**** BEGIN LOGGING AT Mon Sep 09 02:59:58 2013 Sep 09 07:56:28 Hi all, can anyone help me on an X11 issue? Sep 09 07:59:37 what packages do I need to install to get X11 working .. i.e driver and xorg receipe? Sep 09 07:59:59 for the beaglebone black Sep 09 08:00:18 IMAGE_FEATURES += "x11-base" Sep 09 08:00:47 what about the video driver? Sep 09 08:02:46 I see mention of "modesetting"? Sep 09 08:10:30 anyone know the name of the video driver for the BBB? Sep 09 08:14:13 brm: that should be all set up in the machine conf for bbb Sep 09 08:14:43 tf: what do you mean? Sep 09 08:15:33 brm: there is a machine conf for the bbb in the bsp layer, it sets up stuff like that Sep 09 08:16:02 I am using meta-beagleboard, is it in there .. Sep 09 08:25:13 Heya. In the vein of earlier talks about perl module recipes, I wonder - What should be done about core modules that seemingly are not rprovided by the perl recipe? Should I just act as if these would always be part of perl? Sep 09 08:26:19 tf: still not sure what you mean by BSP, do you mean the meta-beagleboard layer? Sep 09 08:31:56 Anyone? Still don't know what video driver to setup for BBB Sep 09 08:47:10 Anyone tell me what the video driver is for the BBB? Sep 09 08:49:16 brm123: according to the .conf file it is xf86-video-fbdev Sep 09 08:51:40 Ndec: thanks .. why does the xorg.conf show driver=modesetter ? Sep 09 08:51:59 i don't know.. Sep 09 08:52:35 Do I change it to fbdev? Sep 09 09:16:01 Hi Sep 09 09:16:22 I get warning when I set LICENSE = "CLOSED" Sep 09 09:16:42 why does it give warning. Its ok in recipes Sep 09 09:16:44 ? Sep 09 09:16:58 its ok in recipe but not image... Sep 09 09:38:12 Hey, have a weird issue. I have an RDEPENDS on perl-module-net-ping, which is a core module provided by the perl recipe itself. The recipe that RPDEPEND's on it builds just fine. However, on do_rootfs for the image itself, it fails claiming that perl-module-net-ping depends on perl-module-posix which does not exist. It seems that the perl recipe only RPROVIDES a few recipes, and that these are _much_ fewer (by orders of magnitudes Sep 09 09:38:13 ) than the recipes whos RPDEPEND's are discussed in the perl-rdedends_5.14.3.inc Sep 09 09:39:33 Any clues? It seems to me, immediately, that the perl recipe does not have the RPROVIDES it should have. Sep 09 09:56:08 Stygia: My guess is that is has something to do with dynamically created packages in the perl recipe, and then something like http://lists.openembedded.org/pipermail/openembedded-core/2012-November/071094.html Sep 09 09:56:42 erbo, Yea,that does seem to be the case. Adding the missing packages as RPROVIDES in a bbappend to perl itself "solves" the issue. Sep 09 09:56:57 But not having read your link (yet) I would conclude that this is an issue with the perl recipe that should be fixed. Sep 09 09:58:18 Stygia: I think the conclusion from my link is that it's a complex problem which isn't easily solved Sep 09 09:59:20 Quite possible. At least if RPROVIDING a dynamic package is "wrong". Sep 09 09:59:27 Either way, it's a working workaround... Sep 09 10:00:38 But IMO it's a bug either way - I definitively should be able to RDEPEND on perl-module-net-ping without errors. Sep 09 10:00:57 I'll try and ask rburton or bluelightening about it later and file a bug if they agree it's an issue. Sep 09 10:02:25 Stygia: I agree :) Sep 09 10:02:50 erbo, Heh, cool. I will almost certainly file it as a bug later, because it seems off. Sep 09 10:15:29 stygia: afaik, perl-module-posix does exist Sep 09 10:16:22 or, no, my bad. i have the same issue Sep 09 10:17:42 huumm NOTE: multiple providers are available for virtual/kernel (linux, linux-dummy) Sep 09 10:18:49 zibri, Oh, yea, it does exist. Sep 09 10:19:00 zibri, That's not the issue. I can RDEPEND on it and build without issue. Sep 09 10:19:09 zibri, It fails once you do_rootfs for the image that actually uses it. Sep 09 10:19:22 why do I get this problem? my PREFERRED_PROVIDER_virtual/kernel ?= "linux" and PREFERRED_VERSION_linux ?= "3.8.13" are correctly set Sep 09 10:29:47 n01 : Are you sure that "linux" is correct value Sep 09 10:31:12 Guest42914: my recipe name is linux_3.8.13.bb Sep 09 10:31:35 so I assume PN=linux and PV=3.8.13 Sep 09 10:33:08 n01 : I think you could rename the recipe like linux-_3.8.13.bb Sep 09 10:44:20 yes, but I don't know why PREFERRED_VERSION_linux is ignored Sep 09 10:53:16 n01 : where have you set that preferred version? Have you created a new disto conf Sep 09 10:53:17 ? Sep 09 10:54:02 Guest42914: it is in my machine conf Sep 09 10:55:05 n01: in mine , have set PREFERRED_PROVIDE... in machine.conf Sep 09 10:55:15 and created new disto conf Sep 09 10:55:26 which sets VERSION... Sep 09 10:57:28 I think you need to create new disto and add version there... Sep 09 10:57:50 uhm, ok let me try ... thank you for the hint Sep 09 10:59:20 actually I have seen several machine.conf setting VERSION ... like here https://github.com/renesas-devel/meta-renesas/blob/master/meta-armadillo800eva/conf/machine/armadillo800eva.conf Sep 09 11:01:55 n01: but that is using linux-yocto. Is that what you are using Sep 09 11:02:06 ? Sep 09 11:02:27 nope ... I'm using a traditional kernel recipe Sep 09 11:02:36 http://pastebin.com/0UE7ac9M Sep 09 11:04:29 n01: I think that PREFERRED_VERSION_linux-yocto ?= "3.4%" just overrides default value in poky.conf Sep 09 11:04:45 n01: I think you need to create the distro . Sep 09 11:05:38 here is example: Sep 09 11:05:38 ok, I'll try, tnx Sep 09 11:05:40 http://pastebin.com/FfbLLb9Y Sep 09 11:05:45 n01: in your machine.conf PREFERRED_PROVIDER_virtual/kernel = "linux" Sep 09 11:05:49 ^ Sep 09 11:06:09 n01: then update you local.conf with that distro Sep 09 11:06:13 ant_work: what's wrong with that? Sep 09 11:06:20 you used ?= Sep 09 11:06:56 ohhhh Sep 09 11:07:44 I'll check that also as soon as my current compilation is done Sep 09 11:08:48 ah ok . maybe its that simple. Sep 09 11:10:43 ant_work: in poky-tiny.conf I read "Distro config is evaluated after the machine config" so my PREFERRED_PROVIDER_virtual/kernel in always overridden? Sep 09 11:11:42 heh, yes, poky-tiny.conf is declaring it here Sep 09 11:13:10 so I cannot configure a preferred provider for the kernel in machine.conf when I uso poky tiny distro? Sep 09 11:13:22 n01: afaik you could still force with a _local override Sep 09 11:35:37 ant_work : if a recipe needs some linux headers for compilation. How do you go about doing that? there is libc headers... Sep 09 11:40:13 anybody Sep 09 11:40:16 ? Sep 09 11:50:29 Can anybody tell me how I can libc headers to compile a recipe Sep 09 11:50:42 i.e. the recipe needs linux headers to compile Sep 09 11:51:28 Guest42914: yes, there are linux-libc-headers Sep 09 11:53:37 i.e. DEPENDS += "linux-libc-headers" Sep 09 11:53:42 I need to use my own version of kernel ant_work Sep 09 11:54:51 then you depends on virtual/kernel Sep 09 11:55:53 s/you/the recipe/ Sep 09 11:56:00 so I manually need to copy the headers to stage then? Sep 09 11:56:11 from the linux recipe? Sep 09 11:56:21 it depends Sep 09 11:57:04 if you plan to add specific headers there are some precautions to take Sep 09 11:57:16 there was a msg on ML last week Sep 09 11:57:20 moment Sep 09 12:03:43 Guest42914: https://lists.yoctoproject.org/pipermail/yocto/2013-September/018083.html Sep 09 12:06:51 Guest42914: are you sure recipe cannot be built against linux-libc-headers and needs a kernel tree ? Sep 09 12:08:28 ant_work : we have own kernel develment so we need to use that. Can I create a new recipe from linux-libc-headers and just change the SRC_URI? Sep 09 12:09:16 well, yes Sep 09 12:09:33 sounds a bit uncommon though Sep 09 12:10:33 so you think I should just depend on my linux recipe.. and then in the linux recipe copy the header files...? Sep 09 12:14:47 see, if the recipe depends on virtual/kernel you have the unsanitized headers to play with in ${STAGING_KERNEL_DIR} Sep 09 12:15:43 if you rely on the copy installed by make headers_install (the linux-libc-headers recipe) you then refer to the headers in ${STAGING_DIR_TARGET}${exec_prefix} Sep 09 12:16:27 I don't know the code you're compiling Sep 09 12:17:26 what I can say, even unusual libs like klibc can be compiled with the sanitized headers Sep 09 12:27:58 rburton: I forgot to add in that opengl/mesa thread that my feeling about "opengl" DISTRO_FEATURE is from times when we didn't have so many PACKAGECONFIGs and it was quite common to build xserver-xorg with dri support and mesa even when you didn't want to use it in runtime Sep 09 12:31:35 ant_work: the problem was the overriding in poky-tiny. Using poky it is just fine. Sep 09 12:32:09 n01: sure. iirc there is a reason about that choice Sep 09 12:32:30 Hey if I wanted to create a recipe for composer (which auto-builds dependencies for php-related things), what's the naming convention, composer-php? Sep 09 12:32:35 It's not a lib. Sep 09 12:33:27 n01: distros usually don't fix the kernel's provider. poky-tiny does Sep 09 12:34:18 ant_work: I saw. this means that I have to write a very very minimal distro by myself Sep 09 12:34:40 n01: you can build distroless :) Sep 09 12:34:41 poky is definitely huge for me Sep 09 12:34:53 ant_work: how? :) Sep 09 12:35:28 DISTRO = "" Sep 09 12:35:29 http://www.openembedded.org/wiki/OE-Core_Standalone_Setup Sep 09 12:36:07 I just don't declare it Sep 09 12:36:14 great :)) thank you! Sep 09 12:38:28 http://cgit.openembedded.org/openembedded-core/tree/meta/conf/distro/include/default-distrovars.inc#n47 Sep 09 12:38:50 default 'distroless' is "oe-core.0" Sep 09 12:39:35 n01: not eyou'll lose th epoky customizations Sep 09 12:39:47 argh this spacebar Sep 09 12:40:25 Hey guys - I've got a non yocto specific question. I'm sure somewhere out there there must be a handy library for reading/writing text config files. You know the type I mean, "name = value" type stuff supporting strings, integers, IP's etc... I must be able to avoid re-inventing the wheel one more time? Sep 09 12:47:05 pev_ : libconfig Sep 09 12:49:59 Guest42914: That looks very interesting, thanks! Sep 09 12:52:58 pev_, if your using c++ you may be interested in the boost libraries (http://www.boost.org/) Sep 09 12:53:35 pev_, they have a boost/program_options class which will do what you want. Sep 09 12:54:41 cfo215: Thanks but I'm in sensible C land - Plus boost is an *immence* add-in for a small feature :-D Sep 09 12:55:31 pev_ kiss Sep 09 13:06:07 cfo215: : Indeed! Sep 09 13:09:25 I want to make sure "linux-firmware-rtl8192cu" gets included in my console-image. What is the best practice for this? create a cfo215-console-image recipe that extends console-image, or just add the package to console-image? Sep 09 14:08:14 Hi :) Sep 09 14:08:36 I'm looking for a (wireless) network manager for my angstrom generated raspberry pi OS Sep 09 14:09:37 I'm using anstrom v2013.06, so it's yocto 1.4 compatible Sep 09 14:10:23 fenrig: oe has connman, which has numerous different UIs Sep 09 14:12:58 rburton: ah nice, a text file is fine enough actually. I'm building a linux for a autonomous car/tank so I don't particulary need a gui. Sep 09 14:14:52 rburton: for my wireless stick to work i probably need to include firmware or some sorts? Sep 09 14:25:18 Heya, what can I do if I get a conflict between files, one of them originating from a package that no longer exists? That package installed a certain file, but it was never necessary, so now it is removed. However, my image recipe fails because of a conflict between them... Sep 09 14:34:17 I cannot install "linux-firmware" because of a missing dependency "linux-firmware-iwlwifi-license" Sep 09 15:16:08 halstead: ping Sep 09 15:23:37 Hey, does anyone know what the corrosponding install command would be to cp -r ${S}/public ${D}/var/www/website/ ? Install -d seemingly didn't install anything, and -m complains about directories. Sep 09 15:23:46 Maybe cp -r is the right way, though. Sep 09 15:23:58 I want to install the directory and all its subfiles and dirs into D Sep 09 15:25:51 afaik install can't recurse, cp is the best bet Sep 09 15:43:47 kergoth, Alright, cool, thanks. Sep 09 15:43:55 Stygia, kergoth just a though, but would 'find' with the '--exec' switch work for you? Sep 09 15:43:59 Anyone remember what the print/warn function was? Sep 09 15:44:17 cfo215, I suppose, but I really want to just copy this entire directory verbatim. Sep 09 15:44:27 cp -r matches the usecase right. Sep 09 15:44:33 Why would you want to use find for sucha thing? cp exists for a reason, it copies, thats waht it does. trying to reinvent it doesn't make sense Sep 09 15:44:36 It was something like bbwarn or bbern or something. Sep 09 15:44:44 bbnote, bbwarn, bberror, bbfatal, yes Sep 09 15:45:09 I call a script and suspect it is failing, but it doesn't tell me anything during do_install, so. Sep 09 15:45:11 just, saying.. if one does work in a particular case, maybe you can work around with an other... Sep 09 15:45:25 cp does work. he was just asking if install could handle that case also Sep 09 15:45:33 to which the answer is no Sep 09 15:45:45 sorry, got into the conversation late. Sep 09 15:46:27 Yup, useable answer, I don't mind using cp, just wondered if I could use the bb tools instead. Sep 09 15:47:06 Hmm can I make bbnote show me the results of a command? When I use bbnote bash script it just prints that Sep 09 15:47:41 install isn't a bb tool :) it comes from coreutils, our use of it is just a convention Sep 09 15:47:54 Running it manually works, but the package isn't populated, so I strongly suspect it's failing somehow. Sep 09 15:47:58 i don't understand the qeustion. any output from the commands you run already goes to the log Sep 09 15:48:01 kergoth, Ah right, fair enough. Sep 09 15:48:10 kergoth, .. huh. Well alright. Sep 09 15:48:22 kergoth, Would it stop compilation if it got a non-zero exit status? Sep 09 15:48:22 so it seems whatever command you're running is either working, or failing silently :) Sep 09 15:48:26 yes Sep 09 15:48:29 all our tasks run with set -e Sep 09 15:48:37 halstead: ping Sep 09 15:48:42 hmm, even when I say "bash ${S}/script"? Sep 09 15:49:11 if anything you run from a task returns a non-zero exit code, the task will fail Sep 09 15:49:19 kergoth, Alright, gotcha. Sep 09 15:50:38 now, if you run a command inside that script (${S}/script) and it returns a non-zero exit code, and that script doesn't set -e, of course that bash command might not return non-zero :) Sep 09 15:51:33 scottrif, pong. Sep 09 15:59:20 kergoth, Ah hmm, the thing is if I CD to that dir and run the script it does what I expect - i.e. populate some subdirs Sep 09 15:59:42 kergoth, good change the guy who wrote it didn't consider our need to specify D, though... Sep 09 16:07:45 halstead: there are some servers in my cube here in JF1. I had to use the internet connection for work. I pulled it from the one that was on. Sep 09 16:09:59 I can't install linux-firmware because of linux-firmware-iwlwifi-license Sep 09 16:10:18 scottrif, No problem. I didn't realize you were back this week. Sep 09 16:10:30 Is there some way to route STDOUT from scripts I call to bbnote? I'd love to be able to see all output from the script I run (external script run by do_install) during the build state, and it seems to be surpressed Sep 09 16:11:24 The real issue I have is that I have a recipe that runs a setup script to install some files. These files do not seem to be anywhere in tmp/ although it works when I cd there and run the script Sep 09 16:11:39 The script sets a BASE_PATH value based on the pwd, so I cd to ${S} before I run it. Sep 09 16:13:37 scottrif, I'll come in to help find a new spot for those. Sep 09 16:17:24 Stygia: again, stdout/stderr of everything you run automaitcally goes to the log Sep 09 16:17:35 Stygia: to see this is true, run 'echo foo' in your task and view the log Sep 09 16:18:28 halstead: thanks Sep 09 16:55:45 oi Sep 09 18:40:57 hello. yocto 1.5 is in the M4 phase. at M5, will recipes for new package versions still be accepted? for example, a new gstreamer release Sep 09 18:45:14 hi I want to add wifi with usb adapter to my raspberry pi (Realtek RTL8188CUS) Sep 09 18:45:25 and it recognizes the device with lsbusb Sep 09 18:45:43 but it doesn't seem to have any driver for it (allthough I included linux-firmware in the image) Sep 09 18:54:53 fenrig, do you have this edimax mini-stick? Sep 09 18:54:54 fenrig: did you enable the appropriate driver in the kernel? Sep 09 18:55:46 fenrig, there are two drivers, one from realtek, one in the mainline kernel. the default rpi kernel uses the realtek driver (it's patched in), which is rather awful Sep 09 18:56:29 fenrig, no additional firmware is necessary for this driver, and it actually works but you need be careful which instructions you follow on the internet :) Sep 09 19:44:48 Hi why does rfkill soft block my wireless adapter (I can't seem to unblock it with rfkill) Sep 09 19:46:59 I am using meta-qt5 layer and when I build an sdk qmake isn't being installed into the sdk. I tried adding TOOLCHAIN_HOST_TASK += "qtbase-native" but it appears that a native package hasn't been built. Has anyone got qtbase-native installed into the sdk? Sep 09 19:57:10 $ runqemu qemux86 gives this error "Unable to parse conf/bitbake.conf" Sep 09 19:57:18 what can be the possible causes Sep 09 20:14:21 gjohnson: you can't install a -native package into an sdk, as -native packages dont' exist. -native recipes are built to be run by other recipes at build time. i expect you want nativesdk, instead. Sep 09 20:16:19 kergoth: you are correct I am looking for a nativesdk for qt5 but I found an irc log from July 22 in which bluelightning indicated there is no SDK support in the qt5 layer yet. Sep 09 20:16:19 "$ runqemu qemux86 " gives the following error "Unable to parse conf/bitbake.conf" Sep 09 20:16:38 any explanations, i cannot decipher it Sep 09 20:17:37 GeekGirl: Do you need to source oe-init-build-env? Sep 09 20:17:54 i have built the image, it was my first build Sep 09 20:18:09 i have my image under tmp/deploy/images Sep 09 20:18:24 and i have set the machine as qemux86 in the localconf file too Sep 09 20:19:02 PS: i had to add the scripts dir in the poky directory to my PATH variable, if this gives some clue Sep 09 20:19:53 as gjohnson says, you likely didn't runqemu in the same terminal you did your builds in, as your env isn't set up properly Sep 09 20:19:59 source the setup script Sep 09 20:20:24 okay, so do i need to run the build again? Sep 09 20:20:31 no Sep 09 20:20:37 no one said you had to run the build again Sep 09 20:20:39 just hte setup script Sep 09 20:20:48 okay... Sep 09 20:21:50 $ source poky/oe-init-build-env, this is it, i suppose Sep 09 20:21:59 15:17 < gjohnson> GeekGirl: Do you need to source oe-init-build-env? Sep 09 20:23:55 stil getting the same error Sep 09 20:23:57 ERROR: Unable to parse conf/bitbake.conf: ParseError at conf/bitbake.conf:18: unparsed line: 'xport exec_prefix = "/usr"' Sep 09 20:24:28 sounds lik your bitbake.conf is broken. Sep 09 20:24:37 any fix for that Sep 09 20:24:38 it didn't ship that way, so likely you were editing it and accidentally deleted a character Sep 09 20:24:42 that line should read 'export', not 'xport' Sep 09 20:24:57 add the missing 'e' and it should be fine, unless there's something else going on Sep 09 20:25:27 okay... Sep 09 20:25:40 and is this the conf dir in my build directory? Sep 09 20:26:28 no Sep 09 20:26:31 poky/conf/bitbake.conf Sep 09 20:26:37 er, poky/meta/conf/bitbake.conf Sep 09 20:26:47 there's no bitbake.conf in your build directory Sep 09 20:28:31 thanks a lot @kergroth, that really helped a lot.. my qemu started up Sep 09 20:28:43 \m/ for u Sep 09 20:28:43 no problem Sep 09 20:29:04 all hail the kernel ;) Sep 09 21:29:54 is there a way to build a SDK *without* a native python? The version included there has some severe problems for my use case, and a working version is available on the host anyways Sep 09 21:30:41 I mean that the SDK does not contain a native python, it's obviously needed for building the SDK Sep 09 22:06:02 imphil: removing python from the list of tools included should work Sep 09 22:08:51 RP, it's not explicitly included, but it goes in as dependency, e.g. for gdb-cross-canadian-arm Sep 09 22:09:34 imphil: you'll probably have to override those dependencies then Sep 09 22:10:07 RP, something like ASSUME_PROVIDED does not exist for this case? Sep 09 22:11:33 imphil: look at populate_sdk_rpm, INSTALL_PROVIDENAME_RPM in particular, assuming you're using rpm Sep 09 22:38:11 RP, this does not seem to work for explicit dependencies of packages. perl is also installed even though it's listed there (it's a dependency of nativesdk-autoconf) Sep 09 22:39:00 RP, I think I'll just go with the hackish way of simply rm usr/bin/python from the SDK again in a SDK_POSTPROCESS_COMMAND Sep 09 22:48:08 imphil: you probably have to list all the dependencies that are pulling in python Sep 09 23:00:12 RP, I got it now to work by overriding the gdb recipe, thanks! interestingly, the gdb-cross is compiled without python, while the gdb-cross-canadian is with python Sep 10 00:35:52 Hi Sep 10 00:36:01 I'm looking for a automatic rtc update program :) Sep 10 00:36:06 for my raspberry pi Sep 10 00:36:10 (so I can use GIT) Sep 10 01:24:02 when I generate a meta-toolchain-qt with angstrom v2013.06 for the raspberry pi, and I try to add '/usr/local/oecore-i686/sysroots/i686-angstromsdk-linux/usr/bin/qmake' to qcreator, qmake seems to crash :/ so I can't add it to qtcreator for cross compilation Sep 10 01:24:34 note that angstrom v2013.06 is yocto compatible, is it maybe because angstrom uses linaro toolchain? Sep 10 02:22:34 fenrig: i'd suspect qmake first, as external toolchains are supported Sep 10 02:23:05 nerdboy: can you explain what you are trying to say? Sep 10 02:23:31 i'd look at qmake first Sep 10 02:24:04 that just comes from 10+ years of patching upstream build systems... Sep 10 02:24:34 so why linaro/angstrom on the pi? Sep 10 02:27:12 i just used poky/meta-openembedded master and started adding things i needed to meta-raspberrypi... Sep 10 02:27:48 i did take a look at angstrom but haven't built anything with it yet Sep 10 02:28:14 maybe i get enough angstrom/oe-classic at work i guess Sep 10 02:31:59 fenrig: did you get any error messages at all? Sep 10 02:32:29 can you try running qmake with strace? Sep 10 02:32:36 nerdboy: I can. Sep 10 02:34:55 https://gist.github.com/fenrig/6504329 Sep 10 02:37:29 looks like you get a core dump Sep 10 02:37:42 you can open it with gdb Sep 10 02:38:04 other than that, nothing horrendous stands out Sep 10 02:38:35 strange Sep 10 02:39:24 i tried to stay away from bloaty stuff Sep 10 02:39:40 in my rpi-xorg image... Sep 10 02:40:23 webkit-gtk is probably the bloatiest thing in it right now Sep 10 02:42:36 I need Qt :D Sep 10 02:42:43 it's a very lovely language Sep 10 02:42:48 because of the signal mapper Sep 10 02:43:08 and the event qeueu Sep 10 02:44:12 what about poky/meta-qt5 ? Sep 10 02:45:32 that should work with the internal toolchain, in theory... Sep 10 02:47:35 there's also a qt demo image of some sort, but i haven't tried building it Sep 10 02:49:01 meta/recipes-qt/images/qt4e-demo-image.bb Sep 10 02:50:01 nerdboy: meta-qt5 doesn't support building a nativesdk Sep 10 02:50:04 yet Sep 10 02:53:06 well, something's dorked in that segfaulting qmake binary Sep 10 02:53:25 running gdb on the core file is my best suggestion Sep 10 02:56:40 nerdboy: I'll look into it tomorrow, better get some sleep. thx for helping **** ENDING LOGGING AT Tue Sep 10 03:00:00 2013