03:58 what's recommended? Oct 07 13:04:32 unless you rely on gcc 4 features (e.g. gcj) I suggest you wait for gcc 4.1 Oct 07 13:05:26 koen: could you restart the monotone server on vanille.de again? Oct 07 13:05:54 It doesn't matter to me ultimately, because I don't know much about the differences. I was just told to try this GCC version several weeks ago when something else wouldn't work Oct 07 13:06:17 k, so if I set PREFERRED_VERSION_gcc in my local.conf, I'll be good? Oct 07 13:06:23 reenoo_: done Oct 07 13:06:54 koen: thanks Oct 07 13:06:59 * koen starts thinking about killfiles Oct 07 13:07:13 ~praise killfiles Oct 07 13:07:22 hmm Oct 07 13:07:27 no ibot it seems Oct 07 13:07:33 reenoo_: What gcc version do you use? Oct 07 13:07:58 03rw 07org.oe.dev * r1b4d1903... 10/: bluez-utils: remove duplicate .inc. apply dbus.patch for -dbus variant. Oct 07 13:10:34 that cleans up the bluez-utils mess Oct 07 13:11:26 dwildes: 3.4.x I believe. like koen said, have a look at familiar-0.8.3.conf or the default preferences for arm or something Oct 07 13:13:21 reenoo_: thanks :) Oct 07 13:13:29 it has 3.4.4 listed here Oct 07 13:14:28 who maintains openomap distro? Oct 07 13:14:43 MP0: Crofton|laptop Oct 07 13:14:50 k Oct 07 13:15:42 Think he'll agree to put the preferred gcc lines into the openomap distro file? Right now I have them in my local.conf, so it doesn't make much difference to me, but someone else might run into the problem. Oct 07 13:16:41 and the arm926ej-s => arm926ejs thing might be nice Oct 07 13:17:04 in -mtune Oct 07 13:17:21 03koen 07org.oe.dev * r2be87031... 10/: packages/btscanner/btscanner_2.0.bb: add btscanner 2.0 Oct 07 13:20:05 reenoo_: thanks for fixing bluez-utils-dubs Oct 07 13:20:10 er, dbus Oct 07 13:20:42 np Oct 07 13:22:03 I figured out why monotone is so slow Oct 07 13:22:11 why? Oct 07 13:22:14 it doesn't draw to the framebuffer directly Oct 07 13:22:19 lol Oct 07 13:22:35 * MP0 chuckles Oct 07 13:22:35 heh Oct 07 13:22:56 ysltLTG Oct 07 13:23:08 oops Oct 07 13:23:24 * koen tries to figure out that acronym Oct 07 13:27:03 MP0: sorry bud - errors over here Oct 07 13:27:34 I'm going to try these new gcc parms that reenoo has suggested and see if that fixes my x86 build problems Oct 07 13:27:59 dwildes, yeah Oct 07 13:28:18 my errors in gcc-4 are due to gcc-4's configure script not having cross compiler joy Oct 07 13:28:35 so I figure hopefully setting the preferred gcc-cross version to 3.4.4 'aught to work Oct 07 13:29:03 maybe for mine as well - be kinda funny both of our problems could be solved with the same thing, huh? :) Oct 07 13:29:12 serendipity ;-) Oct 07 13:29:36 03rw 07org.oe.dev * r99a817d6... 10/: bluez-utils: re-enable dbus.patch for 2.21 as well. Oct 07 13:29:36 Unfortunately I'm still going to have to kill you. Oct 07 13:29:46 * dwildes hides Oct 07 13:29:59 :-) Oct 07 13:30:19 * dwildes locks his RAM in the vault and heads for some unnamed island :) Oct 07 13:30:48 I wouldn't mind a 1.8GHz chip either. Oct 07 13:31:09 oh - uhh, it blew up while compiling...yeah, that's it Oct 07 13:31:17 mreimer_: fixed for real now... Oct 07 13:32:00 oh... so it was really a 400MHz chip clocked up? I don't want it then. Oct 07 13:33:35 you know what would be cool, is if we had a way to setup a compiling farm (similar to Cinelerra) for OE, using multiple computers to do the building :) Oct 07 13:33:49 03mreimer 07org.oe.dev * r4ba84c5c... 10/: Add h2200 support Oct 07 13:34:38 I suggested that they have all of the bitbake and oe.db stuff be server side Oct 07 13:34:48 and just tell clients what they need to do to build something Oct 07 13:34:54 I don't think they liked the idea. Oct 07 13:34:58 inherit += icecream Oct 07 13:35:06 that will make OE distributed Oct 07 13:35:20 icecream is some fancy version of distcc Oct 07 13:35:28 with node monitoring and tooc Oct 07 13:35:33 toolchain distribution Oct 07 13:36:56 koen: awesome!! has anyone tried it? Oct 07 13:37:28 zecke and florian Oct 07 13:37:33 zecke created it Oct 07 13:38:31 mreimer_: for future commits please have a look at this: http://thread.gmane.org/gmane.comp.handhelds.openembedded/5878 Oct 07 13:39:59 thanks reenoo_ Oct 07 13:40:17 :) nice! I have about 4 PCs I could do this with, this would dramatically decrease buildtime for me. It looks simple enough from the docs. I'm going to give this a try tonight Oct 07 13:40:30 thanks koen! Oct 07 13:46:25 ibot! Oct 07 13:46:32 ~botsnack Oct 07 13:46:42 reenoo_: thanks Oct 07 13:48:28 and there was much rejoicing! Oct 07 13:48:53 hehe Oct 07 13:53:16 heh. spiffy Oct 07 13:53:19 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24256 Oct 07 13:53:50 hi there Oct 07 13:54:46 hey giel Oct 07 13:55:32 i've got kerberos running on my slug Oct 07 13:55:39 still finishing up the bb Oct 07 13:58:00 koen: actually you're not quite adhering to the commit standard either. You give too much info. ;-) I'd much rather you list package names than the full file path. Oct 07 13:58:05 koen: btscanner Oct 07 13:58:14 koen: instead of packages/btscanner/btscanner_2.0.bb Oct 07 13:58:27 if someone wants to know what specific file they can look a bit farther Oct 07 13:59:12 there is no 'standard' Oct 07 13:59:18 there's only a proposal Oct 07 13:59:33 true Oct 07 13:59:48 which is why I said I'd personally like just the package name Oct 07 13:59:53 it's supposed to eb a summary, after all Oct 07 13:59:56 details are below Oct 07 14:00:45 (in the commit e-mail) Oct 07 14:00:56 koen: if you disagree with the proposal you are free to respond to oe@. otherwise please follow the policy Oct 07 14:05:09 I extrapolated ": For changes that affect files in conf/ or Oct 07 14:05:09 classes/ you should use the filename of the Oct 07 14:05:10 file." Oct 07 14:05:16 to normal .bbs too Oct 07 14:06:00 name Oct 07 14:06:03 not path Oct 07 14:07:46 what's exactly the problem with using path instead of the name? Oct 07 14:08:02 IMHO efl.bbclass is also acceptable, but classes/efl would be the standard Oct 07 14:08:06 koen: it's more information Oct 07 14:08:14 and? Oct 07 14:08:14 harder to get at what it really is about Oct 07 14:08:31 this is supposed to be a quick summary for those that don't want to look at the files changed Oct 07 14:08:43 and for people to weed things out simply by subject Oct 07 14:08:55 what's exactly the problem with using path instead of the name? Oct 07 14:09:01 if you use the full path then the actual changes are pushed way out Oct 07 14:09:02 koen: tbh because it doesn't fit into the Subject: line in the MUAs I use Oct 07 14:09:26 koen: and I'm on a 1600x1200 screen here Oct 07 14:09:27 and it makes it much harder to see, at a glance, what has happened Oct 07 14:09:42 koen: no way I could read that on my laptop Oct 07 14:10:08 plus the packages/ is redundant the the second directory really adds nothin g Oct 07 14:10:23 all that is needed is the name of the package Oct 07 14:10:59 reenoo_: I'm not on oe-commits, I use cia.navi.cx/stats/project/openembedded/.rss Oct 07 14:11:52 which does not list files Oct 07 14:12:57 it used to list files Oct 07 14:13:38 but cia is down more often than not anyway Oct 07 14:13:54 looking at the bk history we used to list filenames Oct 07 14:14:08 reenoo_: commits get queued for cia Oct 07 14:14:18 well, the policy changed... Oct 07 14:14:27 cia-web is more reliable than cia-bot Oct 07 14:16:17 koen: which bk history are you referring to? the one at oe-devel.bkbits.net doesn't list full paths. Oct 07 14:16:40 reenoo_: I said filenames, not path Oct 07 14:17:03 again. there was no formal policy back then Oct 07 14:17:21 correct Oct 07 14:17:29 and there isn't one now either Oct 07 14:17:30 and if you disagree with the proposal I sent to oe@ please respond to that message Oct 07 14:19:16 with the large amount of contributors with push access we have now there's no room for "I'll do my own thing" any more. the commit messages policy is just a small attempt at bringing a bit of a structure into the chaos Oct 07 14:19:55 think of it as a request for making things easier to follow Oct 07 14:20:31 koen: your commit messages are close to the standard, but they introduce more complecity and make it harder to easily ascertain what is going on Oct 07 14:20:55 koen: I tend to gloss over your commit if I'm not paying attention because it's less apparent what it being changed Oct 07 14:21:11 geez...talk about mistyping... Oct 07 14:21:16 commitS Oct 07 14:21:18 what IS Oct 07 14:23:26 | + set +x Oct 07 14:23:26 | log_check: Using /home/lucas/pda/build/tmp/work/gpe-image-1.0-r17/temp/log.do_rootfs.10431 as logfile Oct 07 14:23:29 | log_check: There were error messages in the logfile Oct 07 14:23:31 | log_check: Matched keyword: [Cannot find package] Oct 07 14:23:34 | Oct 07 14:23:37 | Cannot find package task-bootstrap. Oct 07 14:23:39 | Oct 07 14:23:42 NOTE: Task failed: /home/lucas/pda/build/tmp/work/gpe-image-1.0-r17/temp/log.do_rootfs.10431 Oct 07 14:23:45 NOTE: package gpe-image-1.0-r17: task do_rootfs: failed Oct 07 14:23:47 ERROR: TaskFailed event exception, aborting Oct 07 14:23:49 I should have used pastebin :( Oct 07 14:23:51 NOTE: package gpe-image-1.0: failed Oct 07 14:24:00 what is wrong? Oct 07 14:24:33 ~pastebin Oct 07 14:24:40 rumour has it, pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca/ Oct 07 14:25:18 lucasvo: package task-bootstrap is missing Oct 07 14:25:27 JustinP: but how? Oct 07 14:26:15 reenoo_: mail sent Oct 07 14:27:37 reenoo_: agree with having a policy and all Oct 07 14:28:00 but from the lack of response it's hard to tell wether people agree or don't care Oct 07 14:29:02 I believe it was discussed here in IRC Oct 07 14:29:24 at least somewhat Oct 07 14:31:35 lucasvo: how did you runt hat build? Oct 07 14:32:39 JustinP: bitbake gpe-image Oct 07 14:33:07 lucasvo: well...do you see a task-bootstrap in your tmp/deploy/ipk? Oct 07 14:34:34 task-bootstrap_1.0-r25_pxa250.ipk Oct 07 14:34:52 lucasvo: don't know what to tell you Oct 07 14:34:56 but why is it pxa250 when it should be pxa270? Oct 07 14:35:07 lucasvo: that's likely your problem then... Oct 07 14:35:28 JustinP: so how can I get that task-bootrstrap? Oct 07 14:40:07 lucasvo: I don't know. Sound slike you may have set something up wrong...or there could be an error..what did you set in local.conf? Oct 07 14:47:31 JustinP: I do not know what should be wrong.. Oct 07 14:48:52 lucasvo: ok, so what did you set in local.conf? Oct 07 15:01:57 koen: actually I guess you could just as well use the nslu2-linux rss feed Oct 07 15:02:32 now that they don't use a separate branch anymore Oct 07 15:03:39 that only tracks .dev Oct 07 15:03:45 and adds slug svn Oct 07 15:04:08 ok, then pester mickey to fix the script Oct 07 15:05:22 I still can't see what's wrong with using filenames (not paths!) in the commit message Oct 07 15:05:51 s/filenames/one filename/ Oct 07 15:08:14 filenames for changes to packages are longer, don't add any essential information, and don't scale to the multiple filenames case Oct 07 15:08:42 they do add essential info Oct 07 15:08:49 foo.inc vs foo.bb Oct 07 15:09:20 no, that's either part of the mail body or the stuff cia normally adds anyway Oct 07 15:09:52 if we'd follow that explanation the whole message is useless Oct 07 15:10:04 since the mails will contain a diff Oct 07 15:10:26 dude. I can't read ~50 diffs a day Oct 07 15:10:49 koen: again, the point of the first line of a commit message is a *quick summary* of changes Oct 07 15:11:05 koen: version #'s, inc vs bb, that's all detail which isn't needed there Oct 07 15:11:24 right Oct 07 15:11:54 we want to make it easy for people to track what they want to track and gloss over the rest Oct 07 15:12:01 it makes QA much easier Oct 07 15:12:21 that's the idea Oct 07 15:12:34 people can watch the RSS/webpage/oe-commits list, easily see what concerns them, and then dig in deeper for the things they care about Oct 07 15:12:52 it's a tiny bit of info which clarifies so much Oct 07 15:14:15 but is it needed for the summary? that's the question. If I'm browsing for changes in general do I care what version of a package you changed or whether it was an inb or bb? no. If you do care you can see it in the file list easily enough. Oct 07 15:14:31 or you can always add such things to the "detail" of the message Oct 07 15:14:56 koen: please just do me that one favour. I've been trying to keep up with the huge amount of changes that go in these days but it's very hard... Oct 07 15:15:10 I'm trying that too Oct 07 15:15:22 stuff like the mess pH5 created with bluez-utils just shouldn't happen Oct 07 15:15:47 the more easy I can filter messages in my mailboxes the more patches I can review Oct 07 15:16:02 I don't have the time to go through all of the commits I see everyday for the various projects I work on. But with quick summaries I can chekc what I care/know about fairly easily and get a general overview with the rest. Oct 07 15:16:26 you can filter 'foo: bla' but not 'foo_2.0.bb: bla'? Oct 07 15:16:51 the less info, the easier to filter Oct 07 15:16:53 I *really* want version info in the first line Oct 07 15:17:03 foo: added foo to 2.0 Oct 07 15:17:04 well. I can't visually filter hundreds of foo_2.0.bb: bla Oct 07 15:17:10 exactly Oct 07 15:17:21 which I have to every morning Oct 07 15:17:33 using a filename is the easiest way to do that since you can c/p it from the bottom of the message monotone presents you with Oct 07 15:17:40 words and filenames aren't that different in small doses, but with 100 commits it makes a huge difference Oct 07 15:18:01 we're talking about ease of review, not ease of commit Oct 07 15:18:05 1 person commits Oct 07 15:18:09 100 read Oct 07 15:18:19 the time of the 100 who read is more precious than that of the one dev committing Oct 07 15:18:50 I think we both agree on the essence of the idea Oct 07 15:19:12 would it be ok to make the version mandatory in the short description? Oct 07 15:19:29 "foo: bla in 2.0" ? Oct 07 15:19:54 that would give you the ability to filter and me a version Oct 07 15:20:06 fine with me Oct 07 15:20:29 announce it as mandatory and we should be all set Oct 07 15:20:43 ok Oct 07 15:20:44 fine by me Oct 07 15:20:59 although it should only be mandatory for packages with multiple versions Oct 07 15:21:15 or when you update a version Oct 07 15:22:56 mail sent Oct 07 15:24:22 * CosmicPenguin isn't sure how effective thats going to be, but good luck.. :) Oct 07 15:24:36 The declaring mandatory bit that is Oct 07 15:25:30 OE lacks a clear authority Oct 07 15:25:41 which has good sides and bad Oct 07 15:35:40 re, here is the output I get when building bitbake gpe-image:http://wservices.ch/~lucas/output Oct 07 15:36:23 and my local.conf http://wservices.ch/~lucas/local.conf Oct 07 15:36:33 what the hell is wrong? Oct 07 15:42:35 I'm having exactly de same problem compiling compiling a gpe-image Oct 07 15:42:53 guillermo: ha! Oct 07 15:43:22 guillermo: with the same package? Oct 07 15:43:37 mmm Oct 07 15:43:43 I don't remember right now Oct 07 15:43:45 lucasvo: don't set TARGET_ARCH or TARGET_OS Oct 07 15:43:51 I'm gonna try Oct 07 15:44:01 JustinP: why? Oct 07 15:44:08 lucasvo: do you have the newest bitbake and OE sources? Oct 07 15:44:18 lucasvo: because MACHINE and DISTRO set those. No need for you to. Oct 07 15:45:10 I have no TARGET_{ARCH;OS} but I get the same error Oct 07 15:46:30 what version of bitbake are you using? Oct 07 15:46:34 and have you updated OE? Oct 07 15:46:39 JustinP: my oe is up to date Oct 07 15:47:04 [0:46:57 lucas@wservices <0> monotone.vanille.de]% bitbake --version Oct 07 15:47:04 BitBake Build Tool Core version 1.3.2.1, bitbake version 1.3.2 Oct 07 15:47:15 do I have to be root for compiling? Oct 07 15:47:20 NO Oct 07 15:47:28 ok good Oct 07 15:47:34 DO NOT RUN bitbake AS ROOT Oct 07 15:47:38 EVER Oct 07 15:48:14 phoo Oct 07 15:48:21 I think it's an error in gpe-image.bb Oct 07 15:48:28 * JustinP looks for who to blame Oct 07 15:48:39 JustinP: good.... Oct 07 15:49:13 it could be Oct 07 15:49:29 I wonder why we've never checked the UID and printed a nasty message Oct 07 15:50:12 hmmm Oct 07 15:50:49 JustinP: you got a spitz? Oct 07 15:51:19 have you tried any 2.6 kernel? Oct 07 15:51:48 hmmm....or maybe not... Oct 07 15:51:53 no, I haven't tried the 2.6 kernel Oct 07 15:51:59 too bad Oct 07 15:52:01 I would like sound and CE-RH2 support Oct 07 15:52:13 yeah that'd be nice Oct 07 15:52:25 but I know next to nothing about kernel hacking and don't have the time right now to figure it out Oct 07 15:52:29 but it's not a must for me Oct 07 15:53:25 JustinP: gpe or opie? Oct 07 15:53:51 heh, stupid me. no, it doesn't seem to be an error in gpe0image.bb Oct 07 15:53:58 guillermo: E Oct 07 15:54:28 what's E ? Oct 07 15:55:37 guillermo: http://enlightenment.org/ Oct 07 15:55:42 ooo Oct 07 15:55:46 beautiful Oct 07 15:55:46 e17 Oct 07 15:55:57 but I thoght it was experimental yet Oct 07 15:56:00 is it usable? Oct 07 15:56:04 heh, of course Oct 07 15:56:08 on both counts ;-) Oct 07 15:56:16 awesome Oct 07 15:56:21 technically alpha, but works generally well Oct 07 15:56:31 how could I get an image with that? Oct 07 15:56:32 I maintain the efl/e17 stuff in OE Oct 07 15:56:36 bitbake e-image Oct 07 15:56:43 it's very big, though Oct 07 15:56:48 >90M last I looked Oct 07 15:56:56 e-image-core is ~26M I think... Oct 07 15:57:43 so I guess the picture from the OE from the last OE newsletter was from you Oct 07 15:57:56 JustinP: so do you know what I should do witb my gpe-image? Oct 07 15:58:49 I get : "Cannot find package task-bootstrap" when I try to compile my gpe-image Oct 07 15:58:58 is weird, isn't it? Oct 07 15:59:41 yes Oct 07 15:59:50 I have the same problem Oct 07 16:00:13 lucasvo, guillermo: bitbake task-bootstrap, then bitbake gpe-image Oct 07 16:00:14 JustinP: is E more functional than xfce right now? Oct 07 16:00:36 guillermo: actually I think that was form Mickey's machine, but yeah, I got the image together. Oct 07 16:00:42 guillermo: never used xfce Oct 07 16:00:46 ok Oct 07 16:00:56 guillermo: also be warned, you can't use e-image unless you have a keyboard Oct 07 16:00:58 but isn't it to heavy? Oct 07 16:01:04 it works just fine for me Oct 07 16:01:06 on my spitz Oct 07 16:01:15 what do you mean keyboard? external? Oct 07 16:01:19 no one else has tested on any other machines to my knowledge Oct 07 16:01:23 I have an akita Oct 07 16:01:26 Zaurii have a built-in keyboard Oct 07 16:01:30 then you're fine Oct 07 16:01:33 ah ok Oct 07 16:01:39 you mean the built in keyboard Oct 07 16:01:45 on an ipaq you can't login because entrance doesn't have an on-screen keyboard Oct 07 16:02:12 aha Oct 07 16:02:24 but it could be made to use gpe-login instead... Oct 07 16:02:26 hmmm... Oct 07 16:02:53 I used to just install e over a GPE image Oct 07 16:02:59 but then it wouldn't be so nice Oct 07 16:03:00 :) Oct 07 16:03:21 at least the picture was impressive Oct 07 16:04:46 I tried to bitbake task-bootstrap before gpe-image Oct 07 16:04:50 but it didn't work Oct 07 16:04:56 what happened? Oct 07 16:05:08 for the bitbake task-bootstrap? Oct 07 16:05:22 after task-bootstrap it didn't show any compiling or anything Oct 07 16:06:02 I got 0 attempted packages installed Oct 07 16:06:05 or something like that Oct 07 16:06:20 I'm doing it again to see the exact output Oct 07 16:06:35 NOTE: build 200510080103: completed Oct 07 16:06:35 Build statistics: Oct 07 16:06:35 Attempted builds: 0 Oct 07 16:06:44 this is what it shows Oct 07 16:06:44 exactly that Oct 07 16:07:47 bitbake -cclean task-bootstrap Oct 07 16:07:50 bitbake task-bootstrap Oct 07 16:07:56 or Oct 07 16:07:57 bitbake -i Oct 07 16:08:02 rebuild task-bootstrap Oct 07 16:08:08 -i is nice :-) Oct 07 16:09:18 ok I'm on it Oct 07 16:11:25 ok task-bootstrap worked now Oct 07 16:11:31 let's see gpe-image Oct 07 16:12:43 btw, what other options exists for : TARGET_FPU = "soft" Oct 07 16:13:43 hard? Oct 07 16:13:50 no Oct 07 16:14:10 either "soft" or empty ("") Oct 07 16:14:21 omm Oct 07 16:15:13 reenoo_: what's the main problem for running swing application with OE Oct 07 16:15:31 gnuclasspath or the pdas themselves? Oct 07 16:17:01 ok the old error is gone, but now I get a new one about installing libgpg-error0 Oct 07 16:17:21 is it so hard to compile an image Oct 07 16:17:28 or it's just a misconfig? Oct 07 16:17:28 guillermo: not, it's very simple Oct 07 16:17:40 guillermo: seems you got something corrupted...lost some ipk files or something Oct 07 16:17:51 03justinp 07org.oe.dev * r417b6b1a... 10/: lineak: fix S Oct 07 16:17:54 and how do I clean it? Oct 07 16:17:54 guillermo: I would suggest removing your tmpdir and trying again... Oct 07 16:17:58 ok Oct 07 16:17:59 thx Oct 07 16:18:53 mickey|resting: ewi and vanille are supposed to sync, right? I'm pretty sure I pushed some changes to ewi earlier but they weren't on vanille just now... Oct 07 16:23:38 hey people Oct 07 16:23:44 head out to Hamburg tommorrow Oct 07 16:25:24 I can't :-( Oct 07 16:26:07 what are they so many developers from germany in oe? Oct 07 16:30:40 in EU, yeah Oct 07 16:31:00 guillermo: both Oct 07 16:31:04 s/head/I head Oct 07 16:31:48 reenoo_: so right now imposible to be used properly, oder? Oct 07 16:32:08 guillermo: depends on the application Oct 07 16:32:37 something I want to create Oct 07 16:33:01 I don't find clear documentation about what is working and what is not Oct 07 16:33:28 but using awt it's painfull for anything minimally complex Oct 07 16:33:59 other than $application works to $extent there is no documentation Oct 07 16:34:48 too bad Oct 07 16:35:16 last year I worked with sablevm and ipaqs 3600 Oct 07 16:35:27 and on pdas you may run into additional trouble due to the X server being slightly different (bit-depth, no XKB, etc.) and matchbox being fairly strict (if you use matchbox) Oct 07 16:35:47 and I thought with akitas would be different Oct 07 16:36:10 they are using pretty much the same environment Oct 07 16:36:10 what about E ? Oct 07 16:36:20 would it be a better platform than gpe? Oct 07 16:36:20 it's quite possible to run a full X on these systems Oct 07 16:36:24 with xkb, etc Oct 07 16:36:37 java and e don't get along IIRC Oct 07 16:36:41 not e17 at least Oct 07 16:36:53 java does stupid things...hacks around the WM Oct 07 16:37:14 well. E isn't exactly lightweight Oct 07 16:37:26 so, what is the best option for java ? Oct 07 16:37:33 someone was complaining that their Java apps ended up with no window content in e17 Oct 07 16:37:55 reenoo_: not quite true. The default themes arent lightweight. E itself isn't that bad. Just make a minimal theme and you'd be fine. Oct 07 16:38:10 does OE's python contain autoconf stuff to get the includedir to Python.h ? AM_PATH_PYTHON doesn't include that, as far as i can see, and all autoconf macros i found were not suitable for cross compiling. Oct 07 16:38:17 guillermo: GPE is a palmtop environment Oct 07 16:38:20 guillermo: e is e Oct 07 16:38:36 yeah I know Oct 07 16:38:49 guillermo: GPE is also far more mature Oct 07 16:38:51 JustinP: note that we are talking about _interpreter only_ VMs and a highly inefficient toolkit (swing) here Oct 07 16:38:59 but what is e using? X, fb or what? Oct 07 16:39:00 guillermo: e17 is done well, but it is nowhere near stable Oct 07 16:39:18 JustinP: every single cpu cycle counts there... Oct 07 16:39:45 guillermo: X, e can't run on FB, etc. Only X. the *libs* can run on FB but e (wm) can't Oct 07 16:39:54 reenoo_: sure Oct 07 16:40:50 so, the best option for java in my zaurus would be a full featured X, but how? Oct 07 16:41:04 no Oct 07 16:41:22 fix gnu classpath to not make bad assumptions about the x server or the wm Oct 07 16:42:18 also. it may just work by now Oct 07 16:43:47 I haven't tried any swing apps recently. screenshots of current cvs look promising (on a desktop though) Oct 07 16:43:58 see planet.classpath.org Oct 07 16:44:21 <_guillermo> thanks Oct 07 16:45:07 well - I recompiled from scratch using gcc=3.4.4, and I'm still getting a 'segfault' when trying to run atftpd. :-( Oct 07 16:53:16 Jeu Oct 07 16:53:18 hey* Oct 07 16:53:35 Does OE's cross compiler define __linux? Oct 07 16:57:13 <_guillermo> what about xfce and java, good idea or not? Oct 07 17:17:48 Hrm Oct 07 17:17:52 is ARM little or big endien? Oct 07 17:24:44 arm is le Oct 07 17:24:49 hah! Oct 07 17:24:53 armeb would be be Oct 07 17:25:01 LSB is first then right? Oct 07 17:26:03 http://en.wikipedia.org/wiki/Endianness Oct 07 17:26:59 thanks Oct 07 17:44:05 'night all Oct 07 18:22:51 is there a text-to-speech engine package for OE? Oct 07 18:23:02 like festival Oct 07 18:27:11 is it in packages? Oct 07 18:27:34 the only decent ... er working speach systems i know of are festival and sphinx Oct 07 18:28:27 I've heard of Cepestral too, but neither of them are in the 'packages' directory Oct 07 18:28:57 doesn't mean it doesn't exist though ;) Oct 07 18:34:05 usually it does Oct 07 18:40:55 now taht i think of it you could check the broken repo Oct 07 18:53:20 hey all =) Oct 07 18:53:30 anyone using oe on osx these days? =) Oct 07 18:53:43 no Oct 07 18:54:09 not a soul? Oct 07 18:54:22 nope Oct 07 18:54:23 spose I'll have to give it a shot Oct 07 18:54:32 not sure it even works on BSD yet Oct 07 18:54:43 someone was working on it Oct 07 18:57:32 spose I'll have to file bugs Oct 07 18:57:34 ;) Oct 07 19:16:51 hey Oct 07 19:16:59 does anything provide gprof for the target? Oct 07 19:22:25 is that what binutils-cross-sdk is for? Oct 07 19:27:11 emte: hey - found festival! it's named 'flite' Oct 07 19:35:31 Is there a way to run gprof on profiling output compiled on arm? Oct 07 19:37:25 yes Oct 07 19:37:53 compile -pg, run an arm gprof, it should just work. Oct 07 19:40:02 How do you build an arm gprof with OE/ Oct 07 19:40:13 I don't have an arm gprof thats the problem :\ Oct 07 19:40:26 configure --target=arm Oct 07 19:40:31 sorry, --host=arm Oct 07 19:40:50 Oops, no, it's --target in this case. Oct 07 19:41:48 --build=(unnecessary, what you are building it on), --host=(what it will run on), --target=(where the gprof output comes from) Oct 07 19:43:54 hrm Oct 07 19:44:37 Ill try it Oct 07 19:45:10 so build binutils with --target=arm? Oct 07 19:47:51 dfoz: build gprof with those options. It will probably need some other stuff too. Oct 07 19:48:43 Then you have a gprof which should be able to both read arm elf for the symbol info and read the gprof.out Oct 07 19:49:52 I'm so lost. i downloaded binutils off GCC's website Oct 07 19:49:58 ./configure --target=arm Oct 07 19:49:59 make Oct 07 19:50:02 that won't work? Oct 07 19:50:47 It should do if gprof is in binutils... I'm not sure 'cause the OE build doesn't generate it (I believe) Oct 07 19:51:47 bitbake -c unpack -b binutils_cvs.bb Oct 07 19:52:00 then i'll edit configure.in and make it build gprof Oct 07 19:52:11 and then ill bitbake again and hopefully that will work maybe? :P Oct 07 19:53:14 If you downloaded binutils from the gnu then that's probably a better bet as a starting point. Oct 07 19:53:51 The level of extra indirection building things via OE is probably just going to make things difficult in this case... Oct 07 19:55:39 it complained Oct 07 19:55:47 *** BFD does not support target arm-unknown-none. Oct 07 19:55:51 if i tried right from GNU Oct 07 19:56:17 ./configure --target=arm and then make Oct 07 19:57:04 Try arm-linux - you can work it out from the OE binutils stuff (look in temp/run.do_config* or temp/log.do_config*) Oct 07 19:57:38 Pesky thing apparently thinks it needs to know what the OS is as well... Oct 07 19:57:47 night Oct 07 19:59:23 ;\ Oct 07 19:59:25 Ill try it Oct 07 20:04:35 it's going Oct 07 20:04:37 lets see if it works Oct 07 20:09:53 it still dos it Oct 07 20:09:55 does it* Oct 07 20:10:07 gprof ~/gmon.out Oct 07 20:10:07 gprof: /home/diefool/gmon.out: not in a.out format Oct 07 20:14:26 dfoz: IIRC it expects the name of the executable as an argument Oct 07 20:15:11 gprof.out is just a big pile of data Oct 07 20:17:12 oh. crap. Oct 07 20:19:19 right Oct 07 20:21:09 ./gprof ~/snes9x ~/gmon.out Oct 07 20:21:09 ./gprof: gmon.out file is missing call-graph data Oct 07 20:21:17 :( Oct 07 20:27:46 I just want a native arm gprof Oct 07 20:27:53 is that so much to ask :( haha Oct 07 20:43:32 got it :0 Oct 07 20:43:51 Each sample counts as 0.01 seconds. Oct 07 20:43:51 % cumulative self self total Oct 07 20:43:51 time seconds seconds calls Ts/call Ts/call name Oct 07 20:43:51 50.87 59.64 59.64 S9xGetWord(unsigned int) Oct 07 20:43:51 7.82 68.81 9.17 S9xSetWord(unsigned short, unsigned int) Oct 07 20:43:52 4.53 74.12 5.31 Op8DM1() Oct 07 20:43:54 2.86 77.47 3.36 OpCEM0() Oct 07 20:43:56 2.77 80.72 3.25 Op8DM0() Oct 07 20:43:58 2.44 83.58 2.86 OpADM1() Oct 07 20:44:00 OOPS sorry! Oct 07 20:44:04 i was supposed to paste that in aan IM to someone Oct 07 22:58:49 hmm, monotone.vanille.de doesn't accept pulls? Oct 07 22:58:57 monotone: connecting to monotone.vanille.de Oct 07 22:58:59 monotone: network error: failed to connect: Connection refused Oct 07 22:59:55 heh, thats no good Oct 07 23:01:20 koen is probably drooling on the server again Oct 07 23:01:29 and it shorted out Oct 07 23:06:34 any way to speed up the initial pull? Oct 07 23:06:49 use the snapshot Oct 07 23:07:07 not sure how, seeing as I don't have any files except the db I inited Oct 07 23:07:30 http://ewi546.ewi.utwente.nl/OE/OE.db.bz2 is the one I'm leeching Oct 07 23:08:03 ah Oct 07 23:08:14 then short of buying a super computer ... no Oct 07 23:09:55 snapshot helps Oct 07 23:10:49 ah, I got it =) Oct 07 23:10:58 now just waiting on verify Oct 07 23:32:56 emte: by the way, http://oe.handhelds.org/cgi-bin/moin.cgi/BuildOnOSX Oct 07 23:34:20 yeah i remember someone cursing at quilt Oct 07 23:34:37 hehe Oct 07 23:34:52 emte: is updating to OE via USB possible yet? Oct 07 23:35:04 don't have a CF or SD card here =P Oct 07 23:35:57 for specific chips ... Oct 07 23:36:06 it works in windows Oct 07 23:36:06 taht is bootloader dependant Oct 07 23:36:12 with the official Oct 07 23:36:26 will the official load an OE image? Oct 07 23:36:27 SL-5600 Oct 07 23:36:34 emte: I think it might, I have not tried Oct 07 23:36:54 pulling out the windows laptop currently Oct 07 23:37:21 ah, college, two laptops. Oct 07 23:38:48 lol Oct 07 23:39:47 gonna buy a 1G SD card tomorrow anyways Oct 07 23:39:49 but still Oct 07 23:44:12 coming in at a peasly 1MB/s Oct 07 23:44:13 heh Oct 07 23:44:31 myzaurus.com = slow Oct 07 23:44:58 lol Oct 07 23:47:12 sharp.com gives me 11MB/s =P Oct 07 23:47:30 er sharpusa Oct 07 23:47:32 whatevre it is Oct 07 23:48:23 the updater.sh is different Oct 07 23:48:24 poop Oct 07 23:51:29 it runs update.sh before it has the rom copied this way Oct 08 00:54:48 Is that a key sequence that lets you kill opie and escape to a command line? Oct 08 01:33:28 <_guillermo> I have a problem compiling zaurus kernels Oct 08 01:33:50 <_guillermo> I always get : arm-linux-gcc-2.95: No such file or directoryç Oct 08 01:34:13 <_guillermo> but the file is correctly install under /usr/local/..... Oct 08 01:34:19 <_guillermo> and correctly placed in PATH Oct 08 01:34:45 <_guillermo> I can execute it with no problem, but bitbake seems to have problems finding it Oct 08 02:15:39 03jbowler 07org.oe.dev * rbc8ebf4b... 10/: Oct 08 02:15:39 ixp-osal_2.0.bb ixp4xx-csr_2.0.bb ixp400-eth_1.4.bb: fixed LE implementation Oct 08 02:15:39 This checkin includes a patch to add working le support in these modules Oct 08 02:15:43 03jbowler 07org.oe.dev * r140ea215... 10/: (log message trimmed) Oct 08 02:15:43 nslu2-kernel_2.6.14-rc3.bb: better fix for LE flash access Oct 08 02:15:43 patch "packages/linux/nslu2-kernel/2.6.14/10-ixp4xx-copy-from.patch" Oct 08 02:15:43 Change to use the cfi16 APIs to byte swap the 16 bit values read from Oct 08 02:15:43 the flash as required. Oct 08 02:15:46 patch "packages/linux/nslu2-kernel/2.6.14/defconfig" Oct 08 02:15:48 patch "packages/linux/openslug-kernel-2.6.14-rc3/defconfig" Oct 08 02:40:57 http://pastebin.com/386948 < whats wrong? **** ENDING LOGGING AT Sat Oct 08 02:59:56 2005