**** BEGIN LOGGING AT Thu Feb 22 03:00:05 2018
**** BEGIN LOGGING AT Thu Feb 22 07:39:31 2018
Feb 22 09:56:49 <pn> Hi all, I am trying to build npm http-server. I have custom patch which is applied in the bitbake. But the patch is invalidated/removed during "do_compile" stage (code came back to original/do_unpack) state. How to patch NPM packages in yocto? Thanks in advance!
Feb 22 09:58:30 <pn> Not sure whether it's because of npm cache operations in "npm.bbclass" during do_compile.
Feb 22 10:02:46 <T_UNIX> I'm really puzzled by this: I've two recipes providing something via `RPROVIDES_${PN} = "foo"`. Now, when bitbake fails during `license_create_manifest`because it apprently fails to inversely locate the RPROVIDER `Exception: FileNotFoundError: [Errno 2] No such file or directory: '/cache/yocto/bar/bar-r1/tmp/pkgdata/bar-r1/runtime-reverse/foo'`
Feb 22 10:05:58 <T_UNIX> s;when;;
Feb 22 10:15:50 <T_UNIX> My assumption was that I could add a recipe X to an image and have two `RPROVIDES_${PN} = "X"` in two distinct packages named Y and Z? Do I need to specify anything else? Or follow some naming scheme?
Feb 22 10:20:20 <T_UNIX> brb
Feb 22 12:09:01 <yocti> New news from stackoverflow: Failed to obtain external Linaro toolchain version error <https://stackoverflow.com/questions/48926676/failed-to-obtain-external-linaro-toolchain-version-error>
Feb 22 13:39:17 <yocti> New news from stackoverflow: bitbake SRC_URI file:// <https://stackoverflow.com/questions/27815990/bitbake-src-uri-file>
Feb 22 14:29:22 <lukma> Dear All,
Feb 22 14:29:22 <lukma> What is a recommended way to provide nativesdk specific package group?
Feb 22 14:29:22 <lukma> nativesdk-packagegroup-sdk-host.bbappend
Feb 22 14:29:22 <lukma> RDEPENDS_${PN}_mingw32 = "\ packages "
Feb 22 14:29:52 <lukma> Or packagegroups-XXX-sdk.bb
Feb 22 14:29:53 <lukma> with inherit packagegroup nativesdk ?
Feb 22 14:30:21 <rburton> either works :)
Feb 22 14:30:33 <rburton> i'd do latter if its your own group
Feb 22 14:33:59 <lukma> rburton: But with the second case I do have
Feb 22 14:34:02 <lukma> ERROR: Nothing PROVIDES 'virtual/x86_64-pokysdk-linux-compilerlibs'
Feb 22 14:34:47 <lukma> This shows up when I do try to build image for target (not nativesdk host)
Feb 22 14:35:07 <lukma> Do I need to put some guards for this packagegroup-sdk.bb recipe ?
Feb 22 14:35:20 <lukma> Like PACKAGE_ARCH = "xxxx" ?
Feb 22 15:41:14 <BubuIIC> Can I bbappend a packagegroup recipe?
Feb 22 15:45:10 <kergoth> it's just a recipe like any other, yes, bbappend works on any recipe
Feb 22 15:45:34 <BubuIIC> kergoth: thx... trying it out right now
Feb 22 16:37:03 <armpit> RP or rburton , either one of you going to need the ab or can I fire off a build?
Feb 22 16:37:19 <rburton> armpit: feel free to fire one now imo
Feb 22 16:38:42 <mattsm> are there stables images of builds saved somewhere? i'm just looking to grab the latest stable images for a particular machine (qemux86 in this case)
Feb 22 16:38:51 <mattsm> trying to avoid running the build step as part of some tests
Feb 22 16:39:25 <RP> armpit: go for it
Feb 22 16:40:39 * armpit cool zeddii merge 4.1 changes
Feb 22 16:44:40 * zeddiii wonders how he got an extra 'i' in his nick
Feb 22 16:44:49 <zeddiii> and why freenode won't let me change it to zeddii
Feb 22 16:44:58 * zeddiii curses 30 year old tech
Feb 22 16:45:37 <zeddiii> I must have a rogue IRC running somewhere.
Feb 22 16:46:22 <zeddiii> ahahah
Feb 22 16:48:53 <mattsm> zeddiii two others even
Feb 22 16:50:20 <CoRfr> @mattsm maybe checkout the autobuilder artifacts: http://autobuilder.yoctoproject.org/pub/releases/yocto-2.3.2/machines/qemu/qemux86/
Feb 22 16:51:13 <mattsm> CoRfr nice, just what i was looking for...
Feb 22 16:51:25 <mattsm> could not navigate the maze of that web site myself
Feb 22 16:56:18 <CoRfr> Has anyone tried to work with qemuarm and some other machine/CPU arch than versatilepb? I'm trying to build for cortex-a15 and use -machine virt -cpu cortex-a15 without much luck.
Feb 22 16:58:13 <peacememories> hmm, seems there's no real place to ask repo (the tool) related questions^^'
Feb 22 16:58:35 <peacememories> does anyone here know if i can tell repo to use a different remote for pushing than pulling?
Feb 22 16:59:22 <CoRfr> @peacememories the place would be https://groups.google.com/forum/#!forum/repo-discuss I guess
Feb 22 17:00:13 <CoRfr> I think the push option is defined in the <remote ... review="<https remote>"/>
Feb 22 17:00:52 <CoRfr> So that would be what repo pushes to when doing repo upload
Feb 22 17:05:02 <kergoth> it's 'pushurl'
Feb 22 17:05:04 <kergoth> see https://gerrit.googlesource.com/git-repo/+/master/docs/manifest-format.txt
Feb 22 17:05:40 <kergoth> took me forever to find that link googling
Feb 22 17:07:19 <peacememories> woop, thanks friends :)
Feb 22 17:07:33 <peacememories> apparently i didn't see that option^^'
Feb 22 17:08:07 <peacememories> that helps a lot ^_^
Feb 22 17:08:35 <peacememories> now i only need to find out how i can configure hooks after syncing, and how to best store some helper scripts. and how to configure the manifest repo from the manifest ^^' that's gonna be fun
Feb 22 17:13:38 <kergoth> personally i just use git url replacements to handle pushing
Feb 22 17:15:02 <kergoth> $ git config -l|grep pushinsteadof|grep yocto
Feb 22 17:15:02 <kergoth> url.git@push.yoctoproject.org:.pushinsteadof=git://git.yoctoproject.org/
Feb 22 17:15:31 <kergoth> of course that requires each user add that to their global git config, tradeoff
Feb 22 17:17:23 <peacememories> seems the pushurl does the job very well :)
Feb 22 17:18:07 <peacememories> which makes peacememories very happy, because that's the one feature i was looking for for our repo management
Feb 22 17:29:08 <radsquirrel[m]> is it written down somewhere how exactly patches make their way from oe-core, bitbake, etc into poky?
Feb 22 17:29:44 <radsquirrel[m]> I maintain a yocto based project..we have a bunch of layers.  I'm looking at changing our build process because it doesn't feel optimal but I'm not sure where to start.
Feb 22 17:30:02 <radsquirrel[m]> fishing for ideas.
Feb 22 17:30:38 <radsquirrel[m]> like does every oe-core patch get manually cherry-picked by someone one at a time into poky?
Feb 22 17:32:03 <RP> radsquirrel[m]: the tool used is called combo-layer
Feb 22 17:32:15 <RP> its not something we're entirely happy with
Feb 22 17:32:54 <radsquirrel[m]> nice.  how did I not find this...
Feb 22 17:33:02 <radsquirrel[m]> thanks for the link.
Feb 22 18:10:52 <stephano> rburton: SWAT: MUTMAGEDDON oe-selftest... ross is aware?
Feb 22 19:02:54 * armpit hmm failed to initialize KVM: Cannot allocate memory is back
Feb 22 19:10:08 <JPEWhacker> Hmm.... Doxygen defaults to generating CHM documentation in 1250, but that codepage isn't in the uninative tarball.... Is there a reason we don't include all the codepage conversion tables in uninative? I'm assuming it gets pretty large....
Feb 22 19:53:56 <rburton> JPEWhacker: yes it gets large
Feb 22 19:54:07 <rburton> you can make an argument for specific ones though
Feb 22 19:54:21 <rburton> or, fix doxygen to use an encoding from this millenium ;)
Feb 22 20:00:02 <CoRfr> Has anyone face qemu errors such as "qemu-system-i386: Trying to execute code outside RAM or ROM at 0xc138d060" when running qemu (either x86 or arm).
Feb 22 20:00:47 <CoRfr> I can reproduce it on images from the autobuilder (2.3.2) even though I'm working with 2.2.3 at the moment (and kernel 4.4)
Feb 22 20:01:40 <CoRfr> Weirdly enough, it is much easier to reproduce in the cloud (I tried both on Azure and GCE)
Feb 22 20:11:24 <JPEWhacker> rburton: The problem is that it is the default, so I could patch doxygen to make it something more sane (1252).... or patch every recipe that uses it?
Feb 22 20:30:30 <rburton> JPEWhacker: i'd be patching doxygen to use utf8 by default assuming thats useful for windows machines that read chm files
Feb 22 20:31:11 <JPEWhacker> rburton: Ya, that probably makes the most sense in the long run.
Feb 22 20:32:32 <JPEWhacker> rburton: Honestly, I wish I could turn off doxygen all together... I don't know if it is particularly useful in a build. Do other packages usually generate man pages and such?
Feb 22 20:42:20 <kergoth> we generally have packageconfigs to control it, but it's usually defaulted to off, afaik
Feb 22 21:47:02 <majuk> Hey all, I am setting up a custom recipe [https://pastebin.com/RCCJih90] to fetch a .c file from a git repo and compile it. I can successfully clone the repo. However, invoking `bitbake myrecipe`, the working directory for the recipe has no 'git' folder. log.do_fetch shows nothing interesting -> https://pastebin.com/kQPESd50
Feb 22 21:47:13 <majuk> Any ideas on how to troubleshoot this would be greatly appreciated. Thanks.
Feb 22 22:01:02 <rburton> JPEWhacker: we've a distro feature you can respect and turn off
Feb 22 22:01:07 <majuk> wow, nevermind, I have the dumb
Feb 22 22:32:19 <kergoth> Hmm, would be nice to silence bitbake's message about not using CACHE. If it's not set, it can be for a valid reason
Feb 22 22:54:18 <rburton> zeddii: can we expect the kernel-devsrc reduction patch soon? M3 is veery close
Feb 22 23:16:14 * armpit needs a beer or two
**** ENDING LOGGING AT Fri Feb 23 03:00:01 2018