**** BEGIN LOGGING AT Wed Mar 08 03:00:02 2017 Mar 08 03:21:17 mattsm: nice! Mar 08 03:29:55 bluelightning: a little birdie told me devtool reception at SCaLE was ++ Mar 08 03:30:24 mattsm: ah ok - what was the context there? presentation / tutorial / other ? Mar 08 03:30:37 sorry, moto-timo ^ Mar 08 03:30:40 lol Mar 08 03:30:51 behanw gave a talk/demo Mar 08 03:31:10 so devtool++ and crops++ Mar 08 03:31:17 awesome :) Mar 08 03:31:21 yup! Mar 08 03:31:43 critical mass Mar 08 04:40:04 anybody else try mainline kernel drm/mesa/xorg on imx6q ? Mar 08 04:41:16 ubuntu xenial on top of 4.9.9 kernel w/openbox does glxgears @ ~80 fps Mar 08 04:41:39 * nerdboy just wondering if that's good/bad/mediocre Mar 08 04:42:15 using armade-etnaviv video driver Mar 08 04:43:39 *armada even Mar 08 05:28:05 nerdboy: 80fps is good Mar 08 05:28:13 anything above 24fps is good Mar 08 05:34:18 so you think foss driver/kernel/blah should be snappy enough for qt5 touch ui? Mar 08 06:57:47 if gpu drivers are good then yes Mar 08 07:03:38 seems good enough to package up the missing bits Mar 08 08:01:26 Hi Mar 08 09:25:13 I'm working on layers meta-xilinx with conf zybo-linux-bd-zynq7, someone could explain me what's the difference between two files ? zybo-linx-bd-zynq7.conf and zybo-linux-bd.bb Mar 08 09:32:47 bluelightning: Just saw your ping above. meta-raspberrypi kernel recipes definitely need a bit of streamlining. I think it was discussed a couple of months ago I'll check who raised it Mar 08 10:21:27 Hiya, I've got a Yocto 1.6 environment I'm wondering about building on Ubuntu 16.04 now that 14.04 is EOL. Is there a standard process or HOWTO for using containers as that sounds like potentially a great solution, or am I best just using a VM with 14.04 installed? Mar 08 10:25:49 dp19282: I've struggled with 1.6 and Ubuntu 16.04 due to the compiler upgrades after the Ubuntu 14.04 release Mar 08 10:26:45 paulbarker: Yeah, I'd spotted that. I did try using update-alternatives with gcc and it gets a lot further but its not the whole picture it seems. Mar 08 10:27:05 Ok. Ubuntu 14.04 isn't EOL yet though Mar 08 10:27:37 https://wiki.ubuntu.com/Releases Mar 08 10:28:23 I've been setting up a 14.04 environment using Docker from within Ubuntu 16.04 for this at work Mar 08 10:28:23 Indeed - but its un-avoidably heading that direction, and other newer projects as well mean that 16.04 becomes a sensible development platform for me. Plus I'd like to avoid lots of dual booting :-) Mar 08 10:29:13 Interesting - I've not used docker but am I right in thinking that it sits on top of lxc? Mar 08 10:30:51 It's not lxc, uses the same kernel APIs though. You'd have to look at their docs for more info Mar 08 10:31:07 I do wonder sometimes which the ideal situation is when you realise that you may have to do a re-build in 10-20 years time, not knowing what the environment and legacy support will look like then... :-D Mar 08 10:31:43 Boot an old Linux on an isolated network Mar 08 10:32:01 ...if you have compatible hardware in a cupboard...! Mar 08 10:32:01 Just make sure you have archives of all the sources you use Mar 08 10:32:26 Oh I'd boot it in a VM Mar 08 10:32:30 Oh I see - docker used to use lxc but dropped it... Mar 08 10:52:24 Hey ! I want to add a receipe that forces the bluetooth drivers to be installed in the kernel. I can't find where it is written in the doc, can you help ? Mar 08 10:54:56 (Also I'm about to be mad because when I modify something in the menuconfig it is not saved !!! I tried bitbake -c compile -f virtual/kernel but it works one day and not the other day !) Mar 08 10:56:47 http://www.yoctoproject.org/docs/2.2.1/kernel-dev/kernel-dev.html#generating-configuration-files Mar 08 10:57:50 joshuagl: thanks ! Mar 08 11:10:32 RP: so i'd dropped vulcan from mut as it breaks under musl :) Mar 08 11:12:02 rburton: right... Mar 08 11:12:08 rburton: now you tell me :) Mar 08 11:12:42 rburton: I have a corrected buildtools tarball coming Mar 08 11:14:17 going to send a patch for mingw to use a wildcard append Mar 08 11:14:31 rburton: its fixed Mar 08 11:14:31 rburton: already one in master Mar 08 11:14:35 sorted Mar 08 11:14:47 anyone know what the eSDK error is? Mar 08 11:14:55 https://autobuilder.yoctoproject.org/main/builders/nightly-ppc/builds/1098/steps/BuildImages_2/logs/stdio Mar 08 11:15:04 theres a few non-wildcard appends left in that layer Mar 08 11:15:16 yeah i hit that in mut too Mar 08 11:16:39 i'm suspecting buildhistory.bbclass: Only execute get_extra_sdkinfo when sdk is enabled Mar 08 11:21:30 rburton: the vulkan problem is musl not having secure_getenv(). I guess patching to use getenv() instead in that case is acceptable? Mar 08 11:21:45 this is what systemd does Mar 08 11:22:08 (our systemd, not upstream) Mar 08 11:22:49 * RP submitted that patchelf fix upstream so I guess I should change the patch status now... Mar 08 11:23:26 jku: guess so Mar 08 11:24:44 joshuagl: really puzzled about the eSDK error :( Mar 08 11:24:57 joshuagl: I don't think its anything specific to the changes Mar 08 11:25:02 tenner says its the buildhistory change Mar 08 11:25:22 rburton: did you see it in a non-ppc builder? Mar 08 11:25:33 rburton: why only ppc eSDK ? Mar 08 11:25:45 rburton: and why not on the new cluster which ran the same build? Mar 08 11:26:26 no idea :) Mar 08 11:26:52 RP: will you be repushing next with fixed uninative? SHAs changed. Mar 08 11:27:27 rburton: pushed Mar 08 11:33:18 rburton: ok with merging -next to master Mar 08 11:33:18 ? Mar 08 11:33:41 rburton: I added Martin's gcc-runtime tweak Mar 08 11:34:21 thanks Mar 08 11:38:23 RP: useradd has races Mar 08 11:38:37 got a few failures on my mut run which i suspected that was to blame Mar 08 11:40:19 rburton: you did notice I removed part of that patch? Mar 08 11:40:26 ah, no Mar 08 11:40:49 rburton: As it stands now it just deletes an unused function so should be safe Mar 08 11:41:19 still has teh code that frustrates me Mar 08 11:41:27 (PSEUDO_PASSWD points into recipe sysroot) Mar 08 11:41:39 rburton: but it has to do that? Mar 08 11:41:51 but it doesn't exist when the code runs Mar 08 11:42:01 recipe sysroot is used in a rootfs postinst Mar 08 11:42:20 not the rootfs sysroot, the sysroot of the recipe that wrote the postinst Mar 08 11:43:09 rburton: "This preinstall function can be run in four different contexts" Mar 08 11:43:14 rburton: that is the real issue Mar 08 11:43:44 rburton: that piece should probably be in useradd_sysroot() Mar 08 11:47:59 rburton: something like http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/wip-rss2&id=a23d905df31b797154648685a8e93f69d644b43e ? Mar 08 11:49:17 maaaybe Mar 08 11:49:19 i can give it a go Mar 08 11:50:19 rburton: so ok with -next? Mar 08 11:50:48 i think so Mar 08 11:51:45 v9 for go Mar 08 11:51:53 i'll try a build locally Mar 08 11:53:11 rburton: sounds good Mar 08 12:16:21 Hello all, I am wondering if someone can entertain a question?\ Mar 08 12:17:31 I am quite new to yocto/oe, so I apologise in advance. I have SDK (spit from a yocto build) for cross compiling for arm, the SDK contains Thrift-0.9.3 Mar 08 12:17:44 I want the Thrift compiler, within the SDK, to be able to build on the host x86 Mar 08 12:18:03 Currently building the yocto thrift recipe gives me a Thrift compiler can't run on the host x86 machine (i.e an arm thrift compiler) Mar 08 12:18:22 And in all honesty I am not too sure where to start Mar 08 12:22:20 the SDK needs to contain something ilke nativesdk-thrift Mar 08 12:22:50 where nativesdk means "build for a SDK, running on the host" Mar 08 12:23:38 Right, thrift recipe provided by yocto actually has line ' BBCLASSEXTEND = "native nativesdk" ' Mar 08 12:23:51 Which I would have thought does that Mar 08 12:24:37 (Well provided by oe sorry) Mar 08 12:38:42 that just means it can be built as native or nativesdk Mar 08 12:38:51 but if you bitbake thrift, you get the target one Mar 08 12:41:30 Alright thanks, that clarifies that part. Then perhaps could you point me somewhere to help guide me in getting something like nativesdk-thrift into the SDK? Mar 08 12:42:01 how are you building the sdk? Mar 08 12:45:50 Running the bitbake core-image -c do_populate_sdk command Mar 08 12:50:39 Hi, I am reproducing this bug poky in the branch jethro: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77772 Mar 08 12:50:40 Bug 77772: was not found. Mar 08 12:50:48 Fortunately, it is solved in morty, which uses gdb 7.11.1 instead of 7.9.1. Mar 08 12:50:54 Do any of you know how could I know which commit from gcc (gdb) solved that issue? Would it be possible to backport it to gdb 7.9.1 (using jethro) Mar 08 12:52:07 Luigi_Uccelli: TOOLCHAIN_HOST_TASK+=nativesdk-thrift in your image Mar 08 12:53:10 Thanks a thousand! Mar 08 12:53:31 I believe you need to use TOOLCHAIN_HOST_TASK_append+=nativesdk-thrift Mar 08 12:53:35 notice the 'append' Mar 08 12:53:44 note that append and += doesn't make sense Mar 08 12:54:00 I needed both Mar 08 12:54:35 you're half right, the default is a lazy assign so images need to use TOOLCHAIN_HOST_TASK_append = " nativesdk-thrift" Mar 08 12:54:49 note the leading space in the value Mar 08 12:55:33 noted Mar 08 12:56:01 If you get thsi error when installing "Setting it up...ls: cannot access / opt/mysdk-v2/environment-setup-*: No such file or directory" then try both Mar 08 12:56:10 that is what happened for me, at least Mar 08 12:56:11 How to set --with-xkb-config-root=/usr/lib through do_configure from /poky/meta/recipes-graphics/xorg-lib/libxkbcommon_0.5.0.bb ? Mar 08 12:57:04 Alright, I will keep that in mind. Thanks to both of you. Mar 08 12:57:39 Ramose: add that to EXTRA_OECONF Mar 08 12:58:45 rburton: Sorry but may I know how ? Mar 08 13:00:27 Ramose: EXTRA_OECONF += "—with-xkb-config-root=/usr/lib/" Mar 08 13:00:43 EXTRA_OECONF = "--with-xkb-config-root="/usr/lib" Mar 08 13:01:10 rburton: Oh , right. Mar 08 13:01:36 rburton: Also , do I need to define in do_configure() ? Mar 08 13:02:03 no Mar 08 13:02:10 that variable is what do_configure uses Mar 08 13:02:24 ok, Let me try it Mar 08 13:02:42 one EXTRA_OECONF = "--disable-docs" already defined in same file Mar 08 13:02:48 http://www.yoctoproject.org/docs/latest/ref-manual/ref-manual.html#ref-classes-autotools Mar 08 13:02:55 yes, that's why i said += Mar 08 13:03:06 ok. Mar 08 13:04:03 (or add it to the existing one) Mar 08 13:05:36 rburton: Thanks do_configure has passed now Mar 08 13:09:41 someone could explain me, in meta-xilinx-layers, in bb files, what is the compatible_machine for ? Mar 08 13:13:46 RP: so go builds everywhere apart from ppc now, and fails neatly instead of barfing Mar 08 13:14:29 shall throw at AB just to see what happens Mar 08 13:14:31 rburton: so if we tweak the sato inclusion, we're good? Mar 08 13:14:40 pretty much Mar 08 13:14:49 obviously the sato thing is a testing patch Mar 08 13:17:38 joshuagl: https://autobuilder.yoctoproject.org/main/builders/nightly <— only workers are f24. pretty sure that's not right. can you fix that? Mar 08 13:33:47 yohboy, got specific reference Mar 08 13:36:17 rburton: I've no idea as to the history of that, I can try and work it out though Mar 08 13:39:39 hello Mar 08 13:40:14 i am confused by this line in kbd recipe (poky/meta/recipes/core/kbd): PACKAGECONFIG[pam] = "--enable-vlock, --disable-vlock, libpam," Mar 08 13:40:35 can anybody clarify why is disable-vlock after enable-vlock listed there? Mar 08 13:41:03 http://www.yoctoproject.org/docs/latest/ref-manual/ref-manual.html#var-PACKAGECONFIG Mar 08 13:41:04 (using jethro) Mar 08 13:41:58 cornel: ^ Mar 08 13:42:29 thank you rburton , i'm trying to understand how this works Mar 08 13:46:39 rburton, by default it takes all those parameters? i can;t figure how the 'kbd' ends with with --disable-vlock in the end Mar 08 13:47:07 and there's no bbappend in my setup Mar 08 13:48:21 crofton|work : so if the machine name doesn't match with the compatible_machine, it skip the recipe ? Mar 08 13:51:50 yohboy: any value in MACHINEOVERRIDES, which machine name is in. Also for meta-xilinx SOC_FAMILY is in machineoverrides Mar 08 14:04:35 nrossi: ok thank you ! Also yesterday, you pointed me to a bb file (zybo-linux-bd.bb) to help me to make my own recipe. In my repo, I don't have this file, but I have zybo-linux-bd-zynq7.conf. It looks similar but it seems there is more output fil, provide DTS (MACHINE_DEVICETREE) but not the hdf file. I think the hdf file is loaded in an other recupe : digilent-hdf.bb. My question is: what is exactly the difference between the conf fi Mar 08 14:09:29 yohboy: not sure what digilent-hdf.bb is. But the other two files you mention are different. the .bb is a recipe and the .conf is a machine config. The zybo-linux-bd.bb is a recipe, and its job is to fetch and extract the bitstream and ps7_init files from the hdf only. Mar 08 14:15:56 cornel: because you don't have pam in DISTRO_FEATURES, i'd guess. Mar 08 14:16:12 rburton, i believe i do Mar 08 14:16:18 prove it :) Mar 08 14:16:25 one minute please :) Mar 08 14:16:41 (bitbake -e kdb, then look at the assignment for DISTRO_FEATURES and the history of EXTRA_OECONF) Mar 08 14:18:39 ok Mar 08 14:19:45 in bitbake -e | grep DISTRO_FEAT | grep pam, exists Mar 08 14:20:46 EXTRA_OECONF does not have pam Mar 08 14:20:56 cornel: bitbake kdb -e | grep PACKAGECONFIG | grep pam Mar 08 14:21:14 pastebin Mar 08 14:22:33 nrossi, PACKAGECONFIG="pam" Mar 08 14:22:58 rburton, it's hard to use pastebin since the build is happening on a different system than this one Mar 08 14:23:18 scp? xsel? Mar 08 14:23:41 it;s even in another country but let me try scp :) Mar 08 14:23:58 pretty sure scp doesnt care how far away it is :) Mar 08 14:24:46 rburton, yes, but the company is pretty harsh on tunneling ... Mar 08 14:25:12 rburton, you don't want the whole -e, right? Mar 08 14:26:57 cornel: would make it easier to see the context Mar 08 14:32:30 sorry to say but there's some proprietary info in it :( Mar 08 14:33:07 rburton, nrossi, thank you very much Mar 08 14:33:40 cornel: just cut out the bits around DISTRO_FEATURES and EXTRA_OECONF then, including the history block above each one Mar 08 14:37:15 nrossi: ok, so what's the job of conf file ? Mar 08 14:40:56 yohboy: you should probably read the ref manual, http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#metadata-machine-configuration-and-policy-configuration Mar 08 14:41:37 rburton, https://paste.fedoraproject.org/paste/~J~u30g6-RP0yenNLB1a7l5M1UNdIGYhyRLivL9gydE= Mar 08 14:43:57 nrossi: thank you, I'm going to read it ! Mar 08 14:44:14 cornel: you didn't include the history bits Mar 08 14:44:28 cornel: also, that says --enable-vlock Mar 08 14:44:34 which is what you'd expect with pam enabled Mar 08 14:44:34 rburton, i'm sorry, i don't what this is Mar 08 14:44:52 rburton, yes, but the rpms do not include vlock .... Mar 08 14:45:24 could it be that actually the sources of kbd do not include it .... Mar 08 14:45:43 you said it was passing —disable-vlock Mar 08 14:45:59 if the actual problem is something else then stop looking at configure Mar 08 14:46:16 rburton, you're right, but i was just deducing from the lack of vlock, not really knowing .. Mar 08 14:46:33 when in doubt, look at the logs Mar 08 14:46:42 log.do_configure will tell you exactly what it ran Mar 08 14:46:56 so i'ld check if kdb even ships vlock, and if it does what you need to make it work Mar 08 14:47:48 no log.do_configure in temp Mar 08 14:47:54 let me cleanall and try again Mar 08 14:48:10 bitbake kdb -C unpack is faster Mar 08 14:52:06 rburton, it says --enable-vlock, .... creating vlock Makefiel ,then it says : vlock: no Mar 08 14:52:10 pfff Mar 08 14:52:36 cornel: so, look in configure.ac and find out what it wants to turn it on Mar 08 14:54:14 thank you very much rburton , i'll investigate more and let you know if i find anything interesting Mar 08 14:56:31 strange, for some reason it thinks that VLOG_PROG="no" Mar 08 14:56:49 VLOCK Mar 08 14:56:51 Hi, i compiled yocto CC toolchain, on Ubuntu 14.04. works fine. i tar it, untar on ubuntu 16.10. It doesn't work. Mar 08 14:56:53 * Mar 08 14:56:53 bash: /opt/poky/toolchain/x86_64-pokysdk-linux/usr/bin/arm-poky-linux-gnueabi/arm-poky-linux-gnueabi-gcc: No such file or directory Mar 08 14:57:30 even tho, the file is there. it seems there is some ubuntu multilib issue like this. anyone had similar issue? Mar 08 14:57:52 suvirb, Is it located at the same path as it was before? Mar 08 14:58:14 nopes Mar 08 14:58:18 does that matter Mar 08 14:58:22 i can do that tho Mar 08 14:58:22 It may Mar 08 14:58:29 i'll give it a try Mar 08 14:58:30 suvirb: the toolchain comes in a tarball already, just copy that. on extract it gets relocated. Mar 08 14:58:59 rburton: was looking at that ross/go strip failure and thinking "what now?" but you didn't include recent master changes! Mar 08 14:59:05 You could also run strace to see what it was doing at the time it whinged to see if that gives you a clue as to whats going on Mar 08 14:59:08 RP: whaat Mar 08 14:59:13 swore i rebased Mar 08 14:59:30 rburton: I suspect not Mar 08 15:00:37 RP: ah didn't include the uninative bump Mar 08 15:02:31 rburton: or binutils or patchelf Mar 08 15:02:32 yes, works Mar 08 15:02:36 arm-poky-linux-gnueabi-gcc --version Mar 08 15:02:36 arm-poky-linux-gnueabi-gcc (GCC) 4.9.2 Mar 08 15:02:46 thanks rburton for quick help Mar 08 15:02:52 suvirb, what was it in the end? Mar 08 15:02:57 path Mar 08 15:03:06 ace Mar 08 15:03:09 i created same dir path. Mar 08 15:03:32 like this : /opt/poky/1.8/sysroots/ Mar 08 15:03:45 this is what i had on my machine which compiled the toolchain Mar 08 15:05:20 RP: won't bother re-firing, i'll review the fails manually. i forgot to exclude x32 i see. Mar 08 15:14:59 netcat-openbsd seems to be blacklisted (as seen here http://cgit.openembedded.org/meta-openembedded/tree/meta-networking/recipes-support/netcat/netcat-openbsd_1.105.bb?h=master#n46) but it builds fine on master Mar 08 15:15:33 is it still the case to have it blacklisted? what's the procedure to remove it from the blacklist? Mar 08 15:17:30 plp_: fix is in master-next, you can ping joeythesaint to merge it Mar 08 15:17:50 ok, thanks! Mar 08 15:37:14 nrossi: sorry to bother you, do you have also some docs about device tree ? I'm not sure to understand all the file dtsi and dts include in the recipe zybo-linux-bd-zynq7.bb. We have to generate them with xilinx tools ? then yocto compile them into the DTB ? Mar 08 15:43:52 so yet again our gst git recipes are a major release out of the tarballs Mar 08 15:44:34 1 out of 2 hunks FAILED -- rejects in file screen.c Mar 08 15:44:34 Patch Add-noexit-option.patch does not apply (enforce with -f) Mar 08 15:44:34 ERROR: screen-4.5.1-r0 do_patch: Function failed: patch_do_patch Mar 08 15:44:54 Anyone seen ^^^^ before I dig in and fix it? Mar 08 15:45:34 paulg: ack Mar 08 15:46:26 JaMa: so why don't you inherit remove-la :) Mar 08 15:46:34 JaMa: it makes everything so much nicer! Mar 08 15:48:13 paulg: any special configuration changes at your end ? (didn't fail here) Mar 08 15:49:13 maxin, checking ; probably is a local bbappend that I forgot existed. Mar 08 15:50:43 yohboy: that recipe doesn't handle dts/dtsi, the device-tree recipe (of meta-xilinx) can do the compiling for dts's that are not in the kernel. Also there is no automation of the xilinx tools in meta-xilinx, so you will need to generate the dts/dtsi files externally Mar 08 15:53:13 maxin, yep - a bbappend in meta-overc, so I'll simply have to refresh the patches added from there. Mar 08 15:54:15 paulg: thanks for sharing that info Mar 08 15:56:34 nrossi: my bad there are not in the recipe, there are in the conf. But okay, they must be generate externally, then the device tree recipe can compile them. And so the conf say which one we want to compile ? Mar 08 15:58:19 rburton: I plan to inherit, but our DISTRO is setting own DISTRO_INHERIT so the change in default value in oe-core wasn't applied here and this was the first recipe which failed because lafiles Mar 08 16:00:04 yohboy: you can add the dts/dtsi files via a bbappend on device-tree, then set MACHINE_DEVICETREE in the .conf or in the bbappend with MACHINE_DEVICETREE_. Mar 08 16:05:23 nrossi: ok, thank you very much for help ! I'll try to make my own machine support with my own recipe in the next few day ! Mar 08 16:25:15 rburton: perhaps remove-la should be elevated to defaults Mar 08 16:27:11 khem: it is Mar 08 16:27:30 ok Mar 08 16:34:58 rburton: I tested go on mips and samples compiled and ran on qemumips Mar 08 16:35:17 yay Mar 08 16:35:19 it needed version 1.8 for mips support Mar 08 16:35:42 ppc32/x32 are not supported backends for linux Mar 08 16:35:52 yeah Mar 08 16:36:05 ppc32 is a secondary supported platform anyway Mar 08 16:36:24 x32, there is nacl support Mar 08 16:36:33 hopefully someone will port it to linux Mar 08 16:37:07 but in practice, real use cases are arm/x86 32/64bit Mar 08 16:41:09 here is complete list https://golang.org/src/go/build/syslist.go Mar 08 16:56:22 JaMa: performance with rss is better with la removal ;-) Mar 08 16:56:31 (far fewer file relocations) Mar 08 16:58:10 00:27.59,0:00:17.03,0:00:01.44 Mar 08 16:58:10 00:26.47,0:00:16.35,0:00:01.40 Mar 08 16:58:47 before and after gplv2 removal, parsing 1s faster for cold cache, 0.6s faster for host codeparser cache Mar 08 16:58:56 s/host/hot/ Mar 08 17:01:56 hey is it bad to add a recipe that I didnt create to my layer Mar 08 17:02:20 or should I just point the user to add that recipe on their own Mar 08 17:02:31 in the README Mar 08 17:02:45 ideally, depend on the other layer Mar 08 17:03:16 the recipe isnt automatically installed on AGL Mar 08 17:03:28 and you mean RDEPENDS Mar 08 17:05:23 i mean your layer shouldn't just copy and paste from other layers, instead just add the other layers to your bblayers Mar 08 17:05:28 or list it as a layer dependency Mar 08 19:10:35 RP: usually parsing is worse for folks who use many layers :( Mar 08 19:11:54 rburton: if you only need one recipe and not other 50 from a layer then one needs to use masking Mar 08 19:12:10 khem: whats masking? Mar 08 19:12:20 I think zephyr project had the whitelisting bbclass Mar 08 19:12:35 no it was another intel project ostro I think Mar 08 19:13:11 https://github.com/ostroproject/meta-ostro/blob/master/meta-ostro/classes/whitelist.bbclass Mar 08 19:15:02 wesam: masking means you dont want to parse other recipes from the layer only a handful ones you pick Mar 08 19:15:58 i only need one recipe from meta-webserver but meta-webserver isnt in AGL by default Mar 08 19:16:03 what i should do? Mar 08 19:16:29 what i did was add it to my own layer which i also have a python script i am bitbaking in AGL Mar 08 19:17:34 wesam: I think you can simply add it to your bblayers.conf for start Mar 08 19:19:52 khem: the meta-webserver layer? Mar 08 19:20:40 i want to ship this layer out to people later down the road, should i just put in my README that you should add that layer to your bblayers.conf? Mar 08 19:21:44 yes Mar 08 19:22:18 if it changes behaviour of your distro or adds to parsing time etc. you might look into adding BBMASK Mar 08 19:22:26 for the recipes that are not needed Mar 08 19:28:38 khem: alright, that makes sense, thanks Mar 08 20:22:17 Hi, I am trying to add a patch to gdb, I have created the following file: Mar 08 20:22:19 cat gdb_%.bbapend asdasd SRC_URI += "0001-Do-arm_abi-detection-for-ELFOSABI_GNU-binaries.patch" Mar 08 20:23:45 Sorry I just realized the filename is missing a 'p' Mar 08 22:10:12 I have added a patch to gdb and now it is failing to build with this error: Can NOT get PRAUTO from remote PR service Mar 08 22:10:17 Any idea what could be wrong? Mar 08 22:10:43 This is the structure I created: Mar 08 22:10:48 ./gdb Mar 08 22:10:51 ./gdb/0001-Do-arm_abi-detection-for-ELFOSABI_GNU-binaries.patch Mar 08 22:10:52 ./gdb_7.9.1.bbappend Mar 08 22:11:08 And these are contents of my .bbappend Mar 08 22:11:12 FILESEXTRAPATHS_prepend := "${THISDIR}/${BPN}:" SRC_URI += "file://0001-Do-arm_abi-detection-for-ELFOSABI_GNU-binaries.patch" Mar 08 23:02:56 khem: I've done what I can for parsing speed Mar 08 23:17:34 RP:yes what do you think of https://github.com/ostroproject/meta-ostro/blob/master/meta-ostro/classes/whitelist.bbclass Mar 08 23:20:55 RP: btw. go is working on musl fine here, which arch did you see issues Mar 08 23:53:02 khem: we discussed whitelist a while ago and decided against merging to core Mar 08 23:53:57 khem: ross was testing go, I'm not sure where it is at to be honest, we probably need more test builds... Mar 09 00:53:46 RP: x32/ppc32 are unsupported backend Mar 09 00:54:02 for golang others that we care are now supported Mar 09 01:25:14 khem: thanks, I've updated -next and am running a new test set of builds **** ENDING LOGGING AT Thu Mar 09 03:00:02 2017