**** BEGIN LOGGING AT Sat Aug 03 03:00:00 2013 Aug 03 07:27:36 what language is webhob written in? Aug 03 07:55:11 lpapp: python/django iirc Aug 03 07:57:34 but is it actually handy that end users will need to run a local webserver? Aug 03 07:58:58 lpapp: that is easy enough to do from python Aug 03 08:01:45 well, I have spent 3 full working days to do. Aug 03 08:02:11 it would be a pity if end users had to go through the same trouble... Aug 03 08:02:23 this brings me to the idea of a Qt frontend, really. Aug 03 08:02:34 which you just run and it works. Aug 03 08:05:00 my buildservers don't have any Qt or x11 installed, local webserver sounds like less bloat to me Aug 03 08:05:17 what does it mean meta-oe layer? Aug 03 08:05:32 except that many people do not use buildservers. Aug 03 08:05:49 lpapp: and many people do Aug 03 08:05:50 well, the most people do not use to be fair. Aug 03 08:06:12 lpapp: how many users have you talked to? Aug 03 08:06:16 * JaMa is using minimalistic chroot even on his desktop machine Aug 03 08:06:36 actually, they could already use ssh. Aug 03 08:06:45 since server guys are used to it, anyway. Aug 03 08:07:07 a frontend is a means for end users on local PC in my book. Aug 03 08:07:17 gtk is non-crossplatform. Aug 03 08:07:22 webserver is a hassle for such users Aug 03 08:07:26 you cannot run qt frontend in GNU screen on server :) Aug 03 08:07:40 and for those people, I believe, qt frontend I will work on, is the best solution. Aug 03 08:07:56 JaMa: have you read what I wrote? Aug 03 08:08:00 1) yes, you can. Aug 03 08:08:06 2) I mentioned ssh. Aug 03 08:08:16 no intentionally I'm not reading anything you're writting Aug 03 08:08:16 lpapp: Lets just say that there has been quite a lot of work gone into figuring out what the userbase actually wants/needs through surveys, interviews and so on Aug 03 08:08:35 JaMa: then intentionally do not reply. :) Aug 03 08:08:55 RP: let us just say, many people use local PCs. Aug 03 08:09:22 have you seen any "lpapp:" in my messages? I'm not talking to you, I'm just supporting RP's view Aug 03 08:09:22 lpapp: Have you read what *I* wrote? ;-) Aug 03 08:09:25 and for those, as I already mentioned, qt seems to be the best solution. Aug 03 08:09:35 RP: yes Aug 03 08:10:13 * lpapp really fails to see why qt could not be run over screen Aug 03 08:10:18 that is just technically incorrect. Aug 03 08:10:22 lpapp: for you, qt is perhaps seemingly the best solution, for the whole user base it was not found to be Aug 03 08:10:42 RP: err... no. Aug 03 08:10:42 lpapp: we did actual research for that and there are details in the mailing list archives Aug 03 08:10:56 RP: as I already said, I was primarily talking about local PC builds. Aug 03 08:10:59 lpapp: our surveys found something different? Aug 03 08:11:14 * JaMa off Aug 03 08:11:24 lpapp: You're suggesting I don't understand the survey results? :) Aug 03 08:11:33 RP: yes Aug 03 08:11:44 there is an article somewhere how bad surveys are. Aug 03 08:11:50 and why they are misleading. Aug 03 08:12:04 also, show me the survey, please. Aug 03 08:12:16 and the mailing list thread. Aug 03 08:12:20 lpapp: its all in the mailing list archives Aug 03 08:12:36 show or did not happen. :) Aug 03 08:13:01 lpapp: I'm not interested in being told I can't read or understand things ;-) Aug 03 08:13:25 RP: oh, in that case "we also made a big survey, we just cannot prove, blabla" Aug 03 08:13:54 it is getting unauthentic without proving when being asked for details. Aug 03 08:14:27 lpapp: I'm just not interested in taking this discussion further since I think its pointless Aug 03 08:14:47 right, so I just made a survey Aug 03 08:14:59 a few people any everyone would prefer qt. Aug 03 08:15:09 lpapp: congratulations. Survey sample of one? :) Aug 03 08:15:26 I am not a native speaker. Aug 03 08:15:33 but a few people definitely do not mean one in English. Aug 03 08:16:18 since you cannot back your statement, I am not in a position to conclude out of it. Aug 03 08:17:33 lpapp: https://wiki.yoctoproject.org/wiki/Web_Hob_design_project_with_T%26T has details of some of the work that went into it and links to user testing and so on Aug 03 08:18:24 show me the end result of the survey, and the survey text itself, please. Aug 03 08:18:31 1) mailing list Aug 03 08:18:35 2) a doc file Aug 03 08:26:35 why should anyone work on showing you that? Aug 03 08:26:52 you'll just flame on anyway Aug 03 08:30:40 Bagder: please be civil. Aug 03 08:32:58 lpapp: why are you so concerned that other people are writing a webhob? if you think a qt UI is best, you are free to write one. Aug 03 08:33:01 lpapp: end of story. Aug 03 08:33:22 rburton: I cannot be interested in the details of a claim? Aug 03 08:34:05 of course, I will write my Qt frontend, but unlike some other people I listen to others, and why they have that opinion. Aug 03 08:35:02 it is not like you can throw "random" worlds at the company either without backing your statement. Aug 03 08:36:17 it seems to be like the "overdiscussed gerrit", when I tried to look for information for quite a long while, and turned out I just wasted a lot of time since there was no such a discussion at all in public. Aug 03 08:36:36 sigh Aug 03 08:37:09 lpapp: if you'r really interested in teh webhob research, a minute with google should find it Aug 03 08:37:16 now i'm off to the library with my daughter Aug 03 08:40:43 yes, sure, I need to back others, and not them themselves, sigh. Aug 03 08:41:20 not to mention this has been told for the gerrit stuff as well for which I was looking for hours (!) without any result. Aug 03 08:41:38 ffs Aug 03 08:41:40 https://wiki.yoctoproject.org/wiki/images/9/9d/User_Test_1_Findings-INT-002.pdf Aug 03 08:41:47 top hit on google for "webhob user interviews" Aug 03 08:42:22 and linked from the page I mentioned Aug 03 08:42:30 * RP shrugs Aug 03 08:44:39 oh, so it is not a mailing list Aug 03 08:44:43 and I was looking on the mailing list. Aug 03 08:44:44 sigh Aug 03 08:44:49 so much for correct information .... Aug 03 08:46:20 also, it is a webhob user feedback, not generic ui preference survey! Aug 03 08:46:26 it is a vastly different thing. Aug 03 08:46:35 user interface* Aug 03 08:47:08 8 interviews ... :-) That is a small number, at large. Aug 03 08:47:45 and the original survey is nowhere. Aug 03 08:49:09 so, yes, I could do a Qt user interface feedback, and would probably get positive feedback from Qt users. Aug 03 08:50:11 anyway, it shows the original claim incorrect to me. Aug 03 08:52:01 so what does it mean meta-oe layer? Wasn't the meta layer supposed to be the oe-core layer? There is a non oe-core oe layer? Aug 03 09:16:39 lpapp: meta-oe = meta-openembedded Aug 03 09:16:56 lpapp: http://git.openembedded.org/meta-openembedded/ Aug 03 09:17:51 so non oe-core oe layer Aug 03 09:18:00 is it possible to put fltk in oe-core? Aug 03 09:18:10 lpapp: there were multiple stages and interviews in the webhob process. Those look like some of the later ones, I'm not sure offhand where the earlier ones are archives or if they're online Aug 03 09:19:16 lpapp: You haven't given me much incentive to go and find the information and I'm not sure it would make much difference to you even if I did. There was discussion at each stage on the mailing list about the process. Aug 03 09:19:51 the question would be why is fltk needed in oe-core and why is it not usable from meta-oe Aug 03 09:20:00 * RP -> afk Aug 03 09:20:02 because it is a core embedded graphics stuff? Aug 03 09:20:19 well, you can try to defend yourself by proving yourself would not make a difference for a technical discussion. Aug 03 09:20:27 it looks unreasonable to me, but it is your choice to take. Aug 03 09:21:43 even gnome is in oe Aug 03 09:21:47 and fltk is way less headless Aug 03 09:26:59 mailing list patch is the best way for such proposals? Aug 03 09:27:07 or a separate discussion thread? Aug 03 09:27:19 I would not like to clone yet another full layer just for one single package. Aug 03 09:27:40 and I think fltk is also common on embedded. Aug 03 09:27:52 moving gnome to separate layer was also proposed few times Aug 03 09:28:17 that would make sense as it is a monster Aug 03 09:28:21 fltk is not however. Aug 03 09:28:38 it is probably the most lightweight embedded opengl empowered ui toolkit out there I know. Aug 03 09:29:21 I would like to be able to build a simple ui for a headless board with oe-core, off-hand. Aug 03 09:29:37 just clone another layer, is that so hard? :) Aug 03 09:30:00 for one package to clone 100+? Aug 03 09:30:02 yes, it is so hard Aug 03 09:30:09 and so bloated. Aug 03 09:30:12 and cluttered. Aug 03 09:30:36 but if you include every recipe *you* want in oecore, that will get bloated as well Aug 03 09:30:53 huh? no? Aug 03 09:30:59 it's not like you have to use every recipe in meta-oe either Aug 03 09:31:12 in fact, if you read what I write, I proposed to get the monster out, i.e. gnome. Aug 03 09:31:40 fltk *is* a core ui toolkit for very headless boards. Aug 03 09:31:48 so it is a natural fit for oe-core to me. Aug 03 09:32:06 but of course, you are free to prove me wrong anytime and give an alternative. Aug 03 09:32:13 what to use on a core system. Aug 03 09:32:30 why would a headless board need a gui toolkit? Aug 03 09:32:59 because it has a small display? Aug 03 09:33:04 called "LCD". Aug 03 09:33:11 not the ones i work with :) Aug 03 09:33:23 oh, so if you do not work with it, it does not exist... Aug 03 09:33:28 fltk is also bloated, because different recipes need different fltk versions Aug 03 09:33:38 huh? no? Aug 03 09:33:52 actually, every distribution I know is using one version, and works fine. Aug 03 09:33:58 even meta-oe uses one. Aug 03 09:34:37 see oe-classic discussions about fltk, I know it's old Aug 03 09:35:20 please show a valid discussion. Aug 03 09:36:12 which is a problem _today_. Aug 03 09:45:32 what is "LEAD_SONAME" ? Aug 03 09:57:15 meta-openembedded patches do not go to the oe-core mailing list? Aug 03 10:04:41 what does '2' mean in the qmake2 class? Aug 03 10:06:37 is there a tutorial somewhere how to write class files? Aug 03 10:07:01 lpapp: LEAD_SONAME is used by debian.bbclass for package renaming. Its something we'll probably remove eventually Aug 03 10:07:22 yeah, it looks like a tremendeous hack. Aug 03 10:07:32 lpapp: meta-oe patches go to where the readme says they should go Aug 03 10:09:55 is there a tutorial somewhere how to write class files? Aug 03 10:14:47 I suspect the answer to the specific question is no Aug 03 10:37:44 is there a way to submit a change to a mailing list without subscribing and waiting for moderation? Aug 03 10:52:44 lpapp: if you subscribe you shouldn't hit the moderation Aug 03 11:03:25 lpapp: you can subscribe and turn off mail delivery Aug 03 11:25:11 I do not wanna subscribe. Aug 03 11:43:43 JaMa: how can I pass configure options to qt5? Aug 03 11:43:47 with your layer? Aug 03 11:44:06 to use the feature system for instance. Aug 03 11:44:59 PACKAGECONFIG Aug 03 11:45:52 in some cases it's not so simple (for options with only auto-detection) see qtmultimedia for example Aug 03 11:46:25 JaMa: have you seen my patch btw? Aug 03 11:47:21 yes, it was missing [meta-qt5] prefix in subject but otherwise looks OK Aug 03 11:47:39 JaMa: nope, the second is fine. Aug 03 11:47:48 JaMa: your suggestion in the readme sends two emails Aug 03 11:47:51 no idea why... Aug 03 11:47:57 it asked me the same question twice. Aug 03 11:48:16 id you combine -1 and ? Aug 03 11:48:21 s/id/did/ Aug 03 11:48:41 JaMa: no, I passed 0001... that may have been the reason then. Aug 03 11:49:21 yes it's the same issue as git send-email -1 which sends HEAD-1 + patches Aug 03 11:51:36 rburton: have you seen my question about TOPDIR and so? Aug 03 11:51:41 yesterday evening. Aug 03 14:52:20 tlwoerner, that was a good talk you guys gave at Linaro connect Aug 03 15:04:56 Crofton: thanks :-) Aug 03 15:05:41 i helped more with the lab, which i don't think was recorded Aug 03 15:06:23 the lab was tough to do, despite having DVDs and USB sticks prepared with DL_DIR and sstate-cache Aug 03 15:19:57 labs are tough Aug 03 15:20:28 you should chat with the people that help with the dev day labs Aug 03 15:20:50 figuring out how to do labs wwithout lots of infrastructure is an issue Aug 03 15:36:57 the whole presentation (which was recorded) was done by Nicolas (ndec); he did a great job! Aug 03 17:51:45 this is very nice => https://www.yoctoproject.org/docs/current/poky-ref-manual/poky-ref-manual.html#maintaining-build-output-quality Aug 03 17:52:19 * nerdboy can't wait to try it... Aug 03 18:10:51 someone interested in evolution-data-server? Aug 03 20:54:49 i would assume gnome and/or telepathy folks would be... Aug 03 20:55:02 ? Aug 03 20:55:29 eds is a dependency for a full/modern gnome Aug 03 20:56:15 nerdboy: it's broken for some time, so I'm looking for someone willing to update/fix it Aug 03 20:56:42 nerdboy: and it's better when that someone is actually using it too :) Aug 03 20:59:23 gotcha Aug 03 20:59:53 i thought that was your way of saying you just fixed/pushed it... Aug 03 21:06:43 nerdboy: no it was moved from oe-core to meta-gnome in meta-oe because of lack of testability and interest Aug 03 21:07:00 nerdboy: now it's broken since last glib upgrade Aug 03 21:08:43 happen to know which version of glib broke it? Aug 03 21:10:04 I can check "State of bitbake world" threads to see since when it's broken Aug 03 21:10:33 got a bug number instead? Aug 03 21:12:53 it was broken in May already http://permalink.gmane.org/gmane.comp.handhelds.openembedded/58019 Aug 03 21:13:34 is there a non-git version of evolution-data-server? Aug 03 21:13:49 all i see is evolution-data-server_git.bb Aug 03 21:14:41 so probably broken since upgrade from 2.34.3 to 2.36.0 Aug 03 21:14:48 Date: Sun Apr 21 14:37:16 2013 -0700 Aug 03 21:15:02 that's why I was asking if there is someone actually using it :) Aug 03 21:16:01 at least cheese, contacts, dates are also unavailable since then Aug 03 21:16:16 nerdboy: only git recipe is there Aug 03 21:16:21 eds-3.6.4 works with glib-2.36.3 Aug 03 21:17:16 2.32.3 should be working with it too and is closer to what we have now Aug 03 21:17:30 and on the gentoo side i only see patches for eds-2.3x Aug 03 21:17:56 but there is quite a few patches which will need rebase and testing in runtime Aug 03 21:18:00 2.32.3 with patches... Aug 03 21:18:08 yup Aug 03 21:18:24 so I've started with update and then decided to leave it to someone using it and spend my time fixing other recipes instead Aug 03 21:18:54 right now i'm using as little gnome as i can get away with... Aug 03 21:19:17 plus i have afew requests in the queue already Aug 03 22:45:53 anyone see this before? work-shared/gcc-4.8.1-r0/gcc-4.8.1/gcc/cfgloop.h:20:0: error: unterminated #ifndef Aug 03 22:55:38 JaMa: no, what target? master? Aug 03 22:57:30 master qemux86-64 Aug 03 22:57:39 started after today's oe-core upgrade Aug 03 22:59:06 hmm, I am building master on the AB, but I built MUT earlier this week with that set of changes. Aug 03 23:00:29 maybe it one-of-issue, will try to cleansstate after the build end Aug 03 23:00:29 s Aug 03 23:02:25 *liberally sprinkle "s" where needed... Aug 03 23:06:37 * nerdboy waits for the espresso Aug 03 23:53:47 sgw_: build fine after cleansstate :/ Aug 04 00:00:38 JaMa: right :/ I hate those! **** ENDING LOGGING AT Sun Aug 04 02:59:58 2013