**** BEGIN LOGGING AT Sat Oct 02 02:59:57 2010 Oct 02 03:01:50 greetings internet pals. I'm having some difficulty with my image.. I'm finding that even though I force a 'bitbake -c distclean ', it's still picking up 'older' files. Am I missing something critical? I want to make sure that all the packages (including updated ones) are extracted to the staging fs so it'll pick up all the changes Oct 02 06:38:00 gm Oct 02 06:38:09 cd monotone Oct 02 06:43:46 oops, wrong place, system is too slow to keep up wiht me :-) Oct 02 06:47:02 03Frans Meulenbroeks  07org.openembedded.dev * r2384fbbcf7 10openembedded.git/recipes/monotone/ (5 files in 2 dirs): Oct 02 06:47:02 monotone-6: cleaned old versions Oct 02 06:47:02 as discussed on ML Oct 02 06:47:02 Signed-off-by: Frans Meulenbroeks Oct 02 07:11:35 03Khem Raj  07master * r872d9f9545 10openembedded.git/recipes/ca-certificates/ca-certificates_20090814.bb: Oct 02 07:11:35 ca-certificates: Rename which updates SRC_URI to fetchable tar. Oct 02 07:11:35 Signed-off-by: Khem Raj Oct 02 07:22:16 03Frans Meulenbroeks  07org.openembedded.dev * r6f27f94353 10openembedded.git/recipes/konqueror/konqueror-embedded-20030705/ (flashpart.patch kcookiejar-merge.patch.patch): Oct 02 07:22:17 konqueror : moved unused files to obsolete dir Oct 02 07:22:17 Signed-off-by: Frans Meulenbroeks Oct 02 07:22:22 ok todays clenaing work coming up Oct 02 07:22:39 03Frans Meulenbroeks  07org.openembedded.dev * ra71960e82d 10openembedded.git/recipes/schroedinger/files/Makefile.patch: Oct 02 07:22:39 schroedinger : moved unused files to obsolete dir Oct 02 07:22:39 Signed-off-by: Frans Meulenbroeks Oct 02 07:22:39 03Frans Meulenbroeks  07org.openembedded.dev * rb7c5f0a573 10openembedded.git/recipes/maemo-mapper/files/mapper-hildon1.diff: Oct 02 07:22:39 maemo-mapper/ : moved unused files to obsolete dir Oct 02 07:22:40 Signed-off-by: Frans Meulenbroeks Oct 02 07:22:43 03Frans Meulenbroeks  07org.openembedded.dev * r11840b3ac7 10openembedded.git/recipes/hal/files/ (autoconf.diff dbus-fix-api.patch): Oct 02 07:22:44 hal : moved unused files to obsolete dir Oct 02 07:22:44 Signed-off-by: Frans Meulenbroeks Oct 02 07:22:44 03Frans Meulenbroeks  07org.openembedded.dev * r64cafc7905 10openembedded.git/recipes/rp-pppoe/rp-pppoe-3.8/pppoe-no-timeout.patch: Oct 02 07:22:44 rp-pppoe : moved unused files to obsolete dir Oct 02 07:22:44 Signed-off-by: Frans Meulenbroeks Oct 02 07:22:48 03Frans Meulenbroeks  07org.openembedded.dev * r94dc66b29d 10openembedded.git/recipes/schroedinger/files/configure.ac.patch: Oct 02 07:22:49 schroedinger: moved files dir to more appropriate schroedinger Oct 02 07:22:49 Signed-off-by: Frans Meulenbroeks Oct 02 07:22:49 03Frans Meulenbroeks  07org.openembedded.dev * rc8fb5c6bf7 10openembedded.git/recipes/mamona/gcc-noemu-4.1.2/ (801-arm-bigendian-eabi.patch zecke-no-host-includes.patch): Oct 02 07:22:49 mamona : moved unused files to obsolete dir Oct 02 07:22:49 Signed-off-by: Frans Meulenbroeks Oct 02 07:22:50 03Frans Meulenbroeks  07org.openembedded.dev * raf123cfea4 10openembedded.git/recipes/rp-pppoe/ (4 files in 2 dirs): Oct 02 07:22:50 rp-pppoe: removed version 3.5 Oct 02 07:22:51 this one did not fetch, src is not at DL site Oct 02 07:22:51 we still have 3.8 (which fetches ok) Oct 02 07:23:39 BTW: upstream is at 3.10 maybe we should upgrade Oct 02 07:23:39 Signed-off-by: Frans Meulenbroeks Oct 02 07:23:39 03Frans Meulenbroeks  07org.openembedded.dev * r9b973e6624 10openembedded.git/recipes/matchbox-panel/matchbox-panel-0.9.2/kernel2.6.patch: Oct 02 07:23:39 matchbox-panel : moved unused files to obsolete dir Oct 02 07:23:39 Signed-off-by: Frans Meulenbroeks Oct 02 07:23:39 (9 lines omitted) Oct 02 07:29:19 hi Oct 02 07:29:41 03Khem Raj  07master * rfff02e9a3b 10openembedded.git/recipes/libspiff/ (files/gcc-4.5-compile.patch libspiff_0.8.2.bb): Oct 02 07:29:41 libspiff_0.8.2.bb: Fix C++ files to compile with gcc 4.5 Oct 02 07:29:41 Signed-off-by: Khem Raj Oct 02 07:29:49 khem: I also have this bug : https://bugs.launchpad.net/gcc-linaro/+bug/653316 what is the best workaround you suggest ? Oct 02 07:30:08 ericben: use -Os Oct 02 07:30:25 for time being Oct 02 07:30:55 khem: ok. For the whole distro or only for the failing packages ? Oct 02 07:31:07 only for failing package Oct 02 07:31:29 ok at least udev is failing here Oct 02 07:31:32 you can change makefiles to just change for one file too Oct 02 07:31:36 yes Oct 02 07:31:41 that way you dont lose much Oct 02 07:31:54 khem, looks like you fixed the bug that was locking up garnet Oct 02 07:32:07 ka6sox: yes Oct 02 07:36:50 in an amend.inc : can I do something to get PR = $PR + 1 ? Oct 02 07:41:29 what I'm doing in PR .= "1" but this may not be the best way to handle PR in amend ? Oct 02 07:41:55 ericben: I dont think you can increment it Oct 02 07:42:08 you can override it to something of your own Oct 02 07:43:15 khem: yes, in fact doing .="1" allow my recipe to be able to upgrade when mainoing recipe gets an upgrade Oct 02 07:43:25 mainoing => mainline Oct 02 07:45:19 khem: OK only udev was blocking my image form armv6 with gcc-4.5 (even if udev is not installed on the image), amending its CFLAGS with -OS solved the problem Oct 02 07:45:23 many thanks Oct 02 07:46:30 for the records : it seems gcc 4.3.3 compiles a kernel which generetes data abort & VFP problems on an arm1136jf-s Oct 02 07:47:49 btw, khem, the other day I told you of problems with glibc-initial on a slackware and you suggested me to check the make version : you were right, he was using 3.82 which exhibits these problems, siwtching back to 3.81 solved the thing. Is there a way to have OE check for versions installed on the host ? Oct 02 07:48:19 actually I'm doing this in a shell script which fetches and prepare OE Oct 02 07:51:09 ericben: there are patches floating around to standardize on make version Oct 02 07:51:25 ericben: soon I will fix eglibc to be able to compile with make 3.82 Oct 02 07:51:58 ericben: 4.3.3 compiled kernel fails how about if you compile it with 4.5 Oct 02 07:52:10 khem: actually testing Oct 02 07:52:39 Codesourcery's latest toolchain build a kernel which works fine (they are using gcc 4.4.1) Oct 02 07:52:56 and now I'm testing the kernel compiled with 4.5 Oct 02 07:53:09 ericben: OEs 4.5 is what I am interested in Oct 02 07:53:15 but the problem is not easyu to exhibit as I have to make lot of things on the touchscreen to reproduce it Oct 02 07:53:30 hmmm Oct 02 07:54:06 and for the VFP problem it's quite easy to reproduce with qtdemo : it fails when calculating the animations Oct 02 07:54:40 ericben: hmmm it could be hardware issue too Oct 02 07:55:38 khem: : I've been searching for hardware issues for 3 weeks wthout success. Compiling the kernel with Codesourcery toolchain solves the problem as it has nevere crashed since Oct 02 07:56:08 maybe 4.3.3 doesn't handle all the 1136 bugs ? Oct 02 07:56:15 ericben: ok could be Oct 02 07:56:21 or it's a binutils problem Oct 02 07:56:32 ericben: vfp was still new for gcc 4.3 Oct 02 07:58:08 yes, but in fact here are 2 problems : the VFP one (exhibited by qtdemo animations and "solved" by using tune-arm1136-novfp.inc), an the data abort one (triggered when doing events on the touchscreen). Btw, same touchscreen driver works fine on an ARMV5. Oct 02 07:58:12 ericben: btw if you have any benchmarks that would be awesome to see if it improves with gcc 4.5 Oct 02 07:58:28 khem: will try to setup something asap Oct 02 07:59:26 novfp is disabling it Oct 02 07:59:41 yes, that's why it "solves" the VFP problem :-) Oct 02 08:00:06 ok testing kernel compiled with "gcc version 4.5.2 20100923 (prerelease) (GCC)" Oct 02 08:01:31 hmmm more fun now with the board : qtdemo runs but getty doesn't bring the login prompt any more ! Oct 02 08:03:32 I'm still using the rootfs compiled with gcc 4.3.3, now switching to the new one Oct 02 08:08:25 same thing, OK gcc 4.5 seem to trigger a bug somewhere else Oct 02 08:43:16 ok don't know yet why but getty is now using hardwre flow control on the serial port ! Oct 02 08:43:45 kernel comes from freescale so all the drivers are not really production ready Oct 02 08:45:28 03Holger Hans Peter Freyther  07master * r3ee6ae9e90 10openembedded.git/recipes/qt4/qt4-tools-native.inc: Oct 02 08:45:28 qt4-tools-native.inc: Add more -no options to disable dependencies Oct 02 08:45:28 Make sure we have less dependencies and use less things from Oct 02 08:45:28 the system. For moc/rcc/uic we need the most basic Qt, even Oct 02 08:45:28 for the dbus bindings one mostly needs the xml parser. Oct 02 08:45:32 03Holger Hans Peter Freyther  07master * rf559b19255 10openembedded.git/recipes/qt4/qt4-tools-sdk.inc: Oct 02 08:45:32 qt4-toolsd-sdk.inc: Move to -embedded, disable more features Oct 02 08:45:32 Pass more -no or -qt- options to make sure less things from Oct 02 08:45:32 the system are detected and that there are less dependencies. Oct 02 08:45:32 This disabling has no effect on the tools we want to use. Oct 02 08:45:33 03Holger Hans Peter Freyther  07master * r2edb621ead 10openembedded.git/recipes/qt4/ (qt4-tools-sdk.inc qt4-tools-sdk_4.7.0.bb): Oct 02 08:45:36 qt4-tools-sdk.inc: Switch over to INC_PR Oct 02 08:45:36 Right now only qt4-tools-sdk_4.7.0 use the inc file. Oct 02 08:45:36 03Holger Hans Peter Freyther  07master * rc5b0785812 10openembedded.git/recipes/tasks/task-qte-toolchain-target.bb: Oct 02 08:45:36 task-qte-toolchain-target: Attempt to install QtDeclarative Oct 02 08:45:37 This module is new in Qt 4.7 so put it into the recommendation Oct 02 08:45:37 and hope that it will be installed (it is now). Oct 02 08:56:31 khem: gcc 4.5 + binutils 2.20.1 seems to fix the problem Oct 02 09:01:02 hmm insane.bbclass fails to find bad rpath Oct 02 09:15:10 ok the flow control problem for getty is caused by Freescale's serial driver ... it works fine in mainline kernel Oct 02 09:46:03 libtool is creating so much more problems than it is solving... Oct 02 09:47:11 what's new ? Oct 02 09:47:36 libneon is... Oct 02 09:47:51 if I remove -rpath $(libdir)... it stops creating the .so Oct 02 09:48:22 well can't help you with that, don't know too much about libtool and prefer to keep it that way :-) sry Oct 02 09:58:53 Hm, I wonder how to do correctly tweaks in the image like custom /etc/modutils, /etc/network/interfaces etc. Oct 02 09:59:27 /etc/modutils = /etc/modules Oct 02 10:00:36 I don't want to create my own distro just for this features, but I would like to have as small difference to upstream as possible Oct 02 10:01:04 is creating myimage-configure package with all this tweaks the only option? Oct 02 10:01:25 ynezz: here 'm using an overlay with amend.inc Oct 02 10:02:02 forget to mention, no overlays please :) Oct 02 10:02:20 what is the reason for overlays if there's git and branches Oct 02 10:02:37 what's amend.inc ? Oct 02 10:02:59 well when you have something not mailined I find it easier to support in an overlay using amend.inc thant with git branches Oct 02 10:03:10 INHERIT += "amend" Oct 02 10:03:22 COLLECTIONS = "bla1 bla2 bla3" Oct 02 10:03:27 FILESPATHBASE =. "${@ ':'.join([os.path.join(recipedir, os.path.basename(os.path.dirname( d.getVar('FILE Oct 02 10:03:29 hm, hear about it first time :) Oct 02 10:03:30 ', 1)))) for recipedir in d.getVar('COLLECTIONS', 1).split()])}:" Oct 02 10:03:50 thanks Oct 02 10:04:21 with both these line : you can use amend.inc and if a file in in your overlay/recipes/initscripts/initscripts-1.0/machine_name/ Oct 02 10:04:29 the oe's initscript recipe will take it Oct 02 10:04:50 amend.inc is useful to tweak a recipe without modifying it Oct 02 10:05:08 yes, that's exactly what I want Oct 02 10:05:27 example : I have a busybox recipe where I wan to add a script in etc/mdev Oct 02 10:06:04 in my overly I have recipes/busybox/amend.inc + recipes/busybox/usb.sh Oct 02 10:06:11 and amend.inc is : Oct 02 10:06:28 PR .= "1" (to be able to upgrade packages) Oct 02 10:06:41 oh, now I remember why I dislike overlays, back in the days, every time I've edited recipe in overlay, bitbake found then cache as dirty and reparsed it, which took ages Oct 02 10:06:44 SRC_URI_append_mymachine = "file://mdev.conf file://usb.sh" Oct 02 10:06:58 COMPATIBLE_MACHINE= "(mymachine)" Oct 02 10:07:07 PACKAGE_ARCH = "$mymachine}" Oct 02 10:07:31 that's great info, you should put it on wiki also :) Oct 02 10:07:32 do_install_append_mymachine () { install my mdev.conf and usb.sh } Oct 02 10:07:54 this way if mainline busybox upgrades, my machine's busybox also upgrades Oct 02 10:08:28 thanks, I do get it now and it's exactly what I was looking for Oct 02 10:08:29 andinsead of having r18, I have r181 which keeps the ascending upgrade possible Oct 02 10:08:41 you mean 18.1 Oct 02 10:08:46 adn when mainline moves to r19, I move to r191 Oct 02 10:09:00 no .= '1" concatenate the PR Oct 02 10:09:02 ah it's PR .= not PR = ".1" Oct 02 10:09:08 overlooked it Oct 02 10:09:20 doing .1 made the mess with ditro pr increase Oct 02 10:09:26 ah Oct 02 10:09:31 as the package was r18.1.3 Oct 02 10:09:39 or something like this Oct 02 10:26:43 ericben: this should be enough? http://pastebin.com/aEYWmzM9 Oct 02 10:27:15 ericben: or do I need to specify collection as overlay/recipes/*/*.bb ? Oct 02 10:27:54 here I have : Oct 02 10:28:11 BBFILES = "openembedded/recipes/*/*.bb" Oct 02 10:28:21 BBFILES += "mymachine/recipes/*/*.bb" Oct 02 10:28:28 BBFILE_COLLECTIONS = "openembedded mymachine" Oct 02 10:28:37 BBFILE_PATTERN_openembedded = "openembedded/recipes" Oct 02 10:28:46 BBFILE_PATTERN_mymachine = "mymachine/recipes" Oct 02 10:28:51 BBFILE_PRIORITY_openembedded = 5 Oct 02 10:28:57 BBFILE_PRIORITY_mymachine = 10 Oct 02 10:29:07 ah, cool Oct 02 10:29:09 COLLECTIONS = "${BBFILE_PATTERN_mymachine" Oct 02 10:29:19 FILESPATHBASE =. "${@ ':'.join([os.path.join(recipedir, os.path.basename(os.path.dirname( d.getVar('FILE', 1)))) for recipedir in d.getVar('COLLECTIONS', 1).split()])}:" Oct 02 10:29:37 the OE changed a lot since I've used it last time :) Oct 02 10:30:06 well it's changing a lot each weeks :-) Oct 02 10:30:23 looks so Oct 02 10:30:36 even the mailing list is high traffic now Oct 02 10:33:50 and for BBFILES you don't need the whole path now? Oct 02 10:35:01 ynezz: yes in fact I have a shell variable before what I wrote here : ${MYPROJECTBASE} Oct 02 10:35:12 for BBFILES Oct 02 10:35:45 ok Oct 02 10:46:16 * mrcan is away: Gone away for now Oct 02 10:50:08 ericben: works, thanks a bunch! Oct 02 10:51:38 * mrcan is back. Oct 02 10:55:36 hm, now I wonder how to update /etc/modules using amend.inc Oct 02 10:56:38 it's referrenced in modutils/modutils-collateral.bb, but nor task-base or task-boot is including it Oct 02 10:57:41 in your overlay do a recipeds/modutils/files/youmachine directory and put your modules file inside Oct 02 10:58:01 then rebuild the recipe and it should get your file Oct 02 11:02:29 just that easy?, hm Oct 02 11:03:24 but I should clean/build the package again, right? Oct 02 11:32:26 ynezz: yes clean build Oct 02 11:46:22 is anyone here aware of gcc options for floating point on ARM ? Oct 02 11:46:41 in utne-arm1136jf-s.inc, we have TARGET_CC_ARCH = "-march=armv6j -mtune=arm1136jf-s -mfpu=vfp -mfloat-abi=softfp" Oct 02 11:47:12 but ... if I don't enable floating point emulation in the kernel I sometime get an undefined instruction with QTdemo Oct 02 11:47:22 which seems to mean, qtdemo includes floating point instructions Oct 02 11:48:02 my understanding of -mfpu=vfp -mfloat-abi=softfp is that what can be handled by the vfp goes to the vfp and what the vfp can't handle is in soft float Oct 02 11:48:16 and thus should not use the kernel floating point emulation Oct 02 11:48:23 Am I totally wrong here ? Oct 02 11:49:09 btw this it not directly OE related except if there should'nt be any floating point instruction in the code compiled with tune-arm1136jf-s.inc Oct 02 11:51:09 mmm seems I misunderstood -mfloat-abi=softfp Oct 02 11:51:22 and mixed with -mfloat-abi=soft Oct 02 11:56:02 mmmm still crashed even with kernel FP emulation. This took more time this time ... strange ... Oct 02 12:44:53 03Koen Kooi  07org.openembedded.dev * rc2761b2576 10openembedded.git/contrib/angstrom/sort.sh: angstrom feed sorter: add am389x-evm and c6a816x-evm Oct 02 12:44:53 03Koen Kooi  07org.openembedded.dev * r96c6405186 10openembedded.git/recipes/xorg-lib/ (10 files in 2 dirs): pixman 0.19.4: add 3 more NEON patches Oct 02 13:08:28 03Martin Jansa  07master * r9ca3d13317 10openembedded.git/recipes/evopedia/evopedia_git.bb: Oct 02 13:08:28 evopedia: upgrade to new version using Qt Oct 02 13:08:28 Signed-off-by: Martin Jansa Oct 02 13:26:51 ericben, not not sure if the fp stuff is for emulation Oct 02 13:26:57 or escribes the api used Oct 02 13:27:03 * mrcan is away: Gone away for now Oct 02 13:35:02 Crofton: yes it seems to be for the API Oct 02 13:35:19 i really hate libtool Oct 02 13:35:31 but anyway I seems to have a problem in the kernel Oct 02 13:35:47 and not in the applications Oct 02 13:36:22 I'm writing a mail to linux arm kernel mailing list as it seems disabling preempt solves one of my problems Oct 02 13:37:31 and after this email, I may go away for a real weekend and see if beeing far from the board helps to reduce the problem :-) Oct 02 13:42:10 yeah Oct 02 13:42:21 We are spending the weekend at a lake Oct 02 13:42:25 I am supposed to be sailing Oct 02 13:42:27 but no wind yet Oct 02 13:42:38 just logged in to home and tested some sw though :) Oct 02 13:42:58 I wish I had pointed a webcam at the spectrum analyzer Oct 02 13:45:00 hehe Oct 02 13:47:01 03Martin Jansa  07master * ra734a4012d 10openembedded.git/recipes/freesmartphone/libfreesmartphone-glib_git.bb: Oct 02 13:47:01 libfreesmartphone-glib: bump SRCREV a bit Oct 02 13:47:01 Signed-off-by: Martin Jansa Oct 02 13:47:04 03Martin Jansa  07master * rbfd2e0ce6f 10openembedded.git/recipes/openmoko-3rdparty/mokowm_git.bb: Oct 02 13:47:04 mokowm: bump SRCREV a bit Oct 02 13:47:04 Signed-off-by: Martin Jansa Oct 02 13:48:48 RP: ping? Oct 02 13:50:14 Crofton: here lot of wing which prevented me to go for biking ;-) Oct 02 13:52:08 we had too much yesterday and not enough this morning Oct 02 14:12:58 * mrcan is back. Oct 02 14:13:58 * mrcan is away: Gone away for now Oct 02 14:15:05 * mrcan is back. Oct 02 14:16:47 mrcan seems a little bit undecisive Oct 02 14:19:33 ops sorry :) Oct 02 14:29:52 np :-) Oct 02 14:38:32 03Frans Meulenbroeks  07org.openembedded.dev * r5e71c5f80b 10openembedded.git/recipes/tgt/tgt_1.0.9.bb: Oct 02 14:38:32 tgt: upgraded to 1.0.9 Oct 02 14:38:32 Signed-off-by: Frans Meulenbroeks Oct 02 14:40:31 03Frans Meulenbroeks  07org.openembedded.dev * r25b2db8ccd 10openembedded.git/recipes/tgt/tgt_1.0.8.bb: Oct 02 14:40:31 tgt: remove 1.0.0 Oct 02 14:40:31 oops; did a git mv, but forgot to add the deleted file Oct 02 14:40:31 Signed-off-by: Frans Meulenbroeks Oct 02 14:40:34 03Andreas Oberritter  07org.openembedded.dev * rf70910456e 10openembedded.git/recipes/libcddb/libcddb_1.3.2.bb: Oct 02 14:40:34 libcddb-1.3.2: initial recipe Oct 02 14:40:34 Signed-off-by: Andreas Oberritter Oct 02 14:40:34 Signed-off-by: Frans Meulenbroeks Oct 02 15:13:12 03Frans Meulenbroeks  07org.openembedded.dev * rc4ae6e3ba4 10openembedded.git/recipes/cdrkit/files/xconfig.patch: Oct 02 15:13:12 cdrkit: files dir renamed to cdrkit Oct 02 15:13:12 Signed-off-by: Frans Meulenbroeks Oct 02 15:13:13 03Andreas Oberritter  07org.openembedded.dev * rd3cf329f4d 10openembedded.git/recipes/cdrkit/ (cdrkit_1.1.10.bb cdrkit_1.1.9.bb files/xconfig.patch): Oct 02 15:13:14 cdrkit: update from 1.1.9 to 1.1.10 Oct 02 15:13:14 * 1.1.9 had conflicts with (e)glibc. Oct 02 15:13:14 * 1.1.10 was released on 2009/10/11 and is the latest version. Oct 02 15:13:15 Signed-off-by: Andreas Oberritter Oct 02 15:13:15 Signed-off-by: Frans Meulenbroeks Oct 02 15:30:56 03Khem Raj  07master * r2df0b5cd71 10openembedded.git/recipes/chicken/ (6 files in 3 dirs): Oct 02 15:30:56 chicken: Fix cross and target builds and move 4.4.0 to 4.6.0 release Oct 02 15:30:56 Signed-off-by: Khem Raj Oct 02 16:20:46 03Andreas Oberritter  07org.openembedded.dev * rb672b28ba6 10openembedded.git/recipes/ (3 files in 2 dirs): (log message trimmed) Oct 02 16:20:46 libcdio-0.82: reduce package size, add dependencies, move recipe Oct 02 16:20:46 * Move recipe out of xbmc directory. Oct 02 16:20:46 * Depend on libcddb and ncurses. Oct 02 16:20:46 * Use the GNU_MIRROR variable. Oct 02 16:20:47 * Move executables into new package "libcdio-utils". Oct 02 16:20:47 Signed-off-by: Andreas Oberritter Oct 02 16:26:55 03Frans Meulenbroeks  07org.openembedded.dev * r940d34efcf 10openembedded.git/recipes/libcdio/ (libcdio.inc libcdio_0.82.bb): Oct 02 16:26:55 libcdio: added INC_PR, bumped INC_PR Oct 02 16:26:55 this one was missing a PR bump; also added INC_PR Oct 02 16:26:55 Signed-off-by: Frans Meulenbroeks Oct 02 16:54:52 hi , i m searching for code CPRM protection copying in C/C++ or any other language , i searched google but no result.Someone know about CPRM protection coding or CPRM ? Oct 02 17:00:23 JDuke128: what is a CPRM? Oct 02 17:02:08 content protection for recordable media :) Oct 02 17:03:23 khem: that is not the same a bdb? Oct 02 17:04:13 JDuke128: i am afraid you have to search for closed solutions Oct 02 17:04:41 khem: stupid question.. with gcc 4.5.. do we use LTO by default? Oct 02 17:04:56 khem: or would one still need to pass -flto or such? Oct 02 17:04:59 not default Oct 02 17:05:02 but its enabled Oct 02 17:05:14 so in theory you can enable it for a given package Oct 02 17:05:26 it wont work for many pakcages like kernel glibc etc Oct 02 17:05:33 thats why its not default Oct 02 17:05:47 one needs to bean out package by package and who has time for that Oct 02 17:06:02 zecke: yes you need to pass -flto Oct 02 17:06:16 zecke: with -fwhole-program it will be even more inreresting Oct 02 17:08:03 khem: that is for linking or even on compiling? Oct 02 17:08:22 even on compiling Oct 02 17:08:36 because compiler has to generate info for LTO Oct 02 17:09:28 with flto the .o has information about various IR structures Oct 02 17:09:50 khem: and ofcourse use gcc for linking Oct 02 17:10:15 cprm content protection Oct 02 17:10:16 yes Oct 02 17:10:26 i need help , i searched google alot every where Oct 02 17:10:45 there is no sample code or nothing for CPRM SDK or library or any code for linux window Oct 02 17:10:46 s Oct 02 17:10:49 nothing! Oct 02 17:10:57 zecke: yep Oct 02 17:11:17 JDuke128: because CPRM is flawed, but that is off topic. Search for nagra and DRM Oct 02 17:11:30 zecke: see http://sakrah.dontexist.org/node/1 Oct 02 17:11:37 I blogged few months back on it Oct 02 17:12:06 khem: i tried pre 4.5 for plugins and LTO... LTO choked on webkit. :) Oct 02 17:12:28 zecke: anything will choke in webkit Oct 02 17:12:41 500M of data to process Oct 02 17:12:48 you wont have that much memory Oct 02 17:13:11 zecke: have you looked into making it more managable without stripping Oct 02 17:13:25 that said you might try it without -g Oct 02 17:13:46 nagra and DRM ? Oct 02 17:14:01 khem: choke was... more as internal compiler error Oct 02 17:14:14 JDuke128: and I used google. sorry you are completely off topic here Oct 02 17:14:32 khem: i wonder why process creation suffers from lto? Oct 02 17:15:05 zecke: yeah it was one of things on my plate but in free time you can only do so much :) Oct 02 17:17:00 i know but i think this channel is experienced of programmers Oct 02 17:17:07 thats why i asked here Oct 02 17:17:17 thanks alot Oct 02 17:17:24 u helped me alot by saying DRM Oct 02 17:17:30 i look for DRM now Oct 02 17:17:33 yes it has some API s Oct 02 17:17:36 but Oct 02 17:18:05 JDuke128: at one company we wanted to have licenses for running the software, we went for a german dongle based solution... Oct 02 17:18:33 dongle is expensive Oct 02 17:19:18 in real i need some copy protection , but no DVD and CD has unique id or SD cards Oct 02 17:19:34 i just write a simple software and check unique id Oct 02 17:19:45 if unique id is true , software will run Oct 02 17:19:52 and the unique id cannot be copied Oct 02 17:20:36 when unique id changes or not found , software stops Oct 02 17:22:46 zecke , do u know any kind of cheap solution for this ? Oct 02 17:23:33 JDuke128: no, dongle had best cost/security ratio for our solution Oct 02 17:24:06 how much dongle is ? Oct 02 17:24:07 price Oct 02 17:24:29 JDuke128: i don't remember Oct 02 17:39:24 03Khem Raj  07master * re1abf539af 10openembedded.git/recipes/free42/ (files/free42-bcd-gtk-only.patch free42-vga_1.4.41.bb): Oct 02 17:39:24 free42-vga_1.4.41.bb: Fix compilation issues. Oct 02 17:39:24 It was not having glib-2.0 paths correct and BCD_MATH was not passed Oct 02 17:39:24 into makefile Oct 02 17:39:24 Signed-off-by: Khem Raj Oct 02 17:47:09 jo Oct 02 17:57:33 03Khem Raj  07master * r1b8a89a8ce 10openembedded.git/recipes/openmoko-3rdparty/mcnavi_0.2.10.bb: Oct 02 17:57:33 mcnavi: Upgrade to 0.2.13 Oct 02 17:57:33 0.2.10 is not fetchable anymore Oct 02 17:57:33 Signed-off-by: Khem Raj Oct 02 19:36:09 hm, I'm puzzled with that amend.inc, I've this config http://pastebin.org/2387 Oct 02 19:36:50 in the ts72xx/recipes/netbase/amend.inc is: Oct 02 19:36:52 PR .= ".1" Oct 02 19:36:52 SRC_URI_append_ts72xx = "file://interfaces" Oct 02 19:38:00 bitbake find's the amend.inc and parses it, but the interfaces are not overwritten and PR of package isn't changed Oct 02 19:38:40 it should be netbase_4.21-r48.1 but it's still netbase_4.21-r48 Oct 02 19:40:12 hm sorry I never used amend Oct 02 19:40:58 bitbake debug http://pastebin.org/2496 Oct 02 19:41:28 seems like it's not working Oct 02 19:42:35 it's finding changes in amend.inc, parsing netbase.bb again but probably without the amend.inc stuff Oct 02 19:53:44 ynezz: use bbappend with git/master bitbake Oct 02 19:55:48 khem: I'm using 1.10 and I wonder if I'm brave enough to switch to master :) Oct 02 19:56:44 ynezz: well I think you wont be disappointed Oct 02 19:56:53 I use master ever since Oct 02 19:57:00 ok, no problem then Oct 02 19:57:26 it's first time I've heard about bbappend, what's that? any docs? or example? Oct 02 19:57:39 it was even hard to find anything about amend.inc Oct 02 19:58:10 ah, found on mailing list "A new bitbake extension: .bbappend files" Oct 02 19:58:17 ynezz: I wrote a rant on http://sakrah.homelinux.org Oct 02 19:58:27 ynezz: that will be helpful to you Oct 02 19:58:53 ah, indeed, thanks Oct 02 19:59:14 easy peasy Oct 02 19:59:59 btw what's the future? amend or bbappend? Oct 02 20:00:23 bbappend. Oct 02 20:00:45 amend was created entirely in the metadata, because bitbake didn't support such a thing. it does now Oct 02 20:02:19 should I build from scratch when switching from 1.10 to master? Oct 02 20:02:25 I mean rm -fr tmp Oct 02 20:31:40 ynezz: I dont think so Oct 02 20:34:13 libtool 2.4 could be very nice for us Oct 02 20:35:17 zecke why? Oct 02 20:35:24 and why it isnt in yet? Oct 02 20:35:55 woglinde_: sysroot support Oct 02 20:36:09 woglinde_: I spend all day kicking neon and libtool and i still failed Oct 02 20:36:49 svn neon? Oct 02 20:37:08 zecke: heh I have some prework done for libtool 2.4 here Oct 02 20:37:17 zecke: indeed it will be nice Oct 02 20:37:25 for cross builds Oct 02 20:39:58 03Khem Raj  07master * rf07c5423a1 10openembedded.git/recipes/klimt/klimt_0.6.1.bb: Oct 02 20:39:58 klimt_0.6.2: Add recipe Oct 02 20:39:58 Signed-off-by: Khem Raj Oct 02 20:41:55 woglinde_: yes the http lib Oct 02 20:45:39 woglinde_: many bad rpath's in OE Oct 02 20:46:35 khem: how much? I have a 2.2.10 upgrade (actually not much to do thanks to poky) Oct 02 20:48:37 zecke: I have 2.4 ported in poky style Oct 02 20:48:44 but its not finalized yet Oct 02 20:48:55 I still need to run buil Oct 02 20:48:56 ds Oct 02 20:49:48 khem: thanks, works like a charm :8 Oct 02 20:50:28 bbappend can be treated as a normal recipe right? Oct 02 20:50:56 ynezz: its an addendum to existing .bb Oct 02 20:51:13 you can have a .bb also in the same layer Oct 02 20:52:04 so one can do things like PR := ".1" (I would try it first, but the build is running now) Oct 02 20:52:32 well, I'll better look at source and see how it works Oct 02 20:54:05 yes you can do those stuff Oct 02 21:20:29 zecke: dare to try 2.4 Oct 02 21:20:33 libtool :) Oct 02 21:20:49 I can push it to an user branch Oct 02 21:20:56 khem: once my machine arrives from TW Oct 02 21:21:37 zecke: ok :) Oct 02 21:21:42 that might take some time Oct 02 21:22:01 ~ 2 weeks left Oct 02 21:22:25 hmmm ok Oct 02 21:22:46 zecke: I will probably post the patches before that Oct 02 21:25:02 khem: would be nice Oct 02 21:25:13 just fixed an issue in ncurses Oct 02 21:25:28 ncurses is another hairy beast Oct 02 21:29:55 ynezz: s/:=/.=/ Oct 02 21:30:00 libtool with sysroot support sounds lovely Oct 02 21:31:49 fixing a pam packaging bug from 2008... Oct 02 21:36:12 kergoth_: yes Oct 02 21:36:28 rpath issue in gettext too... :( Oct 02 21:36:41 kergoth_: so far I have compiled libtool 2.4 locally all goes well Oct 02 21:36:59 kergoth_: real hell will break when I will fire a build from scratch Oct 02 21:42:29 kergoth_: now we can remove this sedding installed=yes to installed=no Oct 02 21:42:45 crap... somehow libtool turns the -Lsysroot/.../usr/lib into a -Wl,-rpath -Wl,sysroot... i hate it Oct 02 21:42:46 with proper sysroot support in libtool Oct 02 21:43:14 zecke: we need to remove the logic in classes I think too Oct 02 21:43:24 oh with 2.4 I mean Oct 02 21:43:43 khem: I have only read the NEWS, how does it find the sysroot? does it query gcc? Oct 02 21:45:33 zecke: it needs to be configured with sysroot Oct 02 21:45:45 like gcc and friend Oct 02 21:45:47 s Oct 02 21:45:59 --with-sysroot= Oct 02 21:49:18 khem: when building libtool? or when it gets created in a package? Oct 02 21:51:21 * ant_ still can't believe libtool got sysroot support Oct 02 21:54:08 done by Paolo Bonzini (GNU Smalltalk maintainer) Oct 02 21:56:04 zecke: when building libtool Oct 02 21:56:57 kergoth_: I got rid of .la munging from staging.bbclass and utils.bbclass and repaired insane.bbclass to not check for installed=yes Oct 02 21:57:03 nice Oct 02 21:57:09 any other places I should look into Oct 02 21:58:12 I dont know lets do a bitbake minimal-image and see what falls apart Oct 02 21:58:44 it will prolly speed up build a bit too Oct 02 21:59:10 heh atleast libtool-native is holding on :) Oct 02 21:59:26 khem: it will also fix my todays issue with rpath to sysroot Oct 02 22:00:20 zecke: yes I think so too Oct 02 22:11:47 so, I stop trying to understand this version of libtool Oct 02 22:12:20 :) Oct 02 22:14:31 and insane should check rpaths in a different way... everything outside of ${libdir} is wrong... Oct 02 22:15:33 zecke: right Oct 02 22:16:46 zecke: ideally .la files should not be identical for native or target package Oct 02 22:26:32 well, I'm adding a do_oldconfig task editing do_menuconfig. Anybody did need it before? I see oddities moving from 2.6.34 -> 2.6.35/6 su prefer to do it gradually Oct 02 22:41:46 khem: one more thing, trying to add my own /etc/modules, so bbappending modutils-collateral, it works, there's my version of file in image/etc/modules but it isn't available in final image Oct 02 22:43:25 now I did bitbake -c clean modutils && bitbake my-image, but only image is being build, but without modutils, this seems strange, like some dependency problem Oct 02 22:44:31 maybe some side effect after the switch from 1.10 to master, without rm tmpdir? Oct 02 22:47:29 even adding PR .= ".1" to all my bbappended recipes, bitbake -c cleaning them won't trigger their build Oct 02 22:48:12 will try to reparse Oct 03 00:00:29 ynezz: delete the relevant ipks from deploy Oct 03 00:00:58 kergoth_: I think the SCM exporting patch could be merged into bitbake Oct 03 00:01:08 k Oct 03 00:01:18 kergoth_: my fetching has shown no problems Oct 03 00:01:29 and dl_dir size reduced noticiably Oct 03 00:01:55 unless we merge it more people wont test it :) Oct 03 00:02:22 unfolding libtool is a greater mess Oct 03 00:02:30 now libtool is refusing to add sysroot Oct 03 00:02:32 lets see why Oct 03 00:04:25 you guys should rest :) Oct 03 00:07:10 Crofton: :) Oct 03 00:07:30 I plan to rest Oct 03 00:07:34 we are at some lake Oct 03 00:07:45 out wives did a triathlon Oct 03 00:07:53 someone should send me to jail with an iPad so I can read all those novels Oct 03 00:07:53 I sailed a boat Oct 03 00:08:01 Crofton: nice Oct 03 00:08:18 Crofton: I once upon a time did para sailing Oct 03 00:08:26 cool Oct 03 00:08:30 in Goa/India Oct 03 00:08:30 I windsurf some Oct 03 00:08:38 it was fun Oct 03 00:08:49 yeah Oct 03 00:09:01 now I do hiking/biking Oct 03 00:09:10 no more bungee jumping Oct 03 00:09:10 yeah Oct 03 00:09:18 the water iscold in SFO Oct 03 00:09:48 my wife is more daring she plans to do bungee jumping from helicopter Oct 03 00:09:54 I said not in this life time :) Oct 03 00:10:11 yeah pacific is really cold ocean much like its name : Oct 03 00:10:11 03Andrea Adami  07org.openembedded.dev * r2596310416 10openembedded.git/recipes/linux/ (2 files in 2 dirs): Oct 03 00:10:12 linux-kexecboot_2.6.35: update to latest stable patchset (7) Oct 03 00:10:12 * add c7x0 now (w100 driver fix in the patchset) Oct 03 00:10:12 * unfortunately framebuffer seems broken on c7x0...even no kernel logo Oct 03 00:10:12 * bump PR Oct 03 00:10:22 03Andrea Adami  07org.openembedded.dev * r5899694908 10openembedded.git/recipes/linux/linux-kexecboot/c7x0/defconfig: Oct 03 00:10:22 linux-kexecboot_git: add initial c7x0 defconfig Oct 03 00:10:22 * plagued by same framebuffer issue like in 2.6.35 (bug?) Oct 03 00:10:22 * wip Oct 03 00:10:28 rofl Oct 03 00:10:43 bungee jumping from a helicopter is nuts Oct 03 00:11:01 yeah I know but when I was younger I could have thought about it Oct 03 00:11:35 now my heart will stop in the middle and reboot :) Oct 03 00:12:16 Crofton: was it a day trip ? Oct 03 00:13:33 hmmm btw. we have hard-fp support for arm now in OE along with linaro stuff Oct 03 00:13:51 so we can configure cortex-a8+ toolchains using hard-fp Oct 03 00:14:06 it speeds up stuff quite a bit Oct 03 00:14:29 I want to give it a shot Oct 03 00:47:40 a curious question: Why i can't run OE as root? **** ENDING LOGGING AT Sun Oct 03 02:59:57 2010