**** BEGIN LOGGING AT Sun Nov 08 02:59:56 2009 Nov 08 04:29:09 hey, anyone got experience with the nslu2 and debian? Nov 08 05:17:57 03bzhou * r10864 10optware/trunk/ (4 files in 3 dirs): mysql5: 5.0.45 -> 5.0.86 Nov 08 05:17:59 03bzhou * r10865 10optware/trunk/Makefile: mysql5: promoted Nov 08 08:00:54 03jomaster * r10866 10optware/trunk/platforms/toolchain-fsg3v4.mk: toolchain-fsg3v4: route.h-conflict workaround Nov 08 08:01:35 03jomaster * r10867 10optware/trunk/platforms/packages-fsg3v4.mk: packages-fsg3v4: enable busybox Nov 08 09:09:22 03jomaster * r10868 10optware/trunk/make/libc-dev.mk: libc-dev: replace "$(TARGET_USRLIBDIR)" with "$(LIBC-DEV_USRLIBDIR)" and add little workaround for dead header-links Nov 08 09:12:04 @eno: can you have a look on libc-dev build, please? i did two little changes and test it for some targets. Nov 08 09:23:48 03jomaster * r10869 10optware/trunk/make/gconv-modules.mk: gconv-modules: install "$(GCONV_MODULES_LIB_DIR)" and set paths for vt4 Nov 08 09:31:00 @eno: there are packages at the stable optware feeds which depends on unavailable packages. (how do a package becomes part of stable feeds?) Nov 08 09:36:00 03jomaster * r10870 10optware/trunk/platforms/packages-vt4.mk: packages-vt4: enable gconv-modules Nov 08 16:08:34 jomaster: stable symlinks to unstable right now Nov 08 16:08:45 so the feed should not be any different Nov 08 16:10:47 i don't understand. i mean that there are packages available for vt4 which depends on packages they do not exist. Nov 08 16:23:08 jomaster: that might be true, vt4 just have way too many broken packages to sort out Nov 08 16:24:08 you either fix the broken package, or demote package that depends on it Nov 08 16:25:17 with that many broken packages, I always incline to drop the support for vt4 feed Nov 08 16:29:00 ok Nov 08 16:30:08 at this point, feel free to maintain the vt4 feed Nov 08 16:31:14 i'll take vt4 out of the rotation if it's broken too much, and put it back in if it's fixed Nov 08 16:31:18 i don't have a vt4. so can't test anything Nov 08 16:31:35 i don't have one either Nov 08 16:31:42 i know Nov 08 16:32:06 perhaps i find some testers on openfsg Nov 08 16:32:16 some1 with a vt4 device should step up Nov 08 16:32:57 i mean, my time is as valuable as other's Nov 08 16:33:03 this feed is a time sucker Nov 08 16:33:20 on my machine there is no error on libcurl for vt4 Nov 08 16:34:18 yeah, after certain cleanup of the staging/ it's working Nov 08 16:34:29 so vt4 is back in build rotation right now Nov 08 16:34:30 you are right. we have to use better toolchains for fsg, dt and vt. i will do so for new firmware. Nov 08 16:35:34 but as you might notice, even for mysql5, it cannot build 5.0.86 Nov 08 16:35:52 have to special case vt4, set to 5.0.45 Nov 08 16:36:06 but until now i don't have the knowledge to think about toolchains. Nov 08 16:36:10 i saw it Nov 08 16:36:48 and without libstdc++ so, 5.1 is impossible Nov 08 16:37:04 while on most other platforms, 5.1 compiles just fine Nov 08 16:37:41 is it possible to replace toolchain and get compatible apps? Nov 08 16:38:05 yeah, if the new toolchain is built properly Nov 08 16:39:27 so like dt2 use the apps from cs05q3armel. Nov 08 16:40:03 on with details i should have a look to build a compatible toolchain? Nov 08 16:40:08 which Nov 08 16:41:08 you usually need to stick with the same glibc version and same gcc/libstdc++ version Nov 08 16:42:56 for older gcc3 toolchain, look at using crosstool Nov 08 16:43:08 for newer toolchain, use crosstool-ng Nov 08 16:43:55 ok, i will try it next week Nov 08 16:45:04 make sure fp option and ABI are consistent with firmware Nov 08 16:46:50 fp? ABI? Nov 08 16:50:15 floating point, app binary interface Nov 08 16:50:28 OABI or EABI Nov 08 16:55:13 ok Nov 08 17:36:52 03bzhou * r10871 10optware/trunk/ (5 files in 2 dirs): mysql5: fixed build for openwrt-ixp4xx/slugos4/vt4 Nov 08 18:47:08 @eno: dump question: why there is a dt2-target, when cs05q3armel use the exactly same toolchain? Nov 08 19:16:35 su **** ENDING LOGGING AT Mon Nov 09 02:59:56 2009