**** BEGIN LOGGING AT Fri Apr 27 02:59:57 2007 Apr 27 08:12:00 http://linuxdevices.com/news/NS9425962034.html Apr 27 08:12:02 good morning all Apr 27 08:12:37 morning Apr 27 08:13:02 good morning Apr 27 08:26:03 03hrw 07org.oe.dev * r48cf1d2b... 10/ (5 files in 3 dirs): Apr 27 08:26:03 OpenZaurus time is over - long live Ångström! Apr 27 08:26:03 http://blog.haerwu.biz/2007/04/26/openzaurus-time-is-over-long-life-angstrom/ Apr 27 08:26:03 http://www.openzaurus.org/wordpress/2007/04/26/openzaurus-time-is-over-long-life-angstrom/ Apr 27 08:26:06 03hrw 07org.oe.dev * r02ac25fb... 10/ (4 files in 3 dirs): base-files: drop OpenZaurus support Apr 27 08:26:10 03hrw 07org.oe.dev * r28f1b1d3... 10/ (3 files in 3 dirs): busybox: dropped OpenZaurus config for 1.01 Apr 27 08:26:15 03hrw 07org.oe.dev * r4e1d869e... 10/ (4 files in 4 dirs): initscripts: dropped OpenZaurus specific checkversion script Apr 27 08:26:19 03hrw 07org.oe.dev * rb0e03196... 10/ (3 files in 3 dirs): bluez: dropped OpenZaurus config Apr 27 08:26:24 03hrw 07org.oe.dev * r9f59d345... 10/ (3 files in 3 dirs): openzaurus-version: dropped Apr 27 08:26:29 03hrw 07org.oe.dev * r8a96b7a7... 10/ (1 packages/gtk+/gtk+_2.8.9.bb): gtk+ 2.8.9: dropped OpenZaurus RDEPENDS Apr 27 08:27:33 hrw|gone: s/life/live/ Apr 27 08:31:00 XorA: I know - posts are corrected but it was after commit Apr 27 08:31:02 morning Apr 27 08:31:13 hey hrw, XorA, pH5 Apr 27 08:31:19 good morning, hrw Apr 27 08:33:04 koen: next commits are for you Apr 27 08:39:21 03hrw 07org.oe.dev * r29d984ee... 10/ (3 files in 2 dirs): Apr 27 08:39:21 Familiar: drop distro configurations Apr 27 08:39:21 - org.openembedded.dev is incompatible enough with Familiar 0.8.4 to make that not feasible Apr 27 08:39:21 - Familiar team made it clear that 0.9 won't be using anything OE-based/derived Apr 27 08:39:25 03hrw 07org.oe.dev * r3de0a3bf... 10/ (1 packages/familiar packages/familiar/familiar-version.bb): familiar-version: dropped Apr 27 08:39:29 03hrw 07org.oe.dev * r3267272c... 10/ (5 files in 2 dirs): images: drop Familiar feeds informations Apr 27 08:49:15 hi Rod Apr 27 08:57:54 hey hrw Apr 27 08:58:07 * rwhitby returns from holidays Apr 27 08:58:19 rwhitby: enjoyed? Apr 27 08:58:43 yep. 10 days of perfect beach weather with the family. Apr 27 09:07:27 hi Leon, Holger Apr 27 09:09:09 03MartinDietze 07org.oe.dev * r30f615c6... 10/ (3 files in 3 dirs): mtx-3a.conf, nylon-image-mini.bb: fixes to not conflict with the recent bitbake versions. Apr 27 09:10:17 likewise: I'm starting a armv6 build from scratch to see which rpaths need to be munged Apr 27 09:10:47 I assume that chrpath command cant just remove them Apr 27 09:11:02 Hello, good morning all. Apr 27 09:11:21 XorA: Removing is ok. Apr 27 09:11:42 removing would be nicer than munging Apr 27 09:11:54 speed up the dynamic linker anyway Apr 27 09:11:56 XorA: the dynamic linker will always search the default paths (AFAIK) Apr 27 09:13:00 XorA: but the chrpath-native and chrpath-cross are not target aware... It's a hacky little program. Apr 27 09:13:18 XorA: So instead, I think we should just fix the few packages that are wrong. Apr 27 09:13:31 XorA: chrpath -d Apr 27 09:14:01 koen: yes, but that won't work for most of the host/target combo's Apr 27 09:14:21 koen: I was building overnight. libstc++ and fontconfig are bad, so far. Apr 27 09:18:28 The offender in fontconfig/Makefile: FREETYPE_LIBS = -L/home/leon/sandbox/e520/openembedded/build/tmp/staging/i686-angstrom-linux/lib -Wl,--rpath -Wl,/home/leon/s Apr 27 09:18:30 andbox/e520/openembedded/build/tmp/staging/i686-angstrom-linux/lib -lfreetype Apr 27 09:18:56 heh Apr 27 09:19:11 we fixed an older version of freetype to stop breaking rpath Apr 27 09:19:26 likewise: packages/freetype/files/no-hardcode.patch Apr 27 09:22:26 koen: tnx Apr 27 09:23:59 ~hail NFS Apr 27 09:24:04 * ibot bows down to NFS and chants, "I'M NOT WORTHY!!" Apr 27 09:43:21 koen: do you know the Dutch Bits'n'Chips magazine? Apr 27 09:44:57 yes Apr 27 09:45:08 they interviewed me at fosdem :) Apr 27 09:45:48 (2006) Apr 27 09:47:07 koen: did that appear in BnC? Apr 27 09:47:12 iirc, yes Apr 27 09:47:44 likewise: http://www.bits-chips.nl/artikel/art_view.asp?ac=view&art_id=7219 Apr 27 09:49:05 btw, check http://linuxdevices.com Apr 27 09:49:39 koen: I wonder if we can get a full-spread article on OpenEmbedded in there. Apr 27 09:49:52 Nice :-) Apr 27 09:50:54 koen: nice Apr 27 09:51:34 likewise: in BnC? Apr 27 09:51:58 koen: yup Apr 27 09:52:43 likewise: I guess it would take someone with decent writing skills Apr 27 09:53:01 http://www.linuxdevices.com/articles/AT7065740528.html Apr 27 09:53:41 hrw: OpenEmbedded 2.8 % Apr 27 09:54:05 hrw: TimeSys 1.1 % Apr 27 09:54:17 I wonder why 'kernel.org' and 'uclinux' are listed seperately Apr 27 10:01:42 most interesting is that the use of OpenEmbedded is expected to almost double over the next 2 years Apr 27 10:05:54 is there any known libxcb problem currently? Apr 27 10:06:19 there wasn't any last week Apr 27 10:07:02 ok, must be a local problem on my autobuilder ... thx. Apr 27 10:09:56 hrw: your oz time is over commit email triggered amavis spam filtering: "X-Amavis-Alert: BAD HEADER Non-encoded 8-bit data (char C3 hex): Subject:" Apr 27 10:11:05 ;d Apr 27 10:12:11 did you violate rfc822 or something with that subject? Apr 27 10:16:17 commit mails are autogenerated Apr 27 10:25:02 hrw: Do you think it makes sense to ask people who reported OZ bugs to verify if they are still present in the latest angstrom? Or is it a bit too early for that? Apr 27 10:28:57 03likewise 07org.oe.dev * rc9156882... 10/ (1 packages/freetype/freetype_2.3.3.bb): freetype_2.3.3.bb: Apply no-hardcode.patch and configure.patch. Needed for sane RPATH's in depending libraries (fontconfig/pango/...) Apr 27 10:29:33 Laibsch: I would rather close as WONT FIX with reason please re-report against Angstrom if still true, otherwise we import a million bugs that have long since died and no-one noticed Apr 27 10:29:47 XorA: good idea Apr 27 10:29:59 XorA: I would do that in a second stage Apr 27 10:30:10 You are talking about ~175 bugs. Apr 27 10:30:45 Laibsch: ok, do that first, then give 14 days or so and any bugs that havent moved, WONT FIX them Apr 27 10:30:53 exactly Apr 27 10:30:56 that was my plan Apr 27 10:31:00 Laibsch: thats my opinion, hrw? Apr 27 10:31:58 156 bugs Apr 27 10:32:48 XorA: +1 from me Apr 27 10:33:56 Laibsch: can you make that announcement with holy penguin pee sprayed by hrw, I dont have time today Apr 27 10:34:26 koen: Re: xcb: configure: error: XCB requires xsltproc. Apr 27 10:34:26 <= looks like a depends problem. will investigate further. Apr 27 10:34:38 * hrw promotes Laibsch to OZ BugMaster Apr 27 10:35:22 rwhitby: probably libxslt-native Apr 27 10:35:40 rwhitby: I suspect pH5 and I have xsltproc on our host Apr 27 10:35:45 XorA: You mean to the ML? Apr 27 10:35:51 xsltproc is separate package Apr 27 10:36:01 Laibsch: to MLs Apr 27 10:36:02 Laibsch: yes, make it part of the closing down process of OZ Apr 27 10:36:19 Laibsch: oz-dev/oz-user/oe-dev/oe-user I would suggest Apr 27 10:38:12 OK Apr 27 10:39:02 Laibsch: you have automat for bugtracker? Apr 27 10:39:10 145 comments... Apr 27 10:40:45 ~seen ph5 Apr 27 10:41:33 ph5 is currently on #gpe #htc-linux #oe #openmoko #handhelds.org. Has said a total of 3 messages. Is idling for 2h 5m 58s, last said: 'hey BabelO, kelly'. Apr 27 10:42:22 hrw: should be 156 Apr 27 10:43:04 koen: seems that libxcb_0.9.93.bb has xsltproc-native commented out ... Apr 27 10:43:47 rwhitby: I don't have a xsltproc-native package in my OE tree Apr 27 10:44:14 yeah, neither do I. and your tip of xslt-native seems to have fixed the problem Apr 27 10:44:25 it did? Apr 27 10:44:27 hmmm Apr 27 10:44:53 well, I did a manual bitbake of xslt-native, then tried libxcb again and it worked. Apr 27 10:45:34 I can test again on slugosbe ... Apr 27 10:50:28 rwhitby: lennert is going to try a EABI/EB debian port, again based on angstrom Apr 27 10:50:45 he has an idea where the bug is that you were seeing Apr 27 10:50:56 cool - once that works I can seriously look at SlugOS 5 being EABI. Apr 27 10:53:56 rwhitby, koen: which bug is that actually? Apr 27 10:54:14 messages getting cut off Apr 27 10:54:42 lennert thinks is a regression of the strlen() bug that was in glibc some years ago Apr 27 10:55:00 any objections to adding libxslt-native as a build depends for libxcb ? Apr 27 10:55:57 rwhitby: no Apr 27 11:00:09 ~lart ssh known_hosts Apr 27 11:00:09 * ibot drops a truckload of VAXen on ssh known_hosts Apr 27 11:00:18 koen, rwhitby: the strlen.S fix seems to be in our sources: http://lists.gnu.org/archive/html/bug-glibc/2002-12/msg00056.html Apr 27 11:04:28 03rwhitby 07org.oe.dev * rcb6f5bda... 10/ (1 packages/xcb/libxcb_0.9.93.bb): libxcb: adding DEPENDS on libxslt-native to fix configuration problem building on Debian Sarge host without xsltproc installed. Apr 27 11:12:35 Dr. mickeyl: guten morgen! Apr 27 11:12:42 * mickeyl bows Apr 27 11:12:45 Guten Morgen koen Apr 27 11:12:49 :) Apr 27 11:13:34 Hi Doc Apr 27 11:13:41 :) Apr 27 11:13:50 what's up, Doc? Apr 27 11:14:04 all cool so far Apr 27 11:14:10 (said in a bugs bunny voice) Apr 27 11:14:12 still slightly worried about my hernie, but lets see Apr 27 11:14:21 all cool? Apr 27 11:14:30 rwhitby: with carrot in hand ofcourse? Apr 27 11:14:30 it's 24 degrees C inside here Apr 27 11:14:39 heh Apr 27 11:14:41 well Apr 27 11:14:43 temperature... Apr 27 11:16:44 03koen 07org.oe.dev * rbd7c8e1b... 10/ (1 packages/freetype/freetype_2.3.3.bb): freetype: fix parse error Apr 27 11:43:16 ~curse busybox vi Apr 27 11:43:25 May you be reincarnated as a Windows XP administrator, busybox vi ! Apr 27 11:53:50 how is RPATH this morning? Apr 27 11:54:37 a bit better Apr 27 11:55:27 Crofton: http://rafb.net/p/ghILyP87.html Apr 27 11:58:03 hello, are you modifying linux kernel in oe environment? how to get the patch files? Apr 27 11:58:37 drat, my crystal still isn't back from repairs Apr 27 11:58:56 koen: I know why Apr 27 11:59:05 koen: my has higher priority Apr 27 11:59:22 that must be it Apr 27 12:01:52 Readon: in other words: mostly yes. read our metadate for infos Apr 27 12:02:52 koen: Could we have a MACHINE field in the OE BTS? Apr 27 12:03:36 hrw, do you mean .bb files? Apr 27 12:04:43 yes Apr 27 12:07:23 Laibsch: probably Apr 27 12:12:51 cute, rm_work fixes some rebuild issues Apr 27 12:16:13 koen: It would be nice to have. Can you please implement it? Apr 27 12:16:44 Laibsch: tell me where you want them and I can add them Apr 27 12:18:59 crap, we have 111 machines Apr 27 12:20:28 machine proliferation is hard to prevent :) Apr 27 12:21:23 hrw, perhaps there are some misunderstanding. i want to build images for my device using oe. i want to modify some kernel file and get a patch, so i wonder how to extract patch file from kernel in oe. Apr 27 12:22:24 hi all Apr 27 12:22:29 Readon: man diff Apr 27 12:22:43 Readon: learn quilt Apr 27 12:23:19 there was some discussion sometime back for a meta-package for debian or ubuntu that would create the dev enviroment for OE. Does such a thing materialzed ? Apr 27 12:23:41 does it mean i need manually make two copies of kernel in WORKDIR? Apr 27 12:24:12 Readon: how else do you diff? Apr 27 12:24:41 03koen 07org.oe.dev * r33c4281e... 10/ (1 packages/xcb/libxcb_0.9.93.bb): libxcb: depend on *lib*xslt-native, not xslt-native Apr 27 12:24:55 hmm raining Apr 27 12:25:19 Crofton: surprisingly, it's still not raining here Apr 27 12:25:21 i have expected a better way.:) Apr 27 12:26:03 my termometer shows 40°C in direct sunlight Apr 27 12:27:56 03koen 07org.oe.dev * r953977f7... 10/ (1 conf/checksums.ini): checksums.ini: correct js and xorg checksums Apr 27 12:28:30 hrw : you moved to south Europe ? :) Apr 27 12:29:18 nope Apr 27 12:39:48 what in OE provide 'cmp' (other then busybox)? Apr 27 12:40:59 diff according to http://packages.debian.org/cgi-bin/search_contents.pl?word=cmp&searchmode=searchfiles&case=insensitive&version=stable&arch=i386 Apr 27 12:41:29 koen: "dpkg -S `which cmp`" is easier ;D Apr 27 12:41:47 diffutils in OE Apr 27 12:42:06 * hrw builds ncurses under qemu Apr 27 12:44:50 koen: I want a MACHINE field alongside OS, VERSION, etc. It should be populated with at least the machines that angstrom supports + "other" Apr 27 12:45:42 While we're at it, I think it would be nice if the field names Hardware and OS are worded more clearly in that they refer to the build host (At least that is what I assume) Apr 27 12:50:12 Maybe rename Hardware to "Host" and the new machine field to "TARGET". Although I agree this is not super, either Apr 27 13:14:17 hmmm has anyone tested the debian meta-package for setting up required soft for OE ? Apr 27 13:14:29 the one mentioned in the wiki Apr 27 13:14:35 morning Apr 27 13:14:38 openembedded-essential one? Apr 27 13:14:49 Ifaistos: it only has to depend on required software Apr 27 13:15:40 yes that one Apr 27 13:23:07 Ifaistos: Works like a charm here Apr 27 13:23:09 Thanks, hrw Apr 27 13:23:46 03koen 07org.oe.dev * re3024909... 10/ (22 files in 3 dirs): linux-ezx: add experimental 2.6.21 kernel for a780 Apr 27 13:24:39 re packages/ffmpeg/ffmpeg_0.4.9-pre1.bb, source is unfetchable and upstream seems to have quite making releases. It is all SVN now. There is a bb file for the svn version. OK to drop packages/ffmpeg/ffmpeg_0.4.9-pre1.bb? Apr 27 13:24:39 seems to have quit Apr 27 13:25:56 Laibsch: +1 from me Apr 27 13:28:29 hrw: Sorry, I was wrong. I did not read the screen careful enough. Fetching fails but because of md5sum mismatch. Apr 27 13:28:35 * Laibsch will not drop Apr 27 13:28:49 it Apr 27 13:29:04 NOTE: The MD5Sums did not match. Wanted: '700b66230a8322493c032c04372b8d74' and Got: 'ea5587e3c66d50b1503b82ac4179c303' Apr 27 13:30:32 Should I just update the md5sum? Apr 27 13:30:46 Laibsch: send a mail to the list first Apr 27 13:30:48 shit. looks like source mirror has some archives wrong Apr 27 13:30:59 it looks like the CELF mirror has some broken stuff Apr 27 13:32:31 03xora 07org.oe.dev * r27cc1147... 10/ (1 conf/machine/mx31ads.conf): mx31ads.conf : create tar images Apr 27 13:32:36 03xora 07org.oe.dev * r9609c456... 10/ (1 packages/linux/mx21ads-kernel_2.6.19rc6.bb): mx21ads-kernel_2.6.19rc6.bb : remove bogus inherit package Apr 27 13:32:42 03xora 07org.oe.dev * r59f201d6... 10/ (3 files in 3 dirs): Apr 27 13:32:42 linux-rp_2.6.20.bb : fix loading of 8250 module which makes all those BT/modem/other Apr 27 13:32:42 CF cards people use work. Also fix the dmesg so it prints the correct ttySX device Apr 27 13:32:42 for inserted cards (ttyS4 in practice) Apr 27 13:32:42 The fault was 2.6.20 added an extra pxa serial port (not actually initialised) but Apr 27 13:32:44 sitting in the minor space. Apr 27 13:32:59 Laibsch : thanks for the info on the meta-package :) Apr 27 13:33:19 XorA: 'extra port' being 'hwuart'? Apr 27 13:34:00 btw what's the best way to cellect all the patches a distro needs in order to publish them so that you can comply with GPL ? Apr 27 13:34:26 s/cellect/collect Apr 27 13:34:42 Ifaistos: copy your DL_DIR and OE.mtn database Apr 27 13:35:59 http://www.oesf.org/forums/index.php?showtopic=23749&st=40&p=159951&#entry159951 Apr 27 13:37:11 hrw: the touchscreen statement is not correct Apr 27 13:37:31 hrw: http://lists.linuxtogo.org/pipermail/openembedded-devel/2007-April/001993.html Apr 27 13:37:56 when I last time tested Angstrom it was ok Apr 27 13:37:58 koen: yes Apr 27 13:38:16 hrw: right, but the htc-universal commit last week broke it :/ Apr 27 13:38:23 koen : hmmm ok with the DL_DIR but i need to provide only the patches not the .bb files Apr 27 13:38:43 Ifaistos: technically you need the .bb files as well Apr 27 13:40:36 koen: can you do a rebuild for pxa zauruses Apr 27 13:41:12 koen : not so according to the laywer i talked to. .bb files are optional due to the MIT licence. one needs only to provide the changes for the software that requires it (i.e gpl) Apr 27 13:41:35 Ifaistos: the GPL also talks about the buildscripts uses Apr 27 13:41:39 Ifaistos: the .bb files make changes Apr 27 13:41:46 Ifaistos: it doesn't make that those scripts are MIT Apr 27 13:41:52 s/make/matter/ Apr 27 13:42:35 The changes the bb files are thought the patches Apr 27 13:42:45 The changes the bb files make are thought the patches Apr 27 13:42:51 Ifaistos: not the sed s////g ones Apr 27 13:43:22 Ifaistos: the .bb files are part of the buildsystem Apr 27 13:44:02 xdpyinfo is not building correctly for me... I'm trying to build gpe-image on x86 for x86.... an error log can be found at http://www.cortex-it.com/output_xdpyinfo.txt Apr 27 13:44:03 "The source code for a work means the preferred form of the work for making modifications to it. For an executable work, complete source code means all the source code for all modules it contains, plus any associated interface definition files, plus the scripts used to control compilation and installation of the executable." Apr 27 13:44:23 Ifaistos: section 3 of the gpl Apr 27 13:44:35 it looks like an automake issue. Apr 27 13:44:54 * Crofton goes insane thinking about this Apr 27 13:45:13 Crofton: cypress hill time :-D Apr 27 13:45:18 where is the md5sum info stored? Apr 27 13:45:28 Laibsch: conf/checksums.ini Apr 27 13:45:31 I touched that file once but cannot find it anymoe Apr 27 13:45:37 koen: thanks Apr 27 13:47:40 * Ifaistos hates due diligence for releasing stuff Apr 27 13:47:56 it's so much easier to write device drivers ;) Apr 27 13:48:13 Ifaistos: anyway, you only need to provide them on request, so just burn a dvd with DL_DIR and OE.mtn Apr 27 13:48:43 ah, my plan has paid off :) Apr 27 13:48:46 koen: that only works if you provide a written notice (kinda hard on FTP) Apr 27 13:48:51 * koen looks at merging MC from poky Apr 27 13:49:08 koen: my todays work? :D Apr 27 13:49:12 yes Apr 27 13:49:24 you remind me of something to add Apr 27 13:49:31 Ifaistos: http://www.openembedded.org/gpl-causing-problems-for-derivative-linux-distros would be interesting for you as well Apr 27 13:50:41 koen : thanks. i am aware of that Apr 27 13:51:45 b) Accompany it with a written offer, valid for at least three years, to give any third party, for a charge no more than your cost of physically performing source distribution, a complete machine-readable copy of the corresponding source code, to be distributed under the terms of Sections 1 and 2 above on a medium customarily used for software interchange; or, Apr 27 13:51:47 * Ifaistos needs to look again into release details more Apr 27 13:53:08 XorA : This works nice if you distribute devices + OS stuff Apr 27 13:53:28 Ifaistos: At the end of the day, consider who can come after you: the copyright holder and the copyright holder ONLY Apr 27 13:53:33 Ifaistos: yeah, but not if you use http/ftp site, then you really need to put the source beside it Apr 27 13:53:52 Ifaistos: I doubt they are going to give you a hard time if you made a good effort to provide sources. Apr 27 13:54:03 Ifaistos: users are also not going to complain Apr 27 13:54:14 Ifaistos: Look at what knoppix.org does Apr 27 13:54:16 didnt OE/OZ receive a warning from FSF? Apr 27 13:54:28 Laibsch : it's not users that i am "afraid" of.... Apr 27 13:54:37 XorA: OE? I cannot imagine Apr 27 13:55:04 Ifaistos: knoppix.org provides that offer to ship source on request Apr 27 13:55:14 XorA: that was familiar in the olden days Apr 27 13:55:27 and I have only found reference to it in some mail from Mr Knopper Apr 27 13:55:27 koen: ah, I vagely remebered it Apr 27 13:55:37 it is not too prominent anywhere Apr 27 13:55:38 XorA: short version: FSF is too stupid to use crosstool Apr 27 13:56:22 koen: I wonder what it is that you have that urge to always call people stupid? Apr 27 13:56:59 Laibsch: because those people are stupid? Apr 27 13:57:35 03koen 07org.oe.dev * rbe8a841a... 10/ (1 packages/linux/linux-ezx-2.6.21/pxa27x-udc-support.2.patch): linux-ezx 2.6.21: fix udc patch Apr 27 13:59:28 koen: Even if that was true. I know a couple of people that are probably not the sharpest tools in the shed. It does not make me any "smarter" (for real and in others' perception) to run around calling them stupid, quite the opposite. Get the clue? Apr 27 13:59:52 Being smart is not everything you need in life. Apr 27 14:00:09 smart and 12 inches :-) Apr 27 14:00:34 Laibsch: I don't care Apr 27 14:00:47 I prefer being direct Apr 27 14:02:20 XorA: your wife complain? Apr 27 14:02:21 ;D Apr 27 14:23:12 new logging works nice, but logread doesnt work anymore :-( Apr 27 14:23:49 I'm trying to build xdpyinfo and the most relevant error seems to be from automake: 'Makefile.am:96: LINT does not appear in AM_CONDITIONAL', any hints. Apr 27 14:49:45 Hey all. I've got a general question about oe. What do the different distributions, based on OE have in common? Apr 27 14:50:46 They all use OE. Apr 27 14:50:57 :) Apr 27 14:51:08 Beyond that, it's quite possible (and is being done, in fact) to generate VERY different distros with OE Apr 27 14:52:07 what is OE? Apr 27 14:52:11 For example, OE is used to build the "Unslung" firmware release for the NSLU2 hardware, as well as the SlugOS release for the same hardware. I can't think of anything in the final firmware that would really be a "signature" of OE, per se. Apr 27 14:52:42 mwester, what's NSLU2? Apr 27 14:53:44 OE is an infrastructure for organizing the building of software. The NSLU2 is a Network Storage device manufactured by Linksys Apr 27 14:54:53 also called slug by the big circle of fans, sometimes using it as servers Apr 27 14:55:02 installing different distros on it ^^ Apr 27 14:55:05 wiki0d Apr 27 14:55:28 mwester, so open embedded is only a platform for building software. is that right? Apr 27 14:55:42 Yep. And all the firmware for it, with the exception of the official Debian firmware, is build using OE. Apr 27 14:55:57 "for it" referring to the NSLU2 Apr 27 14:56:12 and with exception of the original linksys system Apr 27 14:56:51 ;) There are some that might object to the "*only* a platform", but I think that is a general description that works. Apr 27 14:58:41 hrw: oops - you're right, I forgot about the Linksys firmware... I guess you can tell I don't use that much! :D Apr 27 15:00:24 dcordes: OE has information on how to cross-compile stuff and put things together in compilation such as images etc. Apr 27 15:00:33 That is my interpretation Apr 27 15:01:51 re: xdpyinfo and LINT - I just commented out the last few lines of the Makefile.in that refered to LINT and it's all off and building again... Apr 27 15:02:07 I have no idea if this will corrupt something further down the food chain... but it's building. Apr 27 15:10:13 Hello all Apr 27 15:10:42 hey sirfred Apr 27 15:10:47 yo sirfred Apr 27 15:10:52 sirfred: hey Apr 27 15:11:15 Back to try to build a working angstrom image for my husky Apr 27 15:11:57 I've found that my yesterday image lacks libgcc. Do you know what recipe should have provided it? Apr 27 15:12:13 gcc-cross Apr 27 15:12:23 hrw: Thanks. Let's see Apr 27 15:13:17 work/armv5te-angstrom-linux-gnueabi/gcc-cross-4.1.2-r2/install/libgcc is empty. That's not a good thing, is it? Apr 27 15:14:41 It doesn't seem to be neither in the staging directory, and there's no ipk for it. Apr 27 15:24:38 When we have in a recipe something like: Apr 27 15:24:40 FILES_libgcc = "${target_base_libdir}/libgcc_s.so.1" Apr 27 15:25:00 I suppose that is the install path? Where do it looks for that file? Apr 27 15:47:06 ~lart RPATH issues Apr 27 15:47:06 * ibot calls RPATH issues on the phone ... the lights are on but nobody's home Apr 27 15:48:34 sirfred: ewi546.ewi.utwente.nl/tmp/chrpath.patch Apr 27 15:49:41 cu Apr 27 15:50:57 koen: Thanks! Apr 27 15:52:51 03xora 07org.oe.dev * rfd264691... 10/ (1 site/x86_64-linux): site/x86_64-linux : fix definition of db_cv_mutex thanks to polynomous Apr 27 15:52:55 03koen 07org.oe.dev * r10d03e8b... 10/ (1 packages/telepathy/libtelepathy_0.0.53.bb): libtelepathy: stage headers (from poky) Apr 27 15:52:59 03koen 07org.oe.dev * r16ebba39... 10/ (1 packages/telepathy/telepathy-mission-control_4.21.bb): telepathy-mission-control: merge from poky Apr 27 15:53:07 hi all Apr 27 15:55:09 koen: tnx for the freetype parse fix. Apr 27 15:55:39 like2wise: that just turned up when searching for rpath in bonsai :) Apr 27 15:56:38 likewise: http://www.openembedded.org/bonsai/query/description/rpath/?scm_id=3 Apr 27 16:44:48 ~convert 200 feet to meters Apr 27 16:44:51 200 feet is approximately 60.96 meters Apr 27 16:52:14 It's a pitty that gpe-conf is only providing 48x48 icons. It would really look nicer with 64x64 ones on a zaurus clamshell Apr 27 16:53:05 Shouldn't be better to relay in an icon theme, isolated from the apps for such generic icons? Apr 27 16:54:00 koen: Watching NCIS again? Apr 27 16:54:21 hvontres|poodle: no, the A-team, followed by Law & order Apr 27 16:54:53 koen: I pitty da fool that doesn't use OE... Apr 27 16:55:03 The A-team... Apr 27 16:55:08 Old times Apr 27 16:55:38 this episode is from 1985 :) Apr 27 16:55:52 The zx-spectrum times Apr 27 16:57:04 Well, I gave up and installed a prebuild angstrom image. Apr 27 17:00:45 * koen looks outside Apr 27 17:00:51 screw L&O Apr 27 17:00:57 ? Apr 27 17:00:57 * koen moves to the balcony with a book Apr 27 17:05:54 Amazing, not bitbake stucked while applying a patch to freetype Apr 27 17:05:59 s/not/now Apr 27 17:06:12 * sirfred feels tired Apr 27 17:07:56 rm -rf ./tmp ./sources... I'm just gonna start all over again... Apr 27 17:08:04 sirfred: I feel your pain. ;-) Apr 27 17:08:09 vervain: ;-) Apr 27 17:08:42 I'm throwing out many hours of compile time just to be sure. Apr 27 17:09:24 vervain: Well, I have suffering a bunch of weird effects. bitbake now compiles a lot of unwanted packages, libgcc package is not generated,... Apr 27 17:09:54 Curious, I just Ctrl-C the build, repeat again and now it passed without a problem the do_patch stage Apr 27 17:11:51 sirfred: Ya - I'm finding a lot of 'curiosities' as well ;) Apr 27 17:13:05 vervain: hehe. Yesterday when a bitbake gpe-image started to build abiword I felt as if it was starting to build the complete universe on my machine. Apr 27 17:23:06 It seems that gcc4 doesn't like snes9x code Apr 27 17:27:10 hey Apr 27 17:27:19 Hey zecke Apr 27 17:41:38 NOTE: Running task 1 of 3507.... Apr 27 17:42:49 03koen 07org.oe.dev * rb6102998... 10/ (1 packages/freetype/freetype_2.3.3.bb): freetype: remove spurious patch Apr 27 17:44:24 Curious, angstrom-x11-image includes gdb. Apr 27 17:44:39 all angstrom images do Apr 27 17:44:51 That's good, so I don't need to install it. Apr 27 17:45:04 DEBUG_APPS += '${@base_conditional("DISTRO_TYPE", "release", "", "tslib-tests tslib-calibrate gdb strace procps",d)}' Apr 27 17:45:40 makes the images a bit bigger, but those apps go away when setting DISTRO_TYPE=release :) Apr 27 17:46:03 Good to know Apr 27 17:46:21 yehh...glibc is done re-building... Apr 27 17:50:00 A curious error building snes9x. In an .h file, they declare an struct. Apr 27 17:50:21 Then, in a cpp file, under a extern "C" clause, they declare an instance of that struct, something like: Apr 27 17:50:27 struct SDSP1 DSP1; Apr 27 17:50:34 code in a .h file? Apr 27 17:50:36 The compiler complains about a linkage mismatch. Apr 27 17:50:40 sounds like a mozilla code Apr 27 17:50:46 koen: No, they declare the struct. Apr 27 17:50:54 ah Apr 27 17:51:11 I have taken away the extern "C" and was able to compile it, but, anyway... Apr 27 17:51:26 I suppose this was not happening under gcc-3.x or perhaps the compile flags are not the same. Apr 27 17:52:36 Anyway, as the whole program seem to be C++, and it is a final executable, I don't understand why that symbols are declared with "C" linkage, perhaps I'll get a surprise trying to run it Apr 27 17:55:36 ipkg is actually a CPU and memory eater Apr 27 17:59:06 Perhaps I have too many feeds. ipkg forks are failing Apr 27 17:59:39 you're out of RAM Apr 27 17:59:58 23% MEM, 96% CPU now Apr 27 18:01:04 The highest number I saw was 31% MEM, and then the fork failure. Apr 27 18:01:41 * sirfred deleting feeds Apr 27 18:03:27 Better now Apr 27 18:03:43 sirfred: how about a rewrite of ipkg with sqlite support? ;-) Apr 27 18:04:09 JustinP: That should be nice. When are you starting? ;-) Apr 27 18:05:07 Talking seriously, it could be a good idea. But, how does sqlite performs memory wise? Apr 27 18:05:36 better than parsing and reading everything into memory manually Apr 27 18:05:52 plus it's a library that is meant to be able to handle that kind of thing Apr 27 18:06:15 people have talked about it but AFAIK it's not been started still....we have this old ipkg cruft lying around... Apr 27 18:06:44 * sirfred writes 'ipkg sqlite support' in his list, after 'IDCT decoding', 'better overlay' and 'world domination' Apr 27 18:06:47 * koen proposes to switch to apt ASAP Apr 27 18:07:31 koen: I never felt apt fast in my desktop machine Apr 27 18:07:44 is 'fast' a requirement? Apr 27 18:07:53 it's probably written better, though.... Apr 27 18:08:07 ipkg is slow because it's doing unneccesary reading/writing/shugging Apr 27 18:08:17 likewise: Well, it would be enough for me if it doesn't fail on forks when I add three repositories. Apr 27 18:08:37 sirfred: add more swap space :) Apr 27 18:08:38 I would like the config scripts to be runnable from a host, so that target images are ready-for-read-only deployment Apr 27 18:09:12 hvontres|poodle: It will be more than zero, I see that I have no swap space. Apr 27 18:09:41 sirfred: heh.. the failing forks are from ipkg going (b)oom Apr 27 18:09:57 * hvontres|poodle knows oom all too well.... Apr 27 18:09:58 Great! Super Mario World is back on my husky! Apr 27 18:10:20 sirfred: nothing like old games to drive innovation.... Apr 27 18:10:52 hvontres|poodle: Well, it gave me a reason to implement xrandr mode switching on Ximageon Apr 27 18:11:05 But, omg, it runs like molasses Apr 27 18:11:59 sirfred: ahhh, not OMG...you'll wake up the koen :) Apr 27 18:12:13 * sirfred hides Apr 27 18:12:13 koen: I like your slides Apr 27 18:12:57 It's unplayable on angstrom... Apr 27 18:13:10 It worked fine on oz. What could be the reason? Apr 27 18:13:57 sirfred: that is illegal C++ thats why it broke gcc 4.1.X Apr 27 18:14:23 XorA|gone: Declaring a C++ struct and using it in a C linkage zone? Apr 27 18:14:48 XorA|gone: Perhaps they need an extern "C" also in the declaring section Apr 27 18:15:39 sirfred: exactly Apr 27 18:15:57 sirfred: otherwise the declaration has C++ name mangling but that doesnt work inside extern "C" sections Apr 27 18:16:59 Humm, there are a lot of alignment traps with snes9x. I remember there were also present under OZ. Could this hit performance worse in ansgtrom? Apr 27 18:17:31 sirfred: I would say that's a good indication of a performance hit.... Apr 27 18:18:00 hvontres|poodle: The traps were already on oz, and the game was faster, much faster. Apr 27 18:18:02 zecke: the slides don't make a lot of sense without the accompanying talk Apr 27 18:18:28 But I'm not sure if I'm getting the same amount of traps, of course Apr 27 18:38:55 Are the default angstrom C optimizing flags different from the used in Openzaurus ? Apr 27 18:48:25 sirfred: -Os for angstrom Apr 27 18:48:45 XorA: Optimize for size? Apr 27 18:48:54 sirfred: ye Apr 27 18:48:55 s Apr 27 18:49:00 That could make a difference. Apr 27 18:49:14 sirfred: override it if needed Apr 27 18:49:22 XorA: What is the right way to do it? Apr 27 18:50:43 koen: right, I assumed the talk to be good as well Apr 27 18:50:54 koen: at least the slides allow to give a good talk :) Apr 27 18:51:09 sirfred: look at angstrom-glibc.inc Apr 27 18:51:20 sirfred: and override those two variables in the package Apr 27 18:52:46 XorA: Thanks Apr 27 18:53:16 I would suggest with _angstrom overides when you finally come to send patches Apr 27 18:53:42 XorA: ok Apr 27 18:55:22 sirfred: you thought of seeing if the gp2x version of snes9x has any interesting optimisations, the gp2x is only 200Mhz Apr 27 18:55:48 XorA: Well, it worked really well on oz Apr 27 18:56:04 sirfred: ah ok cool Apr 27 18:56:24 * XorA makes mental thoughts towards psx4all people to release their work as GPL Apr 27 19:00:17 i have a couple of questions about dependencies that i was hoping someone could shed some light on Apr 27 19:00:40 maybe Apr 27 19:00:56 when i try to build my image with bitbake, i see the following in the debug messages Apr 27 19:01:13 DEBUG: providers for kernel are: ['linux-geodelx', 'linux-xxs1500', 'linux-myppc-2.6.13', 'ipod', 'linux-geodegx', 'linux-tornado-omap2', 'linux', 'linux-h6300-omap1', 'linux-dht-walnut'] Apr 27 19:01:25 NOTE: selecting linux-myppc-2.6.13 to satisfy runtime kernel due to PREFERRED_PROVIDERS Apr 27 19:01:48 however, the other prodiders are then added as dependencies too Apr 27 19:02:13 zecke: I started the talk with "if you have questions, ask them, don't wait for the end" and after slide 2 people started asking questions :) Apr 27 19:02:18 I am getting bad RPATH in gnutls Apr 27 19:02:23 e.g. DEBUG: adding /home/fredric/oe/org.openembedded.dev/packages/linux/linux-geodelx_2.6.11.bb to satisfy runtime kernel Apr 27 19:02:46 I thought likewise is not allowed sleep until all RPATH issues are fixed Apr 27 19:03:29 lfn: a) define a PREFERRED_PROVIDER Apr 27 19:03:47 lfn: b) bitbake will try to use another provider if the original one failed Apr 27 19:04:16 i defined a PREFERRED_PROVIDER in my conf/machines/myppc.conf file Apr 27 19:05:09 output from bitbake: NOTE: selecting linux-myppc-2.6.13 to satisfy runtime kernel due to PREFERRED_PROVIDERS Apr 27 19:06:16 lfn: I need to guess, either your kernel file is not complete Apr 27 19:06:30 lfn: or your preferred provider is wrong, but I can't guess belong this Apr 27 19:08:28 when you say that the kernel file is not complete, do you mean that the .bb file is wrong or that the actual kernel has not been built yet? Apr 27 19:09:23 Crofton__: :-) Restart the build and it will pass by, I'll see into it in a minute. Apr 27 19:10:52 lfn: I can only guess but you might miss a inherit kernel or such Apr 27 19:11:49 likewise, I am just building to find these problems now Apr 27 19:12:06 I am leaving town for the weekend, so no crises Apr 27 19:12:36 zecke: i have an 'inherit kernel' line in my linux-myppc.bb file. i'll go back and take a look at that file though. Apr 27 19:13:54 zecke: is it sufficient to have PREFERRED_PROVIDER_kernel in my conf/machines/myppc.conf file or does it need to be defined elsewhere too? Apr 27 19:14:46 rwhitby, likewise: there is indeed a problem with strlen on EABI/EB Apr 27 19:15:23 lfn: the place is okay, I wonder if the PREFERRED_PROVIDER itself is okay :) Apr 27 19:15:48 zecke: as a datapoint: my ep93xx build just built LAB-kernel for virtual/kernel instead of the ep93xx-kernel Apr 27 19:16:04 it worked fine last week, and hasn't been edited in ages Apr 27 19:16:17 zecke: yeah, that may be the problem. i'm just getting up to speed on oe. thanks for the suggestions! Apr 27 19:26:41 03koen 07org.oe.dev * rb8d9cc0d... 10/ (1 packages/telepathy/telepathy-mission-control_4.21.bb): telepathy-mission-control: add dep on gconf to fix configure Apr 27 19:39:17 http://www.esrac.ele.tue.nl/~leon/ELC2007.xhtml Apr 27 19:39:53 fun with statistics :) Apr 27 19:39:55 ^^^ Relative growth of embedded distro's, extracted from the Embedded Linux Survey Apr 27 19:39:59 :-) Apr 27 19:43:46 03koen 07org.oe.dev * r69a5b54d... 10/ (4 files in 3 dirs): empathy: merge with poky Apr 27 19:44:43 does anyone here use Mono from OE? Apr 27 19:46:08 cbrake: the guy that contributed the recipe did Apr 27 19:46:31 koen: :-) Apr 27 19:46:45 ~seen eblis Apr 27 19:47:10 koen: I'm getting a "The assembly mscorlib.dll was not found or could not be loaded." message. Apr 27 19:47:27 eblis was last seen on IRC in channel #openmoko, 45d 5h 27m 17s ago, saying: 'hmm, are the projects on projects.openmoko.org functional ?'. Apr 27 19:47:37 right, I remember getting that as well Apr 27 19:47:52 I needed to sprinkle some dlls throughout the filesystem Apr 27 19:48:11 * koen does mtn diff on mono/ Apr 27 19:48:22 crap, I only did it on the target Apr 27 19:48:30 * cbrake wonders what mscorlib.dll is ... Apr 27 19:48:44 m$'s version of libc? :) Apr 27 19:49:18 they would have named it "MicroSoftsFundamentalCoreLibraryFullOfIPR.dynamicallylinkedlibrary" Apr 27 19:49:23 but that didn't fit 8.3 Apr 27 19:50:19 zecke: heh. I like how .NET is totally contaminated with ms this, ms that ... Apr 27 19:50:38 likewise: http://rafb.net/p/w5TVpi95.html Apr 27 20:00:28 koen: well, installing mono-dll helps a lot :-) Apr 27 20:05:18 03koen 07org.oe.dev * r64b60930... 10/ (1 packages/telepathy/telepathy-mission-control_4.21.bb): telepathy-mission-control: include dbus service file as well Apr 27 20:07:37 the only mono I kind of like is the audio one :) Apr 27 20:08:05 "and renamed his reverse engineering tool r300reveng to revenge" Apr 27 20:08:09 now that's a cool name Apr 27 20:09:41 * koen reads http://tirdc.livejournal.com/ before trying to get DRI working on his efika Apr 27 20:12:23 koen: thanks Apr 27 20:12:59 koen: I wonder if specifying two .c files on the compiler line will allow the compiler to optimize-away though... Apr 27 20:13:26 * koen has no idea Apr 27 20:15:28 likewise: -funit-at-a-time Apr 27 20:15:45 likewise: but inter procedure support in gcc is 'just' getting extended Apr 27 20:16:05 LLVM is supposed to do these things way better Apr 27 20:16:10 e.g. link time optimisation Apr 27 20:16:27 zecke: interesting. I haven't been reading up on LLVM much... Apr 27 20:17:48 llvm rocks Apr 27 20:18:21 likewise: but yes throwing more code at gcc creates better results and smaller code Apr 27 20:18:41 likewise: e.g. KDE is piping most of the cc files of a unit into an all_foo.cc and then compiles this Apr 27 20:19:04 in theory it work Apr 27 20:30:30 XorA: I'm suspecting that perhaps the problem is libSDL, not snes9x Apr 27 20:34:54 * chouimat bangs his head on the desk ... Apr 27 20:39:39 sirfred: that would make sense, since libsdl seems to be sensitive to optimizations Apr 27 20:40:10 * chouimat curses ZoneAlarm and Winblows Firewall Apr 27 20:40:25 koen: I'm trying with : BUILD_OPTIMIZATION_angstrom = "-O2" and FULL_OPTIMIZATION_angstrom = "-fexpensive-optimizations -fomit-frame-pointer -frename-registers -O2" Apr 27 20:40:50 you could even try -O3 Apr 27 20:41:10 It seems that what is causing a lot of misalignment is sdl sound code Apr 27 20:41:30 When I disable sound in the game, I only get one trap message. Apr 27 20:41:51 Also, sound is corrupted badly. Apr 27 20:42:37 I really wish -Os to be the reason, if it's not... Apr 27 20:45:29 I see that now there are two different keyboards in the angstrom bundled matchbox, but there's no handwriting applet Apr 27 20:47:21 flag changing doesn't fix the problem. Apr 27 20:49:08 With snes9x running, 53,1% usr, 47,9 % sys Apr 27 20:49:08 Too much sys time, isn't it? Apr 27 20:51:05 likewise, libusb has RPATH issues .... Apr 27 20:52:18 Crofton: thanks, I'm still at gnutls :-) Apr 27 20:52:40 woglinde: hi Apr 27 20:53:00 Next week, I'll start putting these in bugzilla Apr 27 20:53:01 hi mrdata Apr 27 20:53:11 I will be out of town this weekend Apr 27 20:53:45 Crofton: hope you have a nice weekend Apr 27 20:53:48 thanks Apr 27 20:53:49 you too Apr 27 20:54:36 woglinde: comes the "request for exclusive IRQ could not be fulfilled" message for using hostap_cs or orinoco_cs ? Apr 27 20:54:55 mrdata it comes form ne200 Apr 27 20:54:56 0 Apr 27 20:55:17 woglinde: okay, then i will look on it Apr 27 20:56:10 woglinde: i getting the same messages while usage hostap_cs and orinoco_cs and it is solved now Apr 27 20:56:41 oh Apr 27 20:56:43 nice Apr 27 20:56:55 morning Apr 27 20:56:57 hm yes I have orinoco too but didnt test it Apr 27 20:57:14 mrdata have you new patches? Apr 27 20:57:45 likewise, apparently mplayer has bad RPATH also Apr 27 20:58:04 I guess nearly all packages have bad rpath Apr 27 20:58:18 not really Apr 27 20:58:34 Crofton: Hmm, reconsidering my job in this :-) Apr 27 20:58:34 woglinde: patches for what? ;-) Apr 27 20:58:39 for bootstrap-image I only see gnutls and libusb Apr 27 20:58:57 so far :) Apr 27 20:58:58 mrdata hm the irq problem Apr 27 20:59:32 woglinde: for hostap and orinoco yes -> http://lists.shmoo.com/pipermail/hostap/2007-January/014859.html Apr 27 21:00:19 likewise, you may need to make it easy to select WARN versus ERROR Apr 27 21:00:29 so when you are exhausted it can be WARN Apr 27 21:00:42 Crofton: Yes, I suggested this on the list. Apr 27 21:00:45 but we can still hunt bad packages by selecting ERROR Apr 27 21:00:58 Crofton: I *only* fixed bad RPATH detection, that's all. Apr 27 21:01:31 mrdata ah than ne2000 fixing should be easy Apr 27 21:01:35 thanks Apr 27 21:01:42 I understand it is bad, but I do not really understand enough to help fix Apr 27 21:01:44 Crofton: or just disable insane.bbclass for your builds Apr 27 21:02:12 I am for pushing really hard to imporve the metadata Apr 27 21:02:26 it really does help us long term Apr 27 21:02:33 Crofton: me too. But it shouldn't affect non-developers. Apr 27 21:02:35 searching the bad paths does waste time Apr 27 21:02:42 koen: any ideda how "bitbake angstrom-console-image" pulls in a whole bunch of gtk and X11 struff as a dependency? Apr 27 21:02:47 anyway, I need to run Apr 27 21:02:48 bye Apr 27 21:02:50 Crofton: conf/distro/include/angstrom.inc:24:INHERIT += "package_ipk debian multimachine sanity devshell angstrom-mirrors insane" Apr 27 21:02:52 Crofton: cu Apr 27 21:03:02 bye crofton Apr 27 21:03:55 * zecke still converts webkit to git Apr 27 21:05:03 hvontres|poodle: yes, avahi, see bugzilla Apr 27 21:05:48 koen: ahhh, thanks. I thought I was going nuts here...started a fresh build last night :) Apr 27 21:06:17 hvontres|poodle: it needs to get fixed, but atm I care too little to spend time on it Apr 27 21:06:55 koen: well, at least that'll speed up later x11-image builds :) Apr 27 21:09:02 any Makefile expert here?: what do the 'at's mean surrounding @LDFLAGS@ in a Makefile? Apr 27 21:09:30 likewise it will filled later Apr 27 21:09:38 or at run time Apr 27 21:09:43 woglinde: ok, thanks Apr 27 21:13:25 rwhitby: we isolated the problem: http://rafb.net/p/w5TVpi95.html (read around the dutch parts) Apr 27 21:13:47 likewise: or is that in a Makefile.in? Apr 27 21:14:23 koen: show me the assembly :) Apr 27 21:14:36 koen: at least we can figure out of we branch to strlen at all Apr 27 21:15:13 zecke, koen: I would re-compile the C code, each unit seperatly, then link them together, just to be sure (or use unit-at-a-time) Apr 27 21:15:29 s/seperatly/seperately Apr 27 21:15:54 JustinP: hmm, good question, should check in a minute Apr 27 21:16:12 likewise: cross/armeb-angstrom-linux-gnueabi/bin/gcc -static -funit-at-a-time -O0 y.c x.c -o foo ? Apr 27 21:16:38 koen: yes Apr 27 21:16:48 same result Apr 27 21:17:14 koen: nice work Apr 27 21:17:22 koen: ok, then the test really tested what lennert wanted it to test, I guess Apr 27 21:17:27 rwhitby: 99% is lennerts work Apr 27 21:17:51 like most of the arm code we thrive on :-) Apr 27 21:17:55 zecke: the non-static version: http://ewi546.ewi.utwente.nl/tmp/strlen.dump Apr 27 21:19:57 koen: does that test pass in EABI/EL ? Apr 27 21:20:01 84c4:ebffff7f __blue119__ 82c8 <.text-0x18> I wonder if that is the strlen call Apr 27 21:20:35 rwhitby: yes Apr 27 21:20:47 * koen hugs multimachine Apr 27 21:21:31 the le version: http://ewi546.ewi.utwente.nl/tmp/strlen.dump.le Apr 27 21:21:34 Hmm - I'll need to look back into it since my holidays, but I thought I saw a string truncation problem in arm-kernel-shim for EABI/EL too Apr 27 21:21:36 koen: qemu-armeb is not guilty? Apr 27 21:22:00 zecke: rwhitby noticed the problem on real hardware as well Apr 27 21:27:59 hm how to merge with vim? Apr 27 21:31:55 dunno Apr 27 21:33:09 3 way merge with vim Apr 27 21:33:38 * JustinP uses emacs Apr 27 21:34:11 vim documentation isnt clear about it Apr 27 21:36:48 woglinde: See bottom of http://www.handhelds.org/moin/moin.cgi/PaulSokolovsky Apr 27 21:37:57 oh okay Apr 27 21:37:59 thanks Apr 27 21:48:23 03rwhitby 07org.oe.dev * rd5da5c91... 10/ (1 conf/distro/unslung.conf): unslung.conf: Cleanup pinned versions and preferred providers. Apr 27 21:49:07 mrdata hm pcnet_cs.c seems to fixed Apr 27 21:51:21 woglinde: that means the error message is not seen anymore? Apr 27 21:54:30 mrdata I check it right now Apr 27 21:55:24 hm Apr 27 21:55:34 woglinde: you make the change in pcnet_probe() to IRQ_TYPE_DYNAMIC_SHARING | IRQ_FIRST_SHARED right? Apr 27 21:55:35 seems I killed my oe repos Apr 27 21:56:11 woglinde: ups :-( Apr 27 21:57:01 mrdata not really but I cant update anymore Apr 27 22:02:24 hi Apr 27 22:02:26 hi coredump Apr 27 22:07:18 mrdata but on the other way it isnt a problem at simpad because you can use only one card at time Apr 27 22:07:45 CoreDump|home: hey there :) Apr 27 22:09:09 woglinde: no pcmcia-card need two or more different irqs Apr 27 22:09:54 woglinde: on slot, at the time, right Apr 27 22:10:38 woglinde: how have you killed your repos? Apr 27 22:10:52 mrdata didnt the 3 mergeway right Apr 27 22:13:24 woglinde: bad, i do my work old style, slow, most with mc Apr 27 22:21:21 woglinde: strg-c early enough Apr 27 22:22:18 zecke what? Apr 27 22:22:55 woglinde: before the merge result gets comitted Apr 27 22:38:14 so Apr 27 22:38:17 repos is fixed Apr 27 22:45:55 CoreDump|home: does the 1.0.x branch have my patches applied or not? Apr 27 22:46:06 it has Apr 27 22:46:23 CoreDump|home: hmmm...it seems to be missig the script... Apr 27 22:46:34 oh Apr 27 22:46:53 CoreDump|home: np. I'll add it in when I get home..:) Apr 27 22:47:18 yeah, I probably forgot to "svn add" it after applying your patch Apr 27 22:47:45 woglinde: fine, and the changes in pcnet_cs.c are not helpfull? Apr 27 22:47:46 * hvontres|poodle needs to learn about svn now... this is his first commit acess :) Apr 27 22:49:00 * JustinP wonders if anyone actually cares about the CE-RH[12] Apr 27 22:49:10 (but him) Apr 27 22:49:21 mrdata hm as far as I read the threads it isnt smp safe Apr 27 22:49:30 but I am going to sleep now Apr 27 22:49:34 nite Apr 27 22:49:39 woglinde: night too Apr 27 22:49:40 night Apr 27 22:50:03 JustinP: I think you are the only one to splash out Apr 27 22:50:32 XorA|gone: you mean shell out? Apr 27 22:50:56 JustinP: that as well :-) Apr 27 22:51:43 JustinP: dosn't RP have one? Apr 27 22:52:12 * hvontres|poodle vaugly remembers him saying something about the remote getting him invlovled in kernel hackint Apr 27 22:52:28 ummm....wasn't that me? Apr 27 22:52:36 it's the only bit of kernel hacking I've done Apr 27 22:52:47 well, time to go home... Have a great weekend y'all :) Apr 27 22:52:59 perhaps he has a CE-RH1 and that started him....would seem awful strange for us both to have that happen that way, though ;-) Apr 27 22:53:04 hvontres|poodle: enjoy Apr 27 22:56:35 JustinP: I dont think RP has a remote, I have the OE project remote here Apr 27 22:58:54 XorA|gone: "project remote"? which one? Apr 27 22:59:12 JustinP: same as you Apr 27 22:59:25 * JustinP would be happy to finish the CE-RH1 driver if he had a CE-RH1 and Cxx0 (or someone with then to run tests) Apr 27 22:59:30 ah....pity Apr 27 22:59:30 JustinP: thats why I was able to give you the method to read the values accurately :-D Apr 27 22:59:53 I should really package my gizmod script too Apr 27 22:59:53 JustinP: I might build myself a RH1 if I get bored Apr 27 23:00:02 that would be very useful Apr 27 23:00:05 JustinP: now that would be cool Apr 27 23:00:10 do you have the schematics? Apr 27 23:00:45 JustinP: I found them on google once Apr 27 23:02:48 http://www.areanine.gr.jp/~nyano/cerh1pad.html Apr 27 23:02:51 bookmarked ;-) Apr 27 23:08:56 JustinP: thats the page, should be like 30 secs to solder that together and I have a C860 to test Apr 27 23:09:17 JustinP: wife is away on a course 13th May onwards, I shall attempt to do it then Apr 27 23:15:30 03rwhitby 07org.oe.dev * rf8b383b3... 10/ (7 files in 2 dirs): ixp4xx-npe: Updated to microcode version 2.3.2 Apr 27 23:16:30 time for me to sleep Apr 27 23:18:07 XorA|gone: I would greatly appreciate it Apr 27 23:18:10 XorA|gone: sleep well Apr 27 23:32:10 night all Apr 27 23:34:27 * rwhitby updates nslu2-linux to bitbake 1.8.2 Apr 27 23:55:28 nite Apr 28 02:45:18 03rwhitby 07org.oe.dev * r4b110499... 10/ (5 files in 5 dirs): ixp4xx-kernel: Added 2.6.21, and set it as default for slugos **** ENDING LOGGING AT Sat Apr 28 02:59:56 2007