| 00:07.41 | florian | good night |
| 00:21.37 | *** join/#oe _keith_ (~keith@CPE-144-131-119-130.nsw.bigpond.net.au) |
| 00:25.42 | *** join/#oe mithro (~tim@202.191.111.52) |
| 00:26.22 | *** join/#oe mithro_ (~tim@202.191.111.52) |
| 00:31.25 | *** join/#oe _keith__ (~keith@bh02i525f01.au.ibm.com) |
| 00:36.58 | kergoth | should switch the OE build of uclibc from cvs over to svn |
| 00:42.11 | keturn | wlan-ng modules don't build targeting uclibc. and I'm wondering if I should bother, or if I can just use hostap modules. |
| 00:42.19 | kergoth | you cant. |
| 00:42.24 | kergoth | for for the sl6000 |
| 00:42.27 | kergoth | er, not for |
| 00:42.37 | kergoth | hostap doesnt support usb |
| 00:43.26 | keturn | I don't think I'll need usb. |
| 00:46.20 | kergoth | then why are you building wlan-ng? |
| 00:46.28 | kergoth | the sl6000's internal wireless _is_ on usb. |
| 00:46.38 | kergoth | if you have an sl6000, you need wlan-ng. |
| 00:47.29 | keturn | I don't have an sl6000. |
| 00:47.40 | kergoth | then why are you building wlan-ng? |
| 00:48.04 | keturn | I want drivers for a Prism 3 chip. |
| 00:48.12 | kergoth | then you want hostap. |
| 00:48.28 | kergoth | and the prism3 firmware support package |
| 00:51.29 | keturn | thanks, didn't know about prism3-support |
| 00:52.01 | kergoth | np |
| 00:52.58 | *** join/#oe tux_mike (~mike@zelda.tuxnami.org) |
| 01:02.33 | *** join/#oe rphillips (~rphillips@rphillips.developer.gentoo) |
| 01:02.50 | chouimat|tv | " he will turn an homophobe gay" |
| 01:03.15 | robilad | who is miguel de icaza? |
| 01:03.26 | robilad | (best salesman, ever) |
| 01:03.51 | kergoth | gnome guy. works for ximian |
| 01:03.53 | kergoth | (iirc) |
| 01:04.10 | robilad | oh, i actually meant that as a joke. |
| 01:04.15 | kergoth | heh |
| 01:04.30 | chouimat|tv | according to somes: an ass |
| 01:04.36 | kergoth | oh, jeapordy answer |
| 01:04.39 | robilad | to the homophobe thing. tunring. good salesman. tv shows. quiz. |
| 01:04.40 | *** join/#oe tux_mike (~mike@zelda.tuxnami.org) |
| 01:04.41 | robilad | yeah! |
| 01:04.43 | robilad | :) |
| 01:05.05 | kergoth | man, doing an svk mirror of the uclibc svn takes a long ass time |
| 01:05.09 | kergoth | they have like 9500 revisions |
| 01:05.13 | chouimat|tv | robilad: it's the ;lyrics from one of the Colonel Claypool's Bucket Of Bernie Brains song |
| 01:05.36 | *** join/#oe tux_mike (~mike@zelda.tuxnami.org) |
| 01:05.37 | robilad | oh, popular culure for 1000? i'm out. |
| 01:05.39 | robilad | :) |
| 01:05.49 | chouimat|tv | aka Primus |
| 01:08.03 | CIA-4 | 03kergoth 07 * r1.3197 10openembedded/ (3 files in 2 dirs): |
| 01:08.03 | CIA-4 | Path variable updates: |
| 01:08.03 | CIA-4 | <PROTECTED> |
| 01:08.03 | CIA-4 | <PROTECTED> |
| 01:08.03 | CIA-4 | <PROTECTED> |
| 01:08.04 | CIA-4 | <PROTECTED> |
| 01:08.06 | CIA-4 | <PROTECTED> |
| 01:12.14 | Cwiiis | this is very odd... Something keeps cards busy when they get added to a unionfs on insertion, but it's fine if added manually after insertion... |
| 01:12.30 | kergoth | check fuser and/or lsof |
| 01:12.33 | kergoth | see whats keeping them busy |
| 01:18.12 | CosmicPenguin | Less for today - check the fucking device matrix before buying a bluetooth serial dongle |
| 01:18.16 | CosmicPenguin | s/Less/Lesson/ |
| 01:32.48 | Cwiiis | kergoth: lsof gave me nothing, but I think cardmgr is the culprit... Turning automatic mounting off and mounting manually is fine |
| 01:38.14 | kergoth | i dont see how. cardmgr isnt what does the mounting, it just runs the 'ide' script. |
| 01:38.19 | Cwiiis | ah, ok |
| 01:38.20 | kergoth | it has no awareness of that |
| 01:38.27 | kergoth | ide is what mounts, if and only if ide.opts is setup that way |
| 01:38.36 | Cwiiis | right |
| 01:39.46 | Cwiiis | what provides ide? |
| 01:39.54 | kergoth | ls /etc/pcmcia/ide |
| 01:40.12 | kergoth | grep '/etc/pcmcia/ide' /usr/lib/ipkg/info/* |
| 01:40.13 | Cwiiis | ah, still pcmcia-cs... Sorry, just missed the file |
| 01:40.15 | kergoth | heh |
| 01:40.17 | kergoth | np |
| 01:47.37 | Cwiiis | well, it's definitely a script in pcmcia-cs that's causing the problem... I'm guessing /etc/pcmcia/shared... Just gotta debug :/ |
| 02:04.16 | Cwiiis | I have no idea why it does this, what the script does is fine... |
| 02:05.26 | *** join/#oe qfh (~qfh@ppp-82-135-64-12.mnet-online.de) |
| 02:08.56 | Cwiiis | I give up, 2.6 kernel development is probably easier :p |
| 02:14.45 | CIA-4 | 03cwiiis 07 * r1.3199 10openembedded/conf/distro/openzaurus.conf: Add TARGET_FPU = "soft" to conf/distro/openzaurus.conf |
| 02:14.50 | CIA-4 | 03cwiiis 07 * r1.3188.1.1 10openembedded/ (8 files in 4 dirs): Rewrite of unionroot and unionroot-utils - Same functionality as before, but less errors and more efficient. |
| 02:18.24 | Cwiiis | Which kernel should I be looking at if I want 2.6 on collie? |
| 02:18.34 | Cwiiis | handhelds-sa? |
| 02:21.43 | Cwiiis | "linux-openzaurus_2.6.11-rc2.bb"... I'd best go to bed. |
| 02:38.15 | keturn | samba-3.0.11 is out. doing "cp samba-3.0.10.bb samba-3.0.11.bb" seems to work just fine. |
| 02:38.35 | kergoth | cool |
| 02:42.08 | *** join/#oe HuntsMan (~hunts@pc-148-118-120-200.cm.vtr.net) |
| 02:45.46 | *** join/#oe gb2 (~gbritton@fog.sekrit.org) |
| 02:45.47 | *** join/#oe [g2] (~g2@rdu57-8-035.nc.rr.com) |
| 02:49.44 | emte | wtf ... |
| 02:49.47 | emte | | checking size of long double... configure: error: cannot compute sizeof (long double), 77 |
| 02:49.56 | emte | thats an odd error and an odd return |
| 02:50.25 | gb2 | what's in config.log? |
| 02:50.33 | Cwiiis | kergoth: 'PREFERRED_PROVIDERS += " virtual/kernel:linux-openzaurus"' - Does that look ok to you? I can't get bb to prefer the linux-openzaurus kernel |
| 02:50.40 | kergoth | emte: make sure its using AC_CHECK_SIZEOF, not a custom size macro |
| 02:50.45 | gb2 | a bunch of architectures don't have long double, but i thought it was defined to just be a double in those cases |
| 02:52.04 | emte | hmm |
| 02:52.36 | kergoth | Cwiiis: the machine .conf is probably overriding you. |
| 02:52.44 | kergoth | Cwiiis: it gets parsed _after_ local.conf, remember. |
| 02:53.03 | Cwiiis | ah, I hadn't thought of checking the machine conf |
| 02:53.56 | Cwiiis | if it used ?=, that wouldn't happen, would it? |
| 02:54.30 | kergoth | ? |
| 02:54.40 | kergoth | dont use PROVIDERS. |
| 02:54.52 | kergoth | you cant accurately control it, because its a space seperated string |
| 02:54.59 | kergoth | use PREFERRED_PROVIDER_virtual/kernel = |
| 02:55.04 | kergoth | and make that use ?= in the machine .conf |
| 02:55.08 | Cwiiis | ok |
| 02:55.09 | kergoth | then yes, you'll be able to override |
| 02:56.04 | emte | hmm looks liek the error might be due to a missing g++ |
| 02:56.46 | emte | time for supper first tho |
| 02:57.21 | gb2 | there's a good chance a missing g++ error is actually a redherring. the test may actually be wrong in those cases. |
| 03:44.43 | *** join/#oe _keith (~keith@CPE-144-131-119-130.nsw.bigpond.net.au) |
| 03:51.51 | CIA-4 | 03kergoth * r144 10bitbake-data/ (ChangeLog src/backend-sqlite3.c tests/check_bitbake_data.c): Start adding bits to rework the sqlite3 tables to facilitate being able to do scope resolution. |
| 03:55.56 | *** join/#oe [g2] (~g2@rdu57-8-035.nc.rr.com) |
| 04:01.31 | kergoth | hmm |
| 04:04.37 | [g2] | #qemu |
| 04:04.44 | [g2] | ooops :) |
| 04:05.03 | kergoth | hehe |
| 04:05.42 | kergoth | [g2]: so having talked a bit more in the portage channel, i'm fairly confident that I can convince the portage developers that libbitbake (our rewrite) will make a suitable backend for the far off portage rewrite. |
| 04:05.46 | kergoth | :) |
| 04:06.12 | [g2] | kergoth cool |
| 04:07.01 | [g2] | I'd hope that metadata tools would either merge or become seamlessly interoperble |
| 04:07.34 | kergoth | well, that'd certainly be nice. libbitbake -can- make use of the existing .ebuild stuff if need be though |
| 04:07.46 | kergoth | i wonder if i can convince them to go with a directly parsed format |
| 04:09.25 | CIA-4 | 03kergoth * r145 10bitbake-data/ (configure.ac doc/doxygen/Makefile.am): Unbork the doxygen doc generation bits. |
| 04:18.12 | CIA-4 | 03kergoth * r146 10bitbake-data/ (. config/ doc/ doc/doxygen/ src/ tests/): Add svn:ignore bits. |
| 04:20.09 | CIA-4 | 03kergoth * r147 10bitbake-data/doc/ (Makefile.am test_scoped_lookups.sql): Add .sql which shows how we're doing scoped lookups w/ sqlite3, for tracking purposes. Add a maintainer-clean-local rule to doc/Makefile.am to remove Makefile.in in maintainer-clean. |
| 04:29.03 | _keith | Has anyone tried to "bitbake konqueror-embedded" lately? It seems to die with all sorts of errors.? |
| 04:31.54 | _keith | Also "bitbake meta-opie" dies when it gets to milage as it appears the package is not available. |
| 04:34.15 | CP|Home | google maps is so damn pretty |
| 04:34.20 | kergoth | yep |
| 04:35.20 | *** join/#oe Marsupilami23 (~Marsupila@152.139-60-66-nokia-static.surewest.net) |
| 04:38.39 | *** join/#oe _keith_ (~keith@bh02i525f01.au.ibm.com) |
| 04:44.19 | chouimat|Zzz | night |
| 04:48.45 | CIA-4 | 03kergoth * r148 10bitbake-data/doc/test_scoped_lookups.sql: Rename some of the fields and tables in the scoped lookups .sql to improve clarity. |
| 05:05.11 | punkass|inclass | anyone know what the best java VM is for oz? |
| 05:11.39 | *** join/#oe qip|Virus|da (~Thomas@pD95DB7ED.dip.t-dialin.net) |
| 05:16.56 | kergoth | hrm |
| 05:17.02 | kergoth | dillemma, this scope resolution method is flawed |
| 05:17.13 | kergoth | or rather, nto flawed, but limited |
| 05:19.39 | *** join/#oe pb_ (~pb@2002:5168:d330:1:20b:6aff:fe40:e27f) |
| 05:23.32 | ljp | does oe support other compilers other than gcc? |
| 05:23.43 | kergoth | doubt its ever been tried |
| 05:24.17 | ljp | hmm |
| 05:24.32 | pb_ | shouldn't think oe itself cares very much. if you supply an appropriate set of CFLAGS, you ought to be in business. |
| 05:29.47 | *** join/#oe joshua_ (joshua@cl-303.ams-04.nl.sixxs.net) |
| 05:41.50 | punkass | kergoth: ping |
| 05:44.36 | punkass | haha...no more questions for you...well not about lego anyways |
| 05:44.43 | punkass | just workin the css stuff |
| 05:44.47 | kergoth | ahh cool |
| 05:44.51 | punkass | id="head-69f66f05c13d3178a5d37ef5d764c37a2dee6ce0" |
| 05:45.01 | punkass | what is that? something moin generates? |
| 05:45.04 | kergoth | ew |
| 05:45.07 | kergoth | must be |
| 05:45.12 | kergoth | google for moin moin info |
| 05:45.31 | punkass | ok just thought i would check to see if you knew |
| 05:49.36 | emte | doobie doobie doo |
| 05:49.43 | emte | found my problem |
| 05:49.52 | emte | not sure how to fix it yet |
| 05:54.55 | kergoth | punkass: i wasnt even the one that setup this moin :) never messed with it that way either |
| 05:55.15 | emte | hmm |
| 05:55.23 | emte | wonder why ld cant find c |
| 05:56.56 | emte | wait thats not the problem ... |
| 05:57.00 | emte | odd |
| 05:57.15 | emte | you ever encounter: |
| 05:57.19 | emte | # ld -lc |
| 05:57.19 | emte | ld: warning: cannot find entry symbol _start; not setting start address |
| 06:06.09 | dave-osk | The QTE patches seem to be missing some +#if defined(QT_QWS_IPAQ) || defined(QT_QWS_SL5XXX) |
| 06:06.28 | dave-osk | which makes builds break on other boards ... known issue? Intentional? |
| 06:06.37 | dave-osk | I'll of course send the patch I end up with... |
| 06:08.17 | kergoth | certainly unintentional, not a known issue |
| 06:08.56 | dave-osk | just frustrating. :) OK, patch to be forthcoming. |
| 06:09.16 | dave-osk | (I didn't know if there was any intent to only support certain hardware just yet.) |
| 06:10.02 | kergoth | oe supports what, 15 target machines right now? probably just no one tried to use qte on the others yet |
| 06:10.16 | kergoth | the goal is to be relatively target os, device, and architecture agnostic |
| 06:11.16 | emte | hmm i wonder wth dietlibc is |
| 06:11.45 | dave-osk | that's what i thought ... it's just that getting opie to build has been annoying, what with qte changes etc |
| 06:12.00 | kergoth | ahh right |
| 06:18.07 | *** join/#oe __law__ (~law@mail.stiftadmont.at) |
| 06:29.30 | punkass | kergoth: is this what you were thinkin? http://70.66.215.174/oe/FrontPage.html |
| 06:29.40 | *** join/#oe raster (~raster@YahooBB219198104088.bbtec.net) |
| 06:31.25 | kergoth | punkass: oh thats much nicer |
| 06:31.42 | kergoth | yeah, thats like what i was thinking, much better fit with our scheme |
| 06:32.01 | punkass | ok good..if you are happy with it ill email ya the css |
| 06:32.29 | emte | heya raster |
| 06:32.32 | punkass | i didnt get moin up and going on my server so i just worked it on this one page |
| 06:33.02 | raster | emte: emmie! |
| 06:39.10 | *** join/#oe rphillips (~rphillips@rphillips.developer.gentoo) |
| 06:43.41 | *** join/#oe zedkatuf (~dfutak@82-32-58-45.cable.ubr08.azte.blueyonder.co.uk) |
| 06:44.18 | punkass | kergoth: so did you want me to mail it to ya? |
| 06:45.15 | kergoth | punkass: yeah |
| 06:45.24 | punkass | ok |
| 06:48.06 | punkass | done. |
| 06:49.00 | raster | kergoth: kergie! |
| 06:57.22 | emte | stupid specifix |
| 07:02.35 | *** join/#oe zilt (~aking@ottawa-dial-206-191-4-22.s-ip.magma.ca) |
| 07:04.41 | *** join/#oe keturn (~kevint@c-24-21-141-200.client.comcast.net) |
| 07:05.12 | *** join/#oe gints (~gints@159.148.213.135) |
| 07:12.29 | *** join/#oe pyrus (~rcagle@crtntx1-ar10-4-3-133-122.crtntx1.dsl-verizon.net) |
| 07:12.45 | *** part/#oe pyrus (~rcagle@crtntx1-ar10-4-3-133-122.crtntx1.dsl-verizon.net) |
| 07:29.47 | *** join/#oe darkschneider (~gab@81-208-36-80.fastres.net) |
| 07:32.55 | *** join/#oe micropal_ (ident@139.80-202-82.nextgentel.com) |
| 07:46.22 | *** join/#oe cedric (~cedric@195.94.59.29) |
| 08:00.14 | koen | good morning all |
| 08:04.55 | punkass | morning...even tho its midnite here :) |
| 08:04.59 | cedric | morning |
| 08:05.51 | *** join/#oe macmac (~mac@f02v-11-213.d1.club-internet.fr) |
| 08:06.03 | *** join/#oe bluelightning (~bluelight@203-173-244-75.akl.ihugultra.co.nz) |
| 08:06.10 | koen | good timezone adjustd time of day whi |
| 08:06.13 | koen | wish |
| 08:06.21 | bluelightning | hi koen |
| 08:06.28 | koen | hey bluelightning |
| 08:06.37 | bluelightning | hello all |
| 08:06.52 | punkass | ello |
| 08:09.47 | hrw | hi |
| 08:09.59 | hrw | today is time to reflash collie and reinstall all on it |
| 08:10.23 | hrw | hopefully meta-opie feed can be stripped to 64MiB so I can save it on cf |
| 08:12.57 | koen | "your item has been received in AMSTERDAM, NETHERLANDS and is progressing through the network for delivery." <- my corgi is coming :) |
| 08:13.23 | bluelightning | :) |
| 08:13.33 | bluelightning | new or 2nd hand? |
| 08:13.39 | koen | 2nd hand |
| 08:15.08 | hrw | bluelightning: there is no such thing as new zaurus (other then sl-c3000) probably |
| 08:15.23 | bluelightning | hmm |
| 08:15.38 | bluelightning | you can still get them new from japan though I think |
| 08:15.44 | punkass | i am working with a new 6000...tho its not mine :( |
| 08:15.46 | koen | or trisoft.de |
| 08:16.00 | bluelightning | I saw a site that will set them up as english and send them to you from there |
| 08:16.02 | koen | if you live in/near germany |
| 08:17.04 | hrw | and have spare money |
| 08:17.33 | koen | *lots* of spare moneyC |
| 08:17.43 | koen | E800 for a Z! |
| 08:18.29 | hrw | let users of sl-c3000 unite and donate one to OZ team ;) |
| 08:18.33 | cedric | the 2.6 bootstrap image is still building sharp-flash-header-c700 on my poodle |
| 08:19.25 | cedric | I've added the jffs2 lines from poodle.conf to openzaurus-pxa-2.6.conf |
| 08:19.48 | hrw | cedric: change zaurus-clamshell.con |
| 08:19.49 | hrw | f |
| 08:19.55 | koen | bluelightning: the ts.conf issue has been solved |
| 08:20.16 | bluelightning | koen: yeah I saw some back-and-forth on the lists... great :) |
| 08:20.32 | hrw | have to go - will be back in 30m |
| 08:20.44 | koen | pb forgot bk diffs doesn't recurse :) |
| 08:23.21 | koen | bluelightning: I'm building a new set of snapshots right now |
| 08:25.01 | bluelightning | koen: ah ok |
| 08:25.20 | bluelightning | koen: just sent a big list of stuff to gpe@ |
| 08:26.06 | koen | Yeah, read it |
| 08:26.21 | koen | 3 of them are related to compositeext |
| 08:26.36 | koen | the nice shadows under your windows |
| 08:26.44 | bluelightning | ah ok |
| 08:27.03 | koen | it will be diabled soon, as kdrive is buggy in that respect |
| 08:31.38 | *** join/#oe mithro (~tim@202.174.36.83) |
| 08:39.23 | *** join/#oe XorA (~dp@81-178-215-10.dsl.pipex.com) |
| 08:40.25 | XorA | morning |
| 08:40.38 | bluelightning | hi XorA |
| 08:42.58 | koen | hey XorA |
| 08:45.31 | hrw|work | re |
| 08:49.54 | Cwiiis | heh, kernel 2.6 ignores the on/off button on collie :) |
| 08:50.44 | Cwiiis | Gonna have to have a play with that when I get back on monday... |
| 08:53.21 | hrw|work | Cwiiis: yep - collie_batswitch is compiled in afaik but ignored |
| 08:58.01 | XorA | Cwiiis: could be worse, 2.6 GPE the on/off switch just restarts X |
| 08:58.22 | XorA | Cwiiis: on husky |
| 09:11.32 | RP | XorA: Does it? Hmmm :-/ |
| 09:13.44 | RP | XorA: The button triggers a APM User Suspend request so iX is probably doing something funny... |
| 09:14.46 | hrw|work | Cwiiis: oops. my mistake.. on/off yes - batswitch also is ignored (or was ignored when I tested last time) |
| 09:18.36 | hrw|work | ipkg -recursive is good thing |
| 09:19.18 | XorA | RP: apm -s locks the Z up completely |
| 09:19.39 | XorA | RP: and I dont have serial cable, I might bodge one together over weekend if I get time, Ive got an old sync cable I can cut |
| 09:19.53 | bluelightning | hehe... "bodge" |
| 09:19.58 | bluelightning | what a brilliant word |
| 09:20.04 | XorA | hrw|work: why does konqueror build think -lpcre is a target? |
| 09:20.09 | RP | XorA: arpm -s works fine from the console so X is doing something weird |
| 09:20.14 | hrw|work | XorA: dont know? |
| 09:20.16 | XorA | RP: most likely |
| 09:20.30 | XorA | RP: hmm, maybe I chvt to conole and try it |
| 09:21.06 | RP | The power button locks up X on mine here as well |
| 09:21.36 | XorA | RP: :-( thats 3 fixes needed before 3.5.3 then :-) |
| 09:22.36 | RP | Well, it doesn't lock the Z up - just X |
| 09:22.58 | RP | The system refuses to suspend and resumes as X is holding some kind of lock... |
| 09:23.07 | RP | XorA: What are the other two? |
| 09:23.30 | XorA | RP: the keymap for 2.4 and 2.6 is different so if you apply the 2.4 xmodmap the keys are all wrong |
| 09:23.48 | XorA | RP: The Xserver file points the TSLIB_DEVICE to /dev/sharp_ts |
| 09:23.52 | robtaylor_ | XorA, RP: its probably that at the console there's just one user of apm (apm -s), and when x is running theres 2 (apm -s, and XfbDev) |
| 09:24.23 | RP | XorA: The keymap difference is to be expected. You can probably use the kernel map for 2.6 |
| 09:25.04 | XorA | RP: I am using the map with no modmap, but if we go with 2.6 for 3.5.3 on clamshells then the xmodmap file needs pulled from ipk |
| 09:25.15 | robtaylor_ | there used to be a bug in the 2.4 kernels that was triggered by two apm users, due to refcounting being broken and hence selecting on the apm device didnt work |
| 09:25.25 | RP | Basically, gpe needs to be taught about the 2.6 kernel in the same way opie was... |
| 09:26.19 | RP | XorA: You make the openzaurus-pxa-2.6 machine type make those changes |
| 09:26.50 | RP | robtaylor_: Does that bug still exist under 2.6? apm was completely rewritten... |
| 09:27.18 | robtaylor_ | RP: Kdrive should just pull the keymap from the kernel. I believe its interpertation of the keymap is buggy. There's a bugzilla bug about it |
| 09:27.39 | robtaylor_ | RP: no, it'll be a different bug, but its quite likely the same kind of problem |
| 09:28.42 | XorA | apm -s from console while X is running gives apm: Resource temporarilly unavailable |
| 09:29.09 | robtaylor_ | XorA: ah well thats something different then ;) |
| 09:29.24 | XorA | ah, from dmesg is seems esd is holding something, RP did have sound as a problem in release notes |
| 09:29.31 | RP | XorA: That just means it doesn't want to suspend X... |
| 09:29.58 | RP | XorA: How recent is that kernel? |
| 09:30.06 | XorA | RP: built yesterday |
| 09:30.44 | RP | Sounds shouldn't be a problem from a suspend point of view. I killed that p[roblem... |
| 09:31.10 | robtaylor_ | XorA: hmm, i thought you said it locked up the Z completely |
| 09:31.11 | XorA | hmm, I cant even kill -9 esd |
| 09:31.23 | robtaylor_ | XorA: whats different this time? |
| 09:31.32 | XorA | robtaylor_: I am at console prompt |
| 09:31.36 | RP | esd? What is esd? |
| 09:31.44 | XorA | RP: sound mixing demon |
| 09:31.52 | koen | enlightened sound daemon |
| 09:32.21 | koen | alsa eliminates most of the uses for esd |
| 09:32.35 | robtaylor_ | XorA: yeah, alsa+dmix, and you can lose esd |
| 09:32.41 | RP | The kernel is very picky about which processes it will suspend. I suspect some applications are going to need fixing to be suspend friendly |
| 09:33.06 | robtaylor_ | having said that, esd is usually suspend freindly, i belive |
| 09:33.32 | XorA | robtaylor_: only if you have esdctl, which gpe doesnt seem to install |
| 09:33.33 | robtaylor_ | (on x86 apm boxen) |
| 09:33.35 | RP | I get a mixture of problems here with suspends and X. It seems you can lock the system into a state where is will no longer suspend... |
| 09:33.40 | robtaylor_ | XorA: ahhhhh.. |
| 09:33.57 | RP | I can't make the Z lock completely though |
| 09:34.14 | RP | SysRq and serial port running thoughout... |
| 09:34.44 | XorA | RP: it locked completely after a few cycles of me hitting the power button, well at least X and network were dead, no way to test anything esle |
| 09:35.02 | RP | Try SysRq... |
| 09:35.14 | XorA | RP: what is sysreq on Z? |
| 09:35.23 | RP | Alt+Home |
| 09:35.44 | XorA | RP: ok I shall do that next time |
| 09:36.47 | RP | I'll add it to my list of things to look into if I get a chance |
| 09:37.48 | RP | XorA: If you can track it down, that would be even better :) |
| 09:38.40 | RP | Keep in mind pressing the power button is a near enough direct equivlent to an "apm -s"] |
| 09:38.47 | RP | but done in kernel space |
| 09:39.12 | XorA | RP: cool, and it does seem like esd has got stuck somewhere, system wont even reboot now |
| 09:39.57 | RP | I found certain programs don't get on with suspends. I binned all the kernel threads as they broke suspends (as well as being horribly ugly) |
| 09:41.34 | RP | SysRq reset? |
| 09:41.42 | XorA | RP: seemd to be dead |
| 09:41.54 | XorA | unless sysreq keys changes in 2.6 |
| 09:42.02 | XorA | havent used it since debugging X on 2.4 |
| 09:42.05 | RP | Try Alt+Home, release Alt, press r |
| 09:42.05 | hrw|work | XorA: clamshells doesnt have 'full reset' button? |
| 09:42.17 | RP | hrw|work: no... |
| 09:42.18 | XorA | hrw|work: nope, you pull battery |
| 09:42.35 | XorA | hrw|work: hence why it takes us days to teach newbies to flash them |
| 09:42.36 | RP | My husky doesn't have the battery cover on it... |
| 09:43.00 | XorA | hrw|work: and diag screen is a total nightmare |
| 09:43.15 | RP | XorA: I know nothing else :) |
| 09:43.25 | hrw|work | ipkg-link suxx totally |
| 09:43.49 | koen | hrw|work: indeed |
| 09:44.01 | koen | symlink farms are prone to failure |
| 09:45.22 | hrw|work | I'm reinstalling my collie now and want to do it without symlinking /opt/ to cf card |
| 09:45.41 | XorA | RP: ok, with esd remove, unit will suspend with X, resume kills X dead though |
| 09:46.23 | RP | "kills X dead in what way?" |
| 09:46.31 | RP | does the screen come back or not? |
| 09:46.39 | XorA | RP: kills the process, then gpe-dm restarts it |
| 09:47.16 | RP | You need to find out why it dies then... |
| 09:47.39 | RP | That's not the kernel's fault... :-/ |
| 09:47.53 | XorA | RP: that I know, bloomin X :-) |
| 09:48.11 | RP | apm's handling of an aborted suspend looks faulty though... |
| 09:48.38 | RP | Although if could be part of the same problem |
| 09:48.38 | XorA | RP: looks like it just calls a resume here |
| 09:48.59 | RP | XorA: it does - resumes anything it suspended |
| 09:49.20 | RP | So I guess it would just make the X problem show up faster |
| 09:50.51 | XorA | RP: ok, I have a clue to the X problem, if you suspend from another VT X works fine, I suspect when X gets the signal to suspend its not letting go of VT like it should |
| 09:51.17 | XorA | I shall have to investigate kdrive, I remeber doing this code on bigX |
| 09:51.18 | RP | XorA: That's a big clue :) |
| 09:51.43 | XorA | RP: what signal does the kernel send to tell you to suspend? |
| 09:51.58 | RP | It tries to freeze the process IIRC |
| 09:52.12 | RP | See the power handling code in the kernel |
| 09:52.18 | *** join/#oe rob_w (~roboo@p213.54.140.102.tisdip.tiscali.de) |
| 09:52.25 | rob_w | hi, |
| 09:52.27 | XorA | RP: ok, I off to investigate (without getting sacked :-) |
| 09:53.14 | rob_w | seems my bitbake system went corrupted .. after parsing files i get errors like : in fire |
| 09:53.14 | rob_w | <PROTECTED> |
| 09:53.14 | rob_w | <PROTECTED> |
| 09:53.14 | rob_w | NameError: global name 'bb' is not defined |
| 09:53.33 | XorA | rob_w: sounds like a partial bbcow patch |
| 09:54.00 | rob_w | i had bbcow applyed long time ago |
| 09:54.09 | hrw|work | bbcow is not updated |
| 09:54.43 | rob_w | i did bk pull yestarday!? |
| 09:54.43 | RP | XorA: See kernel/power/process.c |
| 09:54.45 | XorA | RP: actually I remeber this from x86 suspend to disk, they had to make hack in kernel to return to a non X vt before suspend and reverse that after suspend, no idea if its still there |
| 09:55.17 | RP | XorA: It does that. There is a funny VT it uses for this. |
| 09:55.27 | RP | XorA: X never seems to give up to that though |
| 09:55.44 | RP | morning mickeyl |
| 09:55.54 | rob_w | is there a command to update the bitbake itself ? or does that happen also via bk pull ? |
| 09:56.03 | hrw|work | rob_w: svn update |
| 09:56.06 | mickeyl | morning RP |
| 09:56.08 | mickeyl | hi all |
| 09:56.32 | *** join/#oe CIA-2 (~CIA@flapjack.navi.cx) |
| 09:56.40 | RP | XorA: freeze_processes(void) is where it stops all the processes. It does something with VT's before that gets called though. No idea where it gets called from |
| 09:57.27 | *** join/#oe macmac (~mac@f07v-1-59.d1.club-internet.fr) |
| 10:00.08 | rob_w | hrw|work: that was it .. thnx .. so i guessed into the correct direction since i read in the ChangSet that i get now more information at start |
| 10:03.41 | *** join/#oe jbs (~jbs@213.58.72.17) |
| 10:04.31 | Bernardo | good morning |
| 10:05.54 | XorA | RP: doh, just had realisation, I probably wasnt seeing Z lock, just the RNDIS code hating suspends |
| 10:11.13 | *** join/#oe woglinde (heinold@lin226.mi.fu-berlin.de) |
| 10:11.29 | mickeyl | RP: you code is very clean |
| 10:11.42 | mickeyl | RP: good to learn from. |
| 10:12.10 | mickeyl | i have the first versions of tosa.h and tosa.c done |
| 10:12.23 | mickeyl | will try to boot that thing later tonight |
| 10:16.02 | Bernardo | mickeyl: can you help me a couple of minutes more with my "out of memory" in qpe error? Maybe in #openzaurus, to not piss kergoth? :) |
| 10:16.13 | mickeyl | heh kergoth is asleep |
| 10:16.28 | mickeyl | without gdb we can not do much |
| 10:16.29 | *** join/#oe marvin24 (~marvin24@fb07-iapwap2.physik.uni-giessen.de) |
| 10:16.39 | mickeyl | and even with gdb you probably have no debugging symbols |
| 10:16.44 | mickeyl | so it's of few value |
| 10:16.53 | mickeyl | let's try strace |
| 10:16.55 | mickeyl | install strace and do |
| 10:16.57 | mickeyl | strace bin/qpe |
| 10:17.03 | mickeyl | then paste the output into a pastebin.com |
| 10:17.34 | Bernardo | I have straced it already, I'll paste it in a sec |
| 10:18.55 | Bernardo | http://pastebin.ca/5618 |
| 10:19.23 | Bernardo | it seems it is after trying to load a korganizer icon |
| 10:20.01 | mickeyl | hmm yes, but that's during the normal launcher startup, so doesn't really indicate where the problem is |
| 10:20.11 | RP | mickeyl: Anything is an improvement on sharp's ;-) |
| 10:20.36 | RP | mickeyl: Fingers crossed for a successful tosa boot :) |
| 10:21.38 | mickeyl | RP: thanks. i can't wait to try it out |
| 10:21.55 | *** join/#oe tigrux (~tigrux@styx.suse.cz) |
| 10:22.24 | Bernardo | mickeyl: ok, It looks like I'll have to reflash it again tonight when I get home... |
| 10:23.14 | mickeyl | RP: atm. it's like walking through fog |
| 10:23.19 | mickeyl | i have no idea about the things i'm doing |
| 10:23.26 | mickeyl | which is quite not my style in general |
| 10:23.27 | mickeyl | heh |
| 10:24.00 | RP | mickeyl: Its something I'm rather used to at the moment :) |
| 10:24.24 | RP | Now you know how I feel when I look at bitbake and python ;-) |
| 10:24.35 | mickeyl | heh, can't be _that_ bad :) |
| 10:24.40 | hrw|work | ;)) |
| 10:25.03 | hrw|work | python.. yesterday I had a problem why my "if" is syntax error ;) |
| 10:26.26 | cedric | I've fixed my issue with the c700 header thingy by overwriding the EXTRA_IMAGEDEPENDS, but now I'm getting this --> http://pastebin.ca/5619 |
| 10:26.47 | RP | mickeyl: If I can help with anything, just ask. I've done the kernel learning curve recently and will help if I can... |
| 10:27.26 | mickeyl | RP: thanks for the offer. i'll definitly need it :) |
| 10:27.47 | mickeyl | cedric: is that the complete log? if not, check above for "missing dependencies" (or so) |
| 10:27.56 | mickeyl | bbl, meeting |
| 10:30.33 | cedric | mickey|meeting: yep that's the complete log |
| 10:38.17 | *** join/#oe keturn (~kevint@c-24-21-141-200.client.comcast.net) |
| 10:50.27 | CIA-2 | 03hrw 07 * r1.3200 10openembedded/packages/knights/knights_1.14.bb: knights: added build time dependency on phalanx to get all needed packages ready to use after build |
| 10:57.38 | *** join/#oe ade|desk (~adavey@194.200.143.249) |
| 11:03.34 | *** join/#oe lardman (~lardman@138.38.73.178) |
| 11:04.29 | lardman | I have a package which depends on kernel-modules-usbcore, but I'm not sure whether I can put that dependency in the bb file |
| 11:04.40 | lardman | as I don't know whether bb knows what provides it |
| 11:05.08 | lardman | it's not a build-time dependency, but rather a run-time one (e.g. I want ipkg to handle it) |
| 11:05.12 | lardman | What should I do? |
| 11:11.01 | hrw|work | you can set RDEPENDS |
| 11:11.28 | lardman | That will put the dependency in the control file? |
| 11:11.36 | hrw|work | yes |
| 11:11.41 | lardman | thanks |
| 11:13.50 | *** join/#oe zinga (~arnaudb@186.80-203-227.nextgentel.com) |
| 11:16.21 | koen | I have a small, slightly OT question: Where is that damn reset switch on a corgi? |
| 11:16.38 | hrw|work | koen: there is no reset on clamshell |
| 11:16.51 | koen | http://www.openzaurus.org/web/index.php?option=com_content&task=view&id=22&Itemid=44 |
| 11:16.57 | koen | according to that there is |
| 11:17.03 | hrw|work | 10:42 < hrw|work> XorA: clamshells doesnt have 'full reset' button? |
| 11:17.03 | hrw|work | 10:42 < RP> hrw|work: no... |
| 11:17.03 | hrw|work | 10:42 < XorA> hrw|work: nope, you pull battery |
| 11:17.10 | XorA | koen: there is none, we had this discussion earlier :-) |
| 11:17.15 | XorA | thanks hrw|work |
| 11:17.24 | koen | grrrrr |
| 11:17.38 | bluelightning | koen: I guess it really is inside the battery compartment then :) hahah |
| 11:17.49 | XorA | koen: unless its cleverly hiden |
| 11:18.57 | koen | XorA: http://dominion.kabel.utwente.nl/koen/pda/files/IMG_7025.JPG try to find it :) |
| 11:19.07 | *** join/#oe Cwiiis[uni] (~cil103@banshee.ecs.soton.ac.uk) |
| 11:20.13 | XorA | koen: I could just turn mine over :-) |
| 11:20.47 | hrw|work | koen: can you put larger one? |
| 11:20.48 | lardman | Back to my RDEPENDS - the module I need lives in kernel-module-usbcore_2.4.18-rmk7-pxa3-embedix-r32_shepherd, can I miss off everything after the kernel-module-usbcore bit? |
| 11:21.12 | hrw|work | kernel-module-usbcore is package name |
| 11:21.24 | lardman | cool |
| 11:21.31 | hrw|work | 2.4.18-rmk7-pxa3-embedix-r32 is version |
| 11:21.36 | hrw|work | shepherd is arch |
| 11:23.59 | XorA | but it allows you to chvt between consoles |
| 11:31.14 | Bernardo | http://pastebin.ca/5620 |
| 11:31.33 | Bernardo | how to crash opie with a corrupted saveFile in root/Applications/AlarmServer |
| 11:34.17 | Bernardo | I remebered to remove everything in Settings, didn't remember to check Applications... doh... |
| 11:57.39 | hrw|work | mickey|meeting: without fixing fontsizes in qt/e OZ 3.5.3 cannot be released - opie apps have it fixed but for example ko/pi show shit, so does iqnotes, portabase and probably other non-opie apps |
| 12:12.22 | *** join/#oe qfh (~qfh@ppp-62-245-163-65.mnet-online.de) |
| 12:20.40 | CIA-2 | 03hrw 07 * r1.3201 10openembedded/packages/meta/meta-opie.bb: meta-opie: fixed 'task-opie-todayplugins not found' problem |
| 12:21.04 | *** join/#oe stevenh (~xev@65.167.23.2) |
| 12:40.35 | *** join/#oe dkey (~dkey@L0003P05.dipool.highway.telekom.at) |
| 12:45.10 | cedric | I've understound why my 2.6 rootfs is not building, I've change the MACHINE name therefore OE doesn't regenerate the packages which are machine dependent... |
| 12:45.53 | cedric | ~lard OE |
| 12:47.34 | cedric | how can I regenerate all the machine dependent packages without removing the /tmp directory??? |
| 12:48.07 | koen | cedric: use multimachine.inc |
| 12:48.51 | cedric | koen: how do I do that? |
| 12:49.17 | hrw|work | include multimachine.inc in conf/local.conf |
| 12:49.24 | koen | include it in local.conf |
| 12:50.06 | cedric | and what that will do? |
| 12:50.15 | koen | that still has problems with multi-machine packages like tslib and sysvinit |
| 12:53.07 | koen | cedric: but you will have to clear /tmp after including that |
| 12:55.07 | *** join/#oe TyggerBob (~myname@CPE00045a265f64-CM000f9f7c58fa.cpe.net.cable.rogers.com) |
| 13:01.32 | *** join/#oe Tygger_Bob (~myname@64.26.144.98) |
| 13:02.48 | mickeyl | hrw|work: i have no idea how to fix that in QtE |
| 13:03.05 | mickeyl | hrw|work: this behaviour is triggered by the Qt Version, not by a function in Qte |
| 13:03.18 | mickeyl | hrw|work: which means the only place to fix it are the applications |
| 13:03.24 | mickeyl | hrw|work: or to go back to qte 2.3.9 |
| 13:03.38 | mickeyl | hrw|work: or to make qte 2.3.10 report 2.3.9 (ieek, hack alert) |
| 13:10.52 | mickeyl | RP: ping |
| 13:14.25 | *** join/#oe bluelightning (~bluelight@203-173-244-75.akl.ihugultra.co.nz) |
| 13:19.00 | *** join/#oe Cwiiis[uni] (~cil103@demon.ecs.soton.ac.uk) |
| 13:22.17 | *** join/#oe mac2 (~mac@f03m-4-247.d1.club-internet.fr) |
| 13:23.33 | *** join/#oe raster (~raster@YahooBB219198104088.bbtec.net) |
| 13:25.17 | hrw|work | mickeyl: ah... |
| 13:26.21 | mickeyl | yep, my first assumption that the font bug layed in qte was false |
| 13:26.35 | mickeyl | it was something i had to fix in opie |
| 13:26.45 | mickeyl | i.e. remove the workaround from the old 2.3.2 days |
| 13:29.15 | hrw|work | hmm ipkg .140 is really nasty |
| 13:29.42 | hrw|work | mickeyl: ok - I can look into sources of kdepimpi atleast - can you put an example fix? |
| 13:32.02 | *** join/#oe zilt (~aking@ottawa-dial-206-191-4-22.s-ip.magma.ca) |
| 13:32.23 | Cwiiis[uni] | hrw|work: I don't suppose you'd care to take a look at unionroot-utils sometime? It works perfectly if you manually mount media (DO_MOUNT="n" in /etc/pcmcia/ide.opts), but if it gets mounted automatically, it keeps the device busy... |
| 13:32.36 | mickeyl | hrw|work: i give you an example: |
| 13:34.39 | hrw|work | Cwiiis[uni]: rather not want to look into - first have some other things todo |
| 13:34.47 | hrw|work | I have to create my todo list on website or sth |
| 13:35.05 | Cwiiis[uni] | hrw|work: Fair enough - I've given up on it for now I think, want to move on to kernel 2.6 hacking |
| 13:36.07 | *** join/#oe jbs (~jbs@213.58.72.17) |
| 13:36.08 | XorA | hrw|work: you having ipkg problems? |
| 13:36.18 | hrw|work | yep |
| 13:36.22 | hrw|work | re Bernardo|zzzzz |
| 13:36.23 | Bernardo | re |
| 13:36.29 | XorA | hrw|work: as far as I can 140 is totally busted |
| 13:36.32 | koen | hey Bernardo |
| 13:37.16 | hrw|work | one day I will repackage opera 7.55 for my own usage |
| 13:37.37 | XorA | hrw|work: I get lots of invalid seek errors and @LargeLnk@ errors |
| 13:37.41 | Bernardo | hrw|work: that's the one available from alphaworks? |
| 13:37.46 | hrw|work | Bernardo: from ibm |
| 13:38.02 | koen | XorA: file a bug in hh.org bugzilla |
| 13:38.08 | XorA | koen: they are upto at least 142 |
| 13:38.17 | Bernardo | too bad it is compiled with 2.9x, and no soft-float... |
| 13:38.31 | XorA | koen: not had time to import later one |
| 13:38.37 | koen | ah, ok |
| 13:39.50 | koen | bbl |
| 13:40.21 | hrw|work | btw - is it normal that ipkg leave empty dirs? |
| 13:41.17 | mickeyl | if ( d->m_pointbug ) |
| 13:41.17 | mickeyl | <PROTECTED> |
| 13:41.17 | mickeyl | <PROTECTED> |
| 13:41.17 | mickeyl | <PROTECTED> |
| 13:41.17 | mickeyl | <PROTECTED> |
| 13:41.32 | mickeyl | this is our workaround for the Qt 2.3.2 on SharpROM |
| 13:41.41 | mickeyl | and this the detection patch: |
| 13:41.52 | Bernardo | mickeyl: did you see what was my problem with qpe failing to start? |
| 13:42.06 | mickeyl | - d->m_pointbug = ( qt_version() <= 233 ); |
| 13:42.06 | mickeyl | + d->m_pointbug = ( qt_version() == 232 || qt_version() == 233 ); // SharpROM uses 2.3.2 |
| 13:42.18 | mickeyl | Bernardo: yes, odd. |
| 13:42.25 | mickeyl | Bernardo: please put that into our bugtracker |
| 13:42.42 | Bernardo | ok - login guest/guest? |
| 13:43.01 | mickeyl | no, i mean the opie bugtracker |
| 13:43.04 | mickeyl | opie-bugs.oszine.de |
| 13:43.13 | mickeyl | you can register |
| 13:43.16 | mickeyl | an account |
| 13:43.37 | Bernardo | ok, I'll see if I can get there from work. |
| 13:56.37 | Bernardo | can't find the corrupted file - |
| 13:56.53 | Bernardo | sorry, pressed enter instead of backspace |
| 14:01.13 | Bernardo | mickeyl: the strangest part is that now I have a saveFile in root/Applications/AlarmServer, apparently created by kopi or kopi-applet, which has more gibberish in it (a few ^@) but doesn't kill qpe |
| 14:01.26 | Bernardo | submitted the bug as http://opie-bugs.oszine.de/view.php?id=1564 |
| 14:05.33 | mickeyl | k |
| 14:09.06 | chouimat | morning |
| 14:10.34 | Bernardo | hi chouimat |
| 14:11.53 | *** join/#oe zwi (~zwi@216.88.131.43) |
| 14:12.41 | robtaylor_ | meh, i feel stupid.. can someone remind me how to get boot messages out on serial? |
| 14:13.32 | robtaylor_ | i tried setting KERNEL_CONSOLE in local.conf, but nada.. |
| 14:14.13 | *** join/#oe lardman (~lardman@138.38.73.178) |
| 14:14.18 | RP | mickeyl: pong |
| 14:15.25 | mickeyl | RP: hi. i have two questions. a) could you imaging a problem by adding two scoop devices using your common scoop code ? |
| 14:15.42 | mickeyl | the only difference seem to be the IOaddr of those |
| 14:15.46 | robtaylor_ | ah, its openzaurus-3.5.3.conf's value that gets used, boh |
| 14:15.56 | *** join/#oe reenoo (~rwagner@x04.informatik.uni-bremen.de) |
| 14:16.08 | mickeyl | robtaylor_: yep, oz.conf is parsed after local.conf |
| 14:16.18 | reenoo | afternoon |
| 14:16.53 | *** join/#oe reenoo (~rwagner@x04.informatik.uni-bremen.de) |
| 14:17.09 | RP | mickeyl: The driver wasn't written to handle two scoops and will not work like that. It can easily be altered though |
| 14:17.47 | RP | mickeyl: Its halfway there, you just need to remove the global variables it uses... |
| 14:18.00 | mickeyl | hmm ok, i'll have a look |
| 14:18.13 | RP | Keep in mind, poodle, corgi and collie all use it... |
| 14:18.18 | hrw|work | XorA: I got patch to oz-compat to autocreate /tmp/*/*-0 link |
| 14:18.48 | mickeyl | RP: can you tell me about the sharpsl stuff in corgi.h ? it looks slightly misplaced there as it looks the same as in tosa.h |
| 14:18.53 | RP | There are small problems like how to specify which scoop you want when you call set_scoop_gpio() etc. |
| 14:19.07 | RP | The sharpsl_param stuff? |
| 14:19.27 | mickeyl | yeah there's some flash related defitions in corgi.h |
| 14:19.33 | mickeyl | which is the same as in tosa.h |
| 14:19.42 | mickeyl | so i think that should rather be in sharpsl.h shouldn't it? |
| 14:19.43 | RP | It is *not* flash... |
| 14:19.44 | hrw|work | XorA: http://pastebin.ca/5621 |
| 14:20.03 | RP | Its a header that appears before the kernel in memory before the kernel with some config info in it. The bootloader pulls it from flash and puts it there |
| 14:20.23 | RP | I was talking to JL about this. We're probably going to extract it into a common driver... |
| 14:20.40 | RP | The only important thing in there are the LCD parameters |
| 14:21.21 | mickeyl | hmm |
| 14:21.39 | mickeyl | ok, because the whole "Corgi Parameter Definition" section is identical to what's in tosa.h |
| 14:21.54 | mickeyl | it's also not changed by the SPitz patches |
| 14:22.02 | RP | Its also the same as that in poodle |
| 14:22.10 | mickeyl | ok, good. |
| 14:22.12 | mickeyl | something for later then |
| 14:22.27 | RP | rmk wasn't keen for it to be seperate at first but I think we have arguments for it now |
| 14:22.40 | mickeyl | yes. oh btw., spitz also uses the 2nd scoop, so it would be usable to enhance this. |
| 14:23.06 | RP | Do the GPIOs on both scoops get used? |
| 14:23.56 | mickeyl | yes. have a look at http://www.vanille.de/temp/tosa-patches/tosa-machine-support.patch |
| 14:26.47 | RP | Hmmm How did sharp handle that? Are there calls to something like set_scoop_gpio2() ? |
| 14:27.22 | mickeyl | yes, exactly |
| 14:27.41 | RP | mickeyl: I'd be tempted to just add a second IO_RESOURCE for the second scoop |
| 14:27.43 | *** join/#oe tigrux (~tigrux@styx.suse.cz) |
| 14:28.38 | RP | It needs some thought though. Intially you're fine just to use the first scoop device... |
| 14:28.57 | mickeyl | ok. in two hours i'll try to get it booting .) |
| 14:29.37 | RP | Using two IO_resouces would solve a lot of problems easily. I doubt Russell will accept it due to code duplication |
| 14:30.39 | mickeyl | yep |
| 14:31.16 | mickeyl | something like set_scoop_gpio( unsigned short num, unsigned short bit) would be the alternative |
| 14:32.16 | RP | but how does it know where to get the scoop device details from? :-/ |
| 14:32.48 | cedric | RP: in the corgi_mmc_detect function (corgi.c), is the mod_timer something u come up with? the HZ/4 ??? |
| 14:32.48 | XorA | hrw|work: looks good, as long as gpe follows symlinks, with a hard link I found a restart buggered it |
| 14:33.07 | mickeyl | RP: the device setup in <machine>.c |
| 14:33.18 | mickeyl | RP: there you add either one or two scoops |
| 14:33.26 | RP | cedric_: yes, why? |
| 14:33.46 | mickeyl | bbl, wandering hoem |
| 14:33.48 | mickeyl | home, even |
| 14:34.05 | RP | mickeyl: I know that. But how does set_scoop_gpio know which one to access without some kind of device data? |
| 14:34.07 | hrw|work | XorA: I started opera 7.55 with it |
| 14:34.11 | RP | and where does it get that device data from? |
| 14:34.16 | XorA | hrw|work: cool |
| 14:34.44 | cedric | RP: i was just wondering how u come up with that delay, just by some extensive testing??? |
| 14:34.46 | XorA | hrw|work: 7.55 runs for you? |
| 14:35.05 | RP | cedric: I made it up ;-) Sensible guess |
| 14:35.25 | hrw|work | XorA: yep |
| 14:35.30 | RP | Its the time the card takes to eject from the cardslot given the stength of the spring on corgi |
| 14:35.38 | XorA | hrw|work: grrr, they must have put some collie only crap in it |
| 14:35.48 | cedric | RP: ok |
| 14:36.23 | *** join/#oe [g2] (~g2@rdu57-8-035.nc.rr.com) |
| 14:36.34 | RP | cedric: The problem was that it would try and redectect cards as you were ejecting them which wasn't helpful... |
| 14:37.06 | *** join/#oe linuxwhore (~johnh@70-56-163-89.mpls.qwest.net) |
| 14:37.13 | CIA-2 | 03hrw 07 * r1.3202 10openembedded/packages/oz-compat/ (oz-compat/qt2310.patch oz-compat_0.5.bb): added patch to get oz-compat working with qt/e 2.3.10 |
| 14:37.15 | cedric | RP: I can find the implementation of "pxa_set_mci_info", I 've done a grep on the source and nothing... |
| 14:38.54 | RP | cedric: Look in drivers/mmc... |
| 14:46.59 | *** join/#oe dkey (~dkey@L0009P20.dipool.highway.telekom.at) |
| 14:48.36 | robtaylor_ | mm, i still feel stupid. i've tried console=ttyS0, console=ttyS1 and console=tty1, and non seems to give me a serial (and i've double checked that this is what's going into the .config) .. what am i doing wrong? |
| 15:16.41 | *** join/#oe prpplague (~dave@mailhost2.amltd.com) |
| 15:19.45 | mickeyl | hmm |
| 15:19.47 | mickeyl | Uncompressing Linux......................................................................... done, booting the kernel. |
| 15:19.47 | mickeyl | Error: a |
| 15:19.52 | mickeyl | that's all |
| 15:25.11 | hrw|work | heh.. my collie doesnt want to mount /dev/hda2 ;( |
| 15:26.22 | Bernardo | hrw|work: what format is your second cf partition? |
| 15:26.35 | hrw|work | ext2 ofcourse |
| 15:26.44 | CosmicPenguin | of course? |
| 15:27.05 | Bernardo | did you try "mount -t ext2 /dev/hda2 /mnt/..."? |
| 15:27.06 | hrw|work | CosmicPenguin: what else have a sense on collie? |
| 15:27.10 | hrw|work | Bernardo: sure |
| 15:27.14 | hrw|work | Feb 11 15:25:29 collie user.warn kernel: VFS: Can't find ext2 filesystem on dev ide0(3,2). |
| 15:27.19 | Bernardo | ouch |
| 15:27.22 | hrw|work | will ut fsck etc on it |
| 15:27.29 | Bernardo | fdisk -l /dev/hda2? |
| 15:27.42 | mickeyl | Uncompressing Linux is already output by the kernel isn't it? |
| 15:27.53 | CosmicPenguin | hrw|work: I was just surprised you said *of course* |
| 15:27.58 | hrw|work | CosmicPenguin: I have 256MB cf with 16MiB vfat (flashing) and rest for ext2 (stuff) |
| 15:28.00 | hrw|work | CosmicPenguin: ;) |
| 15:29.03 | hrw|work | Couldn't find ext2 superblock, trying backup blocks... |
| 15:29.03 | hrw|work | /dev/hda2 was not cleanly unmounted, check forced. |
| 15:29.23 | hrw|work | mounted ;) |
| 15:29.28 | hrw|work | /dev/hda2 221.6M 156.8M 53.4M 75% /media/cf |
| 15:31.40 | hrw|work | now time to move pcmcia stuff to be started sooner |
| 15:33.10 | hrw|work | or change opie startup |
| 15:42.37 | mickeyl | yahooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo |
| 15:42.52 | mickeyl | it panics! |
| 15:42.54 | mickeyl | STRIKE |
| 15:43.03 | mickeyl | Linux version 2.6.11-rc2 (mickey@r2d2) (gcc version 3.4.3) #13 Fri Feb 11 16:39:02 CET 2005 |
| 15:43.03 | mickeyl | CPU: XScale-PXA255 [69052d06] revision 6 (ARMv5TE) |
| 15:43.03 | mickeyl | CPU0: D VIVT undefined 5 cache |
| 15:43.03 | mickeyl | CPU0: I cache: 32768 bytes, associativity 32, 32 byte lines, 32 sets |
| 15:43.03 | mickeyl | CPU0: D cache: 32768 bytes, associativity 32, 32 byte lines, 32 sets |
| 15:43.04 | mickeyl | Machine: SHARP Tosa |
| 15:43.11 | RP | mickeyl: nice one :) |
| 15:43.13 | mickeyl | never was so glad to see a panic |
| 15:43.15 | mickeyl | heh |
| 15:43.33 | RP | I had the same moment. "I have a kernel panic" :) |
| 15:44.30 | mickeyl | although i have the feeling that his was the easiest part |
| 15:44.42 | RP | It took me months to get that far... |
| 15:44.52 | RP | pxa linking bug :-/ |
| 15:45.02 | mickeyl | right |
| 15:45.03 | hrw|work | fsck.. |
| 15:45.06 | mickeyl | you have done the pioneer work |
| 15:45.08 | hrw|work | collie doesnt wake |
| 15:45.28 | *** join/#oe kjoe (~tody@bantha.ftk.de) |
| 15:45.30 | mickeyl | hrw|work: doesn't wake up? |
| 15:45.36 | hrw|work | yep |
| 15:45.39 | mickeyl | hrw|work: from within opie? |
| 15:45.43 | hrw|work | yep |
| 15:45.48 | mickeyl | hrw|work: uh oh |
| 15:46.16 | mickeyl | hrw|work: can you do the following: |
| 15:46.44 | mickeyl | hrw|work: i'll point you to a piece of code to add into odevice_zaurus |
| 15:46.59 | mickeyl | then rebuild libopie2core |
| 15:47.44 | mickeyl | hrw|work: look at http://handhelds.org/cgi-bin/cvsweb.cgi/opie/libopie2/opiecore/device/odevice_zaurus.cpp.diff?r1=1.22&r2=1.23 |
| 15:47.46 | hrw|work | I have to go in 15 minutes and 1st want one bug to test |
| 15:47.47 | hrw|work | ok |
| 15:47.54 | mickeyl | ok, then whenever you have a chance |
| 15:48.03 | mickeyl | reintroduce Zaurus::setSoftSuspend into odevice_zaurus.cpp |
| 15:48.09 | mickeyl | and declare it in odevice_zaurus.h |
| 15:48.19 | mickeyl | i have a slight idea that could be responsible |
| 15:48.22 | mickeyl | but i'm not sure |
| 15:51.00 | Bernardo | my susp/resume problems have been caused by sockets left open |
| 15:51.09 | Bernardo | kdepimpi is the worst contender there |
| 15:51.31 | Bernardo | kopi sync always leaves at least one socket in time wait |
| 15:51.44 | Bernardo | this on the collie |
| 15:51.58 | hrw|work | mickeyl: odevice_zaurus.cpp 1.22 odevice_zaurus.h 1.10? |
| 15:52.45 | mickeyl | it makes much more fun developing w/ the sl6000 thanks to its reset button |
| 15:52.45 | mickeyl | heh |
| 15:52.48 | mickeyl | at least those early steps |
| 15:52.48 | mickeyl | without sysrq |
| 15:54.38 | *** join/#oe TyggerBob (~myname@CPE00045a265f64-CM000f9f7c58fa.cpe.net.cable.rogers.com) |
| 15:54.41 | hrw|work | odevice.cpp 1.28 |
| 15:54.44 | *** join/#oe micropal (ident@139.80-202-82.nextgentel.com) |
| 15:55.17 | hrw|work | odevice.h 1.16 |
| 15:55.26 | hrw|work | argh.. cvs usxx |
| 15:57.10 | hrw|work | uf.. build |
| 15:59.35 | hrw|work | heh.. now dont even try to suspend from opie |
| 15:59.48 | hrw|work | but "apm -s" suspends |
| 16:00.20 | hrw|work | and on/off resume |
| 16:00.55 | CosmicPenguin | These quilt permission bugs are pissing me off |
| 16:01.04 | *** join/#oe dkey (~dkey@L0009P20.dipool.highway.telekom.at) |
| 16:01.07 | hrw|work | anyway time to go |
| 16:01.50 | *** join/#oe marcansoft (1337@80.26.222.160) |
| 16:02.14 | Bernardo | bye hrw|work |
| 16:02.25 | hrw|gone | happy weekend |
| 16:27.39 | kergoth | Cwiiis: i'll do some unionroot testing if you'd like. need to play with it some more anyway |
| 16:27.43 | kergoth | monring |
| 16:27.44 | kergoth | morning, even |
| 16:28.58 | koen | hey kergoth |
| 16:47.14 | *** join/#oe rphillips (~rphillips@rphillips.developer.gentoo) |
| 17:05.31 | *** part/#oe ade|desk (~adavey@194.200.143.249) |
| 17:07.43 | *** join/#oe jamie_ (~jamie@modem-2042.porcupine.dialup.pol.co.uk) |
| 17:11.18 | *** join/#oe dkey (~dkey@L0009P20.dipool.highway.telekom.at) |
| 17:13.40 | kergoth | hrm |
| 17:16.57 | CosmicPenguin | hrm indeed |
| 17:18.20 | kergoth | i tried overriding all of our target path variables. out of 10 packages i built, like 7 of them didnt properly obey them |
| 17:18.23 | kergoth | heh |
| 17:18.26 | kergoth | whee |
| 17:19.41 | CosmicPenguin | kergoth: fyi, the i386/uClibc.config has CONFIG_PENTIUM4 selected as the processor type - I'm not sure if thats what you intended |
| 17:20.29 | kergoth | unintentional, but likely doesnt matter since we override the optimizations and shit anyway |
| 17:20.37 | kergoth | feel free to send a patch to the list |
| 17:21.03 | CosmicPenguin | Well, if nobody is complaining |
| 17:21.29 | *** join/#oe reenoo_ (~r@pD9528C5A.dip.t-dialin.net) |
| 17:21.33 | CosmicPenguin | I'm going to check in a i486 target here presently |
| 17:21.36 | CosmicPenguin | as soon as it builds |
| 17:21.49 | kergoth | ccool |
| 17:21.52 | kergoth | s/cc/c/ |
| 17:22.22 | CosmicPenguin | benchmarking libcs - what a fun way to spend a friday afternoon |
| 17:22.38 | kergoth | for i in */*.inc */*.bb; do sed -i -e 's,/usr/bin,${bindir},g; s,/usr/sbin,${sbindir},g; s,/usr/lib,${libdir},g; s,/usr/include,${includedir},g; s,/usr/share,${datadir},g; s,\${datadir}/man,${mandir},g; s,\${datadir}/doc,${docdir},g; s,\${datadir}/info,${infodir},g; s,/var,${localstatedir},g; s,/srv,${servicedir},g; s,/bin,${base_bindir},g; s,/sbin,${base_sbindir},g; s,/lib,${base_libdir},g;' $i; done |
| 17:22.44 | kergoth | whee |
| 17:22.50 | kergoth | who wants to bet that runs rampant and replaces things i dont want replaced |
| 17:23.55 | treke | morning |
| 17:24.40 | kergoth | hey |
| 17:30.40 | *** join/#oe andersee (~andersee@codepoet.org) |
| 17:37.26 | *** join/#oe SirFred (~mteira@85.Red-81-34-170.pooles.rima-tde.net) |
| 17:43.25 | treke | kitten |
| 17:43.28 | treke | it's what's for dinner |
| 17:45.22 | RP | mickeyl: Have you seen the post to OZ-Dev? |
| 17:47.45 | *** join/#oe moa_ (~moa@paris.bluebugs.org) |
| 18:05.37 | *** join/#oe marcanso1t (1337@80.26.245.147) |
| 18:07.32 | *** part/#oe tigrux (~tigrux@styx.suse.cz) |
| 18:17.02 | mickeyl | RP: yep, just read and reply. thanks |
| 18:17.05 | mickeyl | nce timing |
| 18:17.06 | mickeyl | nice, even |
| 18:17.24 | RP | It is :) |
| 18:17.33 | RP | I've also just replied |
| 18:18.48 | koen|tv | OE is up to 0.9% at linuxdevices.com |
| 18:18.58 | mickeyl | hmm |
| 18:19.00 | mickeyl | that would be my vote |
| 18:19.01 | mickeyl | heh |
| 18:19.01 | mickeyl | :) |
| 18:19.11 | mickeyl | please all go and vote |
| 18:19.14 | mickeyl | we need more eyes |
| 18:19.17 | koen|tv | and mine :) |
| 18:19.21 | mickeyl | heh :) |
| 18:19.25 | koen|tv | Number of voters: 216 |
| 18:19.35 | mickeyl | hmm so 0.9% would be more than us two |
| 18:19.51 | mickeyl | ah maybe not |
| 18:21.53 | *** join/#oe beewoolie (~beewoolie@florence.buici.com) |
| 18:22.02 | koen|tv | hmmm |
| 18:22.09 | kergoth | heh, look at hte % of homegrown |
| 18:22.16 | kergoth | thats what we need to sell OE into |
| 18:22.17 | koen|tv | oz isn't multimachine friendly |
| 18:22.18 | mickeyl | too much |
| 18:22.19 | mickeyl | right |
| 18:22.53 | kergoth | i'm skeptical about our ability to do so today, given the performance issues |
| 18:27.58 | Cwiiis | Am I missing something, or is PalmOS missing from that survey...? |
| 18:28.10 | *** join/#oe zap (~zap@217.170.93.9) |
| 18:29.30 | CosmicPenguin | knowing linuxdevices, its amazing CE isn't on the list |
| 18:30.03 | Cwiiis | well, I'd count CE under Windows |
| 18:30.08 | Cwiiis | oh, I get you |
| 18:30.09 | Cwiiis | ignore me |
| 18:43.33 | *** join/#oe Skyhusker (~fiam@80.103.56.147) |
| 18:43.38 | Skyhusker | hi |
| 18:44.25 | Skyhusker | anyone can tell which gdb package do i need to debug arm applications on my x86 desktop? |
| 18:45.04 | kergoth | ? |
| 18:45.25 | CosmicPenguin | Gooooo team! |
| 18:45.44 | *** join/#oe jamie___ (~jamie@softmodem.org) |
| 18:45.51 | CosmicPenguin | Check out x86 - holding its own |
| 18:46.04 | jamie___ | performance issues? |
| 18:47.22 | kergoth | with oe? the parse time and stuff is a real issue |
| 18:48.02 | koen|tv | especially the stuff |
| 18:48.18 | Skyhusker | kergoth: the are gdb and gdb-cross, which is supossed to produce a x86 binary? |
| 18:48.26 | Skyhusker | s:the:there: |
| 18:49.52 | kergoth | -cross, by definition |
| 18:50.05 | jamie___ | such precision! |
| 18:50.16 | kergoth | -native produces binaries that run _on_ the build machine _for_ the build machine. -cross produces binaries that run _on_ the build machine _for_ the target. |
| 18:50.19 | jamie___ | I'm having a hard time with gentoo-embedded and wondering if OE is worth the effort of switching |
| 18:50.26 | CosmicPenguin | Looking for access methods... Unfortunately, your OS is not supported by the PCI Library |
| 18:50.28 | CosmicPenguin | stupid libpic |
| 18:50.33 | CosmicPenguin | s/libpic/libpci/ |
| 18:50.34 | jamie___ | (Hard time == loads of stuff doesn't work without patching) |
| 18:50.54 | jamie___ | (I had the same problem of having to patch loads of stuff with ptxdist) |
| 18:51.00 | Skyhusker | kergoth: ok, thanks :) |
| 18:51.10 | kergoth | jamie___: i'd certainly recommend it. there are over 1300 crosscompilable packages in the upstrea metadata repository at the moment |
| 18:51.13 | kergoth | Skyhusker: np |
| 18:51.36 | jamie___ | Yeah, but the gentoo-embedded folks could make a similar claim. What's it like really? |
| 18:51.45 | Skyhusker | i'm trying to catch a bug in onetwork, and doing it wo/ gdb seems a bit hard |
| 18:52.29 | jamie___ | I.e. do complex things with complex dependencies like xine, mplayer and X.org usually just work? Or are they uncharted territory for OE? |
| 18:53.32 | koen|tv | people are making mythtv images using OE |
| 18:53.32 | kergoth | the gentoo-embedded folks are using an existing set of gentoo metadata. nothing goes into our repository without being patched to build with current autotools, crosscompile and build happily. |
| 18:53.38 | kergoth | should work fine. |
| 18:54.00 | kergoth | OE is used for a number of known sane, working distributions in the field. openzaurus, opensimpad, familiar, and others |
| 18:54.55 | jamie___ | Sounds promising. |
| 18:55.54 | koen|tv | bye bye TMPDIR |
| 18:56.36 | reenoo_ | Skyhusker: it depends on what you actually want. if you want to run a fully blown gdb on the target you'll need to build and install gdb. for gdbserver+gdb-cross you'll need to build gdb and gdb-cross, install gdbserver on the target and gdb-cross on the host |
| 18:57.35 | Skyhusker | reenoo_: i think i need both. i want to debug a signal 11 using the core file |
| 18:57.59 | Skyhusker | downloading gdb 6.3 know, will take a bit since i've 56k modem |
| 18:58.18 | CosmicPenguin | doh |
| 19:00.25 | *** join/#oe mickeyl_away (~LPA@141.2.14.64) |
| 19:30.10 | *** join/#oe badalex (~badalex@216.190.186.238) |
| 19:31.14 | *** join/#oe Laibsch (~leggewie@G136c.g.pppool.de) |
| 19:31.35 | *** join/#oe mallum (~mallum@p21.n-nypop04.stsn.com) |
| 19:42.55 | CosmicPenguin | Whats sad, is that my System.map is larger then my entire /bin directory on my target |
| 19:44.13 | CosmicPenguin | Whats sadder is that my System.map is *almost* bigger then the /lib directory |
| 19:46.35 | *** join/#oe keturn (~kevint@c-24-21-141-200.client.comcast.net) |
| 20:04.54 | CosmicPenguin | wierd |
| 20:05.10 | kergoth`food | whats it doing? |
| 20:05.22 | CosmicPenguin | a --dry-run is only showing the last target |
| 20:05.37 | CosmicPenguin | Its like it can't figure out the dependancies or something |
| 20:05.44 | kergoth`food | aha, i bet i know what that is. |
| 20:05.48 | *** join/#oe rob_w (~roboo@p213.54.140.102.tisdip.tiscali.de) |
| 20:05.49 | kergoth`food | are the others already built? :) |
| 20:05.52 | CosmicPenguin | yeah |
| 20:05.59 | kergoth`food | bitbake no longer fires events on packages its building if it isnt building them. |
| 20:06.12 | kergoth`food | thats why there was a nice performance boost doing builds when things were already done |
| 20:06.15 | CosmicPenguin | well, thats both good and bad, I guess |
| 20:06.19 | kergoth`food | it checks if they're done before it tries now |
| 20:06.25 | CosmicPenguin | bad because I wanted to get a BOM of packages |
| 20:06.28 | kergoth`food | the dry run is now more accurate |
| 20:06.34 | kergoth`food | it really is only showing what its going to do |
| 20:06.35 | CosmicPenguin | nod |
| 20:06.42 | kergoth`food | but yeah, the bitbake command needs a new way to get that information |
| 20:06.43 | rob_w | how do i tell via .bb file where certain libs are ? i create a new .bb file and the compile stops cause of missing xlibs and xaw |
| 20:06.45 | CosmicPenguin | thats pretty easy to work around |
| 20:06.48 | kergoth`food | no more relying on the event handling for that |
| 20:07.11 | kergoth`food | rob_w: you need to ensure that the OE variables get passed into its makefiles. |
| 20:07.16 | kergoth`food | rob_w: thingsl ike CC, CFLAGS, LDFLAGS, etc |
| 20:07.20 | kergoth`food | rob_w: those already set the paths correctly. |
| 20:07.31 | kergoth`food | rob_w: what app is thi? |
| 20:07.32 | kergoth`food | s |
| 20:07.41 | rob_w | that is what mostly is patched at oe , right ? |
| 20:07.50 | kergoth`food | what? |
| 20:07.52 | rob_w | i started trying xboard to get chess for gpe |
| 20:08.03 | kergoth`food | how its done is completely dependent on the buildsystem in question. |
| 20:08.13 | kergoth`food | for autotools, if CFLAGS & friends are in the environment, it automatically picks them up and uses them. |
| 20:08.19 | kergoth`food | (./configure does, that is) |
| 20:08.39 | rob_w | and how about xlibs ? |
| 20:08.44 | kergoth`food | ..? |
| 20:09.03 | kergoth`food | libX11 and friends are all seperate .bb's |
| 20:09.07 | kergoth`food | are you depending on the libs you need? |
| 20:09.21 | rob_w | it needs them to compile at least ! |
| 20:09.30 | kergoth`food | and? |
| 20:09.42 | kergoth`food | add the libs it depends on to its DEPENDS |
| 20:09.51 | kergoth`food | if you dont understand how DEPENDS works, you really ought to read the bitbake users manual. |
| 20:11.05 | rob_w | ok , thnx i think it just doesnt tell the app where the proper xlibs are .. i think i can figure it out now |
| 20:11.47 | kergoth`food | once again, the necessary paths so it can link against the libraries it depends on are all set in our CFLAGS, LDFLAGS, etc. |
| 20:12.55 | *** join/#oe dkey (~dkey@L0009P20.dipool.highway.telekom.at) |
| 20:12.57 | rob_w | have a nice one kergoth`food |
| 20:17.52 | *** join/#oe mallum (~mallum@p21.n-nypop04.stsn.com) |
| 20:19.21 | Laibsch | Hi, this is Rolf Leggewie who is trying to get a Japanese input method called anthy into OZ. Just wanted to let you guys |
| 20:19.21 | Laibsch | know that I just successfully installed bitbake et al. on a machine that does have enough RAM so I can finally get some work done ;-) |
| 20:20.58 | *** join/#oe HuntsMan (~hunts@pc-148-118-120-200.cm.vtr.net) |
| 20:36.29 | *** join/#oe woglinde (woglinde@pD95D1A06.dip.t-dialin.net) |
| 20:56.03 | *** join/#oe TyggerBob (~myname@CPE00045a265f64-CM000f9f7c58fa.cpe.net.cable.rogers.com) |
| 20:58.17 | *** join/#oe marcanso1t (1337@80.26.222.167) |
| 21:10.51 | *** join/#oe marcansoft (1337@80.26.245.251) |
| 21:17.42 | Laibsch | OK, the compilation aborted. |
| 21:18.21 | Laibsch | It got a 404 for quilt_savanah.nongnu.org_$BLAH from www.treke.net |
| 21:18.32 | Laibsch | I am using OE-exported |
| 21:18.44 | kergoth | dont have svn installed? |
| 21:18.49 | kergoth | quilt is pulled out of svn |
| 21:19.03 | kergoth | treke.net is just a mirror. it tried that, then fell back to the normal mechanism (svn) |
| 21:20.38 | Laibsch | I do have it and got bitbake that way |
| 21:20.45 | kergoth | then whats the real error? |
| 21:20.54 | kergoth | after the attempt via treke |
| 21:20.55 | Laibsch | OE-exported I pulled just about 2 hours ago |
| 21:20.56 | kergoth | it ran svn |
| 21:21.46 | Laibsch | Actually, it is not trying svn but cvs which I might not yet have. |
| 21:21.54 | Laibsch | Will check and report back |
| 21:23.08 | Laibsch | cvs was missing. I am installing it ATM. |
| 21:23.12 | kergoth | ah |
| 21:23.21 | kergoth | you might want to read the RequiredSoftware page on the wiki |
| 21:23.26 | kergoth | which, iirc, is linked to from GettingStarted |
| 21:23.48 | Laibsch | Thanx. |
| 21:24.17 | dave-osk | i-gotta-question ... |
| 21:24.45 | dave-osk | OE seems to expect kernel config to synch with userspace options |
| 21:24.58 | dave-osk | but i don't want that ... how to prevent that? |
| 21:24.59 | kergoth | in what way? |
| 21:25.24 | dave-osk | specific example: I want to have bridge utils built for the rootfs |
| 21:25.48 | dave-osk | but have bridging **NOT** be modular ... it complains, it demands modules. |
| 21:26.56 | dave-osk | (so building rootfs dies, etc -- no joy) |
| 21:27.10 | kergoth | sounds like a bug in the bridge utils buildsystem |
| 21:27.21 | kergoth | write up a patch to fix it |
| 21:28.01 | Laibsch | 2nd try. |
| 21:29.28 | dave-osk | the fix would be removing RDEPENDS kernel-module-bridge from that package |
| 21:29.54 | dave-osk | ... but there seems to be a widely adopted (yet undocumented) policy to have those dependencies |
| 21:30.11 | dave-osk | ... and to expect a highly modular kernel. OK, I can change that, but ... |
| 21:31.08 | dave-osk | ... what's the rationale? what hidden assumptions will break? anyone know? |
| 21:31.15 | kergoth | ahh |
| 21:31.18 | kergoth | hmm |
| 21:31.45 | kergoth | well, clearly the tools wont work without bridging support. i'd say kill the dep, and make the user deal with that |
| 21:31.56 | reenoo_ | no |
| 21:32.02 | reenoo_ | make it an RRECOMMENDS |
| 21:32.03 | kergoth | if you install bridge utils and dont have the bridging drivers in some fashion, you deserve what you get |
| 21:32.08 | reenoo_ | RDEPENDS is wrong |
| 21:32.11 | kergoth | ooh thats a good idea |
| 21:32.14 | kergoth | cleaner |
| 21:32.25 | dave-osk | That sounds like a good answer for every place I have this problem. |
| 21:32.49 | dave-osk | So that's what I'll do. Won't promise to scan all the packages though. |
| 21:33.04 | reenoo_ | right. it will have ipkg install the package if available only |
| 21:33.29 | kergoth | and it can be removed without removing the package that recommended it |
| 21:34.02 | dave-osk | ... suggestion: the Wiki really should have info like this in a BitBakeLore page or somesuch |
| 21:35.05 | kergoth | dave-osk: this is OE specific. its our classes that make use of it |
| 21:35.16 | kergoth | dave-osk: we have a bitbake users manual.. but we _need_ an OE users manual |
| 21:35.25 | kergoth | laying out what our classes do and why, what variables are used, etc |
| 21:35.50 | dave-osk | kergoth: so RRECOMMENDS etc is OE-specific? Agreed, there should be an OE manual, but ... |
| 21:36.01 | kergoth | yup |
| 21:36.11 | kergoth | its our packaging classes that take that and put it into the right fields in the ipk |
| 21:36.30 | kergoth | a wiki would be an excellent start on such a thing |
| 21:36.41 | kergoth | move it into docbook or something when it gets fleshed out |
| 21:37.03 | reenoo_ | we used to have that stuff in the old wiki IIRC |
| 21:37.27 | kergoth | some if its there, but a lot of it was just wrong |
| 21:37.33 | kergoth | needs updating |
| 21:37.40 | kergoth | bitrot |
| 21:38.00 | reenoo_ | ~lart bitrot |
| 21:38.19 | reenoo_ | oh |
| 21:38.23 | reenoo_ | and of course |
| 21:38.27 | kergoth | bbiab, meeting |
| 21:38.28 | reenoo_ | ~lart br1 |
| 21:41.12 | *** join/#oe zwi (~zwi@216.88.131.43) |
| 21:41.16 | *** join/#oe mallum (~mallum@p21.n-nypop04.stsn.com) |
| 21:46.41 | CIA-2 | 03reenoo 07 * r1.3203 10openembedded/packages/bridge-utils/bridge-utils_1.0.4.bb: bridge-utils: demote kernel-module-bridge from RDEPENDS to RRECOMMENDS |
| 21:55.16 | *** join/#oe ofauchon (~Olivier@137.100.119-80.rev.gaoland.net) |
| 22:14.17 | Bernardo | hi |
| 22:16.06 | *** join/#oe hedgehair (~hedgehair@user-0c6sdm3.cable.mindspring.com) |
| 22:21.43 | Laibsch | On the first run, OE used 400MB to get through the ~2000 bb files. Now I reran bitbake and the process uses |
| 22:21.43 | Laibsch | 600MB which makes my 512MB system crawl and swap. Is that normal? Should one remove intermediate |
| 22:21.43 | Laibsch | files from a previous run to start with a clean slate? If so, how? |
| 22:22.55 | hedgehair | is there development for a 3700 tslib-conf ? |
| 22:23.22 | kergoth | Laibsch: there's a parse cache in tmp/cache that could be affecting things. you can disable it, but itll slow down the first run. |
| 22:23.59 | hedgehair | i am trying to install qpdf2 but the requirement for libts-0.0-0 (>= 0.0cvs20050208) is also tslib-conf_0.0cvs20050126-r10_h3900.ipk which there is only 3900 series |
| 22:25.31 | hedgehair | mplayer also needs this |
| 22:27.09 | *** join/#oe cedric_ (~cedric@voulx.bluebugs.org) |
| 22:34.58 | Laibsch | kergoth: You are referring to the CACHE setting in conf/local.conf, right? I'll try it. Although I do not quite understand |
| 22:34.58 | Laibsch | how it will slow down the first run rather than slowing down subsequent runs. |
| 22:35.09 | kergoth | heh, right, thats what i meant |
| 22:37.30 | Laibsch | hehe |
| 22:38.18 | kergoth | its friday, what can i say |
| 22:43.20 | Laibsch | TGIF! |
| 22:43.26 | kergoth | ~friday |
| 22:43.27 | ibot | Friday sucks, because it should be the weekend, but you're stuck working anyway. |
| 22:43.53 | Laibsch | Well, come here to Europe, we already have weekend :-) |
| 22:44.29 | kergoth | hehe |
| 22:44.33 | CosmicPenguin | There you go, kergoth |
| 22:44.47 | CosmicPenguin | Of course, you're thinking barbados, where its always weekend |
| 22:44.59 | CosmicPenguin | Or have I been listening to too much Jimmy Buffet again? |
| 22:44.59 | Laibsch | That sounds nice. |
| 22:45.55 | Laibsch | ~sunday |
| 22:45.56 | ibot | Sunday sucks, because tomorrow is monday (See monday). Also, sunday is the day we do all the stuff we should've done Saturday, but didn't. |
| 22:45.58 | ofauchon | hi, could someone point me to an up-to-date tutorial on setting up openzaurus developement environnement thx |
| 22:46.23 | Laibsch | I was just curious ;-) |
| 22:46.36 | Laibsch | You guys covered the whole week, I guess? |
| 22:46.57 | Laibsch | ofauchon: openemebedded.org has a link to getting started |
| 22:47.15 | Laibsch | Here you go http://openembedded.org/cgi-bin/moin.cgi/GettingStarted |
| 22:47.40 | kergoth | CosmicPenguin: hehe, i've been thinking all sorts of things. no idea where i'm going next |
| 22:47.56 | ofauchon | Laibsch, thx |
| 22:48.25 | Laibsch | ofauchon: You're welcome. I am a bleeding newbie myself with this :-) |
| 22:49.02 | Laibsch | Do not hesitate to make additions to the wiki when you stumble and cannot move forward |
| 22:49.07 | Laibsch | on what is provided on the wiki. |
| 22:49.28 | ofauchon | Laibsch, just received my Wifi card this morning, time to start porting some apps tonite :-) |
| 22:50.02 | Laibsch | Good idea. Hope you have a powerful machine with LOOOTTTSSAA RAM. That kept me from |
| 22:50.11 | Laibsch | doing OE stuff until today. |
| 22:50.19 | woglinde | laibsch *g* |
| 22:50.27 | woglinde | 1 gig ram is cheap |
| 22:51.02 | ofauchon | Laibsch, powerfull machine, lots of ram ? mmmm, just for compiling applications ? why ? |
| 22:51.37 | Laibsch | I do not know, it just needs it. 512MB is really the absolute minimum. |
| 22:51.44 | Laibsch | How much do you have? |
| 22:51.56 | ofauchon | 512/ centrino 1.6ghz laptop |
| 22:52.08 | Laibsch | Should be OK, I guess. |
| 22:52.13 | ofauchon | what processes eats the ram ? |
| 22:52.53 | woglinde | package parsing |
| 22:55.07 | reenoo_ | parsing needs CPU. storing the data takes memory |
| 22:55.07 | Laibsch | But still with 2000 packages and 600MB footprint which I just had, that works out to about 250KB |
| 22:55.34 | Laibsch | per package which is something I fail to understand. |
| 22:55.50 | Laibsch | But I have no way to change it anyway so I do not care. :-) |
| 22:56.02 | reenoo_ | ~bbcow |
| 22:56.03 | ibot | bbcow is, like, a couple patches for bitbake and openembedded to use copy-on-write for the metadata handling. It greatly reduces memory consumption. Get them at http://www.frankengul.org/~seb/cowbb/. |
| 22:56.26 | reenoo_ | you _can_ change it :) |
| 22:56.52 | Laibsch | Yes, I know, but |
| 22:57.04 | ofauchon | When does package parsing occurs ? not sure to understand the goal of packages (okok, i'm RTFM :-) |
| 22:57.16 | Laibsch | I'd rather get a feeling for how things work before patching all over the place |
| 22:58.31 | reenoo_ | Laibsch: sound like a chicken--egg situation to me |
| 22:58.49 | reenoo_ | but yeah, the decision is over to you |
| 22:59.18 | Laibsch | ofauchon: AFAIK, the answer is not really documented anywhere. 500MB appears to be a lot and some CS people |
| 22:59.18 | Laibsch | told about this could not imagine how this comes about either. |
| 22:59.23 | *** join/#oe jamie_ (~jamie@softmodem.org) |
| 22:59.38 | Laibsch | reenoo_: Why chicken-egg? |
| 22:59.57 | Laibsch | What is the chicken, what the egg? |
| 23:00.09 | reenoo_ | heh |
| 23:00.36 | reenoo_ | well |
| 23:01.10 | reenoo_ | it's hard to become familiar with bitbake/oe if your system chokes due to the memory usage |
| 23:01.49 | Laibsch | I hope the 512MB I have now shall suffice. |
| 23:02.10 | Laibsch | I hope do as well for ofauchon. |
| 23:02.14 | woglinde | laibsch I hope you have a lot of swap |
| 23:02.21 | kergoth | or use bbcow |
| 23:02.24 | kergoth | ~bbcow |
| 23:02.25 | ibot | i heard bbcow is a couple patches for bitbake and openembedded to use copy-on-write for the metadata handling. It greatly reduces memory consumption. Get them at http://www.frankengul.org/~seb/cowbb/. |
| 23:02.26 | Laibsch | I hope they do as well for ofauchon. |
| 23:02.31 | woglinde | kergoth *g* |
| 23:02.56 | ofauchon | :-) |
| 23:02.57 | Laibsch | Yes, swap is not a prob. (other than speed, of course ;-)) |
| 23:03.42 | Laibsch | I do not want to build the whole thing, just a single package or two and its dependencies. |
| 23:04.03 | kergoth | it cant know what the dependencies are for that package without parsing all the .bb files that exist |
| 23:04.08 | kergoth | due to PROVIDES |
| 23:04.36 | woglinde | kergoth are both patches needed, for bb and for classes? |
| 23:04.42 | kergoth | woglinde: yes |
| 23:05.08 | woglinde | kergoth okay |
| 23:06.14 | ofauchon | time to have a beer while reading this lonnng GettingStarted |
| 23:06.35 | Laibsch | I tried compilation and it failed with the first packge quilt :( |
| 23:06.51 | Laibsch | The step that failed is do_configure |
| 23:07.15 | Laibsch | it says that my compiler cannot create executables. |
| 23:07.19 | jamie_ | Oh. Someone mentioned 512MB. I was about to try OE but I have only 192MB of memory on the development box (it's a laptop). Too little? |
| 23:07.39 | Laibsch | AFAIK that is a very generic and error message. |
| 23:07.53 | CosmicPenguin | jamie_: yeah, thats not going to even come close |
| 23:07.53 | kergoth | Laibsch: your build machine has to have a usable compiler. |
| 23:07.56 | Laibsch | I looked through the necessary SW and AFAIK |
| 23:07.58 | koen | during the build I'm running now pythin uses 55.9% of my RAM (1GB) |
| 23:07.58 | kergoth | Laibsch: go read RequiredSoftware. |
| 23:08.03 | Laibsch | it is on the box. |
| 23:08.08 | kergoth | then it doesnt work |
| 23:08.13 | koen | that's only python, not the actual build |
| 23:08.18 | kergoth | quilt-native is built using your build machine's compiler |
| 23:08.22 | kergoth | nothing crosscompiler related |
| 23:08.29 | ofauchon | i guess http://treke.net/oe/snapshots/openembedded-exported.tar.bz2 is the latest snapshot ? |
| 23:08.48 | Laibsch | I looked through RequiredSoftware. |
| 23:09.20 | treke | ... |
| 23:09.24 | kergoth | read the config.log in tmp/work/quilt-../ |
| 23:09.25 | treke | It's 4 minutes old |
| 23:09.30 | treke | how much newer do you want? |
| 23:09.37 | koen | treke: 1 minute |
| 23:09.40 | jamie_ | Ew. Even gentoo's portage doesn't come close to needing that much. Why so much RAM and is it something easy to fix that nobody's done due to lack of need? |
| 23:09.59 | kergoth | jamie_: because the code sucks. |
| 23:10.06 | kergoth | jamie_: a rewrite is currently underway |
| 23:10.21 | kergoth | can use help on that, if you know C and are willing |
| 23:10.34 | jamie_ | Ah. I recall one of OE's main feature - perhaps an aspiration :) - was the infinite superiority over portage's sucking code. |
| 23:10.37 | woglinde | kergoth in c? |
| 23:10.49 | kergoth | jamie_: its more abut featureset and flexibility than code |
| 23:10.54 | CosmicPenguin | yeah, right know kergoth is written in ruby |
| 23:11.01 | CosmicPenguin | s/know/now/ |
| 23:11.16 | kergoth | jamie_: i know a number of the portage devs that really like some of our decisions. |
| 23:11.25 | kergoth | just the datastore and parsing code isnt very efficient |
| 23:11.27 | jamie_ | I know C well and if I don't find a decent embedded kit that works for me soon I'm likely to "roll my own" like so many before me. |
| 23:11.34 | jamie_ | So helping is a perhaps. |
| 23:11.57 | kergoth | i have a partially complete design code for the new bitbake backend. marc singer wrote a C parser for the existing .bb format |
| 23:12.02 | kergoth | er |
| 23:12.04 | kergoth | design doc |
| 23:12.06 | kergoth | damn fingers |
| 23:12.40 | jamie_ | Isn't the .bb a shell-script like .ebuilds? Or is that something else? I haven't actually dived into OE yet - what with being warned my laptop isn't big enough and such... |
| 23:12.41 | kergoth | anyway, should be able to use his existing parser to get the new bitbake working with the existing metadata |
| 23:12.44 | kergoth | nope, its directly parsed |
| 23:12.48 | kergoth | whichi s one of our advantages over portage |
| 23:12.59 | kergoth | gives us a great deal of control over the data |
| 23:13.02 | kergoth | and flexibility |
| 23:13.28 | jamie_ | and hopefully, no unknown and mysterious environment variables that break things if they're unexpectedly set.... |
| 23:13.50 | kergoth | hehe. OE does inherit the environment, but vars set in oe .conf files override the environment |
| 23:13.51 | jamie_ | Btw, I was discussing cross-compiling over at gentoo-embedded last year. |
| 23:13.55 | kergoth | (not unlike gnu make, in that regard) |
| 23:15.10 | jamie_ | An idea I had was to modify the compiler to generate "cross-marked" object files, modify the linker to accept them and propagate cross-marks to executables/shlibs, and have the installer check and strip the cross-marks. |
| 23:15.41 | kergoth | what's the benefit of that? |
| 23:15.46 | jamie_ | The purpose of this is to verify, at compile time, that cross-compiling for an x86 target on an x86 host (for example) really does only use cross tools. |
| 23:15.59 | jamie_ | It's a way of making sure all the ebuilds are cross-build safe. |
| 23:16.12 | CosmicPenguin | Seems like a lot of work for validation |
| 23:16.37 | kergoth | i'd rather sandbox it and record configure's attempts to touch areas outside of it, or something |
| 23:16.43 | jamie_ | Well a problem is that lots of gentoo maintainers don't test their ebuilds with cross compilers. So lots of ebuilds don't actually cross compile properly when you want them to. |
| 23:16.49 | treke | or run the build on a ppc :) |
| 23:17.15 | jamie_ | treke: Not a practical way to inform maintainers who aren't doing cross-development but are the best people to keep a package up to date. |
| 23:17.44 | kergoth | jamie_: most of it is just flaws in the buildsystems of the packages. for oe, we end up patching nearly every package for that |
| 23:18.00 | CosmicPenguin | Which is something thats always going to have to happen |
| 23:18.33 | jamie_ | But less so, if a build will fail for non-cross-building maintainers when they do something silly like build and run an executable in the build process using $CC |
| 23:19.11 | jamie_ | kergoth: I agree it's flaws in buildsystems. In my recent experience, buildsystems is _big_ maintenance burden. |
| 23:19.15 | kergoth | i've said it before and i'll say it again. most developers make _really_ bad build engineers |
| 23:19.46 | reenoo_ | I'm off to bed. 'night all |
| 23:19.49 | kergoth | night reenoo_ |
| 23:19.49 | jamie_ | Yes, but they exist. |
| 23:19.52 | jamie_ | 'night. |
| 23:19.52 | CosmicPenguin | The packages themselves are more often then not broken, thats something that the .ebuild maintainers will have to fix, regardless if their cross-compiler is generatingthe right code |
| 23:20.45 | treke | look at how heavilly kergoth had to hack at libtool and autoconf to keep them from including native paths |
| 23:21.22 | jamie_ | kergoth might have not needed to do that if they failed to build for the native maintainers due to using those paths... |
| 23:21.24 | kergoth | i dont see this being something you can really easily fix by modifying your tools. you need to fix the assumptions in the minds of the people doing the buildsystems upstream |
| 23:21.46 | jamie_ | I think that's impossible if < 1% of their users are doing cross-dev. |
| 23:21.55 | kergoth | jamie_: nod, i see the point, immediate failure rather than waiting to see if a decision it made based on something in /usr has unintended affects |
| 23:21.56 | jamie_ | There just isn't the energy. |
| 23:22.42 | CosmicPenguin | if the .ebuild maintainers don't have the energy to consider cross compiling, then its clear that the gentoo repository probably isn't the best to base cross compiled builds on |
| 23:22.46 | jamie_ | Precisely. Even doing a cross-build with gentoo, I found a couple of packages inserting Pentium Pro "cmov" instructions when I was targetting a CPU that couldn't run those, for example. That took a week to discover. |
| 23:22.52 | kergoth | CosmicPenguin: well said |
| 23:23.04 | jamie_ | That's true. |
| 23:23.34 | jamie_ | But I have to wonder if there is the energy in OE or elsewhere to manage the more obscure packages that are sometimes wanted for an embedded project. |
| 23:24.03 | jamie_ | It seems a shame to waste the native maintainers' efforts on those, if there is a way to harness it without demanding anything from them. |
| 23:24.11 | kergoth | well, its driven by need. when someone starts using it that needs it, they add it to oe, and leverage the buildsystem skill amongst the OE team to make it happen |
| 23:24.36 | kergoth | naturally we need to do a better job of pushing our patches upstream, but.. |
| 23:24.49 | jamie_ | You say that like it's a quick thing. |
| 23:25.11 | CosmicPenguin | Its getting quicker and quicker |
| 23:25.13 | kergoth | fixing the package to crosscompile? it is, generally speaking. we've seen 95% of the common problems |
| 23:25.26 | kergoth | there's the occasional oddball, but even those get dealt with fairly quickly |
| 23:25.41 | jamie_ | And when the package isn't cross-compiling properly, but the only indication is a SIGSEGV two weeks later...? |
| 23:25.41 | kergoth | as cp says, it gets faster as the package maintainers gain more skill with such things |
| 23:25.51 | jamie_ | (I.e. -march= options being critical..) |
| 23:26.02 | kergoth | OE exerts control over cflags and ldflags. |
| 23:26.08 | jamie_ | In theory. |
| 23:26.09 | kergoth | when the package gets added, we ensure that our variables are obeyed. |
| 23:26.14 | kergoth | if they arent, the package is broken, period |
| 23:26.16 | kergoth | for more reasons than one |
| 23:27.04 | *** join/#oe nninu (~souldadgu@MTL-ppp-154540.qc.sympatico.ca) |
| 23:27.11 | nninu | hi |
| 23:27.19 | kergoth | hm, need to pick someone up from the airport shortly. |
| 23:27.31 | woglinde | bye kergoth |
| 23:27.33 | jamie_ | byebye |
| 23:27.38 | nninu | i need precision about the zaurus c700 model |
| 23:27.46 | *** join/#oe J_ (~john@host-83-146-11-165.bulldogdsl.com) |
| 23:27.49 | Laibsch | Just did a dry-run on the bbcow patches. They both fail patch a has two rejects in make.py. patch b |
| 23:27.49 | Laibsch | has four rejects in data.py. I assume that is reason to worry and not apply the patch? |
| 23:27.50 | treke | nninu: ? |
| 23:28.09 | nninu | about the bug on it processor |
| 23:28.30 | nninu | is the c700 still a good pda for buy |
| 23:28.35 | treke | No. |
| 23:28.40 | treke | next question? |
| 23:28.48 | nninu | why |
| 23:28.49 | Laibsch | Mine ;-) |
| 23:29.00 | Laibsch | Darn, nninu beat me :-) |
| 23:29.08 | woglinde | buy a dell axim |
| 23:29.10 | nninu | ?? |
| 23:29.17 | treke | It uses a pxa250, it has the smaller battery, and it has too little ram. |
| 23:30.20 | ofauchon | Now, i prepared all, should I bitbake opie-image ? |
| 23:30.39 | woglinde | ofauchon go for it |
| 23:30.40 | nninu | i got one for 250 whit wifi card and 256 sd |
| 23:32.02 | nninu | is not it a good deal |
| 23:32.13 | treke | The price is fair. |
| 23:32.37 | ofauchon | connect plus cards . |
| 23:32.38 | ofauchon | ? |
| 23:32.56 | treke | That's a completely seperate issue from whether or not I think it's a device worth buying though |
| 23:33.28 | *** join/#oe bluelightning (~bluelight@203-173-244-97.akl.ihugultra.co.nz) |
| 23:34.02 | nninu | so should i take it ?? |
| 23:34.17 | nninu | or it is better to serch for another model |
| 23:36.02 | woglinde | lol trolltech cannt set the links right ont their website |
| 23:36.15 | woglinde | http://www.trolltech.com/newsroom/announcements/www.samsung.com |
| 23:36.24 | woglinde | http://www.trolltech.com/newsroom/announcements/www.emuzed.com |
| 23:37.13 | SirFred | ERROR: foo.o uses hardware FP, whereas libfoo.so.0.0.0 uses software FP |
| 23:37.22 | SirFred | How can I avoid this? |
| 23:37.39 | treke | SirFred: don't install software that didn't come from the openzaurus feeds. |
| 23:38.18 | SirFred | treke: You mean the toolchain? |
| 23:38.25 | treke | or in your case, don't try building something for openzaurus with a non soft-float toolchain. |
| 23:38.59 | SirFred | treke: I'm using the toolchain generated by oe |
| 23:39.08 | *** join/#oe Pendalar (Pendalar@h232.124.140.67.ip.alltel.net) |
| 23:39.13 | *** join/#oe anomaly_ (~pop@dsl-202-173-142-212.sa.westnet.com.au) |
| 23:39.21 | SirFred | treke: It seems that the files produced by the assembler are hard-float. |
| 23:39.23 | Pendalar | I am so f'kin pissed at some admins here =-/ |
| 23:39.54 | treke | SirFred: If it generated a soft-float toolchain, gcc is going to build everything for soft-float unless you change mid build. |
| 23:40.03 | nninu | or it is better to serch for another model |
| 23:40.17 | treke | unless you are building something written in asm |
| 23:40.25 | SirFred | treke: It is arm-linux-as the one generating ... |
| 23:40.28 | SirFred | treke: That's the point. |
| 23:40.48 | SirFred | treke: Is there any way to change this behaviour? |
| 23:40.49 | treke | then the code simply wont build for soft-float |
| 23:40.53 | treke | no |
| 23:40.59 | ofauchon | could someone briefly explain me the line : |
| 23:41.01 | ofauchon | BBFILES := "${@bb.fatal('Edit your conf/local.conf: BBFILES')}" |
| 23:41.17 | woglinde | sirfred did you change some of the machine.conf files? |
| 23:41.34 | treke | ofauchon: It means you need to read the conf/local.conf file fully |
| 23:41.41 | SirFred | woglinde: No. I'm just using the toolchain generated by oe, for compiling some assembler stuff. |
| 23:41.51 | treke | SirFred: You have to fix your code |
| 23:42.02 | treke | SirFred: It's going to either be written for hard float or soft float |
| 23:42.18 | SirFred | treke: But openzaurus must be soft float, isn't it? |
| 23:42.27 | treke | SirFred: yes. OpenZaurus uses soft-float |
| 23:42.35 | ofauchon | treke, :-) |
| 23:42.56 | treke | ofauchon: and follow the directions in the comments :) |
| 23:43.00 | SirFred | treke: So, there's no way to integrate some assembler on openzaurus? |
| 23:43.15 | SirFred | Because arm-linux-as never will generate soft-float? |
| 23:43.51 | treke | SirFred: arm-linux-as doesn't know anything about soft-float or hard-float. It just assembles whatever you tell it to assemble |
| 23:44.11 | SirFred | treke: So, why is that hard-float? |
| 23:44.22 | ofauchon | treke, is it _realy_ important to understand bitbake, or bitbake is only used for the toolchain compilation ? |
| 23:44.38 | treke | SirFred: Because that's what your code was written to use. |
| 23:44.59 | *** join/#oe aloph1 (~aloph@cnet-cable-160-107-160.canbrasnet.com.br) |
| 23:45.13 | SirFred | treke: What's supposed to be "hard-float" ? That I'm supposed to use FPU code? |
| 23:45.15 | treke | It could either have been written for soft-float or hard-float. If it was written for soft-float it would work on OZ, but not on something that was built using hard float. |
| 23:45.45 | SirFred | treke: I don't understand how an assembler code could be written for hard-float or soft-float. That's my problem. |
| 23:45.55 | woglinde | ofauchon bitbake is like make and makes everything |
| 23:45.57 | treke | SirFred: You'd be using the hardwares floating point capabilities ( or linux's emulation) |
| 23:46.06 | ofauchon | i see |
| 23:46.17 | SirFred | treke: But,... this ARM processors don't have a FPU. |
| 23:46.25 | SirFred | treke: As far as I know. |
| 23:46.27 | treke | SirFred: read the last part of my statement |
| 23:46.33 | treke | "or linux's emulation" |
| 23:46.49 | treke | There are differences in how arguments are passed and returned to functions. |
| 23:47.09 | SirFred | treke: So, perhaps the difference is some assembler directive or something so? |
| 23:48.40 | treke | there is probabe some documentation out there for this |
| 23:48.56 | Laibsch | ~bbcow |
| 23:48.57 | ibot | it has been said that bbcow is a couple patches for bitbake and openembedded to use copy-on-write for the metadata handling. It greatly reduces memory consumption. Get them at http://www.frankengul.org/~seb/cowbb/. |
| 23:49.07 | Laibsch | I think those patches are out of date. |
| 23:49.14 | SirFred | treke: I'm going to look at the as manual. Thanks. |
| 23:49.17 | Laibsch | A dry run gave rejects for both |
| 23:49.47 | treke | look at the rejects. See if they are easy to merge in by hand |
| 23:49.57 | koen | aaaargh |
| 23:50.07 | koen | 5.0 already |
| 23:50.17 | treke | they might be obvious, if not you'll probably need to wait for someone to update the patches |
| 23:50.49 | Laibsch | Is the maintainer of the cowbb involved with OE at all? I assume yes, so it would be worthwhile to make him aware of necessary changes. |
| 23:51.20 | treke | email him |
| 23:53.41 | CIA-2 | 03koen 07 * r1.3199.1.1 10openembedded/packages/rxvt-unicode/rxvt-unicode_4.9.bb: |
| 23:53.41 | CIA-2 | rxvt-unicode_4.9.bb: |
| 23:53.41 | CIA-2 | <PROTECTED> |
| 23:54.29 | Laibsch | Is Holger Schurig the maintainer or seb@calypso.mine.frankengul.org |
| 23:54.32 | Laibsch | ? |
| 23:54.42 | CosmicPenguin | Those crazy rxvt-unicode guys |
| 23:54.50 | woglinde | no schurig ist the maintainer of patcher |
| 23:54.52 | CosmicPenguin | I don't know if they or udev is worse |
| 23:55.27 | Laibsch | woglinde: Thanx. |
| 23:55.47 | ofauchon | Still don't understant what should i set for that ? BBFILES = "/path/to/openembedded/packages/*/*.bb" |
| 23:55.53 | Laibsch | I will mail seb@ |
| 23:56.13 | Laibsch | ofauchon: Where did you install bitbake etc.? |
| 23:56.31 | SirFred | treke: -mfpu=softfpa fixed it. Thanks. |
| 23:56.33 | ofauchon | /usr/local |
| 23:56.37 | woglinde | laibscg not bitbake |
| 23:56.43 | treke | did it? |
| 23:56.46 | ofauchon | ./setup.py install --prefix=/usr/local --install-data=/usr/local/share |
| 23:56.53 | Laibsch | ofauchon: Did you use bitkeeper? |
| 23:56.53 | treke | looks like mithro did *way* more work than necessary then |
| 23:57.11 | ofauchon | sources are /ZAURUS_DEV/bitbake_src |
| 23:57.20 | woglinde | ofachoun its the path where ypu openembedded dir lies |
| 23:57.34 | ofauchon | <PROTECTED> |
| 23:57.57 | woglinde | ofachoun than /ZAURUS_DEV/openembedded/packages/*/*.bb |
| 23:59.06 | ofauchon | # Delete the line below. Then specify which .bb files to consider for |
| 23:59.06 | ofauchon | # your build. Typically this will be something like BBFILES = "/path/to/openembedded/packages/*/*.bb |
| 23:59.17 | ofauchon | woglinde, that's not easy to understand |
| 23:59.25 | woglinde | ofachoun hm |
| 23:59.34 | ofauchon | packages don't exist |
| 23:59.41 | ofauchon | i guess it'll be created ? |
| 23:59.51 | woglinde | ofachoun didnt you use bitkeeper to check out oe? |
