**** BEGIN LOGGING AT Thu May 04 03:00:03 2017 May 04 07:35:21 good morning May 04 08:01:48 Morning May 04 08:15:23 (UGT) May 04 13:33:27 ~ugt May 04 13:33:28 ugt is, like, Universal Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. http://www.total-knowledge.com/~ilya/mips/ugt.html May 04 13:33:40 ~botsnack May 04 13:33:40 hrw: :) May 04 13:39:40 any opinions if AWS or Google cloud is better for a OE build machine? I've been using AWS for years now, and seems to work pretty well. Can use a powerful machine (32-core) and its still cost effective if powered off most of the time (build duty cycle is typically low). So wondering if anyone has experience with google cloud? May 04 14:00:15 Hi, I would like to generate a partition ext4 with only some specific files into. This partition will flash on mtd device and mount at boot (fstab). I don't know how to do that. Could you help me ? May 04 14:00:16 I try to write a recipe, with "inherit image", but I have some filesystem folder, and I don't want them. May 04 14:00:16 Thank you :) May 04 14:02:31 cbrake: What is your time gain if you build on aws ? May 04 14:02:48 vs building on a local i7 May 04 14:03:36 MarcWe: I think its about a wash since the IO (or something) is slower. So its more a convenience of being able to build in cloud. May 04 14:04:44 MarcWe: with one project, customer had a huge C++ app that took 30 minutes to build locally on i7, so throwing 32 cores at it in cloud sped that up some May 04 14:05:33 MarcWe: but in general I've not found AWS to be significantly faster over local. I've not benched in last year or two, so things may have changed May 04 14:07:19 ok ty :) May 04 15:10:24 hm darn, my system image is now 1.5 the size due to python __pycache__ being pulled in by default since May 04 15:10:27 *3f2ab3f7cc | Dominic Sacré | python3: Build and package precompiled modules May 04 15:11:04 sigh May 04 15:15:21 Marex: they'd have been built anyway just on demand May 04 15:17:49 I'm not personnally using it but I have a few customers complaining that the environment-setup script from the SDK doesn't make menuconfig work (basically because it completely masks the natvie compiler) May 04 15:18:24 is this something that someone really cares about? :) May 04 15:18:39 hmm May 04 15:18:48 I'm pretty sure I've done this in the past May 04 15:19:02 menuconfig for building kernel? May 04 15:19:13 rburton: yep, except I can no longer fit my rootfs into my device ;-) May 04 15:19:45 Marex, get a bigger device May 04 15:25:25 Crofton|work: finally a helpful advice, just what I was looking for May 04 15:25:43 :) May 04 15:25:58 I try to help May 04 15:27:27 Crofton|work: you rock May 04 15:37:12 Thanks Marex so do you! Except for the Altera bit May 04 15:39:53 Crofton|work: lol :) May 04 16:51:06 tlwoerner: thanks for the ascii art, what have you used to generate the ascii graphs? May 04 17:30:02 JaMa: wavedrom maybe ? May 04 19:26:30 JaMa: gnuplot, i'll email you the instructions May 04 20:54:13 anyone any idea on how to avoid the cd ${S} in https://pastebin.com/JN2nSTEj May 04 20:54:32 currently I think ends up building in ${S} rather than ${B} May 04 20:56:04 fischerm: is the code can't be built in ${B} and needs to be built in ${S} then replace "inherit autotools" with "inherit autotools-brokensep" May 04 20:56:11 fischerm: s/is/if May 04 20:56:21 oh May 04 20:56:25 tlwoerner: awesome May 04 20:56:45 fischerm: or send a patch upstream ;-) May 04 20:57:05 either way, the "cd ${S}" could then be removed May 04 21:18:12 tlwoerner: thx May 04 21:27:02 fischerm: the proper fix is to fix upstream so that 'make distcheck' works May 04 21:27:18 as any fails are bugs in upstream makefile.am May 04 21:39:06 rburton, thanks ... **** ENDING LOGGING AT Fri May 05 03:00:01 2017