**** BEGIN LOGGING AT Wed Nov 25 02:59:57 2009 **** ENDING LOGGING AT Wed Nov 25 03:35:07 2009 **** BEGIN LOGGING AT Wed Nov 25 03:35:39 2009 Nov 25 12:18:59 ogra: http://launchpadlibrarian.net/35539315/buildlog_ubuntu-lucid-armel.xulrunner-1.9.1_1.9.1.5%2Bnobinonly-0ubuntu1_FAILEDTOBUILD.txt.gz Nov 25 12:19:02 IT ;) Nov 25 12:19:24 clearly Nov 25 12:19:24 yep Nov 25 12:20:05 i wonder if we could just start the transition away from FF already and drop all xul stuff from armel to get the image building Nov 25 12:20:18 hmm ... unlikely ;) Nov 25 12:20:26 i can quickly upload xul Nov 25 12:20:32 for A1 the only target is "make it boot" Nov 25 12:20:38 i just want to know how to best test for target being v7 Nov 25 12:20:47 i could probably check whether we are on lucid Nov 25 12:20:53 + armel Nov 25 12:20:56 right, but you dont know for which files you have to add the IT hack in rules Nov 25 12:21:08 ogra: everywhere ;) Nov 25 12:21:21 i mean ... there are just a few files that have assembler Nov 25 12:21:29 oh ... hmm, indeed Nov 25 12:21:31 i would have thought it doesnt hurt to add it Nov 25 12:21:46 obviously hiding stuff Nov 25 12:21:52 no, it doesnt, and eventually the toolchain will use the IT flag by default anyway Nov 25 12:21:55 maybe i should file an upstream bug so it doesnt get lost ;) Nov 25 12:21:58 right Nov 25 12:22:10 so just on top level rules: if lucid + armel -> use it Nov 25 12:22:14 which makes it only temporary anyway Nov 25 12:22:25 right ... Nov 25 12:22:32 ogra: ok what exactly was that flag? Nov 25 12:22:47 CFLAGS+=-Wa,-mimplicit-it=thumb Nov 25 12:22:56 thx Nov 25 12:23:04 wonder if i should also add that to CXXFLAGS Nov 25 12:23:17 probably better-safe-than-sorry ;) Nov 25 12:23:24 well, given that xul dies in a cpp file ... Nov 25 12:24:05 i wonder when doko will be back ... so the defaults change Nov 25 12:24:28 doko probably takes extensive holiday ;) Nov 25 12:24:34 as usual after UDS Nov 25 12:24:36 well deserved Nov 25 12:24:52 yep Nov 25 12:25:06 but if its to extensive we should actually start a list with packages we added that workaround to :) Nov 25 12:34:14 dyfet``: did you merge the ssd spec in the other blueprint? Nov 25 12:34:28 somehow the url isnt valid anymore ;) ... so unless you say you didnt do anything i will assume its now properly merged? Nov 25 12:35:07 ogra: you think that option would hurt on compilers that dont support it? Nov 25 12:35:22 e.g. can you try on a karmic compiler? Nov 25 12:36:09 hmm it fails becaues of unrecognized option Nov 25 12:36:09 too bad Nov 25 12:58:54 yeah Nov 25 13:00:17 it might work on karmic if you add the other options too ... -march=cortex-a8 at least Nov 25 13:00:55 but then it might produce incompatible binaries, not sure what it implies for the binaries Nov 25 14:47:50 Can anyone suggest the best way for me to suggest resolutions for build failures? Should I raise a bug, or something else? Nov 25 14:50:32 apex and ffmpeg should be built with -marm Nov 25 14:51:00 (apex is a bootloader, with low-level hacks which won't work with Thumb-2 without porting; ffmpeg contains a lot of optimised assembler) Nov 25 15:07:10 dmart, apex is used only on nslu2 ... so i didnt bother to fix it yet Nov 25 15:07:57 i'll add the -marm and i think it also needs a -march option to build for v5 Nov 25 15:11:41 dmart, the proper way is indeed to file a bug, or if you can catch someone here to add a quick fix just use the shortcut through IRC Nov 25 15:53:51 ogra: OK Nov 25 15:55:23 Did we agree a standard tag for the v7/T2 migration issues? Nov 25 15:58:10 I will start tagging with "armv7", though we can change that later if needed. Nov 25 18:47:04 hi all, Nov 25 18:47:04 I open Ubuntu Q&A website - http://www.ubuntu-tr.net **** ENDING LOGGING AT Thu Nov 26 02:59:56 2009