**** BEGIN LOGGING AT Wed Jul 29 02:59:58 2015 Jul 29 07:23:36 good morning Jul 29 07:24:48 morning mckoan Jul 29 07:38:24 hi nrossi Jul 29 08:11:33 morning all Jul 29 08:17:39 hi bluelightning Jul 29 09:28:57 asd Jul 29 09:30:00 howdy Jul 29 09:30:19 if you were to be releasing a new oe device... what version would you choose and why? Jul 29 09:30:54 We are on dizzy... is that too progressive or conservative ? Jul 29 09:33:39 I cannot seem to find any documentation to help gauge what is supposed to be stable and what is bleeding... Jul 29 09:34:02 pompomJuice: all stable releases should be considered stable; the only thing that's bleeding edge is master Jul 29 09:34:18 (and of course contrib branches) Jul 29 09:35:02 the answer is it depends... dizzy is fine, but it may be worth bearing in mind that in October dizzy will no longer be officially supported (since we support two releases back, and release every six months) Jul 29 09:37:39 oh Jul 29 09:37:43 That's interesting Jul 29 09:38:32 Because we have a freescale product with a dodgy network driver or something and the only way to up the kernel is to up the entire thing. Jul 29 09:39:36 How are companies supposed to maintain their product? Do they need one guy that just keeps the thing updated every 6 months or so? Is that what it takes? Jul 29 09:40:06 I think so Jul 29 09:41:17 pompomJuice: i think the term you are after is called "Product Maintainence", after all thats what you put on the invoice ;) Jul 29 09:42:30 But the boss guys paying the big bucks want to finish the thing off and get rid of all the incomprehensible developers saying that it needs maintenance. They even go as far as to request that I make bitbake work without any internet connection. I said that makes no sense but they refuse to understand my POV. What can I do? Jul 29 09:44:45 They do not understand open source at all, all they know is Microsoft. How do I tackle this problem? Jul 29 09:44:47 leave the sinking ship and let everything go kaboom behind you. Jul 29 09:44:57 pompomJuice: you can make bitbake work without network just fine... ensure you have all the sources in a local mirror, no SRCREV = "${AUTOREV}", and set BB_NO_NETWORK = "1" Jul 29 09:45:05 pompomJuice: job done Jul 29 09:45:33 and as long as the source/downloads folder is protect that should work forever? Jul 29 09:45:43 pompomJuice: or the local mirror, yes Jul 29 09:45:58 ok awesome Jul 29 09:46:37 pompomJuice: careful though, if your users want to run the build on newer hosts it might not work :| Jul 29 09:46:58 right, in the absence of host changes Jul 29 09:47:06 yea I would make a backup of the build host aswell Jul 29 09:47:17 turn off auto updates Jul 29 09:47:30 support for newer hosts, security and bugfixes is what we do for supported stable releases Jul 29 09:47:32 one trick which works well is to just ship your development platform as a virtual machine/appliance :) Jul 29 09:47:36 I would even go as far as to put the fridge at 10 Jul 29 09:51:03 bluelightning, what is this local mirror you speak of. How do you make one? Jul 29 09:52:23 pompomJuice: http://www.crashcourse.ca/wiki/index.php/OE_Creating_your_%22own_mirror%22 Jul 29 09:52:37 thanks Jul 29 09:52:44 pompomJuice: you can get away with not doing that, but basically it's some location containing the source files that you point to with PREMIRRORS (or own-mirrors and SOURCE_MIRROR_URL) Jul 29 09:53:13 ok I see Jul 29 09:53:24 so its not some win button I can press... I will research Jul 29 09:53:45 thanks Jul 29 10:54:04 hi, do you remember if in oe-classic anything was needed to be able to INHERIT += from a recipe? Jul 29 10:56:11 you wouldn't use INHERIT += in a recipe, you'd use inherit Jul 29 11:01:22 bluelightning: I need to make it conditional :-/ Jul 29 11:03:11 mnemoc: in OE-Classic there is no means to do a conditional inherit from within a recipe, that's something we added much later in a version of bitbake that almost certainly won't work with OE-Classic Jul 29 11:03:35 another reason to consider moving off it... Jul 29 11:05:37 I wish I could but I'm stuck with a junk-kernel/bootloader that only compiles with gcc 4.3 :| and so use an ancient arago Jul 29 11:06:10 in modern oe, is it possible to use a different toolchain for linux/u-boot than for the rest of the system? Jul 29 11:08:02 2.6.32.17-davinci :'( Jul 29 11:48:08 reminds me time of 2.4 kernels... Jul 29 11:48:39 mnemoc: no KERNEL_CC? Jul 29 11:48:55 * hrw lacks current OE knowledge Jul 29 11:49:13 * mnemoc greps Jul 29 11:50:01 yes! Jul 29 11:52:15 mnemoc: or backport gcc4 kernel patches? Jul 29 11:54:39 hrw: I tried that but the resulting binary didn't even earlyprintk :| Jul 29 11:54:48 ;( Jul 29 11:55:19 mnemoc: TI fired omap and davinci teams and there is no one working on code supporT? Jul 29 11:57:28 afaiwt the soc not supported anymore Jul 29 11:59:50 but not first hand knowledge, from the person who has been working on this product for the last N^many years Jul 29 12:01:17 I'll play with KERNEL_CC and see how far can I get with "fido" Jul 29 13:14:52 how many of you got 'interesting offer' for 'yocto specialist' in Dusseldorf, Germany during last few months? Jul 29 13:16:05 hrw: you mean "Yocto Consultant "? Jul 29 13:16:16 ndec: yep Jul 29 13:16:26 obviously i'm doing something wrong as i didn't! Jul 29 13:16:40 hehe Jul 29 13:17:38 me neither :.( Jul 29 13:18:05 it's probably because Yocto is an ARM thing.. intel profiles don't look that interesting ;-) Jul 29 13:18:20 heh Jul 29 13:19:14 I did get an email the other day saying "based on your extensive experience we'd like to discuss with you an opportunity..." and then it went on to describe an android app developer role; I've never written an android app Jul 29 13:22:28 bluelightning: but you still have an extensive experience. that part was correct ! Jul 29 13:22:41 it just is not and extensive experience on android.. Jul 29 13:23:03 ndec: yeah... it confirmed my opinion of recruitment agents at least ;) Jul 29 13:23:30 ndec: I remember time when my OH paycheck was paid by Intel ;d Jul 29 13:31:09 hrw: those head hunters are definitely annoying Jul 29 13:31:49 I wonder whether they know what they are writing Jul 29 13:32:30 mckoan: doubt it... :P i've seen some recruitment agencies make "videos" for positions Jul 29 13:35:15 BTW best proposals I've received were direct from company's HR Jul 29 13:39:36 mckoan: before headhunters modifications ;D Jul 29 13:39:55 I currently cant find it, but there is a switch somewhere to let OE produce an image with all debug symbols, right? Jul 29 13:40:08 I am aware that the image could then potentially get very large Jul 29 13:42:10 dv_: dbg-pkgs in IMAGE_FEATURES or EXTRA_IMAGE_FEATURES Jul 29 13:43:30 right, thanks Jul 29 15:21:10 is there any specific reason why oe/master is on mesa 10.5 series, not 10.6? or would it be accepted to switch over? Jul 29 20:06:23 wtf, i have LOCALE_GENERATION_WITH_CROSS-LOCALEDEF set to 0, and its running cross-localedef Jul 29 20:06:29 * kergoth tries the empty string instead Jul 29 20:08:30 if the empty string doesn't work, try LOCALE_GENERATION_WITH_CROSS-LOCALEDEF = "NO, PLEASE" Jul 29 20:11:33 afaict all its references explicitly check for == 1, which is why i figured 0 would work, but i must be missing something.. Jul 29 20:12:13 getting localedef failures most likely due to the glibc version mismatch between cross-localedef and the external toolchain glibc, still investigating, though Jul 29 20:18:00 gah **** ENDING LOGGING AT Thu Jul 30 02:59:58 2015