**** BEGIN LOGGING AT Tue Mar 07 10:59:57 2006 Mar 07 11:00:51 XorA: ?? we are one step from need for distro-native and call-to-get-oe-devel-to-get-it-started-at-all Mar 07 11:05:14 btw - did I miss something in last 20 minutes? Mar 07 11:06:28 hrw: http://pastebin.com/588617 Mar 07 11:06:28 XorA: looked at http://www.gibix.net/projects/qemu/ yet ? Mar 07 11:06:43 hrw|work: what do we need from me - the world build or also the images? Mar 07 11:06:43 (a world build has been started some hours ago) Mar 07 11:07:08 XorA: I installed gcc 3 on the machine and made it the default for qemu's benefit. A bit of a cop-out but it got it working :) Mar 07 11:07:35 mickeyl: all wuld be best Mar 07 11:08:14 hrw: i was afraid of you saying that Mar 07 11:08:15 heh Mar 07 11:08:22 RP: it looks like switching DISTRO's is probably my easiest path :-) Mar 07 11:09:02 mickeyl: happens Mar 07 11:09:17 * XorA has no idea why he randomely capitalised DISTRO Mar 07 11:09:27 OE reflexes Mar 07 11:09:41 hrw, mickeyl For 3.5.4.1, I was thinking of branching the release branch and then starting to add the 2.6 bits. For that release, I assume we're just going to do machine images for c7x0, akita, borzoi and spitz? Mar 07 11:10:01 RP: .oz3541 branch from .oz354fam083 sounds good. Mar 07 11:10:12 RP: and yes - 2.6 machines only Mar 07 11:10:23 RP: i agree, that sounds like the way to go Mar 07 11:10:40 FYI: http://handhelds.org/hypermail/familiar-dev/current/0735.html Mar 07 11:10:44 RP: also tosa-2.6 experimental would be good Mar 07 11:10:51 the familiar people will stop using OE Mar 07 11:11:38 koen: aha. good to have an announcement - finally Mar 07 11:11:53 when we will remove familiar-unstable.conf? Mar 07 11:12:22 don't hurry Mar 07 11:12:26 hrw: If do13 provides some tosa updates, we can do that Mar 07 11:12:31 mickeyl: notice the last paragraph where 'hh.org' and 'familiar' become synonymous Mar 07 11:13:28 RP: Dirk wanted to release 2.6 image as experimental Mar 07 11:14:25 hrw: That's fine - we just need to sort the kernel out a bit Mar 07 11:14:53 XorA: what we do in sourcemage is to have a indication on a package if it needs a gcc other than default, then the build system (sorcery) changes env vars when building that package so that it will use the other gcc (mostly 3.4, but a couple of things haven't been updated in ages and need 2.95) Mar 07 11:14:58 koen: what does that mean for 0.8.3 ? Mar 07 11:16:31 mickeyl: that someone has to step up to release it and see that work getting thrown away soon after Mar 07 11:17:18 This looks like the end of familiar in its current form as far as I can see :-/ Mar 07 11:17:23 <_law_> hi all Mar 07 11:17:37 RP: and it seems to burn a lot of bridges: http://handhelds.org/hypermail/familiar-dev/current/0746.html Mar 07 11:20:17 mickeyl: time to start propagandizing EABI and Angstrom :) Mar 07 11:20:32 koen: heh, lets first do some work to get it running ;) Mar 07 11:21:02 indeed Mar 07 11:21:22 Bernardo: I assume you run sourcemage personally then? Mar 07 11:21:28 Bernardo: been googling Mar 07 11:21:34 XorA: I do Mar 07 11:21:52 but my oe build machine is (still) mandriva Mar 07 11:22:20 I was having too much trouble with building on my amd64 machine, so I started using my mandriva domain server Mar 07 11:22:29 Personally, I can see Angstrom taking most of the developers on board and becoming what Familiar should have been... Mar 07 11:22:48 it is still using mandriva on a basis of "it isn't broke, don't fix it" Mar 07 11:23:18 RP: *nod* no work is lost. Mar 07 11:25:43 mickey|lunch: And the fact the work isn't lost is the main thing Mar 07 11:55:10 so it looks like familiar/OE will end but who knows how much devels will stay with familiar/non-OE and how much with non-familiar/OE? Mar 07 11:57:17 I wonder what will happen to familiar 0.8.3... Mar 07 11:57:22 do I get it quite right? Mar 07 11:57:53 brb Mar 07 11:58:27 hrw: sounds like a good conclusion Mar 07 11:59:40 all OE needs to do is to keep momentum with new hardware ports (h6300, hx4700, h2200, palms, htc phones) Mar 07 12:00:16 as far as i can see familiar just killed itself Mar 07 12:02:05 ade|desk: and this is bad.. it was most recognized linuxpda distro Mar 07 12:02:07 gpe,opie etc aren't bound to familiar and most dev stuff for handhelds is in those projects so why stick with familiar as the baseline Mar 07 12:02:12 but who is on familiar/non-oe? Mar 07 12:02:25 Bernardo: France? Mar 07 12:02:27 Besides France? Mar 07 12:02:29 :) Mar 07 12:04:07 good question Mar 07 12:04:38 morning all Mar 07 12:04:53 hi Dirk Mar 07 12:05:05 Hey Richard Mar 07 12:05:18 hi Dirk Mar 07 12:05:41 hi Marcin Mar 07 12:06:17 do13_: OZ 3.5.4.1 - do we release tosa/2.6 experimental with it? (~about one month) Mar 07 12:06:26 hrw: 3.5.4 congrats Mar 07 12:06:52 hrw: I need Liams Wm9712 code Mar 07 12:07:05 do13_: I know Mar 07 12:07:28 do13_: we could do 3.5.4.2 for tosa ;) Mar 07 12:07:48 do13_: Which kernel version are you developing against at the moment? Mar 07 12:07:56 RP: 2.6.15 Mar 07 12:08:43 do13_: ok. At some convinient point, we should probably update the tosa patches in OE then. Mar 07 12:08:56 do13_: Are there any changes you want to get into 2.6.17? Mar 07 12:09:17 RP: agreed. I'am waiting for the wm9712 stuff Mar 07 12:09:27 hi do13 Mar 07 12:09:36 hi pb__ Mar 07 12:09:45 do13_: ok, that makes sense :) Mar 07 12:10:03 RP: I think about pushing the keyboard driver into 2.6.17 Mar 07 12:11:35 do13_: ok. If you want to push anything, I'd advise doing it soon as the 2.6.16 release is imminent and after that you probably have about a week to get patches sent Mar 07 12:13:08 * koen is looking forward to 2.6.16 Mar 07 12:13:10 EABI! Mar 07 12:14:01 koen: 2.6.16-rc5 zaurus kernels work btw Mar 07 12:14:03 koen: you almost have all your peices Mar 07 12:14:13 RP: cool Mar 07 12:14:17 In fact, we could have 2.6.16 in oz-3.5.4.1 :) Mar 07 12:14:43 RP: I might have to move my eabi tests to the Z, since eabi will be armv5 only at this point Mar 07 12:14:54 my ep9312 is armv4t :( Mar 07 12:15:26 :-( Mar 07 12:16:07 *light bulb* Mar 07 12:16:17 the ixp422 is armv5! Mar 07 12:16:27 koen: slug/eabi? Mar 07 12:16:41 hrw: loft/eabi Mar 07 12:17:07 your another arm toy? Mar 07 12:17:10 533MHz/64MB + 2x ethernet + 2x usb + cf + 3x minipci Mar 07 12:17:19 sounds nice Mar 07 12:17:31 http://www.giantshoulderinc.com/hardware.html Mar 07 12:18:28 I just need to finish my report.... Mar 07 12:18:34 how big it is and how much it cost? Mar 07 12:19:35 $350 and it's about the size of 3 dvds Mar 07 12:19:48 ~change 350 usd to pln Mar 07 12:19:53 350.00 United States Dollar (USD) makes 1,135.77 Zloty (PLN) (from http://www.xe.com/) Mar 07 12:20:06 ~change 350 usd to gbp Mar 07 12:20:10 350.00 United States Dollar (USD) makes 201.599 Pound (GBP) (from http://www.xe.com/) Mar 07 12:20:15 RP: Thanks. I'll consider this. All other drivers depends on the tc6393 Mar 07 12:20:36 it's the black thing in the bottomright corner: http://dominion.kabel.utwente.nl/koen/ep93xx/images/ep93xx-setup_small.jpg Mar 07 12:20:51 do13_: Are there problems with getting tc6393 into mainline? Mar 07 12:21:01 koen: hmm. epia is cheaper Mar 07 12:21:46 RP: Hmm. I have to ask mahadri, because he maintains the new driver Mar 07 12:21:52 RP: As I head out the door, bb in 2 hours...familiar has always changed. Once upon a time it was based upon rpms...that is no longer true....it will change in the future... the details of those changes has yet to be discussed....(maybe thats why there is a need for a steering comittee)...Maybe OE will be involved...Maybe it wont....Currently it is off track...It is in bad need of repairs....How to get them done will be the questoin. Mar 07 12:21:52 ..too much fud going on by folks with their own agendas...making up stuff Mar 07 12:22:40 koen: http://www.scan.co.uk/Products/ProductInfo.asp?WebProductID=340118 Mar 07 12:22:41 RP: too say we can not even consider changes is just sad.... Mar 07 12:22:54 RP: heading out the door. Mar 07 12:24:06 * XorA so needs a 800G arm pocket sized box Mar 07 12:25:10 RP: Do you plan to update 2.6.16-rc5 in oe? Mar 07 12:25:47 RP: that one looks great Mar 07 12:25:53 do13_: I'll add it if anyone wants to use it Mar 07 12:26:04 RP: I've been looking for a Gb/s filestorage device Mar 07 12:27:00 although I fear they uses RealTek chips for the ethernet Mar 07 12:27:28 RP: ok. I'll look at this the next days Mar 07 12:28:11 do13_: I have local .bb files for _git, _2.6.15+2.6.16-rc5 and 2.6.15+2.6.16-rc5-mm2 :) Mar 07 12:28:46 RP: does _git compiles? Mar 07 12:29:27 do13_: Most of them have the tosa patches commented out :-/ Mar 07 12:29:48 do13_: I've updated all the other patches where needed Mar 07 12:32:42 do13_: I've put the .bb files I have into http://www.rpsys.net/openzaurus/temp/ Mar 07 12:33:44 -rc5 uses the whole rtc subsystem from -mm which now includes a driver for the rtc on the pxa Mar 07 12:33:58 (as of this morning :) Mar 07 12:36:18 RP: is there a way the Zaurii will benefit from that RTC subsystem? Mar 07 12:38:22 mickeyl: It means you can use /dev/rtc0 as an accurate timer Mar 07 12:38:37 mickeyl: certain video players do this for example Mar 07 12:38:43 RP: which resolution can we get from there฿ Mar 07 12:38:44 ? Mar 07 12:38:57 mplayer likes /dev/rtc Mar 07 12:39:39 mickeyl: as a non-root user, there is one limit, as root, you can get quite impressive resoution. I can't remember the figures offhand Mar 07 12:40:00 RP: fair enough. sounds interesting, i'll take a look at that Mar 07 12:40:24 mickeyl: check out Documentation/rtc.txt in the kernel source Mar 07 12:40:36 mickeyl: There is also a programmable timer there now - that was broken in previous pxa drivers but I've fixed it Mar 07 12:40:43 cool Mar 07 12:41:24 (/dev/rtc is pretty awesome.) Mar 07 12:42:09 njs: can monotone use it to get 1000000 times faster? Mar 07 12:42:11 :p Mar 07 12:42:33 <_law_> any ideas why i get an qemu error on building package-index? http://pastebin.com/588772 Mar 07 12:42:45 no, but it can use it to precalculate that you need to wait 01:23:43.03 for the next pull to compelte Mar 07 12:42:48 complete, even Mar 07 12:43:00 too bad that pxa does not have rtc which will keep datetime ;( Mar 07 12:43:21 koen: it can...... umm... set an alarm to go off when the pull finishes... Mar 07 12:43:25 hrw: back to good old amiga days when you had to buy an RTC :-) Mar 07 12:43:34 _law_: | ERROR: "gcc" looks like gcc 4.x Mar 07 12:43:39 _law_: sounds pretty self explanatory to me Mar 07 12:43:43 (I need an alarm clock that raises an IRQ line instead of like beeping and stuff.) Mar 07 12:43:47 (I would get more sleep.) Mar 07 12:43:51 _law_: oh you were asking about the dependencies, right? Mar 07 12:44:07 XorA: yep ;( Mar 07 12:44:16 _law_: i guess glibc nowadays needs qemu unless you disable that. have a look at pb_'s recent patch Mar 07 12:44:35 mickeyl: resolution wise, it appears it runs in the Mhz region (but non-root users can only request latency of up to 64Hz) Mar 07 12:44:44 _law_: http://www.handhelds.org/~bugzilla/show_bug.cgi?id=1385 Mar 07 12:44:52 koen: in one of my other lives, though, I do psychological experiments, that need accurate timing and control :-) soft-real-time with sub-millisecond deadlines, whee. Mar 07 12:45:02 hi mickeyl Mar 07 12:45:42 RP: Thanks, I'll try this the next days Mar 07 12:45:47 <_law_> mickeyl, but use gcc-cross-3.4.4-r3 Mar 07 12:45:56 do13_: hey dirk! Mar 07 12:46:15 _law_: are you sure? it looks like you're building qemu-native... Mar 07 12:46:27 _law_: which would mean your host compiler version counts Mar 07 12:46:40 _law_: I am in same boat Mar 07 12:46:46 _law_: gcc4 on host Mar 07 12:47:10 <_law_> XorA, right :-( Mar 07 12:47:30 <_law_> XorA, so would could we do? Mar 07 12:47:51 _law_: which distro are you? Mar 07 12:48:06 <_law_> XorA, oz-unstable Mar 07 12:48:13 _law_: on host I mean Mar 07 12:48:21 do the qemu folks know about that yet? Mar 07 12:49:06 <_law_> XorA, Fedora Core 4 Mar 07 12:49:17 _law_: then add BINARY_LOCALE_ARCHES = "" to you conf/local.conf Mar 07 12:49:22 like the comment says Mar 07 12:49:39 _law_: I dont remeber RedHat ever supporting multiple gcc's on same machine :-( Mar 07 12:49:44 _law_: I am FC4 as well Mar 07 12:50:31 jnc: do you have a working prototype of gcc_4.2- available somewhere? Mar 07 12:54:30 XorA _law_ mickeyl : http://www.gibix.net/projects/qemu/ Mar 07 12:56:20 RP: are there plans for poky to move to gcc4 and/or eabi in the forseeable future? Mar 07 12:59:02 ade|desk: right. we should experiment with these patches Mar 07 13:00:06 any opinions on turning on gcj by default for gcc >4.0.0? Mar 07 13:02:22 koen: sounds good to me, how much does it increase build time? Mar 07 13:02:33 koen: poky will probably move to the eabi and gcc4 at the same time and only once it works Mar 07 13:02:56 mickeyl: I'm more concerned about the increased ram usage Mar 07 13:03:01 mickeyl: we'd have to ask zecke Mar 07 13:03:33 RP: that's good to hear Mar 07 13:08:00 koen: Having asked around, I can't find anything about maemo being any more interested in OE than they've ever been :-/ Mar 07 13:08:11 RP: :( Mar 07 13:09:23 koen: just have to make a better than maemo then Mar 07 13:09:33 right what i wanted to say, XorA Mar 07 13:10:26 My thoughts as well. My opinions on scratchbox vs. OE are well known ;-) Mar 07 13:12:22 yeah :) Mar 07 13:14:07 hi all Mar 07 13:14:20 hey florian_kc Mar 07 13:14:59 hi flo Mar 07 13:15:49 too bad you couldn't make it this year to fosdem. next year, i hope? Mar 07 13:17:49 mickeyl: If they manage to keep a way from my wife's and my son's birthday its pretty likely i'll make it. Mar 07 13:23:15 florian_kc: is it the 25th or the 26th? Mar 07 13:26:31 mickeyl: 26th :-) but this time there were a pile of additional reasons... Mar 07 13:50:26 <_guillermo> can anyone tell me what does qemu do exactly in OE? Mar 07 13:51:02 twofold Mar 07 13:51:20 a) it allows testing distributions by setting it as a target machie Mar 07 13:51:39 b) it allows generating stuff which needs to run on the target machine or something reasonably similar Mar 07 13:52:17 b) is mostly a problem that shouldn't exist Mar 07 13:52:27 in a perfect world, yes Mar 07 13:52:37 like localegen OOM on 64M machines Mar 07 13:52:40 yep Mar 07 13:53:18 <_guillermo> so, if I get it right, could we use a pda as host machine and compile x86 compiled code, that's not possible, is it? Mar 07 13:54:27 <_guillermo> RP: what is the problem with scratch box? Mar 07 14:02:20 <_guillermo> I still can't compile qemu-native bacause I have no sdl, what package provides sdl support (as a workaround till a patch is created) ? Mar 07 14:12:40 _guillermo: Add an option to disable it to EXTRA_OEMAKE in the .bb file Mar 07 14:13:02 _guillermo: Scratchbox hacks around too many issues IMO and has no concept of packaging. OE has more structure but doesn't have the development platform (yet) Mar 07 14:13:24 <_guillermo> EXTRA_OEMAKE or EXTRA_OECONF ? Mar 07 14:13:44 _guillermo: Sorry, EXTRA_OECONF Mar 07 14:13:57 <_guillermo> the --disable-gfx-check is the right thing, isn't it? Mar 07 14:15:14 _guillermo: -disable-sdl might be better Mar 07 14:15:21 <_guillermo> ok Mar 07 14:20:35 http://pastebin.com/588897 Mar 07 14:22:59 XorA: http://handhelds.org/scap/port.29351.png <- phear the new OE printing features ;) Mar 07 14:23:30 koen: sorted Mar 07 14:23:52 woot! Mar 07 14:24:04 wow Mar 07 14:24:43 koen: get that working with BT printing and we have a PDA that can print through shop windows :-) Mar 07 14:24:50 <_guillermo> can't believe it. everyday is more difficult to make a different between my akita with oe and my laptop Mar 07 14:25:59 while trying to build with MACHINE=qemuarm : Mar 07 14:26:01 task-bootstrap: unsatisfied recommendation for kernel-module-ipv6 Mar 07 14:26:13 and there is some more. what is missing ? Mar 07 14:26:27 kernel-module-ipv6, at a guess Mar 07 14:26:49 yes, but i can't find a .bb for this Mar 07 14:27:59 it's part of the kernel. check your config file. Mar 07 14:28:20 hey lrg Mar 07 14:28:23 hi lrg Mar 07 14:28:29 cheers liam Mar 07 14:28:30 hey koen , florian_kc Mar 07 14:28:31 ok, thx Mar 07 14:28:35 hi mickeyl Mar 07 14:29:45 Hi Liam Mar 07 14:30:12 hi Dirk, almost ready to sen d a patch for tosa Mar 07 14:30:27 cool. Can't wait :) Mar 07 14:30:45 do13: this will have everything bar the noisey LCD fix Mar 07 14:31:05 do13: I'll do the noise LCD after this release Mar 07 14:31:26 hi Liam Mar 07 14:31:39 hi Richard Mar 07 14:31:50 victorz: Alternatively, find what it RDEPENDing on that module and change it to an RRECOMMENDS Mar 07 14:32:07 lrg: aux adc reading and touch isn't no longer a problem at the same time? Mar 07 14:32:17 RP: think I've figured why the record on corgi (wm8731) wasn't working. Mar 07 14:32:40 RP: it looks like alsa-lib hides some mixers when the names are similar Mar 07 14:32:50 hi liam Mar 07 14:32:55 morning Mar 07 14:32:55 do13: wont be for much longer Mar 07 14:33:06 hey hrw chouimat Mar 07 14:33:29 RP: the modules is already in RRECOMMENDS Mar 07 14:33:41 victorz: its recommendation - you can ignore it Mar 07 14:33:48 ok, cool Mar 07 14:41:12 lrg: Ah, that would explain it. I thought the mixer was looking a little bare but couldn't pinpoint exactly why - I'd assumed it was the merging of the L+R mixers :) Mar 07 14:41:24 lrg: Does this also affect the wm8750? Mar 07 14:46:14 hi all Mar 07 14:46:37 I've got problems while bitbaking opie-image: Mar 07 14:46:44 | ERROR: Cannot satisfy the following dependencies for task-bootstrap: Mar 07 14:46:44 | modutils-initscripts fuser setserial linux-hotplug familiar-version ipkg-collateral ipkg ipkg-link udev module-init-tools modutils-depmod module-init-tools-depmod Mar 07 14:46:51 can anyone help me? Mar 07 14:47:19 Farcaller: you bitbake is proably too old Mar 07 14:48:09 1.3.2 Mar 07 14:48:39 that's not old, that's ancient ;) Mar 07 14:48:49 try using bitbake from svn Mar 07 14:49:47 RP: thanks, I'll check it Mar 07 14:50:07 koen, That's the most new verison in gentoo's portage... ;) Mar 07 14:50:17 so gentoo sucks Mar 07 14:50:19 can you give me the link to svn? Mar 07 14:50:23 big surprise Mar 07 14:50:39 lrg: I'm still a little puzzled as to why I can't record without having playback active too on the wm8750 Mar 07 14:51:34 RP: hmmm. not sure. and I don't have my borzoi any more :( Mar 07 14:51:39 Farcaller: http://oe.handhelds.org/cgi-bin/moin.cgi/GettingStarted Mar 07 14:51:54 RP: I'll try and speak with greentux about a replacement Mar 07 14:52:13 koen, thanks Mar 07 14:52:15 lrg: record only worked on mainstone? (with no sidetone enabled) Mar 07 14:52:19 <_guillermo> I don't think gentoo sucks just because bitbake is old in portage Mar 07 14:53:02 RP: record also worked on borzoi (with and without sidetone) Mar 07 14:53:15 RP: I think it's the same issue Mar 07 14:53:50 RP: I'll check over this and inculde a name change to add the control if its missing. Mar 07 14:54:29 lrg: record with playback active at the same time works fine as does recording whilst the output stage is active due to sidetone. If the output stage is not active, recording doesn't work though :-/ Mar 07 14:56:05 <_guillermo> RP: the EXTRA_OECONF = "--disable-sdl" goes in qemu-native_cvs.bb, doesn't it? I hope not, because it's not working Mar 07 14:56:15 RP: ok, it's not what i said then. I've added some extra debug to the latest code. It might be a good idea to send me the output after the release Mar 07 14:56:37 RP: and I'll get to the bottom of it . Mar 07 14:57:04 RP: although I can't think what's changed :( Mar 07 14:57:37 lrg: I'll try the new release and see what happens and we can take it from there Mar 07 14:57:37 hey Mar 07 14:57:43 hey zecke Mar 07 14:57:45 hi zecke Mar 07 14:57:49 RP cheers Mar 07 14:57:56 hey zecke Mar 07 14:58:01 hi zecke Mar 07 14:58:15 that Alsa thing is still not done? Mar 07 14:58:32 * zecke forseas a bright future for liam giving alsa talks all over the world :) Mar 07 14:58:37 zecke: heh, it's not easy ;) Mar 07 14:59:16 * chouimat is having fun fixing some "forgotten cases" in tailor Mar 07 15:00:39 do13_: did you make any progress on konqe? Mar 07 15:01:14 do13_: I hope to find the alc bug, then do some JMF work and then I can hack on konqe :} Mar 07 15:02:39 zecke: not yet. Mar 07 15:08:29 _guillermo: does --disable-sdl exist for qemu ? Mar 07 15:08:44 <_guillermo> I check the configure and yes, it does Mar 07 15:09:30 <_guillermo> I made it work finally, --disable-gfx-check did the job Mar 07 15:10:09 <_guillermo> I don't know if that should be passed by default Mar 07 15:11:50 _guillermo: It shouldn't as some of us use the sdl grpahics. Instead a nosdl version of the package should be created Mar 07 15:12:27 <_guillermo> ok I understand Mar 07 15:12:54 <_guillermo> maybe disabling dsl and cocoa could have worked Mar 07 15:14:34 I'd expect the disable checks option to disable the checking for specific versions of sdl/cocoa but to bail out from the compile when it finds their not there to link to Mar 07 15:55:50 <[g2-lap]> ipkg: Cannot create directory `/usr/lib/ipkg': Permission denied Mar 07 15:55:58 ignore Mar 07 15:56:01 <[g2-lap]> does that look familiar to anyone ? Mar 07 15:56:41 <[g2-lap]> hrw non-issue ? Mar 07 15:57:01 hrw: ignore, I see it too, no problems Mar 07 15:57:05 [g2]: you can ignore that bug Mar 07 15:58:14 hrw: sorry, meant to point that at [g2-lap] Mar 07 15:58:22 np Mar 07 16:00:24 hmmm Mar 07 16:00:31 cpufreq on c700 = bad idea Mar 07 16:01:43 crash or just not effective? Mar 07 16:02:09 freeze Mar 07 16:02:29 rougly 30 seconds after enabling 'ondemand' it freezes Mar 07 16:03:57 ~lart pxa250 Mar 07 16:03:58 * ibot chops pxa250 in half with a free Solaris 7 CD Mar 07 16:06:27 koen: no, not working. i do know that libssp needs to be --disable-libssp'd Mar 07 16:06:36 koen: libssp is broken currently, per gcc devs Mar 07 16:07:28 that confirms one of my suspicions Mar 07 16:10:39 koen: it's just stack protection stuff, even in Gentoo that sort of thing is highly experimental in nature Mar 07 16:11:00 the default is to enable it because it is actively developed Mar 07 16:19:15 koen: perhaps there should be a mandatory SDK being built? Mar 07 16:19:33 i.e. you cannot build OE unless you build the SDK first Mar 07 16:19:51 and the SDK would have a different configuration file from the distro of the target you're building Mar 07 16:20:01 a standardized SDK has many benefits Mar 07 16:20:14 like meta-sdk? Mar 07 16:20:42 perhaps. i want the confusion of having to build a pre-gcc to be relocated Mar 07 16:21:04 it just means re-organization really, maybe a few changes to bitbake Mar 07 16:21:36 jnc: ?? Mar 07 16:21:45 bitbake interferes with a SDK? Mar 07 16:21:49 when I read all recent OE .dev conversation I feel more and more that I will have problems with starting new build... Mar 07 16:22:07 packages that are built and run on the host machine should not be mixed in with the packages (bb) for the target Mar 07 16:22:14 hrw: meta-opie-all builds Mar 07 16:22:38 zecke: even if I do not have libsdl-dev installed? Mar 07 16:22:52 having gcc_xyz.bb, gcc-cross_xyz.bb, and gcc-cross-initial_xyz.bb all in the same place is confusing, i am saying Mar 07 16:22:55 jnc: where is the relation to the task executor bitbake? Mar 07 16:22:56 jnc: what are you talking about? Mar 07 16:22:57 hmm, anyone else having problems to update bitbake from berlios? Mar 07 16:23:03 it doesn't recognize my password Mar 07 16:23:15 jnc: they ar eseperate....normal vs. -native Mar 07 16:23:21 also, having the -native bb's in the same place as non-native bb's is confusing Mar 07 16:23:24 jnc: and staged to different places (in .dev) Mar 07 16:23:31 mickeyl: you probably had too easy password and got it reset Mar 07 16:23:44 hrw: but i just changed it and it still doesn't work Mar 07 16:23:58 mickeyl: 'just' < 6h? Mar 07 16:24:07 the SDK should be its own distro Mar 07 16:24:22 jnc: no Mar 07 16:24:45 hrw: yep Mar 07 16:24:46 jnc: how you will then 'link' openzaurus-3.5.4 with sdk-for-openzaurus-3.5.4? Mar 07 16:24:57 you just have one SDK Mar 07 16:24:59 mickeyl: then wait until they propagate password on servers Mar 07 16:25:09 hrw: possible Mar 07 16:25:15 jnc: one sdk to rule them all? glibc/uclibc/softfloat/hardfloat etc? Mar 07 16:25:20 yeah Mar 07 16:25:57 mickeyl: do you need tos et SVN_SSH Mar 07 16:26:09 jnc: so if I will build glibc-arm-softfloat distro for device and want to build SDK for it then I will get glibc/uclibc/softfloat/hardfloat sdk for it instead of just glibc-arm-softfloat? Mar 07 16:26:26 the EABI stuff should get rid of the need for soft/hardfloat Mar 07 16:26:30 A large company (not one anyone knows) has done an evaluation on OE and their conclusion is that its too slow. This is looking likely to threaten OH's use of OE :-( Mar 07 16:26:31 jnc: I guess you are covering bandwidth? Mar 07 16:26:40 jnc: arm/mipsel/x86 too? Mar 07 16:26:48 I'm fighting a losing battle :-( Mar 07 16:26:56 RP: :( Mar 07 16:27:02 hi, folks. here is a problem while I bitbake glibc : http://rafb.net/paste/results/neFcBB81.html Mar 07 16:27:07 any ideas ? Mar 07 16:27:14 RP: well, it is pretty slow. i like it though Mar 07 16:27:34 I can't argue about the speed Mar 07 16:27:38 zecke,hrw: do you remember the complete svn command line to download? Mar 07 16:27:54 hrw: sure. you build the same SDK and add additional targets it can build for to the config Mar 07 16:27:55 mickeyl: no - but there is one on berlios Mar 07 16:27:57 mickeyl: svn+ssh://berlios/svnroot/repos/bitbake/bitbake Mar 07 16:28:00 RP: too slow in what regard? Mar 07 16:28:13 mickeyl: svn.berlios.de Mar 07 16:28:28 jnc: have fun :) Mar 07 16:28:41 jnc: now I run 'bitbake gpe-image opie-image bootstrap-image meta-sdk' and go sleep - all in one command Mar 07 16:28:57 RP: is this company aware that software can be improved? Mar 07 16:29:19 leoncamel: you hit recent qemu addon - I do not know how to fix it Mar 07 16:29:47 zecke: Not an option - they need something now Mar 07 16:30:04 can't we just revert that qemu commit? Mar 07 16:30:19 I can't seem to disable at and it breaks my build Mar 07 16:30:20 koen: or make it optional.... Mar 07 16:30:36 I have: Mar 07 16:30:40 Wasn't there a disable option added? Mar 07 16:30:40 hrw, hmm, I bitbake without any error . what do you mean exactly ? Mar 07 16:30:44 USE_BINARY_LOCALE = "0" and BINARY_LOCALE_ARCHES = "" Mar 07 16:30:56 and it still tries to do funky qemu stuff Mar 07 16:31:04 koen: where do you have those? Mar 07 16:31:08 I guess that doesn't remove the qemu dependency? Mar 07 16:31:29 pb__: in local.conf Mar 07 16:31:37 RP: is it mainly the speed of bitbake parsing? Mar 07 16:31:42 RP: or parallel builds? Mar 07 16:31:49 zecke: speed of parsing Mar 07 16:31:51 koen: the definitions in glibc-package.bbclass are probably clobbering yours. Change the former to use ?=, or just edit the assignment in place. Mar 07 16:32:05 RP: do they want to see a roadmap on when it will be fixed? Mar 07 16:32:33 zecke: OE is ruled out now for them period. Mar 07 16:32:34 BINARY_LOCALE_ARCHES ?= "arm" is in glibc Mar 07 16:32:41 okay Mar 07 16:32:50 zecke: I'm just passing on the information I can... Mar 07 16:32:59 ah, is it? In that case, I guess your problem must be elsewhere. Mar 07 16:33:00 RP: bitbake -i wasn't an option? Mar 07 16:33:16 koen: it isn't enough for them Mar 07 16:34:02 RP: ah well, it's not that there aren't alternatives Mar 07 16:34:04 koen: oh, I remember. BINARY_LOCALE_ARCHES is a regexp. you probably need to set it to "none" or something. Mar 07 16:34:11 if you set it to the empty string, I guess it matches everything. Mar 07 16:34:49 * koen sets it to "0" Mar 07 16:35:16 so reenoo was lying: http://cia.navi.cx/stats/project/openembedded/.message/f8b4a Mar 07 16:35:17 leoncamel: if NOTE: package glibc-2.3.5+cvs20050627-r3: task do_package: failed is not an error for you then I have no questions Mar 07 16:35:48 I go home - cu tomorrow guys Mar 07 16:35:52 koen: I doubt he was deliberately lying, but it's possible he was mistaken. Of course, it's equally possible that I am mistaken and he was actually correct. Mar 07 16:38:47 do anybody have a looku at : http://rafb.net/paste/results/neFcBB81.html ? Mar 07 16:40:13 leoncamel: that's clearly a bug, but I guess you have to investigate it yourself. Mar 07 16:41:55 hmm.. i'll try to build the same pretty soon... Mar 07 16:42:25 pb__, hmm, I successfully build glibc-2.3.5+cvs20050627-r3 a few days ago. but I failed today.. :( Mar 07 16:43:06 hi, florian_kc . Mar 07 16:43:12 NOTE: preparing tree for binary locale generation Mar 07 16:43:12 NOTE: generating locale es_NI (UTF-8) Mar 07 16:43:13 sh: qemu-arm: command not found Mar 07 16:43:18 florian_kc, do you encounter this before ? Mar 07 16:43:21 breakage all around Mar 07 16:43:24 leoncamel: i think for now it might be wise to BINARY_LOCALE_ARCHES="none" or revert the tree to get the old behaviour back Mar 07 16:43:25 leoncamel: it was changed a short time ago... Mar 07 16:43:34 03koen 07org.oe.dev * r5d73d97c... 10/packages/glibc/glibc-package.bbclass: glibc: make qemu breakage disasters opt-in instead of opt-out Mar 07 16:43:52 koen: yeah we need to improve Mar 07 16:44:23 koen: oh.. sounds like a missing dependency. Mar 07 16:44:44 florian_kc: no, I explicitly disabled that qemu stuff Mar 07 16:46:30 koen, but why ? why qemu affect the glibc package ? Mar 07 16:47:17 leoncamel: http://cia.navi.cx/stats/project/openembedded/.message/f8b4a Mar 07 16:48:30 leoncamel: we had to come up with a way to generate locales on the host, because the process is OOM'ing on the target system Mar 07 16:48:51 mickeyl, what is OOM ? Mar 07 16:49:06 Out Of Memory killing Mar 07 16:49:13 a 'feature' of the linux kernel Mar 07 16:49:41 :). ok, got it . Mar 07 16:52:04 koen, hmm, can monotone provided a patch on a revision ? Mar 07 16:52:15 * zecke heads home Mar 07 16:52:26 so Access might not use OE now :} Mar 07 16:52:38 well they are hooked with MVista anyway (just guessing) Mar 07 16:53:00 access as in access music? Mar 07 16:53:05 zecke: wind river Mar 07 16:53:10 I would think it would be useful to use the MV tools with OE Mar 07 16:53:12 mickeyl: access as in palm Mar 07 16:53:17 koen: ah Mar 07 16:53:27 Crofton: which ones in particular? Mar 07 16:53:33 compilers Mar 07 16:53:35 libs Mar 07 16:54:20 maybe :) Mar 07 17:14:46 koen, hmm, the glibc package seem ok now .. Mar 07 17:19:11 re Mar 07 17:20:25 <_guillermo> yes, glibc just compiled here Mar 07 17:22:21 do we have a svn+https fetcher ? Mar 07 17:22:54 SRC_URI = "svn://svn.cronos.be/svn/tinymail;module=trunk;proto=https \ Mar 07 17:22:56 ade|desk: we have a svn fetcher Mar 07 17:22:58 ade|desk: so, yes :) Mar 07 17:23:05 ade|desk: and proto can be set to aynthing Mar 07 17:23:36 cheers Mar 07 17:24:18 hey zecke Mar 07 17:24:53 hey pH5 Mar 07 17:25:03 hey koen Mar 07 17:25:28 pH5: jnc is in contact with the gcc people on how to fit 4.2 into OE :) Mar 07 17:25:42 koen: Great! I'm still too dumb to build your gcc 4.1 :-/ Mar 07 17:25:43 ew Mar 07 17:26:10 hey CosmicPenguin Mar 07 17:26:51 greetings Mar 07 17:27:00 assuming you're talking to me and not some other C name.,.. :_) Mar 07 17:28:23 yes, I was talking to mr Crouse Mar 07 17:31:37 pH5: do you have any cool avahi apps to test? Mar 07 17:35:12 koen: not yet on the ipaq. see http://www.avahi.org/wiki/Avah4users#SoftwareMakinguseofAvahi Mar 07 17:43:13 NOTE: package glibc-2.3.5+cvs20050627: completed Mar 07 17:43:25 Looks like it works for me here... Mar 07 17:44:14 florian_kc: qemu has been disabled Mar 07 17:44:29 florian_kc: and you probably has qemu on your host for scratchbox Mar 07 17:44:51 koen: you r latest checkin? Mar 07 17:45:01 florian_kc: yes Mar 07 17:45:02 * zecke continues improving bitbake Mar 07 17:45:09 koen: what were the problems with the qemu stuff? Mar 07 17:45:24 (I didn't try it yet, just wondering if I should) Mar 07 17:45:43 pH5: 1) it doesn't build on my machine 2) it's non obvious how to disable 3) the logic in the .inc is broken Mar 07 17:47:01 koen: i pulled before this :-) Mar 07 17:47:51 koen: which logic is broken? Mar 07 17:48:55 pb__: I put the magic "don't use qemu" incantation in local.conf, so it stopped trying to build qemu, bit it did try to generate the locales Mar 07 17:49:58 which gave an obvious 'qemu-arm: command not found' errorm Mar 07 17:50:17 oops, that sounds wrong indeed... Mar 07 17:53:11 what did you use as the magic incantation? Mar 07 17:54:03 BINARY_LOCALE_ARCHES = "0" USE_BINARY_LOCALE = "0" Mar 07 17:54:24 You shouldn't set USE_BINARY_LOCALE. That is a recipe for disaster. Mar 07 17:56:00 Despite what you might think, putting USE_BINARY_LOCALE = "0" in your local.conf would have the effect of forcing binary locale generation _on_ unconditionally, and defeating the interlock with DEPENDS. It's exceptionally unlikely that this would be the effect you wanted. Mar 07 17:56:19 * pH5 apt-get install qemu ; bitbake glibc Mar 07 17:56:23 sounds like broken logic Mar 07 17:56:37 no, the logic is correct; that variable just isn't intended to be frobbed in local.conf. Mar 07 17:57:15 BINARY_LOCALE_ARCHES = "0" turns of the depend, so frobbing USE_BINARY_LOCALE should have no effect at all Mar 07 17:57:21 The only variable that is intended to be under user control is BINARY_LOCALE_ARCHES. Mar 07 17:57:37 Right, and BINARY_LOCALE_ARCHES = "0" would also turn off locale generation if you weren't setting USE_BINARY_LOCALE. Mar 07 17:57:47 which no one knows, since it isn't documented Mar 07 17:58:10 the only doc (the commit message) is also wrong Mar 07 17:59:33 I guess you should feel free to improve the documentation. Mar 07 18:00:29 I'll leave that to the next victim off this qemu disaster Mar 07 18:01:22 * koen dislikes UNBREAK_ME flags which aren't in local.conf.sample Mar 07 18:02:33 later all Mar 07 18:03:52 koen|away: cu Mar 07 18:05:43 hmm... i wonder what happened here... i don't have qemu on this machine. Mar 07 18:06:50 florian_kc: you shouldn't need it: oe is meant to build it for itself. Mar 07 18:07:43 pb__: then we just need to document how to turn it off? Mar 07 18:07:55 s/off/off properly Mar 07 18:08:02 florian_kc: try setting BINARY_LOCALE_ARCHES = "" in local.conf. Mar 07 18:08:09 or, if that doesn't work, use "0" instead of "". Mar 07 18:09:29 pb__: well... personally i pretty much like to have it working. maybe we just need to put this into the wiki or something like that. Mar 07 18:10:04 yeah Mar 07 18:10:16 I guess someone just needs to test it to find out for sure what the right thing is, and then put it in the wiki. Mar 07 18:10:35 actually, I guess BINARY_LOCALE_ARCHES = "" is definitely the right thing; if that doesn't work then we should just fix the code so that it does. Mar 07 18:12:26 pb__: okay... i think i can try this with the next build.. as soon as my nbp build is working. Mar 07 18:12:54 rock Mar 07 18:13:22 I don't really understand what problem koen was having in the first place that made him want to turn it off. I guess he must have been encountering some kind of bug. Mar 07 18:13:26 * pb__ go home now Mar 07 18:13:27 later all Mar 07 18:14:05 hmm indeed, maybe some bitbake issue causing dependencies to miss... Mar 07 18:14:09 pb__: later Mar 07 18:15:01 pb__: qemu-native refuses to build on my opteron Mar 07 18:16:08 ah... that sounds like a problem. Mar 07 18:16:43 koen|food: does qemu support that platform at all? Mar 07 18:16:52 no idea Mar 07 18:17:10 it should, as reenoo turned it on for *everyone* Mar 07 18:17:26 which means he must have tested it on every buildplatform Mar 07 18:17:36 everyone building for arm that is Mar 07 18:18:21 qemu-native also needs sdl-native (which isn't in OE) Mar 07 18:18:44 koen|food: well, i think dependcy should work, otherwise, i even can't build cross-gcc Mar 07 18:19:52 koen|food: well... it works fine for me without having qemu installed, so that can't be the problem. Mar 07 18:20:05 florian_kc: so I had to guess that glibc-packages.inc was the culprit and had to look into python code to guess how to turn it off Mar 07 18:20:36 look you can expect from an average user Mar 07 18:20:47 Mar 07 18:21:07 it only took me 3 hours to get a build running again.... Mar 07 18:21:22 koen|food: these things happen, but this stuff is pretty important for quality - so it would be much more useful to fix it. Mar 07 18:22:09 florian_kc: while it is getting fixed other people (like me) can continue to work Mar 07 18:22:15 that's why it is checked into .dev instead of any other branch Mar 07 18:22:35 and that's why we had "glibc is broken" complaints all morning here Mar 07 18:24:29 florian_kc: the branch being called '.dev' is no excuse to break it for a large part of the community Mar 07 18:25:41 ah well, it stops people complaining about monotone for a while ;) Mar 07 18:25:57 koen: true... but as reeno did have a different machine he had no chance to notice. Mar 07 18:26:06 :-) Mar 07 18:26:35 koen: *g* Mar 07 18:26:46 florian_kc: stuff like this should be 'opt-in', not 'opt-out' Mar 07 18:28:34 * koen continues to get gcc 4.2 to build Mar 07 18:29:30 jnc: could you put your latest attempt online somewhere (e.g. bugzilla)? Mar 07 18:30:01 * florian_kc -> home, bbl Mar 07 18:30:45 hi Mar 07 18:30:49 hey CoreDump|home Mar 07 18:30:56 hi CoreDump|home Mar 07 18:42:11 koen: BINARY_LOCALE_ARCHS = "" does disable the new code Mar 07 18:42:23 pb_: I changed it to be a list instead of a re Mar 07 18:42:42 a whitespace separated list Mar 07 18:43:27 reenoo: oh right, very good Mar 07 18:44:14 maybe a list of res would be even better but the current arrangement seemed most obvious to me Mar 07 18:44:15 <_guillermo> where is that supposed to go? local.conf? Mar 07 18:44:35 yeah, that's fine Mar 07 18:44:50 I think the original reason I used a re was so that you could say "arm.*" and the like. Mar 07 18:45:04 yeah Mar 07 18:45:18 thus the suggestion to make it a list of res Mar 07 18:45:34 "arm.* mips.*" Mar 07 18:46:09 ARM9ULTRALEETCORE Mar 07 18:46:31 SBC! Mar 07 18:47:55 yup, that would be ideal Mar 07 18:48:59 nah Mar 07 18:49:04 you know what would be ideal? Mar 07 18:49:17 FPGA with 2 PPC cores Mar 07 19:05:01 Does anyone know anything about control of Opie with a keypad? Other than how to compile qte with keypad-mode support? Mar 07 19:05:34 ashak: #opie Mar 07 19:05:53 koen: I asked there this morning, but didn't get any response :( Mar 07 19:07:50 re Mar 07 19:08:13 ashak: well lorn might know, but besides that use the source Mar 07 19:08:59 I think I found something about it earlier, but it was quite confusing. Mar 07 19:12:13 ok, enough of this Mar 07 19:12:25 where is the list of things I need to put in ASSUME_PROVIDED? Mar 07 19:12:58 RP: hey, just to point out, the current patching in the OE .dev branch is not adding int (*get_brightness)(struct backlight_device *bd); and int (*get_power)(struct backlight_device *bd); to struct backlight_properties in include/linux/backlight.h (building for ipaq-pxa270 machine) Mar 07 19:13:42 JustinP: conf/bitbake.conf Mar 07 19:14:35 RP: that is the linux-openzaurus kernel Mar 07 19:14:41 gnome-cups-manager also works: http://handhelds.org/scap/port.2730.png Mar 07 19:15:13 koen: why do you still use rxvt and not gnome terminal? Mar 07 19:16:06 zecke: laziness Mar 07 19:16:28 koen: but the TerminalWidget of Gnome Terminal is so much faster now?! Mar 07 19:18:01 zecke: it still crawls on my dual 2800+ machine Mar 07 19:18:20 and it remaps shortcuts to stupid combos Mar 07 19:18:22 so this persian dude cheated? Mar 07 19:18:28 ctrl-c -> ctrl-shift-c Mar 07 19:18:41 zecke: it's much faster, but still too slow Mar 07 19:18:58 it was supposed to be faster than rxvt and xterm Mar 07 19:19:14 * koen is used to mmx accel'ed eterm Mar 07 19:19:42 damn my body misbehaves... shaking hands and legs Mar 07 19:19:45 speaking of mmx, I still haven't tried iwmmx Mar 07 19:21:56 koen: ah....looks like my ancient local.conf wasn't using += for ASSUME_PROVIDED Mar 07 19:22:18 JustinP: I had the same a while back Mar 07 19:28:48 re Mar 07 19:44:46 florian: wb Mar 07 19:50:13 Aw hell Mar 07 19:50:19 TypeError: __init__() takes at most 3 arguments (4 given) Mar 07 19:50:39 too old bitbake? Mar 07 19:51:01 I'd updated it a few days ago...lots ofupdates Mar 07 19:51:10 just updated it again, 2 updates Mar 07 19:52:05 JustinP: have a installed version of bitbake around? Mar 07 19:52:08 JustinP, do you have some dead meat in /usr/lib/python/...? Mar 07 19:52:26 polyonymous: are you paying those guys? Mar 07 19:52:34 zecke, ;-)))) Mar 07 19:52:56 Yes, to trick you into prepending path, of course. Mar 07 19:53:21 polyonymous: it is prepened already Mar 07 19:53:25 Ah. Mar 07 19:53:40 I haven't updated bitbake recently Mar 07 19:53:42 for ~2 days Mar 07 19:53:56 never touch a running system, you know... Mar 07 19:54:12 right Mar 07 19:54:16 why would I pay those guys then? Mar 07 19:54:27 polyonymous: I might reconsider distributing bitbake only in binary form Mar 07 19:54:37 polyonymous: and have diff's for updates Mar 07 19:54:46 Sounds promising. Mar 07 19:55:09 zecke: ? Mar 07 19:55:09 Especially the binary bit. Mar 07 19:55:16 zecke: is someone stealing bitbake? Mar 07 19:55:19 JustinP: just kidding :) Mar 07 19:55:27 ah Mar 07 19:55:44 It will only work if you restrict any of your users from working on competing build systems Mar 07 19:56:10 JustinP, would you care if someone did? I've recently had my MIT-licensed software "stolen" and distributed binary. They agreed to give credit after I contacted them, though. Mar 07 19:56:12 JustinP: during the last couple of days we had a couple of people falling over an installed version of bitbake Mar 07 19:56:20 But if they didn't I don't think I would care much. Mar 07 19:56:54 right. I write Software because it is fun and I want to share it Mar 07 19:57:12 if some asocial idiots make it proprietary I'm sorry for their soles Mar 07 19:57:12 I wonder what could it be if prepending path doesn't do the trick... Mar 07 19:57:13 yeah, looks like I still had my version installed from gentoo....and I hadn't updated that one in a while Mar 07 19:57:25 if I'd just been using that I'd have been fine I think Mar 07 19:57:28 JustinP: wait Mar 07 19:57:45 Nope, you wouldn't, I think it's too old :) Mar 07 19:57:48 JustinP: do a python -v /path/to/bitbake Mar 07 19:57:55 JustinP: which modules are it loading? Mar 07 19:57:59 polyonymous: no, it updated from svn ;-) Mar 07 19:58:10 zecke: ummm...I already uninstalled the gentoo version... Mar 07 19:58:17 JustinP, ahh, gentoo svn version... Mar 07 19:58:27 JustinP: now we will never know if path prepending works Mar 07 19:58:34 heh Mar 07 19:58:36 JustinP: "but 1.3.2 is the latest in portage, so my bitbake can't be too old!!! OMG WTF SUX0R" Mar 07 19:58:44 talking about gentoo I should restart my glibc on my sharks Mar 07 19:58:48 JustinP: I've been seeing that a lot lately ;) Mar 07 19:58:55 lo Mar 07 19:58:56 l Mar 07 19:59:03 :) Mar 07 19:59:05 koen: I feel for you Mar 07 19:59:18 zecke, well, you can install the older bitbake :) Mar 07 19:59:28 there are l4m3r5 in every distro Mar 07 19:59:49 doesn't that have a z in it somewhere? ;) Mar 07 20:00:11 l4m3r5z? Mar 07 20:00:40 heh, I was thinking more like l4m3rZ or something Mar 07 20:01:46 * CosmicPenguin learns more awk then he believe he should Mar 07 20:02:04 heh Mar 07 20:02:54 hmm, is there a simple way to find out what codes a keypad outputs for each key pressed? Mar 07 20:03:09 ashak: bitbake opie-keyview Mar 07 20:03:18 ashak: ipkg install opie-keyview Mar 07 20:03:28 ashak: also you might want to compile QtE with VNC support Mar 07 20:03:32 yay! http://projects.o-hand.com/matchbox/?p=12 Mar 07 20:03:36 ashak: so your vnc has a mouse at least Mar 07 20:04:16 zecke: I was thinking more along the lines of from within my ssh session.... but vnc session would be cool, how does one do that? Mar 07 20:04:39 ashak: -vnc when configuring QtE Mar 07 20:04:48 doh! Mar 07 20:04:57 ashak: export QWS_DISPLAY=VNC:Fb0 Mar 07 20:05:06 ashak: to display on VNC and Fb0 Mar 07 20:05:14 where does that go? Mar 07 20:05:20 ashak: you could do export QWS_SIZE=1024x768 Mar 07 20:05:26 ashak: export QWS_DISPLAY=VNC Mar 07 20:05:35 ashak: to have one big display Mar 07 20:05:37 Or can I simply set them manually after the phone has booted Mar 07 20:05:54 ashak: in your opie script (/etc/init.d/opie) Mar 07 20:06:04 ashak: just before you start the QtE/Opie/Qtopia apps Mar 07 20:06:22 zecke: ok cool, thanks i'll give that a try Mar 07 20:10:00 is the qemu master around? Mar 07 20:10:12 looks like it should depend on texi2html-native Mar 07 20:10:40 ah yes, that was breakage nr. 4 Mar 07 20:11:21 ah, there we go Mar 07 20:11:41 * JustinP emerged texi2html Mar 07 20:16:20 ashak -> just out of curiosity, which phone do you have? Mar 07 20:18:42 cya, I forgot the power plug in the office rooms :} Mar 07 20:19:09 zecke: doh Mar 07 20:19:54 I should ask my boss for a safety copy of each thing I posess... Mar 07 20:20:30 zecke -> body parts included? Mar 07 20:20:58 awp_: Orange SPV-C500, aka Qtek 8010, I-Mate SP3, Dopod 565, Audiovox SMT5600, Intelleket iQ700 Mar 07 20:21:01 * awp_ could do really well with 4 arms (; Mar 07 20:21:22 ashak -> thanks Mar 07 20:21:54 awp_: just the functional ones Mar 07 20:22:49 ashak -> which one are you running opie on btw? Mar 07 20:23:24 awp_: I have the Orange SPV-C500. But all of those phones are essentially the exact same phone. Mar 07 20:23:49 ashak -> ok thanks Mar 07 20:24:39 awp_: http://wormhole.dsnine.co.uk/~gary/projects/typhoon/ Mar 07 20:25:52 03koen 07org.oe.dev * rbae20f6f... 10/packages/openocd/openocd_svn.bb: openocd: add openocd, a Free and Open On-Chip Debugging, In-System Programming and Boundary-Scan Testing Mar 07 20:26:05 awp_: If you wanted to see a picture :) Mar 07 20:26:58 ashak nice Mar 07 20:27:03 ashak -> you read my mind, thanks, I'm looking for a new phone (; Mar 07 20:28:14 awp_: this one is getting quite old now really. Orange released the C550 a little before I bought this 9 months ago, and they've recently released the C600 Mar 07 20:29:25 woglinde: Heh, thanks, although i've not really done any hard work (with the exception of trying to decipher what these guys tell me and understand the kernel modifications that the guys developing the kernel release) Mar 07 20:30:04 I could never understand why would people want to have pda in the phone. Do you really want your pda and phone run out of battery power at the same time? :) Mar 07 20:30:31 hey zecke Mar 07 20:30:40 re Mar 07 20:30:46 gosh is my desktop loud Mar 07 20:30:48 polyonymous: lol, never really thought of it like that. I simply wanted a project to work on, and Windows Mobile really bugs me. Mar 07 20:31:13 As something to work on - I understand you completely :) Mar 07 20:31:28 ashak -> there is no orange in my country, but I'll search for it's other names, however it's nice to know a phone which runs opie, I doubt you can make phone calls with it at this stage, right? Mar 07 20:31:41 polyonymous: although it's quite useful having all that functionality in one small device. Means you can go out without the need to keep track of your phone AND pda... ;) Mar 07 20:32:37 ashak, well, I have ericsson T39m - old, but I think the last decent PHONE ever produced. It's damn small and bluetooth-enabled, so I can hook it up to pda and I don't need to "track" it really :) Mar 07 20:33:05 awp_: Not just yet, the guys working on the kernel side of things are currently working on talking to the GSM stuff as I understand it, once they have something working I don't see why the 'phone' application for Opie couldn't make calls, as apparently it looks just like a serial modem. Mar 07 20:34:20 ashak -> sweet, thanks for the chat, lesson time, good luck with your project (; Mar 07 20:34:58 njs: how is buildbot handling multiple heads? Mar 07 20:35:22 awp_: No probs, now if you can just tell me how to get the keypad controlling the phone, we'll call it even ;) Mar 07 20:35:30 zecke: with a broadsword and a fishnet Mar 07 20:36:15 handling multiple heads is an ancient problem. hydra had it, but hercules handled them alright... Mar 07 20:36:31 My desktop now looks more like cerberos - exactly three heads ;-) Mar 07 20:40:41 hmmm, how do I know if i've compilted qte with the vnc support... other than trying it? Mar 07 20:41:08 ashak: strings Mar 07 20:41:20 * ashak blinks Mar 07 20:42:11 zecke: strings? Mar 07 20:42:24 strings file | grep -i vnc Mar 07 20:42:25 strings libqte.so.2.3.10 | grep vnc Mar 07 20:42:28 hehe Mar 07 20:42:45 ashak: or use nm or objdump and check for a symbol name VNC or vnc Mar 07 20:42:46 * polyonymous has no idea what case vnc is Mar 07 20:44:23 polyonymous: yeah the part behind the | looks better than mine Mar 07 20:46:31 damn, guess I needed to clean qte before I rebuilt it. Mar 07 20:46:52 ashak: do you have ccache installed? Mar 07 20:47:24 nope... Mar 07 20:47:51 ashak: install it Mar 07 20:47:52 zecke: emerge ccache would probably fix that though Mar 07 20:47:55 zecke: ok Mar 07 20:48:13 zecke, ah, I thought you know ;-) Mar 07 20:48:13 ashak: then recompiling of things like qte, will just fly Mar 07 20:48:39 polyonymous: I Mar 07 20:48:56 polyonymous: I'm pretty sure the symbol looks like QVNCGfx QVNCScreen Mar 07 20:49:25 Ah. Mar 07 20:49:49 zecke: what's the proper way to recompile something? For example, I built opie-image, which obviously built qte... so then after adding -vnc to the qte build file, how do I get it to compile qte again? Mar 07 20:50:16 03koen 07org.oe.dev * ra44a10a7... 10/packages/libmatchbox/libmatchbox_1.8.bb: libmatchbox: update to 1.8 Mar 07 20:50:20 03koen 07org.oe.dev * rfb01cb6b... 10/packages/matchbox-wm/matchbox-wm_1.0.bb: matchbox-wm: add 1.0 Mar 07 20:50:38 rebuild? Mar 07 20:51:08 ashak -> unfortunately I can't help you with the keypad, I'm still learning my way to get linux on the pda, but good luck even though Mar 07 20:51:34 ashak: if you bump the PR of qte Mar 07 20:51:46 ashak: and then do bitbake opie-image, QtE will be rebuild for you Mar 07 20:51:55 zecke: ah ok. Mar 07 20:52:07 ashak: otherwise do bitbake -b qte -cclean and then compile again and then build the image Mar 07 20:52:23 zecke, rebuild won't help? Aren't you keep track of timestamps for dependencies? Mar 07 20:52:36 don't you even. Mar 07 20:52:49 ashak: or you can bitbake qte -cconfigure -f Mar 07 20:53:09 ashak: this will force configuration of QtE and on your next bitbake opie-iamge things should be compiled again Mar 07 20:53:30 polyonymous: in the case of QtE it doesn't matter, it has no PR+1 stamp yet Mar 07 20:53:54 zecke: How do you guys remember all this stuff? ;) Mar 07 20:54:03 I mean, if not bump PR Mar 07 20:54:28 ashak, I don't remember a thing, I've baked my first bit a couple of days ago, I'm just making educated guesses :) Mar 07 20:54:35 ashak: we don't, this is why we need a doc writer! Mar 07 20:55:13 Well, if there was a decent documentation, I don't think I would know about this stuff as much as I do now.. Mar 07 20:55:20 looks like qte is being recompiled this time. Mar 07 20:55:46 polyonymous: I think I could quite agree with that. Mar 07 20:56:08 indeed, zecke is our interactive documentation Mar 07 20:56:17 Always look at the bright side of the lack of documentation :) Mar 07 20:56:35 pb_, personal touch to it :) Mar 07 20:56:50 right! Mar 07 20:56:57 being able to ask a question an often get the exact answer you require is much simpler than reading cryptic documentation ;) Mar 07 20:57:19 ashak, reading cryptic code is much more entertaining on the other hand :) Mar 07 20:57:28 now if I can just get zecke to make this keypad work properly ;) Mar 07 20:57:32 pb_: melvin/zecke your melancholic/depressed bot ready for service Mar 07 20:57:33 polyonymous: heh Mar 07 20:58:39 hmm... since I installed firefox 1.5 I can't help thinking about RAM upgrade... Mar 07 20:58:44 hmm, is it likely that just qte is recompiled and none of the opie applications? Mar 07 20:59:39 damn I'm so stupid... marvin... Mar 07 20:59:56 ashak, even if it's a bug, isn't it a time-saver for you? :) Mar 07 21:00:07 ashak: well, yes for bitbake/OE, yes for this case, generally there could be binary incompabilities Mar 07 21:00:35 zecke: ok, cool, just checking. Mar 07 21:01:32 http://en.wikipedia.org/wiki/Marvin_the_Paranoid_Android <- I love marvin Mar 07 21:01:55 Marvin is okay, but all this adams craze isn't :) Mar 07 21:02:11 YaY, strings opt/QtPalmtop/lib/libqte.so.2.3.10 | grep -i vnc, returns something now, so I guess that worked Mar 07 21:02:14 Before Marvin we've had Eyor, anyway (or what was his name?) Mar 07 21:02:49 eeyore Mar 07 21:02:58 http://pastebin.com/589645 <- just perfect Mar 07 21:03:03 * ashak goes to get his phone, the sd card readr, the usb cable, etc. Mar 07 21:03:11 ~emulate marvin Mar 07 21:03:58 Speaking of emulation and disasters, can one qemulate the whole thing or should one build a native everything to play with. Mar 07 21:04:33 polyonymous: you can qemulate the whole thing right into X Mar 07 21:05:04 is there any zecke on it or do I figure it out myself? :) Mar 07 21:05:33 polyonymous: RP is to blame for this excellent work Mar 07 21:05:37 RP demoed booting an OE built image in qemu at FOSDEM Mar 07 21:05:50 oh damn, I forgot i'm building this stuff on a seperate machine... going to take a while for the rootfs to transfer here Mar 07 21:05:56 ah, so I need to get a time and place machine? :) Mar 07 21:06:04 polyonymous: http://www.rpsys.net/openzaurus/temp/qemuarm-poky1.png Mar 07 21:06:39 It doesn't explain how to approach the whole thing, though :) Mar 07 21:06:47 or http://www.rpsys.net/openzaurus/temp/qemu3.png if you are a console jockey Mar 07 21:07:34 Console is good, but I see no need to play with it at this point, everything's clear with console applications for me now. Mar 07 21:07:56 Is there any .html/.txt about it? Or do I have to be content with .png?:) Mar 07 21:08:17 hey obergix[home] Mar 07 21:08:41 hi koen Mar 07 21:08:45 polyonymous: well it is just another machine :) Mar 07 21:08:49 polyonymous: qemu can Mar 07 21:08:56 It's okay with me if there is no, it's just that I was hoping to save some time. Mar 07 21:09:00 polyonymous: only emulate the integrator board Mar 07 21:09:17 polyonymous: so searching in OE for integrator might be good? Mar 07 21:09:23 Thank you. Mar 07 21:09:40 MACHINE=qemuarm iirc Mar 07 21:09:43 Basically, it means completely different build or does it? Mar 07 21:09:50 Aha, that's what I meant. Mar 07 21:10:04 Thank you. Mar 07 21:23:48 WooHoo, vnc to the phone :) Mar 07 21:24:18 hehe Mar 07 21:24:40 hmm, although it's still dispaying the touch anywhere screen but the actual phone has changed. Mar 07 21:24:50 I should write a book the 1st and 2nd quirk on Linux, Buildsystems and QtE/Qtopia Mar 07 21:26:20 You normally either know the stuff you're talking about or writing a book on it. Mar 07 21:26:29 this is odd, I can click buttons on whatever's being displayed on the phone screen from vnc, but the vnc session is not updating the image Mar 07 21:26:46 ashak: slow network? Mar 07 21:26:57 zecke: usb ethernet Mar 07 21:27:04 the button clicks are almost instant Mar 07 21:27:29 ashak: hmm where did you set the QWS_DISPLAY? Mar 07 21:27:29 vnc server is too lazy to check the screen? Mar 07 21:27:42 in /etc/init.d/opie Mar 07 21:27:46 okay Mar 07 21:27:52 dunno then Mar 07 21:28:07 polyonymous: the Qt VNC implementation gets the changes 'pushed' Mar 07 21:28:24 Yes, after you asked your question I figured my question was silly. Mar 07 21:33:20 * chouimat is away: Out in the scary outside world Mar 07 21:34:36 http://pastebin.com/589729 Mar 07 21:34:44 zecke: what was the app for seeing keypresses? Mar 07 21:34:55 zecke: ah, opie-keyview Mar 07 21:34:58 ashak: opie-keyview Mar 07 21:35:14 ashak: it is an applet (virtual keyboard) Mar 07 21:45:26 hmm, keyview opens (I get 5 boxes, labeled unicode keycodes modifiers isPress autoRepeat. Mar 07 21:45:41 But if I press any keys, I get no sort of output. Mar 07 21:46:45 ashak: is the keyboard working? Mar 07 21:47:05 ashak: this applet should see every key pressed (well I have not tried with the keypad mode...) Mar 07 21:47:10 zecke: when in the console, if I press keys, I get some form of output yes. Mar 07 21:47:45 ashak: I will go to bed now :) Mar 07 21:47:53 'night zecke Mar 07 21:47:59 ashak: find out if any keys are received Mar 07 21:47:59 zecke: heh, ok, night, thanks for the helo Mar 07 21:48:14 ashak: ps xua | grep qpe (save the pid) Mar 07 21:48:32 ashak: ls -la /proc/THE_PID/fd Mar 07 21:48:39 ashak: and see which devices it has opened Mar 07 21:48:53 ashak: what is hapening when you enter a key over VNC? Mar 07 21:50:09 zecke: in /proc/1815/fd/ there are 0-13, 16. Mar 07 21:50:21 ashak: these are symlinks :) Mar 07 21:50:26 ashak: is /dev/tty opened? Mar 07 21:50:33 zecke: Nothing output in keyview when I press a key in vnc, Mar 07 21:51:22 zecke: Ah, erm.... no /dev/tty Mar 07 21:51:46 you need to debug yourself somehow Mar 07 21:51:48 zecke: some /dev/null, some sockets, /dev/fb/0, /dev/vc/0 Mar 07 21:52:08 I want to get past "How to write Effective Use Cases" page 2 Mar 07 21:52:19 (reading) Mar 07 22:14:30 OMFG....this is going to take forever.... Mar 07 22:16:50 how can I limit the generated locales again....? Mar 07 22:17:05 * JustinP really only needs en_US (AFAHK) Mar 07 22:19:55 JustinP: I'm afraid you can't atm Mar 07 22:20:20 ah... Mar 07 22:20:34 yay...bonary locale generation is so *fast* why would I want to anyway?? Mar 07 22:20:40 s/bonary/binary/ Mar 07 22:20:57 well, you can disable the binary locale generation by setting BINARY_LOCALE_ARCHES = "" Mar 07 22:21:04 looks like around a minute per locale.... Mar 07 22:21:15 heh Mar 07 22:21:19 10 seconds here Mar 07 22:21:23 * polyonymous thinks it's absolutely unimportant to have anything but en_US in initial setup. As well as sokoban, etc. :) Mar 07 22:21:29 < 10 seconds Mar 07 22:21:38 en_US == C, more or less, so you might as well not even install any locales if that's all you want. Mar 07 22:22:30 true. Mar 07 22:22:34 I don't care about *installing* locales, I just don't want to build them... Mar 07 22:22:51 then do like reenoo said and set BINARY_LOCALE_ARCHES to the empty string. Mar 07 22:22:51 Well, it's worse than installing - it's configuring :) Mar 07 22:24:01 * reenoo heads off to doing some project work Mar 07 22:24:01 pb_: it will still build them....just not as binary Mar 07 22:24:16 sure, but building them as source takes negligible time Mar 07 22:24:19 certainly not a minute each Mar 07 22:24:21 my poor little 700Mhz Celeron is being obsoleted further now :-( Mar 07 22:24:37 it's not negligible, I'd say, but yeah, it takes much less Mar 07 22:28:08 I guess it might not be completely negligible on a 700MHz machine, yeah, but it's not as if you build glibc every day. Mar 07 22:29:10 If you wanted to submit a patch to limit the list of generated locales, or even turn them off altogether, that would be fine as well. Mar 07 22:32:08 heh Mar 07 22:41:58 can anone try to compile diet-x11 Mar 07 22:42:43 it's giving errors in patching : errordb-keysymdb-path-fix.patch is problematic Mar 07 22:59:21 hop, un goto de plus. Mar 07 22:59:41 oups, sorry. Mar 07 23:13:16 * champ is now accepting zaurii donations. Mar 07 23:36:38 .me keeps waiting for locale packages to be built.... Mar 08 01:26:50 hmmm - is it me, or it is strange that the cross compiler isn't making profiling data files? Mar 08 01:34:01 Here's a tip boys and girls - don't do ccache + gcov Mar 08 01:34:03 argh! Mar 08 01:34:44 The good news is I figured out why the data files aren't in there Mar 08 03:08:12 * france is back (gone 39:46:52) Mar 08 03:38:17 03mwester 07org.oe.dev * r942da6c3... 10/packages/nslu2-binary-only/unslung-rootfs-2.3r63/ (NOTES README): unslung-rootfs: Updated README and NOTES Mar 08 04:34:20 wow ... monotone has sped up quite a bit Mar 08 04:49:26 emte: how so? Mar 08 05:08:52 jnc, well it now takes ~20 seconds/rev on my system Mar 08 05:09:43 monotone 0.24? Mar 08 05:11:29 0.24-1+b1 according to debians revs anyway Mar 08 05:12:36 it included a few boost changes tho Mar 08 05:12:36 wich i am guessing the +b1 references Mar 08 06:15:48 * france is away: Away Mar 08 07:00:16 emte: "+b1" means the it got automatically rebuilt by one of the debian build daemons, in this case because the libraries it depends on got rebuilt with a newer g++, so it had to be rebuilt as well. Mar 08 07:00:19 Mar 08 07:01:45 the libraries got upgraded a full version Mar 08 07:02:27 i am now curious if the performance cahnge was monotone or the differenece in boost libraries Mar 08 07:02:32 change* Mar 08 07:03:31 anyway i am happy, 20seconds a rev is a nice change from 2min a rev Mar 08 07:04:19 interesting, on c3100, /dev/mtdblock2 is the D+B recovery linux rootfs Mar 08 07:05:35 is that before or after sharps wonky block allocation (now i wonder if taht exists on the c3100) Mar 08 07:06:27 i'm in the D+B recovery linux boot looking at it Mar 08 07:06:55 i see some QtPalmtop and such in a mount of /dev/mtdblock3 Mar 08 07:07:46 it would appear that the mtd flash on C3100 hardware is segregated into four areas, 0 1 2 3 Mar 08 07:08:02 areas 2 and 3 are ext2 filesystems i think Mar 08 07:08:18 not sure about areas 0 and 1 Mar 08 07:09:16 i vaguly remember mention of a secton of specially assigned sharp blocks part way in between 0 and 1 Mar 08 07:09:30 on some of the sharp devices Mar 08 07:11:32 nevermind the increase in flash space, the c3100 looks more like what i wanted originally Mar 08 07:11:35 :) Mar 08 07:12:00 yeah its a nice looking device Mar 08 07:12:01 kieellll kill killll .. the white iPoddish look Mar 08 07:12:08 bit pricey tho Mar 08 07:12:12 uh huh Mar 08 07:12:20 i don't like blowing that kind of money on stuff Mar 08 07:12:20 tahts noting an airbrush wont fix :) Mar 08 07:12:22 3100 or 3200 ? Mar 08 07:12:28 3100 and 3000 here Mar 08 07:12:33 not going to bother with the 3200 Mar 08 07:13:09 if i'd had the money i guess i would have preferred the 3100 over my 1000 Mar 08 07:13:17 sure Mar 08 07:13:25 i'm selling my C3000 i think Mar 08 07:13:53 the C3000 had a couple of dead pixels Mar 08 07:13:55 if i had the money ... i could pay my tuition Mar 08 07:14:15 yeah. need to save for school soon Mar 08 07:15:10 my schooling days are nearly over and i'm paid up :) Mar 08 07:15:25 ade|desk: nice :) Mar 08 07:15:25 the c31/32 is a decent alternative to a sub-notebook tho Mar 08 07:15:37 emte: hah, i am still not convinced about that Mar 08 07:15:41 just my part time masters still on going Mar 08 07:15:50 subnotebook would have usb 2.0, and at least a PCMCIA slot Mar 08 07:15:57 indeed Mar 08 07:16:00 for wifi 802.11G/A that works Mar 08 07:16:11 possibly Mar 08 07:16:12 or cheaper access to working CDMA/GPRS card Mar 08 07:16:51 * emte shrugges Mar 08 07:17:04 dont really see the use for a gprs card here Mar 08 07:17:09 i think a pxa270@ 600+ MHz 800x600 would be a good start for a subnotebook Mar 08 07:17:41 if i had the contract the koen is on i'd go GPRS too Mar 08 07:17:54 s/the/that Mar 08 07:18:26 with an enfora card? Mar 08 07:19:01 GPRS is probably more useful in europe Mar 08 07:19:25 the data bill would kill you here Mar 08 07:19:28 Cingular/AT&T is the largest network carrier in the USA Mar 08 07:19:30 all GSM Mar 08 07:19:31 would be good Mar 08 07:19:49 lowest price i've saw is 5cents/kb Mar 08 07:20:00 unlimited plans for 80bucks / month Mar 08 07:20:01 hmm GPRS isn't so big over there Mar 08 07:20:26 yeah canada is a bit slow on the free beer packages Mar 08 07:20:33 ah Mar 08 07:20:59 UK sucks for GPRS Mar 08 07:22:00 leaving bluetooth out of the cxx00 series was a dumb move in any country Mar 08 07:22:21 thats another thing that is fairly rare here Mar 08 07:22:23 * jnc larts sharp Mar 08 07:22:38 why is bluetooth not used much in japan ? Mar 08 07:22:52 about the only bt you see are cell earpeices Mar 08 07:23:03 it's not all that common in the states, but it is super easy to implement Mar 08 07:23:20 there's clearly space inside the cxx00 shell to have a bt xmitter Mar 08 07:26:32 maybe i'll have a new device for us to play around with ... Mar 08 07:26:52 will see if the contract pans out Mar 08 07:27:37 build is at kernel ... thats good Mar 08 07:28:08 hopefully the new wpa_supplicant will compile for me Mar 08 07:28:27 i'd like to see if it plays well on my ipaq Mar 08 07:30:15 if you have trouble with wpa_supplicant please let me know Mar 08 07:30:24 i know the config file and init scripts to be broken Mar 08 07:30:40 i am not sure if it will solve my issue or not Mar 08 07:30:46 i am hoping Mar 08 07:30:46 but it should at least compile and install, and you can run the wpa_supplicant by hand Mar 08 07:31:07 i actually need to figure out what WISE uses Mar 08 07:31:11 i used that to auth against WPA2 CCMP WPA-PSK on my router Mar 08 07:31:12 :) Mar 08 07:31:57 i have to do a bunch of extra stuff by hand anyway Mar 08 07:32:16 i've a WCF56G Mar 08 07:32:24 which has the nice issue of no MAC Mar 08 07:32:32 is that the agere one? Mar 08 07:32:33 by default Mar 08 07:32:39 apperently Mar 08 07:32:52 works with the wavelan driver Mar 08 07:33:01 wcf56g does not result in anything Mar 08 07:33:04 you meant 54g? Mar 08 07:33:13 whoops, yeah Mar 08 07:33:21 that's the linksys G-compatible one Mar 08 07:33:32 i am very happy to hear someone else using one Mar 08 07:33:34 how do you like it? Mar 08 07:33:37 its the cisco-linksys G one Mar 08 07:33:58 not sure if it changed when cisco got put on the case or not Mar 08 07:34:29 works fine so far with the exception of trying to make it work on a WISE network in familiar Mar 08 07:34:42 what is a WISE network Mar 08 07:35:22 i truly dont know, thats just what the techs keep telling me when i ask for the network key Mar 08 07:35:34 i am guessing its some windows acronym Mar 08 07:36:14 Windows, I Shit-you-not-has-become-involved-in Everything. Mar 08 07:36:40 lol Mar 08 07:37:01 its harder to lose lawsuits that way Mar 08 07:42:09 ah fk ... Mar 08 07:42:22 * emte kicks myself in the head Mar 08 07:42:30 i forgot to patch fakeroot :( Mar 08 07:50:35 Oh my. Mar 08 07:51:22 Even with everything going on with oe & familiar, I was able to build meta-sdk Mar 08 07:51:39 for familiar-0.8.3 and ipaq 3600 Mar 08 07:51:59 Got to try to do that again :) Mar 08 07:53:08 lol Mar 08 07:53:21 unless you have my problem it should build fine Mar 08 07:53:34 Trust me, I had my share of problems Mar 08 07:54:17 I've been trying to build it for the past week. And tell you the truth, I'm not sure why it compiled this time :) Mar 08 07:56:09 sleep to try to make this patch work ... Mar 08 07:56:15 sleep i guess Mar 08 07:56:21 wife says so Mar 08 07:56:22 :) Mar 08 07:56:42 Tell her that you have priorities Mar 08 07:57:05 "The fate of the world depends from this patch" Mar 08 07:58:03 my priority is to sleep in the same bed with her and not make her mad :) Mar 08 07:59:02 Weak :) The code is an asteroid, you patch is a nuclear explosive and you are Bruce Willis. Tell her that, I'm sure she will understand :) Mar 08 07:59:27 lol Mar 08 07:59:49 actually the patch is fine, its the test taht needs work Mar 08 08:00:04 as usually Mar 08 08:00:16 i'll have to look that the way jnc 's gcc test works again Mar 08 08:00:30 see if i can understand it enough Mar 08 08:01:22 Cockroach-: you can avoid the word patch Mar 08 08:01:29 substitute the word "Multipass" Mar 08 08:03:18 lol Mar 08 08:03:29 you been watching 5th element again jnc ? Mar 08 08:03:51 "lulu multipass" Mar 08 08:08:13 like, how can you not heart jovovich Mar 08 08:08:23 lol Mar 08 08:08:24 i heart her, times ten. Mar 08 08:08:36 just wasnt sure if thats what you were watching Mar 08 08:09:42 i wish bitbake had more docs .... Mar 08 08:10:05 okay, i put a new flag in local.conf Mar 08 08:10:19 how do i export it when the conf is parsed ... Mar 08 08:10:33 something to do with one of the classes but which one ... Mar 08 08:13:43 AHA! Mar 08 08:25:29 Hmm, I wonder if there is any information about using meta-sdk Mar 08 08:27:14 good morning all Mar 08 08:27:27 Morning Mar 08 08:27:28 morning Mar 08 08:41:33 I hate to bother with my beginner problems, but I've had my share of banging the head to the wall :) as koen suggested, I built the meta-sdk, but I'm not sure if it will help me as much as I would need it to. Mar 08 08:42:49 Since if I would like to create new software from scratch and be sure that it works with the same library versions as available in the target hardware, I would first need the x86 versions. meta-sdk did seem to build the correct arm libraries Mar 08 08:43:06 so, should I also build the meta-sdk with MACHINE set to "x86" Mar 08 08:55:18 morning Mar 08 08:56:12 hey XorA Mar 08 08:56:22 koen: you tried that new evince I put in? Mar 08 08:56:27 XorA: not yet Mar 08 08:56:44 koen: to me it seems significantly faster Mar 08 08:57:03 pdfs gone to being readable, and it no longer eats all my RAM Mar 08 08:57:15 cool Mar 08 08:57:28 bitbake r390 seems to be very, very broken Mar 08 08:57:36 koen: ut oh, what broke? Mar 08 08:57:51 better_exec() by the looks of it Mar 08 08:58:04 and it raise execptions while parsing Mar 08 08:59:03 reminds me, must find out why ALL gpe apps segfault on my machine now, most annoying when I was trying to use sketchbook for ADND Mar 08 08:59:04 so you get zillions of "foo only builds for ARCH bar" Mar 08 09:00:59 * koen moves back to r384 Mar 08 09:01:27 koen: Im at 384 , think Ill avoid an upgrade Mar 08 09:01:40 I'm at r329 Mar 08 09:02:21 also koen got any ideas why gconf always gets built as well as gconf-dbus, is totally screws up builds Mar 08 09:02:23 ibot: botmail for zecke: bitbake r390 doesn't work with the current OE repo Mar 08 09:02:46 XorA: no idea, but over here it builds dbus 0.23.1 and 0.61 as well Mar 08 09:03:03 so I locally dropped all dbus versions execpt 0.61 Mar 08 09:04:05 koen: only builds 0.61 here Mar 08 09:06:20 you could try doing 'monotone drop gnome/gconf_2*' Mar 08 09:07:03 koen: I let it build. remove the ipkg and shlibs files then the stamp files prevent it building again Mar 08 09:07:18 koen: Im used to this as ever mozilla based build requires this before you build another Mar 08 09:09:09 http://www.oesf.org/forums/index.php?showtopic=18104&hl= - kexec for 2.4 Mar 08 09:09:28 Bernado just didnt listen when told to use bin-config did he? Mar 08 09:10:54 Bernardo|away: your patch would be much simpler if you just inherit bin-config in your .bb that takes care of copying xxx-config to the right dir Mar 08 09:11:53 hi Mar 08 09:13:10 XorA: it could also have to with the fact that gconf-dbus has '0.0' as PV Mar 08 09:13:22 XorA: I suspect that should be 2.12 or something Mar 08 09:14:31 2.12+svn${SRCDATE}? Mar 08 09:18:01 something like that Mar 08 09:26:15 http://handhelds.org/hypermail/familiar-dev/current/0751.html Mar 08 09:26:22 now who's FUD'ing? Mar 08 09:30:31 XorA: building evince right now Mar 08 09:31:34 koen: Im quite annoyed by that message Mar 08 09:31:45 koen: it seems the debianites are taking control Mar 08 09:32:03 koen: and its not even my project Mar 08 09:34:04 sweet Mar 08 09:34:23 I can now print photo's from my handhelds to my CP-400 Mar 08 09:34:28 koen: wicked Mar 08 09:34:54 over wifi :) Mar 08 09:34:57 2/cr Mar 08 09:35:51 koen: any luck with an EABU build yet? Mar 08 09:36:03 no, still missing glibc Mar 08 09:36:12 and I'm too lazy to try uclibc Mar 08 09:36:42 koen: oh, message you posted a couple of days ago looked like a cvs glibc would EABI Mar 08 09:36:59 teh cvs version doesn't build Mar 08 09:37:07 koen: fairy nuff :-) Mar 08 09:37:15 gcc4 + -Os is working nicely, though Mar 08 09:37:31 * XorA should move to angstrom soon Mar 08 09:37:51 ut oh, busybox upgrade Mar 08 09:38:08 koen: that familiar-dev thread is nice Mar 08 09:38:29 florian wrote something nice: http://handhelds.org/hypermail/familiar-dev/current/0750.html Mar 08 09:42:51 there are a lot of people getting extremely pissed of at hh.org in general Mar 08 09:44:26 still no objections to my RFC Mar 08 09:47:54 time for Recognizable 0.8.3 :-) Mar 08 09:57:45 morning all Mar 08 09:58:10 hi lrg Mar 08 10:01:55 good morning Mar 08 10:05:50 hey lrg & Bernardo Mar 08 10:06:51 <_guillermo> morning all Mar 08 10:07:01 koen: thanks for the tip yesterday on neon, helped me to get it and subversion fixed Mar 08 10:13:48 shit, blew away my / :-(I Mar 08 10:14:09 <_guillermo> what is the main source of this familiar-OE conflict? I'm missing part of the movie Mar 08 10:15:42 _guillermo: the old familiar which hasn't done anything in ~2.5 years decided that OE is crap Mar 08 10:15:45 XorA: rm -rf /? Mar 08 10:15:59 _guillermo: and those are the same people with root access on handhelds.org Mar 08 10:16:03 Bernardo: no, Z crashed on ipkg upgrade and has left ext2 needing a fsck Mar 08 10:16:16 Bernardo: a fsck it cant get to because ext2 is fucked Mar 08 10:16:39 XorA: card reader + linuxbox? Mar 08 10:16:48 koen: yup, have to be Mar 08 10:16:49 <_guillermo> koen: that is pretty sad Mar 08 10:16:58 koen: I must buy me own SD reader for work Mar 08 10:17:06 the one in my lappy isnt recognised in linux Mar 08 10:19:51 sd readers are dirt cheap nowadays Mar 08 10:21:47 I usually bring with me a sd reader Mar 08 10:22:00 <_guillermo> is ansgtrom usbale right now? Mar 08 10:22:01 I had that kind of trouble when I had a collie Mar 08 10:22:10 <_guillermo> s/usbale/usable Mar 08 10:22:38 Bernardo: my Sd reader at home is built into my shuttle Mar 08 10:23:26 _guillermo: it builds and installs, but it is missing the tweaks OZ and familiar have for certain packages Mar 08 10:24:35 and the ABI isn't finalized yet Mar 08 10:25:15 * XorA siffers #1503 again Mar 08 10:27:49 hi Mar 08 10:28:22 hi mickey|office Mar 08 10:28:44 hey mickey|office Mar 08 10:28:50 did we lament about the SL-C3200 yet? :D Mar 08 10:29:00 mickey|office: not yet Mar 08 10:29:05 is it just a software upgrade? Mar 08 10:29:20 mickey|office: do you know someone @trisoft? Mar 08 10:29:28 hrw|work: yes, i do Mar 08 10:29:49 XorA: it's a hardware upgrade. Sl-C3200 = SL-C3100 with s/4GB/6GB hd/ Mar 08 10:30:16 mickey|office: oh, they ran out of 4GB in other words Mar 08 10:30:27 XorA: right Mar 08 10:30:42 better than a kick in the nuts I suppose Mar 08 10:30:54 mickey|office: are you building the oz images with http://ewi546.ewi.utwente.nl/tmp/build-image-from-feeds.diff ? Mar 08 10:31:05 ya, well, the good news is they're still making zaurii Mar 08 10:31:30 hopefully with the death of Ipaq they will become more popular Mar 08 10:31:37 I'm curious who will get c3200 first - trisoft or users... in other words: who will tell us first what is dogname and pcmcia info about uDrive Mar 08 10:31:39 koen: if it's not in branch i don't. what's thins change doing? Mar 08 10:31:44 s/thins/this/ Mar 08 10:31:55 mickey|office: assemble the images from the feeds instead of deploy/ipk Mar 08 10:32:14 koen: hmm.... which feeds? there are no feeds ? Mar 08 10:32:27 mickey|office: the feeds specified in oz354.conf Mar 08 10:33:03 hmm. i can't quite follow. so how's the workflow supposed to be then? do a world build, upload, and then build images? Mar 08 10:33:13 mickey|office: yes Mar 08 10:33:22 eek Mar 08 10:33:27 mickey|office: it's nicer from a QA POV Mar 08 10:33:48 mickey|office: since you already have tested the feeds with the RC images, right? Mar 08 10:34:10 mickey|office: actually i didn't. /me looks to hrw Mar 08 10:34:20 ~lart windows backup solutions Mar 08 10:34:21 * ibot makes a balloon animal out of windows backup solutions Mar 08 10:34:40 * hrw|work whistles innocently... Mar 08 10:34:44 mickey|office: someone with OE/python skills should make that patch conditional so we can put it in .dev Mar 08 10:34:57 what the hell do you do when you find the backup software that legacy has left you doesnt actually appear to be able to restore tapes Mar 08 10:35:10 XorA: pray Mar 08 10:35:22 koen: it sounds like a good idea, but it needs to integrate with a script to build for multiple machines or else it won't work Mar 08 10:35:28 koen: we already suffered one office fire, you think IT guy at the time would learn Mar 08 10:35:41 give tapes for not-liked child? Mar 08 10:35:41 koen: I have 200+ backup tapes I have no idea how to read Mar 08 10:36:00 koen: and cant afford Computer Associates technical support :-( Mar 08 10:36:14 mickey|office: no need to Mar 08 10:36:39 XorA: cry Mar 08 10:36:44 mickey|office: see http://pastebin.com/590519 Mar 08 10:37:19 mickey|office: that patch worked great when I built the last incarnation of Familiar Mar 08 10:38:04 XorA: what is the backup software? winbackup? Mar 08 10:38:20 koen: ok, then how are the mutiple ipk for different machines are supposed to come from? Mar 08 10:38:53 Bernardo: winbackup for some, Brightstor for others Mar 08 10:39:44 winbackup is tricky, that crap would say it had backedup successfully and the tapes would be 100% unreadable Mar 08 10:39:46 mickey|office: you start a build-run for all images, do a world build, split the resulting feeds usig http://ewi546.ewi.utwente.nl/tmp/viewmtn/revision.psp?id=f3e9a55fd47b949228a6bb81b0b349fd683df1a9 during an RC period. When it's release time, you activate the patch and everything works as expected :) Mar 08 10:40:46 * XorA is very tempted to connect the tape drives to linux and find some decent backup software Mar 08 10:40:47 koen: ah, i'm getting an idea about this release procedure. sounds pretty neat. Mar 08 10:41:21 <_guillermo> I don't see where meta-gpe is being used anymore, is its use deprecated? Mar 08 10:42:40 XorA: good luck Mar 08 10:42:59 I once had the same problem, and never got data back from one of those tapes Mar 08 10:43:16 Bernardo: well at least I am pro-active, I havent burned the office down yet Mar 08 10:43:18 mickey|office: it makes sure the feeds are sane and tested before the release Mar 08 10:43:40 XorA: yep, it's best to find that out before they are needed than after Mar 08 10:44:10 damn, the svn and neon ipks I send my friend with the collie seem corrupted... :( Mar 08 10:44:39 Bernardo: btw you svn patch, why does it not inherit binconfig to automatically stage neon-config Mar 08 10:45:17 because binconfig copies the neon-config to staging-bindir Mar 08 10:45:51 that was fixed by pb ages ago, wtf has happened Mar 08 10:50:52 no idea, but svn has to find it in ${STAGING_DIR}/${HOST_SYS}/bin/ so it then looks for the libs in ${STAGING_LIBDIR} Mar 08 10:51:10 if it isn't there, it assumes there is no neon support Mar 08 10:51:52 that sounds like someone accidently regressed binconfig Mar 08 10:52:40 because it used to have magic to make -native and ${ARCH} builds to end up in different dirs Mar 08 10:54:01 this is on the oz354 branch Mar 08 10:54:08 maybe it is fixed in the dev branch Mar 08 10:54:14 Bernardo: I just looked in dev, and that is missing in there as well Mar 08 10:54:59 Bernardo: I only remeber because its the only bit of OE build side that I ever submitted a patch for Mar 08 10:55:01 then there's the stupid part of subversion's autoconf Mar 08 10:56:01 looking for neon includes in /usr/include/neon instead of ${STAGING_INCDIR}/neon Mar 08 10:56:12 XorA: that patch had to be reverted as it turned out to make more problems than it solved. Mar 08 10:56:33 XorA: it's using another method to differenciate between -config and -config-native nowadays. Mar 08 10:56:51 XorA: have a look at all the E packages, they rely heavily on that mechanism Mar 08 10:57:16 mickey|office: fair enough Mar 08 10:58:11 mickey|office: totally missed that sometime then Mar 08 10:59:01 XorA: it was during the days when JustinP and me started to finally get somewhere with the E stuff. Mar 08 10:59:16 which needs all those -native packages Mar 08 10:59:22 (unfortunately) Mar 08 10:59:25 btw, E doesn't build anymore Mar 08 10:59:45 mickey|office: Ill have to look out for that when playing with -native then, I assume lots of --with-xx-config=blah-config-native is needed Mar 08 10:59:45 undefined reference to libiconv_bla **** ENDING LOGGING AT Wed Mar 08 10:59:56 2006