**** BEGIN LOGGING AT Tue May 12 02:59:59 2015 May 12 04:30:23 vvgljnudtnnnckbhurbljgrenvvgurihncejdidrlece May 12 04:30:48 sorry... May 12 05:25:42 hi May 12 05:25:50 help anyone ? :p May 12 08:01:26 morning all May 12 08:54:26 <[w00t]> Hi, I'm having trouble installing my core-minimal build in a virtual machine environment. The build is a simple core-minimal for genericx86 without GPLv3, with the only addition grub-0.97. I've also removed the options for efi and live image. The target is stored as a tarball and I install using a separate minimal gentoo image (just to boot up the May 12 08:54:26 <[w00t]> system and manage disks). As I boot up I set up the disks with fdisk (/dev/sda1 as ext2 and /dev/sda2 as ext3). I mount them in temporary directories (where sda2 is the "fake" root and sda1 is mounted in the "fake" boot directory) and extract the image generated by yocto. After setting up my grub.conf file mounting /dev and /proc in the fake ro May 12 08:54:26 <[w00t]> ot, I chroot to the virtual environment and try to run grub-install on /dev/sda.. but I get a lot of segmentation faults.. however grub-install later prints out that the installation was successful, but I'm unable to boot. Does anyone know what might be wrong? May 12 11:03:21 hi. anyone care to help me with creating a recipe? i need dbus-python so i've created a recipe that installs it using autotools. during configure it finds the python headers, but then reports that they are not sufficient ... May 12 11:03:56 if i read the m4 macros correctly, the test is made by creating a C++ file that contains only an include of "Python.h", and that seems to fail... May 12 11:05:27 ionte: python-dbus is in oe-core already May 12 11:06:45 rburton: it is? i searched for it... May 12 11:07:00 oh.. there it is... May 12 11:11:11 rburton: thanks, it installed cleanly... i think it was one of the dependencies i was missing, most likely python-threading... May 12 11:26:08 anybody know how to do conditional inherits/requires in bitbake recipes? May 12 11:26:23 the goal is to check variables coming in from outside before using their values in commands. May 12 12:50:04 bluelightning: hi, is there any chance someone will look at the fitImage patches and give me some feedback please ? May 12 12:50:22 bluelightning: is there anything I can do to help it ? May 12 12:50:49 Marex: it's not my area... but I can ping some people maybe May 12 12:53:56 bluelightning: that'd really help a lot, sorry to put this onto you May 12 12:54:03 bluelightning: thank you! May 12 12:54:32 good afternoon May 12 12:55:03 INIT: /etc/inittab[31]: id field too long (max 4 characters) -> I am confused why I am getting this while migrating from dylan to daisy. May 12 12:55:25 we use the id "ttyS1" which is more than four characters, but it used to work like that. May 12 12:56:07 to be fair, I would need to check what was generated with dylan in the inittab... but is there any expected change about this with sysvinit? Actually, I think we ought to even use busybox's stuff rather than sysvinit... May 12 13:14:27 lpapp: I recall one of the entries in the inittab was malformed at one point May 12 13:16:48 Marex: how did you fix that? May 12 13:17:01 and also, what created that entry? May 12 13:17:07 I cannot for my life figure that out. May 12 14:16:57 dylan generated this for us in inittab at the end: foo:5:respawn:/sbin/foo.sh May 12 14:17:25 S:234:respawn:/sbin/getty -L 38400 ttyS1 May 12 14:17:46 but daisy generates this one instead of those two in addition on top of the "stock" inittab: ttyS1:12345:respawn:/sbin/getty ttyS1 38400n May 12 14:18:13 there surely must be something that has changed. I can understand why ttyS1 is invalid identifier for sysvinit. May 12 14:22:09 lpapp: if you look at the code in meta/recipes-core/sysvinit/sysvinit-inittab_2.88dsf.bb, it drops the "tty" from the start of the device name May 12 14:22:16 I don't know how you could be getting the full name in there May 12 14:32:47 bluelightning: not from the beginning because of the semi-colon, or is that some special syntax element? May 12 14:33:06 lpapp: which semicolon? May 12 14:33:39 label=`echo ${i} | sed -e 's/^.*;tty//' -e 's/;.*//'` May 12 14:34:26 hmm, this is probably the relevant line though? echo "$n:2345:respawn:${base_sbindir}/getty 38400 tty$n" >> ${D}${sysconfdir}/inittab May 12 14:36:19 bluelightning: how to start debugging this issue? May 12 14:36:40 can I somehow make sure that yocto always fetches a repository in the fetch step and uses the latest commit of a branch if SRCREV points to a branch name? it seems that this is not always done May 12 14:37:10 jaeckel: have a look at AUTOREV in the manual, that does what you want May 12 14:37:20 basically, SRCREV=${AUTOREV} May 12 14:37:49 lpapp: you can see the inputs, and the inittab that gets created - that should be enough... May 12 14:37:53 and how do I then define which branch it should rely on? May 12 14:38:54 through the branch syntax in SRC_URI? May 12 14:40:22 bluelightning: how can I see the inputs? May 12 14:40:37 lpapp: you can see the variables it's using May 12 14:41:21 bluelightning: you mean SYSVINIT_ENABLED_GETTYS ?= "1" May 12 14:41:23 ? May 12 14:41:53 lpapp: is that the only variable reference in that recipe? May 12 14:42:24 sure May 12 14:42:34 hint: it isn't... May 12 14:43:04 hmm, it is for me. May 12 14:43:16 oh, gosh, I am looking at the dylan stuff May 12 14:43:25 actually, no, I take that back. May 12 14:43:29 by variable reference I mean ${ ... } May 12 14:43:30 so the other use is the for loop May 12 14:43:38 but that is not initialised May 12 14:43:42 initialisation* May 12 14:45:28 so I have this ./build/tmp/work/foo-linux-gnueabi/foo-image/1.0-r0/rootfs/etc/inittab with ttyS1:12345:respawn:/sbin/getty ttyS1 38400n at the end. May 12 14:48:51 bluelightning: perhaps it is coming from somewhere else, but where would that come from? May 12 14:54:36 lpapp: have you perhaps modified meta/recipes-core/sysvinit/sysvinit-inittab/inittab ? or do you have a bbappend which does the same? May 12 14:56:02 nope May 12 14:56:47 well, unless you have some custom mangling going on, the sysvinit-inittab recipe has a default inittab file which it then modifies as you can see, that is all there is to it May 12 15:04:06 bluelightning: yeah, I agree. I do not know where the issue is. May 12 15:06:47 bluelightning: cannot SERIAL_CONSOLE play a role in this? May 12 15:07:11 lpapp: it does, yes... that's what I meant about inputs May 12 15:07:32 we have SERIAL_CONSOLE = "ttyS1 38400n" May 12 15:07:40 I thought you meant SYSVINIT_ENABLED_GETTYS May 12 15:09:04 bluelightning: did that line become wrong later than dylan? May 12 15:09:32 lpapp: I believe it's supposed to be 38400 ttyS1 May 12 15:09:40 that is what I was going to say May 12 15:09:47 http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#var-SERIAL_CONSOLE May 12 15:09:56 SERIAL_CONSOLE = "115200 ttyS0" is in the example May 12 15:10:01 also, this is now obsolete anyway? May 12 15:10:26 SERIAL_CONSOLES is supposed to replace it yes May 12 15:10:50 SERIAL_CONSOLES currently defaults to SERIAL_CONSOLE with ' ' replaced with ';' May 12 15:10:55 since I have to touch this line, I probably ought to append the plural just as well? May 12 15:11:35 also, it is in our machine config. I am not sure whether that is the best place for this variable. May 12 15:12:07 I believe that's where people typically put it since it is machine-specific May 12 15:12:48 ok, thanks! May 12 15:31:18 #/j dbus May 12 17:26:29 Hello guys! I am trying build image with qt and sql support in it. I've added this to local.conf PACKAGECONFIG_pn-qtbase += " sql-mysql", but when after it I do "bitbake -c configure qtbase", I look into log.do_configure and see that MySQL ................ no, i.e. sql driver was not enabled May 12 17:27:06 how can I beat it? May 12 17:30:55 ncgs: you could try PACKAGECONFIG_append_pn-qtbase = " sql-mysql" May 12 18:03:17 hi, quick question, is there anyway that I can make the git fetcher to preserve the origin pointing to the SRC_URI ? I'm using devtool and right now is pointing to the local download from the fetcher (git.indirectionsymlink). Same for devshell May 12 18:30:48 can I force output of do_compile to stdout ? if yes how May 12 18:43:28 khem`: not afaik May 12 18:46:54 yeah thought so May 12 21:56:24 can anyone point me to where, I assume in a class, that changes an empty password field in etc/passwd to *, or tell me how to ensure passwords remain empty? May 12 21:57:57 rschaef: "a" password? or the root password? May 12 21:59:16 the root password. but it looks like i found it. enable IMAGE_FEATUERS debug-tweaks allows empty root passwd? May 12 22:00:10 rschaef: that's correct yes May 12 22:00:42 it does a couple of other things as well May 12 22:00:45 in 1.8 and later there is also a fine-grained feature just for that part May 12 22:01:12 any big thigns debug-tweaks does that I should look out for? or a reference to look at? May 12 22:01:50 rschaef: http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#ref-features-image May 12 22:02:43 (that is the "current" i.e. 1.8 manual) May 12 22:03:16 okay thanks May 13 02:33:54 hi. I wish to submit recipes to openembedded.org so more people can use them. Anyone has idea on how I can get started? **** ENDING LOGGING AT Wed May 13 02:59:58 2015