**** BEGIN LOGGING AT Tue Apr 08 02:59:58 2014 Apr 08 06:17:23 <_valle_> Thanks bluelightning! Apr 08 07:29:20 good morning Apr 08 08:37:34 morning all Apr 08 10:00:07 hey guys ... are we getting the heartbleed fix for yocto 1.5 by any chance ? Apr 08 10:02:28 Marex_: it's being worked on as we speak Apr 08 10:02:35 Marex_: thanks for the reminder though :) Apr 08 10:02:55 (for master first, then we can do a backport for 1.5) Apr 08 10:05:58 bluelightning: I expected no less :) Apr 08 10:06:00 bluelightning: thanks Apr 08 10:14:34 heartbleed ? **** BEGIN LOGGING AT Tue Apr 08 10:27:10 2014 Apr 08 10:46:54 <_valle_> bluelightning: Thank you for the alsa tip yesterday Apr 08 10:47:06 _valle_: np Apr 08 10:55:40 <_valle_> Next problem... If I add qtquickcontrols-qmlplugins to IMAGE_INSTALL I get --> ERROR: qtquickcontrols-qmlplugins not found in the base feeds Apr 08 10:56:14 that means that package was never produced Apr 08 10:57:55 _valle_, I do have "qtquickcontrols-qmlplugins" and "qtdeclarative-qmlplugins" in my IMAGE_INSTALL, together with "qtbase-plugins" Apr 08 11:12:52 <_valle_> kroon: I got the "qtbase-plugins" and "qtdeclarative-qmlplugins" and that works. But when I add the "qtquickcontrols-qmlplugins" it fails Apr 08 11:14:37 _valle_, ah yes.. you might need a PACKAGECONFIG_append_pn-qtbase = " accessibility" in your distro conf, otherwise the quickcontrols package wont be built IIRC Apr 08 11:15:21 yes, see comment in qtbase.inc Apr 08 11:16:05 not very straightforward this, unfortunately.. Apr 08 11:30:09 dany1: http://heartbleed.com/ read this :) Apr 08 12:01:02 <_valle_> kroon: Thank you, solved the problem! Apr 08 13:08:20 Hi. I bbappended the following to a recipe : do_configure_virtclass-nativesdk[noexec] = "1" // but bitbake still execute a task "configure" and fails in addition. Do you know why ? Apr 08 13:09:36 bitbake -e tells me that the do_configure[noexec] is properly expanded ;) Apr 08 14:08:25 _valle_, glad I could help Apr 08 14:14:58 halstead: ping Apr 08 14:19:34 scotrif pong Apr 08 14:23:16 do_pingpong[noexec] = "1" Apr 08 14:23:50 halstead: ignore ping - was having key lock problems but I figures them out. Apr 08 14:24:14 No problem. Apr 08 14:31:31 rbuchmann: flags aren't propagated across overrides, at least they haven't in the past. that wouldn't result in do_configure[noexec] = "1". It'd stay on the _virtclass-nativesdk variable Apr 08 14:31:32 * kergoth yawns Apr 08 14:33:17 thx Kergoth Apr 08 15:00:05 YPTM: Scott Rifenbark is on the call Apr 08 15:00:08 YPTM: Vali Cobelea is in hte conf Apr 08 15:00:10 YPTM: Dial-in number: 1.972.995.7777 US Toll Free number: 1.877.561.6828 BlackBerry users, click this link to join your conference as a participant: 1.972.995.7777x42001078# Apr 08 15:00:35 YPTM: Stephen just joined. If you joined earlier please re-announce yourseves. Apr 08 15:00:46 YPTM: Tom Z on the call Apr 08 15:00:53 YPTM: Jefro is here Apr 08 15:01:00 YPTM: Matthew on the call Apr 08 15:01:05 YPTM: Belen is on the call Apr 08 15:01:21 YPTM: AlexG on the call :) Apr 08 15:01:32 YPTM: Mark is here Apr 08 15:01:36 YPTM: Denys is here Apr 08 15:02:14 YPTM: Richard is here Apr 08 15:02:25 YPTM: Darren is on Apr 08 15:02:44 YPTM: Saul is on Apr 08 15:02:47 Alex d here Apr 08 15:03:17 YPTM: Paul Eggleton is on Apr 08 15:03:38 YPTM: Nitin is on Apr 08 15:04:11 meta-cgl Apr 08 15:04:30 CGL = Carrier Grade Linux (I'm assuming) Apr 08 15:04:42 that would make more sense :-) Apr 08 15:04:52 I was just trying to quickly expand it to *something* Apr 08 15:04:54 :-) Apr 08 15:04:57 thanks bluelightning Apr 08 15:04:58 AFAIK there isn't a meta-cgl anywhere... Apr 08 15:05:02 YPTM: current 1.6_M5.rc2 Status: https://wiki.yoctoproject.org/wiki/1.6_QA_Status Apr 08 15:05:05 but ya.. Carrier Grade... Apr 08 15:05:30 YPTM Michael on Apr 08 15:05:30 YPTM: hi all Apr 08 15:06:54 YPTM: Sean Hudson from Mentor joined the YPTM Apr 08 15:09:06 hmm, is it me or did his voice go away? Apr 08 15:09:13 did for a sec.. but came back Apr 08 15:09:29 cut out again.. Apr 08 15:09:34 * darknighte nods Apr 08 15:12:25 http://www.openembedded.org/wiki/OEDAM Apr 08 15:14:07 Thanks, Sean - it is important to note also that OEDAM is NOT training. It is a meeting for OE and YP developers to solve some specific issues and plan for future releases. Just want to make sure everyone understands the purpose. Apr 08 15:15:00 I'm only saying this because I have fielded many questions about training in particular. I guess we need to continue to provide DevDays at future events! But this isn't one. Apr 08 15:15:05 Jefro: I forgot to make that point. Thanks. For the interested, this is intended to be a hackfest Apr 08 15:15:41 I wonder if there is a way to poll for interest in a DevDay later this year... Apr 08 15:15:59 yup.. people really like the DevDay.. but it's expensive.. Apr 08 15:16:34 fray: you're right on that point. Apr 08 15:17:15 We are planning for a DevDay event at ELCE in October Apr 08 15:17:48 * darknighte should probably catch up on his yocto-advocacy ml reading... Apr 08 15:18:18 info on the updater I was referring to: http://comments.gmane.org/gmane.linux.embedded.yocto.general/17082 Apr 08 15:18:51 thanks! Apr 08 15:19:50 some related bugs: Apr 08 15:19:52 https://bugzilla.yoctoproject.org/show_bug.cgi?id=5584 Apr 08 15:19:52 Bug 5584: enhancement, Medium, 1.6, tom.zanussi, NEW , Evaluate adding Stefano's update software Apr 08 15:19:56 https://bugzilla.yoctoproject.org/show_bug.cgi?id=3429 Apr 08 15:19:56 Bug 3429: enhancement, Medium+, Future, saul.wold, ACCEPTED , create infrastructure to support direct image upgarding Apr 08 15:20:01 fray: you mentioned making deployment a priority Apr 08 15:20:18 (the latter mentions nitra / fenrus, but I don't think that avenue has ultimately lead anywhere) Apr 08 15:20:20 for 1.7. RP thoughts? Apr 08 15:20:29 DarkKnight: what about it? Apr 08 15:20:37 darknighte dunno about 'priority', but 'goal'.. Apr 08 15:20:58 IMHO the product has transitioned from "lets build a system" to "lets enable the developer", and one of the key points of that is deployment.. Apr 08 15:21:07 the 'wic' stuff is one of the first pieces of the deployment puzzle Apr 08 15:26:54 Thanks for the BZ references. I'm taking a look, very useful. Apr 08 15:29:30 ya.. white paper vs code may be what is appropriate right now Apr 08 15:31:06 fray+1 Apr 08 15:31:35 fray: I like that idea. Apr 08 15:31:43 like I said 'goal'.. not priority.. priority for me is we have to devote resources (we already have).. goal (we want to do this, but I don't know if we can) Apr 08 15:32:16 fray: agreed Apr 08 15:33:05 YPTM is done. Apr 08 15:40:12 fray: that really does sum it up. I think that was what RP was saying as well. Apr 08 17:31:03 Hmm, wonder if i should patch/sed site.py to disable USER_SITE in python-native, or use a sysroot wrapper Apr 08 18:06:32 guys, is there any patch for meta-qt5 to add qtmultimedia examples? Apr 08 18:14:23 Looking for Stefan Stanacar here... not sure of handle. Apr 08 18:17:03 RP, bluelightning: does https://github.com/MentorEmbedded/meta-mentor/pull/127/files seem reasonable for oe-core, or should I alter Lib/site.py in a patch or do_patch postfunc instead of do_install? I figured since it shouldn't affect the python-native build process, might as well do it as far along in the tasks as possible Apr 08 18:17:08 * kergoth gets food Apr 08 18:17:41 kergoth: I don't know enough about this particular area to comment I'm afraid Apr 08 18:18:31 no worries, maybe richard will have an opinion. just needed a second set of eyes to make sure i'm not entirely insane :) Apr 08 18:20:10 * kergoth preps rfc patch for the list to discuss it there instead Apr 08 18:30:16 now the world has another bigger bug (openssl) the question is: how do I deal with it with yocto? Apr 08 18:30:44 it's fixed in oe-core master Apr 08 18:31:00 not sure about the stable branches, haven't checked Apr 08 18:31:08 but just updated our layers at work and saw the fix come down Apr 08 18:31:22 I just sent out a backport of the fix for dora Apr 08 18:31:25 ah, nice Apr 08 18:31:34 I'm just build-testing for dylan Apr 08 18:31:51 Saul also sent out the workaround which just disables the functionality Apr 08 18:32:45 volker-: so you have three choices: use the patch to upgrade to 1.0.1g, apply the patch to add the fix only (as I've done for dora and soon dylan), or apply the workaround Apr 08 18:34:24 bluelightning: so it always ends in some manual work and delays? Apr 08 18:34:58 volker-: if you have an alternative suggestion I'd love to hear it... Apr 08 18:35:03 heh Apr 08 18:35:54 bluelightning: no, just checking. yocto is a reference distribution, I don't know how stuff is managed there. Other distros (which are indeed NOT reference systems) have security advisories. So just curious how to deal with it with yocto :) Apr 08 18:36:36 to clarify, yocto isn't a reference distribution, poky is :) Apr 08 18:37:18 yeah, sorry, all a big blobb that looks from me from the outside like the same ;-) Apr 08 18:37:51 hehe. fair enough Apr 08 18:39:36 bluelightning: ok, my suggestion, lets call it a dream, would be having security advisories for the core poky reference distro a lot of people seem to build there own code up on :) Apr 08 18:41:40 volker-: if we had the resources for a dedicated security person, that would be the ideal situation yes Apr 08 18:42:25 Does anyone here work on the yocto book or know where I can find the TOC? Apr 08 18:42:56 volker-: the contents? surely it's at the top of the document? Apr 08 18:43:04 volker-: which manual are you referring to? Apr 08 18:43:11 bluelightning: http://www.packtpub.com/embedded-linux-development-with-yocto-project/book Apr 08 18:43:44 oh, interesting... I wasn't aware that was out yet Apr 08 18:43:49 otavio is one of the authors Apr 08 18:44:13 bluelightning: it is not out yet, amazon says april 6th, but not available Apr 08 18:44:28 amazon lists also another one for nov 5th Apr 08 18:46:17 bluelightning: are there any available books out there? Apr 08 18:51:44 I think this would probably be the most up-to-date one that I'm aware of Apr 08 18:56:01 let's see if otavio response :) Apr 08 19:00:06 volker: otavio and daiane are both attending the freescale tech forum (FTF) which just started today in Texas, so they're probably travelling/getting settled/at the conference Apr 08 19:03:08 kk :) Apr 08 19:03:23 volker-: otavio is on irc a lot, i don't know if daiane hangs out on irc, but they're both on the meta-fsl mailing list quite a bit too Apr 08 19:03:38 volker-: thanks for the link to the book, i wasn't aware of its existence :-) Apr 08 19:04:20 tlwoerner: I would be happy if it would be on the market when I started ;-) Apr 08 19:04:32 tlwoerner: so make the yocto page announce it ;-) Apr 08 19:05:23 is there any basic ntp client available with the poky distro? Apr 08 19:09:13 volker-: try the metadata index: http://layers.openembedded.org/layerindex/branch/master/recipes/ Apr 08 19:09:29 volker: busybox can be configured to build ntp, or there's ntpd in meta-oe Apr 08 19:11:53 Is there a document that covers all of the flags passed to the compiler for the target system (i.e. CC, CFLAGS, LDFLAGS etc)? The Reference Manual only mentions CFLAGS. Apr 08 19:19:47 rburton: thanks for the hint. I will modify busybox :) Apr 08 19:20:07 how do you gusy use the openembedded layers when it comes to releases? Apr 08 19:20:15 s/gusy/guys/ Apr 08 19:21:13 volker-: you might want to expand that question Apr 08 19:21:59 rburton: when you release, you want to have a repeatable procedure to re-create something 1:1. So one way is to copy the downloaded files and only use this folder on builds. Apr 08 19:22:35 rburton: that's why you dont use git HEAD to build but a specific hash or even better use a tar.gz that doesn't change. Apr 08 19:22:57 rburton: so I am curious how other people deal with using other layers like the ones found on openembedded.org Apr 08 19:25:05 right, i tend to recommend people use git clones Apr 08 19:25:09 and track the stable branches Apr 08 19:25:15 and do mass-tagging when you make a release Apr 08 19:25:28 use git submodules to track the layers, or repo, or something else Apr 08 19:25:56 but fetch it all from git so you can replicate exactly, but still have an easy way to upgrade to a slightly newer stable release Apr 08 19:27:32 hmmm Apr 08 19:27:59 right now I think I will go with the poky tar for the release Apr 08 19:28:16 but maybe the repo itself would be the better way Apr 08 19:56:07 busybox comes with its own defconfig. From my understanding this defconfig is only used because of do_prepare_config() in busybox.inc Apr 08 19:56:37 what would be an slim way to just modify one line in the default defconfig without copying the entire file. Apr 08 19:56:52 with the kernel you just seem to need to create another .cfg Apr 08 19:57:28 you can use a .cfg in busybox too Apr 08 20:05:32 strange. If I watch the .config file and do a -c clean followed by a -c compile the variable I overwrite gets first added in do_configure, then removed and added with a time delay in do_compile Apr 08 20:58:38 in what manner git repos work with yocto? if i have a git url with a specific branch and do a bitbake, later i commit a newer version to the git. what happens if i perform a new bitbake? the sources are cached, so the newer git version is not included in the new image or does the parsing recipes process check if the local git version is older than the git version of the repo??? Apr 08 20:58:57 it depends on the recipe Apr 08 20:59:19 SRCREV defines the commit to check out. if you set it to ${AUTOREV}, it'll follow the tip of the specified branch, updating every time it parses Apr 08 20:59:44 updating the commit hash every time it parses, that is, not the sources, obviously :) Apr 08 20:59:52 (runs git ls-remote to examine the remote) Apr 08 21:03:35 so, i got AUTOREV Apr 08 21:04:39 that means that the newer bitbake automatically merges all new commits … well rebuilds all the outdated sources? Apr 08 21:07:39 you also need to include SRCPV in PV, thereby ensuring that the recipe version changes increments when teh sources change Apr 08 21:07:43 s/increments// Apr 08 21:16:41 kergoth, thanks, that coincedentally answers my #oe question, I'm not setting SRCPV.. Apr 08 21:21:39 kergoth: The only thing I tend to worry about is sed lines stopping applying in later versions Apr 08 21:22:10 kergoth: Patches are much more deterministic in that regard, I know the python stuff is full of sed though Apr 08 21:30:11 RP: true, that's a good point. I'll prep as a patch instead Apr 08 21:35:03 kergoth: we should probably document the preference of patches over sed somewhere as I'm seeing a lot of it and I don't think it may be a good thing Apr 08 22:02:32 Is there a list of environment variables (e.g. CC, CFLAGS, LDFLAGS) that bitbake provides to a Makefile? Apr 08 22:10:11 daren, erm... OEMAKE OEMAKE_EXTRA I think? See the reference manual Apr 08 22:16:48 EXTRA_OEMAKE is what you use to pass extra arguments on the make command line Apr 08 22:16:52 which may be what you want Apr 08 22:17:33 if you just want to see the full environment exported for a particular task (after running it), look at run.do_xyz in the temp directory under the work directory for the recipe Apr 08 22:18:11 bluelightning: run.do_xyz might be what I want Apr 08 22:18:25 dvhart: Didn't have much luck with the reference manual :) Apr 08 22:19:37 daren, I meant to dig more now that you had some variable names to look for. But, bluelightning's response is likely the better path. Apr 08 22:27:01 bluelightning and dvhart, I think run.do_compile.### is what I wanted..... basically a list of what is exported (CC, PATH, AR, etc.) so I can use them in the Makefile Apr 08 22:27:31 I have a mess of a Makefile, I need to get rid of the hard-coding and let bitbake work it's magic Apr 08 22:35:42 is beth here? Apr 08 22:46:28 maxtothemax, doesn't look like it Apr 08 22:46:39 k Apr 08 23:20:20 the dev- and reference-manual do not talk about oe_runconf? Apr 08 23:21:37 oe_runconf seems to inject options that not every configure likes. How can I remove them without modifying all other builds? Apr 08 23:23:30 see autotools.bbclass Apr 08 23:25:03 kergoth: is there somewhere some hidden documentation about autotools.bbclass? Apr 08 23:28:01 volker-: possibly, no idea, i just read the classes, they're often not that complex, particularly in this case, it's largely just shell functions Apr 08 23:32:49 kergoth: thanks for the hint. I think I found what I need there Apr 08 23:32:59 just need to find out how to remove only one line out of CONFIGUREOPTS Apr 08 23:40:55 volker-: if you're using recent oe-core/poky, use _remove. CONFIGUREOPTS_remove = '--with-foo' Apr 08 23:49:22 kergoth: cool. was looking to do something similar to that, didn't know it works there Apr 08 23:49:57 what confsuses me is that CONFIGUREOPTS = "$CONFIGUREOPTS" doesn't work (that doesn't make sense itself, but is a good way to "test" it Apr 08 23:51:41 ok, it seems like I can't use CONFIGUREOPTS_remove in the .bb file **** ENDING LOGGING AT Wed Apr 09 02:59:58 2014