**** BEGIN LOGGING AT Fri Dec 05 02:59:58 2014
Dec 05 04:05:37 <chankit1> how do bitbake do_fetch work with checking out svn repository with http prefix?
Dec 05 04:14:17 <kergoth> chankit1: svn:// url, but add ;protocol=http to the end of the url
Dec 05 05:28:08 <chankit1> kergoth: will try it out
Dec 05 05:28:09 <chankit1> thanks
Dec 05 05:29:56 <chankit1> kergoth: now it's complaining about "..missing required parameter 'module'"
Dec 05 07:51:06 <mckoan> good morning
Dec 05 09:15:18 <irontia> good morning all
Dec 05 09:17:29 <irontia> I've added pkgconfig pc files to my library (automake setup). At the moment i install the file with a do_install_append() function, with this cmd: install -m 0644 ${WORKDIR}/build/libpsx.pc ${D}${libdir}/pkgconfig/ . It feels wrong, i probably should install the file in the sources make install, no?
Dec 05 09:36:16 <cassidy> so I'm still trying to use my SSTATE_MIRRORS. What would be the easiest way to test if it's actually used? Delete my local sstate-cache directory, run a build and check the number of 'Cached Tasks' in toaster?
Dec 05 09:38:32 <bluelightning> morning all
Dec 05 09:38:57 <bluelightning> cassidy: that would be one way yes
Dec 05 09:39:37 <cassidy> bluelightning, is there a way to enable some extra debugging to see cache hit attempts from Yocto or something like that?
Dec 05 09:47:39 <bluelightning> cassidy: short answer, other than toaster, not really
Dec 05 09:47:46 <bluelightning> cassidy: you might get something out of -DDD but that spews quite a lot of other debugging output
Dec 05 09:51:22 <cassidy> bluelightning, ok. It's still not that clear to me when the cache is actually used. For example, on a build of 1360 tasks, I only have 5 cache attempts and all of them are "File not in cache"  (and this is before removing my local sstate-cache dir)
Dec 05 10:07:32 <bluelightning> cassidy: what does bitbake print at the end of the build for the task summary?
Dec 05 10:07:40 <bluelightning> cassidy: and what value are you setting for SSTATE_MIRRORS ?
Dec 05 10:08:42 <cassidy> something like "NOTE: Tasks Summary: Attempted 3814 tasks of which 3797 didn't need to be rerun and all succeeded."  (not the actual number of this specific build). Bitbake doesn't even do a catch attempt for the "no need to rerun" tasks?
Dec 05 10:09:09 <cassidy> SSTATE_MIRRORS ?= "\
Dec 05 10:09:09 <cassidy> file://.* https://my-server.com/sstate-cache/PATH;downloadfilename=PATH"
Dec 05 10:09:31 <cassidy> where https://my-server.com/sstate-cache/ contains 0a 0b etc directories
Dec 05 10:09:41 <cassidy> (the root of the cache dir)
Dec 05 10:13:00 <bluelightning> cassidy: it will be doing a cache check for all of the tasks which are marked as being sstate tasks (probably about a third of them)
Dec 05 10:13:11 <bluelightning> whether or not it correctly reports that via toaster is a different issue
Dec 05 10:15:14 <cassidy> looks like my https server isn't in the "URI(s) searched" reported by toaster
Dec 05 10:17:20 <cassidy> is file:////mnt/sstate/qemux86hyb/PATH a default thing as well? I don't have any /mnt/sstate
Dec 05 10:17:54 <bluelightning> doesn't mean anything to me I'm afraid, but it doesn't sound right
Dec 05 10:18:52 <bluelightning> it's probably worth noting that there is a small amount of guesswork as far as what toaster captures for sstate reuse, it doesn't directly connect to the part of the build system that does the cache fetching for all of it
Dec 05 10:19:24 <bluelightning> so if it's reporting strange things it's more likely that the data collection is incorrect than the cache is not functioning (at the moment, anyway)
Dec 05 10:20:05 <cassidy> I'm trying with SSTATE_MIRRORS commented out to see if this URL is still reported
Dec 05 10:20:56 <belen> cassidy: let us know what comes out. I can file a bug if Toaster is reporting the wrong thing
Dec 05 10:22:02 <cassidy> yep
Dec 05 10:23:13 <cassidy> file:////mnt/sstate/qemux86hyb/PATH
Dec 05 10:23:13 <cassidy> is still there even without SSTATE_MIRRORS, so it's not confusing it
Dec 05 10:23:38 <cassidy> so I guess there are 2 questions here a) why do we have this URI? b) why isn't it trying my mirror
Dec 05 10:24:41 <belen> cassidy: which version are you using? 1.6 or 1.7?
Dec 05 10:26:52 <cassidy> belen, where can I see that? (not my setup)
Dec 05 10:27:59 <bluelightning> cassidy: was it fetched from git? if so what branch are you on?
Dec 05 10:28:54 <cassidy> it's been branched at some point I think, HEAD is 8e386a710023e000a504e05c13da0106df0c7f3a if that helps
Dec 05 10:29:10 <bluelightning> cassidy: also, can you run bitbake -e | less and then search for SSTATE_MIRRORS and see if somewhere else is setting SSTATE_MIRRORS ?
Dec 05 10:29:55 <cassidy> ahah !
Dec 05 10:29:57 <cassidy> SSTATE_MIRRORS=" file://.* file:////mnt/sstate/qemux86hyb/PATH \n "
Dec 05 10:30:08 <cassidy> it's starting to make some sense :)
Dec 05 10:30:12 <bluelightning> aha!
Dec 05 10:30:14 <bluelightning> ;)
Dec 05 10:31:22 <cassidy> let's see where this come from
Dec 05 10:32:15 <bluelightning> the history reported by bitbake -e should report that
Dec 05 10:38:46 <cassidy> so I have a custom specific file doing a -SSTATE_MIRRORS += "\
Dec 05 10:38:47 <cassidy> -file://.* file:////mnt/sstate/${MACHINE}/PATH \n \
Dec 05 10:39:02 <cassidy> removing this line seems to not override my mirror anymore, running a build now to test with it
Dec 05 10:39:23 <cassidy> shouldn't I use a += instead of ?= in my local.conf instead? (so I won't have to modify this file)
Dec 05 10:41:41 <cassidy> ah I see setscene tasks now !
Dec 05 10:43:48 <bluelightning> cassidy: ?= sets if it isn't already set, so if it's already set that statement won't do anything
Dec 05 10:44:02 <bluelightning> cassidy: so += (or =+) would work better here, yes
Dec 05 10:45:56 <cassidy> I see lots of "ERROR: QA Issue: Package version for package ldd went backwards which would break package feeds from (0:2.19-r0.3 to 0:2.19-r0.1)"  I guess that's because I ran more build on my local machines than on my build slaves, right?
Dec 05 10:56:19 <bluelightning> cassidy: sounds like you are using a different PR server or a different PR server database resulting in the PR increments (the .x in the version) being reset
Dec 05 10:57:20 <bluelightning> you can ignore the errors if you don't care about version upgrades on the target from the previous versions (at this exact moment, anyway)
Dec 05 10:57:31 <cassidy> yeah
Dec 05 11:03:06 <cassidy> bluelightning, belen: cool, looks like my cache is finally used now ! Thanks a lot, you've been very helpful :)
Dec 05 11:07:03 <bluelightning> cassidy: no worries :)
Dec 05 11:18:10 <cassidy> is removing the tmp, cache and sstate-cache dirs a safe and good way to start a clean build from scratch?
Dec 05 11:20:16 <bluelightning> cassidy: yes
Dec 05 11:20:38 <bluelightning> though note that if you have your mirrors set up it will fetch from those, that may or may not be what you want
Dec 05 11:20:54 <cassidy> yeah that's what I want, see how much this mirror speed up things (or not)
Dec 05 12:06:28 <mcfrisk> how to clear sstate cache for a single package? I have a build error with binutils native which seems to depend on compiler version on host but this oddly goes away by second or third compilation attempt..
Dec 05 12:06:52 <bluelightning> mcfrisk: bitbake -c cleansstate <recipename>
Dec 05 12:08:15 <mcfrisk> bluelightning: thanks!
Dec 05 12:08:19 <bluelightning> mcfrisk: what was the error?
Dec 05 12:09:12 <mcfrisk> some undefined macro/variable thingy with gcc 4.9, doesn't happen with gcc 4.6
Dec 05 12:09:41 <mcfrisk> but build isn't breaking every time, which is odd..
Dec 05 12:12:26 <mcfrisk> bluelightning: http://pastebin.com/raw.php?i=3gXyAhE2
Dec 05 12:13:05 <bluelightning> mcfrisk: hmm, ok... I've not seen that one before, I was wondering if it was another issue I had seen very occasionally
Dec 05 12:13:42 <mcfrisk> and after sstateclean got a successful build, sigh...
Dec 05 12:24:42 <jaeckel> is there a way to create a recipe that creates a package from source code that is located in a folder on the machine?
Dec 05 12:25:34 <jaeckel> I thought about using one git repository for package and recipe/layer
Dec 05 12:26:14 <jaeckel> or is this probably a stupid idea? ;)
Dec 05 12:32:41 <frsc> jaeckel: Maybe you can use the first example here as reference: http://www.yoctoproject.org/docs/current/dev-manual/dev-manual.html#new-recipe-testing-examples?
Dec 05 12:36:29 <jaeckel> frsc: yeah I tried that, both with relative and absolute paths to my source code, but I arrived not yet to get it to fetch the files
Dec 05 12:43:24 <soderstrom> jaeckel: What is your SRC_URI, LICENSE and LIC_FILES_CHKSUM variables? Also can you post your <arch>/<recipe name>/<software version>/temp/log.do_fetch log? Or just look in it and see what it is saying.
Dec 05 12:46:03 <jaeckel> SRC_URI = "file://path/to/src/*"
Dec 05 12:46:19 <jaeckel> LIC* are from the example
Dec 05 12:46:48 <jaeckel> and the log.do_fetch does not complain that something went wrong
Dec 05 12:47:09 <jaeckel> oh
Dec 05 12:48:15 <soderstrom> jaeckel: if you are using the checksum from the example and not one generated from a file in the package it is very unlike for these to match ;)
Dec 05 12:49:03 <jaeckel> LIC_FILES_CHKSUM = "file://${COMMON_LICENSE_DIR}/MIT;md5=0835ade698e0bcf8506ecda2f7b4f302"
Dec 05 12:51:59 <jaeckel> I think that's probably just a bad idea to have one folder with the source code and parallel to that one folder with the layer
Dec 05 12:53:11 <jaeckel> either I should put the source code in the layer or clearly separate them
Dec 05 13:00:25 <soderstrom> jaeckel: Actually the location of the source shouldn't matter that much (it can be a local file/dir,tar-file or external ftp,svn,git location). The source will be extracted to the S-variable location in the unpack task.
Dec 05 13:08:38 <jaeckel> okay, got it
Dec 05 13:08:50 <jaeckel> thanks for your help
Dec 05 13:10:15 <soderstrom> jaeckel: No problem, I'm only starting out with Yocto myself, hope to see you around :) Maybe you will help me next time ;)
Dec 05 13:17:26 <manuel__> heya, is there a way to pass PR and PV down to the built program, so that i can set version and revision strings in the program itself?
Dec 05 13:22:23 <manuel__> i guess EXTRA_OEMAKE is what i am looking for, also i use QMAKE
Dec 05 13:28:37 <soderstrom> manuel__: I haven't used qmake that much but I'm unsure if the EXTRA_OEMAKE will be passing anything to the qmake. Specially since there are variables name EXTRA_QMAKEVARS_POST and EXTRA_QMAKEVARS_PRE. So my guess is that you want one of those..
Dec 05 13:29:33 <soderstrom> manuel__: http://www.yoctoproject.org/docs/1.7/ref-manual/ref-manual.html#var-EXTRA_QMAKEVARS_POST
Dec 05 13:34:24 <manuel__> yap, can pass stuff down
Dec 05 13:34:33 <manuel__> thanks you
Dec 05 13:35:05 <soderstrom> manuel__: Your welcome :)
Dec 05 13:37:18 <soderstrom> Sometimes I just love my english.. Ofc it should be "you are" or "you're" welcome ^^. Anyways have a great day!
Dec 05 13:39:28 <manuel__> soderstrom: works perfectly, thank you
Dec 05 13:53:38 <bluelightning> mcfrisk: which host distro are you building on?
Dec 05 13:58:21 <mcfrisk> bluelightning: debian unstable
Dec 05 13:58:35 <bluelightning> hmm, ok
Dec 05 13:59:02 <bluelightning> it's odd that you would get this kind of error intermittently
Dec 05 14:00:25 <mcfrisk> yep, but this is in 1.5.3 yocto, 1.6 seems better so... sigh, I wish we could move to the later releases but...
Dec 05 14:02:35 <bluelightning> ah right... hmm
Dec 05 14:02:45 <bluelightning> well at least you were able to get past this one
Dec 05 14:03:37 <mcfrisk> yes, and I have fingers on cherry-pick button to take the BUILD_CC fixes in and force builds to gcc-4.6, but then the problem just goes away..
Dec 05 20:58:46 <Crofton|work> ANyone know what license this is?
Dec 05 20:59:03 <Crofton|work> https://github.com/richards-tech/RTIMULib/blob/master/LICENSE
Dec 06 00:16:48 <halstead> Okay everyone. The planned outage will start interrupting services now.
Dec 06 01:02:06 <kergoth> Crofton|work: MIT
Dec 06 01:02:12 <kergoth> Crofton|work: http://opensource.org/licenses/MIT
Dec 06 01:02:16 <kergoth> at least, thats what it looks like to me
Dec 06 01:02:17 * kergoth yawns
Dec 06 01:10:26 <halstead> Most services are restored.
Dec 06 01:12:54 <halstead> All services are restored.
**** ENDING LOGGING AT Sat Dec 06 03:00:00 2014