**** BEGIN LOGGING AT Mon Jun 20 23:59:56 2005 Jun 21 01:31:00 good morning all Jun 21 01:35:23 morning koen Jun 21 01:36:15 hey do13 Jun 21 01:38:10 morning all! Jun 21 01:39:43 has anyone made a package for csoap yet? Jun 21 01:40:12 not that I know off Jun 21 01:40:26 ok, so I keep on editing :) Jun 21 01:51:48 moring Jun 21 01:51:52 morning, too Jun 21 01:53:11 morning mickey Jun 21 01:53:23 hi dirk, moin cedric Jun 21 01:54:04 morning Jun 21 02:00:08 morning Jun 21 02:02:01 hi hrw|work Jun 21 02:03:08 hrw|work: ping Jun 21 02:03:16 proti: pong Jun 21 02:03:30 hrw|work: Hi, how do you reproduce bug #88 Jun 21 02:03:51 proti: w8 Jun 21 02:04:14 proti: bitbake -i;parse;build something Jun 21 02:04:22 proti: edit something.bb Jun 21 02:04:30 proti: rebuild something in bitbake Jun 21 02:04:41 without doing parse again Jun 21 02:05:02 Ok, I will try to reproduce and have a look at it. Jun 21 02:05:05 ok Jun 21 02:05:12 hrw|work: thx Jun 21 02:05:22 np Jun 21 02:06:06 hrw|work: morning. in which .bb file should I add the USE_NLS stuff for uclibc ? Jun 21 02:06:32 all gcc-cross? Jun 21 02:06:37 gcc-cross-initla? Jun 21 02:06:40 initial even Jun 21 02:06:43 morning BigAl Jun 21 02:06:46 morning Jun 21 02:06:59 mickeyl: gcc-cross Jun 21 02:07:16 hrw|work: all of them ? Jun 21 02:07:35 or gcc3-build-cross.inc ? Jun 21 02:07:45 mickeyl: gcc3-build-cross.inc will be best Jun 21 02:07:58 morning all Jun 21 02:08:19 hrw|work: right. will do it there. USE_NLS = "1" ? Jun 21 02:08:21 morning RP Jun 21 02:08:46 mickeyl: USE_NSL="no" Jun 21 02:08:54 s/NLS/NSL Jun 21 02:08:55 RP: when you have a chance, please check 0.15rc2. I've added a patch that appeared on the orinoco list that's supposed to fix the firmware loading. Jun 21 02:09:00 morning RP Jun 21 02:09:04 hrw|work: k Jun 21 02:09:16 mickeyl: USE_NLS_gcc-cross = "no" Jun 21 02:09:28 mickeyl: That sounds good. I'll test when I get a chance Jun 21 02:09:31 thats from my local.conf - found in conf/documentation.conf Jun 21 02:09:42 morning RP Jun 21 02:09:48 hi RP Jun 21 02:09:50 mickeyl: I think I know what the suspend/resume problem is with the c7x0 and its not apm :) Jun 21 02:10:42 RP: d'oh. what's it then ? Jun 21 02:11:18 mickeyl: I think its the keyboard driver - now the device is so fast, the keyboard driver sees the resume button pressed and resuspends... Jun 21 02:12:02 RP: wooah Jun 21 02:12:07 that's kind of funny :) Jun 21 02:12:59 ;)) Jun 21 02:13:17 Of course the serial port gets turned off over a suspend so you don't see that part of the log and get very confused :) Jun 21 02:13:17 hrw: USE_NLS = '${@base-conditional( "TARGET_OS", "linux-uclibc", "no", "", d )}' Jun 21 02:13:32 hrw|work: into gcc-cross-build, so we don't need to add _cross Jun 21 02:13:40 hrw|work: please test, when you have a chance Jun 21 02:13:43 ok Jun 21 02:14:01 i'll push in a minute Jun 21 02:14:12 RP: congrats for finding _that_ one! Jun 21 02:15:04 mickeyl: will do - my attempt to build gcc-cross 4.0.0/mipsel-uclibc failed again (will look does it java again or not) Jun 21 02:15:14 mickeyl: I just need to work out a fix now :) Jun 21 02:16:37 mickeyl: bug in bitbake shell? Jun 21 02:16:38 BB>> filebuild gcc-cross- Jun 21 02:16:38 gcc-cross-initial_3.3.3.bb gcc-cross-initial_3.4.4.bb gcc-cross-kernel-3.3.3_3.3.3.bb Jun 21 02:16:43 gcc-cross-initial_3.3.4.bb gcc-cross-initial_csl-arm-20050416.bb gcc-cross-kernel-3.3.4_3.3.4.bb Jun 21 02:16:45 gcc-cross-initial_3.4.3.bb gcc-cross-initial_csl-arm-20050603.bb gcc-cross-sdk_3.3.4.bb Jun 21 02:16:48 no other choices.. Jun 21 02:17:13 ops.. my bug sorry Jun 21 02:17:20 *phew* :) Jun 21 02:17:23 have to finish my morning 0.5l of coffee Jun 21 02:17:27 hehe Jun 21 02:18:12 mickeyl: did you see the branch naming convention email I sent you? Jun 21 02:18:25 mornin' Jun 21 02:18:39 RP: congrats too. How many bugs did you catched? Jun 21 02:19:04 hi pb Jun 21 02:21:03 hi hrw|work Jun 21 02:21:14 pb_: any comments on the monotone branch naming conventions? Jun 21 02:22:23 rwhitby: I don't really know enough about monotone to have any opinion on that yet. Jun 21 02:22:37 mickeyl: there is 'peek' - where is 'poke'? I want to change one var in gcc-cross inside shell Jun 21 02:22:45 BB>> peek gcc-cross PARALLEL_MAKE Jun 21 02:22:46 -j 4 Jun 21 02:22:55 rwhitby: i didn't receive anything from you yet Jun 21 02:22:56 how to change it to '' in shell Jun 21 02:23:11 hrw|work: poke is currently outcommented because it doesn't work due to some things i don't understand in the smart_dict implementation :/ Jun 21 02:23:20 need to pester zecke on that Jun 21 02:23:23 aha Jun 21 02:23:31 you can poke, but the change is not recognized Jun 21 02:23:39 so it's of little use atm. Jun 21 02:23:40 :/ Jun 21 02:24:20 rwhitby: please resend it to mickeyl@handhelds.org, perhaps it got lost Jun 21 02:24:24 hi koen Jun 21 02:24:28 hey mickeyl Jun 21 02:25:18 rwhitby: our great uni seems to have misplaced the server.... Jun 21 02:25:39 so we need to find it first to press the reset button :( Jun 21 02:25:43 urhgs Jun 21 02:26:13 gcc 3.x: set USE_NSL=no if TARGET_OS=linux-uclibc Jun 21 02:26:30 mickeyl: thx Jun 21 02:26:33 np Jun 21 02:26:36 hey hrw|work Jun 21 02:26:47 hi koen Jun 21 02:29:30 koen: thx for your reply on the monotone branch email Jun 21 02:29:48 "Speicherzugriffsfehler" Jun 21 02:30:03 German is such a funny sounding language ;) Jun 21 02:30:10 yes Jun 21 02:30:26 "I'm not sure what a Speicherzugriffsfehler is, but it doesn't sound good" - very nice sentence Jun 21 02:30:55 this is great Jun 21 02:31:03 gave me the first lough today :) Jun 21 02:31:26 you gotta love british humour Jun 21 02:34:55 koen: mickeyl: if you want to send me a monotone pubkey you can use our mtn.nslu2-linux.org server for testing while they are finding the server at koen's uni :-) Jun 21 02:35:40 rwhitby: i will do that, thanks. I hope that I have some time today to setup the environment Jun 21 02:36:44 I will open up write access completely to all keys while we are testing Jun 21 02:37:35 anyone know the comment character in lua ? Jun 21 02:38:21 -- Jun 21 02:38:33 for one liners Jun 21 02:38:35 for blocks: Jun 21 02:38:40 --[[ Jun 21 02:38:41 ... Jun 21 02:38:42 ... Jun 21 02:38:44 ]] Jun 21 02:38:56 ugh :-\ Jun 21 02:39:18 that's what I thought as well Jun 21 02:39:31 rwhitby: mail sent to your nslu2-linux.org address Jun 21 02:40:51 ok Jun 21 02:41:14 I'll set it up to serve the org.* collection Jun 21 02:41:30 do13: The problem with 2.6.12-rc6-mm1 was the git-arm-smp patch - its fixed in oe now Jun 21 02:41:47 do13: The rest of the problems are issues within corgikbd :-/ Jun 21 02:42:11 koen: scratch that, it's currently serving the openembedded.* collection, and we'll stick with that (with the understanding that it will change if it goes into production) Jun 21 02:42:13 rwhitby: did you find a monotone versions for 'bk changes -R'? Jun 21 02:42:31 nope, but haven't had time to look (been at work all day, just got home) Jun 21 02:42:37 * RP chases the input people about the previouis corgikbd patches Jun 21 02:42:59 RP: speaking about the input people... Jun 21 02:43:12 please give me blinkin' lights Jun 21 02:43:18 :) Jun 21 02:43:38 I really need to sort that out :) Jun 21 02:45:21 RP: Did you tried 2.6.12-mm1? Jun 21 02:45:58 do13: Not yet. I should really give that a try. See what's broken next :) Jun 21 02:49:56 RP: Currently I'am compiling a new Image with this kernel. I will test after work and give you a feedback. Jun 21 02:50:44 do13: Ok, thanks. Hopefully all the arm-smp problems were resolved by the time that came out! :) Jun 21 02:57:29 http://lists.gnu.org/archive/html/monotone-devel/2005-06/msg00034.html contains some nice use cases Jun 21 03:03:26 yeah right Jun 21 03:03:44 mickeyl: rebuilding gcc-cross 3.4.4 Jun 21 03:06:47 koen: you should be right to write to mtn.nslu2-linux.org openembedded Jun 21 03:08:45 monotone: read from fd 4 (peer mtn.nslu2-linux.org) failed, disconnecting Jun 21 03:11:02 try again Jun 21 03:11:23 rwhitby-away: I think we should recreate the repos with the correct namespace Jun 21 03:11:31 it works now Jun 21 03:11:33 (had to restart to give access) Jun 21 03:11:45 want me to restart it with org.* ? Jun 21 03:12:03 (I agree about the recreating) Jun 21 03:12:04 yes, pleaes Jun 21 03:12:25 restarted Jun 21 03:12:34 bbia Jun 21 03:12:36 bbiab Jun 21 03:17:43 mickeyl: NOTE: [12:11] package gcc-cross-3.4.4-r0: task do_build: completed Jun 21 03:17:50 mickeyl: thx for last change Jun 21 03:18:49 hrw|work: :) yet another success story for ${@base_conditional} :)) Jun 21 03:23:03 ;) Jun 21 03:26:51 hi zecke Jun 21 03:27:34 moin zecke Jun 21 03:27:48 hey Jun 21 03:31:26 mickeyl: did you want to send a monotone key too? Jun 21 03:31:50 koen: server is back up with org.* collection served (so nothing in it at the moment) Jun 21 03:31:57 ok Jun 21 03:32:21 I'll make a new monotone repo with the right namespaces Jun 21 03:32:45 rwhitby: sure, i'm wading through the tutorial again... it's been some weeks since I did that the last time Jun 21 03:32:48 * mickeyl mt init Jun 21 03:32:51 * mickeyl mt genkey Jun 21 03:32:51 koen: any chance you can do a last pull from bk://nslu2-linux.bkbits.net/openembedded before you convert? Jun 21 03:34:09 we're considering stopping all nslu2-linux bk changes early, so that you guys don't have to worry about us when you're converting the db Jun 21 03:34:29 we've just tagged OpenSlug 2.0, so now would be a good time for us to freeze before the transition Jun 21 03:34:36 pulling Jun 21 03:36:21 rwhitby,koen: mt key sent Jun 21 03:36:35 i will send another key for my home repository later tonite Jun 21 03:37:00 mickeyl: you sent the fingerprint, use pubkey to get the whole key Jun 21 03:37:02 mickeyl: can bitbake-shell do "build one two three"? Jun 21 03:37:10 rwhitby: oops, ok Jun 21 03:37:30 monotone --db=~/koen.db pubkey koen@handhelds.org | mail foo@bar.org Jun 21 03:39:00 hmm.. I think that one day I will need to setup scratchbox to populate mipsel site a bit.. Jun 21 03:40:01 | checking if trio library is needed... configure: error: cannot run test program while cross compiling Jun 21 03:41:38 hrw|work: unfortunately not yet. the command parser only features a fixed set of params yet Jun 21 03:41:48 hrw|work: >TODO :) Jun 21 03:43:36 WTF Jun 21 03:48:35 hey zecke Jun 21 03:48:54 koen, did that final nslu2-linux pull go okay? Jun 21 03:49:00 yes Jun 21 03:49:04 Cool! Jun 21 03:49:10 it is in my working copy of monotone Jun 21 03:49:24 Sounds good. Jun 21 03:49:30 I was about to merge it into org.openembedded.koen branch Jun 21 03:49:33 I'm still getting myself up to speed on monotone. Jun 21 03:49:57 * koen has the tutorial and manpage opened Jun 21 03:51:18 Some of the commands are sort of familiar, but others are kind of vague. Jun 21 03:52:24 viewmtn is a really nice tool to overview it Jun 21 03:52:41 did the nslu2 web person manage to set it up yet? Jun 21 03:53:52 i think he looked at it; but didnt get around to it yet. Jun 21 03:57:01 mickey|lunch: key added to mtn.nslu2-linux.org Jun 21 04:00:30 hrw|work: I always ask myself is it really hard to create qpf fonts? Jun 21 04:01:36 rwhitby: is the server up and running? Jun 21 04:01:47 zecke: you need to create fontdir, set QTDIR and start makeqpf in all rotations Jun 21 04:01:57 hrw|work: or use makeall Jun 21 04:02:01 zecke: it quite easy and can be automated Jun 21 04:02:24 zecke: but takes time, need qvfb or fb access etc.. Jun 21 04:02:27 qpf fonts suxx Jun 21 04:02:46 rwhitby: monotone --db=../OE-koen.db push mtn.nslu2-linux.org org.openembedded does not succeed Jun 21 04:04:29 dominion.kabel.utwente.nl should be serving org.openembedded now Jun 21 04:07:47 koen: the server is not giving a good error :-( Jun 21 04:08:23 maybe I need to pull from yours first Jun 21 04:10:29 tip for young players - don't delete your only copy of your local db, as it contains your private key and then you need to generate new ones :-) Jun 21 04:14:30 rwhitby: reading of my tree isn't guarded by keys Jun 21 04:14:58 koen: sent you a new nslu2-linux key for writing, and will also send a new rwhitby key too Jun 21 04:15:14 ~lart openwrt for lack of ipkg status entries Jun 21 04:15:14 * ibot lowers openwrt's priority for lack of ipkg status entries Jun 21 04:15:30 koen: I will also delete our current nslu2-linux db, and start again from yours Jun 21 04:15:38 ok Jun 21 04:19:56 rwhitby: can you try pulling from my repo? Jun 21 04:21:06 monotone: connecting to dominion.kabel.utwente.nl Jun 21 04:21:06 monotone: network exception: failed to connect: Connection refused Jun 21 04:21:30 koen: I sent you my new pubkey Jun 21 04:21:49 hmmm Jun 21 04:21:58 can you nmap me? Jun 21 04:22:50 monotone@nudi:~$ telnet dominion.kabel.utwente.nl 5253 Jun 21 04:22:50 Trying 130.89.193.158... Jun 21 04:22:50 telnet: Unable to connect to remote host: Connection refused Jun 21 04:24:18 linux-oz-2.6: Add rmk's armsmp fix, fix collie typo, revise corgi_ts pmu patch for upstream, revise corgi_kbd suspend key handling - should fix all known suspend/resume button problems :) Jun 21 04:25:47 RP: 2.6.12-mm1? Jun 21 04:26:12 rc6-mm1 Jun 21 04:26:14 hrw|work: We're still using 2.6.12-rc6-mm1 but its near enough the same thing... Jun 21 04:26:30 RP: I have rc5-mm now Jun 21 04:26:58 The suspend/resume button handling is so much nicer with that patch Jun 21 04:27:04 no infinite loops :) Jun 21 04:27:14 or having to press the button quickly Jun 21 04:28:20 koen: working now Jun 21 04:28:21 ok, it was a rogue /etc/host entry Jun 21 04:28:39 syncing org.openembedded Jun 21 04:28:46 what bk changeset is that at ? Jun 21 04:30:00 OE 1.3581.1.2 Jun 21 04:30:11 and i suspect nslu2 cset 1.3580 Jun 21 04:31:54 this one: ? Jun 21 04:32:00 ChangeSet@1.3580, 2005-06-19 15:01:54-07:00, jbowler@marsco.kalmiopsis Jun 21 04:32:00 OpenSlug 2.0-beta tag Jun 21 04:32:00 (ready for testing) Jun 21 04:32:00 TAG: OpenSlug-2.0-beta Jun 21 04:33:15 that will be our final bk changeset. I've put the word out for no more nslu2-linux changes in our repo. Jun 21 04:33:45 we will wait until after the transition is complete before we make the next change Jun 21 04:34:02 ("we" and "our" refers to the nslu2-linux project) Jun 21 04:34:49 rwhitby: that one Jun 21 04:35:04 anything that was in 20 minutes ago should be in monotone Jun 21 04:35:57 * koen hugs bk export -tpatch r Jun 21 04:45:23 koen: got it, try syncing back Jun 21 04:45:46 monotone sync mtn.nslu2-linux.org org Jun 21 04:56:09 RP: excellent Jun 21 04:56:48 RP: how did you solve the key issue? Jun 21 05:02:18 raster: hey Jun 21 05:02:24 rwhitby: monotone sync mtn.nslu2-linux.org org.openembedded fails Jun 21 05:03:06 try just org Jun 21 05:03:13 moooh! Jun 21 05:03:39 monotone: read from fd 4 (peer mtn.nslu2-linux.org) closed OK after goodbye Jun 21 05:03:39 monotone: [bytes in: 424] [bytes out: 500] [revs in: 0] [revs out: 0] Jun 21 05:03:47 koen: cool Jun 21 05:04:03 we're synced then Jun 21 05:04:11 grrrr, viewmtn requires apache2 Jun 21 05:05:28 woohoo! server found Jun 21 05:05:35 mickeyl: I've made it register one press and then ignore the key for 1 second afterwards Jun 21 05:06:04 When it suspends, keys get disabled. When it resumes, the keys get reactivated but the 1 second lock gets applied to the power key Jun 21 05:06:55 RP: ah very good. nice and easy :) Jun 21 05:06:57 | fs/fat/dir.c:1269: fatal error: closing dependency file fs/fat/.dir.o.d: No space left on device Jun 21 05:07:00 argh.. Jun 21 05:07:12 rwhitby,koen: sorry, i goofed my database. I sent both of you my new key *sigh* Jun 21 05:07:34 heh Jun 21 05:07:59 note to self: don't remove $HOME/*.db Jun 21 05:08:52 mickeyl: that's two of us so far Jun 21 05:09:05 13:10 < rwhitby> tip for young players - don't delete your only copy of your local db, as it contains your private key Jun 21 05:09:09 and then you need to generate new ones :-) Jun 21 05:09:10 aaah Jun 21 05:09:12 hehe Jun 21 05:09:25 overlooked that Jun 21 05:09:50 mickeyl: you can pull from dominion.kabel.utwente.nl or mtn.nslu2-linux.org Jun 21 05:10:00 koen: ok, i'M trying that now with your server Jun 21 05:10:07 it's mt sync, not pull, right Jun 21 05:10:08 ? Jun 21 05:10:20 sync also tries to write Jun 21 05:10:26 ah ok Jun 21 05:10:32 very good. no alias necessary then :) Jun 21 05:11:10 less brain needed to switch :) Jun 21 05:12:09 monotone: warning: not currently serving requested collection 'org.openembedded.mickeyl' Jun 21 05:12:16 ah Jun 21 05:12:23 how's your branch called again? Jun 21 05:12:28 org.openembedded.koen? Jun 21 05:12:32 yes Jun 21 05:13:27 koen is serving org.openembedded, and mtn.nslu2-linux.org is serving org Jun 21 05:14:01 (while testing. in production mtn.nslu2-linux.org will be serving org.nslu2-linux) Jun 21 05:15:42 mickeyl: you can try adding the missing csets using bk export Jun 21 05:15:50 hmm Jun 21 05:15:58 still working on a pull Jun 21 05:16:03 mickeyl: use just "org" as the last arg to mtn.nslu2-linux.org Jun 21 05:16:20 how big would be monotone with the imported oe tree? Jun 21 05:16:29 hmm there's something wrong Jun 21 05:16:41 mickey@gandalf:/local/pkg/oe$ mt pull mtn.nslu2-linux.org org Jun 21 05:16:41 mickeyl: try sync instead of pull Jun 21 05:16:41 monotone: warning: doing anonymous pull Jun 21 05:16:41 monotone: connecting to mtn.nslu2-linux.org Jun 21 05:16:41 monotone: rebuilding merkle trees for collection org Jun 21 05:16:41 monotone: read from fd 4 (peer mtn.nslu2-linux.org) failed, disconnecting Jun 21 05:16:47 try sync Jun 21 05:16:58 ah that's better Jun 21 05:17:11 I guess I should enable anonymous access Jun 21 05:17:36 that'd be gooduntil we figured out how to serve it on oe.org Jun 21 05:18:21 I think you just have to sync first and can pull after that Jun 21 05:18:22 mickeyl: I sent a RFC to admin@ yesterday Jun 21 05:18:32 koen: excellent. Jun 21 05:19:11 mickeyl: collection -> org.openembedded or org, branch -> org.openembedded. Jun 21 05:19:43 koen: do you really want individual branches, or just all use the org.openembedded branch? Jun 21 05:20:03 if the latter, then mtn heads and mtn merge works automatically Jun 21 05:20:08 (I think) Jun 21 05:20:39 we need to try that out. automagical merges would be very important Jun 21 05:20:54 my idea was that the 'org.openembedded' branch would be the gold standard Jun 21 05:21:10 so maybe org.openembedded.dev ? Jun 21 05:21:12 and we do our thing in our own branches Jun 21 05:21:13 org.openembedded.bugs Jun 21 05:21:32 so we merge to o.oe.dev and sync Jun 21 05:21:39 koen: then you would need to propagate each time you committed Jun 21 05:21:41 I always found that my own work benefits from using two branches Jun 21 05:21:48 so we have one that is "outgoing to public" Jun 21 05:21:52 org.openembedded.dev Jun 21 05:22:01 and some of us have one additionally Jun 21 05:22:06 like me, org.openembedded.dev.mickey Jun 21 05:22:08 or so Jun 21 05:22:35 org.openembedded.dev.hrw org.openembedded.dev.hrw.not-to-send-to-public Jun 21 05:22:39 right Jun 21 05:22:46 "hey I need some help with foo, can you pull my branch?" Jun 21 05:22:57 hah Jun 21 05:23:17 "pull my finger ..." Jun 21 05:23:21 :D Jun 21 05:23:44 one common dev branch for all outgoing things sounds good to me Jun 21 05:23:56 what we do below this common branch is up to the individual dev Jun 21 05:23:59 so koen will have org.oe.koen.break-opie Jun 21 05:24:19 * mickeyl grin Jun 21 05:24:28 hrw|work: is it normal qpf will use 99% cpu for a while? Jun 21 05:24:49 monotone: successful exchange with mtn.nslu2-linux.org Jun 21 05:25:04 zecke: makeqpf? dont remember Jun 21 05:25:13 zecke: regarding your size question: Jun 21 05:25:16 -rw-r--r-- 1 mickey abvs 27653120 Jun 21 14:23 mickey.db Jun 21 05:25:56 does my tree need to be available from outside or can I have it totally without any external access? Jun 21 05:26:27 hrw|work: you can have it without external access Jun 21 05:26:42 cool Jun 21 05:27:24 mickeyl: complete oe history? Jun 21 05:27:35 mickeyl: what is that in human values? Jun 21 05:27:37 zecke: no no. fresh like a virgin Jun 21 05:27:39 27MB Jun 21 05:27:48 we agreed to ditch history for the new tree Jun 21 05:28:05 we will put the old history read-only somewhere in a cvs Jun 21 05:28:41 the old bk tree was 221MB Jun 21 05:28:45 that is with 3 years of history Jun 21 05:28:46 mickeyl: or a seperate monotone cvsimport tree Jun 21 05:29:00 koen: ya, would be an option Jun 21 05:29:09 koen: just something with a web interface Jun 21 05:30:01 viewmtn does a pretty goed job Jun 21 05:30:12 yo can get diffs and whole files Jun 21 05:30:15 you* Jun 21 05:30:16 2.6.12-mm1 works nicely so I'll upgrade the kernel tree Jun 21 05:30:21 RP: excellent Jun 21 05:30:36 * mickeyl ponders whether to take the shepherd or the spitz to LinuxTag Jun 21 05:31:26 re monotone: so is this still testing, or is this it now ? Jun 21 05:31:47 http://ewi546.ewi.utwente.nl/tmp/viewmtn/ Jun 21 05:31:52 (scp'ed the db) Jun 21 05:33:22 rwhitby: there is no real alternative that would give us the same level of painlessness Jun 21 05:33:46 rwhitby: so i think this is it. the only thing to do is to export a fresh tree on first of july and put it on oe.org as the main repo Jun 21 05:34:08 perhaps not even that Jun 21 05:34:17 mickeyl: or keep these tree up to data Jun 21 05:34:18 we can work until then in parallel Jun 21 05:34:20 right Jun 21 05:34:24 bk export -t patch :) Jun 21 05:34:30 -tpatch even Jun 21 05:34:40 mickeyl: Lets try and do it a few days beforehand, just in case we need the bk program for anything Jun 21 05:34:52 nslu2-linux is going into hibernation for the next 10 days :-) Jun 21 05:35:16 RP: *nod* Jun 21 05:36:08 i don't think the bk binary will cease to work Jun 21 05:36:29 it's more likely that just push, pull, commit will stop to work Jun 21 05:36:33 mickeyl: my understanding is that your lease to run the bk binary will expire on 1 July and not be renewed Jun 21 05:37:08 mickeyl: I think it will be a gradual failure but I'd hate to find that wasn't the case... Jun 21 05:37:11 ah. so they will sue me when i use bk export -tpatch in my cellar on 2nd july :) Jun 21 05:37:36 RP: sure. better safe than sorry Jun 21 05:37:42 linux-oz-2.6: upgrade to 2.6.12-mm1 Jun 21 05:37:45 mickeyl: no, each time you run the bk binary it is actually using a lease which is frequently renewed by contacting bitkeeper Jun 21 05:38:03 if you've ever run bk when away from the network for a time, the lease expires and you can't run it Jun 21 05:38:08 until you connect again Jun 21 05:38:13 urhgs. Jun 21 05:38:17 that's even worse than i thought Jun 21 05:38:25 ~lart Larry and his paranoia Jun 21 05:38:25 * ibot beats Larry and his paranoia senseless with a 50lb Unix manual Jun 21 05:41:03 man have you guys not transitioned yet?! Jun 21 05:41:12 RP: thinking about that key issue... could we have the same problem on SL5500 ? Jun 21 05:41:28 suspend/resume has been pretty unreliable there since 3.5.3 Jun 21 05:42:37 iirc we changed the compiler from 3.5.2 to 3.5.3. perhaps the new compiler has better optimization hence apm userland has different timing Jun 21 05:43:19 http://ewi546.ewi.utwente.nl/tmp/viewmtn/revision.psp?id=dd842877375f56e48b8de130e6a0fc9c613c3917 Jun 21 05:43:29 mickeyl, rwhitby: can you try to sync? Jun 21 05:45:02 hmm Jun 21 05:45:20 collection - org.openembedded Jun 21 05:45:27 righto Jun 21 05:46:17 seems to work Jun 21 05:46:26 koen: which machine? Jun 21 05:46:39 rwhitby: dominion Jun 21 05:50:15 koen: failed from home, worked from mtn.nslu2-linux.org Jun 21 05:50:18 mickeyl: That's quite possible Jun 21 05:50:43 mickeyl: In this case, it was tempramental before. With the copypage speedup, it just broke completely... Jun 21 05:50:58 A compiler change might do the same thing Jun 21 05:51:20 koen: did you add the rwhitby key to your db? Jun 21 05:51:49 (the new one I sent) Jun 21 05:52:00 I think so Jun 21 05:52:04 you need to run dropkey first, then read the new one Jun 21 05:52:15 did all that Jun 21 05:52:20 I was wrong before - the warning means that the new key is not added Jun 21 05:52:45 b5e9ff9b4d151ac20ee094620afa045e86a375d0 rwhitby@nslu2-linux.org Jun 21 05:52:52 is that a good fingerprint? Jun 21 05:52:53 does the fingerprint match? Jun 21 05:52:58 mickeyl: when will we switch? Jun 21 05:53:04 yes, that's correct Jun 21 05:53:32 mickeyl: will we import the old cvsroot into a monotone db? Jun 21 05:53:44 koen: does the server say anything when I try to sync ? Jun 21 05:54:34 * rwhitby suspects koen is restarting the server now ... Jun 21 05:55:23 hmm - looks like you actually have to make a change to be able to create a new branch ... Jun 21 05:58:08 zecke: i think we'll switch on 1st of july Jun 21 05:58:25 zecke: we just have started setting up trees and will work on parallel to gain experience with MT Jun 21 05:58:48 zecke: as for the old history, i think kergoth (as the official tree admin) has to pester Bitmover to convert it into a cvs tree Jun 21 05:59:01 mickeyl: anyone can do that Jun 21 05:59:07 takepatch: 1403 new revisions, 1 conflicts in 227 files Jun 21 05:59:09 wow... Jun 21 05:59:15 http://kernel.bkbits.net/cgi-bin/cvsexport.cgi Jun 21 05:59:48 it took about half an hour when I requested the nslu2-linux repo Jun 21 06:00:43 ok Jun 21 06:00:46 i'm requesting it now Jun 21 06:01:10 RP: I just finished 2.6.12rc6-mm build... Jun 21 06:01:27 done Jun 21 06:03:27 rwhitby: what kind of URL do they expect? I've submitted the http URL, not the bk URL Jun 21 06:03:48 I gave them bk://nslu2-linux.bkbits.net/openembedded Jun 21 06:04:17 I expect the http would probably work too, as they only have to read from it Jun 21 06:04:27 right. lets see Jun 21 06:05:05 ok, restarted the server Jun 21 06:05:42 monotone: read from fd 4 (peer dominion.kabel.utwente.nl) failed, disconnecting Jun 21 06:05:42 monotone: [bytes in: 209] [bytes out: 216] [revs in: 0] [revs out: 0] Jun 21 06:06:12 works from here Jun 21 06:06:17 with collection org.openembedded Jun 21 06:06:55 works from the nslu2-linux machine too, and pull from dominion works at home, just not sync from home Jun 21 06:07:06 heh.. my poor network... again linux-*.tar.gz... Jun 21 06:07:09 ;) Jun 21 06:07:22 heh Jun 21 06:08:02 rwhitby: readded your new key Jun 21 06:12:39 works now, thx Jun 21 06:17:46 monotone list missing and monotone list unknown are my new friends Jun 21 06:18:28 have you guys considered having multiple "modules" in the one database? we're considering having org.nslu2-linux.openembedded.* branches and org.nslu2-linux.something-completely-different.* branches in the same database Jun 21 06:18:53 with completely disjoint and colliding path names Jun 21 06:19:23 (since the server can only serve one database per port) Jun 21 06:21:17 sounds interesting. didn't know that this is possible Jun 21 06:21:31 just about to start testing it Jun 21 06:22:00 we had an nslu2-package-symlinks repo which we used instead of BBMASK, so I'm going to commit that to the org.nslu2-linux.oe-symlinks branch Jun 21 06:24:01 hrw|work: At least we only do it once every few months now ;-) Jun 21 06:25:12 RP: yep Jun 21 06:25:19 mickeyl: jamey is ok with hosting a monotone repo, but we'll have to work out some details Jun 21 06:25:27 koen: never used "bk sfiles -i"? Jun 21 06:25:30 koen: good to hear. Jun 21 06:25:31 since you can have only one db per port Jun 21 06:25:46 hrw|work: what does that do? Jun 21 06:25:46 another tip for young players: monotone add follows symlinks ... Jun 21 06:25:49 koen: like rwhitby says, multiple modules could be a possibility Jun 21 06:26:41 koen: show 'interesing' files Jun 21 06:28:37 ok. kernel and hostap rebuilt Jun 21 06:30:41 hmm - major issue (maybe not for openembedded, but for others): you cannot monotone add a symlink to a directory without it following the symlink and adding everything in the target directory Jun 21 06:31:39 rwhitby: you might want to report that to the mt list. monotone is still in flux, so i think there's a good chance now to request things Jun 21 06:33:15 does monotone do 'bk changers -R' in some form? Jun 21 06:33:57 who wants to be a hero and write a mail to oe@ about monotone and the test repos? Jun 21 06:35:04 * rwhitby should probably subscribe to oe@... Jun 21 06:35:05 If someone writes some intructions, I'll test them ;-) Jun 21 06:35:30 monotone: beginning service on ewi546.ewi.utwente.nl : 5253 Jun 21 06:35:37 and another server joins the club Jun 21 06:36:36 one which does have a reasonable uplink Jun 21 06:38:13 RP: http://dominion.kabel.utwente.nl/koen/blog/pyblosxom.cgi/Handhelds/monotone.html Jun 21 06:47:19 Which version of monotone are you all using? Jun 21 06:47:49 0.19 Jun 21 06:47:57 the one in debian :) Jun 21 06:48:07 monotone 0.19 (base revision: ac4b3dd3c60eb95aa18e01d0f525a5dcfa85567c) Jun 21 06:48:17 ubuntu breezy Jun 21 06:48:38 RP: 0.19, everything older is *yawn* pretty slow Jun 21 06:48:51 * koen hints some python heroes to a monotone fetcher for OE Jun 21 06:49:13 * rwhitby wonders if monotone would run on the nslu2 .... Jun 21 06:49:23 (cross-compiled from OE) Jun 21 06:49:48 * mickeyl hides Jun 21 06:51:42 * RP compiles Jun 21 06:54:04 * hrw|work pulling Jun 21 06:54:55 Good afternoon Jun 21 06:55:25 hey SirFred Jun 21 06:55:29 hi SirFred Jun 21 06:57:02 mickeyl: ping Jun 21 06:57:17 hi SirFred Jun 21 06:57:49 Hi. Jun 21 06:58:22 mickeyl: Well, I like to listen your opinion about how to handle rotation regarding tslib usage from qte. Jun 21 07:00:38 mickeyl: After a fast look, coordinates returned from the touchscreen are transformed using the same functions are with screen coordinates. So, it's assumed that the touchpad and the screen have the same orientation. This is not going to be true, as far as we are rotating the framebuffer instead of transforming the screen elements to change the orientation. Jun 21 07:03:20 koen: can that "--db=hrw.db" be omitted in some way? Jun 21 07:03:38 hrw|work: alias mt='monotone --db=~/work/mickey.db' Jun 21 07:03:41 Humm, my english is getting worse day by day. Jun 21 07:03:42 from my bashr Jun 21 07:03:43 c Jun 21 07:03:48 SirFred: pong Jun 21 07:04:04 hrw|work: when you run monotone inside a working dir, it remembers the db Jun 21 07:04:10 mickeyl: Hi. Jun 21 07:04:20 rwhitby: uf.. Jun 21 07:04:37 SirFred: Hum you need to get the point of the screen where the user taped. In which case do you need the tslib coords != screen coords ? Jun 21 07:04:50 SirFred: I'm afraid I can't be of any help regarding that. Your best bet would be the resident tslib experts zecke, pb_, kergoth Jun 21 07:05:03 proti: When the framebuffer gets rotated. Jun 21 07:05:27 proti: Instead of using a transformed approach, I would like to rotate the framebuffer. Jun 21 07:05:35 mickeyl: OK, thanks. Jun 21 07:06:03 mickeyl: Well, I see it more a problem of how to implement it in qte, than tslib itself. Jun 21 07:06:20 proti: On the Zaurus we can hardware rotate the screen to run in 480x640 instead of 640x480. SirFred's problem is the touchscreen coords also then need to be rotated to make sense Jun 21 07:06:35 RP: thanks. Jun 21 07:07:10 proti: The problem is that from qte, the usual way to do this is transform all the coordinates,pixmaps, images instead of rotate the framebuffer. Jun 21 07:07:25 proti: In that way, the screen and touchscreen orientation are allways synchronized. Jun 21 07:07:59 SirFred: You want it the other way ? Don"t transform framebuffer coords, but the tslib ones ? Jun 21 07:08:29 proti: Well, transform them or just instruct tslib to rotate as we make with the framebuffer. Jun 21 07:09:04 proti: I suppose than some transformation is involved to go from tslib primitive events to coordinates. Jun 21 07:09:32 There is, else you wouldn't need any calibration. Jun 21 07:09:51 proti: I've seen that there is a file with the calibration information. Jun 21 07:10:07 /etc/pointercal Jun 21 07:10:20 koen: Yes. Jun 21 07:10:28 I suspect you'll probably have to extend tslib. I'll bet it could easily transform the data with a rotation module Jun 21 07:10:46 night 'toners Jun 21 07:10:54 'night rwhitby-asleep Jun 21 07:10:54 Your transformation function lies somewhere around. Jun 21 07:11:03 RP: That's my fear. I wouldn't like to extend any API if not neccesary. Jun 21 07:11:23 the problem is you don't tell tslib 'hey, I rotated everything' Jun 21 07:11:35 SirFred: You will have to add something to signal tslib that you want rotated coords Jun 21 07:11:38 koen: Is there a way to do so? Jun 21 07:11:43 not that I know off Jun 21 07:11:58 Well,this evening I will look further into tslib. Jun 21 07:12:24 you could make 4 different pointercals and re-init tslib Jun 21 07:12:44 * RP crys at the idea Jun 21 07:13:05 koen: what do you mean, to have different calibration info for every orientation. Jun 21 07:13:10 yes Jun 21 07:13:19 koen: I think that's not the way. Jun 21 07:13:20 SirFred: Just ignore him ;-) Jun 21 07:13:25 koen: One calibration have to be enough. Jun 21 07:13:44 koen: You have to be able to calculate the rotation, but you don't need more calibration info. Jun 21 07:13:59 The translations are easy enough - you just need to create a module and find a way to signal it Jun 21 07:14:13 koen: To say you the truth, i was considering that solution yesterday night. But I feared to be burned here. Jun 21 07:14:18 koen: ;-) Jun 21 07:14:28 RP: And the worse thing. Jun 21 07:14:56 RP: The encapsulating classes for tslib, qwsmousesomething need to provide a method to make that. Jun 21 07:15:19 SirFred: Patching qte isn't a new thing for oe ;-) Jun 21 07:15:28 RP: But that change is going to break compatibility. Jun 21 07:15:44 SirFred: How so? Surely its just an extra function? Jun 21 07:15:45 RP: I don't know how is that managed. Jun 21 07:15:52 SirFred: that's no problem Jun 21 07:16:09 RP: Isn't an extra method in a class going to break the compatibility? Jun 21 07:16:28 SirFred: Not in my book ;-) Jun 21 07:16:36 SirFred: You should see what zecke did to the keyboards ;-) Jun 21 07:17:02 RP: Well, at least adding members to the classs changes it's size. Jun 21 07:17:09 RP: Perhaps it's not the same with methods. Jun 21 07:17:17 cya later Jun 21 07:17:22 cya zecke Jun 21 07:17:23 RP: You're making me doubt. Jun 21 07:17:27 * zecke visits the dentist to get stoned Jun 21 07:17:28 bye zecke Jun 21 07:17:31 heh Jun 21 07:18:06 SirFred: mickeyl knows more about qte so he's the one to ask but I don't see any problem... Jun 21 07:18:42 SirFred: don't worry about breaking compatibility in qte. Our users will be more than willing to pay that price for W100 accelleration. Jun 21 07:18:53 95% of OZ users always reflash when a new distro comes out Jun 21 07:19:00 5% upgrade Jun 21 07:19:01 then fail Jun 21 07:19:03 then reflash Jun 21 07:19:03 heh Jun 21 07:19:04 mickeyl: :) Jun 21 07:19:05 :) Jun 21 07:19:18 mickeyl: Anyway, about the technical item itself, I'm interested. Jun 21 07:20:14 mickeyl: Shouldn't changing a class (adding a member, for example) breaks the compatibility? Jun 21 07:20:44 For example, if my program is using that class, it's allocating a given amount of memory. Jun 21 07:20:48 qte isn't the only victim of this problem Jun 21 07:20:58 SirFred: In C++ you can't add or remove a class Jun 21 07:21:02 so extending tslib would be an option Jun 21 07:21:14 SirFred: that's why most class libraries use a virtual method hack Jun 21 07:21:23 mickeyl: ok Jun 21 07:21:24 SirFred: an additional indirection to add methods Jun 21 07:21:30 SirFred: but like i said, don't worry :) Jun 21 07:21:39 mickeyl: That's what I wanted to hear. Jun 21 07:21:42 s/class/method/ Jun 21 07:21:50 mickeyl: Yes I supposed that. Jun 21 07:22:18 koen: Well, extending tslib probably will be necessary, but also changing apis in qte. Jun 21 07:22:30 *nod* Jun 21 07:22:35 koen: At least the class that encapsulates tslib usage. Jun 21 07:22:48 hehhe.... Jun 21 07:22:58 mickeyl: For reference, we're on track to see an input switch event being added to the kernel to solve our keyboard hinge switch problem Jun 21 07:23:25 RP: To rotate the cursor keys ? Jun 21 07:23:56 Well, I'm hungry. Lunch time. Jun 21 07:23:58 SirFred: We did have it all working until the input people broke things... Jun 21 07:24:06 RP: :) Jun 21 07:24:09 koen: i don't think i like what jamey has written in that mail. Jun 21 07:24:19 bbl Jun 21 07:24:21 mickeyl: what do you not like? Jun 21 07:24:26 ah here you are Jun 21 07:25:05 it sounds like you just want to incorporate in all the work we did to support Zauruses, SIMpads et. all. into the familiar distribution Jun 21 07:25:34 mickeyl: email on familiar* ML? Jun 21 07:25:47 who is going to maintain this familiar distribution if we agree on a new merged thing? Jun 21 07:26:09 hrw|work: familiar@hh.org Jun 21 07:26:20 ah.. Jun 21 07:26:43 mickeyl: we are hoping to find developers such as yourself you would like to incorporate support for zaurus, simpad, etc into Familiar Jun 21 07:26:59 s/you would/who would/ Jun 21 07:27:32 if we agree on a new merged distribution, there would be a merged set of developers Jun 21 07:27:39 jamey: so you are essentially against giving up your name? Jun 21 07:27:46 who would all get credit for their contributions Jun 21 07:28:05 we do want to keep the name and we want to refocus on stability Jun 21 07:28:09 I don't know if you watched the OE community lately, but we already have merged the developers and 99% of the distributions Jun 21 07:29:13 hmm ok, then we probably should pick a new name for the merged OE-based distribution and then see familiar as a stable-tree for that Jun 21 07:29:39 I'm going to go out on a limb and speak for the other old timers without them being here Jun 21 07:30:07 with all due respect to the old timers, the ones who did most of the work in the last 3 years already are in favour of a merger Jun 21 07:30:31 as I said, we are happy to continue hosting the distribution if you want to create a new one Jun 21 07:30:44 and maybe that will result in Familiar withering away in favor of the new one Jun 21 07:31:05 or maybe it will result in two distributions, one more like gentoo and one more like a traditional linux distro, but for handhelds Jun 21 07:32:22 I see open embedded as a build system and a meta distribution Jun 21 07:32:22 from which distributions can be produced Jun 21 07:32:22 depending on package versions and configurations Jun 21 07:32:35 so I think it's confusing to merge OE and Familiar Jun 21 07:32:38 hmm it's a bit of a disappointment to me, because in the end it will sound like you are not commited to the simpad, zaurus, ipaq people joining forces among the hat of a new name. Jun 21 07:32:45 huh? Jun 21 07:32:53 no one spoke of merging OE. OE is not a distribution Jun 21 07:33:03 we agreed to merge OpenZaurus, Familiar Linux, OpenSIMpad Jun 21 07:33:32 that means removing the subtle little differences in distro policies Jun 21 07:33:47 after all, those three are 99% the same already Jun 21 07:33:54 perhaps we should talk about what the differences are in the distro policies Jun 21 07:34:08 the major one being the soft float vs. hard float thing Jun 21 07:34:33 ah, that is a config thing and not a policy thing Jun 21 07:35:01 we always regarded this as a distro policy, but ok Jun 21 07:35:03 cal it config Jun 21 07:35:05 call, even Jun 21 07:35:15 it does need to be discussed if the distros are to merge Jun 21 07:35:25 another related issue is kernel versions and config Jun 21 07:35:42 I think a distro supporting all these platforms will have to go through a transition period with multiple kernel versions Jun 21 07:35:57 because the newest devices are 2.6 only and the oldest ones are 2.4 only Jun 21 07:36:03 hmm it's really funny that you now start to revive this discussion when already did that weeks ago. why didn't you come up with these fundamental concerns back then? Jun 21 07:36:21 where did you have these discussions? Jun 21 07:36:38 familiar@hh.org, openzaurus-devel@hh.org, simpad-general@lists.sf.net Jun 21 07:36:49 on all three major lists Jun 21 07:37:02 besides, it was mentioned in the OOO newsletter and on the oesf forums Jun 21 07:37:04 I have been behind on my mailing list reading Jun 21 07:38:15 but I should have been paying more attention Jun 21 07:39:16 jamey: For reference, openzaurus has been shipping with both 2.4 and 2.6 kernels for some devices. The next release will be 2.6 only as the port is now complete and stable (ish :) Jun 21 07:39:31 very good Jun 21 07:39:40 RP: 2.6 on pxa, 2.4 on collie Jun 21 07:40:38 RP: or we will never release next OZ without collie or never Jun 21 07:40:48 hrw|work: I was meaning the c7x0 being 2.6 only - yes some devices are stuck with 2.4 but c7x0 will be 2.6 only from now on... Jun 21 07:42:11 as it stands, koen and me want to continue with our merger for the three major OE based distros since it makes no sense to duplicate all the release work et. al. If you want to introduce a policy change in how the Familiar Linux distribution is developed in the future to focus on stability, that sounds good and won't infere with our merge plans. Jun 21 07:42:38 mickeyl: ok, what does that mean? Jun 21 07:43:25 jamey: I would be careful, this sounds like another 'one ring to rule them all.." scam. I heard they're even thinking of calling it MyPreci-OS... Jun 21 07:44:03 hmm.. OE-WRT is needed Jun 21 07:44:10 root@OpenWrt:~# ldconfig Jun 21 07:44:10 -ash: ldconfig: not found Jun 21 07:44:12 argh Jun 21 07:44:26 Sorry to chime in with something off the current topic: I understand that with 2.6 on the c7x0 under Opie the hinge sensor & open/close switch return keypress events. I can't detect any using xev under GPE, do I need additional kernel patches which GPE doesn't use by default? Jun 21 07:44:49 lardman|busy: you need 2.6.11 Jun 21 07:45:05 jamey: that means that I see the us releasing this new distro with EABI and support for all the hardware that has been currently supported in openzaurus, familiar, and opensimpad. I don't know exactly what your future plans are for familiar then, but as long as it continues to be OE-based all work on it will also benefit the common distribution. Jun 21 07:45:09 hrw|work: will 2.6.12 not work? Jun 21 07:45:31 lardman|busy: yes Jun 21 07:45:43 mickeyl: ok Jun 21 07:45:57 lardman|busy: The input people don't want switches being represented as keys. The 2.6.12 kernels therefore add a new switch event type Jun 21 07:46:09 * mickeyl wonders about koen being so silent Jun 21 07:46:20 RP: ah, how would I go about detecting it? Jun 21 07:47:31 lardman|busy: Look for EV_SW events from the keyboard Jun 21 07:47:51 RP: okay, thanks Jun 21 07:50:20 mickeyl: just got back from the kitchen Jun 21 07:50:27 * koen scrolls up Jun 21 07:51:13 mickeyl: does opie-pcmciaapplet works now? (20050621 cvsdate) Jun 21 07:51:29 hrw|work: it should do something. not all functionality is in place yet Jun 21 07:51:55 hrw|work: configuring works, detecting card works, the actions don't work yet Jun 21 07:52:04 mickeyl: ok. Jun 21 07:52:08 hrw|work: i'll complete that during the weekend Jun 21 07:52:29 mickeyl: all I need is eject/ident now Jun 21 07:53:49 hrw|work: that should work. i'd appreciate some tests Jun 21 07:54:44 mickeyl: ok. - lets go back to discussion about merging distros with koen Jun 21 07:55:01 hrw|work: i just commited a fix for the applet pixmap to not use the cardmon's pixmap Jun 21 07:55:55 mickeyl: that pixmap problem whole opie has Jun 21 07:56:16 i don't think there's much more of a discussion necessary. it looks like the veterans want to pick up Familiar with a different style of working to produce more stable releases. which sounds pretty good to me since for sure all OE based distros benefit from it. Jun 21 07:56:54 hrw|work: *nod* something to fix when the theming interface debuts Jun 21 07:57:03 hrw|work: when we have to go over all pixmaps anyway Jun 21 07:57:35 * mickeyl wanders home, it's too frickin' hot in this office Jun 21 07:58:23 * koen makes some dinner, will be back in ~30 minutes Jun 21 08:01:03 * jamey back later Jun 21 08:03:13 hi france Jun 21 08:03:42 hi mithro Jun 21 08:19:35 * koen looks at irssi Jun 21 08:19:43 bbl Jun 21 08:19:57 hi all Jun 21 08:31:04 02hi out there. Jun 21 08:31:04 anybody with gentoo hosted oe experience? Jun 21 08:39:07 02hm, seems to be a chill channel, ok, bye... Jun 21 08:39:37 uf.. he parts.. Jun 21 08:39:47 heh Jun 21 08:39:48 now I understand why gentoo irc users are awful... Jun 21 08:40:09 since I couldn't read two-third of the stuff he wrote, that doesn't seem like much of a loss Jun 21 08:46:46 cu Jun 21 09:08:40 gentoo users must have short attention spans .... Jun 21 09:09:21 of course he had to do it in color to be noticed Jun 21 09:09:37 yeah - which is really astounding given that they have to wait that often ;) Jun 21 09:36:21 sup people? Jun 21 09:40:03 * RP pulls from koen :) Jun 21 09:40:56 monotone? Jun 21 09:41:10 koen: yes Jun 21 09:42:21 send me you key if you want to poke at it Jun 21 09:48:46 anyone want turl.ws? Jun 21 09:50:33 koen: sent, thanks :) Jun 21 09:51:16 evening Jun 21 09:54:21 RP: added to the ewi machine Jun 21 09:58:46 koen: thanks - will have a "play" later :) Jun 21 09:58:57 * RP gets some food Jun 21 09:59:17 RP: if the ewi box goes down, you can also use dominion.kabel.utwente.nl Jun 21 10:24:01 RP: eep/ping Jun 21 10:30:10 ~lart kde Jun 21 10:30:11 * ibot follow's kde with a gauntlet and ... scratch ... HUMILIATION Jun 21 10:31:52 hey zecke_ Jun 21 10:33:20 hehe I think usability and stability was greatly improved Jun 21 10:34:41 oh that is not #handhelds.org :} Jun 21 10:35:34 hi zecke Jun 21 10:38:17 zecke_: I just had too laugh real loud at openuseability.org this morning Jun 21 10:39:03 koen: what did they do? Jun 21 10:40:30 it seems kde discoverd useability after years of ridiculizing gnome for it Jun 21 10:40:50 Humm, I'm starting to have the feeling that qte is really hacked. Jun 21 10:40:56 KDE is usable :) Jun 21 10:41:20 it still has nouns instead of verbs on the dialogs Jun 21 10:41:22 I hate that Jun 21 10:41:37 ? Jun 21 10:41:48 koen: hehe, I meant familiar Jun 21 10:41:56 koen: 0.5, 0.6.x were just horrible broken Jun 21 10:42:15 0.7.2 was pretty good *after* fixing ipkg Jun 21 10:42:28 um... KDE dialogs have both nouns *and* verbs :) Jun 21 10:42:56 koen: my kde wallet decided not to accept my password anymore Jun 21 10:43:08 I think this due KDE (debin) -> KDE (ubuntu) switch... Jun 21 10:43:10 zecke: that's a bug, not usability =p Jun 21 10:43:11 re Jun 21 10:43:12 is that some sort of keyring manager? Jun 21 10:43:16 hey hrw Jun 21 10:43:18 ~lart 2.6.12-mm1 Jun 21 10:43:18 * ibot nukes 2.6.12-mm1 with a single large nuke Jun 21 10:43:24 | drivers/built-in.o(.text+0x666c): In function `elpp_render_char': Jun 21 10:43:24 | pxamci.c: undefined reference to `fb_sysmove_buf_aligned' Jun 21 10:43:35 mickeyl: I have it on husky Jun 21 10:43:40 why do they need to touch the fb API every time Jun 21 10:43:42 *sigh* Jun 21 10:43:50 mickeyl: because they can Jun 21 10:43:52 mickeyl: he? Jun 21 10:44:13 * luke-jr_ got FATAL: This kernel is too big for the c7x0 and will destroy your machine if you flash it!!! Jun 21 10:44:24 koen: it is wallet holding your passwords Jun 21 10:44:47 that's what I meant Jun 21 10:45:12 zecke: I suggest doing monthly wallet backups or such Jun 21 10:45:18 zecke: same w/ address book Jun 21 10:45:21 "keyring manager" is the gnome term and "keychain" the osx term Jun 21 10:45:40 koen: keyring is for public/private keys in KDE Jun 21 10:45:45 luke-jr_: don't let me get started on kdepim... Jun 21 10:45:49 mickeyl: elpp? Jun 21 10:46:11 hrw: the logo / init progress patch Jun 21 10:46:21 gets broken with every tiny subrelease of the kernel Jun 21 10:46:26 well, that keeps me busy Jun 21 10:46:27 heh Jun 21 10:46:50 elpp doesn't qualify for upstream submission? Jun 21 10:46:59 no way Jun 21 10:47:00 http://lists.kde.org/?l=kde-devel&m=111930225607761&w=2 Jun 21 10:47:06 luke-jr_: linus will refuse it Jun 21 10:47:11 more logo stuff won't be accepted Jun 21 10:47:21 nonsense Jun 21 10:47:34 painting directly to the fb isn't fluff or bloat, ask ljp Jun 21 10:47:36 zecke: which thread is that? Jun 21 10:47:41 koen: heh Jun 21 10:47:56 luke-jr_: ARM devel Jun 21 10:48:19 oooh, jenoptik Jun 21 10:48:34 that are my heros when it comes to binoculars Jun 21 10:49:24 koen: not even the GDR/DDR was capable of ruining it Jun 21 10:49:51 but kde-devel is a pretty funny place for arm kernel dev Jun 21 10:50:25 koen: specially we've so many people actually have seen a PDA Jun 21 10:50:47 koen: but yes KDAB Employees (lutz) are highly qualified embedded linux engineers as well... Jun 21 10:50:59 koen: he enlightened me with on-site teaching at ROAD... Jun 21 10:51:06 no doubt Jun 21 10:51:17 Hmm. What can be removed from OPIE_DEPENDS_24? Because the rootfs is to big for tosa. Jun 21 10:51:29 do13: how's that ? Jun 21 10:51:37 do13: how much flash do you have on tosa? Jun 21 10:51:44 hmm Jun 21 10:51:47 the rootfs has 28MB Jun 21 10:51:54 my kde-devel ML folder doesn't have an ARM dev thread :| Jun 21 10:51:58 do13: how big is it atm.? Jun 21 10:51:59 that's 1mb more as the c700 Jun 21 10:52:02 <[g2]> anyone familiar with zreladdr, xscale and 2.6 ? Jun 21 10:52:11 unless it's the job offer post thing Jun 21 10:52:20 luke-jr_: it is the job offer Jun 21 10:52:27 *hot* Jun 21 10:52:28 do13: I'd be very surprised if it was suddenly larger than 24MB. Jun 21 10:52:34 do13: that sounds like a bug then Jun 21 10:52:51 do13: I'll start a build later tonite and check Jun 21 10:52:51 ah Jun 21 10:52:52 30228480 2005-06-21 19:25 opie-image-tosa-20050621171734.rootfs.jffs2 Jun 21 10:52:54 * koen remembers a corgi 2.4 bug :\ Jun 21 10:53:19 ? Jun 21 10:53:29 the mic one? Jun 21 10:54:07 no, the sudden increase of space in my root partition Jun 21 10:54:15 mickeyl: the 2.6 version needs more space? Jun 21 10:54:39 the kernel decided that I had a 57 mb /root instead on 27mb Jun 21 10:54:59 do13: that surely can't account for more than 6 MB Jun 21 10:55:06 do13: or did you install all modules available ? Jun 21 10:55:20 koen: isnt that just parameters? Jun 21 10:55:41 could be Jun 21 10:56:10 my root is 53 MB... Jun 21 10:56:11 mickeyl: Not all:) Jun 21 10:56:53 do13: change ROOT_FLASH_SIZE to 23 in your tree Jun 21 10:56:53 i Jun 21 10:56:58 I'll look into it Jun 21 11:01:58 do13: You don't have the system set to not strip binaries or anything like that? Jun 21 11:02:07 luke-jr_: pong Jun 21 11:02:39 luke-jr_: I see you've built a kernel that's too big :) Jun 21 11:02:39 RP: FATAL: This kernel is too big for the c7x0 and will destroy your machine if you flash it!!! Jun 21 11:02:44 RP: /me nods Jun 21 11:02:57 Was this c7x0? Jun 21 11:03:47 It must have been to get that message. Its strange as mine built within the limits :-/ Jun 21 11:06:14 :\ Jun 21 11:06:19 i don't get the message either Jun 21 11:06:21 there's been a revision since then Jun 21 11:06:24 so maybe it was fixed Jun 21 11:06:27 koen: why do people try to build a kernel for arm when they don't know what a compiler is? Jun 21 11:06:38 * luke-jr_ is waiting on the retry Jun 21 11:06:49 zecke: wtf? =p Jun 21 11:07:12 zecke_: gentoo Jun 21 11:07:24 koen: yeah gentoo users... Jun 21 11:07:29 luke-jr_: Which compiler are you using? Jun 21 11:07:34 that was to easy.... Jun 21 11:07:36 I'm using gcc 3.4.3... Jun 21 11:07:45 RP: GCC 3.4.4 Jun 21 11:07:46 koen: we should randomize our linker script when compiled on Gentoo Jun 21 11:08:01 luke-jr_: I was refering to a post on familiar Jun 21 11:08:04 mickeyl: Which gcc are you using? Jun 21 11:08:17 Does 3.4.3 work better w/ OE? Jun 21 11:08:45 RP: All binaries are striped. Jun 21 11:08:57 luke-jr_: Are you using the csl compiler branch or not? Jun 21 11:09:27 It could even be that your problem and Dirks are related - I wonder if we have a compiler bug causing excessive file size? Jun 21 11:09:28 koen: can one easily build a maemo sdk with oe? Jun 21 11:09:45 sdk,no, image, yes Jun 21 11:09:51 RP: not CSL, no Jun 21 11:10:01 RP: CSL was being a problem =p Jun 21 11:10:10 RP: default Jun 21 11:10:24 zecke_: but we're working on it :) Jun 21 11:10:27 using DISTRO=openzaurus-3.5.4 Jun 21 11:11:08 luke-jr_: The kernel has just built fine for me with openzaurus-3.5.4 and the 3.4.3 compiler... Jun 21 11:13:34 luke-jr_: Looking at the kernel size, mine came out at 11348 bytes under the limit so its just the different toolchain Jun 21 11:13:58 Looks like we'll need to prune some modules again :) Jun 21 11:14:27 Maybe some recent drivers are being compiled in by default and shouldn't be... Jun 21 11:14:30 wait until do_compile on the latest rev finishes =p Jun 21 11:14:51 any idea if GCC 4.0 would work? Jun 21 11:15:02 ok, latest is still too big Jun 21 11:16:10 luke-jr_: No idea. I'd be interested in someone trying it but I've enough problems without that :) Jun 21 11:16:33 luke-jr_: There looks to be some things defaulting to "Y" in the defconfig that shouldn't be there - I'll try and trim it back Jun 21 11:18:55 Bloat Jun 21 11:19:00 Fluff Jun 21 11:19:13 Hmm. I removed opie-reader and the games and now the image size is 20Mb Jun 21 11:19:29 that's the best part! Jun 21 11:19:33 do13: pretty odd Jun 21 11:19:39 lol Jun 21 11:27:44 mickeyl: does not latest opie-reader needs flite? Jun 21 11:28:49 there was some buzilla activity about it Jun 21 11:29:58 hrw: uh oh, that could be. i wasn't involved in tweaking that Jun 21 11:30:22 mickeyl: Indeed. opie-reader needs 8MB. Jun 21 11:30:27 tsss Jun 21 11:30:45 * koen is not going to say fluff Jun 21 11:30:50 let me check Jun 21 11:31:27 yeah, of course :/ Jun 21 11:31:40 With and without opie-reader: Jun 21 11:31:41 20955136 2005-06-21 20:24 opie-image-tosa-20050621182106.rootfs.jffs2 Jun 21 11:31:41 30228480 2005-06-21 20:29 opie-image-tosa-20050621182709.rootfs.jffs2 Jun 21 11:31:59 someone needs to package the codecs individually Jun 21 11:32:11 in the meantime i remove it Jun 21 11:33:44 RP: 2.6.12-mm1 is very fast Jun 21 11:35:28 does that get built by default? Jun 21 11:36:09 beware one thing... -rc6 are newer then -mm1 Jun 21 11:37:31 koen: default on tosa is 2.4 Jun 21 11:37:32 mickeyl: Finally, I didn't need to change tslib neither qwsmousehandler. Jun 21 11:37:42 do13: and for c7x0? Jun 21 11:37:53 2.6 Jun 21 11:37:59 SirFred: woot! how that? Jun 21 11:38:08 mickeyl: I've made all the transformations in QWSServer::setMouse Jun 21 11:38:24 remove opie-reader from OPIE_DEPENDS_24 until the codecs get packaged more fine granular Jun 21 11:38:31 SirFred: excellent! Jun 21 11:38:31 * koen fires up a build Jun 21 11:38:33 mickeyl: Anyway, I had to add some methods to QScreen. Jun 21 11:38:45 mickeyl: Well, I'm not really proud of what i've perpetred. Jun 21 11:38:56 how is monotone like? Jun 21 11:39:03 mickeyl: The bad thing is that QScreen has isTransformed() and isRotated(). Jun 21 11:39:42 mickeyl: Just to tell the difference between a transformation (screen with software transformation) and rotation( a hardware one) Jun 21 11:39:49 mickeyl: It works, at least. Jun 21 11:40:48 SirFred: that's really good news. well done Jun 21 11:41:03 mickeyl: Now I have to implement hot-rotating. Jun 21 11:41:03 * mickeyl sees OZ 3.5.4 with kernel 2.6 + W100 accell. Jun 21 11:41:38 mickeyl: Just now, the rotation only can be specified using QWS_DISPLAY environment var. Jun 21 11:42:00 http://ewi546.ewi.utwente.nl/tmp/viewmtn/revision.psp?id=8819f290c560829da531349dc5b5b453608ac7fc Jun 21 11:42:03 SirFred: tried X11 on w100? it has bloody fast rotate Jun 21 11:42:05 mickeyl: I think that the way to make that rotation is another hack. I've seen some qt_transformed_screen global pointer in some place. Jun 21 11:42:17 SirFred: ok, so you need to implement whatever happens on setRotation() Jun 21 11:42:32 hrw: No. You mean the accelerated w100 x11 driver? Jun 21 11:42:41 hrw: The one from pdaxrom ? Jun 21 11:42:49 SirFred: no - fbdev one from kdrive Jun 21 11:43:14 hrw: I didn't saw it. Jun 21 11:43:31 hrw: A question. You was testing the w100 qte driver, wasn't you? Jun 21 11:43:38 SirFred: and I have it now Jun 21 11:43:44 hrw: The last version? Jun 21 11:43:52 hrw: Well, the one in bitkeeper. Jun 21 11:43:59 SirFred: yes - suspend/resume is fscked Jun 21 11:44:13 hrw: Yes, I know. It's the next scheduled job. Jun 21 11:44:19 SirFred: http://hrw.one.pl/tmp/PICT1851.JPG Jun 21 11:44:44 hrw: Did you tried the sysinfo gfx bench? Jun 21 11:45:04 SirFred: yes - its better now - synchronized Jun 21 11:45:06 hrw: I see, I have the same behaviour. Jun 21 11:45:17 hrw: And better numbers? What numbers are you getting? Jun 21 11:45:42 root@c7x0:~# /etc/init.d/opie wstart Jun 21 11:45:53 moment.. Jun 21 11:46:04 bbl, need to watch confederation cup Jun 21 11:46:22 mickey|ccup: nice hint Jun 21 11:46:24 hrw: Because the numbers you say the other day were too high compared with mine. Jun 21 11:46:51 SirFred: hmm.. now I have it rotated (top on left side) Jun 21 11:47:06 hrw: Strange. Jun 21 11:47:08 hrw: Transluency Jun 21 11:47:18 hrw: Perhaps your vt2 was rotated. Jun 21 11:47:31 hrw: That version don't touch the rotation of the screen, as far as i know. Jun 21 11:47:36 zecke_: :) Jun 21 11:47:45 SirFred: it was not - I killed old opie after resume and started opie again Jun 21 11:48:33 hrw: Can you kill opie again, and go to the second vt using Fn + arrow ? Jun 21 11:48:38 hrw: It should be rotated. Jun 21 11:49:06 SirFred: will do - first started tests on 480x480 part of screen Jun 21 11:49:29 will install fbgrab Jun 21 11:50:03 2047gops in gfx test, 557 in txt one, 46/25 in ram test Jun 21 11:50:08 hrw: scap Jun 21 11:50:36 hrw: 2047, that's more than I'm getting. Jun 21 11:50:51 I'm getting 1800/1900. Jun 21 11:50:59 And 1200 in software mode. Jun 21 11:51:29 The text mode is a little strange. Jun 21 11:51:39 do13: It is much faster, isn't it :) Jun 21 11:51:47 It gots faster every time you use it. Jun 21 11:52:05 Maybe some cache is involved. Jun 21 11:52:35 qte 2.3.10 has glyph caching Jun 21 11:53:15 koen: So, that is the reason. Jun 21 11:53:26 koen: yes but who uses it Jun 21 11:53:32 koen: After some iterations, the text benchmark is incredible fast. Jun 21 11:54:33 SirFred: >20k even Jun 21 11:55:22 ok. last try Jun 21 11:55:45 RP: what is the magical incantation to use regular gcc for the kernel? Jun 21 11:56:57 koen: KERNEL_CCSUFFIX="-3.4.3" ? Jun 21 11:57:14 luke-jr_: I've pushed some changes to the defconfig which with luck will get you below the magic size Jun 21 11:57:46 http://handhelds.org/scap/port.22986.png Jun 21 11:57:50 scap sux Jun 21 11:58:18 no, you have and old version of scap Jun 21 11:58:47 use the script on top of http://handhelds.org/scap/ Jun 21 11:58:48 hmm.. OE has newer one? Jun 21 11:59:28 http://hrw.one.pl/tmp/shot.png Jun 21 11:59:37 anyway on Z it looks different Jun 21 11:59:46 OE has the most recent one Jun 21 11:59:58 check if fbset is in your path Jun 21 12:00:11 it is Jun 21 12:00:16 hrw: 2047 gops/s. Jun 21 12:00:59 hrw: If will be better when the blitting work with transparent pixmaps. Jun 21 12:01:04 hmm. chvt 1 and it is rotated ;( Jun 21 12:01:06 it was not Jun 21 12:01:25 how to unrotate it? Jun 21 12:01:35 hrw: fbset -xres 640 -yres 480 Jun 21 12:01:40 hrw: Wait Jun 21 12:01:43 hrw: fbset -n -xres 640 -yres 480 Jun 21 12:01:51 koen: god the mameo installer is broken... Jun 21 12:01:55 koen: look: Jun 21 12:01:56 root@c7x0:~# ipkg info scap Jun 21 12:01:56 Package: scap Jun 21 12:01:56 Version: 0.1-r1 Jun 21 12:02:05 koen: the same is in OE... Jun 21 12:02:12 hrw: It sets the current vt. Jun 21 12:02:27 hrw: I don't know if there's a way to set other than the current. I use to do it from a remote ssh session. Jun 21 12:02:52 SirFred: fbset -a set all Jun 21 12:02:59 hrw: Good to know. Jun 21 12:03:05 anyway I have it rotated still Jun 21 12:03:17 hrw: Rotated or flipped? Jun 21 12:03:20 rotated Jun 21 12:03:30 hrw: And fbset don't change it? Jun 21 12:03:35 but I rotated to 480x640 and back to 640x480 so it changed Jun 21 12:03:46 starting opie again Jun 21 12:04:39 SirFred: 525 txt, 1872gops Jun 21 12:05:18 hrw: That's similar to mine. Jun 21 12:06:12 maybe previuos was better because of rotation was wrong so only part of screen was visible Jun 21 12:06:35 cu Jun 21 12:06:40 hrw: Bye Jun 21 12:08:51 hmm... GCC 4.0 deps on some gmp-native that is x86-only... Jun 21 12:15:44 cdbot: part Jun 21 12:16:08 CoreDump|home: any news on the akita? Jun 21 12:17:07 yeah, turns out the neighbour took indeed the parcel. He left me a note in the letter box this morning. He'll be back home in a week...bastard Jun 21 12:17:26 <[g2]> Anyone know what 0xE1A00000 means in armv5b machine code ? Jun 21 12:17:38 CoreDump|home: lol Jun 21 12:17:56 ohoh my notebook fan sounds.... Jun 21 12:25:00 [g2]: gotta buy David Seal's book... E is always, not that that helps much ;-) Jun 21 12:25:34 objdump will run on arbitrary binaries if you tell it that the type is binary and that it is arm big-endian. Jun 21 12:25:40 ~lart scratchbox Jun 21 12:25:40 * ibot blames scratchbox for all the evil in the world Jun 21 12:26:23 anyone know a good reference for automake/autoconf? Jun 21 12:26:38 Which directory does a compile task usually execute in? Jun 21 12:26:53 ${S} Jun 21 12:27:52 Crofton: see the manuals at gnu.org Jun 21 12:27:57 the atilib code is running gcc with a -L. option and then not finding the library... Jun 21 12:28:12 thanks reenoo_ Jun 21 12:28:27 * chouimat is away: out of Coke Jun 21 12:28:45 zecke_: playing with the maemo sdk? Jun 21 12:29:18 zecke_: I was thinking of building sb rootstraps with OE Jun 21 12:30:10 koen: yeah... playing... Jun 21 12:30:18 koen: I'm stopping it and doing my assigments now Jun 21 12:30:46 installing Xephyr lessens the hurt Jun 21 12:31:51 koen: /scratchbox... my '/' is only 80mb... Jun 21 12:32:29 PARALLEL_MAKE strikes again :) Jun 21 12:37:43 zecke_: yeah, that sucks Jun 21 12:41:55 Ciao all Jun 21 12:41:55 buon giorno Pigi Jun 21 12:42:36 koen, improving, every day ! I guess I should start learning dutch .... Jun 21 12:43:21 :) Jun 21 12:46:13 CoreDump|home: cdbot doesn't obey me :| Jun 21 12:46:53 hmm? Jun 21 12:47:12 what did you try to do? Jun 21 12:48:30 told it to join #oe Jun 21 12:48:31 =p Jun 21 12:49:16 hah Jun 21 12:49:58 sharp-binary-only/sharp-aticore-oss_1.0.1.bb <-- what does 'oss' mean? o.O Jun 21 12:50:57 'OMG Stupid Sharp' Jun 21 12:51:05 or something of equal value Jun 21 12:51:42 ibot oss Jun 21 12:51:43 O.o; Jun 21 12:51:43 Provides sound card drivers for most popular sound cards under Linux. URL: http://www.opensound.com/ Jun 21 12:51:43 mmm i'm experimenting kernels for sl-5600 someone want to give one? :) Jun 21 12:51:54 aha... that OSS Jun 21 12:51:54 * chouimat is back. Jun 21 12:52:23 RP: new kernel works Jun 21 12:54:03 koen: I'd guess open source solution. Which is a bit ironic :) Jun 21 12:54:36 I'll stick with my version Jun 21 12:54:51 RP: france/ibot suggested Open Sound System :) Jun 21 12:55:11 I was thinking "binary only + only source? wtf?" :) Jun 21 13:06:51 what toolchain is suggested for building a kernel for 5600? Jun 21 13:07:25 i have found docs suggesting still 2.95, is that still correct (look ancient to me) Jun 21 13:07:27 2.95.3 I guess for 2.4 Jun 21 13:07:38 quality lineo code Jun 21 13:07:46 Yes, they're 2.95.3 only Jun 21 13:08:05 mmm then i have to get that... but that will never compile my app :( Jun 21 13:08:13 How do I get normal x11 instead of diet? Jun 21 13:08:16 ok, let's try Jun 21 13:08:21 tx Jun 21 13:08:33 luke-jr_: set the right preferred_provider Jun 21 13:08:38 koen: tried Jun 21 13:08:48 PREFERRED_PROVIDER_x11 = "x11" Jun 21 13:09:01 darkschneider: Compile the kernel with 2.95.3 - compile everything else with a modern gcc Jun 21 13:09:20 ? on x86 objects are not compatible Jun 21 13:09:36 or i remeber bad? Jun 21 13:09:45 kernel != userspace Jun 21 13:09:45 darkschneider: objects != binaries Jun 21 13:09:55 darkschneider: final C binaries are compatible Jun 21 13:10:14 (note not C++) Jun 21 13:10:17 ok so only kernel with 2.95.3 and rest with mdern Jun 21 13:10:23 yes Jun 21 13:10:24 * luke-jr_ nods Jun 21 13:10:26 ok that makes sense.. sorry Jun 21 13:10:33 koen: so how about PREFERRED_PROVIDER? Jun 21 13:10:50 it works for me Jun 21 13:10:54 i imemgiatly wougth of building a lib with 2.95.3 and then some app using it with 3.xx Jun 21 13:10:57 exactly what I have works? Jun 21 13:11:00 but I have it set to diet-x11 all over the place Jun 21 13:11:11 diet is the default >.> Jun 21 13:11:12 local.conf and $DISTRO.con Jun 21 13:12:34 * luke-jr_ tries PREFERRED_PROVIDER_virtual/x11 Jun 21 13:12:48 (since PREFERRED_PROVIDER_virtual/xserver is needed) Jun 21 13:24:36 ~lart HIGHMEM Jun 21 13:24:36 * ibot takes large quantities of Krispy Kream donuts and stuffs them one after another down HIGHMEM's throat until HIGHMEM puts on 150lbs Jun 21 13:49:43 ibot: oss is also open source software Jun 21 13:49:44 okay, mickeyl Jun 21 14:03:24 koen: how can I mess with the monotone tree? Jun 21 14:03:48 send me your monotone pubkey Jun 21 14:03:56 mickeyl: thre are any knwon stability problems with the kernel in 3.5.3 dir? because i used both it and the one CoreDump|home provvided and they both give the same crashy behavour, but i think they have the cache thing enabed (no bug) and notting other special Jun 21 14:04:56 darkschneider: no known instabilities except when running a pxa255 kernel on a pxa250 device. Jun 21 14:05:03 mickeyl: hey Jun 21 14:05:05 then again... testing it has been extremely limited Jun 21 14:05:06 eheh Jun 21 14:05:08 zecke_: and read http://dominion.kabel.utwente.nl/koen/blog/pyblosxom.cgi/Handhelds/monotone.html Jun 21 14:05:13 http://ln-s.net/5Iq (at dominion.kabel.utwente.nl) Jun 21 14:05:15 since... you know... the poodle situation Jun 21 14:05:18 cdbot: shut up Jun 21 14:05:19 mickeyl: Don't know, don't care Jun 21 14:05:38 mickeyl: check out for poke Jun 21 14:05:51 zecke: I've uncommented it so you can easily play with it Jun 21 14:06:04 mickeyl: i have tested both kernels in 3.5.3 dir, one shoudl be for 255 and other for 250, anyway they give the same problem, that is the same the one from core gave. anyway thanks. if you need testing just ask Jun 21 14:06:30 darkschneider: well, I don't know what to tweak, so I don't know what we could test Jun 21 14:06:47 darkschneider: how do the crashes manifest ? Jun 21 14:06:50 something in logread ? Jun 21 14:06:59 nope :( Jun 21 14:07:05 it's real hard crash Jun 21 14:07:47 i can generate 2 types.. one is just pinging an interface or accessing to it with ssh, the other is just using one app (zrally) crash allways it in 5-10 mins Jun 21 14:07:58 hmm Jun 21 14:08:03 that sounds like faulty hardware to me Jun 21 14:08:09 could you play with a sharp rom for some days ? Jun 21 14:08:10 eekk Jun 21 14:08:14 sure Jun 21 14:08:20 just to make sure Jun 21 14:08:29 oki, rigth do not wanna to make you lose time Jun 21 14:08:45 nah, i'd like to help but my ideas on that are just limited Jun 21 14:08:53 (i used my Z really a few!!! so it's really new.. sayit has 100 hours of use max) Jun 21 14:10:20 hmm Jun 21 14:10:23 something comes to mind Jun 21 14:10:35 zecke_: you can mess with it alrady, but I'm interested in testing merges and stuff, so send me the key to give you write access Jun 21 14:10:38 you told me that you have WE unmatching Jun 21 14:10:45 is that with the wellenreiter from the upgrades feed? Jun 21 14:11:23 mickeyl: yes Jun 21 14:11:27 nope Jun 21 14:11:27 uh oh. can you do a uname -a for us Jun 21 14:11:31 it's all from stick rom Jun 21 14:11:37 sure Jun 21 14:12:56 2.4.18-rmk7-pxa3-embedix #1 thu,25 sep 2003 12:18:42 +0100 aemv5tel unknown Jun 21 14:13:11 hah! Jun 21 14:13:15 * mickeyl dances a jig Jun 21 14:13:18 wellenreiter is from the opie default image Jun 21 14:13:20 like i was suspecting Jun 21 14:13:24 you are running a wrong kernel Jun 21 14:13:25 eeh.. ok what i did wrong? Jun 21 14:13:27 all things can happen Jun 21 14:13:33 really? Jun 21 14:13:33 don't you get the WARNING on boot ? Jun 21 14:13:39 mm this is the 3rd kernel i try Jun 21 14:13:49 aye.. i said that yesterday Jun 21 14:13:53 hmm Jun 21 14:13:56 not to me Jun 21 14:14:00 it's crystal clear then Jun 21 14:14:06 if the kernel doesn't match Jun 21 14:14:12 hell can freeze Jun 21 14:14:30 and debian can release Jun 21 14:14:34 indeed.. i can take the logs and copy my sentence from yesterday "is that warning message bad... because it looks so" Jun 21 14:14:37 uname -a needs to look like that: Jun 21 14:15:04 the fact is that this kernelis from the openzaursus dir Jun 21 14:15:45 no way Jun 21 14:15:48 look at the date Jun 21 14:15:50 2003 Jun 21 14:15:55 we didn't release openzaurus 3.5.3 in 2003 Jun 21 14:16:01 you kernel didn't flash Jun 21 14:16:03 indeed.. that looked strange to me too.. Jun 21 14:16:05 probably misrenamed or so Jun 21 14:16:11 mumbles Jun 21 14:16:11 that's why i invented the WARNING Jun 21 14:16:24 that was giid invention i say Jun 21 14:16:30 good Jun 21 14:16:33 heh Jun 21 14:16:50 and that the second one you have :) (at least that i knwo) Jun 21 14:16:51 the "real" kernel would tell you it's from Nov. 2005 and something like Jun 21 14:16:54 2.4.18-embedix-3.5.3 Jun 21 14:16:56 why not put the kernel in the rootfs? Jun 21 14:16:58 * koen runs Jun 21 14:17:03 ~shoot koen Jun 21 14:17:05 * ibot shoots koen in the eye with a glue gun! Jun 21 14:17:19 now i have to discover why that thing didnt0 flash Jun 21 14:17:27 mickeyl: so yous till have the updater script decoder? Jun 21 14:17:35 sure. grab it from vanille.de/tools Jun 21 14:17:38 tx Jun 21 14:17:57 pxa-updater-encsh Jun 21 14:18:07 aye.. tx Jun 21 14:20:16 ah, both zimage.bin and zImage will work...didnt0 known Jun 21 14:21:01 RP: please remember to add modules to the machine config when you change the defconfig Jun 21 14:22:16 mickeyl: Ah, sorry. I thought only the essential ones needed to be listed there? Jun 21 14:22:23 that's right Jun 21 14:22:24 however Jun 21 14:22:25 -CONFIG_PCMCIA_PCNET=y Jun 21 14:22:25 +CONFIG_PCMCIA_PCNET=m Jun 21 14:22:30 i consider that as essential :) Jun 21 14:22:38 Right :) Jun 21 14:23:26 ~lart mickeyl Jun 21 14:23:27 * ibot strangles mickeyl with a 9-pole serial cable Jun 21 14:23:35 heh Jun 21 14:23:37 what did i do? Jun 21 14:23:58 surely a 25 pin would be more effective? Jun 21 14:24:04 most likely Jun 21 14:24:06 or a scsi cable Jun 21 14:24:13 rs485 :) Jun 21 14:25:04 Is there a nice graphical gps program that will pan maps nicely as you move around? Jun 21 14:25:15 in windows Jun 21 14:25:19 gpsdrive should do that Jun 21 14:25:21 opie: qpegps; x11: gpsdrive Jun 21 14:25:22 for ~300 Jun 21 14:25:29 or pygps Jun 21 14:25:38 unfortunately both won't probably work with recent gpsd Jun 21 14:25:47 CIA-23: bloody poke command should work now Jun 21 14:25:51 or roadster Jun 21 14:25:54 yeha Jun 21 14:25:56 ~praise zecke Jun 21 14:25:57 All hail zecke! Jun 21 14:26:07 virtually nothing works with a recent gpsd Jun 21 14:26:18 bahh! then the dotted bar did not got removed... grrr.. Jun 21 14:26:26 Can I use an old gpsd? Jun 21 14:26:42 dotted bar? Jun 21 14:27:06 emte: sorry was a rant.. the dotted bar indicating the process of flashing.. Jun 21 14:27:49 mickeyl: ok for not it's not booting :) Jun 21 14:27:49 the zmodem status in minicom? Jun 21 14:28:11 emte: nm...i just talk too much Jun 21 14:28:40 hey me too Jun 21 14:28:45 .. when i am alive Jun 21 14:28:49 * mickeyl play with peek and poke Jun 21 14:29:10 or i just make mickeyl's life miserable by checking in bad things Jun 21 14:29:13 mickeyl: it could work Jun 21 14:29:38 mickeyl: can I force loading of a module? Jun 21 14:29:40 * darkschneider mumles at this nice "openzaurus" image on his PDA wondering if to use it as a picture Jun 21 14:29:44 zecke: you mean reloading? Jun 21 14:29:48 mickeyl: yes Jun 21 14:29:57 zecke_: reload( module ) Jun 21 14:30:10 mickeyl: can I inject code into a instance? Jun 21 14:30:19 d = {} Jun 21 14:30:33 zecke: sure. scoping order first searches for instance methods, then for class methods Jun 21 14:30:37 d['methods']['my_app'] = my_imp Jun 21 14:31:08 s/scoping/NRO/ Jun 21 14:31:59 * mickeyl watches bitbake parsing and taps his foot Jun 21 14:33:29 * emte finds an oversized branding iron to mount the smd ... oh wait thats not a branding iron thats my soldering iron Jun 21 14:34:11 must be time to etch an adapter board Jun 21 14:34:40 bbl Jun 21 14:34:48 time to sleep Jun 21 14:34:51 night all Jun 21 14:34:57 g'night pigi Jun 21 14:37:10 * koen monotone commit --message="csets 1.3591.1.1, 1.3594 and 1.3595: RP does stuff to the kernel and libaticore" Jun 21 14:38:45 BB>> poke nano FOO bar Jun 21 14:38:45 OK Jun 21 14:38:45 BB>> peek nano FOO Jun 21 14:38:45 bar Jun 21 14:38:47 excellent Jun 21 14:39:10 mickeyl: now try with CACHE disabled ;) Jun 21 14:39:21 * mickeyl does that Jun 21 14:39:54 bitbake-ng will be even better in this regard Jun 21 14:40:06 with scopes you know which variable came from where (scopes) Jun 21 14:40:10 woooot Jun 21 14:40:13 you could poke and regenerate the file Jun 21 14:40:14 * kergoth has his car back from service :D Jun 21 14:40:31 kergoth: rock Jun 21 14:40:33 *cheers* Jun 21 14:40:40 hm, that reminds me, I should take my car for service. Jun 21 14:40:43 nearly had to pay the full $250 for a new tire Jun 21 14:40:48 theres no road hazard warranty on these Jun 21 14:40:49 it's about a year overdue, which is rather poor form Jun 21 14:40:58 just warranty against defects, and they were claiming it was from an impact Jun 21 14:40:58 :\ Jun 21 14:41:02 pb_: doh, thats no good Jun 21 14:41:34 * emte thinks ... all technical people and how many know how to chaing thier oil? Jun 21 14:41:42 change* or spell Jun 21 14:42:04 heh. well, knowing how to change the oil is one thing, but _wanting_ to do it is quite another. Jun 21 14:42:06 i can change the oil, though my car needed quite a bit more for its service than just that Jun 21 14:42:08 that too Jun 21 14:42:13 i'd rather just pay and not have to deal with it Jun 21 14:42:29 cars have oil in them? Jun 21 14:42:32 heh Jun 21 14:42:58 Only european ones Jun 21 14:43:00 jacques: well, maybe yours doesn't any more Jun 21 14:43:30 I just wish I could get the brakes to work on mine... Jun 21 14:43:35 ! Jun 21 14:43:35 I thought the oil pressure gauge on zero was a good thing Jun 21 14:43:56 lol Jun 21 14:43:57 yeah, that's right. less stress on the pump that way Jun 21 14:43:58 I want my oil to be calm, not under any pressure Jun 21 14:44:08 pressure is bad, right? Jun 21 14:45:53 yes, you dont want ANY blood pressure Jun 21 14:46:01 it might get to your brain that way :P Jun 21 14:47:47 ewwwww Jun 21 14:47:51 blood in my brain?? Jun 21 14:47:55 that sounds nasty Jun 21 14:50:39 emte: I can... sortof change my oil Jun 21 14:50:53 emte: does it include anything other than checking amount and pouring it in? Jun 21 14:51:40 lol Jun 21 14:52:15 * emte sets up an "Techie Oil Change" station ... Jun 21 14:55:13 * zecke_ leaves the code unfinished and heads to bed Jun 21 14:55:31 'night zecke_ Jun 21 14:55:37 g'night zecke Jun 21 14:55:51 * mickeyl follows his example Jun 21 14:55:52 g'night Jun 21 14:56:02 * koen does the same Jun 21 14:56:04 'night all Jun 21 15:09:24 * pb_ has flat battery Jun 21 15:09:25 later all Jun 21 15:46:23 03koen 07 * r1.3596 10openembedded/packages/libgpevtype/libgpevtype_0.12.bb: libgpevtype_0.12.bb: remove old install_append Jun 21 17:11:40 hmm. why does gdb-cross-6.3 eat the do_stage task? Jun 21 17:18:10 it's hungry Jun 21 17:20:05 'night all Jun 21 17:40:28 argh, I hate it when I spend hours looking at a problem and then find out I have a *different* problem Jun 21 17:40:46 the problem I thought I had: usbnet bombs with my code and gets messed up Jun 21 17:41:11 the real problem: my code is a hardware controller and the device it enables creates some horrible EMI and THE CABLE GOES RIGHT BY THE USB CABLE Jun 21 17:42:15 I should've though about that when I started getting strange phantom mouse clicks on my PC when the cable was near the mouse Jun 21 17:45:38 marcan: what kind of device? o.o Jun 21 17:46:01 actually it is a PWM controlled CCFL inverter Jun 21 17:46:15 o.O wtf is that? Jun 21 17:46:25 PWM = digital power controlling Jun 21 17:46:36 i.e. quicly turning on and off to simulate a % power Jun 21 17:46:56 and a CCFL inverter is one of those things that up 12V to 1000V or more to run one of those nifty PC mod lighting tubes Jun 21 17:47:03 (sorta like fluorescent lighting, but thinner) Jun 21 17:47:29 so really I'm just contrlling the brightness of the thing with my Z Jun 21 17:47:29 * luke-jr_ ponders what that could possibly be used for =p Jun 21 17:47:46 making some kind of glowing case? Jun 21 17:48:11 basically, but instead of going the standard route (add lotsa expensive hardware, lotsa light, lotsa everything) I go the creative route Jun 21 17:48:20 I embed my old semibroken SL-5500 in the front of the case Jun 21 17:48:24 and use it as a system controller Jun 21 17:48:32 add a custom I/O board and it can do about anything Jun 21 17:48:59 I made a really cool control panel as an SDL app to run on it Jun 21 17:50:41 the funny thing here Jun 21 17:50:47 is now my *computer case* runs linux Jun 21 17:50:50 geeky. Jun 21 18:00:41 problem solved: changed the PWM frequency. It seems much more harmless now. Jun 21 18:30:15 marcan, taht doesnt sound the most efficient way to do things ... Jun 21 18:31:00 curious how much that impacts the life of your tubes tho Jun 21 18:31:57 you could have just use the A/D and adjusted the voltage/current levels Jun 21 18:32:43 suppose it depends on the tube type tho ... Jun 21 18:33:46 but a nifty reuse of your z Jun 21 18:34:20 generally PWM is better suited for fans or motors Jun 21 18:35:16 emte: why would the tube care? Jun 21 18:35:24 the PWM is smoothed out by a capacitor anyway Jun 21 18:35:32 the big advantage of PWM is near zero power dissipation Jun 21 18:35:41 since it is only on/off Jun 21 18:35:47 if it is of the flourecent type it has to "ignite" on every pulse shortening it's life Jun 21 18:36:07 it doesn't. It's fast enough. Jun 21 18:36:55 the tube is always fed on AC anyway. Same for normal fluorescents. So technically they reignite constantly under normal operation. Jun 21 18:37:32 hmm, i always thought flourecents used a ac-dc tranformer Jun 21 18:37:37 nop Jun 21 18:37:41 transformer* Jun 21 18:37:46 they run on AC Jun 21 18:38:22 the cause of the deterioration under constant turning on and off (as in playing with the switch) is because during ignition they are fed a high voltage pulse and this deteriorates the electrodes Jun 21 18:38:44 but once they are ionised that doesn't happen Jun 21 18:39:19 its a good thing i am going into lighting :P Jun 21 18:39:30 they do turn on and off really fast, but they don't need reignition since the plasma regenerates easily :P Jun 21 18:39:47 anyway, my dektop fluorescent lamp died yesterday Jun 21 18:39:53 lol Jun 21 18:39:55 it made a nice disco style fast blinking display Jun 21 18:40:02 unrealated i am guessinfg Jun 21 18:44:34 and that one had no PWM through it :P Jun 21 18:44:36 I s'pose a faulty starter or electrodes Jun 21 18:44:36 BTW, it is actually bad for fluorescents to run on DC Jun 21 18:44:36 the cathode electrode is always harmed more than the anode Jun 21 18:44:36 so if they are run on DC, the cathode deteriorates faster Jun 21 18:44:38 they are always run on AC to balance the deterioration on both electrodes Jun 21 21:00:08 I have a few newbies questions; went through the openembedded website and am confused. Jun 21 21:00:25 can i ask a few stupid questions over here ? Jun 21 21:00:45 what is open embedded ? Jun 21 21:01:00 ... Jun 21 21:01:09 The website doesn't answer that? Jun 21 21:01:20 it's a repository of package metadata for constructing embedded linux distributions. and that information is most definately in the wiki Jun 21 21:02:00 first i will tell u what i know of open embedded. please correct me if i am wrong . Jun 21 21:02:38 i think it is a collection of sources for compiler, binutils and library sources Jun 21 21:02:48 not at all Jun 21 21:02:54 i am using cross tools now Jun 21 21:02:54 more like instructions for building those Jun 21 21:03:10 and what they depend on, etc Jun 21 21:03:15 so, u dont have the sources with open embedded. Jun 21 21:03:20 nope Jun 21 21:03:24 ram123: generally OE downloads and patches the necessary sources Jun 21 21:03:34 from where ? Jun 21 21:03:40 kergoth: BitBake is considered part of OE? Jun 21 21:03:43 upstream, generally. the official download locations Jun 21 21:03:49 ram123: OE refers to the locations of sources Jun 21 21:03:51 no, bitbake is a tool that acts on the OE metadata Jun 21 21:04:01 so what does oe consist of a few shell scripts ? Jun 21 21:04:04 bitbake is to OE as gnu make is to makefiles Jun 21 21:04:23 what is oe metadata ? Jun 21 21:04:25 kergoth: so that should be "generally BitBake downloads..." Jun 21 21:04:34 ram123: OE metadata is OE itself, now Jun 21 21:04:37 ram123: variables holding information like urls, descriptions, etc, and then yes, "tasks", which are executable shell script tidbits Jun 21 21:04:55 luke-jr_: well, its ambiguous, because its OE's base.bbclass that defines the fetch task. Jun 21 21:05:02 ah Jun 21 21:05:08 kergoth: Hey Chris, How are ya 8-) Jun 21 21:05:25 not too bad, finding this whole "being broke" thing doesnt sit well with me Jun 21 21:05:27 hehe Jun 21 21:05:28 you? Jun 21 21:05:48 i am using cross tools now. Jun 21 21:06:07 i download the required lib and cross compile. how different is oe from this Jun 21 21:06:09 * luke-jr_ wonders what 'cross tools' is Jun 21 21:06:11 Pretty good, Very busy at work. Lucky you. Did you find any interesting jobs, or are you still not looking Jun 21 21:06:21 ram123: oe isnt just the toolchain. its thousands of packages. Jun 21 21:06:31 ram123: it is used to build numerous linux distros for embedded devices Jun 21 21:06:43 are these packages part of open embedded Jun 21 21:07:02 ram123: metadata (scripts to build, info about them) for those packages are in OE. Jun 21 21:07:18 which means, bitbake+oe knows how to obtain, patch, and build those things for any arbitrary target platform Jun 21 21:07:34 ram123: you run 'bitbake gpe-image' and BitBake+OE output a complete image file Jun 21 21:07:59 Geo_KM: sorta looking. mostly only looking around here, since i'm not ready for another move, not even knowing what i want to do. hopefully i can find a local linux job to pay the bills, or find some short term contract work thats telecommute or something. Jun 21 21:08:00 * kergoth shrugs Jun 21 21:08:13 might check a technical temporary job service type place for something Jun 21 21:08:14 that is when i run this command u mean to say it downloads cross compiles and gives u the required exe Jun 21 21:08:17 kergoth: I thought you were moving *for* work? O.o Jun 21 21:08:32 ram123: except that there's no exe... Jun 21 21:08:44 ram123: it emits binary packages for all those things, and knows how to construct a root filesystem with them, and an image from that. Jun 21 21:09:13 inside the file system there will be exe ?. Jun 21 21:09:20 i do that manually. Jun 21 21:09:26 I'm not aware of any Windows targets.. Jun 21 21:09:38 i am using omap 5912 ? Jun 21 21:09:42 ram123: in the filesystem will be binaries and configuration files and whatever else, for all the packages you're building Jun 21 21:09:50 OE supports omap today. it'll build you a rootfs and kernel Jun 21 21:10:00 does it support X windows also ? Jun 21 21:10:04 it can build X, yep Jun 21 21:10:10 and numerous X applications Jun 21 21:10:17 kergoth: I finally got around to upgrading my SL-6000 to OZ-3.5.3. Works great, just the few known bugs so far. Jun 21 21:10:34 Geo_KM: glad to hear it :) i need to blow the dust off the units i have here and load a recent oz myself Jun 21 21:11:00 so, u mean to say i give what i want say OMAP Kernel 2.6.11, X windows, GPE and it gives u a file system containing these and an Kernel Image Jun 21 21:11:19 kergoth: Yeah Believe it or not I got so used to actually using it as a PDA I never seemed to get around to the upgrade. 8-) Jun 21 21:11:21 What is OMAP? Jun 21 21:11:23 it downloads these from the required locations and cross compiles and does that job. Jun 21 21:11:32 yup Jun 21 21:11:40 luke-jr_: its a TI processor that has both an ARM core and a DSP Jun 21 21:11:48 luke-jr_: used in some pdas and countless mobile phones Jun 21 21:11:52 hmm... OE supports it? O.o Jun 21 21:11:54 when is the cross compiler and bin utils built ? Jun 21 21:12:05 ram123: one of hte first steps in the build process. it builds the tools it needs to build the rest Jun 21 21:12:09 ram123: part of the dependency process Jun 21 21:12:10 does it do it every time i need to build a package ? Jun 21 21:12:25 ram123: only if there's a new version or something Jun 21 21:12:25 it builds it once, it tracks what it has already done and doesnt do it again unnecessarily Jun 21 21:12:28 yeah Jun 21 21:12:53 if i want a specific package that is not part of oe list of packages ? Jun 21 21:13:11 then you'd create a new file so oe knows how to build it Jun 21 21:13:15 for most packages its pretty painless Jun 21 21:13:30 if it uses a normal ./configure, make, make install type thing for example, we have classes that abstract out common build processes Jun 21 21:14:25 heh... even kdelibs is taking less work than I expected ;) Jun 21 21:15:12 hehe Jun 21 21:15:43 ... mickey asked for a patch to disable Qt's depending on non-diet X11 Jun 21 21:15:46 ram123: i'd recommend taking a look at some of the OOO newsletters, they usually have pointers to examples of our .bb metadata Jun 21 21:15:56 it seems to be a simple -DQT_NO_XIM Jun 21 21:16:17 ah Jun 21 21:16:20 cool Jun 21 21:16:43 now the question is... is there a way to make it optional? o.o Jun 21 21:17:18 what i do now to build Linux Disto is download bin utils, gcc sources and glib sources and some librariy sources and get a list of cross tools (gcc, ld, ar) and cross compiled libraries Jun 21 21:18:03 i take a new package say X windows and its friends cross compile them using the cross tools and build the file system myself Jun 21 21:18:07 kergoth: is PR="r5.1" valid, BTW? Jun 21 21:18:14 ram123: sounds time consuming Jun 21 21:18:16 ram123: oe will save you a lot of headaches Jun 21 21:18:20 i am looking at open embedded becuase i heard it is simple Jun 21 21:18:27 luke-jr_: should be, afaik Jun 21 21:20:00 ram123: An example of how simple, I have an OE-based distro for Cxx0 Zaurii that has daily image releases :) Jun 21 21:57:32 i was looking for gpe. gpe has 50 dependend libs. so i have to download each one of then cross compile and then cross link and it takes lots of efforts. so someone told me look at oe it will simplify Jun 21 21:57:58 the task. Jun 21 21:58:13 can u tell me how open embedded works in this case ? Jun 21 21:59:52 gpe is already in OE and builds for most deices Jun 21 21:59:59 devices* Jun 21 22:00:03 ram123: oe will download, patch, and build each of them, obeying dependencies to ensure they're built in order. Jun 21 22:00:40 how do i specify what packages i need ? Jun 21 22:00:49 does it do by tself ? Jun 21 22:00:52 it provideds the build system, compilation system, image generations , as wwll as extra package generation Jun 21 22:01:15 how does it find out what libs are needed ? Jun 21 22:01:32 dependent libs ? Jun 21 22:01:44 folows the image metadata Jun 21 22:02:46 simular and more effective than emerge/apt-get source/etc Jun 21 22:12:25 checking whether we can use openpty... configure: error: cannot run test program while cross compiling Jun 21 22:12:37 Is there a normal way to get around such? Jun 21 22:26:39 so i have to give it the meta data required for each packages and its dependency. but what about the location where the sources are there ? Jun 21 22:27:06 that's in the metadata too Jun 21 22:27:19 you just tell it a directory to get metadata from and it will do the rest Jun 21 22:28:31 ram123: just read GettingStarted on the wiki and try it :) Jun 21 22:28:48 so what if the existance of lib is not known to gpe say inetutils ? Jun 21 22:29:07 then GPE obviously doesn't depend on it... o.O Jun 21 22:29:24 or you will need to create, test, and add the lib Jun 21 22:29:31 sorry i was referring to oe Jun 21 22:29:51 have you looked to see what is in the OE package list? Jun 21 22:29:55 if oe does not know the existance of a lib say inetutils ? how does it built for it Jun 21 22:30:03 what do you mean? Jun 21 22:30:07 i am just taking an example ? Jun 21 22:30:07 i dont understand the question Jun 21 22:30:10 ram123: then you add that lib at the same time as you add the program depending on it Jun 21 22:30:11 you will need to create, test, and add the lib Jun 21 22:31:15 let us assume there exists a library source with me and i want to cross compile it and put it into the file system. how does oe handle it ? Jun 21 22:31:30 ram123: you make a new metadata file for it Jun 21 22:32:58 what is the hard disk size required for open embedded + bit blake ? Jun 21 22:33:24 does all the 2000 lib sources need to be stored in my hard disk ? Jun 21 22:33:34 BitBake uses about 2 MB Jun 21 22:34:24 if i want compiler, linker; it downloads whatever is required and makes the required executables ? Jun 21 22:34:40 does oe knows about TI OMAP 5912 ? Jun 21 22:34:53 OpenEmbedded (metadata) uses about 125 MB Jun 21 22:35:18 omap5912osk seems to have an OE config Jun 21 22:35:58 Machine configuration for the OMAP Starter Kit with a 5912 processor Jun 21 22:36:25 ram123, you need ~8gb Drive space Jun 21 22:36:40 ram123, kergoth might Jun 21 22:36:47 he worked with TI for a while Jun 21 22:37:01 u mean to say how to compile 2000 packages where to download them is stored in 125 MB for 2000 packages ? Jun 21 22:37:12 ram123: 125 MB is the metadata Jun 21 22:37:29 ram123: source tarballs for my system are 1.6 GB, but that varies on how much you build Jun 21 22:37:41 build directories get quite large Jun 21 22:37:52 glib, gcc, etc **** ENDING LOGGING AT Tue Jun 21 23:59:56 2005