**** BEGIN LOGGING AT Mon Jun 27 23:59:56 2005 Jun 28 00:55:48 morning all Jun 28 01:00:10 morning Jun 28 01:06:48 lol how many times have you typed 'shitdown -h now' ? Jun 28 01:06:55 morning? afternoon ! Jun 28 01:07:11 shitdown? Jun 28 01:07:26 rather than shutdown Jun 28 01:07:34 hey hrw|work Jun 28 01:11:47 heh.. my oz 3.5.3 builddir broke ;( Jun 28 01:16:18 good morning all Jun 28 01:17:08 hi koen Jun 28 01:18:05 hey hrw|work Jun 28 01:18:25 hey koen Jun 28 01:22:03 hey gints|wrk Jun 28 01:32:51 morning all Jun 28 01:32:59 hey RP Jun 28 01:34:00 koen: There's a proper tslib fix in xserver cvs now btw thanks to mallum Jun 28 01:35:48 RP: old new ;) Jun 28 01:36:39 RP: I have mallum on jabber Jun 28 01:37:20 koen: fair enough :) Jun 28 01:39:41 morning Jun 28 01:39:48 hey BigAl Jun 28 01:48:57 hi BigAl Jun 28 01:55:50 * BigAl hates colds. Jun 28 02:01:31 I think I have one about to start as well... Jun 28 02:03:07 * ade|desk joins the club Jun 28 02:05:40 RP: sysfs works now ok Jun 28 02:06:00 RP: I did the same as in http://pastebin.ca/15541 and all works **** BEGIN LOGGING AT Tue Jun 28 10:26:28 2005 Jun 28 10:50:29 * mr_claus is away: dinner :)))) Jun 28 10:53:01 Could someone explain me what do this mean in the kernel startup messages? Jun 28 10:53:12 Scanning device for bad blocks Jun 28 10:53:13 Bad eraseblock 5540 at 0x05690000 Jun 28 10:53:13 Bad eraseblock 5740 at 0x059b0000 Jun 28 10:53:31 sifred bad flashspeicher? Jun 28 10:53:41 ups storage Jun 28 10:54:01 woglinde: Could this be a unrecoverable error? I mean, a physical damage? Jun 28 10:54:14 sirfred I think yes Jun 28 10:54:26 :-( Jun 28 10:56:13 SirFred: it just means that your flash has some invalid blocks. Jun 28 10:56:13 pb_: yeah, something in libstdc++ Jun 28 10:56:22 pb_: I figure that andersee will catch it eventually Jun 28 10:56:23 SirFred: that's quite normal, and no cause for alarm Jun 28 10:56:29 CosmicPenguin: righto Jun 28 10:56:32 pb_: Unrecoverable? Jun 28 10:56:39 pb_: Physical damage, you mean. Jun 28 10:56:46 pb oh okay Jun 28 10:57:30 SirFred: "unrecoverable" in the sense that there is nothing you can do about it, yes. Jun 28 10:57:39 but, like I say, it is no cause for alarm. Jun 28 10:58:14 pb_: Well, thanks. Jun 28 10:59:13 SirFred: almost all NAND flash chips have a few invalid blocks. Jun 28 10:59:36 pb_: I suppose two bad blocks is not a great amount of size, is it? Jun 28 10:59:40 indeed not. Jun 28 10:59:48 jffs2_get_inode_nodes(): Data CRC failed on node at 0x009cb394: Read 0x5f316d2e, calculated 0x77661043 Jun 28 10:59:57 one block is 16k, so you have lost a total of 32768 bytes. Jun 28 10:59:59 And what about that one? Jun 28 11:00:27 That just indicates that a block of flash contains some scrambled jffs2 data, probably because you rebooted while there was data in the writebuffer. Jun 28 11:00:36 Again, that's harmless: jffs2 will eventually recycle that block. Jun 28 11:01:04 pb_: Great. Jun 28 11:04:17 This is very strange. I think that something on my machine gots corrupted. Jun 28 11:04:55 The fact is that qte is trying to allocate room for a pixmap with geometry 38966x38916 Jun 28 11:05:23 I was suspecting that perhaps something got corrupted. Jun 28 11:20:13 woglinde: http://www.qtopia.net/modules/newbb_plus/viewtopic.php?topic_id=48&forum=5 Jun 28 11:20:19 woglinde: thanks for that URI Jun 28 11:20:32 woglinde: bitbake --update-opie-to-qtopia opie-image Jun 28 11:20:34 new bitibake plus? Jun 28 11:20:42 woglinde: would be the right answer Jun 28 11:20:58 woglinde: or Morons, but time has changed since Sadam is in heaven Jun 28 11:25:08 cu Jun 28 11:26:43 hi Jun 28 11:34:00 172f4e97f70e08af3bea06d5ea36a1dfb6724fdd is the monotone key, right? =p Jun 28 11:36:12 is ewi546.ewi.utwente.nl still correct? Jun 28 11:38:07 03mickeyl 07 * r1.3639.4.5 10openembedded/packages/opie-taskbar/opie-taskbar/qpe.conf: opie-taskbar: adjust qpe.conf also for collie Jun 28 11:38:09 03mickeyl 07 * r1.3639.4.4 10openembedded/packages/ (8 files in 7 dirs): opie-taskbar/qpe.conf: Bye bye QPF, hello TTF Jun 28 11:38:11 03mickeyl 07 * r1.3639.4.3 10openembedded/packages/opie-ttf-support/ (files/fontdir opie-ttf-support_1.0.0.bb): add opie-ttf-support metapackage containing a fontdir. to do it the 'right way', we need a script that updates qtfontdir for TTF Jun 28 11:38:14 03mickeyl 07 * r1.3639.4.2 10openembedded/packages/python/python-pyxml_0.8.4.bb: upgrade python-pyxml to 0.8.4 Jun 28 11:38:18 03mickeyl 07 * r1.3639.4.1 10openembedded/packages/python/ (python_2.4.1.bb python-2.4.1-manifest.inc): add _strptime to python-datetime and adjust dependencies Jun 28 11:38:36 phrw Jun 28 11:38:41 heh Jun 28 11:38:51 mickeyl: cheers Jun 28 11:39:03 first step to eliminate qpf from the planet Jun 28 11:39:10 now we have a size problem though Jun 28 11:39:14 * koen get's reminded of the "corefonfs ghetto" remark Jun 28 11:39:21 only 300K left on collie rootfs Jun 28 11:39:31 need to remove something Jun 28 11:39:36 :/ Jun 28 11:39:38 640k is enough for everybody Jun 28 11:39:43 heh Jun 28 11:39:46 jffs2 != everybody Jun 28 11:39:57 mickeyl: hmm Jun 28 11:40:04 mickeyl: I somehow don't like that Jun 28 11:40:52 mickeyl: TTF might look more pretty but we a) can not update the font.dir b) do not use font config yet c) do not really share glyphs d) have less space Jun 28 11:40:59 mickeyl: do you have the power to revoke keys at bkbits? Jun 28 11:41:12 mickeyl: a) can be solved by a program from hrw Jun 28 11:41:25 mickeyl: c) can be solved to built freetype into Qt Jun 28 11:42:00 d) is only an issue on collie Jun 28 11:42:41 + all other 16MB ipaq Jun 28 11:42:50 koen: gues who has a collie ;) Jun 28 11:42:53 16mb ipaqs actually have 16MB Jun 28 11:42:59 ah right Jun 28 11:43:02 not 14.5 Jun 28 11:43:23 well 16MB - kernel Jun 28 11:43:30 mickeyl: ship happens if you do not use a metric system Jun 28 11:44:22 * koen finds that difficult to imagine Jun 28 11:44:31 but I'm a european engineer :) Jun 28 11:44:52 koen: something funny happened at the swiss border Jun 28 11:45:02 koen: we decided to build a bridge across the border Jun 28 11:45:21 koen: when the bridges met there was one meter diference in the altitude... Jun 28 11:45:43 koen: germans meassured height from the 'meeres spiegel', swiss did something differently... Jun 28 11:45:53 i am aware of the issues and I think we shouldn't ship oz 3.5.4 unless we have sorted some of 'em out. On hires models, ttf even takes less space than those large qpf's. For collie we might reconsider it, however hrw told me he's using ttf since long ago and has never experienced any performance problems. that leaves more or less just the space problem on collie. Jun 28 11:45:54 heh Jun 28 11:46:20 mickeyl: that means Qt did not get slower ;) Jun 28 11:46:39 * mickeyl watches the Qt4 Dance Video Jun 28 11:46:50 mickeyl: We've been dancing all day... Jun 28 11:46:51 do the qt4 dance Jun 28 11:46:59 mickeyl: and popler got Arthur Painter already Jun 28 11:47:11 mickeyl: I hope Mozilla will get it during the next days as well Jun 28 11:47:44 zecke: did you see the mail from zack about obsoleting XAA? Jun 28 11:47:56 koen: no Jun 28 11:48:12 koen: got a uri to the xorg ml archive? Jun 28 11:48:18 http://lists.freedesktop.org/archives/xorg/2005-June/008356.html Jun 28 11:48:38 damn I should have bought the magnifier as well Jun 28 11:48:54 mickeyl: if we cross our fingers kdrive will have a (stable) release after Xorg 7.0 Jun 28 11:49:27 rock Jun 28 11:49:32 *crossing fingers* Jun 28 11:51:12 zecke: poppler is in OE, but the qt backend is disabled, if someone with QT skills could have a look at that, I'd be very happy Jun 28 11:52:16 koen: I've written my own poppler-qt at ROAD Jun 28 11:52:21 koen: the basic version is just crap Jun 28 11:52:35 koen: building it with Qt2 is simple Jun 28 11:52:44 damn I really should have bought the magnifier Jun 28 11:53:38 koen: today (KAA) sounds really nice Jun 28 11:54:23 now that zack gets paid, I hope he finishes his work Jun 28 11:54:28 and doesn't stop tomorrow Jun 28 11:54:48 and doesn't get distracted as easily as keithp Jun 28 11:55:12 "enhanching tslib supp.... SHINY OBJECTS!!!!" Jun 28 11:55:25 -h Jun 28 12:00:07 koen: LOL Jun 28 12:00:23 koen: Heroin can talk? Jun 28 12:00:25 * koen searches for a dozen ipaqs to do something like http://tungstengraphics.com/TVCds.html Jun 28 12:00:58 koen: that looks nice Jun 28 12:01:16 koen: I would take the system left to him Jun 28 12:01:28 yeah Jun 28 12:02:03 but that would take so much power I couldn't plugin a fridge anymore Jun 28 12:02:53 mickeyl: do you have a thunder or tiger system? Jun 28 12:03:56 * chouimat is back. Jun 28 12:05:39 koen: thunder i7505 Jun 28 12:06:30 ah, the intel one Jun 28 12:06:42 thunder k7 over here Jun 28 12:07:39 more than 4 years old and still being fast :) Jun 28 12:08:02 ~lart 1.5 year turnaround time Jun 28 12:08:02 * ibot whips out a hot clue gun and makes sure that 1.5 year turnaround time is stuck to the floor Jun 28 12:09:22 mickeyl: could you revoke all keys in bkbits on thursday? Jun 28 12:09:33 koen: i could do that. Jun 28 12:10:06 *sigh* this'll be a pretty busy week Jun 28 12:10:45 yeah Jun 28 12:11:14 I realised there's no way to finish my project before thursday, so I'm taking it easy at the moment Jun 28 12:11:24 when are you leaving? Jun 28 12:11:29 thursday? Jun 28 12:11:32 friday 6 am Jun 28 12:11:55 thursday afternoon we're going to my paren't to save an hour of traveling time Jun 28 12:12:09 * koen gets another coffee Jun 28 12:13:37 hey ggilbert Jun 28 12:13:41 mickeyl: Hi. Jun 28 12:14:23 mickeyl: I was looking at fixing the suspend/resume problem in the w100 qte driver. Jun 28 12:14:41 any ideas on how to name our monotone identities? Jun 28 12:15:56 firstname.lastname@openembedded.org, @openembedded.org, @openembedded.org? Jun 28 12:16:59 SirFred: hi. any luck yet? Jun 28 12:17:03 mickeyl: My idea was to trigger w100 init on some qcop message, you said that QPE/System aboutToSuspend() and QPE/System returnFromSuspend() were used on suspend/resome. Jun 28 12:17:26 SirFred: there may be a better way than that Jun 28 12:17:28 mickeyl: Well, my first question is know if those messages are reaching qapplication. Jun 28 12:17:33 mickeyl: Tell me. Jun 28 12:17:51 SirFred: I'm leaning more towards a direct call from ODevice to the qtscreen or the qws Jun 28 12:18:11 Would it be kosher if I added a FEED_URIS line, but put a '#' before the name of the feed to effectively comment it out? Jun 28 12:18:23 koen: i think i'm for @oe.org Jun 28 12:18:24 mickeyl: Something like having ODevice methods to suspend/resume? Jun 28 12:19:00 SirFred: yep. ODevice_Zaurus already overrides the default suspend/resume, so that'd be the natural place to wake up something Jun 28 12:19:02 the problem Jun 28 12:19:13 that won't work when you suspend/resume via button Jun 28 12:19:16 or apm -s Jun 28 12:19:24 in that case there's no such message Jun 28 12:19:26 That's a big problem. Jun 28 12:19:28 (returnFromSuspend) Jun 28 12:19:59 which makes me think it needs to be in the driver Jun 28 12:20:02 mickeyl: Perhaps we need some kind of direct access from the driver? Jun 28 12:20:03 not in userland Jun 28 12:20:21 RP: what do you think? Jun 28 12:20:29 the solution should benefit all GUIs using the w100 Jun 28 12:20:37 so even the qte driver would be too highlevel Jun 28 12:20:42 mickeyl: You mean the kernel driver. Jun 28 12:20:46 yep Jun 28 12:20:54 the kernel driver will get notified in any case Jun 28 12:20:58 on resume Jun 28 12:21:04 so that looks like the really natural place to solve it Jun 28 12:21:48 mickeyl: I suppose that could work as long as we are not using hardware rotation anymore to rotate the screen. :) Jun 28 12:21:51 koen: is ewi546.ewi.utwente.nl still correct? Jun 28 12:22:01 Luke-Jr: yes Jun 28 12:22:12 strange... :\ Jun 28 12:22:22 I get an error trying to pull Jun 28 12:22:38 mickeyl: Did you receive my today patches? Jun 28 12:22:40 http://pastebin.ca/16125 Jun 28 12:23:18 try pulling org Jun 28 12:26:26 mickeyl: What are you asking here? You want the kernel driver to restore all the accel registers? Jun 28 12:29:04 RP: Hi. Jun 28 12:29:17 hi SirFred Jun 28 12:29:51 RP: Now that I read you, I'm not sure if any operation at kernel driver will be enough to avoid the problem. Jun 28 12:30:08 The kernel can't fix this Jun 28 12:30:14 RP: The problem is actually that suspend/resume freezes the AtiCore functionality. Jun 28 12:30:28 RP: And you're right, that cannot be fixed at kernel driver. Jun 28 12:30:46 SirFred: Can you be specific - when does it freeze? Jun 28 12:30:48 at suspend or resume? Jun 28 12:30:54 RP: At resume. Jun 28 12:31:10 RP: I'm not sure why, but it start to give timeouts for every operation. Jun 28 12:31:36 I'd guess it needs reinitalising Jun 28 12:31:38 RP: Perhaps the mmapped memory gets corrupted, or something so. Jun 28 12:31:50 RP: What happens to shared memory on a suspend? Jun 28 12:32:06 All the w100's registers are destroyed. The memory should be preserved... Jun 28 12:32:15 RP: Well, anyway, the shared memory looses sync with the w100 itself. Jun 28 12:32:33 Actually, I think we only save the active video memory - not any of the surrounding memory Jun 28 12:33:11 RP: Perhaps that's the problem, assuming invalid state ( stored in shared memory) Jun 28 12:33:23 RP: What do you mean with surrounding memory? Jun 28 12:34:06 We save all the memory - I've just checked Jun 28 12:35:22 Once upon a time, w100fb just use to save the active frambebuffer itself Jun 28 12:35:23 RP: Well, anyway, could be some sort of sync lost. Jun 28 12:35:38 SirFred: Every register in the w100 is destroyed over a suspend/resume Jun 28 12:35:44 RP: Wasn't the shared ram used to store some register value? Jun 28 12:36:21 No. We move all the video memory into main memory to preserve it, then turn off its power Jun 28 12:37:05 RP: I'm talking about the shared segment used by AtiCore. Jun 28 12:37:27 OT: FSF moved again, update your license headers Jun 28 12:37:38 Yes, that does cache values. That's a good point Jun 28 12:37:52 You probably need to tear down everything and start again :-/ Jun 28 12:38:03 RP: That's doable. Jun 28 12:38:18 RP: But I need to know when the device is resumed. Jun 28 12:38:27 RP: I don't know how to do that. Jun 28 12:38:46 SirFred: Opie should really tell you. It can find out from apm_bios Jun 28 12:39:14 load /home/pigi/.xchat2/identyfy_and_join Jun 28 12:39:16 sorry Jun 28 12:39:18 Ciao all Jun 28 12:39:40 damn xchat Jun 28 12:40:44 Because I'm too lazy to RTFG, is there a tool to htmlize a Packages file? Jun 28 12:40:45 mickeyl: Perhaps opie should read apm_bios and get informed of suspend/resume events, ever if they are triggered by the on/off button? Jun 28 12:41:04 CosmicPenguin: perl Jun 28 12:41:11 * koen hides Jun 28 12:41:17 * CosmicPenguin lols Jun 28 12:41:20 good show Jun 28 12:41:41 hmm Jun 28 12:41:42 well Jun 28 12:43:48 yeah Jun 28 12:44:00 either that or just with a resume.d script Jun 28 12:44:13 koen: org seems to be working... O.o Jun 28 12:44:55 SirFred: for the time being, let's assume that we call the reinitializing from ODevice in ODevice::suspend() Jun 28 12:45:11 i'll think about the method how to call that on an arbitrary resume later Jun 28 12:46:05 mickeyl: Well, that method is supposed to be actually suspending now, isn't it? Jun 28 12:46:11 mickeyl: Calling apm -suspend or something so? Jun 28 12:46:35 mickeyl: Presumably you mean ODevice::resume() ? Jun 28 12:47:02 RP: Well, if ODevice::suspend() is suspending, on resume, execution will follow at that point. Jun 28 12:47:17 RP: no the suspend is assumed to be synchronus Jun 28 12:47:25 Luke-Jr: collection != branch Jun 28 12:47:38 RP: prepare_to_sleep; sleep; we're back a life; Jun 28 12:47:46 Ah, right. I don't understand the workings of this high level stuff :) Jun 28 12:47:47 Arjan: ping Jun 28 12:47:54 koen: I was just going off your blog =p Jun 28 12:48:06 koen: so I need to have a local copy of all the org stuff just for OE.dev? Jun 28 12:48:26 sort of Jun 28 12:48:30 :\ Jun 28 12:48:41 Perhaps using qt_screen at that level ( ODevice) is a little dirty. Jun 28 12:48:50 a branch is a virtual thing Jun 28 12:49:08 Of course how you handle this when something else suspends the device is interesting... Jun 28 12:49:14 merging is just putting the right sha1 hashes in a manifest Jun 28 12:49:32 RP: :) Jun 28 12:50:01 RP: I suppose that be on time can not be guarantied. Jun 28 12:50:42 RP: Well, one thing is suspending and another one is be notified for an extern suspend. Jun 28 12:51:25 Perhaps ODevice::suspend() could call ODevice::notifySuspend() , that will also be called when an external suspend was detected. Jun 28 12:52:08 Or maybe you listen for something like notifyResume() ? Jun 28 12:52:47 RP: A qcop message? Jun 28 12:53:14 Well, I have to leave now. Jun 28 12:53:28 See you Jun 28 12:53:38 SirFred: bye! Jun 28 12:54:16 mickeyl: are we going monotone 0.19 or 0.20rc ? Jun 28 12:54:44 i'll probably try 0.20rtc Jun 28 12:54:46 rc even Jun 28 12:57:02 (for H in `monotone automate heads`;do monotone ancestry_difference $H `cat MT/revision`; done) monotone automate toposort -@- '| (while read R; do monotone log -r$R --depth=1) Jun 28 12:57:23 according to the montone devs that should give you all cset comments Jun 28 12:58:00 handy Jun 28 13:00:26 "can I do bk changes -R in monotone" "sure, do: (for H in `monotone automate heads`;do monotone ancestry_difference $H `cat MT/revision`; done) monotone automate toposort -@- '| (while read R; do monotone log -r$R --depth=1) " Jun 28 13:00:35 #monotone is scary :) Jun 28 13:03:53 hehe Jun 28 13:03:53 expoerts Jun 28 13:03:53 i'm sure we don't appear much less scary to outsiders Jun 28 13:06:11 bitbake -obscure-option | evil sed script | python foo Jun 28 13:06:27 heh, yeah Jun 28 13:07:01 thanks god we don't have perl devs hanging out here Jun 28 13:07:38 I do like perl :) Jun 28 13:08:02 that's probably why I do like ipkg ? Jun 28 13:08:08 or better, coding ipkg ? Jun 28 13:16:29 re Jun 28 13:16:31 hi mickeyl Jun 28 13:16:37 how's your c3000? Jun 28 13:16:39 hi bernado Jun 28 13:17:32 hi woglinde Jun 28 13:17:59 how are things Jun 28 13:18:04 ? Jun 28 13:19:01 hi. goes fairly well, seldomly used because of the exciting things happening with 2.6 and w100 accel on c7x0 Jun 28 13:20:40 accell driver = bigger binary = bloat = !PIM = fluff Jun 28 13:20:59 koen: breaking compat Jun 28 13:21:12 o yeah, forgot that one Jun 28 13:21:16 koen: imaging apps assuming not being accelerated Jun 28 13:23:39 xtank :) Jun 28 13:24:47 I had to go back to sharp's rom on my c1000, as I didn't even have a dialup line to download software... now I've found a open router, and I should have adsl again in another month or two Jun 28 13:25:18 is there a bb file that can be used to build the whole and only cosscompile ambient ? Jun 28 13:26:25 bitbake gcc-cross glibc Jun 28 13:27:04 only ? Maybe my question was not complete. Jun 28 13:27:46 I would like to avoid recompile all the stuff like gcc and so on that are used for the build of an image Jun 28 13:28:19 this way I can ASSUME_PROVIDED all the stuff ( I'm not sure I have been clear this time too ) Jun 28 13:56:51 ~herring holger Jun 28 13:56:52 * ibot whacks holger on the side of the head with a large red herring named alfred Jun 28 13:57:43 I'm scratching my head from where to get 3.3V for /SHDN... and my MAXIM3232 does not have it... Jun 28 14:06:37 mickeyl: no way to make a kernel from recent oe boot on my poodle, can be the image and not the kernel? Jun 28 14:06:48 (it stopes just after the splash screen) Jun 28 14:19:53 darkschneider: unlikely, since other people have success with that. Jun 28 14:20:02 darkschneider: i recommend you building a kernel with console=tty1 Jun 28 14:20:09 and then have a look at the output Jun 28 14:20:16 k Jun 28 14:20:21 or ttyS0, when you have a serial cable Jun 28 14:20:28 (i was reporting my failure :) ) Jun 28 14:21:14 a fact i noted is that notting much has changed in the kernel since the 3.5.2 release Jun 28 14:21:30 just a tosa patch and a variable that has been changed Jun 28 14:27:00 ~praise marcan Jun 28 14:27:00 All hail marcan! Jun 28 14:27:07 marcan: I haven't been able to lock up the 5500 any more Jun 28 14:27:14 we can consider this issue solved Jun 28 14:27:34 now it would be handy if someone could patch it for 5600 and 6000 as well :)) Jun 28 14:27:44 "case closed" Jun 28 14:35:50 evening Jun 28 14:36:02 hi Jun 28 14:36:06 g'night Jun 28 14:36:26 'night mickeyl Jun 28 14:36:42 'night mickey|zzZZzz Jun 28 14:58:57 So, we're going to stop submitting stuff to BK on thursday? Jun 28 14:59:08 time to sleep Jun 28 14:59:11 nite all Jun 28 15:16:56 'night all Jun 28 15:20:30 'night all Jun 28 15:36:33 does tinyx or kdrive do anti-aliased fonts? Jun 28 15:36:39 sure Jun 28 15:37:27 kdrive has Render, Fixes, Composite extensions as well Jun 28 15:37:46 ljp: some people would even say it will sit your spouse... Jun 28 15:37:51 alpha blending? Jun 28 15:39:04 ljp: I do not see who should prevent when from doing it Jun 28 15:39:25 ljp: ask Lars and Zack ;) Jun 28 16:03:03 huh Jun 28 16:03:14 so now I know why other Z's don't have the 5500 bug Jun 28 16:03:24 the apm scripts don't even get CALLED Jun 28 16:03:46 apmd is never notified Jun 28 16:03:49 ~lart lineo Jun 28 16:03:50 * ibot beats lineo to within 2.54cm of his life Jun 28 16:07:55 marcan: well lineo has little to do with the code in oz :p Jun 28 16:08:13 apart from the fact that we work around some apm kernel bugs on the 5500 Jun 28 16:09:37 treke: it seems to me that the kernel doesn't even send the events Jun 28 16:09:41 it doesn't seem to be userland Jun 28 16:10:01 in fact, it works if I do a software suspend (apm -s) Jun 28 16:10:15 check apm Jun 28 16:10:19 make sure it isnt a wrapper Jun 28 16:10:27 it is I think Jun 28 16:10:30 which is what I mean Jun 28 16:10:37 it calls apmd and does it right Jun 28 16:10:57 do probably it doesn't work with normal on/off-button suspend because the kernel doesn't send the events Jun 28 16:10:57 eh? Jun 28 16:11:05 the kill signal to apmd exists to work around the bug Jun 28 16:11:14 wait - let me check Jun 28 16:11:15 I dont recall what the work around was clearly :) Jun 28 16:11:37 the bug is when multiple apps have the apm_bios device open, on the 5500, during a suspend/resume cycle, it never resumes Jun 28 16:11:42 so you cant do the _proper_ method of calling the apm binary to signal the kernel Jun 28 16:11:55 kergoth: a single app kills it too Jun 28 16:12:03 eh? Jun 28 16:12:21 run apmd, suspend using the button. Boom. Jun 28 16:12:24 at least on recent OZs Jun 28 16:12:45 and what do you think is catching the button press? Jun 28 16:12:47 hm? Jun 28 16:12:50 The kernel. Jun 28 16:13:23 thats only true in console. in opie, it catches the press and calls the apm binary Jun 28 16:13:26 iirc. Jun 28 16:13:31 I was running in console Jun 28 16:13:33 checking the code Jun 28 16:13:35 and it hung nonetheless Jun 28 16:13:53 the problem with the kernel was it shutting down half the system and then sending the signal to apmd and waiting for a response Jun 28 16:14:07 very silly of course, since unless apmd was very fast and responded fast enough, it got halted Jun 28 16:14:21 and thus the kernel hangs around waiting for a response that will never get there Jun 28 16:14:30 until resume of course, which necer happens because suspend never completes either Jun 28 16:14:59 the 2.4 apm and pm in general is a fucking joke Jun 28 16:15:06 :\ Jun 28 16:15:13 try to read the code Jun 28 16:15:26 it is horrid Jun 28 16:15:29 i've been through every line of lineo code in the sharp 2.4 5500 codebase. Jun 28 16:15:39 so yes, i know Jun 28 16:15:41 heh Jun 28 16:15:43 heh Jun 28 16:15:47 definitely Jun 28 16:16:02 Now, there's a line for your resume Jun 28 16:16:02 I hope we move to 2.6 soon Jun 28 16:16:05 at least for tosa Jun 28 16:16:08 haw Jun 28 16:16:14 I would hire you on the spot Jun 28 16:16:20 lol Jun 28 16:16:27 "read the lineo code and left with mind somewhat intact!" Jun 28 16:16:44 "(NOTE: in similar circumstances, expects hazard pay) Jun 28 16:16:44 " Jun 28 16:16:45 :P Jun 28 16:17:57 all this also expects the wlan crashes on suspends Jun 28 16:18:03 it doesn't even call the scripts Jun 28 16:18:09 *explains Jun 28 16:33:02 great - now the tosa kernel won't compile Jun 28 16:35:31 arm-linux-ld-2.11.2: final link failed: No space left on device Jun 28 16:35:31 great **** BEGIN LOGGING AT Tue Jun 28 16:39:15 2005 Jun 28 16:41:23 so... which Monotone branch is equivalent to HEAD? .koen or .dev? Jun 28 16:43:10 marcan: 'we' who? SL-Cxx0 are already mostly on 2.6, if you mean "Zaurii images" =p Jun 28 16:43:34 I mean tosa :P Jun 28 16:43:40 i.e. get tosa 2.6 ready Jun 28 16:43:46 (wifi, sound, suspend, etc) Jun 28 16:44:07 Once I get the very basic big bugs of spm on tosa on the embedix kernel fixed, I'm diving straight into getting the 2.6 kernel ready Jun 28 16:44:16 *apm Jun 28 16:50:45 heh Jun 28 16:50:55 marcan: a stock 2.4 kernel calls the resume events in the same order as the suspend events Jun 28 16:51:00 * kergoth wonders how that _ever_ worked Jun 28 16:51:15 so nearly every embedded patch set ends up flipping the resume order Jun 28 16:51:16 heh Jun 28 16:51:16 heh Jun 28 16:51:18 pathetic Jun 28 16:57:22 kergoth: a lot of older stuff didn't do much on suspend/resume, it was the cross-your-fingers strategy really Jun 28 17:17:49 * kergoth nods Jun 28 17:24:15 argh - keyboard brakes after a while on tosa Jun 28 17:24:22 that's it - I'm leaving the 2.4 kernel Jun 28 17:24:34 let's get 2.6 ready ASAP Jun 28 17:24:44 * marcan compiles current 2.6 Jun 28 21:13:51 booga booga Jun 28 21:14:13 lets crunch some evas code :P Jun 28 21:16:00 kcall? Jun 28 21:16:44 ? Jun 28 21:17:11 someone named Eva made kcall =p Jun 28 21:17:45 and a guy named raster made evas and a whack more goodies Jun 28 21:18:23 whats it do? Jun 28 21:19:04 you ever hear of enlightenment? Jun 28 21:19:39 guess not Jun 28 21:20:07 raster, i guess there is one guy on the planet that doesnt know about you :P Jun 28 21:20:36 i'll add him to my jeelly bean jar Jun 28 21:20:37 :) Jun 28 21:20:40 lol Jun 28 21:20:44 how goes it? Jun 28 21:20:50 emte: enlightenment == evas how? =p Jun 28 21:20:51 better today than yesterday? Jun 28 21:21:13 * Luke-Jr used to run Enlightenment on his old 133 MHz laptop since it wouldn't run much else Jun 28 21:22:01 well evas happens to be the canvas library Jun 28 21:22:12 ... and quite a bit more Jun 28 21:22:47 i c Jun 28 21:23:38 dont worry about it Jun 28 21:25:02 not sure i have enough guinness tho :( Jun 28 21:25:24 * emte sets up a "donate to emte's guinness fund" Jun 28 21:27:26 Guinness powered programming? Jun 28 21:27:53 emte: busy busy ... hungry too Jun 28 21:27:54 yeah ... didnt raster tell you thats how he does it? Jun 28 21:28:00 :) Jun 28 21:28:27 * emte sends raster a nice big prime rib pepper steak Jun 28 21:28:31 Makes a change from the usual coffee powered programming. Jun 28 21:28:57 ... and peopel wonder wht enlightenmnet is creative :P Jun 28 21:29:01 why* Jun 28 21:29:11 and i cant spell Jun 28 21:29:14 :) Jun 28 21:30:58 odd ... Jun 28 21:31:18 On Semi is a backwards company ... Jun 28 21:31:37 hahahaha Jun 28 21:31:38 moving fabrication from malaysia to pheonix?? Jun 28 21:31:41 lunch time! Jun 28 21:31:41 werd Jun 28 21:31:55 must be cheaper to hire mexicans Jun 28 21:33:02 Maybe the cost of bribes got too high. Jun 28 21:34:08 * emte shrugges Jun 28 21:34:23 i only read the headline and the first paragraph Jun 28 21:34:50 you can goto http://www.eetimes.com if you wanna read about it .,.. Jun 28 21:55:28 hey raster, find me a cheap/free flybook .... Jun 28 21:55:45 i need something decent and small enough to program on the bus Jun 28 21:56:23 arnt people supposed to throw stuff like that in the trash in JP? Jun 28 21:56:38 month old and out of date Jun 28 21:56:55 maybe i'll just buy a lotto ticket instead Jun 28 21:57:07 sounds more likly Jun 28 21:59:52 * emte goes to get another guinness instead Jun 28 22:06:16 mmm guinness Jun 28 22:36:39 :( Jun 28 22:36:49 * emte is out of guinness :( Jun 28 22:37:03 now its down to wine or pernode ... Jun 28 22:37:09 hmm Jun 28 22:37:25 maybe pernod and wather later ... Jun 28 22:37:34 water* Jun 28 22:42:40 hmm ... well i got an A in embedded systems and C Jun 28 22:42:55 not too shabby Jun 28 22:43:21 only a b+ in digital tho ... Jun 28 22:43:54 prob because my BCD clock Lab was late ... Jun 28 22:44:05 stupid thing Jun 28 22:44:45 hello, I try to build a bootstrap-image for wrt54g, and get this warning Jun 28 22:44:45 NOTE: exceptions.NameError:name 'linux_module_packages' is not defined while evaluating: Jun 28 22:44:45 ppp wireless-tools ${@linux_module_packages('ipv6 ipsec mip6-mn tun loop sunrpc nfs af_packet ppp-async ppp-deflate ppp-mppe ip-gre ip-tables ipip input uinput ')} Jun 28 22:44:48 what should I invest in OE? Jun 28 22:45:29 not sure Jun 28 22:45:39 sounds liek a bad macro for some reason Jun 28 22:45:40 can i play mp3 sound in gpe ? Jun 28 22:45:45 how do i do that ? Jun 28 22:46:02 ram123, use duct tape and a straw? Jun 28 22:46:21 night kergoth`zzz Jun 28 22:46:59 does gpe has an mp3 player ? Jun 28 22:48:52 gpe-nmf Jun 28 22:49:09 where do i find it ? Jun 28 22:49:14 bitbake gpe-nmf Jun 28 22:49:16 from the main menu ? Jun 28 22:49:19 ram123, congrats that is a much wiser question Jun 28 22:50:01 I think "Audio player" on the menu Jun 28 22:50:05 after installation Jun 28 22:50:20 i gave bitbake gpe-image Jun 28 22:50:31 that won't include it, last I checked Jun 28 22:50:32 does it get added into it Jun 28 22:50:58 ... Jun 28 22:51:52 funny how he is like you were Luke-Jr ... Jun 28 22:52:02 good thing you have gotten better Jun 28 22:52:03 :) Jun 28 22:52:46 I wasn't like that =p Jun 28 22:53:02 I usually stick around after a question is answered Jun 28 22:53:09 unless it's like #linux or something spammy Jun 28 23:05:59 ram123, you really do not need to leave after each question ... you do know that right? Jun 28 23:06:31 sorry. Jun 28 23:06:35 more than one person my give you an awnser Jun 28 23:06:39 may* Jun 28 23:06:41 my system got rebooted Jun 28 23:06:47 ah Jun 28 23:07:30 once i do bitbake gpe-nmf how do i add it to bitbake gpe-image ? Jun 28 23:07:30 ie adding it to the main image ? Jun 28 23:07:36 anyway like Luke-Jr said nmf is not a std package Jun 28 23:07:50 you build it, dld it and install it Jun 28 23:08:02 or get it form the build world for your distro Jun 28 23:08:32 adding it to the image requires you to edit a couple meta files and add it Jun 28 23:09:50 can i put the executable into the filesystem directly Jun 28 23:10:06 what is the ouput of bitbake gpe-nmf Jun 28 23:10:10 you can dld the package and install it , yes Jun 28 23:10:30 ram123: the output is an IPK Jun 28 23:13:06 where does it get stored ? Jun 28 23:13:36 in /deploy/images ? Jun 28 23:14:12 almost Jun 28 23:14:20 images are just that Jun 28 23:14:37 deploy/ipk :) Jun 28 23:15:32 IPKGDIR=tmp/deploy/ipk/; tmp/staging/x86_64-linux/bin/ipkg-make-index -l "${IPKGDIR}"/Packages.filelist -p "${IPKGDIR}"/Packages -v "${IPKGDIR}" Jun 28 23:15:34 adjust as needed Jun 28 23:25:35 ERROR: Nothing provides virtual/kernel-wrt Jun 28 23:25:36 ERROR: dependency virtual/kernel-wrt (for task-bootstrap) not satisfied Jun 28 23:26:51 sounds like an error in your config Jun 28 23:27:03 I include packages/*/*.bb and there is a linux-wrt_2.4.20.bb in ./linux/, what should I do? Jun 28 23:27:36 it doesnt know you want linux-wrt_2.4.20.bb Jun 28 23:29:37 hue, you told it you want linux-wrt but not specifically a version or where to get it Jun 28 23:31:46 emte, but I have do this ""BBFILES := "/stuff/openembedded/packages/*/*.bb" Jun 28 23:33:33 emte, what else should I add to local.conf to specify that I'd like linux-wrt_2.4.20.bb ? Jun 28 23:33:36 PREFERRED_PROVIDER_virtual/kernel = "handhelds-sa" Jun 28 23:33:51 PREFERRED_PROVIDER_virtual/kernel = "linux-wrt" Jun 28 23:33:59 try that Jun 28 23:34:08 emte, thanks. Jun 28 23:34:39 you should have that in your machine config ... Jun 28 23:50:05 blech .. i'll chanse bugs for breakfast, night Jun 28 23:57:53 emte, there is a line BOOTSTRAP_EXTRA_DEPENDS = "virtual/kernel-wrt ppp wireless-tools hostap-modules" Jun 28 23:57:53 in wrt54.conf **** ENDING LOGGING AT Tue Jun 28 23:59:57 2005