**** BEGIN LOGGING AT Mon Aug 26 02:59:58 2013 Aug 26 07:40:40 prserv/serv: Fix pid file removal.. heh I was assuming that those hundrets of files were from cases where I had to kill bitbake with -9 (when it was stuck) Aug 26 08:18:48 rburton: hey, have you seen 5046 Aug 26 08:23:40 hi all, how to clean only one recipie? Aug 26 08:24:26 drasko: bitbake -c clean recipe-name Aug 26 08:25:07 JaMa, I want to clean only binutils part in order to reconfigure it and rebuild it Aug 26 08:25:26 how do I find recipe-name for this? Aug 26 08:25:42 drasko: check for the relevant .bb file. Aug 26 08:25:54 Is it under some variable? Aug 26 08:26:04 What's the name of this variable to grep? Aug 26 08:26:11 ./recipes-devtools/binutils/binutils_2.23.2.bb Aug 26 08:26:49 not sure what variable you mean. Aug 26 08:26:53 so, I have to give a link to bb file? Aug 26 08:27:08 bitbake -c clean ./recipes-devtools/binutils/binutils_2.23.2.bb Aug 26 08:27:10 like this? Aug 26 08:27:15 no Aug 26 08:27:23 bitbake -c clean binutils Aug 26 08:27:34 hm... this does not work for me Aug 26 08:27:49 Summary: There was 1 ERROR message shown, returning a non-zero exit code. Aug 26 08:28:21 "does not work" does not work for help. ;-) Aug 26 08:28:28 It looks like it is not called binutils Aug 26 08:29:15 drasko: http://www.macieira.org/blog/2012/05/doesnt-work-doesnt-work/ Aug 26 08:29:26 how to find name of the recipe? Aug 26 08:30:13 it's called binutils Aug 26 08:30:22 what's wrong is written in that ERROR message Aug 26 08:30:43 JaMa is this name kept in some variable within ./recipes-devtools/binutils/binutils_2.23.2.bb? Aug 26 08:30:49 Because I can not see it Aug 26 08:31:53 drasko: why do you keep talking about variable? Aug 26 08:32:00 why don't you just check the link above? Aug 26 08:32:26 you cannot say to the doctor "you feel ill", but not provide more information. :) Aug 26 08:32:46 lpapp, how does yocto detemine that `binutils` is related to binutils_2.22.bb etc... Aug 26 08:33:07 drasko: come on ... you got an error, and you have not even pasted it... Aug 26 08:33:15 ERROR: Execution of event handler 'external_linaro_toolchain_version_handler' failed Aug 26 08:33:43 paste is not equal to grabbing out a line you may think is the most important. Aug 26 08:33:47 paste the whole. Aug 26 08:33:53 http://pastebin.com/eQ4Ecsiw Aug 26 08:33:59 trust the doctor to judge the important bits, do not try to play his role. :) Aug 26 08:34:09 OK doc Aug 26 08:34:12 ;) Aug 26 08:34:26 I am likely not the person to help, but at least the experienced people do not need to ask the same. Aug 26 08:36:37 any diagnose? Aug 26 08:36:42 CmdError to bb.process.CmdError Aug 26 08:36:54 you are using old meta data, apparently. Aug 26 08:37:06 what do you mean? Aug 26 08:37:18 see this is the problem. Aug 26 08:37:31 you think you pasted the error until you pasted the whole, and your line was irrelevant. :) Aug 26 08:37:35 drasko: variable PN is automatically assigned from filename Aug 26 08:37:51 here is what I did : Aug 26 08:37:54 bitbake virtual/kernel Aug 26 08:38:00 binutils breaks Aug 26 08:38:17 i changed the way it is configured trying to add --disable-doc Aug 26 08:38:30 you're probably talking about binutils-cross not target binutils Aug 26 08:38:32 drasko: http://lists.linaro.org/pipermail/linaro-dev/2013-June/016395.html Aug 26 08:38:55 drasko: do you have an up to date version? Aug 26 08:39:01 Now when I restart bitbake virtual/kernel it executes do_install of binutils, instead to go to beginning (configure, compile, etc...) Aug 26 08:39:05 and because you're using external-linaro-toolchain it's not provided by binutils-cross recipe but by external-linaro-toolchain recipe (prebuilt toolchain installed somewhere on your host) Aug 26 08:39:14 so I want to force it to do this by cleaning binutils somehow Aug 26 08:39:35 drasko: you are using an external toolchain, prebuiltr. Aug 26 08:39:38 prebuilt.* Aug 26 08:40:53 https://git.linaro.org/gitweb?p=openembedded/meta-linaro.git;a=tree;f=meta-linaro-toolchain/recipes-devtools/binutils;h=111819820577cb9d274b8ddedd21ec17c6aad162;hb=master presumably this would be cleaned. Aug 26 08:41:06 but then again, either use the latest version, or apply the fix manually. Aug 26 08:44:28 drasko: also, you might wanna update the external toolchain layer more regularly. Aug 26 08:46:48 true missing my fix is bad :) Aug 26 08:47:36 JaMa: I found a flaw bot unly for meta-micro atm ;) Aug 26 08:48:03 JaMa: you could hopefully restart the rebuild of klibc now Aug 26 08:48:45 JaMa: you are using the external linaro toolchain, too? Aug 26 08:54:31 Anyone here familiar with the busybox recipe? Aug 26 08:55:14 I was just wondering if there's a recommended way to append config options to it, and the recipe looks fairly complex. Aug 26 08:55:41 I was thinking I'd just sed the config file, but if there's a better or more standard way I'd love to do that instead. Aug 26 08:56:35 I somewhat am. Aug 26 08:56:47 Stygia: you can use config fragments. Aug 26 08:57:01 Stygia: see the relevant mailing list discussion I initiated. Aug 26 08:57:03 sec. Aug 26 08:59:06 Stygia: https://git.linaro.org/gitweb?p=openembedded/meta-linaro.git;a=blob;f=meta-linaro/recipes-core/busybox/busybox_1.21.1.bbappend;h=0fd4571e7b09161c3fd8a2ccd42cd9cfea9babb6;hb=HEAD Aug 26 08:59:12 that is an example how to use it. Aug 26 08:59:52 Stygia: I asked for documentation on the issue tracker, but the people closed it. Aug 26 08:59:56 they did not wanna document it. Aug 26 08:59:59 it would be nice ... Aug 26 09:01:04 Stygia: see the last post, https://bugzilla.yoctoproject.org/show_bug.cgi?id=3379 Aug 26 09:01:05 Bug 3379: enhancement, Low, 1.4 M5, Qi.Chen, VERIFIED FIXED, Add config fragment support to busybox Aug 26 09:05:25 lpapp: FWIW it was closed after we did add an example to meta-skeleton ? Aug 26 09:05:50 lpapp: http://git.yoctoproject.org/cgit.cgi/poky/commit/?id=ff10f72abda52315b892c64f711096772d9a4c1c Aug 26 09:06:24 RP, no that is a different thing. Aug 26 09:06:35 4964 was about _documentation_. Aug 26 09:06:49 note, how example in the source code and documentation are not equivalent. Aug 26 09:07:03 also, that example was added waaaay after my report. Aug 26 09:07:08 and my report was closed before that. Aug 26 09:07:10 lpapp: well, its better than having nothing at all Aug 26 09:07:22 lpapp: well, my point is *something* was done about it Aug 26 09:07:26 RP: better than having nothing at all does not make my request solved. Aug 26 09:07:53 and again, my report was closed way before that example added. Aug 26 09:08:02 (i.e. there was not even an example for the time) Aug 26 09:08:14 now, tell me whether it was closed rightfully. Aug 26 09:08:24 or mistakenly. Aug 26 09:08:58 (and no, I am not blaming anyone, just simply pointing out objectively that a report has been closed before solving the issue, actually) Aug 26 09:10:15 lpapp: I do not know the details of that so I will not comment on it Aug 26 09:11:26 lpapp: All I do know is that an example was added so some kind of response to your request did happen Aug 26 09:12:14 lpapp: yes in one project Aug 26 09:12:20 RP: yes without giving a thanks. :) Aug 26 09:12:34 or marking my report as "solved" rather than "invalid" which is well ... incorrect. Aug 26 09:12:55 RP: https://bugzilla.yoctoproject.org/show_bug.cgi?id=4964 Aug 26 09:12:56 ant_work: it's running with your patch from yesterday Aug 26 09:12:56 Bug 4964: enhancement, Low, Future, saul.wold, RESOLVED INVALID, busybox: missing documentation for config fragments Aug 26 09:13:29 RP: that is a typical example of why I do not like devs closing bugs without allowing the reporters to even speak up or reply. Aug 26 09:13:39 RP: because devs or triages will not simply get back to it. Aug 26 09:14:08 JaMa: ok, is only broken for micro distro atm (/usr -> /) Aug 26 09:14:09 ant_work: and if it's building I can live with that until you send v2 :) Aug 26 09:14:14 ;) Aug 26 09:15:13 I somehow thought kernel is always in /usr but those hardcodes are only in the klibc build tree Aug 26 09:15:20 lpapp: that bug changed meaning several times over and several of the topics in there are invalid. If you want consistent clean answers to bug reports, keep them to a simple topic, rather than keeping changing their meaning Aug 26 09:15:57 RP: there was no meaning changed. Aug 26 09:16:12 RP: the bugreport is about allowing people to customize the config of things, like busybox. Aug 26 09:16:20 RP: people suggested to open a feature request, so did I. Aug 26 09:16:31 lpapp: You changed the subject a short while ago, it was "Simplify the mapping between PACKAGECONFIG and .bb/.inc file" which was correctly marke INVALID Aug 26 09:16:44 once someone presented there is a feature (why are features added without documentation btw?!), I suggested to turn that into documentation. Aug 26 09:16:55 RP: exactly. Aug 26 09:17:15 because just like with any other similar reports, a missing feature is turned into a documentation issue. Aug 26 09:17:48 also, why is this not a project policy to add features only with documentation? Aug 26 09:18:07 it is a good and working practice in several big projects. Any reason why not to do so in here? Aug 26 09:18:45 lpapp: funnily enough it is a policy we're adopting in future, it just taked time to do these things Aug 26 09:19:03 lpapp: Instead of working on this or the release I'm here trying to explain to you how to use a bugzilla though :( Aug 26 09:19:24 you do not need to explain how to use bugzilla. Aug 26 09:19:34 you do not need to explain how to use bugzilla.0 Aug 26 09:19:36 especially not according to the practice which Jeff also promoted. ;- Aug 26 09:19:37 ) Aug 26 09:19:47 against the practice* Aug 26 09:20:16 this seems to be a standard practice due to the lack of quality for that matter, that missing documentations are added once the feature is revealed to be present. Aug 26 09:20:40 RP: also, if you do not have time to hang on IRC for such topics to be discussion, do not do it. You are not obliged. I will not mind. Aug 26 09:20:48 to be discussed* Aug 26 09:22:53 IMO, the long discussion in the thread just shows that how much we need documentation because it is hard to figure out what is going on without it. Aug 26 09:24:59 lpapp: One of the key things to do in making a good bug report is to have one clear issue and to clearly explain that issue. 4964 mostly talked about PACKAGECONFIG which in fact had nothing to do with "buzybox config fragements not documented" Aug 26 09:25:14 lpapp: bugs should not "mutate" into something else in their lifecycle Aug 26 09:25:48 lpapp: If it turns out that you later understand you need something different I recomment a new clear single topic bug report in future. I will talk to Jefro about this and we will agree to do this in future Aug 26 09:27:21 RP: what about smthg like this to simplify our live to build rare images? http://tinyurl.com/lsjxetp Aug 26 09:27:58 RP: low hanging fruit, just unpackaged Aug 26 09:28:23 ant_work: don't we use update-alternatives for that? Aug 26 09:29:27 RP: no alternatives/conflicts once cramfs removed from meta-openemdebbed/meta-filesystems Aug 26 09:30:41 RP: probably there are more unpackaged utils... Aug 26 09:32:33 ant_work: it looks ok to me then Aug 26 09:32:51 ant_work: surely we'd get an unpackaged file warning? Aug 26 09:33:22 I don't remember any Aug 26 09:38:08 RP: see, mkfs.bfs is SCO so we can probably ignore it Aug 26 09:46:48 RP: that was the idea for the feature, but then the feature was present already. I do not see how it is irrelevant. Aug 26 09:47:31 RP: you would only need to talk to Jeff, but a lot of people who act differently than you. Aug 26 09:47:42 perhaps you should accept their way rather than changing everyone else around. Aug 26 09:48:16 and "add a feature or document it" is one topic: I wanna have a feature with documentation. Actually there is even a documentation flag nowadays to the features proposed. Aug 26 09:48:30 and as you said, that should be even more atomic in the future, so I do not see how you would be correct here. Aug 26 09:48:53 lpapp: What I'm saying is don't change bugs from one thing into something completely different Aug 26 09:49:19 RP: I agree about, but that is a different discussion. Aug 26 09:49:37 when a user wanna have a feature he wanna have that present _and_ documented. Aug 26 09:49:50 lpapp: in an ideal world, yes and I agree Aug 26 09:49:55 if the feature is present, and it is only a documentation part missing, then solving the issue is simpler. Aug 26 09:49:59 I do not really see the problem. Aug 26 09:50:05 lpapp: I also agree we do have some missing documentation Aug 26 09:50:30 lpapp: however I do think having a config fragment example for busybox does go a long way to helping people understand how to use it Aug 26 09:50:41 I agree in a perfect world we should have documentation too Aug 26 09:50:49 RP: actually, not, the bare minimum is close that report as done as suggested. Aug 26 09:51:04 that example could have been added at the point of implementing the feature. Aug 26 09:51:16 whether this QA gate is missing in Yocto, that is another discussion. Aug 26 09:51:41 so, all I am saying, PACKAGECONFIG is just a low-level detail there; do not be concerned about that much. Aug 26 09:52:02 the main point of the documentation "add a documentation configuration feature more than the existing PACKAGECONFIG" Aug 26 09:52:08 of the report* Aug 26 09:52:09 lpapp: We have a lot of existing functionality which is not documented. We cannot magically have documentation for it all. As you see, we are adding QA processes to check documentation Aug 26 09:52:24 and then it was explain why that would be necessary. Aug 26 09:52:28 explained* Aug 26 09:53:12 RP: I do not see how that affect the bugreports. The bugreport is about adding a documented feature. If it is only the documentation missing, it is easier to solve the bug report, so the dev can even be happier, and the user earlier. Aug 26 09:53:17 affects* Aug 26 09:54:13 lpapp: You need to work on the clarity of your bug reports since I can't understand exactly what the problem is, or what kind of a solution would make you happy Aug 26 09:54:15 as I said, when you request a new feature, the documentation flag was even integrated. Aug 26 09:54:27 so let us not break us the good way of moving forward. Aug 26 09:54:33 up* Aug 26 09:55:57 RP, Do you know if we actually have a config fragment example for busybox? I'm tampering with busybox right now, and saw the bug that was resolved an hour or so ago. Aug 26 09:57:46 JaMa: log.world.20130825_083725.log/bitbake.log looks *good* Aug 26 09:58:18 Stygia: he already pasted the link, so did I one in linaro. Aug 26 09:58:26 Stygia: http://git.yoctoproject.org/cgit.cgi/poky/commit/?id=ff10f72abda52315b892c64f711096772d9a4c1c Aug 26 09:58:28 Stygia: and the bugreport also gave one at the bottom what I mentioned. Aug 26 09:59:02 lpapp, Yea sorry, I don't monitor this channel 24/7. Aug 26 09:59:06 RP, Thanks. :) Aug 26 09:59:13 Stygia: it was 10-20 minutes ago or so. :) Aug 26 09:59:19 with highlighting you specifically. Aug 26 09:59:51 lpapp, I don't think so, the xchat icon flashes if someone says my name. Aug 26 09:59:53 But sure. Aug 26 10:00:04 Either way,the link's what I needed, so thanks RP Aug 26 10:00:19 Hi all, I am trying to update my met-linaro, but I am confued. Here are the differences between my meta-linaro.OLD, which is current meta-linaro, and the one I just clend from git repo : http://pastebin.com/6HC9m2rX Aug 26 10:00:46 Can anyone explain how this update can be done, because I guess that I have to keep the same directory structure Aug 26 10:01:16 Stygia: are you saying you did not receive my posts when I tried to help you, and used your nick? Aug 26 10:02:03 drasko: git pull Aug 26 10:02:08 lpapp, , Eh. I come here to check whenever the icon flashes. Any line where someone says "stygia" is red. I haven't seen any red lines until just now. Aug 26 10:02:11 Nor flashing icons. Aug 26 10:02:29 drasko: or you need to archive the latest first, and then uncompress. Aug 26 10:02:43 Stygia: do you have this in the history: Aug 26 10:03:08 lpapp, what do you mean? I did git clone and the directory structures are different... Aug 26 10:03:12 lpapp, Does it matter? Aug 26 10:03:20 http://paste.kde.org/~lpapp/p68989b0f/ Aug 26 10:03:43 Stygia: whether it matters you are not getting the help you request? Aug 26 10:04:08 drasko: why would that be a problem? Aug 26 10:04:27 lpapp, As I said to RP, the example he send me seemed to be quite sufficient. Adding CONFIG_SHA512SUM=y to something.cnf and adding that file to SRC_URI is simple and logical. Aug 26 10:04:37 lpapp, what is the update procedure? Aug 26 10:04:40 lpapp, Just leave it. I'm fine, I'll say otherwise. Aug 26 10:04:51 copy folders from new dir to old one? Aug 26 10:04:58 I do not get it quite Aug 26 10:05:13 Stygia: ok, then I will not try to help in the future. Aug 26 10:05:20 lpapp, Spare me the sarcasm. Aug 26 10:05:22 I thought that I can just replace cold meta-linaro with new cloned dir Aug 26 10:05:29 if you are not getting the posts for some reason. Aug 26 10:05:34 Stygia: there is no sarcasm. Aug 26 10:06:15 drasko: for meta-sourcery, I just use git pull personally. Aug 26 10:06:21 drasko: I do the same with meta-networking. Aug 26 10:06:38 drasko: but if you are inside a git repository, archiving might be cleaner. Aug 26 10:08:34 What I am even more confused is that binutils used during the bitbake virtual/kernel are located in meta/recipes-devtools/binutils Aug 26 10:08:48 RP, I'm not super adapt with GIT, did that commit go into master? And was support for it _just_ added, or was it just the example that was added there? Having that ability to just use a something.cnf file like that would save me ages. Aug 26 10:08:49 and they should be in meta-linaro... Aug 26 10:08:52 I do not get this Aug 26 10:09:02 RP, I don't immediately see it in git log for my meta-openembedded checkout Aug 26 10:09:16 it is not in meta-openembedded ... Aug 26 10:09:21 have you checked where it was committd? Aug 26 10:09:23 committed* Aug 26 10:09:45 it is meta-*skeleton* Aug 26 10:09:46 meta-skeleton... Aug 26 10:09:52 Ah, I thought that was a subdir of meta-openembedded Aug 26 10:09:57 Right, I'll checkout skeleton, thanks. Aug 26 10:10:20 np Aug 26 10:10:22 But... meta skeleton, is that examples? Is support for this already in there? Aug 26 10:10:39 Stygia: read my paste. :) Aug 26 10:10:45 it was mentioned it is an example. Aug 26 10:11:12 lpapp, The paste.kde one? Aug 26 10:11:16 obviously, the feature is implemented if there is an example using it. Aug 26 10:11:17 yes Aug 26 10:11:37 lpapp, Right then. Aug 26 10:12:03 kinda Aug 26 10:12:25 a documentation would still be nice... see the classes, there are examples, but I would not know how to write a class file, what the guidelines are, etc. Aug 26 10:12:37 however, this case is pretty simple, so yeah. Aug 26 10:16:35 JaMa: do you see any remaining issue with the stunnel change? Aug 26 10:17:36 Any idea why binutils is located meta/recipes-devtools/binutils instead meta-linaro? Aug 26 10:18:03 drasko: because you need binutils without external toolchain when you are fine with building on the fly. Aug 26 10:21:17 lpapp, can I replace this meta/recipes-devtools/binutils directory with this one : https://git.linaro.org/gitweb?p=openembedded/meta-linaro.git;a=tree;f=meta-linaro-toolchain/recipes-devtools/binutils;h=111819820577cb9d274b8ddedd21ec17c6aad162;hb=82fa48fb82e94a908f8be74ebddcd85a6ca306e9 in order to update binutils? Aug 26 10:21:35 drasko: you should not, no. Aug 26 10:21:43 Stygia: its part of oe-core so you should already have it if your oe-core or poky is up to date Aug 26 10:22:00 lpapp, but what can I do then - this one is breaking on build Aug 26 10:22:28 drasko: you have not pasted the error. Aug 26 10:23:00 lpapp, there are errors realted to texinfo, I will paste them later Aug 26 10:23:11 I am just interested how can we update these binutils Aug 26 10:23:31 drasko: do not talk about anything else until you fix the errors you have. Aug 26 10:23:33 and why these binutils are used, and not one from external toolchain? Aug 26 10:23:39 RP, Right I found it. :) Aug 26 10:28:02 Hmm. Outside of any function, in a bbappend without anything but an SRC_URI, why does $THISDIR not expand to the location of the bbappend? It seems to expand to the location of the actual recipe. Aug 26 10:28:24 Stygia: what is the problem at hand? Aug 26 10:28:45 drasko: because *you* request so when using an external toolchain. Aug 26 10:28:48 lpapp, I have a bbappend file that adds a file.cnf. The file.cnf is in the same directory as the bbappend Aug 26 10:29:04 Stygia: good, but what is the problem? Aug 26 10:29:11 But it expands to the location of the actual recipe, not the directory the bbappend file is in, and fails with the file not being found. Aug 26 10:29:49 Even if I do a SRC_URI += "file://${THISDIR}/file.cnf" <- this expands to the path of the actual recipe, not the bbappend Aug 26 10:29:59 file://file.cnf doesn't help, either. Aug 26 10:30:32 Stygia: add meta-skeleton to the bblayers.conf and try building busybox from there. Aug 26 10:30:39 bitbake busybox -> does that work? Aug 26 10:31:09 lpapp, ... why would I want to add meta-skeleton? Aug 26 10:31:20 Stygia: it is called debugging ... Aug 26 10:31:33 .. I need one variable. The variable that means "The directory this file is located in". Aug 26 10:31:39 I thought that was THISDIR. It's not. Aug 26 10:31:45 come on... you are doing something that is supposed to work Aug 26 10:31:50 and it does not work. Aug 26 10:31:57 Anyone else have any ideas? Aug 26 10:31:58 how about verifying whether what is supposed does work? Aug 26 10:32:18 it is a usual debugging ceremony AFAICT. Aug 26 10:32:37 and it takes a few minutes maximum. Aug 26 10:32:52 adding one line, and then running a short command, maybe two to clean up before. Aug 26 10:34:43 Stygia: you need to use := for immediate variable expansion Aug 26 10:34:57 Stygia: I suggest you look at one of the bbappends, e.g., in meta-yocto Aug 26 10:35:30 erbo: yeah, that is also in the example and works for me fine. Aug 26 10:35:34 I just checked. Aug 26 10:36:15 erbo, Fantastic, thanks. :) I saw := did something, but for some reason, when I used that, then the main recipe couldn't find it's LICENSE file... I'm investigating. Aug 26 10:36:20 Stygia: but you might also want to follow the "standard" and use FILESEXTRAPATHS_prepend Aug 26 10:37:04 Stygia: See? If you had run the example in a few minutes you would have figured out where the problem was. Aug 26 10:37:44 RP: do you have any idea for the custom bblayers.conf sample issue? Aug 26 10:38:30 Stygia: just paste the example as is, and then customize the bits you need. It is simpler than starting from scratch. Aug 26 10:39:38 lpapp: no, but meta-networking changes are applied by Joe Aug 26 10:39:43 RP: https://bugzilla.yoctoproject.org/show_bug.cgi?id=5046 Aug 26 10:39:44 Bug 5046: normal, Undecided, ---, scott.m.rifenbark, NEEDINFO , Document that CPPFLAGS override behavior and possible workaround Aug 26 10:40:03 JaMa: yes, sure. Thanks for the previous review anyway. Aug 26 10:40:05 Stygia: Stygia sound like lpapp or someone else gave you an example already, otherwise looks at http://git.yoctoproject.org/cgit/cgit.cgi/meta-ivi/tree/recipes-core-ivi/eglibc/eglibc_2.18.bbappend Aug 26 10:40:22 Stygia: that's the recommended way of doing it Aug 26 10:40:31 erbo, Yup, I had an example long ago, I just wondered why ${THISDIR} didn't expand as I expected. :) Aug 26 10:40:34 erbo: if you copy and paste the example it works. :) Aug 26 10:41:21 erbo: do you have any iea for the CPPFLAGS issue? Aug 26 10:42:36 erbo, The issue is that when I mimic the examples, such as the one you linked, the file can't be found. Aug 26 10:42:37 lpapp: unfortunately no, if I think of anything I'll let you know though Aug 26 10:42:53 Stygia: where have you put the file? Aug 26 10:43:13 in meta-company/recipes-company/busybox/busybox_1.20.2.bbappend Aug 26 10:43:27 In that same directory I have my file.cnf Aug 26 10:43:46 it should be in a subdir called busybox Aug 26 10:43:48 Presumably I have to move it to something that matches ${THISDIR}/${PN}... so that being a subfolder named after the package, i.e. Aug 26 10:43:55 erbo, ... yea exactly. :P Aug 26 10:44:17 erbo, I could just remove ${PN} from the append I suppose, but still. The funny thing is that it worked, but then the main recipe failed complaining about the LICENSE file. But ah,well. Aug 26 10:44:56 or you could create that directory to stay as close to the "standard" as possily Aug 26 10:45:36 erbo, Yup, I'm gonna. Aug 26 10:47:33 unused-debug-types -fPIC -c /media/sdb5/eel/build/tmp/work/x86_64-poky-linux/libpam/1.1.6-r3/Linux-PAM-1.1.6/libpam/pam_dynamic.c -fPIC -DPIC -o .libs/pam_dynamic.o Aug 26 10:47:33 | /media/sdb5/eel/build/tmp/work/x86_64-poky-linux/libpam/1.1.6-r3/Linux-PAM-1.1.6/libpam/pam_prelude.c:13:32: fatal error: libprelude/prelude.h: No such file or directory Aug 26 10:47:33 | #include Aug 26 10:47:33 | ^ Aug 26 10:47:33 | compilation terminated. Aug 26 10:47:40 anyone else seeing this ? Aug 26 10:48:42 Stygia: if you copy the busybox example it does not work? Aug 26 10:48:59 PN is different to P so the two examples are not interchangable. Aug 26 10:49:09 dany: paste.kde.org Aug 26 10:51:32 lpapp: sry about that. Aug 26 10:52:17 dany: also, you need to paste more, including the command you executed. Perhaps, you could share a few words about your setup, too. Aug 26 10:55:56 wheb issuing bitbake -k core-image-lsb with poky master, libpam seems to have autodetected libprelude.h, Aug 26 10:56:00 Stygia: THISDIR needs to use immediate expansion to work (:=). Usually you set the FILESEXTRAPATHS_prepend := "${THISDIR}" and then SRC_URI will find the file without a path Aug 26 10:56:18 Equally you could _prepend or _append to SRCURI and use a := I guess Aug 26 11:00:15 which is installed on the host machine, but does not exist in the sysroot. Aug 26 11:02:23 dany: then you need to package it maybe? Aug 26 11:02:54 or fix the detection... so it is not detected for core-image-minimal for instance? Aug 26 11:05:48 lpapp: pam is not built/installed for core-image-minimal. Aug 26 11:06:39 dany: yeah, but I mean any other? Aug 26 11:07:08 dany: alternatively, what happens when you execute "bitbake pam"? Aug 26 11:34:30 RP, Yup, that's exactly what I ended up doing. Aug 26 11:35:02 RP, Though it does seem like it didn't quite "catch"... And that busybox hasn't included my config setting. But I'll investigate. Aug 26 11:47:47 RP: even with "process: Improve exit handling and hangs" I have bitbake hanging after Ctrl+C Aug 26 11:48:29 JaMa: I'm not claiming it fixes them all, it did seem better in my testing though Aug 26 11:49:04 RP: http://pastebin.com/MDwvt8Sc Aug 26 11:49:06 JaMa: I suspect the best way to fix the hanging issue is to fix all the Ctrl+C ones though :/ Aug 26 11:49:23 yes I'm writting only because you said that you're interested in feedback on that :) Aug 26 11:49:39 JaMa: totally, its useful data thanks :) Aug 26 11:49:47 not meaning to be negative :) Aug 26 11:49:55 in this case it was also from a bit weird state Aug 26 11:50:16 So that was Ctrl+C during parsing? Aug 26 11:50:18 when I accidentaly started bitbake with dylan branch checkouted in oe-core repo and while it was running I've checked out master Aug 26 11:50:37 and pressed Ctrl+C while it was still parsing, yes Aug 26 12:16:20 JaMa: I tried using test-dependencies.sh but I guess it doesn't do what I intended to do Aug 26 12:17:02 JaMa: i just want to have cross-toolchain, not requiring me to re-compile, and try the build failure by minimal dependencies Aug 26 12:17:27 after the package is done, I want to have something like "go to that point and start again", in that case have cross toolchain ready Aug 26 12:17:45 would you mind giving some introduction to the script? Aug 26 12:20:31 lpapp, how can I upgrade just meta/recipes-devtools/binutils? Can this be done, or whole meta directory has to be replaced? Aug 26 12:24:13 drasko: you are using an external toolchain. For you that recipe is meaningless as previously written. Aug 26 12:24:33 lpapp, however - this recepie is executed and breaks the build Aug 26 12:24:51 there is a bug in Binutils2.22 Aug 26 12:24:53 https://sourceware.org/bugzilla/show_bug.cgi?id=15183 Aug 26 12:24:54 Bug 15183: was not found. Aug 26 12:25:05 drasko: the problem is that with this whole conversation, you claim things which are out of your league Aug 26 12:25:18 and you do not wanna paste an error rather than telling the issue to the doctors. :) Aug 26 12:25:35 telling your analyzes* Aug 26 12:25:43 so, how about pasting the error? ;) Aug 26 12:26:06 lpapp, OK, OK... But I just wanted to learn how to upgrade one packet in meta... Aug 26 12:26:10 is there a way? Aug 26 12:26:27 why would you need to learn that? Aug 26 12:26:43 well - I would like to upgrade my Binutils to 2.23 Aug 26 12:27:00 but again, you are using an external toolchain because you selected it. Aug 26 12:27:00 again, this bug : https://sourceware.org/bugzilla/show_bug.cgi?id=15183 Aug 26 12:27:01 Bug 15183: was not found. Aug 26 12:27:15 do not jump into conclusions, just tell the issues. Aug 26 12:27:20 Bug 15183 - texinfo-5.0: bfd.texinfo: unknown command `colophon' Aug 26 12:27:21 Bug https://bugzilla.yoctoproject.org/show_bug.cgi?id=15183 was not found. Aug 26 12:27:24 so what is the problem? Again, can you paste the error? Aug 26 12:27:56 lpapp, I will have an error log in a little while - it is building Aug 26 12:28:07 eren: I have about 60 patches in my queue from yesterday evening, so please wait for new status email Aug 26 12:28:11 problem is that I have binutils 2.22 in my meta Aug 26 12:28:23 and I have texinfo 5.1 installed on my machine Aug 26 12:28:38 due to the bug Bug 15183 - texinfo-5.0: bfd.texinfo: unknown command `colophon' this build failes Aug 26 12:28:40 there is no any problem with it Aug 26 12:28:45 JaMa: okkie. when I fix the problem that I mentioned, I can start working on the packages Aug 26 12:28:47 I am having the same scenario Aug 26 12:28:50 drasko: http://www.macieira.org/blog/2012/05/doesnt-work-doesnt-work/ Aug 26 12:28:58 eren: in order to let every recipe to detect as many dependencies as possible the script is building "world" first Aug 26 12:29:01 JaMa: currently, O remove tmp-eglibc and sstate to build the package from scratch Aug 26 12:29:16 eren: keep sstate-cache directory Aug 26 12:29:34 eren: and just do bitbake -c cleansstate before rebuilding it Aug 26 12:29:36 JaMa: I don't want to detect dependencies. These are already been detected by your build. I just want to build only that package, without any additional packages that were built so that I can get the same error as you Aug 26 12:30:07 how do we deal with binutils that break during the build? We add patches, or we can replace binutils with newer version? Aug 26 12:30:22 I prefer to replace binutils from my meta directory to a newer version Aug 26 12:30:24 JaMa: am I clear on the problem? Aug 26 12:30:32 drasko: paste the error, please. Aug 26 12:30:42 if you cannot paste the error, we can only guess, and that is time consuming. Aug 26 12:30:48 lpapp, unrelated to this error Aug 26 12:30:50 eren: for detecting missing mandatory dependencies yes Aug 26 12:30:55 last time you pasted an irrelevant line and we spent X time with it. Aug 26 12:30:55 I am asking - in general Aug 26 12:31:02 how do we deal with these situations? Aug 26 12:31:10 when you actually pasted the real error, I got the patch for you in a few seconds. Aug 26 12:31:11 eren: as I said, just remove tmp-eglibc, keep sstate-cache and do bitbake -c cleansstate recipe-name; bitbake -c build recipe-name Aug 26 12:32:01 drasko: not to mention, what you are asking is in the manual. Aug 26 12:32:02 JaMa: okkie, no need to run test-dependencies.sh then Aug 26 12:32:07 (as it is a basic stuff) Aug 26 12:33:27 drasko: look for .bbappend, but pretty please: do not start solving this issue yourself because I am afraid you will need to drop your work later. Aug 26 12:33:34 eren: yes I pointed you to test-dependencies.sh to show you how it's being tested (and it's the same as what I said now - for minimal deps) Aug 26 12:33:52 let us discuss the error output first, and how it is different to my setup because I have the same and it works. Aug 26 12:33:56 JaMa: okkie, thanks. I am waiting for new email Aug 26 12:34:33 also, 15183 does not exit, so I am not sure what you are referring to at this point. Aug 26 12:34:52 the bot seems to have claimed the same twice. Aug 26 12:35:35 eren: you can check the jenkins link from yesterday, but it will take 10+ hours to finish Aug 26 12:35:42 lpapp, here is the error I am facing Aug 26 12:35:44 http://pastebin.com/GGxyYuTQ Aug 26 12:35:48 drasko: sorry, I meant to say another recipe, not .bbappend. Aug 26 12:35:52 finally! Aug 26 12:36:23 it took some time for compilation to reach this moment of breaking ;) Aug 26 12:36:26 Bug 15183 - texinfo-5.0: bfd.texinfo: unknown command `colophon' Aug 26 12:36:27 Bug https://bugzilla.yoctoproject.org/show_bug.cgi?id=15183 was not found. Aug 26 12:36:32 drasko: you still have not fixed the error we pasted you. :O Aug 26 12:36:56 wait a second... Aug 26 12:37:01 It was an old log Aug 26 12:37:05 I have new one Aug 26 12:38:25 here : http://pastebin.com/dqX5xYaj Aug 26 12:38:59 drasko: why don't you use a supported distribution? Aug 26 12:39:24 e.g. wheezy. Aug 26 12:39:55 drasko: anyway, use poky master. Aug 26 12:40:38 RP: another data point can be cacheprogres.update sometimes showing backtrace on ^C http://pastebin.com/VydR2Ejv Aug 26 12:40:38 that is the only thing which would make me think that you do not have the texinfo fix. Aug 26 12:42:31 It's not my choice - I have got this one to work on Aug 26 12:42:32 colophon and cygnus are seen as unknown commands - this is a binutils 2.22 texinfo related bug Aug 26 12:42:33 I need either a patch that correct tis Aug 26 12:42:33 or replacement to a newer version of binutils Aug 26 12:42:34 But I do not know how to do this :( Aug 26 12:42:58 have you tried master? Aug 26 12:43:06 JaMa: I'll merge http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/wip76&id=02bb58747f8dd8e89c306fb8424ac88d34696e2a which should help Aug 26 12:43:15 what environment is that where you do not choose your Linux? :( Aug 26 12:43:29 anyway, one texinfo has been fixed in master, so do not use dylan, but master. Aug 26 12:43:35 or backport the fix with a .bbappend file. Aug 26 12:43:52 drasko: also, be prepared there might be a lot of issues for that testing debian Aug 26 12:44:05 at least archlinux got several patches in master... as for debian testing, probably not many. Aug 26 12:46:00 RP: I'm going to cherry-pick it right now, thanks :) Aug 26 12:46:55 JaMa: hmm, it seems there is a qtserialport recipe in meta-oe. Aug 26 12:47:18 RP: I got catched in a nice infinite loop yesterday on do_package of inexistent relative symlinks. Ctrl+C did help in that case Aug 26 12:47:56 lpapp: yes there is qt4 version of it Aug 26 12:48:17 JaMa: 5.1.0 Aug 26 12:49:24 hmm, the name is a bit bogus, but yes, built against qt4 from version qt 5.1.0 Aug 26 12:49:24 lpapp, so how do I replace/fix these binutils? How to backport the fixes with .bbappend? Aug 26 12:49:31 I meant built with qt4-tools-native (qmake2) Aug 26 12:50:10 drasko: have you tried master Aug 26 12:51:22 lpapp, I have tried Yocto poky-dylan-9.0.1 and it builds fine Aug 26 12:51:36 however, project I have received uses older version Aug 26 12:52:44 drasko: then .bbappend for backporting the fix in your layer Aug 26 12:52:58 in your distro or/and bsp layer. Aug 26 12:54:10 .bbappend what? How to use this? Aug 26 12:54:33 please read the reference manual for details. Grep for it, and see how it is used. Aug 26 12:54:39 Anyone know if there's a variable like ${PN} which won't include things like -native? Aug 26 12:54:48 or even the docs about the own layer creation. Aug 26 12:55:09 drasko: http://www.yoctoproject.org/docs/current/dev-manual/dev-manual.html#using-bbappend-files Aug 26 12:55:37 drasko: you will need to find the fix. Aug 26 12:55:59 Stygia: BPN Aug 26 13:09:22 I'm having the issue that in my bbappend for PHP I have a php.ini file, in a php subfolder. I've appended this file to SRC_URI and FILES_${PN}, and yet, I can't find the files in my do_install_prepend. I have FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}:" - How do I properly specify the files I want in a do_install_prepend in my bbappend, which is located in a different layer? Aug 26 13:10:18 http://pastebin.com/ttXciLKM <- The bbappend I'm trying to make work. Aug 26 13:10:57 Stygia: point to the dir that contains the file in do_install_prepend, e.g ${S}/php.ini Aug 26 13:12:15 erbo, As far as I can tell, it doesn't wind up anywhere in tmp/ (work or otherwise. Aug 26 13:12:52 Stygia: then you should have got an error during do_fetch Aug 26 13:13:19 are you sure it reran do_fetch after you edited the bbappend? Aug 26 13:13:52 erbo, Nope. I just normally do "bitbake recipe". Aug 26 13:14:13 erbo, Did it just now. Still, the php.ini file only apperas in subfolders of the (full, not append) recipe that uses the file.. Aug 26 13:14:54 erbo, Sorry. The recipe that _used_ to include the file. It belongs in a php bbappend, but we used to have it as part of our company site recipe. Aug 26 13:15:11 I still see it in that subfolder when I do find tmp/ -name php.ini Aug 26 13:15:19 But not in a folder with php in it. Aug 26 13:15:43 try a clean build of the recipe, bitbake -c cleansstate and then built it again Aug 26 13:15:47 erbo, The recipe is right here http://pastebin.com/ttXciLKM Aug 26 13:15:49 erbo, Will try. Aug 26 13:18:35 erbo, No change. Aug 26 13:19:18 what does the actual recipe look like? Aug 26 13:20:34 Stygia: so you have no php.ini in tmp/work/yada/${PN}? Aug 26 13:21:21 erbo, No, it doesn't look like it. At least, find tmp/ -name php.ini only shows the file in locations that it's been installed to in the path. Aug 26 13:21:25 erbo, And it's the vanilla php recipe Aug 26 13:22:31 is the .bbappend being detected properly? try bitbake-layers show-appends | grep php Aug 26 13:22:43 erbo, Well it sure fails if I have a type in it. Aug 26 13:22:48 erbo, But sure thing, just a sec. Aug 26 13:22:53 ok, good :) Aug 26 13:23:18 erbo, Yea, it does. Aug 26 13:23:36 The bbappend is located in meta-company/recipes-company/php/, and the files in a php/ subfolder to that. Aug 26 13:23:54 erbo, It may be worthwhile to note that it fails trying to build php-native. Aug 26 13:24:24 erbo, And from that command it does look detected. Aug 26 13:24:36 Stygia: how does it fail? Aug 26 13:25:28 tf, by complaining that it can't find php.ini Aug 26 13:25:40 tf, That's the file I'm trying to install into /etc/php5/ in my do_install_prepend Aug 26 13:25:59 tf, It's looking for it in tmp/work/x86_64-linux/php-native/5.4.14-r4.0/ Aug 26 13:26:20 But either way, what's weird is that even after a (successful) do_fetch I don't find a php.ini file anywhere in tmp/. Aug 26 13:27:51 Stygia: replace the ${PN} with ${BPN} Aug 26 13:27:58 in the files path Aug 26 13:28:14 tf, I take it you mean FILESEXTRAPATH_prepend? Aug 26 13:28:15 or just php Aug 26 13:28:19 yes Aug 26 13:28:49 tf, Will do. And _that's_ the variable to get PN without -native postfixes etc... thanks. :P Aug 26 13:29:52 tf, Hmm it still can't find it during do_install, though. Aug 26 13:30:01 try just 'php' Aug 26 13:30:31 Which is... weird. do_fetch is fine, and both php.ini and the other file is in src_uri, and yet, not found in tmp/ Aug 26 13:30:33 tf, I'll try Aug 26 13:30:50 typ in the bbappend Aug 26 13:30:56 SRC_URI is spelled SRI_URI Aug 26 13:30:57 :) Aug 26 13:31:08 but BPN should work Aug 26 13:31:28 erbo, Ah yup. In the recipe it's SRI_URI, though. Aug 26 13:31:36 tf, Still no luck. Aug 26 13:31:45 Stygia: but it should be SRC_URI Aug 26 13:31:55 or? Aug 26 13:32:07 erbo, Oh right, I misunderstood your comment, then. Aug 26 13:32:18 erbo, Seemed like you wanted s/SRI_URI/SRC_URI/g Aug 26 13:32:23 yes Aug 26 13:32:26 definately Aug 26 13:32:47 erbo, And there... Aug 26 13:32:54 erbo, God dammit... programming and slight dyslexia. Aug 26 13:32:58 Sorry for wasting your time. Aug 26 13:33:13 no worries :) Aug 26 13:33:18 erbo, Things are sane, it does work exactly like it should .:P Aug 26 13:33:27 erbo, I had a feeling I was missing something obvious, hmm. Aug 26 13:33:37 erbo, I should make like a spellchecker for this... or basic sanity checker for recipies. Aug 26 13:34:36 Could be pretty cool as an emacs extension, really. Aug 26 13:39:59 except for the emacs part :) Aug 26 13:40:46 do we want consolekit with polkit dependency or is it worth it patching configure to allow disabling it? Aug 26 13:40:48 * tf was waiting who would rise to that :) Aug 26 13:41:28 ah we have to patch it as polkit is in meta-oe Aug 26 13:41:59 JaMa: iirc ross was trying to get rid of that dependency Aug 26 13:42:24 JaMa: vague memory, but todays is holiday in the UK, so he is not around Aug 26 13:44:55 * JaMa testing patch for it already Aug 26 13:45:06 I remember seeing some discussion about it Aug 26 14:02:34 tf: I told Stygia the same half an hour ago. Aug 26 14:02:43 for some reason he makes his life harder. Aug 26 14:03:35 (in fact more than an hour) Aug 26 14:31:18 RP, Can I get advice about what to do when bitbake shows libtool version mismatch errors and tells me to recreate aclocal.m4 and run autoconf again? This is in relation to gnutls. Aug 26 14:31:40 x86_64-linux-libtool: Version mismatch error. This is libtool 2.4.2, but the Aug 26 14:31:48 definition of this LT_INIT comes from an older release. Aug 26 14:31:59 You should recreate aclocal.m4 with macros from libtool 2.4.2 Aug 26 14:32:09 and run autoconf again. Aug 26 14:32:42 halstead: is this a clean build? Aug 26 14:33:10 erbo: sorry to spoil the fun, but if everything goes OK, then there are only 4 undeterministic dependencies now :) Aug 26 14:33:20 RP, I ran bitbake -c clean buildtools-tarball before but I have used the build dir in the past. Also I am using sstate. Aug 26 14:33:39 halstead: bitbake gnutls -c clean might get you past this. Aug 26 14:33:47 RP, Thanks! Aug 26 14:34:01 halstead: failing that it might be safer to remove tmp/ and rebuild Aug 26 14:34:27 halstead: If you're doing what I think you're doing we've probably changed a lot since you last built in that directory :/ Aug 26 14:34:44 RP, I've removed tmp. Aug 26 14:35:49 tf: yeah, today is a bank holiday in here. Aug 26 14:35:54 RP, much has definitely changed and my last build was a MUT. I've just been lucky in the past I guess. Aug 26 14:36:08 RP, why aren't you out riding? Aug 26 14:39:09 Crofton|work: someone scheduled a release freeze on a bank holiday? Aug 26 14:39:27 halstead: in theory upgrading should work, it just doesn't always and we can't test every combination Aug 26 14:39:35 We need a global list of holidays Aug 26 14:39:47 Crofton|work: one of the bikes is also has its engine in bits on the work bench :( Aug 26 14:39:52 I am certain the Americans wanted the freeze out of the way before the labour day weekend Aug 26 14:40:06 well, you should be working on that :) Aug 26 14:40:19 Crofton|work: I need a lot of expensive parts :( Aug 26 14:40:31 RP, I've been improperly using -c clean commands when I should have been deleting tmp. But it's worked until now. Aug 26 14:40:40 * RP is semi looking at parts catalogues atm Aug 26 14:40:47 I switched to offlineimap and realized that oe-commits list is so big. The process hasn't ended :( Aug 26 14:40:59 I would have deleted oe-commits mails if I had known Aug 26 14:43:46 * RP wishes he had a trip to the US coming up soon as the parts are cheaper over there :/ Aug 26 15:11:39 rp, I could bring them to ELCE :) Aug 26 15:11:43 unless they are heavy Aug 26 15:17:03 Crofton|work: they are probably what you'd class as heavy and I'd like to work again before ELCE if I can :) Aug 26 15:17:17 I fogured that :) Aug 26 15:17:21 <[simar|on]> Hey guys, I guess this is more of a python question than yocto but would anyone know of an alternative to have python multiprocessing without /dev/shm? Currently bitbake just fails out if /de/shm isn't found/unwritable. Aug 26 15:17:38 [simar|on]: I had the same issue, but no. Aug 26 15:17:50 my bugreport has been closed, and there will be no alternative backend implementation sadly. Aug 26 15:17:54 I had the same concerns. Aug 26 15:18:13 [simar|on]: are you using chroot? Aug 26 15:18:30 <[simar|on]> lpapp: on a Mac. Aug 26 15:18:43 [simar|on]: blame RP. :p Aug 26 15:20:02 lpapp: Anyone is free to step up and maintain such an implementation. I do not have the time or inclination to do so Aug 26 15:20:14 [simar|on]: https://bugzilla.yoctoproject.org/show_bug.cgi?id=4925 -> you can write your opinion in here. Aug 26 15:20:15 Bug 4925: enhancement, Low, ---, scott.m.rifenbark, RESOLVED WONTFIX, Hard to set up shard memory IPC in chroot for supported distributions Aug 26 15:20:25 RP: that does not mean you should close as invalid. Aug 26 15:20:46 lpapp: it wasn't closed as invalid, was closed as WONTFIX Aug 26 15:20:47 or "wontfix" because they are technically incorrect and form the basis for thinking that this would not be accepted. Aug 26 15:21:08 RP: same for us end users. Aug 26 15:21:10 we suffer. Aug 26 15:21:15 lpapp: If someone comes along and wants to implement *and* support it, we can repoen and talk about it Aug 26 15:21:24 Until then, it is the correct resolution Aug 26 15:21:32 I disagree. Aug 26 15:21:35 it is a task that makes sense. Aug 26 15:21:40 it just has a very low priority. Aug 26 15:21:45 (for you) Aug 26 15:21:53 lpapp: its low enough in priority its been closed Aug 26 15:21:55 live with it Aug 26 15:22:11 [simar|on]: please post your issue and explanation in there. Aug 26 15:24:18 [simar|on]: I tried to make it work for weeks. Aug 26 15:24:34 I gave up on it. If you have any solution that may help me, I will be happy to listen to you. Aug 26 15:24:49 I am currently just using an ubuntu ssh as the work around ... Aug 26 15:25:57 [simar|on]: admittedly, I was frustrated in that bugreport so please forgive that for me. Aug 26 15:26:00 :) Aug 26 15:26:42 I literally tried it for days without any success, but perhaps it is simpler to get done on Mac. I am sure you will not get much documentation from here though, sadly. Aug 26 15:27:34 <[simar|on]> I'm no software architect but the rationale given by the maintainers seems appropriate. I don't see this as a yocto issue if it has been clearly stated what distributions are supported. Aug 26 15:27:53 [simar|on]: I was using a supported distribution. Aug 26 15:28:00 (as written) Aug 26 15:28:37 hmm, looks like toolchain-layer hasn't cuaght up to the gcc restructuring yet Aug 26 15:30:30 lpapp: chroots are not on the official support list Aug 26 15:30:54 RP: where is it documented it is not supported? Aug 26 15:31:01 can you show me the documentation? Aug 26 15:31:14 lpapp: can you point me at where we say it is? Aug 26 15:31:28 sure, debian wheezy is said to be supported. Aug 26 15:31:30 as is. Aug 26 15:32:09 and even that, we do not wanna have it documented, but we wanna get the broken core technology replacable. Aug 26 15:32:12 optionally. Aug 26 15:32:47 lpapp: It says in the documentation that debian wheezy as a chroot is supported? Aug 26 15:33:22 kergoth: :/. Hopefully there shouldn't be anything too painful, mostly just combining of includes Aug 26 15:33:35 RP: why would it? If it does not mention, everything is included, VB, VMware, chroot, kvm, etc Aug 26 15:33:46 if it says debian is supported, then it is supported. Aug 26 15:34:16 lpapp: er, no. None of these are tested configurations Aug 26 15:34:19 lpapp: I don't see the point to argue on that. The error message I saw was easy enough to point out what was going wrong. Aug 26 15:34:21 anyway, let us not focus on the wrong kinda argument. Aug 26 15:34:37 the idea is not to get documented you are blocked. Aug 26 15:34:44 * kergoth nods Aug 26 15:34:46 the idea is to be able to replace the broken core technology. Aug 26 15:35:52 lpapp: please go and talk to the python people about their "broken technology" ;-) Aug 26 15:36:04 simar: do you seriously think if you get an error message for something that is officially meaning it is supported or not? Aug 26 15:36:10 We're just using standard python modules Aug 26 15:36:13 simar: perhaps you should check what an unsupported host says. Aug 26 15:36:22 simar: in fake, yocto reported that my distribution is supported! Aug 26 15:36:27 fact* Aug 26 15:36:57 RP: I have not seen this causing any issues the last 10 years Aug 26 15:37:01 including python. Aug 26 15:37:07 for me, Yocto is the first project. Aug 26 15:37:38 actually this conversation inspired me to add another post in there that yocto actually reported supported distribution which is totally bogus. Aug 26 15:40:12 done Aug 26 15:44:27 Anyone using a chroot should be setting up the proper bind mounts to make the chroot actually useful. without proc, sys, dev/shm, etc, it's pretty useless. Aug 26 15:44:34 schroot makes this easier Aug 26 15:45:10 I tried everything, really, even arch-chroot which is I guess similar to schroot Aug 26 15:45:20 but you are welcome to give the working commands. I could not manage it in days. Aug 26 17:36:51 [simar|on]: have you been able to work the issue around? Aug 26 17:38:05 <[simar|on]> lpapp: I moved over to supported linux distro in the meantime. Aug 26 17:38:21 <[simar|on]> I just experimenting if building was possible on a Mac. Aug 26 17:38:29 ok, good for you. Aug 26 17:38:47 it was a blocker for me, and still not solved. I still use ssh for daily yocto development. Aug 26 17:40:10 otherwise, I could use debian stable in chroot. Aug 26 17:43:16 lpapp: you can still use a chroot, we've been over this many times Aug 26 18:36:59 hello Aug 26 18:40:27 is eglibc supposed to compile fine with callconvention-hard ? Aug 26 18:41:52 http://pastebin.com/HgtKbHnz Aug 26 18:49:09 builds fine here Aug 26 18:49:50 nerdboy: which branch are you using? Aug 26 18:50:30 master Aug 26 18:51:01 hmmm Aug 26 18:51:17 had to do my own tuning in the rpi layer, but it's been building for several weeks Aug 26 18:51:32 although i do need to pull again... Aug 26 18:51:32 i'm new to yocto, any idea how I can try to work out why yours is building and mine isn't? Aug 26 18:51:42 i'm using the fsl layer Aug 26 18:53:08 i'll update after this image build and see what happens Aug 26 18:55:29 it's doing webkit-gtk now and has about 1000 tasks left, so it might take a little while... Aug 26 18:55:52 nerdboy: np, just ping me whenever Aug 26 18:56:20 nerdboy: when you say 'had to do my own tuning to the rpi layer', what were those tunings? Aug 26 18:57:14 to get all the hardfloat stuff right Aug 26 18:57:43 nerdboy: ok, so fixes for the rpi layer specifically, therefore unrelated to eglibc Aug 26 18:58:22 yup, nothing specific to libc Aug 26 19:03:31 is there anyone else facing the following error Aug 26 19:03:33 Fetcher failure for URL: 'http://downloads.sourceforge.net/oprofile/oprofile-0.9.8.tar.gz'. Unable to fetch URL from any source. Aug 26 19:04:01 i am building this for the first time, and it has been 6hours since it started, Aug 26 19:04:14 how will this error effect the final output? Aug 26 19:05:37 can't build if it can't fetch Aug 26 19:05:53 check it manually? Aug 26 19:06:08 ahmedammar: this is what i meant Aug 26 19:06:12 https://github.com/sarnold/meta-raspberrypi/commit/d796222d99fb43b37c296555994ba7c1159ddd64 Aug 26 19:06:38 nerdboy: thanks for that Aug 26 19:06:54 but you glibc missing symbol error doesn't seem related to that Aug 26 19:07:06 yeah i realize that Aug 26 19:07:23 i'm rebuilding again after bumping poky to master-next Aug 26 19:07:34 they are nscd/nss symbols which is part of glibc Aug 26 19:08:00 maybe it just didn't build it for some reason Aug 26 19:08:22 there we go, same crash again .. Aug 26 19:08:24 did the configure args enable it? Aug 26 19:08:39 so, is there a way i can download it manually and begin with the build? Aug 26 19:08:41 check the do_configure log Aug 26 19:08:48 or should i run the bitbake command again Aug 26 19:09:17 ArunKumar: if you can find it, download and touch the .done file Aug 26 19:10:26 can u give me a little more detail please Aug 26 19:10:31 nerdboy: http://paste.ubuntu.com/6029828/ Aug 26 19:11:03 http://paste.ubuntu.com/6029829/ Aug 26 19:11:10 do_configure/do_compile Aug 26 19:14:37 try http://downloads.sourceforge.net/oprofile/oprofile/oprofile-0.9.8/oprofile-0.9.8.tar.gz Aug 26 19:17:08 diff -uN build/tmp.old/work/armv7a-vfp-neon-poky-linux-gnueabi/eglibc/2.18-r0/build-arm-poky-linux-gnueabi/.tmp.config.old build/tmp/work/cortexa9hf-vfp-neon-poky-linux-gnueabi/eglibc/2.18-r0/build-arm-poky-linux-gnueabi/.tmp.config.old | pastebinit Aug 26 19:17:08 http://paste.ubuntu.com/6029848/ Aug 26 19:17:12 nerdboy: ^^ Aug 26 19:17:17 that looks odd to me Aug 26 19:18:33 1) i don't see a config option for nscd, 2) --enable-kernel=2.6.16 looks weird for such a new version of gcc Aug 26 19:18:38 er, glibc Aug 26 19:20:03 and the last part dosen't look good either Aug 26 19:20:25 i think it Aug 26 19:20:38 's the last part that's causing it Aug 26 19:20:43 no idea why though Aug 26 19:26:28 otavio: if i use meta-fsl-arm master-next with everything else being at master is it expected to work? Aug 26 19:34:15 ok nevermind, i think i found it, it was my fault Aug 26 19:37:38 incorrect use of DISTRO_FEATURES in local.conf it seems Aug 26 19:38:36 nerdboy: thanks for your time Aug 26 19:43:15 unrelated question: is there a reason the autobuilders don't publish Packages.gz for the debs? Aug 27 00:29:39 hiya folks, can anyone tell me how to enable an spi device on my beaglebone black? Aug 27 01:36:11 hiya folks, can anyone tell me how to enable an spi device on my beaglebone black? Aug 27 01:42:37 and where do I find the device tree source file? Aug 27 01:46:49 I am getting "ERROR: Can NOT get PRAUTO, exception timed out" and "ERROR: Function failed: package_get_auto_pr" after updating to master:4517cdbfafdc7c5fa16bb5d09ed70c8e0da6a8c8. Any ideas? Aug 27 01:47:30 in local.conf I have PRSERV_HOST = "localhost:0" **** ENDING LOGGING AT Tue Aug 27 02:59:58 2013