**** BEGIN LOGGING AT Fri May 26 03:00:04 2017 May 26 05:59:29 I have tried the the latest poky from git, I found the populate_sdk won't include the pc files of the pkgconfig anymore May 26 05:59:42 should I use the Extensible SDK? May 26 08:01:38 Anyone has any clues: i have a project which recently migrated to morty, in the process the build broke in ci which explicitly uses source mirroring from a https host which also required authentication.. SOURCE_MIRROR_URL was set to "https://user:pass@host/path/" format, which failed in do_fetch phase. After getting rid of user:pass, fetching works again. May 26 09:44:40 hi all May 26 09:49:26 Hi. If i have an autotools based recipe, is it possible to install aditional files along with the ones installed by autotools? (i have an autotools application and i want to add a systemd service for it, but when i add a do_install_append task the QA complains that i have installed files not in any package) May 26 09:50:16 which means you need to set FILES_${PN} or whatever package you want the files to be in May 26 09:50:26 also note there is a systemd class which helps with systemd bit May 26 09:52:08 can i append files to FILES_${PN}? (i assume that autotools adds files to FILES_${PN} and that when i try to add my files to it i overwrite the variable) May 26 09:52:17 yes, use += May 26 09:52:23 Ok May 26 09:52:27 Will try that now May 26 09:52:28 if you use = then yes, you'll be overwriting May 26 09:52:29 Thanks! May 26 09:52:38 i suggest you read the docs on how recipe syntax works May 26 10:07:56 rburton: is my impression correct that OE-core's meta/class/sanity.bbclass oecore_update_bblayers() contains code that is only applicable to the Poky distribution? May 26 10:10:21 Hmm, not quite. It is specific to OE-core's meta/conf/bblayers.conf.sample. May 26 10:11:08 But the end result is the same: the class doesn't do the right thing for other distros which happen to reach lconf >= 4. May 26 10:11:31 For example, in refkit I now get a "Your conf/bblayers.conf has been automatically updated." although the necessary changes have not been made at all. May 26 10:11:46 How have other distros dealt with that? May 26 10:11:53 Surely we are not the first to hit this? May 26 10:14:41 hm May 26 10:14:51 maybe other distros haven't been using layer versions May 26 10:18:47 rburton: the code which runs after the initial version comparison must be taken from the same place as local.conf.sample. Any suggestions on how to do that? May 26 10:19:03 hm May 26 10:19:14 nothing simple comes to mind May 26 10:19:39 Perhaps via local.conf.sample + INHERIT. May 26 10:19:54 And the in the inherited class provide the additional code. May 26 10:20:53 yeah that should work May 26 10:24:42 I just suspect that it is too late for refkit: current build configurations will be handled incorrectly because local.conf won't have such an INHERIT. May 26 10:26:27 is there a place in the mega-manual where the syntax for variables like this: May 26 10:26:29 UBOOT_CONFIG ??= "emmc sdcard" May 26 10:26:29 UBOOT_CONFIG[emmc] = "myBoard_emmc_defconfig,,u-boot-with-spl-pbl.bin" May 26 10:26:29 UBOOT_CONFIG[sdcard] = "myBoard_sdcard_config,,u-boot-with-spl-pbl.bin" May 26 10:26:35 is explained ? May 26 10:26:56 i don't really understand the list like syntax and how it impacts the build system May 26 10:27:22 I'm trying to separatelly build U-boot without bitbake May 26 10:27:48 and try to fing out all dependencies May 26 10:28:17 in other words, do a reverse engineering of the build process May 26 10:28:31 thats specific to the uboot recipe so you'll want to read that. looks like its doing the same sort of thing as PACKAGECONFIG, ie UBOOT_CONFIG selects options (emmc and sdcard) and then the values in [emmc] and [sdcard] are used May 26 10:28:39 the recipe is the only place that will tell you what those values actually mean May 26 10:29:00 i'll have a look on it May 26 10:35:08 is there a way to find the location of a recipe exept doing a find in the location indicated by BBLAYERS? May 26 10:35:32 e.g. i want to see where a.bb is located May 26 10:36:03 from within bitbake doing something like bitbake -e a | grep ^somevar= May 26 10:36:52 ... if this is the right way May 26 10:37:01 $ bitbake-layers show-recipes u-boot -f May 26 10:37:01 === Matching recipes: === May 26 10:37:01 /home/ross/Yocto/poky/meta/recipes-bsp/u-boot/u-boot_2017.01.bb May 26 10:37:23 (-f is —filenames) May 26 10:37:26 nice ... forgot about this one May 26 10:37:43 thanks Ross May 26 10:44:48 anyone here done change requests from gerrit to yocto builds ? Top level layers are easy as i can just do a checkout just before the build but for recipes which do git checkouts its a bit more troublesome .. May 26 11:56:17 rburton: do you understand how the oecore_update_bblayers() code is meant to work? Does it get called repeatedly? For example, the code for "if current_lconf == 4 and lconf_version > 4" just does one change and bumps from 4 to 5, then returns. May 26 11:58:05 Also, is that code really still working? The code for "current_lconf == 5 and lconf_version > 5" calls "status.addresult()" but "status" is undefined in the oecore_update_bblayers() function. May 26 11:58:36 Looks like the code was copied from elsewhere - check_sanity() perhaps. May 26 13:02:42 hello guys, is it correct recipe http://ix.io/v8P if I use several repositories in one repo? What value SRCREV should contain in this case? May 26 13:04:28 Ox4: what do you mean by several repositories in one repo? May 26 13:04:52 Ox4: you should use only use just one repo per recipe May 26 13:05:12 Ox4: in case of many, not sure which will pick.... May 26 13:05:35 in one recipe* May 26 13:05:37 sorry May 26 13:06:01 the SRC_URI contains several git repositories May 26 13:06:35 and I want to use particular commit for each source code May 26 13:06:50 Ox4: but why do you want several repositories? one recipe one source tree, that is the common situation May 26 13:07:37 Ox4: create multiple recipes, each with one pointing to a specific repo and commit ID, that is the safest May 26 13:08:15 lsandov: because this application has dependencies May 26 13:08:19 Ox4: name them and you can set srcrev per-repo May 26 13:08:30 hm, ok May 26 13:08:31 thank you May 26 13:08:55 see cross-localedef-native.bb May 26 13:13:33 ed2: around? May 26 13:13:51 rburton: but in that case, recipe is just changing the SRCREV but the SRC_URI remains static May 26 13:14:28 well setting a single srcrev in a recipe with many git clones is meaningless May 26 13:14:42 the rev= bits can be removed and replaced with a series of SRCREV_name assignments May 26 13:15:03 Ox4 may then want to look up SRCREV_FORMAT to control what happens to the version May 26 13:26:53 And make sure your values for name are prefix-free w.r.t. to each other if you use SRCREV_FORMAT. Don't try app_a and app_a_libs together in SRCREV_FORMAT... May 26 13:27:05 (Not sure if that was fixed meanwhile, it was broken a while ago) May 26 13:48:42 rburton: what do you mean? May 26 14:17:16 Is it OK to have non-alphanumeic characters in a recipe name (i.e. "dbus-c++")? May 26 14:17:49 archive.bbclass fails in this specifc example, but I wonder if that is a bug or expected May 26 14:19:43 JPEWhacker: No, + in specific doesn't work due to variants May 26 14:20:16 Errr, nevermind, wrong packaging system. May 26 14:21:40 JPEWhacker: Pretty sure ++ in recipe names work, we have common-api-c++.bb common-api-c++-dbus.bb May 26 14:22:16 Ok, thanks. I will submit a patch. May 26 14:30:39 neverpanic: can I skip SRCREV_FORMAT at all? May 26 14:34:48 Probably, if you have one default-named source, in which case that default source will be used. May 26 14:35:04 There may be a sanity check that forces you to set it, though. May 26 14:35:09 I would recommend setting it. May 26 15:05:14 neverpanic: what value should be set? May 26 15:07:02 Ox4: That depends on what you want the version to look like. For SRC_URI = "url1;name=name1 url2;name=name2" SRCREV_FORMAT="name1_name2" and SRCREV_name1 = "abc", SRCREV_name2 = "def", it will be "abc_def" May 26 15:07:40 If there's a hierarchy between your different sources, sort in order of descending importance May 26 15:08:17 neverpanic: understood, thank you May 26 15:22:47 hm, I have got an error :-( May 26 15:26:35 http://ix.io/v9w May 26 15:29:10 Marex: can you resend the u-boot openssl patch with all the correct headers? missing signedoff and upstream-status. May 26 15:30:07 (your sob in the patch header, and they're either Submitted or Accepted depending on whether you grabbed them from the list or if they're in the repo already) May 26 15:59:15 rburton: sure, I plan to do that May 26 15:59:23 rburton: I saw your email, I'll process that in bulk May 26 15:59:32 rburton: can you at least apply the u-boot 2017.05 and qemu updates ? May 26 16:01:02 yeah they're all in mut now May 26 16:01:14 going to hit the AB shortly May 26 16:01:28 rburton: super, I'll wait for them to be pushed, then rebuild and retest May 26 16:02:12 cool May 26 16:15:34 rburton: sent couple of patches one of them is for glibc May 26 16:15:41 in preparation for glibc 2.26 May 26 16:28:09 khem: bets on how much the hardening will break stuff? May 26 16:28:23 ie is it worth running that on the AB standalone to isolate any breakage May 26 17:51:28 we should really make bitbake's better_compile/better_exec actually be better May 26 17:51:43 when the failure is in python, it's inferior to the default exception traceback. doesn't show the failed line, doesn't show the column it failed on, etc May 26 17:52:02 RP1: ^ May 26 17:52:27 had to add a logger.exception() to actually find the cause of my exception May 26 18:24:18 is it possible to build the esdk with meta-qt5 ?! May 26 20:43:24 kergoth: I agree fwiw May 26 22:49:26 where do I specify the IMAGE_FEATURES ? what file does that go in? it says it can fo local.conf but that it's a bad idea. WHere does it go? May 26 23:03:45 I search for where to set IMAGE_INSTALL and IMAGE_FEATURES, it only tells me where NOT to set them May 26 23:03:53 i would like to know where to set them May 26 23:04:40 just create your own image May 27 01:00:58 not a bad idea for testing/developing **** ENDING LOGGING AT Sat May 27 03:00:01 2017