**** BEGIN LOGGING AT Wed Mar 19 02:59:57 2008 Mar 19 03:55:18 wat characterizes an http sub-request? Mar 19 03:55:27 er what... Mar 19 03:56:57 I'm trying to work out a way to identify any direct request to view an image file, and reqrite the request to wrap some HTML around the image insead of just displaying the image in the browser as one would expect for a direct request Mar 19 04:07:10 ive seen that done on lots of sites Mar 19 04:07:13 but dont know how they do it Mar 19 04:09:04 well, as best as I can figure it would be done a rewrite and SSI script which then inserts the image into the document which is the target of the rewrite Mar 19 04:10:00 because an SSI script can evaluate 'DOCUMENT_NAME' indicating the document into which the SSI is embedded Mar 19 04:10:44 so the way this would work is, resuest comes in for /images/photo.gif Mar 19 04:11:21 gets redirected to /wrapper/image/photo.gif Mar 19 04:11:40 if you run apache you can do that easilly using mod rewrite Mar 19 04:12:08 XorA: the same can be done via lighttpd (rewriting) Mar 19 04:12:30 I would re-write to a small bit of php that generates the HTML you want Mar 19 04:12:42 but the trick is to wortk out the optimal way to identify when an image request is coming from being embedded within the wrapper file Mar 19 04:12:45 because you can re-write to wrap.php?image.gif Mar 19 04:13:47 XorA: more or less. I don't want it to be an external redirect, and I wouldn't use php, but esentially that's the idea Mar 19 04:14:24 the problem is, evaluating the cases where you don't want the redirect to occur Mar 19 04:14:53 hillct: mod rewrite can do internal redirects Mar 19 04:14:57 hillct: [p] Mar 19 04:15:11 dont know about lighttpd Mar 19 04:15:17 XorA: yes, I know. I've been doing this a LONG time :) Mar 19 04:15:50 hillct: so your the expert, quit asking get writing :-) Mar 19 04:16:01 for example, when an image is embedded in a document on the local domain or one from a select set of remote domains such as Google. It may be sufficient to check referrer, but perhaps not Mar 19 04:16:41 XorA: it's not a matter of can it be done, so much as a matter of the optimal approach Mar 19 04:16:44 hillct: put a cryptographic ? after the rewrite when the image is loaded from the wrapper, and then redirect requests which don't have the correct ? arg Mar 19 04:17:12 rwhitby: I like that... Mar 19 04:17:57 give or take the crypto bit depending on how draconian you want to be ... Mar 19 04:18:16 but there are plenty of cases where we don't want to make that a requirement, or else I would take an approach like trigger_b4_download Mar 19 04:18:32 yeah, maybe Mar 19 04:20:35 does the terget request subsequent to a rewrite-once () getthe initial request URL as a referer or does it get the referer, of the initial rewuest unmodified? (does the question even make sense?) Mar 19 04:22:59 rwhitby: thanks. I just realized I was musing about this in entirely the wrong window Mar 19 04:23:24 could have sworn I was in #lighttpd not #oe Mar 19 04:24:06 * * OE Bug 1634 has been RESOLVED (WONTFIX) by xjqian(AT)gmail.com Mar 19 04:24:08 * * gnome-desktop is no buildable due to lack of scrollkeeper-config Mar 19 04:24:10 * * http://bugs.openembedded.org/show_bug.cgi?id=1634 Mar 19 04:24:17 * hillct goes to sit in the corner for a while Mar 19 04:34:58 evening everyone Mar 19 04:38:31 03xjqian 07org.oe.dev * r173acf09... 10/ (7 files in 2 dirs): Mar 19 04:38:31 libbonobo: upgrade broken 2.14.0 to 2.22.0, unify Mar 19 04:38:31 * verifed the pthread patch is still needed to prevent configure from complaining about cross-compile Mar 19 04:48:52 I am having trouble compiling qt-embedded Mar 19 04:49:06 No static library support for platform/compiler Mar 19 06:16:01 ah I need to specify which qte is preferred if not defined it picks up the qte-mt-static recipe bt default Mar 19 06:16:18 static recipe is prolly not what I should use Mar 19 06:17:47 anyway PREFERRED_PROVIDERS_libqte2 = "qte" seems to get me past this issue Mar 19 06:19:59 * XorA kicks kdiff 3 for being a useless merger Mar 19 06:22:02 XorA: try emacs merger :D Mar 19 06:22:28 * XorA normally uses meld Mar 19 06:51:29 03xora 07org.oe.dev * rc7613c5f... 10/ (1 packages/perl/libxml-parser-perl-native_2.34.bb): libxml-parser-perl-native_2.34.bb : change ordering to make staging work Mar 19 06:56:21 03xora 07org.oe.dev * rf40ae929... 10/ (1 packages/tasks/task-base.bb): Mar 19 06:56:21 task-base.bb : dont RRECOMAND explicitly on alsa-states and by default Mar 19 06:56:21 alsa-state already does this and some DISTRO might not want this. Mar 19 06:56:29 XorA: How do u chose the diff managers in mtn ? Mar 19 06:56:42 MTN_MERGE = "meld" for example Mar 19 06:56:53 in environment Mar 19 06:57:04 XorA: ah you are my hero I hate ediff Mar 19 06:57:17 yeah I like meld Mar 19 06:57:34 so .bashrc is prolly right place Mar 19 06:57:35 * XorA has all this stuff setup at home, I forget just quite how OE optimised my own machine is Mar 19 06:57:46 Khem: yup, just export that variable Mar 19 06:58:24 XorA: thx Mar 19 07:55:03 morning Mar 19 08:15:49 03xora 07org.oe.dev * r15406433... 10/ (1 packages/mpeg2dec/mpeg2dec_0.4.0b.bb): mpeg2dec_0.4.0b.bb : ${PN} is already in packages, dont put it in there again Mar 19 08:37:59 03koen 07org.oe.dev * r57372f44... 10/ (1 packages/tzdata/tzdata_2007k.bb): tzdata: every entry in PACKAGES is by default in PROVIDES, so remove the bogus PROVIDES Mar 19 08:38:04 03koen 07org.oe.dev * r552396a9... 10/ (1 classes/testlab.bbclass): testlab: remove bashism, spotted by XorA Mar 19 08:56:00 bonjour Mar 19 08:56:45 bonjour Mar 19 08:57:47 morning Mar 19 09:00:07 Tartarus: That patch was applied if I remember rightly Mar 19 09:01:08 morning Mar 19 09:23:33 03koen 07org.oe.dev * rdf4bff9c... 10/ (1 packages/openmoko2/settingsgui_0.8.bb): settingsgui: update to 0.8 Mar 19 09:31:01 g'morning Mar 19 09:36:14 yo ant|work :) Mar 19 09:38:46 bjour ! Mar 19 09:40:08 03koen 07org.oe.dev * r299687c8... 10/ (3 files in 3 dirs): illume: add untested recipe that isn't likely to work (this one does need RDEPENDS) Mar 19 09:48:12 hi, I'm getting this on the angstrom 2007.12 branch Mar 19 09:48:36 NOTE: Task failed: localedef returned an error (command was PATH="/data/oe/build/tmp/staging/i686-linux/bin/arm-angstrom-linux-gnueabi:/data/oe/build/tmp/staging/i686-linux/bin:/data/oe/build/tmp/staging/i686-linux/bin:/data/oe/build/tmp/cross/bin:/data/oe/build/tmp/staging/i686-linux/bin:/data/oe/bitbake-1.8.10/bin:/usr/local/apache-maven-2.0.8/bin:/usr/local/apache-maven-2.0.8/bin:/usr/lib/qt-3.3/bin:/usr/kerberos/bin:/usr/lib/ccache:/usr/local/bin:/usr/b Mar 19 09:48:36 in:/bin:/usr/X11R6/bin:/home/cedric/bin" I18NPATH="/data/oe/build/tmp/work/armv5te-angstrom-linux-gnueabi/glibc-2.5-r9/locale-tree//usr/share/i18n" qemu-arm -r 2.6.16 -L /data/oe/build/tmp/work/armv5te-angstrom-linux-gnueabi/glibc-2.5-r9/locale-tree /data/oe/build/tmp/work/armv5te-angstrom-linux-gnueabi/glibc-2.5-r9/locale-tree/bin/localedef --force --old-style --no-archive --prefix=/data/oe/build/tmp/work/armv5te-angstrom-linux-gnueabi/glibc-2.5-r9/locale- Mar 19 09:48:36 tree --inputfile=/usr/share/i18n/locales/es_NI --charmap=UTF-8 es_NI). Mar 19 09:48:38 NOTE: package glibc-2.5-r9: task do_package: failed Mar 19 09:48:40 ERROR: TaskFailed event exception, aborting Mar 19 09:48:42 NOTE: package glibc-2.5: failed Mar 19 09:48:44 ERROR: Build of /data/oe/org.openembedded.angstrom-2007.12-stable/packages/glibc/glibc_2.5.bb do_package failed Mar 19 09:48:46 ERROR: Task 93 (/data/oe/org.openembedded.angstrom-2007.12-stable/packages/glibc/glibc_2.5.bb, do_package) failed Mar 19 09:48:49 NOTE: Tasks Summary: Attempted 444 tasks of which 361 didn't need to be rerun and 1 failed. Mar 19 09:48:50 03koen 07org.oe.dev * r797469a4... 10/ (5 files in 2 dirs): Mar 19 09:48:50 update-alternatives: move to opkg Mar 19 09:48:50 * u-a source needs to get moved to opkg-svn as well, but I'll leave that to the opkg dudes Mar 19 09:48:51 ERROR: '/data/oe/org.openembedded.angstrom-2007.12-stable/packages/glibc/glibc_2.5.bb' failed Mar 19 09:48:51 ~pastebin Mar 19 09:48:52 [~pastebin] A "pastebin" is a web-based service where you can paste anything over 3 lines without flooding the channel. Here are links to a few : http://www.pastebin.com , http://pastebin.ca , http://channels.debian.net/paste , http://paste.lisp.org , http://www.rafb.net/paste Mar 19 09:49:08 XorA: sorry Mar 19 09:49:21 cedric_: just saving your life :-) Mar 19 09:49:35 from you guys... Mar 19 09:49:53 from more anal peopel Mar 19 09:50:18 cedric_: it looks like gcc-native is not pulled before qemu...I thought this was backported Mar 19 09:51:22 ant|work: so should I wait for a fix? or is there a simple solution to fix this? Mar 19 09:51:56 cedric: gcc-3.4.4 is/was needed...try bitbake gcc-native Mar 19 09:53:23 Which task calls do_stage? There doesn't seem to be a task called stage?? Mar 19 10:04:29 good monring Mar 19 10:08:51 florian: "guten mogren" Mar 19 10:10:16 ant|work: same thing Mar 19 10:10:45 ant|work: ah... that happens frequently - I'll try again after the second coffee :-) Mar 19 10:11:01 ant|work: or should I rebuild qemu? Mar 19 10:11:02 cedric_: clean the previous glibc build... Mar 19 10:11:17 ant|work: that's what I've done Mar 19 10:11:21 cedric_: and verify in the logs you're using gcc-3.4.4 Mar 19 10:12:11 cedric: you can't imagine how much I hate that es_NI locale ;-) Mar 19 10:12:23 :-) Mar 19 10:13:18 cedric: qemu should have been built by gcc-native, BTW Mar 19 10:13:39 it even did not compile with gcc-4.x months ago... Mar 19 10:15:49 florian: good monring Mar 19 10:16:52 Could anyone recomend a variable which would bet set to the arch's staging dir please? E.g. ${STAGING_DIR}/arm-angstrom-linux-gnueabi/ Mar 19 10:17:41 03koen 07org.oe.dev * ref4c5705... 10/ (1 packages/sysvinit/sysvinit_2.86.bb): sysvinit: prepend shebang to shell scripts and fix formatting of those Mar 19 10:18:04 ant|work: I think it uses 4.1.2-r13, that's what I've got in my tmp/work/armv5te-angstrom-linux-gnueabi Mar 19 10:20:33 re Mar 19 10:22:28 cedric_: mmh, what's the gcc on your host? gcc3.x ? Mar 19 10:22:53 tcooksey: STAGING_DIR_TARGET Mar 19 10:23:23 hrw: Thanks! Mar 19 10:24:06 ant|work: I've got gcc (4.1.2) and gcc34 (3.4.6) Mar 19 10:24:31 tcooksey: for future: "bitbake -e -b some-recipe.bb|grep ^STAGING" Mar 19 10:27:00 hrw: seems there's more than one variable which points to the same directory (STAGING_DIR_HOST also works). I assume STAGING_DIR_TARGET is the one to use. Mar 19 10:27:47 cedric_: are you on Gentoo? Mar 19 10:28:37 ant|work: fedora 8 Mar 19 10:29:47 ant|work: when I've compiled qemu, it complain I didn't have gcc 3, so I've installed it... Mar 19 10:29:53 cedric_: the whole point of having gcc-native (preferred version 3.4.4) is that you don't need gcc-3x on host to properly compile qemu and glibc-locales Mar 19 10:31:58 03koen 07org.oe.dev * rbe700ed0... 10/ (1 packages/sysvinit/sysvinit_2.86.bb): sysvinit: add missing ! Mar 19 10:32:58 cedric_: you could try to uninstall gcc34 from host and clean/rebuild qemu (I had issues with gcc34 on Gentoo) Mar 19 10:33:15 ^3.4.6 Mar 19 10:33:47 tcooksey: read *names* of vars? Mar 19 10:35:07 hrw: yup... TARGET obviously the right one... not sure why host is set the same, surely that should point to x86_64-linux? Mar 19 10:35:27 tcooksey: You have BUILD, HOST and TARGET Mar 19 10:35:55 03koen 07org.oe.dev * r85e676eb... 10/ (1 classes/rootfs_deb.bbclass classes/rootfs_ipk.bbclass): rootfs_{deb,ipk}.bbclass: the o-hand dudes went overboard with sed, and opkg-native overwrites our u-a-native, so we need to set both IPKG_OFFLINE_ROOT *and* OPKG_OFFLINE_ROOT Mar 19 10:38:46 ant|work: nope qemu complained I don't have gcc 3... Mar 19 10:39:57 cedric_: bitbake gcc3-native then Mar 19 10:43:47 hrw: wasn't this backported? in .dev seems to work... Mar 19 10:46:10 hrw: same thing... Mar 19 10:46:35 RP: gcc-native broken after gcc's recipes clean-up? Mar 19 10:50:45 cedric_: http://lists.linuxtogo.org/pipermail/angstrom-distro-devel/2008-February/001901.html Mar 19 10:52:02 ant|work: going to move to .dev Mar 19 10:52:34 cedric_: but has been backported...pls try gcc-native one more time Mar 19 10:52:55 if it fails could be a new bug :-( Mar 19 10:56:17 ant|work: nope Mar 19 10:59:00 cedric_: nope = qemu does'nt compile or glibc-locale-generation error? Mar 19 10:59:33 ant|work: qemu complained about gcc 3 Mar 19 11:01:04 cedric_: gcc-native built fine? Mar 19 11:01:09 ant|work: yep Mar 19 11:01:42 ant|work: NOTE: package gcc-native-3.4.4: completed Mar 19 11:03:37 cedric_: PATH and BBPATH are ok? Mar 19 11:04:15 ant|work: I've used source-me.txt on the angstrom site Mar 19 11:05:32 ant|work: echo $BBPATH --> /data/oe/:/data/oe/build/:/data/oe/org.openembedded.angstrom-2007.12-stable/ Mar 19 11:05:56 ant|work: echo $PATH --> /data/oe/bitbake-1.8.10/bin Mar 19 11:06:03 ant|work: look ok to me... Mar 19 11:07:14 ant|work: I could try doing --> rm -rf tmp; bitbake gcc-native; bitbake x11-image Mar 19 11:07:26 cedric_: exactly... Mar 19 11:08:17 cedric_: but actually you should not need gcc-native build (automatically pulled in) Mar 19 11:08:38 cedric_: at least in .dev works like this Mar 19 11:09:11 cedric_: I'd just try "bitbake console-image" Mar 19 11:09:58 ant|work: gcc-native was built before, therefore when you asked me to build it --> I have to do a clean... Mar 19 11:11:40 cedric_: first do "rm -rf /data/oe/build/tmp" then bitbake an image Mar 19 11:12:41 ant|work: that's what I'm doing at the mo Mar 19 11:14:00 cedric_: It will take some time... :=} Mar 19 11:15:19 ant|work: with a nice core2duo... :-) Mar 19 11:23:36 ant|work: same pb with qemu... Mar 19 11:24:03 cedric_: you build on rhel4? fedora3? Mar 19 11:29:06 hrw: he said fedora 8 Mar 19 11:29:55 hrw: I can retry angstrom-stable with gcc-4.1.2 on host later on... Mar 19 12:01:59 03xora 07org.oe.dev * rcc0974f3... 10/ (3 files in 3 dirs): Mar 19 12:01:59 openmoko.conf : openmoko is not Angstrom even if we steal everything. Mar 19 12:01:59 xserver-kdrive-common_0.1.bb : PR bump because of above Mar 19 12:03:08 I want OE in git - maybe then people will start to write one line summaries of commits... (due to git shortlog) Mar 19 12:03:10 now openmoko is a distro? Mar 19 12:03:29 openmoko always a distro Mar 19 12:04:26 but now doesn't depends on angstrom Mar 19 12:05:47 methril: it does Mar 19 12:06:19 openmoko.conf : openmoko is not Angstrom even if we steal everything. Mar 19 12:06:23 ? Mar 19 12:06:38 methril: well read the commi Mar 19 12:06:39 t Mar 19 12:14:29 ant|work: please let me know... Mar 19 12:19:29 cedric_: If you feel brave enough try the .dev branch in the meantime Mar 19 12:20:05 cedric_: did you successfully build OE recently? Mar 19 12:21:09 cedric_: there have been some heavy changes last week... Mar 19 12:21:26 can someone help me? I get the following error:http://nopaste.org/p/alC6TpbSR Mar 19 12:22:59 ant|work: I haven't touch OE since last year... Mar 19 12:25:15 edistar_: see log Mar 19 12:25:25 in bbclass , where should i name do_ fonction and where it's not needed ? Mar 19 12:26:38 hrw: I don't understand it Mar 19 12:29:58 we do not see it Mar 19 12:33:37 by a convention , where should we prefix do_ in bbclass fonction ? Mar 19 12:54:15 hrw: could you have a look at the config.log for me? Mar 19 12:56:47 edistar: pastebin broken log Mar 19 12:57:00 edistar: /home/edistar/OE/build/tmp/angstrom/work/i686-linux/glib-2.0-native-2.12.4-r1/temp/log.do_compile.4655 Mar 19 12:57:16 hrw: http://nopaste.org/p/azJTZKdNR Mar 19 12:57:45 hrw: config.log too? http://nopaste.org/p/aKmUHY5tn Mar 19 12:58:44 edistar: what you run as distro? Mar 19 12:58:49 fedora1? Mar 19 12:58:52 suse 7? Mar 19 12:58:55 hrw: Debian Mar 19 12:58:58 lenny Mar 19 12:59:06 very strange Mar 19 12:59:48 Genesis: When its a task Mar 19 13:00:10 hrw: everything else built well.. only glib-2.0-native didn't Mar 19 13:00:40 RP : which difference you make between task and fonction ? Mar 19 13:00:58 Genesis: tasks are things bitbake runs - fetch, unpack, compile Mar 19 13:01:01 * hrw -> shop Mar 19 13:01:41 All tasks have an associated function but not all functions are tasks... Mar 19 13:01:52 it's subjectiv ok Mar 19 13:02:12 because i thought what fct i could export in my bbclass Mar 19 13:02:27 not easy to guess what an user would use and what is private to the bbclass Mar 19 13:05:50 03rpurdie 07org.oe.dev * rf0a8faaa... 10/ (1 site/powerpc-common): site/powerpc-common: Add va_copy entry for dpkg, fix other incorrect va_copy entry Mar 19 13:05:55 03rpurdie 07org.oe.dev * rdc81d002... 10/ (1 packages/syslog-ng/syslog-ng_2.0.5.bb): syslog-ng: Enable dynamic linking Mar 19 13:06:00 03rpurdie 07org.oe.dev * re3883b9b... 10/ (1 packages/thttpd/thttpd_2.25b.bb): thttpd: Fix .debug file packaging Mar 19 13:06:05 03rpurdie 07org.oe.dev * r1d9bbee9... 10/ (5 files in 2 dirs): quagga: Fix SRC_URIs Mar 19 13:06:10 03rpurdie 07org.oe.dev * r15351304... 10/ (1 packages/net-snmp/net-snmp_5.4.1.bb): net-snmp: Fix -dev RDEPENDS and -dbg RRECOMMENDS since doesn't exist Mar 19 13:06:15 03rpurdie 07org.oe.dev * r043da0b3... 10/ (1 packages/ncftp/ncftp_3.2.0.bb): ncftp: Fix SRC_URI Mar 19 13:06:20 03rpurdie 07org.oe.dev * r5861933e... 10/ (3 files in 2 dirs): man-pages: Fix SRC_URI, add version 2.79 Mar 19 13:06:24 03rpurdie 07org.oe.dev * rf52d7d3a... 10/ (3 files in 2 dirs): libtool: Only apply rpath-control patch to libtool-cross, not libtool-native Mar 19 13:06:29 03rpurdie 07org.oe.dev * r47cf4379... 10/ (4 files in 3 dirs): libpcre-native: Fix linking issues when the build system doesn't have pcre installed Mar 19 13:06:34 03rpurdie 07org.oe.dev * r8b250862... 10/ (1 packages/maradns/maradns_1.0.41.bb): maradns: Add 1.0.41 Mar 19 13:16:10 03rpurdie 07org.oe.dev * raaf2a55c... 10/ (1 classes/autotools.bbclass): Mar 19 13:16:10 autotools.bbclass: Sync with poky - don't add libtool-cross to sdk package Mar 19 13:16:10 depenedncies, use maintainer mode for reconfiguring packages, make sure Mar 19 13:16:10 STAGING_INCDIR exists before installing to is, fix issues with autostaging Mar 19 13:16:11 functions where .la files in subdirectories would be ignored Mar 19 13:16:20 salut Mar 19 13:18:10 bonjour zecke :) Mar 19 13:21:21 wow 73 patches a gui for git-rebase -i would be cool now Mar 19 13:22:18 ~seen thesing Mar 19 13:22:20 thesing was last seen on IRC in channel #oe, 5d 12h 44m 32s ago, saying: 'gnight'. Mar 19 13:25:40 cedric_: tried .dev branch? Mar 19 13:26:23 ant|work: sorry doing some development work... Mar 19 13:30:23 cedric_: np, I'll tryest stable later this evening Mar 19 13:30:45 s/tryest/test/ Mar 19 13:31:24 ant|work: thx Mar 19 13:31:55 Good day! I build my angstrom distro image for at91sam9263 but when i load it to board using the same offsets with original board config i get `Bad Magic Number` Mar 19 13:32:02 Do oe angstrom config do U-boot mkimage generating uImage? Mar 19 13:34:03 strikov: not automatically, it can do it though Mar 19 13:34:15 aTc: if that happens - send me one, I have a remote without the battery back cover Mar 19 13:34:17 oops wrong window :) Mar 19 13:34:54 zecke: And why it calls outbut image uImage? I think it is connected with `u-boot bootable config` Mar 19 13:35:17 strikov: ah if you have uImage in deploy, it should be a uImage :) Mar 19 13:35:59 zecke: in deploy i have uImage-2.6.20.***-blah-blah Mar 19 13:36:01 :) Mar 19 13:36:37 zeche: is it correct that Bad Magic Number it is bad u-boot header? Mar 19 13:39:11 no idea :) Mar 19 13:39:21 strikov: yes, I got it once Mar 19 13:39:37 strikov: with NO header... Mar 19 13:40:04 03jeremy_laine 07org.oe.dev * rdb59da02... 10/ (4 files in 3 dirs): libc-client: fix build error by quoting CC Mar 19 13:41:17 ant|work: u meen u get the same error when load smth without header or u get incorrect u-boot image from oe? :) Mar 19 13:42:05 strikov: u-boot did not recognize the kernel and the complaint was "bad magic number" Mar 19 13:43:37 strikov: I was doing some dirty unsupported things, and I created my uImage by hand from the zImage Mar 19 13:48:19 strikov: mkimage -C none -A arm -O linux -T kernel -a 0xA0008000 -e 0xA0008000 -n "Linux 2.6.23" -d ./zImage-2.6.23 uImage Mar 19 13:48:43 strikov: that was for a kernel image (no initramfs) Mar 19 13:49:21 ant|work: thanks! i try to do uImage myself, withou oe helper :) Mar 19 13:50:14 ant|work: with .dev and bitbake-1.8 from svn --> http://pastebin.com/m24eecd9f Mar 19 13:50:41 ant|work: sorry, but do u know where i can find all pre-date of `deploy` version? Mar 19 13:51:09 there was a gcc/uclibc related bug, that was fixed for gcc 4.2.2, woglinde pointed me to it, and damn - I forgot the number and can't find it in the tracker Mar 19 13:51:15 same happens for gcc 4.3.0 Mar 19 13:51:20 but can't find which patch it was :) Mar 19 13:51:27 does anyone remember? Mar 19 13:53:08 strikov: sorry I don't understand whath you mean with "pre-date" Mar 19 13:53:39 ant|work: e.g. zImage or U-boot pre-build mkimage utility Mar 19 13:54:19 cedric_: you need git (and perhaps psyco) Mar 19 13:55:16 strikov_: mkimage is in /work/ ..../u-boot/tools IIRC Mar 19 13:55:23 MTN_MERGE = "meld" for example Mar 19 13:55:47 XorA|gone: Thanks! I was gonna ask just such a question this morning... :) Mar 19 13:56:38 heh Mar 19 13:56:56 its all in the manual Mar 19 13:57:50 :) I know, but it's far easier to be lazy and ask... :D Mar 19 14:02:18 cedric_: pls update bitbake like described here http://www.openembedded.org/wiki/GettingStarted Mar 19 14:03:32 ant|work: that's what I've followed... Mar 19 14:03:38 cedric_: and please use /oe/bitbake (or check twice your source-me script) Mar 19 14:03:50 ant|work: yes Mar 19 14:04:07 ant|work: it's building stuff at mo Mar 19 14:07:33 cedric_: git can not be launched? install git? Mar 19 14:08:06 * * OE Bug 4101 has been created by andrew.paulsen(AT)gmail.com Mar 19 14:08:07 * * ipaq-sleep cleanup, some improvements Mar 19 14:08:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4101 Mar 19 14:08:23 zeche: thx, I've done it with ant great support! Mar 19 14:11:09 RP: should we add git to he programs in sanity.bbclass? the git-native depends is not good enough :( Mar 19 14:19:00 Does git-native not build a local version of git, then? Mar 19 14:19:57 mwester: well, bitbake when parsing is executing git (I rest my case) Mar 19 14:20:15 Ah. That would be a problem, then. Mar 19 14:20:37 actually from cedric's log we only execute git-ls-remote :) Mar 19 14:21:07 So there's really not much point in git-native, then - unless you absolutely wish to guarantee that your build envs all have the same git version. Mar 19 14:21:08 RP: touchscreen problem remains even with opie :( Mar 19 14:21:36 Khem: Opie Qt/E is using tslib as well Mar 19 14:22:14 zecke: yeah I wanted to see if this is a problem with X or gpe I tried x11-image before Mar 19 14:22:25 zecke: So something is wrong elsewhere Mar 19 14:23:01 secondly is qemu-0.9.1 working for anybody ? I am getting Could not initialise SDL errors Mar 19 14:23:20 zecke: Thats a good argument for adding git-native to assume_provided :/ Mar 19 14:23:52 Khem: Its something wrong with the environment scripts. It will probably affect all GUIs... Mar 19 14:26:19 RP: yeah so what packages should I start with Mar 19 14:26:50 RP: I tried to downgrade busybox, seems to have no effect Mar 19 14:26:51 Khem: You have a known good image and a bad one? Mar 19 14:26:58 I have both Mar 19 14:27:12 a good image is about 2-3 months old Mar 19 14:27:15 Khem: ok, so compare them. Copy bits of one into the other until it starts/stops working Mar 19 14:29:35 RP: yeah I was thinking of doing so Mar 19 14:29:53 Khem: does ts_print_raw works? Mar 19 14:30:08 a big change ofcourse is gcc 4.2 jump from gcc 4.1 Mar 19 14:30:10 Khem: if it does not works then no sense on checking x11 or opie Mar 19 14:30:40 hrw: I can see touchscreen data with od Mar 19 14:30:53 hrw: so kernel is ok I think Mar 19 14:30:54 zecke: thanks 4 your hint bout hand-made uImage - now U-boot starts to load it Mar 19 14:31:01 hrw: tslib works, the environmental variable TSLIB_TSDEVICE isn't getting set Mar 19 14:31:08 yes Mar 19 14:31:11 thats the problem Mar 19 14:31:33 I forced it in /etc/X11/Xserver script and it started to work Mar 19 14:35:59 zecke: about tzdata, don't you think a distro-postinstall-script should create /etc/localtime ? As it is, the user has to do it... Mar 19 14:36:22 Khem: you have /dev/input/touchscreen0 of course? Mar 19 14:36:35 yes Mar 19 14:36:46 its a link to /dev/input/event1 Mar 19 14:36:52 I am on qemu Mar 19 14:36:56 hrw: The problem is the environment, that much has been established :/ Mar 19 14:37:25 RP: ok, but /etc/profile.d/tslib.sh is supposed to set TSLIB_TSDEVICE if /dev/input/touchscreen0 is present Mar 19 14:38:27 03jeremy_laine 07org.oe.dev * r769fde4a... 10/ (1 packages/linux/linux-2.6.24/mpc8313e-rdb/defconfig): linux-2.6.24: build TUN/TAP module for mpc8313e-rdb Mar 19 14:38:30 hrw: I see what you're thinking. Mar 19 14:38:54 hrw: here is another experiment I did on a system which works, I removed tslib.sh and started a new shell it did not have TSLIB_TSDEVICE set so it means everytime I start a new shell it reads /etc/profile.d/tslib.sh Mar 19 14:40:15 Khem: /etc/profile do run-parts on /etc/profile.d/ Mar 19 14:40:38 zeche: NOTE: package glibc-2.6.1: completed Mar 19 14:40:40 but in bad case this variable is not set in env when I start new shell even tslib.sh exists Mar 19 14:40:56 ant|work: NOTE: package glibc-2.6.1: completed Mar 19 14:41:03 Khem: even when you "source /etc/profile.d/tslib.sh"? Mar 19 14:41:05 =) Mar 19 14:41:29 hrw: no when I manually source it then it works Mar 19 14:42:04 which busybox? Mar 19 14:43:34 1.9.1 Mar 19 14:43:55 hrw: Same busybox works with the existing working env Mar 19 14:44:04 no idea then Mar 19 14:44:10 so I guess its not busybox hopefully Mar 19 14:44:32 etc/profile comes from busybox ? Mar 19 14:47:25 oday Mar 19 14:49:31 U-Boot loads kernel OK but after `Starting kernel ...` a get nothing. Can u presume why? Mar 19 14:51:28 florian: before FOSDEM I mentioned that I wanted to do some work on ipaq-sleep. I got some of it done and it appears to be in a useable state. I put it in oe bug 4101 if you want to take a look. Mar 19 14:51:58 apaulsen: ah cool, that's good news - i'll take a look Mar 19 14:54:57 strikov: the kernel should be decompressed at that point... Mar 19 14:55:06 strikov: wrong entry address? Mar 19 14:55:49 strikov: look if you have "ADRESS" (wrong) or ADDRESS in your conf Mar 19 14:55:52 hrw: I search for any file thats calling tslib.sh in /etc and I did not find any Mar 19 14:56:11 Khem: /etc/profile do that Mar 19 14:56:48 hrw ok at what point etc/profile is called during boot ? Mar 19 14:56:57 Load Adress / EntryPoint is 20008000 -> my SDRAM adress space -> I think its ok? Mar 19 14:57:32 Khem: x11 session scripts? Mar 19 14:57:55 hrw: in opie ? Mar 19 14:58:37 then opie initscript Mar 19 14:59:51 hrw: ok I see in K20opie Mar 19 15:00:02 ant|work: and what conf u meen? Mar 19 15:00:17 Khem: no. you see it in /etc/init.d/opie Mar 19 15:00:32 strikov: machine.conf? Mar 19 15:00:36 I really do not understand why people look in /etc/rc*.d/ instead of /etc/init.d/ Mar 19 15:03:50 ant|work: i had only /build/conf/local.conf another .conf i think comes from conf/machine/at91sam*** but there is no ADDRESS string there Mar 19 15:04:27 but there are UBOOT* vars? Mar 19 15:08:10 apaulsen: ... as soon as the bugtracker decides to reply Mar 19 15:09:23 florian: it works for me... http://bugs.openembedded.net/show_bug.cgi?id=4101 Mar 19 15:10:37 apaulsen: i just tried to log in and it takes ages... Mar 19 15:11:11 florian: no hurry, I've been sitting on this code for a few weeks anyway :) Mar 19 15:15:06 I updated the tree and now gui doesnt even start forget touchscreen Mar 19 15:15:45 apaulsen: :) Mar 19 15:20:10 strikov: lookinf for u-boot and at91sam http://www.linux4sam.org/twiki/bin/view/Linux4SAM/U-Boot#U_boot_Main_Commands Mar 19 15:20:47 strikov: boot_addr 0x21400000 Mar 19 15:23:33 ant|work: sorry I'm getting the same stupid "locale es_NI (UTF-8)" in .dev (glibc-2.6.1), the previous build was maybe the native one or something like that... Mar 19 15:26:13 ant|work: this phase is OK i think, U-Boot find my kernel image and uncompress it to 20008000 Mar 19 15:26:38 ant|work: but then don't wonna go to entry point as i think Mar 19 15:27:28 03koen 07org.oe.dev * r358ee536... 10/ (1 classes/autotools.bbclass classes/native.bbclass): native and autotools bbclass: use autotools_stage_all instead of oe_runmake install for staging native packages Mar 19 15:40:50 cedric_: that's odd, gcc-native workaround was ok a few days ago. I'll check again and if it repeats we'll have to bother RP ;-) Mar 19 15:41:33 ant|work: ok Mar 19 15:42:30 Please tell me one fundamental thing (or a link, please) how can i tell to kernel that i have some rootfs image in SDRAM already? How kernel parse memory for data? Mar 19 15:45:25 03koen 07org.oe.dev * rc31a0840... 10/ (1 conf/distro/include/angstrom.inc): angstrom: set QA_LOG to 1 Mar 19 15:45:30 03koen 07org.oe.dev * r440982b1... 10/ (1 packages/gettext/gettext_0.14.1.bb): gettext: switch to autotools_stage_all Mar 19 15:53:00 strikov: this is a log of u-boot booting http://bugs.openembedded.org/attachment.cgi?id=5189 Mar 19 15:53:22 strikov: do you reach Uncompressing Linux... ? Mar 19 15:57:29 03rpurdie 07org.oe.dev * r8438ae3b... 10/ (1 classes/insane.bbclass): insane.bbclass: Only check installed status for packages which aren't native/cross, find desktop-file-validate from PATH so it can be ASSUME_PROVIDED (from poky) Mar 19 16:03:52 bye Mar 19 16:03:53 ant|work: we've solved this problem together :) thanks....problem was in linux4sam project config! Mar 19 16:08:13 can someone solve my problem? this is my configure.log: http://nopaste.org/p/aKmUHY5tn Mar 19 16:08:31 s/configure.log/config.log Mar 19 16:10:12 ant|work: now, i'm trying to set where in DRAM kernel should locate FS, somehow Mar 19 16:15:38 strikov: cmdline is compiled in the u-boot sources usually Mar 19 16:16:05 has anybody here been able to compile lzo-native on x86_64 lately? Mar 19 16:18:01 ant|work: I see but what should i send as root=... if i wanna mount SDRAM image (e.g. from 0x00 to 0x10) Mar 19 16:18:37 it fails with: x86_64-linux-libtool: link: `lzo1x_d3.lo' is not a valid libtool object Mar 19 16:19:16 strikov: I think you could create a kernel+initramfs uImage Mar 19 16:19:29 does anyone know when the packaged-staging status reported on the list will go in .dev? Mar 19 16:20:05 (full log @ http://www.pastebin.ca/949072) Mar 19 16:21:04 hi gang - just double checking - in bitbake, the form "var ?= value " means assign if var is not already assigned, right? Mar 19 16:24:12 yes Mar 19 16:24:37 hrw, thanks. Mar 19 16:24:45 strikov: http://www.denx.de/wiki/view/DULG/KernelCrashesWithRamdisk and http://www.denx.de/wiki/DULG/UBootEnvVariables -> initrd_high Mar 19 16:24:48 ant|work: can u give me short step list to do it from builded .jffs2 image and initramfs-less uImage Mar 19 16:32:58 anyone toolchain experts around? Mar 19 16:32:59 http://pastebin.mozilla.org/373316 Mar 19 16:33:16 with gcc-4.3.3 I get an error that limits.h is missing and indeed the layout in cross is a little different Mar 19 16:33:21 what is this include-fixed directory?Y Mar 19 16:35:28 is it really supposed to be /lib/gcc/mips-angstrom-linux-uclibc/4.3.0/include-fixed/limits.h Mar 19 16:35:29 ? Mar 19 16:35:51 ...compare gcc 4.2.2: ./lib/gcc/mips-angstrom-linux-uclibc/4.2.2/include/limits.h Mar 19 16:37:57 4.3.3? Mar 19 16:38:25 oops 4.3.0 Mar 19 16:38:41 did I say 4.3.3? Mar 19 16:39:02 19 17:38 < Jin^eLD> with gcc-4.3.3 I get an error that limits.h is missing and indeed the layout in cross is a little different Mar 19 16:39:12 oh yeah, sorry :) s/gcc-4.3.3/4.3.0/g :> Mar 19 16:39:28 crap, even my sed thingie is wrong ;) Mar 19 16:39:28 anyway Mar 19 16:39:59 any idea if the layout in cross should be different i.e. include instead of include-fixed or if another -I should be added to uclibc? Mar 19 16:47:58 ant|work: thanks, i'am reading :) Mar 19 16:50:05 * * OE Bug 4102 has been created by  Mar 19 16:50:07 * * sdcc-native-2.5.0-autobuild Mar 19 16:50:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4102 Mar 19 16:52:18 strikov: http://www.denx.de/wiki/view/DULG/RootFileSystemDesignAndBuilding Mar 19 16:57:34 ah opie boots now someone fixed something for me I owe someone a beer Mar 19 17:00:54 bbl Mar 19 17:09:50 03rpurdie 07org.oe.dev * rf2f68aaf... 10/ (3 files in 3 dirs): libxslt: Remove Koen's gross hack and replace with a proper patch Mar 19 17:10:21 Editorial comments in the changelog :D Mar 19 17:10:45 I should probably go see what folks had to say as they corrected some of my messes... :D Mar 19 17:10:56 mwester: heh... koen can take it :) Mar 19 17:12:13 Anybody have any ideas on why lzo-native fails: http://www.pastebin.ca/949072 ? Mar 19 17:12:55 hvontres|work: same problem with pkg-config-native here. Mar 19 17:13:07 Seems that autoconf-native did not stage properly. Mar 19 17:14:12 dottedmag: hmmm, thanks. I think koen pushed a fix for that today... Mar 19 17:14:39 hi all Mar 19 17:14:45 hvontres|work: I'm trying to make it work on current tip, still no luck. Mar 19 17:15:02 * hvontres|work will need to pull again tonight Mar 19 17:30:09 oh... mtn pull ... lol Mar 19 17:47:45 ant|work: Thanks!!! Mar 19 18:04:55 ant|work: Article is really very informative, but is it possible to mount `already in SDRAM` (using u-boot) image? Mar 19 18:05:12 bye Mar 19 18:14:32 oh hmmm downgrading busybox and now I see that touch screen it working Mar 19 18:21:02 re Mar 19 18:27:38 bye Mar 19 19:41:37 03koen 07org.oe.dev * r04060d9b... 10/ (1 classes/autotools.bbclass classes/native.bbclass): autotools, native bbclass: try harder to stage native binaries Mar 19 19:57:41 What causes this error: I try to run "/bin/kill.procps" on my target and get "ERROR: no "kill.procps" support.". ("/bin/ps.procps" works, btw). Mar 19 20:00:09 likewise: perhaps the binary is one of those that looks at argv[0 Mar 19 20:00:10 I have initrd image, ready for u-boot (using mkimage), how can i use it during Linux kernel boot? Mar 19 20:00:16 ] to see what it should do? Mar 19 20:00:43 i load kernel from NAND to SDRAM and then boot it, should i copy initrd from RAM too? Mar 19 20:02:25 mwester: thanks. Mar 19 20:02:56 mwester: also, the shell had a "kill" command, so I had three kills: busybox, procps and shell's... Mar 19 20:03:13 You're well-armed,then. Mar 19 20:03:29 (or would that be "well-ARMed"... Mar 19 20:03:31 ) Mar 19 20:05:10 oh, geeze Mar 19 20:19:35 mwester: well-POWERed, actually :-) Mar 19 20:19:50 ... Mar 19 20:21:10 night all Mar 19 20:21:18 03koen 07org.oe.dev * rbc4a8e8d... 10/ (1 classes/autotools.bbclass classes/native.bbclass): Mar 19 20:21:18 native, autotools bbclass: check if bindir exists before trying to access it Mar 19 20:21:18 * hardcoded since native.bbclass redefines ${bindir} and friends :( Mar 19 20:21:53 Is something like http://cgi.ebay.de/NEU-NAS-Server-3-5-HDD-GEHAUSE-SAMBA-FTP-USB-2-0-LAN_W0QQitemZ200205555507QQihZ010QQcategoryZ139518QQssPageNameZWDVWQQrdZ1QQcmdZViewItem likely hackable? Mar 19 20:22:00 serial console? reflashable? Mar 19 20:24:00 Laibsch: I would open the case and look at the content of the HDD on another computer Mar 19 20:27:39 Laibsch: english description of it anywhere? Mar 19 20:27:55 (my German is not very good) Mar 19 20:37:06 * * OE Bug 4103 has been created by  Mar 19 20:37:08 * * epdf-0.1.0+cvs20071206-autobuild Mar 19 20:37:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4103 Mar 19 20:43:01 03koen 07org.oe.dev * r3868e38f... 10/ (1 classes/autotools.bbclass): autotools bbclass: sync with poky and expand on it Mar 19 20:43:25 ok, who broke gettext-native Mar 19 20:44:10 ScaredyCat: My guess would be it's related to packaged staging and autotools :) Mar 19 20:44:34 :) Mar 19 20:44:48 ScaredyCat: many native packages are broken Mar 19 20:45:08 again? Mar 19 20:45:12 or still Mar 19 20:45:31 Yes. Every time I rebuild from scratch after pull today it breaks earlier and earlier :) Mar 19 20:46:05 btw, I have several typos silently breaking things fixed. How do I communicate them back to the devs? Mar 19 20:46:11 ScaredyCat: The commit from koen above should help Mar 19 20:46:28 Something went in which broke things quite badly and unexpectedly :( Mar 19 20:46:29 dottedmag: bugzilla is a good way Mar 19 20:46:35 good stuff RP Mar 19 20:46:51 about the fix, obviously, not the breaking Mar 19 20:49:53 Laibsch: It might be...unless they used OTP parts instead of flash.... Mar 19 20:52:47 chouimat|work, mwester, hvontres|work: Thanks for the replies Mar 19 20:52:53 I don't own the device yet Mar 19 20:53:14 It is fairly cheap and therefore I will likely buy it even if non-hackable Mar 19 20:53:34 how much? Mar 19 20:54:07 If it was hackable, that would be just fantastic (adding NFS, for example and a couple of command line programs/daemons) Mar 19 20:54:18 chouimat|work: depends on the auction price of course Mar 19 20:54:23 but around 25$ Mar 19 20:54:26 35$ Mar 19 20:54:38 hmmm Mar 19 20:55:02 * hvontres|work thinks the dollar is dropping faster and faster these days :( Mar 19 20:55:20 one pays about that much for an external HDD case, this one has the added benefit of NAS Mar 19 20:55:47 So if the device is non-hackable, it is just an external HDD case to me Mar 19 21:13:32 Anyone got opkg-native built? Mar 19 21:14:18 It breaks for me on do_fetch, because can't do 'svn co -r 1 http://svn.openmoko.org/trunk/src/target/opkg' - and yes, there was not opkg at revision 1 in this repo. Mar 19 21:15:07 * * OE Bug 4104 has been created by  Mar 19 21:15:09 * * openmoko-dialer2-0.1.0+svnr3902-autobuild Mar 19 21:15:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4104 Mar 19 21:22:18 03koen 07org.oe.dev * re9a35786... 10/ (1 packages/openmoko-projects/illume_svn.bb): illume: fix PV Mar 19 21:56:24 03koen 07org.oe.dev * r93a53a4a... 10/ (1 classes/native.bbclass): native.bbclass: fix ] spacing Mar 19 22:24:58 03koen 07org.oe.dev * r0b604857... 10/ (1 classes/insane.bbclass): Mar 19 22:24:58 insane bbclass: add '8' (.la references workdir) to list of fatal errors Mar 19 22:24:58 * this will also trigger on broken libtool archives already present in staging. If you get hit by this: rebuild the offending recipes or do a clean rebuild Mar 19 22:26:12 no wonder koen hates libtool :) Mar 19 22:27:06 * * OE Bug 4105 has been created by  Mar 19 22:27:08 * * libconic-0.13-autobuild Mar 19 22:27:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4105 Mar 19 22:43:57 I guess I'll just keep waiting a few more hours before I just sync up and rebuild the world... Mar 19 23:34:43 Hi all, I'm trying to build a console-image with Angstrom 2008.01, but kernel fails requesting for a strange deploy/uclibc/image, though I'm using glibc ... Mar 19 23:42:11 ljp: hey Mar 19 23:45:49 hi zecke, probably u might be of help Mar 19 23:46:01 I'm trying to build a console-image with Angstrom 2008.01, but kernel fails requesting for a strange deploy/uclibc/image, though I'm using glibc ... Mar 19 23:46:02 not in my state *tired* :) Mar 19 23:46:07 lol Mar 19 23:46:31 tsdogs: what is the kernel requesting? Mar 19 23:46:31 lisppaste7: url Mar 19 23:46:31 To use the lisppaste bot, visit http://paste.lisp.org/new/oe and enter your paste. Mar 19 23:48:12 tsdogs pasted "uclibc kernel ?" at http://paste.lisp.org/display/57683 Mar 19 23:49:46 glibc_2.6.1.bb failed Mar 19 23:50:16 ho Mar 19 23:50:57 ups, I guess I'm more tired than u :) thx Mar 19 23:51:17 tsdogs, what machine? Mar 19 23:51:32 Fedora 8 x86_64 Mar 19 23:51:50 no, what is MACHINE in local.conf? Mar 19 23:52:08 ho htcuniversal Mar 19 23:52:21 sorry I'm not familiar with that one Mar 19 23:53:15 no prob, I'll check why it didn't build glibc Mar 19 23:55:52 hmm only error I find is Mar 19 23:55:55 log.do_install.23705:make[2]: [/home/alex/PDA/OE/build/tmp/work/armv5te-angstrom-linux-gnueabi/glibc-2.6.1-r3/image/usr/libexec/pt_chown] Error 1 (ignored) Mar 19 23:56:32 but it might be related to the fact that during the build I interrupted it and then restarted it later... I'll try rebuilding glibc Mar 20 00:06:43 ljp: eek, QPasswordDialog is lacking a d-pointer :( Mar 20 00:24:20 n8 Mar 20 00:27:21 zecke: ! Mar 20 00:27:23 g'niht Mar 20 00:48:20 * mwester , with great trepidation, syncs up, deletes his TMPDIRs, and begins a full rebuild after todays commits... Mar 20 00:48:34 good luck :) Mar 20 01:45:25 tsdogs: The glibc error you posted is not your problem Mar 20 01:45:32 tsdogs: this error is ignored Mar 20 01:45:39 there is something else going on Mar 20 01:55:11 morning Mar 20 02:13:09 XorA: morning Mar 20 02:13:50 yo Fique **** ENDING LOGGING AT Thu Mar 20 02:59:56 2008