**** BEGIN LOGGING AT Fri Jun 17 23:59:56 2005 Jun 18 00:54:12 morning all Jun 18 00:54:27 morning Jun 18 00:55:27 morning Jun 18 00:56:27 RP: 2.6.12-rc6 kernel does not come back to live, once switched off or suspended. Is taht a known issue ? Jun 18 00:59:52 RP: ping Jun 18 01:16:09 uv1: Its a known issue Jun 18 01:16:17 uv1: Someone needs to find out why... Jun 18 01:16:39 It only started with 2.6.12-rc6-mm1... Jun 18 01:20:26 RP: Recognized that also. 2.6.12-rc5 has also a major issue when returning from sleep. Jun 18 01:21:11 uv1: I suspect this is the power key repeat problem as it doesn't do it for "apm -s" Jun 18 01:21:41 2.6.11 did have that as well, but this did not often occur. Waking up by button sometimes does not fully startup the device. Instead of this, the vt screen gets stuck and you have to reboot. Jun 18 01:22:44 2.6.12-rc5 does have this problem obviously also when woken up by alarm. This is bad, as the device does not go back to sleep and drains the battery. Jun 18 01:22:45 uv1: I think that is a different problem. I have seen it but never managed to find out what that problem was... Jun 18 01:23:38 Would you suggest to go back to 2.6.11 then ? Jun 18 01:23:39 uv1: I need to find some time to debug it. Maybe I'll find that this weekend but I can't promise anything Jun 18 01:23:48 Or someone else can have a go :) Jun 18 01:24:21 RP: Will be back on irc tomorrow. So if you need test help, I can do it. Jun 18 01:25:58 uv1: Ok, I'll keep that in mind, thanks. If I commit some "power button" fixes to the file, some reports on what problems remain will be handy. I suspect we have several problems with the kernel at the moment... Jun 18 01:58:20 RP: Hi. Jun 18 01:59:46 RP: Do you have an idea of how suspending/resuming interacts with opie? I've found that the w100 driver timeouts after resuming. I've also found that there's a AtiCore_WakeUpCall() that perhaps could be useful for this. Jun 18 02:15:40 SirFred: The w100 looses *all* state over suspend/resume Jun 18 02:15:50 SirFred: Hi btw :) Jun 18 02:15:56 RP: Hi. :) Jun 18 02:17:23 RP: So, we'll need some kind of reinitialization? Jun 18 02:17:24 SirFred: You will probably have to totally re-setup the device Jun 18 02:17:37 RP: buff Jun 18 02:17:49 SirFred: Yes, probably from scratch - that's what w100fb does Jun 18 02:19:42 RP: I need to find out if some function in qte is called on resuming. Jun 18 02:20:16 RP: Any knowledge about Aticore_WakeUpCall() ? Jun 18 02:21:09 SirFred: There was a call like that which did very little - I suspect its probably a different WakeUp to the one you have in mind Jun 18 02:21:39 Trust me when I say the ATI will have *no* internal state left and will need re init'ing Jun 18 02:21:49 RP: Of course, I trust you. Jun 18 02:21:51 RP: :) Jun 18 02:22:02 RP: That function name just give me some hopes. Jun 18 02:22:06 s/give/gave/ Jun 18 02:22:32 RP: I've made some progresses in the driver. Corrected a pair of bugs found by hrw. Jun 18 02:22:54 SirFred: Excellent. You also mentioned you knew the problem with arcs? Jun 18 02:22:58 RP: Now I'm getting ~1800 gops/s in the sysinfo benchmarc. Jun 18 02:23:09 RP: Yes. It was my mistake. Jun 18 02:23:28 What is the non accel gops/s? Jun 18 02:23:42 RP: polyline opcode was not accelerated due to a software mistake. Jun 18 02:23:52 RP: I'm getting ~1200 gops/s without acceleration. Jun 18 02:24:12 RP: Take in mind that the blitting test is never accelerated, because the pixmap has transparent background. Jun 18 02:24:12 That's not a bad improvement then :)) Jun 18 02:24:47 RP: I made a little test program copying the sysinfo benchmarking code, to test the different opcodes individually. Jun 18 02:24:52 That makes the figure even more impressive :) Jun 18 02:25:00 RP: The numbers were: Jun 18 02:25:09 Lines: 5128 lines/s -- 7194 lines/s Jun 18 02:25:17 Rects: 271 rects/s -- 533 rects/s Jun 18 02:25:27 Arcs: 387 rects/s -- 700 rects/s Jun 18 02:25:38 And pixmaps more or less the same speed, of course. Jun 18 02:25:51 RP: We need that Aticore_TransBitblt. :-( Jun 18 02:26:09 I was just thinking that Jun 18 02:26:47 I need more time to spend on things :) Jun 18 02:27:14 hi all Jun 18 02:27:16 RP: don't we all... Jun 18 02:27:21 hi bluelightning Jun 18 02:27:26 RP: I've also fixed a sinchronyzation problem between accel and non accel functions. Jun 18 02:28:08 SirFred: The driver appears to be moving along well :) Jun 18 02:31:31 RP: Well, slowly Jun 18 03:28:53 03pb 07 * r1.3562 10openembedded/packages/cyrus-imapd/ (3 files in 2 dirs): update cyrus-imapd to 2.2.12; fix FILES per bug 87 Jun 18 03:39:54 03pb 07 * r1.3563 10openembedded/packages/mythfront/ (mythfront-session/mythfront.sh mythfront-session.bb): also link mysql.txt Jun 18 03:58:58 Anybody know in what package could I find the "QPE/System" QCop server implementation ? Jun 18 04:47:26 afternoon Jun 18 06:23:31 03pb 07 * r1.3564 10openembedded/packages/gpe-dm/gpe-dm_0.47.bb: update gpe-dm to 0.47 Jun 18 06:24:44 * zecke is too stupid to even remember his bugzilla account name Jun 18 06:25:24 on handhelds.org, or openembedded.org? Jun 18 06:26:34 03pb 07 * r1.3565 10openembedded/packages/mythfront/ (mythfront-session/mythfront.sh mythfront-session.bb): fix some mythfront problems Jun 18 06:26:52 pb_: good hint Jun 18 06:30:02 zecke: heh Jun 18 06:32:40 I tried zecke@handhelds.org at openembedded... I should have used freyther@math.fu-berlin.de Jun 18 06:34:01 OT: Does El Cattivo mean anything bad? Jun 18 06:36:20 pb_: is it right(tm) to create a 'release bitbake-1.3.1' bug and make it depend on known issues? Jun 18 07:08:31 zecke: yeah Jun 18 07:09:25 zecke: have a look at the bugs listed in the page footer on http://handhelds.org/~bugzilla/, for example Jun 18 07:16:07 l Jun 18 07:16:36 indeed Jun 18 07:29:07 pb_: thanks Jun 18 07:44:33 zecke: El cattivo? Jun 18 07:47:54 zecke: I suppose it's italian. In my mother language, galician, cativo means bad, with bad meaning bad quality. Jun 18 07:48:34 03zecke123 * r263 10bitbake/lib/bb/build.py: Jun 18 07:48:34 bitbake/lib/bb/build.py: Jun 18 07:48:34 -Fix the task_cache as reference assumption as described Jun 18 07:48:34 in bug report #60 Jun 18 07:48:34 Patch courtsey keturn Jun 18 07:49:05 probably not ideal to name bitbake-1.3.1 that way... Jun 18 07:51:32 zecke: cattivo? Jun 18 07:51:48 zecke: Who proposed that? Jun 18 07:52:44 zecke: Well, In some regions in my country, it's also used as "young person", so perhaps italians use it in that way. Jun 18 07:58:04 SirFred: it is a song by a german music group Jun 18 07:58:20 "The Evil always wins" is the chorus... Jun 18 07:58:25 :) Jun 18 07:58:33 And the song is named cattivo Jun 18 07:58:34 ? Jun 18 07:58:42 yupp Jun 18 07:58:46 El Cattivo is a cowboy Jun 18 07:58:58 an evil cowboy though Jun 18 07:59:18 It gets some spaghetti western flavour. Jun 18 07:59:45 I supposed italians should say "Il Cattivo" Jun 18 07:59:49 Not sure. Jun 18 08:00:02 moin Jun 18 08:01:11 moin Jun 18 08:01:22 mickeyl: Hi. Jun 18 08:01:22 hi mickey Jun 18 08:01:23 SirFred: It's the opie-taskbar (QPE/System) Jun 18 08:01:57 mickeyl: Well, I was looking into implementing rotation in the driver. But I'm a little confused. Jun 18 08:02:01 zecke: listening to Die Ärzte ? Jun 18 08:02:30 mickeyl: Actually 'Transformed' driver is used to draw on the screen. Jun 18 08:02:57 mickeyl: This is just a linuxframebuffer but with some functions to transform the coordinates given an screen orientation. Jun 18 08:03:16 mickeyl: I'm not sure of how should I implement rotation for the accelerated drivet. Jun 18 08:03:23 s/drivet/driver Jun 18 08:03:29 i know what you mean Jun 18 08:03:40 i would need to look into the code to give an educated answer Jun 18 08:03:42 One idea should be just say Transformed to inherite from QW100Screen. Jun 18 08:04:13 mickeyl: But that solution will shadow the transformed unaccelerated driver. Jun 18 08:04:37 mickeyl: My first doubt is how the rotation works now on opie. Jun 18 08:04:38 i think you rather need to tell qt that you are a transformed driver and then handle the transformation in the driver Jun 18 08:05:05 try asking that on qt-embedded-interest@lists.trolltech.com Jun 18 08:05:15 mickeyl: There's a "hardware" way to rotate the framebuffer. Jun 18 08:05:17 mickeyl: what make you think that? Jun 18 08:05:22 maybe once in a lifetime one gets a usable answer Jun 18 08:05:24 heh Jun 18 08:05:39 mickeyl: I would like to know if that is used now in opie. Jun 18 08:05:59 mickeyl: Because it change things a little. Jun 18 08:06:04 SirFred: we set QWS_DISPLAY Jun 18 08:06:25 zecke: Yes, I've also seen that, but I've seen that it's set as Transformed:*:0 Jun 18 08:06:36 zecke: That's no good for the accelerated driver. Jun 18 08:07:19 Perhaps the basename of the driver should be used for that. Jun 18 08:07:27 zecke: do you have any idea why the h1940 would need to set QWS_MOUSE_PROTO=TPanel in tslib.sh? Jun 18 08:07:43 I mean, if the driver name is W100:Rot0:0, on a rotation, it should be set to W100:Rot90:0 Jun 18 08:08:07 i don't see a problem in that Jun 18 08:08:10 But I think that it's hardwired to Transformed as the basename, no matter what was its original name. Jun 18 08:08:18 ah, that's no biggy Jun 18 08:08:22 pb_: no, it sometimes happen ;) Jun 18 08:08:30 zecke: heh Jun 18 08:08:37 pb_: what MACHINE is the h1940? Jun 18 08:08:40 mickeyl: biggy? Jun 18 08:08:45 SirFred: no big problem Jun 18 08:08:48 mickeyl: Oh Jun 18 08:08:50 mickeyl: :) Jun 18 08:09:12 we already have API to query the default rotation Jun 18 08:09:16 zecke: currently it has its own MACHINE, but I would like to abolish that if possible Jun 18 08:09:18 we can add an API to query the default driver Jun 18 08:09:40 mickeyl: Perhaps that should be better. Jun 18 08:09:53 afaict, the only difference between the h1940 and h3600 parts of tslib is that the h1940 sets that variable and h3600 doesn't. Jun 18 08:10:00 yep. i wouldn't like to see Transformed be abused Jun 18 08:10:05 mickeyl: So, the idea is to handle the rotation from the W100 driver. I don't remember how it was done in the sharp rom. Jun 18 08:10:16 pb_: I think -DQWS_TSLIB makes TPanel the default driver... Jun 18 08:10:48 mickeyl: BTW, I have corrected a pair of bugs that hrw notified a pair of days ago. Jun 18 08:10:51 pb_: I've never looked into it too carefully Jun 18 08:10:52 is there a bash "god" here by chance? I need to redirect STDIN (ie: the keyboard) inside a shellscript so that keypresses are no longer written into the shell. Can it be done? Google wasn't much of a help Jun 18 08:10:57 zecke: okay Jun 18 08:11:04 mickeyl: Could I send you a new w100.patch ? Jun 18 08:11:28 CoreDump|home: using `echo` you should be able to turn it off Jun 18 08:11:51 mickeyl: The bad performance on arcs and the bad synchronization on drawing pixmaps and rects are fixed. Jun 18 08:11:54 zecke: hmm how so? Any links? Jun 18 08:12:44 SirFred: sure Jun 18 08:12:57 SirFred: excellent news Jun 18 08:13:05 pb_: found it Jun 18 08:13:13 setting the fgcolor to black doesn't count :) Jun 18 08:13:17 mickeyl: I'd rather send you the whole w100.path file, and not a patch for the old w100.patch file. Is that fine for you? Jun 18 08:13:29 SirFred: yep Jun 18 08:13:35 mickeyl: Sending... Jun 18 08:14:07 if QT_QWS_[IPAQ,C700,CUSTOMPANEL,SL55xx,YOPY,CASIOPAI] is defined the default mousedriver is TPanel Jun 18 08:14:19 otherwise 'Mouse' (if mouse is compiled in) or None Jun 18 08:19:04 SirFred: got it. i'm checking compilage then i'll commit Jun 18 08:19:14 mickeyl: ok. Thanks. Jun 18 08:24:26 CoreDump|home: stty -echo Jun 18 08:25:23 03mickeyl 07 * r1.3562 10openembedded/packages/qte/qte-2.3.10/c7x0-w100-accel.patch: qte-2.3.10: update w100 patch. this version fixes the synchronization between software pixmaps and hardware rects as well as polyline accelleration. Patch courtesy Manuel Teira - thanks a lot. Jun 18 08:25:26 03mickeyl 07 * r1.3550.2.1 10openembedded/packages/soundtracker/soundtracker-gtk2_0.6.7.bb: add gtk2-port of the soundtracker, an AMIGA style module tracker program Jun 18 08:26:08 mickey|brb: :) Jun 18 08:26:16 03mickeyl 07 * r1.3563 10openembedded/packages/qte/qte_2.3.10.bb: qte-2.3.10: bump revision Jun 18 08:26:25 damn, I've forgotten my root pw on the Z Jun 18 08:26:26 mickey|brb: You're fast. Jun 18 08:28:19 hehe de Raadt is funny Jun 18 08:29:50 reenoo_: will try, thanks Jun 18 08:37:50 ~lart busybox Jun 18 08:37:50 * ibot lowers busybox's priority Jun 18 08:48:01 kergoth: hello, are you there? Jun 18 09:03:51 03CoreDump 07 * r1.3564 10openembedded/packages/sharp-binary-only/ (2 files in 2 dirs): sdcontrol: Remove Samba error message Jun 18 09:05:48 Hi Jun 18 09:06:21 good afternoon all Jun 18 09:06:31 hi koen & darkexe Jun 18 09:06:48 hey CoreDump|home Jun 18 09:20:07 reenoo_: ah, need to patch out the XIM stuff. Jun 18 09:23:10 hey offroadgeek Jun 18 09:27:58 hi coredump|home Jun 18 09:30:25 coredump|home: do you know who "RICHTER" is? (UPS says the package was signed by this name) Jun 18 09:31:30 offroadgeek: of course. I don't know him personally but he's about 60+ yo Jun 18 09:31:46 and I've been trying to reach him since thursday Jun 18 09:32:11 I guess he's in holliday or something. Jun 18 09:32:28 so, the signature is 99% fake Jun 18 09:33:17 ahh... so it wasn't him... just someone who signed his name Jun 18 09:33:30 mickey|brb: ping Jun 18 09:33:36 oh somebody stole a package? Jun 18 09:33:53 zecke: yep Jun 18 09:34:00 offroadgeek: yes Jun 18 09:34:37 and stupid UPS didn't check for an ID. They never do Jun 18 09:35:36 heh Jun 18 09:35:37 basically any name will do as long as there's a matching name-tag on the door _somewhere_ Jun 18 09:35:42 UPS itself stole my ipaq a year ago Jun 18 09:36:02 ouch, hows that? Jun 18 09:36:17 ipaq broke -> repairs -> empty box Jun 18 09:36:25 and -> new ipaq! Jun 18 09:36:37 heh Jun 18 09:36:50 i wish i _had_ an empty box lol Jun 18 09:38:55 * offroadgeek tries to call UPS Germany... this should be interesting Jun 18 09:39:22 CoreDump|home: I think it is more likely UPS stole it directly Jun 18 09:39:25 offroadgeek: i did that yesterday. They - of course - say it's not their problem and point to their TOS Jun 18 09:39:41 possibly Jun 18 09:39:47 pb_: I'm afraid that's a bit beyond my knowledge of X. I can send you my autotools patches if you want to give it a try Jun 18 09:40:02 coredump|home: ok... then I guess it will be up to trisoft to resolve (since they were the actual shipper) Jun 18 09:40:04 CoreDump|home: that is point in their customer policy Jun 18 09:40:41 CoreDump|home: did anyone know you were expecting the package? Jun 18 09:40:48 no one Jun 18 09:41:12 maybe someone left the building and saw the UPS guy Jun 18 09:44:44 CoreDump|home: How is that? You don't have a unique postal address ? Jun 18 09:45:09 CoreDump|home: I mean, a different number, flat, letter than your neighbours? Jun 18 09:45:15 nah, there are like um 20-30 flats in this house Jun 18 09:45:26 oh my name is unique all right Jun 18 09:45:46 CoreDump|home: And you don't have an unique door? Jun 18 09:45:51 fact is, UPS didn't deliver to me (i missed UPS by 10 minutes after waiting all day) Jun 18 09:46:10 and UPS didn't try to deliver at a later date Jun 18 09:46:20 but instead gave the parcel someone else WTF Jun 18 09:47:33 CoreDump|home: Sorry Jun 18 09:47:40 :\ Jun 18 09:48:17 CoreDump|home: What was in the package? Jun 18 09:48:34 a brand new shiny akita Jun 18 09:49:14 CoreDump|home: Do they allow you to talk with the shipper ? Jun 18 09:49:22 * offroadgeek want's to fly to germany to beat up coredump's UPS driver Jun 18 09:49:23 nope, tried that Jun 18 09:49:29 offroadgeek: :) Jun 18 09:49:39 We should go to the UPS office, to burn it. Jun 18 09:49:52 lol Jun 18 09:49:53 burn her, burn her Jun 18 09:50:28 CoreDump|home: So, what are you going to do now? Jun 18 09:51:06 I'll go to the police on monday. In the meantime I'll post a letter in the house asking the bastard to give it back. Jun 18 09:51:25 maybe someone saw who took it Jun 18 09:51:53 CoreDump|home: Good luck. Jun 18 09:51:59 thanks! Jun 18 09:55:21 coredump|home: do you want the invoice with the serial number? that way you can prove it was destined to you? Jun 18 09:56:04 oh a call to UPS will do that just fine :) It was -after all- addressed to me. Jun 18 09:56:28 + the tracking system shows the delivery Jun 18 09:58:03 I guess I meant for the serial number of the Z... in case you see someone in your building with an akita and you tackle them... Jun 18 09:58:24 he will use your ROM anyway... Jun 18 09:59:04 offroadgeek: ahh Jun 18 09:59:13 i didn't know it _had_ a serial Jun 18 09:59:19 that would be very useful indeed Jun 18 09:59:29 ok... I'll email it to you Jun 18 09:59:42 great! Jun 18 10:01:21 damn... and I was hoping for a new release of 3.5.3 for the akita by monday ;) Jun 18 10:01:51 hmm, anyone of the people working on bitbake/bitbake-ng here? Jun 18 10:02:04 :) Jun 18 10:02:21 fwiw, I have a new untested image lying around here :) Jun 18 10:02:29 TheCount: bitbake-ng probably only kergoth Jun 18 10:02:46 zecke: that's the sad part Jun 18 10:02:48 zecke: 'kay, tried contacting him, seems like the idle time is genuin ;) Jun 18 10:02:53 uh, genuine Jun 18 10:02:58 i wonder if one can read the serial off of the device Jun 18 10:03:04 hehe Jun 18 10:03:10 *implements a ROM bomb for this serial Jun 18 10:03:44 highly unlikely anyone will have any use for the contents of the package but thought it was something sellable Jun 18 10:04:15 TheCount: what content? what package? Jun 18 10:04:35 mickey|brb: the nano-1.2.1 part is tricky Jun 18 10:04:41 thecount: good point... we should check ebay.de... Jun 18 10:05:08 zecke: see offroadgeek/CoreDump Jun 18 10:05:16 offroadgeek: was it insured? Jun 18 10:05:26 TheCount: ah hehe Jun 18 10:05:27 TheCount: correct Jun 18 10:05:38 i dunno. trisoft will know Jun 18 10:06:53 i'm sure it was... I can't imagine trisoft would send a Z without insuring it Jun 18 10:07:09 right Jun 18 10:07:40 we'll at least it was a discounted Z ;) Jun 18 10:08:07 :D Jun 18 10:08:18 nice Jun 18 10:08:35 * koen waits for the 770 to arrive Jun 18 10:09:45 waits for the 770 discount code to arrive Jun 18 10:09:56 770 discount code? Jun 18 10:10:03 Can you buy 770's from retail yet? Jun 18 10:10:06 n770 Jun 18 10:10:07 nokia 770 for 99 euruos Jun 18 10:10:10 ahh Jun 18 10:10:15 CoreDump|home: I think one can Jun 18 10:10:17 euros even Jun 18 10:10:19 how do you get the discount code? Jun 18 10:10:31 yo bad the $ is worthless Jun 18 10:10:41 http://maemo.org/news/25052005.html Jun 18 10:10:44 ~convert 99euros to usd Jun 18 10:10:45 Crofton|laptop: a) invent time machine b) travel some weeks back c) send mail Jun 18 10:10:52 heh Jun 18 10:11:01 wow... the C1000 goes for 569 EUR at trisoft... Jun 18 10:11:12 ~convert 99 eur to usd Jun 18 10:11:19 thanks Jun 18 10:11:25 hmm Jun 18 10:11:25 hmmm, 800x480 screen, bluetooth, wifi for 99 euros Jun 18 10:11:32 versus 569 for the z Jun 18 10:11:34 ~convert 99 eu to usd Jun 18 10:11:45 ~change 99 eur to usd Jun 18 10:11:57 99.00 Euro (EUR) makes 121.755 United States Dollar (USD) (from http://www.xe.com/) Jun 18 10:12:11 I guess the 500 have been spoken for? Jun 18 10:12:21 I saw that, but wasn't on the ball then Jun 18 10:12:58 all that omap goodness for such a small price :) Jun 18 10:14:43 zecke: only 13 days left till Q3 Jun 18 10:14:50 heh Jun 18 10:15:00 I need to 64M of RAM Jun 18 10:15:11 I'm swapping my guts out in the OSK's 32M Jun 18 10:31:34 .o( /sbin/gutswapperd ) Jun 18 10:32:12 sounds painful Jun 18 10:36:44 wb ar_ Jun 18 10:36:55 hi Jun 18 10:37:29 I am a newbie in all this oe stuff.. I just builded Jun 18 10:37:47 built the task-bootstrap.. but there is no kernel in there Jun 18 10:37:59 How can I add one? Jun 18 10:38:28 Sorry for asking such dumbass questions ;) Jun 18 10:38:35 hmm IMO bootstrap should build one Jun 18 10:38:36 bitbake virtual/kernel Jun 18 10:40:48 CoreDump|home: only if the machine and/or package specify it in the depends Jun 18 10:41:18 ah right Jun 18 10:42:00 bootstrap-image.bb will probably build one Jun 18 10:43:32 Alright.. thanks. that seems to work.. Jun 18 10:45:52 bootstrap-image rocks Jun 18 10:46:03 at least for your first few builds :) Jun 18 10:46:07 darkexe: what machine are you building for? just building some random kernel won't help much. Jun 18 10:46:37 ipaq h3850 Jun 18 10:47:17 darkexe: what machine did you set in local.conf? Jun 18 10:48:32 darkexe: what you want is MACHINE="h3600" Jun 18 10:49:32 alright, thanks.. :) Jun 18 10:50:12 what kernel should i choose? 2.6? Jun 18 10:50:32 2.4 Jun 18 10:54:24 * TheCount continues hacking on bakerman Jun 18 11:02:46 Alright. got to go.. thanks for your help Jun 18 11:12:02 mickey|brb: I hate you for reporing #103 Jun 18 11:14:27 sorry about that but i use that feature a lot and it's definitly a regression bug Jun 18 11:14:38 so if we can't fix it sane we have to revert 235->236 Jun 18 11:15:16 mickeyl: do we have DEPENDS = 'foo-version'? Jun 18 11:15:25 yes Jun 18 11:15:27 zecke: yes, some packages do that Jun 18 11:15:41 mickeyl: I'm considering making foo+version be a special case Jun 18 11:16:05 mickeyl: the issue is with nano-1.2.1 we ask bestProvider() for nano Jun 18 11:16:13 mickeyl: and the best is 1.3.5 ;) Jun 18 11:16:20 which is wrong Jun 18 11:16:26 yeah Jun 18 11:16:28 mickeyl: yes it is wrong Jun 18 11:16:30 (no doubt) Jun 18 11:16:31 vi is the best provider Jun 18 11:16:50 koen: vi is for people too stupid to use sed Jun 18 11:16:58 what is r235 doing differently when gathering the best provider? Jun 18 11:17:08 zecke: vi is sed with a gui :) Jun 18 11:17:56 mickeyl: we ask bestProvider('nano') (see version is not there any more) Jun 18 11:18:09 mickeyl: and bestProvider does only check PREFERRED_PROVIDERS ;) Jun 18 11:18:28 zecke: so we have to readd that code to buildPackage Jun 18 11:18:35 showdata doesn't care about that Jun 18 11:20:01 * zecke just got his ubuntu cds Jun 18 11:20:42 zecke: ubuntu or kubuntu? Jun 18 11:20:54 ubuntu Jun 18 11:21:27 mickeyl: If we should build foo-1.3.5 we should build it? Jun 18 11:21:33 mickeyl: and nothing else? Jun 18 11:21:42 mickeyl: or should we assume foo-1.3.5 could be a virtual? Jun 18 11:21:50 or someone provides it? Jun 18 11:22:59 if we are requested to build foo-bar we should just build foo-bar and don't look for any other providers Jun 18 11:23:32 mickeyl: alternatively we can make bestProvider to carry a 2nd/3rd parameter Jun 18 11:23:45 mickeyl, I added another patch to bug 93 to address your concerns Jun 18 11:23:59 zecke: yeah, why not. add some more logic to best provider would be ok Jun 18 11:24:06 Crofton: k Jun 18 11:24:20 thanks for your help Jun 18 11:24:51 mickeyl: the question is do we need to find the bestProvider if we've foo-version Jun 18 11:25:03 mickeyl: or is foo-version already the best provider Jun 18 11:25:12 zecke: foo-version is already the best Jun 18 11:25:13 however Jun 18 11:25:17 we have to ensure the provider exists Jun 18 11:25:23 and we have to make the alternatives logic apply Jun 18 11:25:30 if one provider fails, then we try to build another one Jun 18 11:25:34 e.g. Jun 18 11:25:36 foo-bar-r0 Jun 18 11:25:37 foo-bar-r1 Jun 18 11:25:56 mickeyl: even more todo... Jun 18 11:26:05 yeah, but until now we don't have more than one revision Jun 18 11:26:12 so we can ignore that case for now Jun 18 11:26:44 mickeyl: ok I'm going to change bestProvider Jun 18 11:27:04 zecke: thank you! Jun 18 11:27:17 i will adapt the shell afterwards Jun 18 11:27:23 shell uses this feature, too Jun 18 11:27:28 method, even Jun 18 11:27:57 mickeyl: no need to (I will use a default parameter) Jun 18 11:28:10 zecke: alright, even beter Jun 18 11:28:13 better as well Jun 18 11:29:35 *siesta* Jun 18 11:30:20 good idea Jun 18 11:30:25 i'm totally exhausted by squash Jun 18 11:30:33 we played 3 hours today Jun 18 11:30:38 my bones will thank me tomorrow Jun 18 11:30:39 heh Jun 18 11:31:58 mickeyl: you still have bones Jun 18 11:32:10 these old fashioned seniors, titan is the future Jun 18 11:32:32 3and after 20050701 we even get our souls back! Jun 18 11:33:21 koen|tv: what should we do with it? Jun 18 11:33:43 zecke: sell them to novell of course Jun 18 11:33:53 and get filthy rich Jun 18 11:34:37 koen|tv: could idea? I think Novell should create an European Ximian Mobile Division Jun 18 11:34:46 I would love to be a stupid monkey Jun 18 11:34:59 zecke: in the "nano-1.2.1" case, why are you asking for the best provider for "nano" rather than the best provider for "nano-1.2.1"? Jun 18 11:35:13 I don't think there should really be any need to special-case the version thing. Jun 18 11:35:34 KeyError: nano-1.2.1 Jun 18 11:37:00 zecke: pardon? Jun 18 11:37:22 nano-1.2.1 is definitly in the list of available providers Jun 18 11:37:28 i can see it in bbshell nano Jun 18 11:37:43 since bitbake.conf contains PROVIDES=PN-PV Jun 18 11:37:49 We've bestProvider uses the 'PackageName -> [Version]' array Jun 18 11:37:55 zecke: http://udu.wiki.ubuntu.com/%C2%B5buntu Jun 18 11:38:32 is pivotboot-image maintained thse days ? Jun 18 11:38:34 zecke: hm, that doesn't seem right. Surely it should be considering everything in ${PROVIDES}, not just the package name Jun 18 11:38:44 and, as mickey says, PROVIDES contains ${P} Jun 18 11:38:51 I'm looking for an image type which could be easily modified into an initrd Jun 18 11:39:17 pb_: no doubt that r236 introduced some bugs Jun 18 11:39:37 ~praise RP for 2.6.12-rc6 on c7x0 being much faster than previous versions Jun 18 11:39:39 All hail RP for 2.6.12-rc6 on c7x0 being much faster than previous versions! Jun 18 11:39:48 Jun 18 11:40:01 GPE or qutopia for environment Jun 18 11:40:50 mickeyl: very good! what did he do? Jun 18 11:41:36 koen|tv: did he ever hold an ipaq in his hands? Jun 18 11:41:51 zecke: probably not Jun 18 11:42:06 pb_: there was something in copy_user regarding preempt Jun 18 11:42:12 pb_: i don't remember details Jun 18 11:42:16 zecke: and they seem to have a bunch of clueless zaurus users in their ranks Jun 18 11:42:17 perhaps too many locks Jun 18 11:42:28 mickeyl: removing the preempt lock from copy_page Jun 18 11:42:57 excellent. booting nearly 1.5 as fast Jun 18 11:43:10 pb_: It was Russell removing the preempt lock from clear_page that makes the biggest difference although the changed locking on copy_page will help as well. I just made his patches actually work... Jun 18 11:43:36 mickeyl: Just don't try suspend/resume :-( Jun 18 11:43:47 zecke: maybe we should point them to handhelds.org and say "wheel! don't reinvent" Jun 18 11:43:48 mickeyl: It is so much more responsive though :) Jun 18 11:43:54 who needs suspend/resume? :) Jun 18 11:43:59 pb_: 1st it checks nano-1.2.1 in providers Jun 18 11:44:01 pb_: and it passes Jun 18 11:44:13 pb_: 2nd it builds a list of PNs providing nano-1.2.1 Jun 18 11:44:19 pb_: it finds 'nano' Jun 18 11:44:29 pb_: 3rd it asks what is the best provider for nano Jun 18 11:44:33 pb_: it gets 1.3.5 Jun 18 11:44:42 in 2nd there is the bug Jun 18 11:44:46 nano doesn't provide nano-1.2.1 Jun 18 11:44:50 or at least shouldn't Jun 18 11:45:02 yeah Jun 18 11:45:11 I think this is similar to the gcc-cross-kernel thing we had the other week. Jun 18 11:45:29 isn't that the 5 lines of priority sorting the provider list that 235-236 eliminated ? Jun 18 11:45:46 if you asked for gcc-cross-kernel-3.3.3, bitbake would go "oh, right, gcc-cross-kernel provides that, and the newest one is 3.3.4", so you'd get the wrong compiler and the build would break. Jun 18 11:46:44 pb_: hehe why did you not file a bug report against bitbake ;) Jun 18 11:47:10 dunno, I guess I was suffering from a temporary bout of stupidity. Jun 18 11:47:15 heh Jun 18 11:48:56 like mickey says, I think step 2 is bogus. Generating a PN -> provider map involves destroying information, because you can have multiple bbfiles with the same PN and different PROVIDES. Jun 18 11:50:16 in fact, a package close to your heart is a good example of this. Jun 18 11:50:25 qte/qte_2.3.10.bb:PROVIDES = "virtual/qte virtual/libqte2" Jun 18 11:50:25 qte/qte_4.0.0-snapshot.bb:PROVIDES = "virtual/qte4 virtual/libqte4" Jun 18 11:50:42 clearly it would be disaster if you asked for virtual/libqte4 and bitbake gave you qte 2.3.10. Jun 18 11:51:22 * koen|tv sees the horrendous crosspost Jun 18 11:51:31 though it would be the other way round wouldn't it? Jun 18 11:51:59 pb_: Qt/E is broken anyway? right? Jun 18 11:52:00 well, it depends. normally, yeah, but PREFERRED_VERSIONs and package sets might change that. Jun 18 11:52:16 koen|tv: crosspost? Jun 18 11:52:31 the perl problem from mr harpin Jun 18 11:52:33 * zecke is losing some more hairs and idling Jun 18 11:52:52 for i in mailinglists ; cc:$i ; done Jun 18 11:52:53 zecke: no doubt, but you should try to put that out of your mind Jun 18 11:52:55 zecke: you're young, you can afford losing some hair ;) Jun 18 11:53:02 mickeyl: lol Jun 18 11:53:11 no kidding I will bold before getting thirty Jun 18 11:53:16 will be even Jun 18 11:53:43 mickeyl: maybe qtconfig now works for me Jun 18 11:54:02 mickeyl: its mainwidget was at least 2000000x2000000 Jun 18 11:55:31 03pb 07 * r1.3565 10openembedded/packages/linux/ (linux-epia-2.6.12/epia_defconfig linux-epia_2.6.12.bb): update linux-epia to 2.6.12; disable in-kernel drm Jun 18 11:55:39 zecke: heh Jun 18 11:55:51 zecke: qtdemo doesn't work here :/ i just get the framework, but not the apps Jun 18 11:56:38 mickeyl: they're slow over xdmcp Jun 18 11:56:45 pb_: master of the X Jun 18 11:57:15 pb_: can an app find out if the XServer is somewhere else than the client? Jun 18 11:57:25 without using getenv('DISPLAY') Jun 18 11:57:45 zecke: no. this is the essence of network transparency. Jun 18 11:57:55 you could try requesting shm Jun 18 11:58:09 even DISPLAY won't work, since you might be talking to a local proxy like sshd. Jun 18 11:58:45 ah okay Jun 18 12:01:46 hmm Jun 18 12:01:57 a c7x0 without suspend/resume is slightly suboptimal Jun 18 12:02:02 * mickeyl gets back to the battery thing Jun 18 12:03:25 hmm Jun 18 12:03:32 who removed the vtapplet ? Jun 18 12:03:51 mickeyl: the no-fluff gang Jun 18 12:03:56 grrr Jun 18 12:04:20 that would be my first guess Jun 18 12:05:34 aah Jun 18 12:05:36 false alarm Jun 18 12:05:44 Ciao all Jun 18 12:13:04 ~praise SirFred for the accellerated w100 driver Jun 18 12:13:05 All hail SirFred for the accellerated w100 driver! Jun 18 12:15:36 mickeyl: obviously 2nd is wrong. We should find the best provider from a set of possible providers? Jun 18 12:15:55 righto Jun 18 12:17:12 * zecke watches werner Jun 18 12:17:49 mickeyl: I have to admit I did not know the album "Debil" but El Cattivo really rocks Jun 18 12:18:31 zecke: heh. listen to the rest of the album, most things are really cool. i love this early phase of that band Jun 18 12:18:46 mickeyl: they were funny Jun 18 12:19:05 mickeyl: Erna P. Jun 18 12:19:06 03pb 07 * r1.3566 10openembedded/packages/lirc/ (lirc_0.7.1.bb files/lircd.init): check for lircd in ${sbindir} Jun 18 12:19:27 oh that isn't on Debil Jun 18 12:22:37 * mickeyl needs to watch some tv to make the headache go away (or make it increase) Jun 18 12:23:51 OT: does anyone here own a Diablo? Jun 18 12:24:15 zecke, do you mean "Lamborghini" Diablo ? Jun 18 12:24:27 Maybe in the next 20 lifes :) Jun 18 12:24:44 Pigi: http://www.jonglerie.de/ Jun 18 12:24:45 If I'm able to pass my money from one life to the other . Jun 18 12:25:34 zecke: uh. why? Jun 18 12:25:38 03pb 07 * r1.3567 10openembedded/BitKeeper/deleted/ (2 files): remove some old, broken xservers Jun 18 12:26:10 want to get me one Jun 18 12:27:06 zecke, can't find it, Maybe my german need to improve before I can surf that site ;) Jun 18 12:28:12 zecke: I think I do Jun 18 12:28:58 how much do you think a good 3970 used would cost ? Jun 18 12:29:27 Pigi: I once again pasted the wrong link Jun 18 12:29:52 Pigi: E150 or so Jun 18 12:30:28 wow. I'm following an auction on ebay, but I think it would go a bit up on this. Crossing my finger Jun 18 12:30:58 Pigi: you can get a new ipaq for E200 orso, keep that in mind Jun 18 12:31:08 s/orso/or so/ Jun 18 12:31:18 where can I get a new one for 200E ? Jun 18 12:31:31 not in italy, for sure :) Jun 18 12:32:22 an h1900 series should be cheap Jun 18 12:32:31 and a hx2110 is E 275 here Jun 18 12:32:55 koen|tv, you're right, but the familiar/linux support is yet to be finished, if I'm right ... Jun 18 12:33:24 and I don't really wan't the evil microsoft on any of my achine ;) Jun 18 12:33:27 if the 2110 is like the 2750 the support is quite good Jun 18 12:33:52 Pigi: see http://handhelds.org/moin/moin.cgi/SupportedHandheldSummary Jun 18 12:33:56 that should be up to date Jun 18 12:34:25 I think the only ipaq that you can easily buy new for 200 euros would be the rz1710, which doesn't have much support yet. Jun 18 12:34:43 the h1940 would indeed be quite cheap if you can find one, but the support even for those isn't the best. Jun 18 12:35:26 we should really port faster :( Jun 18 12:35:50 * zecke throws away buildProvider Jun 18 12:35:59 ! Jun 18 12:41:16 03koen 07 * r1.3564.1.1 10openembedded/ (2 files in 2 dirs): Jun 18 12:41:16 Move kdrive 20050610 - 20050617 Jun 18 12:41:16 Add ipaq-pxa270.conf, this should go away after we can build a mixed pxa25x and pxa27x kernel Jun 18 12:46:18 I wonder if would be a good choice to get a sharp 5500 or similar. I think this auction would go too high :) Jun 18 12:46:38 Pigi: is you like having 14 of flash.... Jun 18 12:46:42 14MB even Jun 18 12:46:50 s/is/if/ Jun 18 12:46:55 * koen|tv gets some coffee Jun 18 12:47:00 heh Jun 18 12:47:57 gosh... 14 mb is really few.... Jun 18 12:48:09 * Pigi remove a lot of item from "MyEbay" Jun 18 12:48:22 sharp and sane hardware design just doesn't combine Jun 18 12:48:59 no, they really don't. Jun 18 12:49:16 although the cx60 units are nice Jun 18 12:49:36 the rest either lacks ram or connectivity Jun 18 12:49:58 anyone use pantheon here? Jun 18 12:50:12 pantheon? Jun 18 12:50:41 i've visited it when I was in rome Jun 18 12:50:56 koen|tv: how far is the maemo integration? Jun 18 12:51:06 The cx60 units are lovely :) Jun 18 12:51:08 zecke: about 95% Jun 18 12:51:14 I'm not at all biased :) Jun 18 12:51:46 zecke: the remaining 5% deals with stabbing initscripts in the eye Jun 18 12:52:48 zecke: bitbake maemo-image should do the job, after you fix #106 Jun 18 12:53:42 koen|tv: That's nice to know :) Jun 18 12:54:15 RP: florian did a good job with spamming their bugzilla Jun 18 12:54:51 koen|tv: excellent :) Jun 18 12:55:36 koen|tv: Do you know if you're getting a 770? Jun 18 12:55:44 RP: and we have our inside guy at nokia: luc Jun 18 12:55:49 RP: Q3.... Jun 18 12:56:36 Inside guys are good :) Jun 18 12:56:50 http://www.intercept.com/html_files/products/pantheon/pantheon_main.htm Jun 18 12:56:57 pantheon is an EDA Jun 18 12:57:18 a group of developers who split from cadence/orcad Jun 18 12:57:39 aaargh! Jun 18 12:57:42 orcad!!! Jun 18 12:57:53 I wasted 4 weeks of my life with that crap Jun 18 12:58:11 especially with the 'save' bug Jun 18 12:58:38 lol Jun 18 12:58:43 I think I ended up using protel last time I was doing that kind of thing... Jun 18 12:58:52 well cadence bought orcad back in 98/99 Jun 18 12:59:06 yeah i am looking for a linux EDA Jun 18 12:59:12 protel is windows only Jun 18 12:59:21 and now ~10k/license Jun 18 13:00:08 eek! Jun 18 13:01:01 yeah it used to be 2k Jun 18 13:01:20 then they added a huge amount of stuff Jun 18 13:01:41 In my case, it was provided to do a job... Jun 18 13:01:45 i curse at it every time it loads up when i dont want it to Jun 18 13:02:03 usually trying to launch mplab or quartus Jun 18 13:02:42 ah well, lets see if intercept has a good student price/demo Jun 18 13:05:54 RP: I'm still waiting for the kernelsources, so I can add them to OE.... Jun 18 13:06:20 koen|tv: You and everyone else... Jun 18 13:07:16 RP: and farsight + libmimic integration, of course Jun 18 13:07:58 * koen|tv should poke robtaylor about adding that to OE Jun 18 13:09:06 03zecke 07 * r1.3557.1.10 10openembedded/packages/opie-reader/ (opie-reader_cvs.bb opie-reader.inc): Jun 18 13:09:07 opie-reader: Jun 18 13:09:07 build Flite Speach output for the CVS version Jun 18 13:09:07 Patch courtsey Patrick Ohly Jun 18 13:09:10 03zecke 07 * r1.3557.1.9 10openembedded/packages/flite/flite_1.2.bb: Jun 18 13:09:10 flite: Jun 18 13:09:10 use oe_libinstall to stage libraries. This is courtsey Jun 18 13:09:11 Patrick Ohly from #82 Jun 18 13:09:31 anyon nkow if galaxyhack compiles in bb? Jun 18 13:09:44 requires boost-filesystem Jun 18 13:10:35 * koen|tv gets some snacks and settles in for the English Detective night. Jun 18 13:11:06 * RP gets annoyed with the -rc6 kernels Jun 18 13:17:39 Its 2.6.12-rc6-mm1 that's the problem, not 2.6.12-rc6... Jun 18 13:40:07 03pb 07 * r1.3571 10openembedded/packages/xserver/xserver-xorg_6.8.99.11.bb: update xserver-xorg to 6.8.99.11 Jun 18 13:40:14 03pb 07 * r1.3567.1.1 10openembedded/packages/xserver/xserver-xorg_6.8.99.10.bb: enable DRI and XvMC Jun 18 13:54:20 it's time to sleep now Jun 18 13:54:25 'night all Jun 18 15:12:16 lol that was easy Jun 18 15:12:46 RP: *sorry* could you ask ibot once more to give the uri to the bitbake c parser? Jun 18 15:16:08 ~bbp Jun 18 15:16:09 [bbp] A preliminary .bb parser written in C using 'lemon' for the parser. See http://projects.buici.com/oe/. Jun 18 15:16:50 Guess who's fault the problems in 2.6.12-rc6-mm1 are? :) Jun 18 15:16:51 * zecke hopes not to delete tmp/ this time Jun 18 15:16:57 RP: yours? Jun 18 15:17:03 :} Jun 18 15:17:13 zecke: rmk's :) Jun 18 15:17:26 here we go again... Jun 18 15:17:40 hehe, who uses intel cpus anyway Jun 18 15:18:33 RP: tell me more Jun 18 15:20:04 Its some arm5 smp code which somehow breaks suspend/resume. I have no idea why without spending hours working out what it does. I'll complain in the first case, just in case its something obvious Jun 18 15:21:14 03zecke123 * r264 10bitbake/bin/bitbake: Jun 18 15:21:15 bitbake/bin/bitbake: Jun 18 15:21:15 -Possible fix for #103 make findBestProvider operate Jun 18 15:21:15 on a given, limited pkg_pn (the one we created). Jun 18 15:21:48 btw, is there a webinterface for bitbake (yet)? Jun 18 15:21:54 mickey|tv: after having idled so long... Jun 18 15:22:00 TheCount: webinterface for what? Jun 18 15:22:02 i.e. for managing it/the project(s)/the releases Jun 18 15:22:14 TheCount: hehe god no ;) Jun 18 15:22:25 TheCount: bitbake 1.x.x are managed by bugs.openembedded.org Jun 18 15:22:49 zecke: that's not what I meant ;9 Jun 18 15:22:49 TheCount: the TODOs will be there as bug reports as well Jun 18 15:23:04 zecke: ok, then I've not been doing duplicate work the last months. Jun 18 15:23:19 TheCount: so tell us what you did? Jun 18 15:25:35 zecke: we did a 'Fallstudie' regarding 'Build- und Releasemanagement fuer Embedded Systeme auf OpenSource Basis' for our academy title, Jun 18 15:26:17 zecke: and we ended up replacing CVS with Subversion, a self-made buildsystem with bitbake, and a non-existing release-process with bakerman, a webinterface for bitbake, Jun 18 15:26:43 zecke: basically handling contact information, developers, machines, distributions, packages, their options, the projects they form, and the builds that result Jun 18 15:27:18 TheCount: sounds interesting Jun 18 15:28:04 zecke: oh, it is. addresses a bunch of management problems in a developement context Jun 18 15:28:05 TheCount: packages is something koen wanted to do and I'm currently trying to create some kind of Q&A structure Jun 18 15:28:17 zecke: uh? Jun 18 15:28:23 but my knowledge of Q&A is pretty limited Jun 18 15:29:00 TheCount: We've bitbake_qa it is a module (using bitbake) to lint, find missing sources, find packages not applying patches properly Jun 18 15:29:09 zecke: I think you mean QA - Q&A = Questions and Answers :) Jun 18 15:29:13 TheCount: also we will soon have a tinderbox + tinderclients available Jun 18 15:29:24 I'm so freaking confused Jun 18 15:29:46 It would be nice if some of you could start a fresh build, I'm going to release 1.3.1 soon Jun 18 15:29:59 zecke: ah, ok. that's not what we're doing. we're basically listing the packages from the openembedded tree, creating a meta-package for the project which depends on them, and provide include-.bbs which can send e.g. configure-options for them Jun 18 15:30:27 zecke: I'll build collie from scratch Jun 18 15:31:21 TheCount: koen will be happy to see your web intergace Jun 18 15:31:25 interface even Jun 18 15:31:48 zecke: it's supposed to be done next saturday, when we'll present it in the academy Jun 18 15:32:01 zecke: the month after that we'll enter our paper Jun 18 15:32:37 zecke: and inbetween we can take care of some integration to bitbake itself - so far, it's just the webinterface without real interaction with bitbake Jun 18 15:32:39 TheCount: how did you cope with the latest bitbake changes? Jun 18 15:32:50 zecke: didn't have to ;) Jun 18 15:33:00 ah good Jun 18 15:33:49 zecke: what were the latest changes? Jun 18 15:35:52 TheCount: some API changes Jun 18 15:36:00 TheCount: random breakage Jun 18 15:37:14 zecke: ah. the stuff interfacing the database to bitbake wasn't even started yet, so .. no problem there. got to graciously avoid mentioning that during the presentation, thou Jun 18 15:37:20 03zecke123 07bitbake-1.3.1 * r265 10/: Tag/Copy bitbake-1.3.1 now, it might be the final one... Jun 18 15:37:52 TheCount: at which faculty do you work? Jun 18 15:38:07 zecke: no faculty, a business academy Jun 18 15:39:02 (pun intended) Jun 18 15:41:26 03RP 07 * r1.3572 10openembedded/packages/linux/linux-openzaurus_2.6.12-rc6-mm1.bb: linux-oz-2.6: Revert the git-arm-smp.patch as it breaks suspend/resume Jun 18 15:43:03 mickey|tv: Even the suspend/resumes are snappy :) Jun 18 15:43:17 Perhaps now I can try and fix some of our bugs instead of other peoples... Jun 18 15:51:17 * TheCount coughs on his saturday evening bootdrink Jun 18 15:51:56 zecke: VWA (Verwaltungs- und WirtschaftsAkademie), studying for 'Wirtschaftsdipl.Informatik-Betriebswirt' Jun 18 15:59:00 'night all Jun 18 16:10:33 ok sleep well guys Jun 18 16:11:43 'night zecke Jun 18 16:15:32 RP: are you going to attend the LinuxTag in Karlsruhe? Jun 18 16:19:09 zecke: As things stand, I've not planned to... Jun 18 16:21:19 Sadly it would conflict with other things I have planned anyway Jun 18 16:23:10 ah okay Jun 18 16:23:12 cya Jun 18 16:23:30 well 2,6,12 is finally out I see Jun 18 16:28:13 jacques: Yes, at last :) Jun 18 16:28:21 03RP 07 * r1.3573 10openembedded/packages/linux/linux-openzaurus_2.6.12-rc6-mm1.bb: linux-oz-2.6: Add patches for oprofile backtrace on arm, preempt fixes for mwfpe and corgi_ts pmu sharing Jun 18 16:32:27 RP, how many of your fixed made it in? Jun 18 16:32:30 fixes Jun 18 16:38:10 jacques: Enough of them so that 2.6.12 should actually run on pxa/zaurii ;-) Jun 18 16:38:20 RP, wow that's really good news Jun 18 16:38:42 maybe we'll try it on the ixp420 (nslu2) after all :-) Jun 18 16:38:51 jacques: There's a lot of code queued up for inclusion in the kernel post 2.6.12 so I'm going to be busy submitting patches... Jun 18 16:39:07 RP, aha, so we should wait a bit then Jun 18 16:39:18 jacques: I haven't tried 2.6.12 but I know 2.6.12-rc6 worked... Jun 18 16:39:38 jacques: No, I'd go for it. THe patches are more zaurri related Jun 18 16:39:56 You can see them in the above .bb file Jun 18 16:40:00 thanks for the info, now I have something to tell the eager kernel beavers in #nslu2-linux Jun 18 16:41:03 jacques: The copypage/clearpage changes are a massive speedup on the c7x0 - 2.6.12 should have all the appropriate code in it Jun 18 16:41:25 RP, the nwfpe preempt thing shouldn't affect softfloat users ? Jun 18 16:41:59 ooh I wanna run some benchmarks then, see if the copypage thing helps us (we don't call them slugs for nuthin') Jun 18 16:54:18 jacques: If you use softfloat, the nwfpe thing is irrelavent Jun 18 16:56:01 RP, thanks, that's what I hoped Jun 18 20:20:17 RP, you around? did something change with serial port during boot for 2.6.12 ? Jun 18 20:20:27 I no longer get boot messages **** ENDING LOGGING AT Sat Jun 18 23:59:56 2005