**** BEGIN LOGGING AT Mon Oct 15 03:00:02 2012
Oct 15 07:15:38 <afournier> hi there !
Oct 15 07:16:39 <afournier> is it a good practice to make a recipe that will install a public key for the root user ? 
Oct 15 07:21:34 <silvio> morning all
Oct 15 07:32:35 <rob_w> i dont get the deal with the LIC_FILES_CHKSUM  ... am i supposed to provide a COPYING file in each pacakge ?
Oct 15 07:32:43 <mckoan> good morning
Oct 15 07:41:01 <silvio> hi afournier , hi mckoan 
Oct 15 09:00:22 <ericben> good morning
Oct 15 09:01:51 <ericben> what could prevent git fetcher to generate a tar.gz after fetching sources from the git server ? I have one recipe where the .tar.gz is not generated and for the moment I don't see where is the problem here.
Oct 15 09:05:13 <bluelightning> ericben: AFAIK only BB_GENERATE_MIRROR_TARBALLS controls that...
Oct 15 09:08:42 <ericben> bluelightning: that's waht I was thinking and it's enable as other tarballs are generated but not this one !
Oct 15 09:09:51 <ericben> that's very strange. I'll try to find some time to investigate this
Oct 15 09:10:30 <bluelightning> indeed
Oct 15 09:10:58 <bluelightning> throw a few debug prints into the code in lib/bb/fetch2/ and see what code is being executed
Oct 15 09:55:15 <afournier> is it possible to populate a mysql database with oe before build an image ? like running mysql-native first, executing a mysql script, and then copying the database files to the image ?
Oct 15 09:58:59 <bluelightning> afournier: you can put pretty much whatever you like into a shell function and then call it from ROOTFS_POSTPROCESS_COMMAND
Oct 15 09:59:59 <afournier> ha ok
Oct 15 10:32:44 <pb_> morning all
Oct 15 10:34:03 <woglinde> he pb
Oct 15 10:36:58 <bluelightning> hi pb_, woglinde
Oct 15 10:38:57 <pb_> hi woglinde, bluelightning
Oct 15 10:39:43 <woglinde> hi bluelightning
Oct 15 10:52:28 <hrw> I am getting tired of gnu-config updates
Oct 15 10:52:39 <hrw> today modphp got added to list
Oct 15 10:53:36 <woglinde> who needs modphp?
Oct 15 10:55:19 <hrw> my lamp users
Oct 15 10:55:27 <pb_> well, it's pretty vital if you're trying to create that authentic late-1990s web experience
Oct 15 10:58:11 <stefan_schmidt_w> pb_: make sure you also mimic all the security bugs :)
Oct 15 10:58:42 <pb_> right, I'm working on a recipe for Linux 1.2 at present
Oct 15 10:59:16 <pb_> and I will soon be sending patches to add libc4 and DLL support to oe-core
Oct 15 11:00:47 <bluelightning> heh
Oct 15 11:03:33 <bluelightning> I think "late-1990s" is a bit harsh though, PHP's heyday continued for at least a few years after that
Oct 15 11:04:30 <pb_> yeah, true enough
Oct 15 11:04:31 <hrw> pb_: libc4... Sounds too scary for me to even ask
Oct 15 11:04:49 <pb_> I was actually still writing php code as recently as 2010, so I guess I can't really cast too many aspersions.
Oct 15 11:05:08 <hrw> when I started Linux I read FAQ entries about Watchtower being awful and it was libc5...
Oct 15 11:05:41 <hrw> I got job as PHP programmer in 2001 
Oct 15 11:06:40 <bluelightning> hmm, I see php 5.3.17 is already out... probably should update the recipe in meta-webserver (already)
Oct 15 11:08:14 <hrw> bluelightning: want recipe for 5.4.4?
Oct 15 11:08:26 <hrw> bluelightning: but for normal php, not modphp
Oct 15 11:08:51 <bluelightning> hrw: I had intended to move php-cgi from meta-oe to meta-webserver, so yes
Oct 15 11:09:10 <bluelightning> (yes please :)
Oct 15 11:09:49 <hrw> bluelightning: http://tygrysek.juszkiewicz.com.pl/~hrw/oe/
Oct 15 11:31:46 <Zagor> why does the bzip2 recipe add a Makefile.am and configure.ac that doesn't exist upstream?
Oct 15 11:32:34 <Zagor> just to more easily piggyback on autotools.bbclass?
Oct 15 11:32:46 <Zagor> it looks a bit strange to me
Oct 15 11:37:16 <woglinde> becaue it was easier than to fix the stupid makfefile for crosscompiling?
Oct 15 11:44:06 <Zagor> woglinde: that's one theory, but the Makefile actually looks very simple
Oct 15 11:45:09 <Zagor> or, sorry, are you saying that was the actual reason?
Oct 15 11:57:21 <hrw> can layers have dependencies on other layers in automated way?
Oct 15 12:00:36 <bluelightning> hrw: with LAYERDEPENDS yes
Oct 15 12:00:53 <hrw> cool
Oct 15 12:01:05 <bluelightning> it's not widely used yet, but it does work
Oct 15 12:01:11 <bluelightning> meta-intel uses it at least
Oct 15 12:01:36 <jackmitchell> hrw: are you planning on pushing it to meta-oe?
Oct 15 12:01:47 <hrw> jackmitchell: php?
Oct 15 12:01:51 <jackmitchell> hrw: yes sorry, php
Oct 15 12:02:13 <hrw> jackmitchell: this recipe was not even tested does it work
Oct 15 12:02:17 <hrw> it builds
Oct 15 12:02:43 <hrw> so it may land in my contrib tree to post it later
Oct 15 12:03:00 <jackmitchell> hrw: ah ok, I could possibly test it sometime today if I can get latest git to build for me
Oct 15 12:03:20 <hrw> cool
Oct 15 12:03:33 <bluelightning> jackmitchell: if you have a moment could you add the LSB headers to the initscript for haiwatha and then send a v2 of your patch to add it to meta-webserver?
Oct 15 12:04:13 <jackmitchell> bluelightning: yes, that is on my list for today - I have no experience with the LSB headers so I need to have a bit of a read
Oct 15 12:04:43 <jackmitchell> bluelightning: I assume they're the headers that can be converted for automatic systemd service creation, yes?
Oct 15 12:04:45 <bluelightning> jackmitchell: neither did I until Koen complained the ones for apache were missing ;)
Oct 15 12:04:57 <bluelightning> jackmitchell: I understand it helps for that, yes
Oct 15 12:07:22 <hrw> bluelightning: looks like LAYERDEPENDS is good only when all layers share repo
Oct 15 12:07:35 <bluelightning> hrw: no, that's not a requirement
Oct 15 12:08:03 <jackmitchell> while I'm still not 100% convinced on systemd I'm sure not a fan of init scripts
Oct 15 12:08:17 <bluelightning> jackmitchell: I'm not using systemd yet either
Oct 15 12:08:39 <hrw> bluelightning: I have: meta-{aarch64,linaro,oe} + openembedded-core and use meta-{aarch64,linaro,oe,webserver} + toolchain-layer + oe-core layers. have to play
Oct 15 12:08:51 <jackmitchell> bluelightning: i'm looking forward to making the jump on my embedded project when 1.4 brings in official support
Oct 15 12:09:16 <jackmitchell> bluelightning: I suppose I should also transition on my host system too in order to get a good foothold ;)
Oct 15 12:09:27 <bluelightning> jackmitchell: yes I will be pleased when we have support for it in core, a lot of people ask for it
Oct 15 12:10:01 <bluelightning> hrw: it will work fine as long as the collection names used in each layer's layer.conf are distinct
Oct 15 12:11:19 <jackmitchell> bluelightning: I'm a bit worried about the hardcoded sbin path in the hiawatha init file, is this ok or can I replace it with a variable nicely (aka without an iffy sed command)
Oct 15 12:11:19 <hrw> bluelightning: but how bitbake will find out where to check for layers?
Oct 15 12:11:40 <hrw> bluelightning: I will add pastebin
Oct 15 12:11:54 <bluelightning> hrw: it looks at whatever layers you have configured in bblayers.conf
Oct 15 12:12:26 <hrw> bluelightning: when I add all needed layers in bblayers I do not need LAYERDEPENDS 
Oct 15 12:12:28 <bluelightning> jackmitchell: sed is really the only way to handle that effectively
Oct 15 12:12:45 <bluelightning> hrw: that's what LAYERDEPENDS forces you to do, that's the purpose of it...
Oct 15 12:12:57 <bluelightning> hrw: if a needed layer is missing it bails out straight away
Oct 15 12:13:07 <hrw> it is not
Oct 15 12:13:14 <hrw> ERROR: Nothing PROVIDES 'modphp'
Oct 15 12:13:34 <jackmitchell> bluelightning: /me expected as such - over to the sed tutorials again ;)
Oct 15 12:13:37 <hrw> it is not "ERROR: There is no 'webserver' layer"
Oct 15 12:13:52 <bluelightning> hrw: so, how have you set LAYERDEPENDS then?
Oct 15 12:14:37 <hrw> bluelightning: bblayers has: meta-aarch64, meta-oe, oe-core. meta-aarch64 layerdepends on linaro, webserver. linaro depends on toolchain-layer
Oct 15 12:14:59 <hrw> bluelightning: when I add all 6 layers into bblayers I do not need layerdepends.
Oct 15 12:15:19 <hrw> bluelightning: if I skip one of required ones I will get 'error: nothing provides'
Oct 15 12:15:29 <bluelightning> then LAYERDEPENDS has not been set properly
Oct 15 12:15:41 <hrw> LAYERDEPENDS = "toolchain-layer"
Oct 15 12:15:55 <bluelightning> you need to do LAYERDEPENDS_collectionname
Oct 15 12:16:03 <hrw> ok
Oct 15 12:16:08 <bluelightning> same as when setting BBFILE_PATTERN etc.
Oct 15 12:16:35 <hrw> thx
Oct 15 12:16:50 <gizero> hi! trying to figure out how distrodata class works. The goal is to automate tracking upstream releases of sources I've provided a recipe for. I'd like to have some kind of automatic alert when a new release comes out. Not yet read all the source but noticed that upstream version check for one of those sources fais. Looks like check_new_version() is unable to list download dir content from SRC_URI, which is on github. I susp
Oct 15 12:16:53 <gizero> ect this could be affecting any source archive coming from github. Anyone here with experience on distrodata internals?
Oct 15 12:17:22 <hrw> bluelightning: now it works. thanks
Oct 15 12:17:22 <gizero> s/fais/fails ;-)
Oct 15 12:17:27 <bluelightning> hrw: np
Oct 15 12:21:38 <hrw> bluelightning: can you add 'gnu-configize --force' into do_configure of modphp?
Oct 15 12:22:13 <bluelightning> hrw: sure
Oct 15 12:23:02 <hrw> thx
Oct 15 12:23:12 <hrw> one bbappend less to write
Oct 15 12:31:09 <jackmitchell> my sed skills are majorly lacking, can anyone see why this doesn't work?   sed -i 's/sed_sbin_path/${sbindir}/' ${WORKDIR}/hiawatha-init
Oct 15 12:31:30 <jackmitchell> do I have to escape the ${} or something odd?
Oct 15 12:31:50 <bluelightning> no, that should be fine... the variable substitution happens long before it gets executed
Oct 15 12:32:12 <jackmitchell> it seems to break a sed further down the line and I hit this message
Oct 15 12:32:30 <jackmitchell> | sed: -e expression #1, char 18: unknown option to `s'
Oct 15 12:33:40 <bluelightning> jackmitchell: oh, I see what's going wrong
Oct 15 12:33:50 <bluelightning> you need to use a different separator character instead of /
Oct 15 12:33:55 <bluelightning> since the path contains /
Oct 15 12:34:01 <bluelightning> e.g. ! or :
Oct 15 12:34:08 <afournier> :)
Oct 15 12:35:00 <jackmitchell> bluelightning: ah, good shout - I did wonder why all the other examples I looked at used odd seperators
Oct 15 12:35:25 <bluelightning> jackmitchell: the separator can be almost anything you like basically, it's whatever char follows s
Oct 15 12:43:37 <pb_> of course, whatever character you pick, you will lose if the path happens to contain that character :-}
Oct 15 12:45:20 <bluelightning> pb_: sure, but anyone who has ! in a path deserves everything they get IMO ;)
Oct 15 12:46:09 <pb_> yeah, true enough
Oct 15 12:46:26 <bluelightning> we're pretty unforgiving on paths in general in OE, we don't even allow spaces
Oct 15 12:46:30 <pb_> it might be nice to try to standardise on ! or something so that there aren't too many different random separators in use
Oct 15 12:46:45 <bluelightning> yes, that would be good
Oct 15 12:47:14 <pb_> right now I think !, :, @ and no doubt other things are all used by some or other piece of code, and obviously the more different ones are involved the greater the chance that someone will encounter a path that uses one of those characters.
Oct 15 12:47:50 <jackmitchell> I used ',' as it seemed less offensive than using ! which to me, carries certain connotations 
Oct 15 12:48:23 <jackmitchell> ':' would have also been a good choice
Oct 15 13:53:17 <rob_w> whats the command to fully(!) clean a build of a recipe .. seems -c clean is not strong enough anymore ? 
Oct 15 13:54:34 <Noor> rob_w:  you want to remove sharedstate as well?
Oct 15 13:55:01 <afournier> -c cleansstate
Oct 15 13:56:04 <Noor> yeah cleansstate cleans the package and its sharedstate as well
Oct 15 13:57:05 <rob_w> hmm ok
Oct 15 14:04:31 <jackmitchell> hrw: on your php recipe I get the error : Applying patch php_exec_native.patch
Oct 15 14:04:31 <jackmitchell> patching file configure.in
Oct 15 14:04:31 <jackmitchell> Hunk #1 FAILED at 1216.
Oct 15 14:04:31 <jackmitchell> 1 out of 1 hunk FAILED -- rejects in file configure.in
Oct 15 14:04:31 <jackmitchell> Patch php_exec_native.patch does not apply (enforce with -f)
Oct 15 14:05:46 <jackmitchell> hrw: I'm just using the php.inc included with meta-oe could this have something to do with it?
Oct 15 14:07:48 <hrw> jackmitchell: I copied those files from my meta-oe dir. php.inc was not changed
Oct 15 14:08:16 <hrw> jackmitchell: I assume you fetched php-5.4.4 directory as well
Oct 15 14:09:21 <jackmitchell> hrw: ok I think it was my problem, I already had that file in my php directory and it had renamed your version to php_exec_native.patch.1 and  so was using the old version, it seems to be chugging along now
Oct 15 14:09:33 <hrw> ;)
Oct 15 14:09:45 <rob_w> i dont get it .. i am working on denzil branches .. created my own image recipe and add packages .. now for what reason ever , packages like udev, opkg etc are not showing up in the image ?
Oct 15 14:09:52 <hrw> jackmitchell: it was named php-5.4.4 for a reason ;)
Oct 15 14:10:12 <jackmitchell> hrw: yeah; call me lazy :P
Oct 15 14:10:13 <hrw> rob_w: 'your own image recipe' is at fault?
Oct 15 14:10:43 <hrw> rob_w: you lack packagemanagement group/feature and set /dev manager wrong?
Oct 15 14:10:44 <rob_w> i guess so .. but i dont see what i can do wrong while extind the IMAGE_INSTALL =" " stuff
Oct 15 14:10:59 <hrw> rob_w: IMAGE_INSTALL +=
Oct 15 14:11:23 <rob_w> well i use = and add a long list .. not good ?
Oct 15 14:11:37 <hrw> rob_w: make yourself a favour and do not write image from scratch - use one of exising ones and add packages to it
Oct 15 14:11:48 <hrw> been there, did that ;D
Oct 15 14:12:05 <rob_w> well i based on core-minimla-image .. basicly
Oct 15 14:12:41 <hrw> rob_w: IMAGE_INSTALL can be changed by image.bbclass - so you first inherit classes and then adds to it
Oct 15 14:13:50 <rob_w> so i start again .. who know .. is core-minimal-image a good reference ?
Oct 15 14:15:02 <bluelightning> nothing wrong with it as a base
Oct 15 14:16:34 <bluelightning> but you almost certainly need to include task-core-boot in your IMAGE_INSTALL (as it does)
Oct 15 14:16:37 <bluelightning> rob_w: ^
Oct 15 14:20:29 <rob_w> hmm i copie core-image-base.bb again and it rebuilds quite some packages now .. wonder if i screwed with whitespaces and fucked up the previous recipe
Oct 15 14:29:21 <gchiii> does anyone have any hints on what it takes to get X to support multitouch on an oe-classic 
Oct 15 14:29:26 <gchiii> build
Oct 15 14:32:47 <autopilot> hi, is anybody active?
Oct 15 14:32:59 <rob_w> no
Oct 15 14:36:12 <autopilot> trying to boot with an uncompressed fs rather than lzma, getting the following error: EXT2-fs (ram0): error: ext2_check_page: bad entry in directory #557: : unaligned directory entry - offset=0, inode=453005403, rec_len=19803, name_len=0
Oct 15 14:36:29 <autopilot> would anybody know what this message means?
Oct 15 14:39:54 <afournier> autopilot: have you done an fsck on it ?
Oct 15 14:42:21 <autopilot> afournier: i'm not sure this busybox was built with ext2 support: fsck: fsck.ext2: No such file or directory
Oct 15 14:43:17 <afournier> you should grab the image, and fsck it on another running linux
Oct 15 14:45:06 <afournier> it looks like your fs is corrupted
Oct 15 14:45:39 <rob_w> i am missing tinylogin , udev opkg and such basic stuff .. so i miss ?
Oct 15 14:45:48 <autopilot> afournier: it's a ramdisk image, which means it has to be mounted, is that right? it's not in a state that we can unmount it
Oct 15 14:46:22 <bluelightning> rob_w: do you have task-core-boot as I mentioned above?
Oct 15 14:46:41 <afournier> autopilot: you should grab that ramdisk image, save it somewhere on your linux, losetup /dev/loop0 this-image; fsck -t ext2 /dev/loop0; losetup -d /dev/loop0
Oct 15 14:48:06 <rob_w> http://pastebin.com/TxLjiSsJ  <-  bluelightning 
Oct 15 14:49:43 <rob_w> should i mention that i run my own MACHINE ?
Oct 15 14:56:18 <bluelightning> rob_w: note that to get package management you will need package-management in IMAGE_FEATURES
Oct 15 14:56:49 <bluelightning> that will install opkg (assuming you have selected ipk packaging) and ensure the package manager db is left in place
Oct 15 14:59:14 <rob_w> to select ipk packageing i do PACKAGE_CLASSES = "package_ipk" in my local.conf , right ?
Oct 15 14:59:15 <autopilot> afournier: fsck didn't fix the issue. i believe the fs is good, but isn't being loaded/init properly somehow, just not sure how. it boots fine when it's compressed, but it always shows above errors when it boots uncompressed
Oct 15 14:59:19 <bluelightning> rob_w: yes
Oct 15 14:59:31 <bluelightning> rob_w: so you're doing += after the inherit core-image, that should mean you don't need to add task-core-boot or indeed udev, initscripts etc
Oct 15 14:59:50 <rob_w> right
Oct 15 15:00:00 <rob_w> but i was missing the full IMAGE_FEATURES thing 
Oct 15 15:00:02 <bluelightning> rob_w: you can check what IMAGE_INSTALL is actually being set to using bitbake -e imagename | grep ^IMAGE_INSTALL=
Oct 15 15:00:35 <afournier> autopilot: can you load it at different physical address in the linux config ?
Oct 15 15:00:51 <afournier> autopilot: maybe a part of it get overwritten somehow
Oct 15 15:03:15 <autopilot> afournier: i suspect the same thing. unfortunately i'm not very informed on the boot process and documentation seems to be lacking in this subject. nothing (obvious) seems to need the space we've assigned to the filesystem (it is exactly 8MB, residing in a 11MB partition)
Oct 15 15:04:10 <afournier> autopilot: it's small enough to me, but the interesting thing would be to know where it gets loaded in the physical memory...
Oct 15 15:04:29 <afournier> autopilot: i think there is an option/offset you can change in the linux menuselect
Oct 15 15:04:46 <afournier> autopilot: what arch is it ?
Oct 15 15:05:09 <autopilot> arm7
Oct 15 15:05:32 <autopilot> do you know which menu in the kernel config that option would live?
Oct 15 15:06:23 <afournier> i don't remember exactly, but it should be next to initrd stuff
Oct 15 15:06:41 <afournier> try typing "/" when running menuselect
Oct 15 15:06:47 <afournier> and look for INITRD
Oct 15 15:07:33 <afournier> how do you specify your initrd ?
Oct 15 15:07:41 <afournier> using the kernel cmdline from the bootloader ?
Oct 15 15:08:18 <jackmitchell> hrw: the php recipe seems to work, I am now using it for a very basic site (some html generation and sqlite queries), through the php-cgi binary
Oct 15 15:12:25 <afournier> hrw, jackmitchell, could i take a look at your php recipe ?
Oct 15 15:16:07 <otavio> Hello
Oct 15 15:16:23 <otavio> I've been using the meta-toolchain output ...
Oct 15 15:16:33 <otavio> and it seems not easy to override the sysroot to use
Oct 15 15:16:53 <otavio> this is easy to do using eclipse but not using command line
Oct 15 15:17:22 <autopilot> afournier: it's specified on a redboot exec script
Oct 15 15:17:31 <jackmitchell> afournier: it's the one out of meta-oe
Oct 15 15:17:40 <bluelightning> otavio: are you using the setup script and variables that it sets (CC etc.)?
Oct 15 15:17:52 <otavio> yes
Oct 15 15:17:52 <jackmitchell> afournier: hrw made an update a later version of PHP that I was testing
Oct 15 15:18:06 <afournier> ah ok
Oct 15 15:18:07 <bluelightning> otavio: perhaps you could describe the problem further then?
Oct 15 15:18:09 <afournier> i have to go
Oct 15 15:18:11 <jackmitchell> afournier: http://tygrysek.juszkiewicz.com.pl/~hrw/oe/php-5.4.4/
Oct 15 15:18:13 <otavio> bluelightning: but OECORE_TARGET_SYSROOT is copied in many variables
Oct 15 15:18:23 <afournier> thanks :)
Oct 15 15:18:24 <jackmitchell> afournier: it still requires the php.inc from meta-oe however
Oct 15 15:18:40 <afournier> ok
Oct 15 15:18:52 <bluelightning> otavio: are you using the relocatable version now available in master or an earlier one?
Oct 15 15:18:54 <afournier> see you later
Oct 15 15:19:19 <otavio> bluelightning: relocable ... but I also checked the source
Oct 15 15:19:33 <otavio> bluelightning: let me get the source point for you
Oct 15 15:22:08 <otavio> bluelightning: http://git.openembedded.org/openembedded-core/tree/meta/classes/toolchain-scripts.bbclass#n8 lines: 14 is an example, but it should use the variable defined in 44
Oct 15 15:25:30 <hrw> jackmitchell: good to know it works fine
Oct 15 16:06:52 <hrw> bluelightning: once apache2 and modphp are in place and apache is running. how to get it working without editing /etc/apache2/modules.d/*php*.conf to remove ifdefine PHP5?
Oct 15 16:08:18 <bluelightning> hrw: the version of apache currently in meta-webserver adds -D PHP5 on the command line so you don't have to do that
Oct 15 16:10:49 <hrw> <?php phpinfo() ?> returned
Oct 15 16:12:31 <bluelightning> hrw: is the file called .php ?
Oct 15 16:13:08 <hrw> yes
Oct 15 16:13:26 <bluelightning> well, not sure... it works here...
Oct 15 16:14:53 <hrw> let me reboot qemu
Oct 15 16:16:14 <hrw> sorry, works now
Oct 15 16:16:24 <hrw> nice job bluelightning 
Oct 15 16:16:47 <bluelightning> hmm, so why didn't it work without rebooting?
Oct 15 16:18:05 <hrw> lets say pebkac
Oct 15 16:19:25 <hrw> bluelightning: now I need to get mysql in modphp ;)
Oct 15 16:19:34 <hrw> plan is to have it in lamp image
Oct 15 16:19:56 <bluelightning> should be as easy as ensuring mysql is in PACKAGECONFIG for modphp
Oct 15 16:20:13 <bluelightning> I've done it in local.conf but it can also be done in a bbappend
Oct 15 16:20:32 <hrw> packageconfig is new var for me ;)
Oct 15 16:20:55 <hrw> should I set it: PACKAGECONFIG = "mysql" or otherway?
Oct 15 16:22:38 <hrw> and will it add mysql to DEPENDS?
Oct 15 16:22:53 <hrw> bluelightning: btw - for pgsql you have mysql5 as last argument
Oct 15 16:23:04 <bluelightning> oops :(
Oct 15 16:23:49 <bluelightning> if you're setting it from local.conf you should do PACKAGECONFIG_pn-modphp = "mysql"
Oct 15 16:24:49 <bluelightning> hrw: before I fix that I guess I should check that the postgres support actually works...
Oct 15 16:26:26 <hrw> I did bbappend
Oct 15 16:27:14 <hrw> bluelightning: I prefer to not check it - postgresql needs --disable-spinlocks on aarch64 which is another performance loss.
Oct 15 16:27:29 <hrw> but we still have qemux86 for that ;D
Oct 15 16:27:30 <bluelightning> hrw: I'll test it here
Oct 15 16:28:19 <hrw> bluelightning: you have lot of --disable-this-or-that :( I want wordpress running ;)
Oct 15 16:28:37 <hrw> bluelightning: can we get php-cgi and modphp ~synced in configuration?
Oct 15 16:28:43 <bluelightning> hrw: most of that is from the OE-Classic recipe
Oct 15 16:28:51 <bluelightning> it would make sense for them to be the same yes
Oct 15 16:29:06 <bluelightning> in fact I wonder if there ought to be a common inc file
Oct 15 16:31:44 <hrw> bluelightning: would be great
Oct 15 16:32:06 <hrw> bluelightning: I do wonder why php is fine with autotools.do_configure and modphp is not
Oct 15 16:38:29 <bluelightning> hrw: don't know, but I just tried, it still breaks horribly
Oct 15 16:39:44 <bluelightning> from the output it looks like the configure script is being mangled
Oct 15 16:42:25 <hrw> I saw breakage and it was awful
Oct 15 16:53:58 <hrw> anyway time to end work
Oct 15 16:54:03 <hrw> have a nice day people
Oct 15 17:08:25 * rob_w figured out that something is wrong with udev script  not beeing placed in rcS.d/
Oct 15 17:08:52 <rob_w> without that most of the system is a mess
Oct 15 18:23:15 <woglinde_> hi effem
Oct 15 18:23:30 <eFfeM> hi woglinde, all
Oct 15 18:56:48 <Crofton> what the class I add to get a list of packages in the image these days?
Oct 15 18:57:33 <Crofton> ah output moved
Oct 15 18:59:16 <Crofton> hmm
Oct 15 19:00:17 <kergoth> use buildhistory
Oct 15 19:00:20 <kergoth> it's highly recommended
Oct 15 19:00:30 <kergoth> https://wiki.yoctoproject.org/wiki/Buildhistory
Oct 15 19:01:07 <Crofton> yeah, I am getting confused with build stats
Oct 15 19:01:28 <kergoth> buildstats is task-level info about build time, etc. not going to be helpful for the image, i expect
Oct 15 19:05:23 <Crofton> right
Oct 15 19:05:28 <Crofton> I was just going over the output
Oct 15 19:07:58 * kergoth nods
Oct 15 20:48:54 <Crofton> MACHINE_EXTRA_RRECOMMENDS = "kernel-modules"
Oct 15 20:49:36 <Crofton> with this in a machine file, would I expect all images to have all kernel modules in them?
Oct 15 20:50:56 <woglinde> dont know
Oct 15 20:51:01 <woglinde> glob or regexp
Oct 15 20:57:49 <GNUtoo-desktop> I guess it's included or in images or in tasks
Oct 15 20:58:01 <GNUtoo-desktop> which is now named packagegroup
Oct 15 21:02:53 <Crofton> heh
Oct 15 21:03:22 <Crofton> the kernel recipe makes a package kernel-modules that RDEPENDS on all the moduls :)
Oct 15 22:15:01 <Crofton> do we have sshfs avilable anywhere?
Oct 15 22:20:52 <bluelightning> Crofton: I can't see it in my local index at least
Oct 15 22:30:48 <Crofton> yeah, grep not finding it
Oct 15 23:30:29 <EricBlade> hello #oe peoples :)   silly? question: the sanity checker is complaining about bad RPATHs .. what exactly does that refer to?
Oct 15 23:31:49 <kergoth> it refers to rpaths.
Oct 15 23:38:13 <EricBlade> kergoth: i.. gathered that.  more specifically, what is a bad rpath versus a good rpath? and where do the rpaths come from?
Oct 15 23:39:28 <kergoth> googling for rpath for 2 seconds will tell you quite a lot about what rpaths are and how they're set
Oct 15 23:39:40 <bluelightning> EricBlade: https://fedoraproject.org/wiki/Packaging/Guidelines#Removing_Rpath
Oct 15 23:39:43 <EricBlade> looking at the insane.bbclass, it appears that it merely checks to see if the directory in question exists, which it does appear to
Oct 15 23:44:34 <EricBlade> hmm. interesting
Oct 15 23:48:11 <EricBlade> so, it doesn't really make sense that an RPATH would contain ..../sysroots/system/usr/lib 
Oct 15 23:48:20 <EricBlade> s/would/should
Oct 15 23:49:39 <bluelightning> indeed, that's obviously bad since it would never be useful on the target
Oct 15 23:51:11 <EricBlade> interesting. ok, i think i've located it in a makefile.  now to figure out where the makefile comes up with it
Oct 15 23:53:57 <EricBlade> thanks for that link, bluelightning .. never would've found much of anything searching for it as it relates to bb/oe
Oct 15 23:54:23 <bluelightning> EricBlade: np
Oct 15 23:55:07 <bluelightning> usually there's either a --no-rpath or --disable-rpath option to the configure script; or it's libtool "helpfully" adding it
Oct 15 23:56:10 <bluelightning> ok time to get some sleep, night all
Oct 15 23:56:16 <EricBlade> in this case it appears to be something someone added to one of qmake's scripts .. but removing the -rpath declaration did not fix the error :O crap
**** ENDING LOGGING AT Tue Oct 16 03:00:01 2012