**** BEGIN LOGGING AT Sun Feb 24 02:59:58 2013 Feb 24 10:59:58 hmm, i'm seeing some confusion over which libjpeg to use. libcups, for example. if i add +libjpeg-turbo to DEPENDS, no complaints. Feb 24 11:00:10 +libjpeg and it barfs Feb 24 11:31:00 also, it looks like libgmp's gmp.h isn't getting installed Feb 24 11:31:12 so that other packages can find it Feb 24 11:32:20 russell--: better silently applied then siltently dropped/ignored ;) Feb 24 11:32:22 *than Feb 24 11:33:37 KanjiMonster: indeed, /me happy! Feb 24 11:34:02 juhosg seems to be a bit behind; the patches aren't marked as applied in patchwork either Feb 24 11:51:32 hmm. i see in in staging_dir now, but for some reason php4's gmp modules isn't finding it or something. Feb 24 11:56:43 russell--: check if they expect it in e.g. ../gmp/gmp.h and it's installed as ../gmp.h or the other way round Feb 24 12:00:42 ah, i think i've discovered the problem, it's finding the include alright, but it's referencing something that's not (or no longer) there Feb 24 12:01:02 /aux/src/openwrt/build_dir/target-mips_uClibc-0.9.33.2/php-4.4.9/ext/gmp/gmp.c:1162:69: error: '__GMP_BITS_PER_MP_LIMB' undeclared (first use in this function) Feb 24 12:03:10 looks like it just needs changing to GMP_LIMB_BITS Feb 24 12:03:25 to cope with modern gmp Feb 24 12:15:24 yes, that looks like it's fixing it Feb 24 12:40:36 did anyone ever settle on a solution for the libjpeg problem (conflicting with libjpeg-turbo)? Feb 24 12:41:15 cups, php4 and php5 all complaining about it Feb 24 12:42:36 unfortunately there is no real conflict/provide handling in openwrt yet Feb 24 12:44:10 what's the pragmatic solution? Feb 24 12:44:26 disable one of them, i suppose Feb 24 12:44:33 yes Feb 24 12:44:46 see the libelf vs libelf1 problem Feb 24 12:45:27 but the dependencies all have to be adjusted too Feb 24 12:45:58 yes, the current way of fixing it is to only use one Feb 24 12:46:03 and remove the other Feb 24 12:46:31 * russell-- has already forgotten if the turbo version falls back in the absense of acceleration hardware Feb 24 12:46:44 or whatever it does Feb 24 12:46:57 uses available instructions to speed things up Feb 24 12:48:39 http://libjpeg-turbo.virtualgl.org/ Feb 24 12:49:08 "On other types of systems, libjpeg-turbo can still outperform libjpeg by a significant amount, by virtue of its highly-optimized Huffman coding routines. In many cases, the performance of libjpeg-turbo rivals that of proprietary high-speed JPEG codecs." Feb 24 12:49:52 * russell-- is in a weak position to set openwrt policy on this one Feb 24 12:50:05 (or any one) Feb 24 12:50:40 * russell-- considers maintain a local branch that uses libjpeg-turbo Feb 24 12:51:06 so that stuff compiles Feb 24 12:58:37 of course, what compiles depends on which version was built/installed last, package/jpeg or package/libjpeg-turbo Feb 24 13:19:22 [LUA] 'local num = line:match("Value=(%d+")' is the (%d+) works fine with number like '27.71'? Feb 24 13:19:38 'local num = line:match("Value=(%d+"))' Feb 24 13:58:33 maybe ([%d%.]+). testing... Feb 24 16:40:51 can someone put compat-wireless-2013-02-22.tar.bz2 on downloads.openwrt.org as mirror2.openwrt.org is down atm? Feb 24 17:55:26 hauke r35764 trunk/package/broadcom-diag/src/diag.c * broadcom-diag: use new in kernel board detection code Feb 24 17:56:17 hauke r35765 trunk/package/broadcom-diag/src/diag.c * broadcom-diag: Detection of WNDR3400v1 Feb 24 17:57:45 hauke r35766 trunk/package/broadcom-diag/src/diag.c * broadcom-diag: add support for e3200 v1 Feb 24 18:01:23 hauke r35767 trunk/package/broadcom-wl/patches/913-avoid-dbe-on-ifs_ctl-readw-hack.patch * broadcom-wl: hack to work around a bug in the PCIe controller Feb 24 18:02:18 <3JTAANC26> build #180 of rb532 is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/180 Feb 24 18:05:50 <3JTAANC26> build #180 of ppc44x is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/180 Feb 24 18:08:20 juhosg r35768 trunk/target/linux/ adm5120/files/drivers/net/adm5120sw.c adm5120/files/drivers/ata/pata_rb153_cf.c adm5120/files/drivers/usb/host/adm5120-drv.c * adm5120: remove __dev{init,exit} annotations Feb 24 18:08:31 juhosg r35769 trunk/target/linux/ (18 files in 5 dirs) * adm5120: add support for 3.8 Feb 24 18:10:59 hauke r35770 trunk/target/linux/brcm47xx/patches-3.6/260-MIPS-BCM47XX-add-board-detection.patch * brcm47xx: use same board detection code as kernel 3.8 Feb 24 18:42:36 <3JTAANC26> build #185 of uml is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/185 Feb 24 19:25:07 <3JTAANC26> build #182 of x86 is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/x86/builds/182 Feb 24 19:44:01 <3JTAANC26> build #170 of sibyte is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/170 Feb 24 19:46:51 updated openwrt/upstream, https://home.comcast.net/~sdwalker/uscan/index.html Feb 24 21:46:30 swalker: somehow googlecharts is broken; e.g. the pie chart coloring for "Maintainer breakdown for the 984 packages in packages" does not match the actual numbers Feb 24 22:17:34 KanjiMonster: yes, I'm not sure what the issue is for that one chart Feb 24 22:22:22 swalker: it looks like it's treating values > 100 as 100, so 107 vs 877 becomes 100 vs 100 - same problem on the chart above Feb 24 22:26:23 swalker: it seems to accept floats, too, so putting the percentages instead of the raw numbers into the chd field should fix it Feb 24 23:19:32 KanjiMonster: both packages charts are fixed, thanks Feb 25 01:57:40 <3JTAANC26> build #176 of avr32 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/176 **** ENDING LOGGING AT Mon Feb 25 02:59:57 2013