**** BEGIN LOGGING AT Fri Feb 13 12:02:41 2009 Feb 13 13:11:24 jo Feb 13 13:14:29 anyone here building something for x86 ? Feb 13 13:14:39 hm Feb 13 13:14:40 nope Feb 13 13:14:48 whats the problem? Feb 13 13:15:21 http://pastebin.com/m7f460dc5 Feb 13 13:15:31 bitbake fails with a "too many open files" error Feb 13 13:15:38 uh Feb 13 13:15:54 i tried different kernel versions.. didn't matter Feb 13 13:16:03 2.6.20 / 2.6.25 / 2.6.28 Feb 13 13:16:05 this comes from your system Feb 13 13:16:16 what are you running? Feb 13 13:16:31 hardly... my build for cmx270 works just fine Feb 13 13:16:37 woglinde: I have it on my debian too Feb 13 13:16:38 <_diego__> gilligan_, what does ulimit say? Feb 13 13:16:51 hrw hm oh Feb 13 13:16:56 I never saw this Feb 13 13:17:12 hrw was this happening the last days? Feb 13 13:17:12 _diego__: here it says "unlimited" Feb 13 13:17:24 ulimit says unlimited as it should .. Feb 13 13:17:27 woglinde: ~month iirc Feb 13 13:17:31 hm Feb 13 13:17:33 woglinde: did not built it before Feb 13 13:17:38 here it build fine Feb 13 13:17:43 on ubuntu too Feb 13 13:17:53 /proc/sys/fs/file-max is set to 254434 ( I increased it ) Feb 13 13:18:21 371435 Feb 13 13:18:24 I don't quite believe it's really a to many open files error Feb 13 13:18:26 that here Feb 13 13:19:00 <_diego__> which command file with this error? Feb 13 13:20:13 gilligan_ no alone, but when you have other programms running which had many files opentoo Feb 13 13:20:14 gilligan_: I had the same problem with x86, use i586-generic instead Feb 13 13:20:17 the sum it coulb be Feb 13 13:22:35 gilligan_: /proc/sys/fs/file-max won't change this issue Feb 13 13:24:51 i'll use i586 then Feb 13 13:24:52 thanks Feb 13 13:25:07 i'm using i586 right now and it is almost working Feb 13 13:31:06 03woglinde 07org.openembedded.dev * rac34125ba1 10openembedded.git/packages/alsa/files/alsa-utils-autoconf-strl-funcs.patch: alsa-utils: QA-fix for m4 syntax Feb 13 13:31:22 * mckoan is wondering what could be broken in x86 machine Feb 13 13:34:29 that would be interesting to know Feb 13 13:35:40 hm x86.conf Feb 13 13:35:43 has TARGET_ARCH = "i486" Feb 13 13:40:55 that should be correct Feb 13 13:41:40 x86 is for devices that are not i586 Feb 13 13:42:05 vortex86sx for example Feb 13 13:42:23 hm the tunefile for 586 is inccorect Feb 13 13:42:33 it hast i386, but nptl cannt work with i386 Feb 13 13:47:12 hi mickey|fsomeet Feb 13 13:50:21 he pb Feb 13 14:00:11 hi woglinde Feb 13 14:11:33 woglinde, hi Feb 13 14:12:06 hi gnutoo Feb 13 14:13:47 woglinde, I've a bug: http://rafb.net/p/dyH2Xh14.html Feb 13 14:14:14 where? Feb 13 14:14:48 woglinde, did you look at the pastebin? Feb 13 14:14:52 yes Feb 13 14:14:57 its config for bug device Feb 13 14:15:05 ah lol Feb 13 14:15:05 woglinde, yes I've got one Feb 13 14:15:05 yes Feb 13 14:15:06 sorry Feb 13 14:15:10 lol Feb 13 14:15:38 30 seconds later I thought how misundetstanding it would be to have the bug device Feb 13 14:15:40 i was looking at the pastebin and wondering how that could have been related to the report of a bug Feb 13 14:15:42 and talk about it Feb 13 14:15:52 time to get a coffee Feb 13 14:16:08 I won it at fosdem Feb 13 14:16:15 gnutoo we have one too at work Feb 13 14:16:17 he cool Feb 13 14:16:24 did you meat ken? Feb 13 14:16:36 woglinde, no but I talked to him on irc Feb 13 14:16:44 woglinde, I won it with the fosdem lottery Feb 13 14:16:52 ah Feb 13 14:16:56 woglinde: one BUG? Feb 13 14:17:07 hrw one bug device Feb 13 14:17:19 woglinde: you have serial cable? Feb 13 14:17:41 hrw unfornatly robert took it at home as the most other devices too Feb 13 14:17:58 so cant anwser you Feb 13 14:18:04 * hrw waits for wifi module Feb 13 14:19:07 hrw, what do you think Feb 13 14:19:08 oops Feb 13 14:19:32 hrw, what is the simpliest way to get your serial cable? (if I pay a little more money I don't care) Feb 13 14:20:27 Gnutoo: mail me address Feb 13 14:20:38 hrw, ok thanks I will Feb 13 14:20:52 I think that I will handle some mails via post office on monday so will send also connectors Feb 13 14:21:55 ok Feb 13 14:24:33 03acid-burn 07org.openembedded.dreambox * r4702c80b54 10openembedded.git/packages/autofs/ (3 files in 2 dirs): small cleanup and bump revision Feb 13 14:24:34 03acid-burn 07org.openembedded.dreambox * re25bdc2b68 10openembedded.git/packages/samba/ (files/opendreambox/smb.conf samba_3.0.24.bb): make new samba happy, bump revision Feb 13 14:24:35 03Mladen Horvat 07org.openembedded.dreambox * rdcb0a7056a 10openembedded.git/: Merge branch 'org.openembedded.dreambox' of git://git.openembedded.net/openembedded into org.openembedded.dreambox Feb 13 14:24:35 03Mladen Horvat 07org.openembedded.dreambox * r38c3b8beb7 10openembedded.git/: Merge branch 'org.openembedded.dreambox' of git://git.openembedded.net/openembedded into org.openembedded.dreambox Feb 13 14:24:38 03Mladen Horvat 07org.openembedded.dreambox * r74f552fd9c 10openembedded.git/: Merge branch 'org.openembedded.dreambox' of git://git.openembedded.net/openembedded into org.openembedded.dreambox Feb 13 14:24:41 03Mladen Horvat 07org.openembedded.dreambox * r97ccc30b61 10openembedded.git/: Merge branch 'org.openembedded.dreambox' of git://git.openembedded.net/openembedded into org.openembedded.dreambox Feb 13 14:24:44 03Mladen Horvat 07org.openembedded.dreambox * rdef34f8c6f 10openembedded.git/: Merge branch 'org.openembedded.dreambox' of git://git.openembedded.net/openembedded into org.openembedded.dreambox Feb 13 14:24:47 03acid-burn 07org.openembedded.dreambox * ra4205dfee7 10openembedded.git/packages/initscripts/ (initscripts-1.0/umountnfs.sh initscripts-opendreambox_1.0.bb): properly unmount network filesystems on shutdown Feb 13 14:24:53 03acid-burn 07org.openembedded.dreambox * r41e5ebc33a 10openembedded.git/packages/ (enigma2/enigma2.bb images/dreambox-image.bb): remove nonexistent packages Feb 13 14:24:56 03Felix Domke 07org.openembedded.dreambox * r2dab9513c4 10openembedded.git/packages/enigma2/ (enigma2-plugins.bb enigma2-skins.bb enigma2.bb): switch to enigma2-2.6 Feb 13 14:39:15 * mckoan is hearing talinkg Polish at bust stop in front of office window :-) Feb 13 14:41:02 mckoan is learning polish by osmosis >:) Feb 13 14:41:42 gremlin[it]: I couldn't do it in a different way :-D Feb 13 14:42:47 * gremlin[it] whant learn japanese by osmosis ... Feb 13 14:42:58 hehe Feb 13 14:43:05 for all the anime stuff? Feb 13 14:44:20 yes also ... but not only ... i study japanese years ago ... but without practice i think i forgot about all Feb 13 14:44:35 I am considering changing checksums.ini format Feb 13 14:46:57 why do we save these checksums in this global file anyway? Feb 13 14:47:15 to protect from trojans Feb 13 14:47:31 it would be simpler to merge custom projects with OE upstream then Feb 13 14:47:35 mwester: that's the reason for the checksums, but not really a good reason for having them in a global file Feb 13 14:47:43 if checksums are in package files Feb 13 14:48:12 or in digest files (like in Gentoo) Feb 13 14:48:18 yeah, indeed. I consider checksums.ini to be my enemy. Feb 13 14:48:32 freebsd ports store checksums in distinfo file inside port directory Feb 13 14:48:33 lol Feb 13 14:48:36 bitbake does actually support checksums in SRC_URI already, it just happens that oe doesn't use that feature. Feb 13 14:48:55 pb some recipe used it Feb 13 14:48:55 pb_: or we stopped using them rather Feb 13 14:49:01 hrw yes Feb 13 14:49:03 we can store all package-based checksums inside packages//checksums Feb 13 14:49:03 every time I pull upstream stuff I need to resolve the conflict. It's not a pleasure to me. Feb 13 14:49:04 dont know why Feb 13 14:49:11 maybe koen knows or mickeyl Feb 13 14:49:49 or better store checksum inside *.bb file :) Feb 13 14:50:10 Jay7: that's what seems to be the right thing (we don't want to see hundreds of digest files in our subdirs) Feb 13 14:50:13 there is an other problem with checksum.ini Feb 13 14:50:31 ftp.de.debian.org != ftp.pl.debian.org != ftp.debian.org Feb 13 14:50:51 and each of them is DEBIAN_MIRROR so can be selected by user Feb 13 14:51:12 yes, we should store only checksums of source files Feb 13 14:51:21 Jay7: that's also a good thing. Specific digests are not I think. But anyway any other solution will be simpler than what we have now Feb 13 14:51:25 jay7 yes Feb 13 14:51:35 hrw coukd write a proposal to the dev ml? Feb 13 14:51:42 Jay7: [ftp://ftp.debian.org/debian/pool/main/a/apmd/apmd_3.2.2.orig.tar.gz] Feb 13 14:51:52 I vote to store checksum inside *.bb file Feb 13 14:52:00 Jay7: it is source and we have it 3 times - debian, pl.debian, us.debian Feb 13 14:52:12 hrw: 'basename' it :) Feb 13 14:52:32 if we already have bitbake support for in-files checksums than please use them. Feb 13 14:53:22 Jay7: apmd_3.2.2.orig.tar.gz on my server has 3 extra files in it. how you support such situation if only basename is used? Feb 13 14:53:39 hm.. well Feb 13 14:54:23 time to write summary mail to oeml... Feb 13 14:54:37 freebsd peoples store that separate: URL without source file name and source file name Feb 13 14:55:25 (that was very simplified view about freebsd ports) :) Feb 13 14:56:37 hrw: you can use patches or extra distfile e.g. to supply that extra files.. Feb 13 14:56:44 but it is workaround Feb 13 14:56:49 not a general solution Feb 13 14:57:05 Jay7: it was hipotetic Feb 13 14:57:20 (but it seems to me better than having multiple urls with distfiles with same name) Feb 13 15:00:59 bb in ~1h Feb 13 15:19:06 morning Feb 13 15:19:45 03Diego Dompe 07org.openembedded.dev * r1bd17718ed 10openembedded.git/packages/ (9 files in 4 dirs): Feb 13 15:19:45 dsplink: add dmai 1.20 support Feb 13 15:19:45 linux-omap 2.6.28: add patch for ISP resizer, needed for dmai Feb 13 15:29:21 Greetings Feb 13 15:32:59 I don't want to install any kernel files (in /lib/modules). I don't want to compile any kernel. How can I achive this? Feb 13 15:33:21 in Poky Linux I have 'linux-dummy' package which installs "nothing". What is an OE way? Feb 13 15:33:59 the problem is that task-boot REQUIRES virtual/kernel provider Feb 13 15:34:11 the problem is that task-boot REQUIRES virtual/kernel provider Feb 13 15:52:15 g'day kergoth Feb 13 15:52:42 booxter: either add it to ASSUME_PROVIDED, or do what you describe with poky, or don't use task-boot. Feb 13 15:54:24 task-boot's mission is to encapsulate the stuff that is required to boot a traditional unix-type system, and a kernel is indeed necessary for booting purposes so this requirement is not inherently unreasonable. Feb 13 15:56:38 pb_: I use my own kernel (built with custom bitbake file in a separate build system). I tried to make it a provider for virtual/kernel but with no result. Feb 13 15:56:53 no result? Feb 13 15:57:08 I think it's because I don't inherit kernel class or other such stuff Feb 13 15:57:17 but i may be wrong Feb 13 15:57:31 Does anybody know what might be wrong if I get http://www.pastebin.ca/1336145 errors when booting? Feb 13 15:58:19 thesing: your /etc/inittab has wrong format Feb 13 15:59:26 maybe your /sbin/init expects another stuff in the file Feb 13 16:00:19 it looks fine to me: http://www.pastebin.ca/1336147 Feb 13 16:00:31 booxter: inheriting kernel.bbclass shouldn't be necessary for your package to be recognised as a provider of virtual/kernel. Feb 13 16:00:43 thesing, real init or busybox? Feb 13 16:01:16 busybox I guess. Its a normal console-image. Feb 13 16:01:18 re Feb 13 16:01:25 bitbake doesn't have that much introspection; if you mention virtual/kernel in PROVIDES, and your bb file isn't obviously disqualified by things like COMPATIBLE_HOST, it will be eligible. there's no logic in there to try to determine whether it actually does build a kernel. Feb 13 16:01:35 kergoth: yeap, no result. I pushed PROVIDES += "virtual/kernel" into my kernel bb, I specified PREFERRED_PROVIDER for it in my machine config file. Then build system for my minimal-image weirdly tried to build at91 (?) kernel and failed Feb 13 16:01:47 plain old console-image should use real init Feb 13 16:01:54 last i knew Feb 13 16:01:57 which was a while ago Feb 13 16:02:20 Ok. I will try to find out. Feb 13 16:02:42 pb_: is it enough to make my package a provider? Feb 13 16:02:56 If yes then I'll try to make it again Feb 13 16:03:13 yes, it should be Feb 13 16:03:29 ok, I'll try it again Feb 13 16:06:19 hiya Feb 13 16:08:10 03Otavio Salvador 07org.openembedded.dev * rd31d5a45da 10openembedded.git/packages/cups/ (cups-1.3.8/use_echo_only_in_init.patch cups_1.3.8.bb): cups: add 1.3.8. Feb 13 16:08:11 03Otavio Salvador 07org.openembedded.dev * r26c982aa38 10openembedded.git/packages/rt2x00/rt61-firmware_1.2.bb: rt61-firmware: add 1.2 Feb 13 16:08:12 03Otavio Salvador 07org.openembedded.dev * ra045dfacf6 10openembedded.git/classes/ (package.bbclass package_ipk.bbclass): package.bbclass, package_ipk.bbclass: add dynamic depends on locale packages (ipk specific) Feb 13 16:08:13 03Otavio Salvador 07org.openembedded.dev * r4eeb691678 10openembedded.git/packages/sun-jre/sun-jre6_1.6.0.10.bb: Feb 13 16:08:16 sun-jre6: add 1.6.0.10 Feb 13 16:08:18 Signed-off-by: Otavio Salvador Feb 13 16:08:20 03Otavio Salvador 07org.openembedded.dev * r1a783e6205 10openembedded.git/packages/parted/ (6 files in 2 dirs): Feb 13 16:08:23 parted: drop versions up to 1.8.8 Feb 13 16:08:25 Versions older then 1.8.8 are not maintained upstream and looks to Feb 13 16:08:27 have no users inside of OE hence makes no sense to be kept. Feb 13 16:08:29 Signed-off-by: Otavio Salvador Feb 13 16:08:31 03Otavio Salvador 07org.openembedded.dev * r168ba1763b 10openembedded.git/ (3 files in 2 dirs): Feb 13 16:08:34 flash-plugin: add 10.0.12.36 Feb 13 16:08:36 Signed-off-by: Otavio Salvador Feb 13 16:08:38 03Otavio Salvador 07org.openembedded.dev * r9e4e61a440 10openembedded.git/: (log message trimmed) Feb 13 16:08:45 Merge branch 'master' of git://projetos.ossystems.com.br/git/users/otavio/org.openembedded.dev into org.openembedded.dev Feb 13 16:08:48 * 'master' of git://projetos.ossystems.com.br/git/users/otavio/org.openembedded.dev: Feb 13 16:08:50 flash-plugin: add 10.0.12.36 Feb 13 16:08:52 sun-jre6: add 1.6.0.10 Feb 13 16:08:54 parted: drop versions up to 1.8.8 Feb 13 16:08:56 package.bbclass, package_ipk.bbclass: add dynamic depends on locale packages (ipk specific) Feb 13 16:18:21 mickey|zzZZzz mickey|fsomeet ping? Feb 13 16:25:47 Anyone knows what the error bellow is? Feb 13 16:25:48 NOTE: Checking Package: aufs-tools Feb 13 16:25:49 ERROR: QA Issue: No GNU_HASH in the elf binary: '/home/otavio/hacking/ossystems/oe/tmp/osthindesktop-lite/work/oppitz-netclient-533-ossystems-linux/aufs-cvs20081203-r3/install/aufs-tools/usr/sbin/aulchown' Feb 13 16:25:53 NOTE: DONE with PACKAGE QA Feb 13 16:26:05 This has happen now that I've merged with master again Feb 13 16:26:55 mips or no? Feb 13 16:27:46 Tartarus: no; i686 Feb 13 16:31:49 /WC/quit Feb 13 16:32:02 03Tom Rini 07org.openembedded.dev * raf5b13bb6c 10openembedded.git/packages/udev/ (udev-124/init udev-124/permissions.rules udev_124.bb): (log message trimmed) Feb 13 16:32:02 udev-124: Sanitize permissions.rules and remove warning about leading /. Feb 13 16:32:02 We created /etc/dev.tar by 'tar cf /etc/dev.tar /dev' and then when we Feb 13 16:32:02 extract tar complains about a leading /, but since we do 'cd /;tar xf ..' Feb 13 16:32:04 drop the leading / in creation. Feb 13 16:32:06 Go and take out all of the GROUP='...' lines in permissions.rules for Feb 13 16:32:08 groups that we currently do not have, and no package addgroup's. Matches Feb 13 16:32:13 03Tom Rini 07org.openembedded.dev * rebb3272a2d 10openembedded.git/packages/udev/ (7 files in 5 dirs): Feb 13 16:32:13 udev-{092,097,100}: Sanitize permissions.rules Feb 13 16:32:15 Go and take out all of the GROUP='...' lines in permissions.rules for Feb 13 16:32:19 groups that we currently do not have, and no package addgroup's. Matches Feb 13 16:32:21 udev-118. Feb 13 16:52:58 pb_: so I get this fucking at91 kernel build - which failed :( Feb 13 16:53:12 PROVIDES hasn't fixed the problem Feb 13 16:54:31 whereshould I look for dependency tree? Feb 13 16:54:36 I get the following: Feb 13 16:54:38 ERROR: Multiple .bb files are due to be built which each provide virtual/kernel (/home/booxter/winbox/oe//openembedded/packages/linux/linux-nxp_2.6.24.bb /home/booxter/winbox/oe//openembedded/packages/linux/at91_2.6.28.bb). Feb 13 16:54:38 This usually means one provides something the other doesn't and should. Feb 13 17:02:05 * mckoan wants to share this celebration tonight http://www.embedded.it/?q=content/timet-structure-celebration Feb 13 17:02:57 mckoan: http://coolepochcountdown.com/ Feb 13 17:03:23 mattface: cool! Feb 13 17:03:45 anyone can help me with that kernel provider problem? Feb 13 17:03:47 booxter choose one kernel Feb 13 17:04:07 dont know why you seleted two diffrent kernels for your board Feb 13 17:04:16 woglinde: I chose the one. What do you mean? Feb 13 17:04:33 where have I done it? :) Feb 13 17:04:46 PREFERRED_PROVIDER_virtual/kernel="linux-nxp" Feb 13 17:04:49 ? Feb 13 17:04:57 booxter: linux-2.6.28 has recently been added, so I suspect OE is tryiong to use it as latest one Feb 13 17:04:58 assuming you want linux-nxp Feb 13 17:05:00 yeap, that's what I have in my machine conf Feb 13 17:05:14 my linux-nxp is 2.6.24 Feb 13 17:05:52 mckoan: why does build system do it? Feb 13 17:06:02 booxter: set PREFERRED_VERSION_linux = "2.6.24" Feb 13 17:06:10 I don't want vanilla kernel even if it has larger version number Feb 13 17:06:25 at91_2.6.28.bb is no vanilla kernel Feb 13 17:06:28 ow, that's the point. Tnx Feb 13 17:06:32 it does this way because is OE standard behaviour Feb 13 17:06:49 otavio ? Feb 13 17:06:51 besides that no machine mayber besides x86 uses vanilla Feb 13 17:07:12 woglinde: linux-2.6.28 is a vanilla kernel Feb 13 17:07:27 I have only 2 or 3 patches applied for basic board stuff Feb 13 17:07:33 http://pastebin.com/m683c6654 Feb 13 17:08:10 look at the SRC_URI_append in the recipes Feb 13 17:08:15 otavio, missing : Feb 13 17:08:34 otavio, building again to see if anything else is amiss Feb 13 17:08:40 hrw: maybe latest patches had broken Linux4sam stuff? Feb 13 17:09:00 tartarus pleas look at the dev-ml there is the stuff discussed a bit Feb 13 17:09:20 at91_2.6.28? Feb 13 17:09:52 woglinde, wrt? Feb 13 17:10:09 ok, my fault Feb 13 17:10:51 Tartarus: oh ok Feb 13 17:10:58 Tartarus: ping me please Feb 13 17:11:01 otavio, I'll push in a sec Feb 13 17:11:05 mckoan: it tries to compile linux-2.6.24... Feb 13 17:11:07 Tartarus: great Feb 13 17:11:09 why do I need it? Feb 13 17:11:14 I don't want it :) Feb 13 17:11:32 I want to have 'linux-nxp' - that's all Feb 13 17:12:10 booxter, then you're missing something still in your recipe Feb 13 17:12:18 03Tom Rini 07org.openembedded.dev * rfec216f25e 10openembedded.git/classes/package_ipk.bbclass: package_ipk.bbclass: Add missing : from a045dfacf6f2541e0f410d9d109581f181303756 Feb 13 17:12:21 bitbake -g can help tell you what's being pulled in for what dependancy Feb 13 17:12:26 booxter: show me 'bitbake -e -b /home/booxter/winbox/oe//openembedded/packages/linux/linux-nxp_2.6.24.bb|grep ^PREF' output Feb 13 17:13:13 Tartarus: oh ok; I found the missing : Feb 13 17:13:14 03Marcin Juszkiewicz 07org.openembedded.dev * r3064dea9c5 10openembedded.git/packages/linux/at91_2.6.28.bb: Feb 13 17:13:14 at91: dropped not used kernel recipe Feb 13 17:13:14 it was added by mistake Feb 13 17:13:34 otavio, yes, fixed :) Feb 13 17:13:42 Tartarus: pushed? Feb 13 17:13:52 what CIA just said :) Feb 13 17:14:01 hrw: http://pastebin.com/m7069526d Feb 13 17:15:03 booxter: show me linux-nxp recipe Feb 13 17:15:50 I wonder how could nxp try to use AT91 linux Feb 13 17:16:55 booxter: why do you say "I get this f*** at91 kernel build" ? Feb 13 17:17:11 mckoan: because it failed :) Feb 13 17:17:19 I think it's fixed recently Feb 13 17:17:25 hrw: http://pastebin.com/m45657862 Feb 13 17:18:40 booxter: and your MACHINE=stb225 right? Feb 13 17:18:50 yeap Feb 13 17:18:52 booxter: btw - sick recipe it is ;) Feb 13 17:19:14 hrw: of course: it's just another build system fetched and built :)) Feb 13 17:19:47 booxter: and what error you get? Feb 13 17:19:59 that proprietary stuff is madness so I think it's easier to stay with this obsolete one for kernel stuff :) Feb 13 17:20:13 booxter: "bitbake -g console-base-image" will tell you which app require other kernel Feb 13 17:20:32 booxter: I think that problem is due to not using kernel.bbclass where something require something from it Feb 13 17:21:36 hrw: hm. What should I do with this 'depends.dot' file? Feb 13 17:21:53 look at it with text viewer or pass though dot Feb 13 17:22:32 have to go Feb 13 17:22:40 tnx for help Feb 13 17:22:44 np Feb 13 17:34:55 hi ant Feb 13 17:40:40 hello woglinde Feb 13 17:44:04 With BBPATH I can create external overlay with my custom packages. How should I externally merge my conf/checksum.ini, classes etc? Feb 13 17:44:19 checksum.ini is the hard one Feb 13 17:44:26 no way? Feb 13 17:44:28 hrw|gone just posted about what to do about that file Feb 13 17:44:57 then how do people use this BBPATH feature?.. Feb 13 17:45:08 03Tom Rini 07org.openembedded.dev * r264f51739d 10openembedded.git/packages/busybox/ (8 files in 2 dirs): Feb 13 17:45:08 busybox: Add a 'mountall' file and install for CONFIG_FEATURE_MOUNT_FSTAB=y Feb 13 17:45:08 For all versions of busybox that have the above enabled, bump PR and include Feb 13 17:45:08 the generic mountall (taken from sysvinit). In the inc file, if enabled Feb 13 17:45:09 make ${D}/etc/defaults/ and put this in there. Feb 13 17:45:11 03Tom Rini 07org.openembedded.dev * rc91ad1caf6 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Feb 13 17:50:37 so making a 'hard overlay' (copying custom files on the fetched tree) is the only real way to work with it. Bad-mad... Feb 13 17:50:41 Tartarus: there was a long discussion in Gentoo about their GLEP-0044 Feb 13 17:50:44 http://www.gentoo.org/proj/en/glep/glep-0044.html Feb 13 17:50:53 I see valid points there Feb 13 17:51:18 ok Feb 13 17:51:31 my biggest concern is wrt overlays Feb 13 17:51:40 I don't want to have to disable checksumming Feb 13 17:51:51 So I'm doing the fork checksums.ini thing Feb 13 17:51:54 Which only mostly sucks Feb 13 17:52:06 Tartarus: on Gentoo you create your own manifest for the locale overlay Feb 13 17:52:25 I see a lot of similarities Feb 13 17:52:53 Is that a "fork" tho or a merge A and B? Feb 13 17:53:09 ... into C Feb 13 17:53:29 well, instead of having one file per recipe now there is only one Feb 13 17:53:51 they saved approx 10% disk-space Feb 13 17:54:00 which is never bad Feb 13 17:54:17 We've only got one file that counts now :) Feb 13 17:54:29 It's just that anyone with an overlay has to fork it Feb 13 17:54:38 (like any other file they enhance/upgrade) Feb 13 17:55:40 wel, now is better than I remembered: "reducing the apparent tree size by about 20%" Feb 13 17:57:08 booxter: personally I just turn off checksums and ignore the checksum.ini Feb 13 17:57:11 oh, heh, he quit Feb 13 17:57:54 pb__, certainly the easy way :) Feb 13 17:58:53 yeah. until checksums.ini is abolished I think it is about the only way :-} Feb 13 18:00:01 then, big help comes keeping the sources on disk... Feb 13 18:01:11 * pb_ go home Feb 13 18:01:13 later all Feb 13 18:01:30 going too, bbl Feb 13 18:15:41 03Koen Kooi 07org.openembedded.dev * rd4e26970c9 10openembedded.git/packages/dsplink/dmai_1.20.bb: dmai: apply patches Feb 13 18:19:10 hm what causes this: ERROR: 119:SyntaxError:' if not '-locale-' and not '-dbg' and not '-dev' in pkgname' Feb 13 18:19:24 in do_package_write_ipk Feb 13 18:22:27 looks like an extra quote on the end Feb 13 18:23:04 eFfeM, pull again Feb 13 18:23:06 fixed already Feb 13 18:23:34 Tartarus: thanks ! pulled earlier today so I assumed I was up to date Feb 13 18:50:45 03Mike Westerhof 07org.openembedded.dev * r966a2682c0 10openembedded.git/conf/distro/include/slugos.inc: SlugOS: slugos.inc - nail boost preferred version down since newer one won't build Feb 13 18:52:37 03Roman I Khimov 07org.openembedded.dev * r97b340db7c 10openembedded.git/ (3 files in 2 dirs): tnef: Add 1.4.5 Feb 13 18:53:25 gpe-mini-browser[2] seem to be broken for me in the latest OE. Anyone else see it ? (GLib-ERROR **: The thread system is not yet initialized.). adding g_thread_init(NULL), seems to fix it. Feb 13 18:54:58 eve. Feb 13 18:55:38 re Feb 13 19:00:31 mwester, since samba > 3.2.4 doesnt have smbmount and a crap swapper script is its only replacement, package smbfs is going Feb 13 19:00:52 timtimred, ugg Feb 13 19:00:59 effectively no change Feb 13 19:01:04 Is it still in, in older versions? Feb 13 19:01:10 3.0.34 Feb 13 19:01:19 use mount.cifs instead Feb 13 19:01:26 its deprecated for a reason :) Feb 13 19:01:49 ok. Feb 13 19:01:53 3.0.34? ok, i'll drop down too for now, thanks Feb 13 19:02:31 3.2.4 never hit OE git, i cant remember if its in foonas git or not. Feb 13 19:02:41 if it is its in the history there. Feb 13 19:02:53 Is cifs a functional replacement? (i.e. can it be said that wiki documentation is the only real issue for end users?) Feb 13 19:03:12 yes, its a much better replacement Feb 13 19:03:14 :) Feb 13 19:03:56 plus you might save space if you are carrying both in your distro Feb 13 19:04:00 Then we'll drop it from the feeds, and let the first user who asks update the wiki. (I can't imagine any size/space differences that would matter) Feb 13 19:04:25 good stuff. Feb 13 19:04:33 :) Feb 13 19:05:47 i got netatalk to appear as a time capsule! Feb 13 19:06:06 however, it lacks a couple of things it actually needs to initialise properly Feb 13 19:06:16 (to support it actually using it as one) Feb 13 19:06:47 also it needs a special MDNS record including the devices MAC and other bits Feb 13 19:08:24