**** BEGIN LOGGING AT Fri Oct 23 10:59:57 2020 Oct 23 12:20:43 g0hl1n: btw you're maintaining the meta-java / Richard ? :) Oct 23 13:36:47 marex, correct ;-) Oct 23 15:20:08 g0hl1n: oh, yeah, cool Oct 23 15:20:32 g0hl1n: thanks for testing that LAN8710 PHY patch, it seems that is a particularly crappy PHY Oct 23 16:15:48 marex, yeah, I had problems too with it on our first PCB... ah, now your nick makes sense, so it's you Marek I guess? ;-) Oct 23 16:19:54 marex, so do you plan/have time for investigating the BB_NUMBER_THREADS/PARALLEL_MAKE issue? ... regarding the zeus state: do you plan to stay on zeus? Should I investigate in backporting to it? Oct 23 16:49:52 g0hl1n: so part one is this -- sed -i "s@BB_NUMBER_THREADS@PARALLEL_MAKE@" recipes-core/openjdk/openjdk-8-cross.inc recipes-core/ecj/libecj-bootstrap.inc ; git add -u ; git commit -sm "..." Oct 23 16:50:34 g0hl1n: part two, do_install[vardepsexclude] = "PARALLEL_MAKE" ... I guess for this case, it should be OK Oct 23 16:51:09 g0hl1n: and yes, I am stuck with non-LTS already-almost-EOL zeus for reasons ( = xilinx ) Oct 23 17:35:23 marex: if it makes you feel any better, I'm currently building an older AGL release based on thud to debug something Oct 23 17:50:20 I'm not sure if this is a good idea .... https://lwn.net/SubscriberLink/834682/a0a0a23ce2bfc34d/ Oct 23 18:07:04 halstead: patchwork is awol again Oct 23 18:09:13 Thanks khem. Looks like this will be an ongoing issue. I'm not sure what causes the bouncing. Oct 23 18:09:45 Crofton|cloud: 2.70 does, unfortunately, include a few places where backward compatibility with 2.69 could not be preserved. Oct 23 18:09:52 that will be PITA Oct 23 18:31:52 * moto-timo happily stops using java Oct 23 20:12:33 Anyone know of any bitbake memory usage issues that are fixed in master but not dunfell, by chance? I'm getting a lot of oom conditions lately Oct 23 20:13:50 it seems like we need more ram per bitbake thread than we used to Oct 23 20:13:55 anyone else seen this sort of behavior? Oct 23 20:14:27 I don't have any hard numbers, just seems this way Oct 23 20:42:49 huh, didn't know about expanded_urdata, will have to use that when we upgrade past dunfell finally Oct 23 21:12:14 Crofton|cloud: is there anything better than autoconf ? Oct 23 21:23:09 i need to play with meson more Oct 23 21:23:24 marex: all of them suck equally, actually we move from one disaster to another when it comes to build systems Oct 23 21:23:44 kergoth: yeah its a new recipe for disaster :) Oct 23 21:23:56 hah Oct 23 21:24:41 problem is at some point component build logic gets complex and none of these build systems cope with it well Oct 23 21:25:05 they are not designed using some language paradigm and semantics behind it Oct 23 21:25:32 its more iterative, it does one thing better than previous one but not a whole lot Oct 23 21:30:36 meson is certainly an improvement for me tooling support and community would be key and I think CMake has that in good shape, many IDEs support it out of box etc. meson is kind of same as CMake barring syntactic preference Oct 23 21:34:23 khem: so, why don't we use bitbake instead of all of that ? Oct 23 21:35:10 :-) Oct 23 21:35:35 bitbake isn't designed for that, it's designed to call into something that operates at the file level Oct 23 21:36:01 kergoth: shell scripts ? Oct 24 01:38:56 marex: my suggestion is pick your poison and deal with it Oct 24 01:39:06 none of them are stellar Oct 24 10:46:58 khem: yep **** ENDING LOGGING AT Sat Oct 24 10:59:57 2020