Re: [yocto] 1.4_M6.rc2
On Wed, 2013-04-17 at 13:14 -0700, Flanagan, Elizabeth wrote: We're delayed slightly for the 1.4_M6.rc2 build. The build should be soonish, once RP gives me the thumbs up. Please adjust your schedules accordingly, we're hoping to get this started in a few hours. We do have a 1.6_M6.rc2 candidate build which is revision 87eb70243749e0e2431f98770746fa42002f4441 (dylan branch). I suggest we start testing this with particular focus and prioritisation on the systemd and sysvinit testing due to the changes that have gone in last minute. If any problems are discovered, please flag them ASAP as the plan would be to abort this rc and rebuild if any issues in those areas are discovered. I should stress that DISTRO_FEATURES should not be changed in existing build directories (reusing sstate is fine). Several of the init issues we've see have been tracked down to this. Its never been supported although is probably something we'll improve in 1.5 since it catches so many people out. This note needs to be in the release notes as well. Cheers, Richard ___ yocto mailing list yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/yocto
Re: [yocto] writeup on the mechanics of FILESEXTRAPATHS and bbappend files
On 2013-04-17 16:26, Trevor Woerner wrote: On Wed, Apr 17, 2013 at 11:12 AM, Gary Thomas g...@mlbassoc.com wrote: I also think it would be good to point out the use of := in the FILESEXTRAPATHS assignment and why it's important. Knowing that the := operator will put your directories first is only useful if you also know that the underlying mechanism will take the first such files from the given directories it encounters :-) What's to say bitbake doesn't look at all the directories/files and take the last ones? This has more to do with how the path is constructed (using _prepend). My point, which was totally unclear to me when using this line, was that with ':=' ${THISDIR} evaluates to the directory that contains the .bbappend file (what you want/need) whereas with '=' it evaluates to the directory containing the main .bb file. Incredibly confusing if you miss it... -- Gary Thomas | Consulting for the MLB Associates |Embedded world ___ yocto mailing list yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/yocto
Re: [yocto] writeup on the mechanics of FILESEXTRAPATHS and bbappend files
On Thu, 18 Apr 2013, Gary Thomas wrote: On 2013-04-17 16:26, Trevor Woerner wrote: On Wed, Apr 17, 2013 at 11:12 AM, Gary Thomas g...@mlbassoc.com wrote: I also think it would be good to point out the use of := in the FILESEXTRAPATHS assignment and why it's important. Knowing that the := operator will put your directories first is only useful if you also know that the underlying mechanism will take the first such files from the given directories it encounters :-) What's to say bitbake doesn't look at all the directories/files and take the last ones? This has more to do with how the path is constructed (using _prepend). My point, which was totally unclear to me when using this line, was that with ':=' ${THISDIR} evaluates to the directory that contains the .bbappend file (what you want/need) whereas with '=' it evaluates to the directory containing the main .bb file. Incredibly confusing if you miss it... ah, quite so, must make a note of that. i wonder how many other places that point should be made forcefully. rday -- Robert P. J. Day Ottawa, Ontario, CANADA http://crashcourse.ca Twitter: http://twitter.com/rpjday LinkedIn: http://ca.linkedin.com/in/rpjday ___ yocto mailing list yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/yocto
Re: [yocto] runqemu: Could not initialize SDL(No available video device) - exiting
On Thu, Apr 18, 2013 at 12:50 AM, Rudolf Streif rstr...@linuxfoundation.org wrote: Removing sstate triggered a rebuilt of qemu-native which also includes running the configure task again whose output sets the switch whether or not Qemu is built with video support. Apparently the first time you built it it could not find libsdl-devel on your host but the second time it could. But of course I cannot tell you why that happened. Do you still have the logs? Did you install libsdl-devel between the two attempts? Unfortunately the logs are long lost; I clean out TMPDIR quite regularly (but retain sstate). This makes perfect sense and is most likely what happened; I probably noticed I was missing libSDL-devel when I built qemu from git by hand :-) In any case restoring my old sstate and doing $ bitbake -c cleansstate qemu-native $ bitbake -c cleansstate nativesdk-qemu cleaned everything up for me. Should YP be trying to build its own copy of libSDL (with the -devel) if it finds some part of libSDL (i.e. the header files) is missing from the host? This is almost like a host contamination issue, no? ___ yocto mailing list yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/yocto
Re: [yocto] writeup on the mechanics of FILESEXTRAPATHS and bbappend files
Sorry if I'm asking a really stupid question, but if so, it means I'm way off the mark :-) On Thu, Apr 18, 2013 at 6:23 AM, Gary Thomas g...@mlbassoc.com wrote: My point, which was totally unclear to me when using this line, was that with ':=' ${THISDIR} evaluates to the directory that contains the .bbappend file (what you want/need) whereas with '=' it evaluates to the directory containing the main .bb file. Incredibly confusing if you miss it... So it's (sort of like) a matter of timing? By evaluating THISDIR now (though the use of ':='), we get the directory we are in when this directive is encountered. Whereas with '=' it will be evaluated sometime in the future when we are no longer in THISDIR but rather in SOMEOTHERDIR? ___ yocto mailing list yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/yocto
[yocto] [poky] bash: oe-init-build-env: line 33: syntax error near unexpected token `elif'
Hi, The following commit http://ares/gitweb/?p=tools/poky.git;a=blobdiff;f=oe-init-build-env;h=68af7b5193b73627cd6ebf6196adfa685a787c38;hp=67eddcd295bd9d1adac21fc4e97506fca05f4bfa;hb=813127247a1100b1abe179dfba25795560eac864;hpb=a468b0d5579148771da716eaf357e5d45fc3211c is causing a build error while sourcing the oe-init-build-env build script $ cd /tool/yocto/poky;source oe-init-build-env build : command not found : command not found bash: oe-init-build-env: line 33: syntax error near unexpected token `elif' 'ash: oe-init-build-env: line 33: ` elif [ -n $ZSH_NAME ]; then This is on Ubuntu 12.10 x86 64-bit. Elvis Dowson ___ yocto mailing list yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/yocto
[yocto] Package woes
This morning, I had built an image successfully using a slightly older Poky/Yocto master (af9ed97e30ac1a3a2b4e1062f2f8aee255d05972). Then I updated to master (90b98764555945a186562ca8d501a9585ce2b23f) Now, when I try to build an image (any 'rootfs' image), I get a ton of errors like these: | * opkg_install_pkg: Package python-modules md5sum mismatch. Either the opkg or the package index are corrupt. Try 'opkg update'. | * opkg_install_cmd: Cannot install package python-modules. | * opkg_install_pkg: Package libffi6 md5sum mismatch. Either the opkg or the package index are corrupt. Try 'opkg update'. | * opkg_install_cmd: Cannot install package gst-plugins-base-app. | * opkg_install_pkg: Package libx11-6 md5sum mismatch. Either the opkg or the package index are corrupt. Try 'opkg update'. | * opkg_install_cmd: Cannot install package gst-plugins-base-xvimagesink. There are also a lot of messages like these: | Found /home/local/imx6_poky/tmp/deploy/ipk/armv7a-vfp-neon/locale-base-sw-ke_2.17-r0_armv7a-vfp-neon.ipk in Packages, but mtime differs - re-reading | Found /home/local/imx6_poky/tmp/deploy/ipk/armv7a-vfp-neon/locale-base-sw-tz_2.17-r0_armv7a-vfp-neon.ipk in Packages, but mtime differs - re-reading | Found /home/local/imx6_poky/tmp/deploy/ipk/armv7a-vfp-neon/locale-base-szl-pl_2.17-r0_armv7a-vfp-neon.ipk in Packages, but mtime differs - re-reading I don't understand this one at all because I tried rebuilding virtual/libc-locale and I still get them. Any ideas what's happening? Sadly, I seem to be totally stuck with this problem :-( -- Gary Thomas | Consulting for the MLB Associates |Embedded world ___ yocto mailing list yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/yocto
Re: [yocto] [poky] bash: oe-init-build-env: line 33: syntax error near unexpected token `elif'
On Apr 18, 2013, at 8:12 PM, Elvis Dowson elvis.dow...@gmail.com wrote: On Apr 18, 2013, at 7:58 PM, Trevor Woerner twoer...@gmail.com wrote: On Thu, Apr 18, 2013 at 11:37 AM, Elvis Dowson elvis.dow...@gmail.com wrote: http://ares/gitweb/?p=tools/poky.git;a=blobdiff;f=oe-init-build-env;h=68af7b5193b73627cd6ebf6196adfa685a787c38;hp=67eddcd295bd9d1adac21fc4e97506fca05f4bfa;hb=813127247a1100b1abe179dfba25795560eac864;hpb=a468b0d5579148771da716eaf357e5d45fc3211c Do you have a better link? http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/oe-init-build-env?id=813127247a1100b1abe179dfba25795560eac864 I reverted the commit, but the error still persisted. It's something else, probably due to a recent update to Ubuntu 12.10. I tried reverting to known commit ids but all of them don't work. My shell is correctly configured to use bash instead of dash. When I ran the debug mode for the bash shell, I found out that the : command not found : command not found errors were due to the new lines on line 2 and line 20. Deleting the new lines made the : command not found errors go away, but I get the following error with the older commit bash: ./oe-init-build-env: line 45: syntax error: unexpected end of file and the following error with the current master bash: ./oe-init-build-env: line 33: syntax error near unexpected token `elif' 'ash: ./oe-init-build-env: line 33: ` elif [ -n $ZSH_NAME ]; then Man, this is so un-expected!! :-) Okay, false alarm, sorry about that. I noticed that when I pull the latest updates, something went wrong and there were a lot of modified files after the update in red. I did a git clean -df and git reset --hard, and it solved the problem. Don't know how that happened though, but problem fixed by resetting the git repo and cleaning all the dirty files. Best regards, Elvis Dowson ___ yocto mailing list yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/yocto
Re: [yocto] [poky] bash: oe-init-build-env: line 33: syntax error near unexpected token `elif'
Hi, I think I know what happened. I briefly setup autocrlf=true in the global .gitconfig last week, and when I pull in all the latest commits, it ended up adding CRLF, and ended up causing issues. Even after disabling it, the damage already done to the local poky repository. Luckily I had backups. This shouldn't have happened to with git! Elvis Dowson ___ yocto mailing list yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/yocto
[yocto] Warning message building python packages
Hi, I have posted my question on Stack Overflow: http://stackoverflow.com/questions/16090550/building-python-packages As the presence of OE users is on SO is not quite high, I though I might post the link here also. Cheers, -- Ilya ___ yocto mailing list yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/yocto
Re: [yocto] runqemu: Could not initialize SDL(No available video device) - exiting
Should YP be trying to build its own copy of libSDL (with the -devel) if it finds some part of libSDL (i.e. the header files) is missing from the host? This is almost like a host contamination issue, no? You are raising an interesting point to which I have no answer. YP has a recipe to build libsdl and it specifies BBCLASSEXTEND = nativesdk only, not BBCLASSEXTEND = native nativesdk as one would expect if were to build it for Qemu. The qemu.inc recipe also does not list libsdl in DEPENDS_class-native but it does so in DEPENDS_class-nativesdk and in DEPENDS. I cannot tell you why that is but it may be worthwhile trying to add libsdl to DEPENDS_class-native of qemu.inc and native to BBCLASSEXTEND in libsdl_1.2.15.bb. I wonder if bitbake qemu-nativesdk would build even if you don't have libsdl installed on your host. Rudi ___ yocto mailing list yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/yocto