**** BEGIN LOGGING AT Tue Apr 18 03:00:01 2017 Apr 18 06:42:41 good morning Apr 18 08:07:02 morning Apr 18 08:08:09 Jin^eLD: good morning Apr 18 11:52:45 khem: ping Apr 18 12:00:05 khem: nevermind, misread the date of a commit you made. Apr 18 17:04:46 what would be the easiest way to implement something like bb.utils.onlycontains() ? Apr 18 18:06:03 in bitbake, what is "d"? is it possible to do something like d.count() or d.sizeof() ? Apr 18 18:08:12 tlwoerner: d holds all the metadata. functions, variables, everything Apr 18 18:08:18 count of what? size of what? Apr 18 18:09:21 i'd like to find a way to figure out if a variable contains a string and only that one item Apr 18 18:09:40 first, a variable *is* a string Apr 18 18:09:53 if it's conceptually used as a list, then you'd have to first split the string into a list Apr 18 18:10:01 tlwoerner: you are the guy who split the chromium recipes into x11 and wayland, right? Apr 18 18:10:35 kergoth: somethine like bb.utils.containonly('GALLIUMDRIVERS', 'swrast' ...) Apr 18 18:10:44 dv_: yes Apr 18 18:11:01 ${@'foo' if '${GALLIUMDRIVERS}'.strip() == 'swrast' else 'bar'} Apr 18 18:11:03 done Apr 18 18:11:20 :) Apr 18 18:11:37 "contains only this" is basically "is equal to this, other than whitespace" Apr 18 18:12:12 tlwoerner: have you been watching the ozone-integration progress? Apr 18 18:12:39 dv_: haha, no, that's why i split them out so i could follow just the x11 stuff :-) Apr 18 18:18:46 kergoth: hmm bitbake didn't like that :-( https://pastebin.com/2LVTLrg2 Apr 18 18:20:32 i only want to add the --enable-texture-float flag to the mesa build if building a gallium-enabled driver that has hardware acceleration Apr 18 18:22:49 tlwoerner: first of all, you didnt' put an else, so if it's not true, it'll put "None" in your packageconfig Apr 18 18:22:56 second, you didn't actually say what "didn't like that" is supposed to mean Apr 18 18:23:01 what error? Apr 18 18:23:13 'foo' if bar else '' would be what you want Apr 18 18:23:20 most likely, anyway Apr 18 18:23:51 https://pastebin.com/TAnLWPHD Apr 18 18:24:36 start by adding the missing else clause Apr 18 18:25:37 kergoth: \O/ Apr 18 18:26:08 it passed parsing, just seeing how the ./configure goes... Apr 18 18:28:30 kergoth: oops, i got the logic wrong, i *want* the flag in the else case and not the true case :-S Apr 18 18:29:24 tlwoerner: hm well chromium itself is getting ozone support, ozone-wayland seems to be deprecated Apr 18 18:29:45 so eventually we'll be able to merge it back together again & simplify the recipes Apr 18 18:30:58 dv_: chromium needs more third-party libraries as part of its source like it needs.... Apr 18 18:31:10 ? Apr 18 18:31:34 I mean that ozone integration is now part of the chromium codebase Apr 18 18:31:43 not sure if it is feature complete yet Apr 18 18:31:53 dv_: the chromium sources are massive enough already because everything is "in house" Apr 18 18:32:24 dv_: they're all collected under the "third-party" subdirectory Apr 18 18:33:29 yeah Apr 18 18:33:30 dv_: i'm just trolling a little, it's nice to hear wayland will be better supported Apr 18 18:33:45 but at least ozone stuff does not have be patched-in, which was painful to get to work Apr 18 18:33:52 dv_: agreed Apr 18 18:34:08 and yeah, the chromium sources are ridiculously massive Apr 18 18:34:25 which is why i left it behind, i'm glad others stepped up and have been maintaining chromium-wayland Apr 18 18:35:29 kergoth: you are awesome! thanks so much :-D Apr 18 18:35:45 np, glad it worked for you Apr 18 18:36:09 the non-lite tarball for chromium 59 : 519.6M the lite tarball: 413.2M Apr 18 18:37:40 now OE users can get opengles 3.0 without thinking about it on platforms where patents aren't a problem instead of being stuck at 2.0 Apr 18 18:38:36 patents? Apr 18 18:40:45 dv_: apparently so Apr 18 18:41:57 you mean patents are holding back gles3.0? Apr 18 18:43:30 dv_: yes. according to rob clark "without texgture-float you can't expose some extension which became required somewhere along the way to gl3/gles3" Apr 18 18:44:05 dv_: "but that really only applies to sw drivers (since it is all handled in hw for real gpus, and the gpu vendor has already paid the license fee)" Apr 18 21:00:51 so nios2 now has this funny option , where one can pass -march=1 or -march=2 to select between nios2 r1 and r2 , which are binary incompatible Apr 18 21:01:05 should I introduce new architecture (nios2r2) or should that r1/r2 be tunables ? Apr 18 21:01:31 I cannot run nios2r1 on nios2r2 or vice-versa because they changed the instruction encoding Apr 18 21:34:28 anyone ? Apr 18 21:34:56 what is your question ? Apr 18 21:36:46 khem: see above ... Apr 18 21:37:53 * rburton is curious to the answer as he has never studied machine/tune interaction properly Apr 18 21:41:17 rburton: yeah, that, pretty much Apr 18 21:41:42 rburton: well while I still have you here, should we add the nios2 tunes etc to oe-core or keep them in meta-altera ? Apr 18 21:42:01 * Marex always hits some obscure crap, sigh Apr 18 21:47:51 Marex: there is no above for someone who might have joined just now Apr 18 21:48:32 23:06 < Marex> so nios2 now has this funny option , where one can pass -march=1 or -march=2 to select between nios2 r1 and r2 , which are binary incompatible Apr 18 21:48:35 23:07 < Marex> should I introduce new architecture (nios2r2) or should that r1/r2 be tunables ? Apr 18 21:48:38 23:07 < Marex> I cannot run nios2r1 on nios2r2 or vice-versa because they changed the instruction encoding Apr 18 21:48:41 khem: now there is , see above Apr 18 22:01:49 r1 and r2 should be tunables Apr 18 22:02:14 mips1 and mips2 are not compatible isa either but we control that with tunable Apr 18 22:07:59 khem: why tunables and not different arch ? Apr 18 22:08:05 khem: what about mips r5 and r6 ? Apr 18 22:38:54 so ... ? Apr 18 22:44:48 what do you intend is the loss Apr 18 22:48:58 khem: errr, what ? Apr 18 22:50:19 tunable vs arch Apr 18 22:50:50 r5 and r6 have different ABIs Apr 18 22:53:14 r5 and r6 have different instruction encoding too Apr 18 22:53:28 so what's the point of arch then ? Apr 18 22:54:07 and when does one use arch and when tunable ? Apr 18 22:54:22 is this documented somewhere or is that just some ad-hoc decision of some ad-hoc OE higher-up ? Apr 19 00:43:27 Marex: if you want to share feeds among closely designed sub families of arch Apr 19 00:43:33 use tunable Apr 19 01:07:41 * tlwoerner welcomes Marex to Wonderland Apr 19 01:15:43 its not all wonderland some of it is mystery too Apr 19 01:17:16 khem: i was referring to the Wonderland from "Alice in Wonderland" ;-) Apr 19 01:30:05 doing OE at times I also feel like being a kid in a nude bar, I agree **** ENDING LOGGING AT Wed Apr 19 03:00:03 2017