| 00:01.27 | *** join/#openembedded devnull_cat (~devnull_c@pcp01823855pcs.howard01.md.comcast.net) |
| 00:03.48 | gb2 | WARNING: autoconf execution failed. |
| 00:03.53 | gb2 | NOTE: package gcc-initial-uclibc-cross-3.3.2-r0: task do_configure succeeded |
| 00:12.51 | kergoth | gb2: autotools execution doesnt abort the build, generally. |
| 00:13.20 | kergoth | itd probably be a good idea to make aclocal failure or autoconf failure abort it, but not autoheader/automake, since they might not be used |
| 00:15.11 | gb2 | it looks like autotools now does the wrong thing attempting to build gcc |
| 00:15.27 | kergoth | dont run them against binutils, gcc, or glibc. itll break the build |
| 00:15.36 | gb2 | heh |
| 00:15.38 | kergoth | you'll note i made do_configure() for binutils just call oe_runconf |
| 00:15.39 | kergoth | heh |
| 00:15.39 | gb2 | i used to use those :P |
| 00:15.47 | gb2 | ok |
| 00:16.02 | kergoth | we dont have the necessary m4 macros to do the autotools run |
| 00:16.06 | chouimat | hmmm need to upgrade my network tomorrow ... |
| 00:16.34 | gb2 | do_configure () { ... stuff ... ; oe_runconf } |
| 00:16.35 | gb2 | ? |
| 00:16.42 | gb2 | i used to do a do_configure_prepend |
| 00:16.46 | kergoth | yep |
| 00:18.18 | W8TVI | just dont mix it up with do_configure_pretend |
| 00:18.37 | gb2 | hehe |
| 00:18.59 | gb2 | kergoth: + still can't be used in a package name, right? |
| 00:19.24 | kergoth | it is.. |
| 00:19.33 | gb2 | odd |
| 00:19.54 | kergoth | the annoying thing about what chars we use in variables is |
| 00:19.57 | gb2 | i tried to do a FILES_uclibc-libstdc++ and it did bad things |
| 00:19.58 | kergoth | shell is less tolerant |
| 00:20.13 | kergoth | yep |
| 00:20.21 | kergoth | shell cant deal with a lot of chars in variable names |
| 00:20.31 | kergoth | so when it exports it, our run script goes bleh |
| 00:20.44 | gb2 | neat, gcc built |
| 00:20.51 | kergoth | nice |
| 00:22.40 | gb2 | and my current uclibc built.. now it's building the full cross compiler |
| 00:22.53 | gb2 | oh.. my loop turned out to be the typical multiple providers problem :| |
| 00:23.20 | kergoth | ahh yeah |
| 00:23.22 | kergoth | hate that |
| 00:24.07 | gb2 | i changed the dep so that it was headers-initial instead of the initial libc just being "headers" and the full one being libc and headers |
| 00:24.15 | gb2 | lame, but it works |
| 00:25.35 | kergoth | ahhh yeah |
| 00:26.32 | gb2 | i'm also running oemake on a subset right now.. i figured the multiple versions of gcc we currently have might have been causing problems |
| 00:27.21 | gb2 | now.. it's going to cross compile the gcc for the native system |
| 00:27.25 | gb2 | :) |
| 00:27.27 | gb2 | it's so smart |
| 00:27.41 | ljp | thats inconceivable! |
| 00:27.46 | gb2 | too bad that failed |
| 00:27.59 | kergoth | heh, i did that |
| 00:28.14 | kergoth | crosscompiled a gcc3 chain with my 296 to use to build the cross 3.3.2 |
| 00:28.21 | kergoth | that was a pain in the ass, let me tell you |
| 00:28.52 | gb2 | + export 'CPP=i386-linux-gcc -E' |
| 00:28.53 | gb2 | + CPP=i386-linux-gcc -E |
| 00:28.57 | gb2 | wonder what told it to do that |
| 00:29.15 | kergoth | dunno, default BUILD_PREFIX should be "" |
| 00:30.53 | gb2 | i'm cross compiling |
| 00:31.15 | gb2 | BUILD_PREFIX should be ${BUILD_SYS}- |
| 00:31.15 | gb2 | right? |
| 00:31.19 | gb2 | s/BUILD/TARGET? |
| 00:31.20 | gb2 | rather |
| 00:31.40 | kergoth | yep |
| 00:31.51 | gb2 | hm |
| 00:32.05 | gb2 | TARGET_ARCH = i386 |
| 00:32.05 | gb2 | TARGET_OS = linux |
| 00:32.05 | gb2 | TARGET_SYS = ${TARGET_ARCH}-uclibc-${TARGET_OS} |
| 00:32.06 | gb2 | i'm doing that |
| 00:32.20 | gb2 | hm |
| 00:33.22 | kergoth | showdata, look around |
| 00:33.34 | gb2 | what's the correct incant to do showdata again? |
| 00:33.39 | gb2 | i can never get it quite right |
| 00:33.42 | kergoth | its just a task |
| 00:33.44 | dan2003 | i unpack some source on my x86 desktop, do i need only export paths to make it build for arm or is there more to it? ( i have the toolchains) |
| 00:33.46 | kergoth | oebuild showdata /path/to/.oe |
| 00:34.08 | kergoth | dan2003: heh. theres a shitload more to it |
| 00:34.40 | icefox | http://kaylee.meyerhome.net:83/~ben/ben/articles/linux_pda.php <- a top 100 wishlist for a linuxpda |
| 00:34.49 | icefox | An article I am putting together |
| 00:35.16 | gb2 | HOST_PREFIX="i386-linux-" |
| 00:35.29 | dan2003 | i thought so! im a bit confused as to how the buildroot works, it seems to just grab a package make a few patches and build, |
| 00:35.36 | kergoth | gb2: whats BUILD_PREFIX? |
| 00:35.48 | kergoth | gb2: the native/cross classes should be setting HOST_PREFIX to BUILD_PREFIX when it does the switch |
| 00:35.49 | gb2 | "" |
| 00:35.58 | kergoth | double check the classes |
| 00:36.03 | gb2 | this isn't cross/native |
| 00:36.07 | gb2 | this is normal |
| 00:36.18 | kergoth | heh, whats TARGET_PREFIX then? |
| 00:36.21 | gb2 | so HOST_PREFIX should be TARGET_PREFIX |
| 00:36.30 | gb2 | TARGET_PREFIX="i386-uclibc-linux-" |
| 00:36.40 | kergoth | ah |
| 00:36.42 | kergoth | i know what it is |
| 00:37.01 | kergoth | HOST_PREFIX isnt TARGET_PREFIX, its HOST_SYS-, which in turn is HOST_ARCH, which is TARGET_ARCH |
| 00:37.05 | kergoth | its happening at the low level, not the high |
| 00:37.12 | kergoth | set TARGET_VENDOR to uclibc- |
| 00:37.16 | kergoth | instead of setting SYS directly |
| 00:37.20 | gb2 | heh |
| 00:37.26 | kergoth | either that, or set HOST_SYS as well |
| 00:37.26 | gb2 | right |
| 00:37.38 | gb2 | setting host sys sounds like a recipe for breakage |
| 00:37.38 | kergoth | vendor's best |
| 00:37.53 | gb2 | - goes at the end? |
| 00:37.53 | kergoth | yep |
| 00:37.58 | kergoth | iirc yeah |
| 00:38.00 | kergoth | can doublecheck oe.conf |
| 00:38.11 | gb2 | yeah |
| 00:38.23 | gb2 | # NOTE: _VENDOR fields must be either empty or start with a - |
| 00:39.39 | dan2003 | what version of util-linuz should i use, the util-linux_2.11z.orig.tar.gz is gone from source loaction there an util-linux_2.11n.orig.tar.gz or a 2.12 |
| 00:40.12 | kergoth | pick one. |
| 00:43.43 | gb2 | hrm, now my build didn't get as far as before |
| 00:43.56 | *** join/#openembedded cyn (~cyn@adsl-155-129-59.mem.bellsouth.net) |
| 00:44.23 | *** join/#openembedded Shdwdrgn (Picard@sourpuss.net) |
| 00:45.02 | gb2 | crap |
| 00:45.06 | gb2 | i've seen this one before |
| 00:45.30 | *** part/#openembedded devnull_cat (~devnull_c@pcp01823855pcs.howard01.md.comcast.net) |
| 00:46.13 | W8TVI | gb2, same shit, different day? |
| 00:46.29 | gb2 | yeah |
| 00:47.38 | gb2 | checking host system type... i686-uclibc-linux-gnu |
| 00:47.52 | gb2 | i guess |
| 00:47.59 | *** join/#openembedded TMM1 (~aman@000-358-167.area3.spcsdns.net) |
| 00:48.07 | gb2 | urm |
| 00:48.07 | gb2 | wait |
| 00:48.20 | gb2 | that looks like BUILD_VENDOR got polluted |
| 00:49.09 | *** mode/#openembedded [+o TheMasterMind1] by ChanServ |
| 00:49.12 | *** topic/#openembedded by TheMasterMind1 -> OpenEmbedded/OpenZaurus developer lounge | User support in #openzaurus, #openembedded-ipaq, and (if applicable) #opie | Snapshots available at http://treke.net/~bk/ | Python stuff also | http://www.hackles.org/cgi-bin/archives.pl?request=2 |
| 00:50.36 | gb2 | urm. |
| 00:51.42 | gb2 | crap |
| 00:51.48 | gb2 | BUILD_SYS="i686-linux" |
| 00:51.56 | gb2 | HOST_SYS="i686-uclibc-linux" |
| 00:51.58 | gb2 | that's not right |
| 00:52.27 | gb2 | TARGET_VENDOR="-uclibc" |
| 00:52.33 | gb2 | BUILD_VENDOR="" |
| 00:52.38 | gb2 | HOST_VENDOR="-uclibc" |
| 00:52.47 | gb2 | HOST_VENDOR got set to the wrong one |
| 00:53.28 | gb2 | heh |
| 00:53.35 | gb2 | cross.oeclass is broken |
| 00:53.49 | kergoth | heh |
| 00:53.52 | kergoth | it doesnt touch vendor, does it? |
| 00:54.02 | gb2 | right |
| 00:55.23 | gb2 | still touches CROSS too |
| 00:55.29 | *** part/#openembedded icefox (~ben@ool-18b87f53.dyn.optonline.net) |
| 01:00.04 | gb2 | adding the vendor one fixed that bug btw |
| 01:01.17 | kergoth | cool |
| 01:01.51 | *** join/#openembedded maxik (~maxik@217.66.146.252) |
| 01:03.11 | gb2 | i'm going to have to remmeber how to commit changes again soon |
| 01:04.42 | *** join/#openembedded maxik (~maxik@217.66.146.252) |
| 01:08.30 | *** join/#openembedded maxik (~maxik@217.66.146.252) |
| 01:12.55 | *** join/#openembedded darkschneider (~gab@62-101-126-192.fastres.net) |
| 01:15.44 | gb2 | NOTE: build 200311291955 completed |
| 01:15.46 | gb2 | :) |
| 01:18.11 | darkschneider | gb2 help |
| 01:18.34 | gb2 | ? |
| 01:18.35 | darkschneider | mmm ops... gb2 is not the bot |
| 01:18.37 | gb2 | telling me about bk help? |
| 01:18.39 | darkschneider | sorry... |
| 01:18.47 | darkschneider | ehm... |
| 01:18.53 | darkschneider | :)) |
| 01:18.53 | gb2 | i am NOT a bot :P |
| 01:19.07 | darkschneider | eheh yeah indeed, or if you are you are the best ever made :) |
| 01:19.14 | ljp | hehehe |
| 01:19.29 | gb2 | not sure if that's a compliment or not |
| 01:19.57 | darkschneider | btw just passign by to say microdrive is quite sweet |
| 01:20.16 | darkschneider | if i only could watch movies on my Z :( |
| 01:21.01 | kergoth | why cant you? |
| 01:21.29 | darkschneider | kergoth you knwo notthing much working on the 5600 |
| 01:21.44 | kergoth | huh? |
| 01:21.50 | darkschneider | many app crash, expecialy thsoe dealing with the fb device |
| 01:21.59 | kergoth | um |
| 01:22.12 | kergoth | funny, i've used op2 on mine |
| 01:22.23 | darkschneider | like image viewers (crash with big images, or after fome small) |
| 01:22.34 | darkschneider | opie worked with the experimental 1st version rom |
| 01:22.46 | darkschneider | but stopped working when i tested the 2nd (and last) version |
| 01:22.58 | kergoth | you tested 3.3.5? |
| 01:23.00 | kergoth | the release? |
| 01:23.02 | darkschneider | anyway in that version audio stated working |
| 01:23.03 | darkschneider | nope |
| 01:23.07 | darkschneider | goign to test it now |
| 01:23.13 | darkschneider | that's why ic onencted :) |
| 01:23.19 | kergoth | cool |
| 01:23.26 | treke|home | guh. My build died hours ago |
| 01:23.35 | darkschneider | treke :( |
| 01:24.59 | gb2 | kergoth: i'm fixing the cross class and making uclibc compile again |
| 01:25.19 | darkschneider | mmm do you need a mirror for images= |
| 01:25.43 | darkschneider | i noriced images have no mirrors, i can provvide a file only mirror.. i have so much unused space at diniciacci.org |
| 01:26.33 | pokryfka | is there any work being done on porting gpe? |
| 01:26.49 | darkschneider | ok can i have a link to the files to 3.3.5 so i try to get a working image of new version for poodle |
| 01:27.16 | CIA | repository=03oe user=11gbritton@doomcom.org |
| 01:27.17 | CIA | 07ChangeSet@1.315 fix vendor for cross class |
| 01:27.20 | mickeyl | no idea about that "working", but you may try @ oz.org/official/unstable |
| 01:27.37 | Twiun | mickeyl: whoa! what are you still doing up? ;) |
| 01:27.39 | darkschneider | s/get/compile |
| 01:27.48 | mickeyl | Twiun: relaxing :) |
| 01:28.01 | mickeyl | Twiun: but you're right... it's way over bed time |
| 01:28.37 | mickeyl | was cought in a chat trying to help someone getting wellenreiter to run |
| 01:28.56 | CIA | repository=03packages user=11gbritton@doomcom.org |
| 01:28.56 | CIA | 07ChangeSet@1.343 fixes for cross/prefix build changes |
| 01:28.57 | gb2 | there. |
| 01:29.30 | Twiun | yeah, that happens to me sometimes too. |
| 01:29.39 | Twiun | I'm trying to get Swing running on fbAWT ;) |
| 01:29.47 | mickeyl | quite a job ;) |
| 01:30.07 | Twiun | yeah, but if it works it should be much faster than what Jeode could accomplish |
| 01:30.34 | mickeyl | and afterwards port the backend to QDirectPainter? should be not a big deal. |
| 01:30.36 | treke|home | mickeyl: do you have an ipaq? |
| 01:30.47 | Twiun | it's amazing how much the repainting system shifts if the whole component stuff is lightweight |
| 01:30.51 | mickeyl | treke: yeah. i even wanted to do a build, but i couldn't build the kernel... |
| 01:30.54 | gb2 | lesse how far this gets |
| 01:30.57 | treke|home | mickeyl: fun. |
| 01:31.04 | treke|home | mickeyl: So you can test it if I build? |
| 01:31.13 | mickeyl | treke: yeah |
| 01:31.15 | mickeyl | 3650 here |
| 01:31.16 | treke|home | cool. |
| 01:31.20 | treke|home | I don't have any |
| 01:32.57 | gb2 | sofar.. zlib and ncurses |
| 01:32.57 | mickeyl | it was weird - the kernel make just stopped with an error but i didn't see any. never happened before so i was out of ideas |
| 01:33.05 | gb2 | dropbear |
| 01:33.08 | gb2 | this is sweet |
| 01:35.55 | gb2 | failed because tmake fetch failed |
| 01:36.36 | gb2 | busybox built, but that's to be expected :) |
| 01:37.16 | gb2 | this is pretty cool how so much just builds with uclibc |
| 01:39.58 | *** join/#openembedded rlh (~rlh@p377-tnt10.syd.ihug.com.au) |
| 01:41.21 | darkschneider | grrr i just hate my Z when he does not want to get recognized by the kernel.. grrr |
| 01:44.04 | gb2 | hm. |
| 01:47.05 | gb2 | kergoth: the i686 site file you have.. |
| 01:47.17 | gb2 | could i get that from you? |
| 01:48.05 | kergoth | yeah |
| 01:48.08 | kergoth | i forget to check it in? |
| 01:48.17 | gb2 | i don't think it's in there, no |
| 01:48.24 | kergoth | k |
| 01:48.35 | *** join/#openembedded icefox (~ben@ool-18b87f53.dyn.optonline.net) |
| 01:48.56 | gb2 | btw, did you fix the bug where site configs must be in OEDIR rather than just anywhere in OEPATH? |
| 01:50.15 | darkschneider | any idea on how to convince my Z to publicize himself on the net (not it does that _only_ if i turn it on when it's allready on the docking) |
| 01:52.19 | kergoth | gb2: nah, easy fix though |
| 01:52.21 | kergoth | in oe.conf |
| 01:52.26 | kergoth | export CONFIG_SITE = ${@oe.which(oe.data.getVar('OEPATH', d, 1), "site/%s" % oe.data.getVar('HOST_SYS', d, 1))} |
| 01:52.31 | kergoth | will take care of it |
| 01:52.59 | gb2 | ah |
| 01:52.59 | gb2 | k |
| 01:53.49 | CIA | repository=03oe user=11kergoth@direwolf.ppp.ti.com |
| 01:54.24 | kergoth | k there |
| 01:54.46 | kergoth | hmm |
| 01:54.50 | kergoth | i lost my i686-linux |
| 01:54.54 | kergoth | need to script populating it again |
| 01:58.10 | *** join/#openembedded TheMasterMind1 (~aman@pcp04370982pcs.nrockv01.md.comcast.net) |
| 01:58.12 | *** mode/#openembedded [+o TheMasterMind1] by ChanServ |
| 02:02.35 | chouimat | hmmmmmmmm boring |
| 02:02.40 | gb2 | ok, i'll use the arm one for now.. that ought to mostly be ok for this purpose |
| 02:05.37 | chouimat | night |
| 02:17.54 | gb2 | <PROTECTED> |
| 02:17.55 | gb2 | hm |
| 02:19.51 | treke|home | this is impressive |
| 02:20.07 | treke|home | so far a whole buildroot build has progressed to Opie without much trouble. |
| 02:22.24 | Twiun | pretty nice |
| 02:24.16 | darkschneider | kergoth: can you rename the kernel image file on the site with a uppercase I, just so people notice that (i'm talking of the one in official/unstable/3.3.5/SL5600 ) |
| 02:24.23 | gb2 | this is really weird |
| 02:26.17 | CIA | repository=03packages user=11kergoth@direwolf.ppp.ti.com |
| 02:26.57 | *** join/#openembedded vvff (~victor@ip16.124-173-207.eli-du.nwlink.com) |
| 02:30.12 | gb2 | odd |
| 02:31.10 | *** join/#openembedded mithro (~tim@dsl1-83.gw1.adl1.airnet.com.au) |
| 02:32.12 | gb2 | wonder why this isn't building the c++ compiler |
| 02:39.57 | gb2 | kergoth: why do we have both 3.3.2 and 3.3.1 gcc's in there anyway? |
| 02:40.39 | treke|home | wait. darkscheider is doing the pxa build? |
| 02:40.51 | treke|home | we need to work out who is doing this shit :) |
| 02:41.55 | kergoth | gb2: no reason, really. just in general dont remove metadata unless we need to |
| 02:42.04 | kergoth | course it causes issues right now with the multiple provides problem.. |
| 02:43.04 | gb2 | yeah |
| 02:43.16 | gb2 | i think i'm seeing non-deterministic behavior |
| 02:54.27 | cyn | NP-complete eh |
| 02:56.06 | Twiun | g'night |
| 03:04.06 | gb2 | lesse if this gets further |
| 03:10.38 | gb2 | aha, first real glibc/uclibc bug. |
| 03:11.07 | kergoth | what failed? |
| 03:11.17 | gb2 | rdisc.c:1507: error: `sys_errlist' undeclared (first use in this function) |
| 03:11.18 | gb2 | :) |
| 03:11.21 | gb2 | in iputils |
| 03:11.22 | kergoth | ahh |
| 03:11.30 | kergoth | well thats not surprising |
| 03:11.33 | gb2 | yeah |
| 03:11.39 | gb2 | an easy fix |
| 03:13.14 | gb2 | is there any way to run oemake (ignoring failures) and giving me a log of all the failed packages? |
| 03:13.26 | gb2 | or do i just run it without -a and tee a log |
| 03:13.48 | kergoth | i think tee is about the only option atm |
| 03:13.52 | kergoth | logging still needs rewroking |
| 03:13.54 | kergoth | bleh |
| 03:14.59 | gb2 | a lot of this stuff seems to build great with uclibc |
| 03:17.02 | gb2 | kergoth: does everything in the current packages (not counting nonworking) build for you? |
| 03:17.05 | gb2 | with your glibc build |
| 03:21.21 | kergoth | gb2: yup |
| 03:22.04 | gb2 | kergoth: cool. it's about to try building qt-embedded |
| 03:22.14 | kergoth | nice |
| 03:22.19 | gb2 | sofar, only iputils and libvorbis (vorbis was an failure in do_install) |
| 03:22.24 | *** join/#openembedded bisho (~bisho@136.Red-213-97-191.pooles.rima-tde.net) |
| 03:22.46 | gb2 | kergoth: ... libtool problem |
| 03:23.07 | gb2 | .../i386-uclibc-linux/bin/ld: cannot find /lib/libc.so.6 inside /usr |
| 03:23.21 | gb2 | qt-embedded didn't even configure :) |
| 03:23.57 | gb2 | <PROTECTED> |
| 03:24.00 | gb2 | interesting |
| 03:24.06 | kergoth | heh |
| 03:26.22 | *** join/#openembedded cyn (~cyn@adsl-155-129-59.mem.bellsouth.net) |
| 03:40.38 | *** join/#openembedded cyn (~cyn@adsl-155-129-59.mem.bellsouth.net) |
| 03:48.07 | CIA | repository=03packages user=11gbritton@doomcom.org |
| 03:48.08 | CIA | 07ChangeSet@1.345 staging fixes for uclibc |
| 03:57.58 | treke|home | cool. build done |
| 03:59.04 | gb2 | i should have a hitlist for uclibc fixing in a little while |
| 04:02.17 | *** join/#openembedded Dis-Z (~dis@arent.you.special.you.gotontheinter.net) |
| 04:21.15 | *** join/#openembedded vvff (~victor@207.173.124.16) |
| 04:21.33 | Pendalar | can you extract certain files out of a tar but not all? |
| 04:22.31 | Dis-Z | yep. |
| 04:22.32 | treke|home | yes. |
| 04:22.44 | Pendalar | I thought so but I'm having a 'fun' time doing it =-/ |
| 04:22.53 | Dis-Z | try the man page. |
| 04:22.56 | treke|home | tar xf file.tar filename |
| 04:23.01 | Pendalar | been looking |
| 04:23.44 | Pendalar | I'm wanting more of a path and sub thing... like say it's a full HD backup, would "tar xf file.tar /root/*" get root and down? |
| 04:23.55 | Dis-Z | just root/ |
| 04:23.58 | treke|home | erm. a directory is a file |
| 04:24.04 | Pendalar | true |
| 04:24.04 | treke|home | tar xf file.tar root/ |
| 04:24.05 | Dis-Z | no leading / and no * |
| 04:24.32 | Pendalar | oh heh I was doing tar xf file.tar.gz files |
| 04:24.43 | Pendalar | need to uncompress there first dumbass.. |
| 04:24.46 | Pendalar | =-{ |
| 04:24.54 | Pendalar | or add the parameter to do so |
| 04:26.20 | Pendalar | sorry to bug you guys with that =-/ I was doing the extracting part right but not the right parameters for decompressing |
| 04:37.37 | gb2 | hrm |
| 04:37.44 | gb2 | it's trying to build glibc now |
| 04:37.50 | gb2 | wonder if i should have inhibited that |
| 05:04.44 | *** part/#openembedded icefox (~ben@ool-18b87f53.dyn.optonline.net) |
| 05:09.53 | *** join/#openembedded devnull_cat (~devnull_c@pcp01823855pcs.howard01.md.comcast.net) |
| 05:13.31 | devnull_cat | Hi all. I'd like to set up an OE build environment, and have followed the instructions on the wiki, but apart from patcher everything dies at the do_compile stage. Are others having the same problems, or am I missing something? |
| 05:14.18 | kergoth | didnt add your toolchain bin dir to your path, i'd say. |
| 05:14.53 | kergoth | e.g. export PATH=/usr/local/arm/3.3/bin:$PATH |
| 05:14.54 | kergoth | whatever |
| 05:18.16 | devnull_cat | kergoth: That wasn't in the instructions, but I did figure out that step. The most common error I'm getting is "cannot specify -o with -c or -S and multiple compilations" - while this seems a valid error, it's possible there's a step I'm missing that would avoid the error. |
| 05:18.45 | kergoth | paste your conf/local.conf here |
| 05:19.03 | kergoth | the instructions are for building OE. they arent specific to crosscompilation |
| 05:19.11 | kergoth | i use OE to build native rpms and the like as well |
| 05:21.04 | devnull_cat | kergoth: TARGET_ARCH=arm; DL_DIR=/usr/src/OpenEmbedded/downloads; OEDIR=/usr/src/OpenEmbedded/oe; OEFILES="/usr/src/OpenEmbedded/packages/*/*.oe" |
| 05:21.14 | devnull_cat | kergoth: read the semicolons as newlines |
| 05:22.09 | kergoth | looks fine. oebuild showdata glibc/glibc_2.3.2.oe|grep CC= |
| 05:23.13 | devnull_cat | export BUILD_CC="ccache gcc"; export CC="ccache arm-linux-gcc" |
| 05:23.24 | kergoth | looks fine. what toolchain? |
| 05:23.25 | devnull_cat | kergoth: that doesn't look self-consistent... |
| 05:23.29 | kergoth | hm? |
| 05:23.35 | kergoth | thats exactly what it should be |
| 05:23.58 | devnull_cat | I'm using the cross-3.3.1 toolchain from as referenced in the Wiki |
| 05:24.00 | kergoth | CC should really be HOST_CC, but packages generally expect CC, so thats what we use |
| 05:24.07 | kergoth | so /usr/local/arm/3.3/bin/arm-linux-gcc exists? |
| 05:25.08 | devnull_cat | Yes, it exists - it's all untarred in /usr/local/arm, and in the PATH. |
| 05:25.27 | kergoth | well i dunno what to tell you, i did a full build from scratch across all packages yesterday |
| 05:25.32 | kergoth | rm -rf tmp |
| 05:25.40 | kergoth | then, build patcher first, then autoconf-native, automake-native, libtool-cross |
| 05:25.45 | kergoth | then build glibc |
| 05:26.09 | devnull_cat | I'll try in exactly that order and get back to you... |
| 05:26.14 | kergoth | k |
| 05:26.58 | kergoth | ugh, organizing my mp3s is a pain in the ass |
| 05:27.03 | kergoth | even with mass tagging and mass renaming tools |
| 05:29.55 | treke|home | cool. found a metric assload of pens |
| 05:30.21 | kergoth | nice |
| 05:30.26 | kergoth | i'm always losing them |
| 05:31.25 | Dis-Z | kergoth: poke |
| 05:31.44 | kergoth | heh, this time it wasnt sleep, it was laziness |
| 05:31.57 | treke|home | kergoth: If I lose them at the rate of 1 per day, I'll now be safe through 2004 :) |
| 05:32.46 | Dis-Z | well tomorrow is about all the time I have. work is gonna start really suckong again and so forth. my time is shot, was really hoping to do this thurs....... |
| 05:33.19 | kergoth | alright i'll get it out now then |
| 05:33.28 | Dis-Z | cool thanks. |
| 05:33.53 | treke|home | hehe. kergoth is a lazy bastard, isn't he? :p |
| 05:35.28 | treke|home | shit. SUBDIRS not SUBIRDS |
| 05:36.19 | ljp | and not SUBTURDS either |
| 05:36.19 | Dis-Z | heh |
| 05:44.21 | kergoth | okay if its mapped to a literal in the yacc code, which gets the key set to K(KT_LETTER, KVAL(yyvsp[0])) |
| 05:44.42 | kergoth | so take the example letter 'a', stuff hte ascii value into an int, apply the KVAL macro.. |
| 05:45.45 | kergoth | which is just the lower byte |
| 05:45.49 | ljp | huh? |
| 05:45.56 | kergoth | dont mind me, just rambling about linux keyboard crap |
| 05:46.00 | ljp | :)O |
| 05:46.22 | ljp | do you know where the source for showkeys might be? |
| 05:47.11 | kergoth | console-tools sources |
| 05:47.28 | kergoth | you mean showkey? |
| 05:47.29 | kergoth | heh |
| 05:47.51 | kergoth | or do you mean the busybox ones, which use a binary format, not the standard .map files? |
| 05:48.18 | ljp | I thought it was showkeys |
| 05:48.23 | ljp | hmm |
| 05:49.06 | ljp | heh, maybe thats why I couldnt frickin find it |
| 05:53.10 | devnull_cat | kergoth: I got all the way to the end, then at the very end of glibc I get again "arm-linux-gcc: cannot specify -o with -c or -S and multiple compilations." |
| 05:55.36 | kergoth | devnull_cat: hmm, builds fine here.. send the log to the openembedded-devel list, i'll take a look in a bit |
| 05:56.02 | kergoth | Dis-Z: sent. if its confusing, which it very well may be, let me know. I gave you an example key, showed the old methodology vs what we want. |
| 05:57.59 | kergoth | Dis-Z: oh, an aside. some keys dont translate directly to qt keysyms.. for example, there's a type of keycode you can map which deals with international chars, and theres another type to just dump a string.. guess for now those are best left unhandled.. |
| 05:59.17 | *** join/#openembedded bisho (~bisho@136.Red-213-97-191.pooles.rima-tde.net) |
| 06:08.51 | gb2 | aha |
| 06:08.54 | gb2 | my build finally finished |
| 06:09.03 | kergoth | nice |
| 06:09.09 | kergoth | making a list of the known uclibc failures? |
| 06:09.24 | gb2 | yup |
| 06:09.38 | gb2 | fog% grep ERROR loggish |grep 'build.failed' |wc -l |
| 06:09.38 | gb2 | <PROTECTED> |
| 06:09.53 | gb2 | some of those are stuff like: |
| 06:09.57 | gb2 | ERROR: package base/gcc-initial-uclibc-cross-3.3.1-r0: build failed |
| 06:10.21 | kergoth | hehe |
| 06:10.24 | kergoth | yeah, dupes |
| 06:11.11 | gb2 | fog% grep 'do_build succeeded' loggish|wc -l |
| 06:11.12 | gb2 | <PROTECTED> |
| 06:11.19 | gb2 | more than 50% succeeded :) |
| 06:13.16 | Disconnect | kergoth: thats not the most clear thing I've read but I think all the data is there. |
| 06:13.23 | kergoth | hehe |
| 06:13.36 | Disconnect | well the explanation seems clear, its just not the most clear procedure ;) |
| 06:14.10 | kergoth | for the 5500, we have a .map and a qt keymap that we know go together |
| 06:14.24 | kergoth | so could at least script the letters and possibly a couple others |
| 06:15.06 | Disconnect | the part that confuses me is here: New methodology: Read raw value 01, pass through the kernel's keymaptables |
| 06:15.13 | Disconnect | wouldn't it have gone through those before we see it? |
| 06:15.23 | kergoth | no, qte opens the console in RAW or MEDIUMRAW mode |
| 06:15.29 | kergoth | qt itself needs to see both press and release events |
| 06:15.35 | Disconnect | ah. damn. good point. |
| 06:15.37 | kergoth | which means *qt* has to push it through the kernel keymap |
| 06:15.40 | kergoth | same thing tinyx has to do |
| 06:15.42 | Disconnect | ok. |
| 06:15.50 | kergoth | so we ask the kernel for it, keep it around, and do it ourselves |
| 06:16.11 | Disconnect | so would the easiest thing be to map all -possible- kernel keys, similar to how X does it..? |
| 06:16.34 | kergoth | yeah. the kernel map could map a raw key to *anything* |
| 06:16.37 | Disconnect | (eg tinyx on the ipaq could do F14.. which - IIRC - I had mapped to record at one point for some or anothter purpose) |
| 06:16.47 | kergoth | so we should fully map out the kenrel keycode (K()) to qt keysym relationships |
| 06:16.51 | kergoth | for everything we can |
| 06:16.57 | kergoth | then itll be happy whatever we throw at it |
| 06:16.59 | kergoth | on whatever device |
| 06:17.56 | kergoth | then we make an opie kernel keymap manipulator thatll let us easily remap |
| 06:18.01 | kergoth | :) |
| 06:18.09 | Disconnect | yah but thats gui work for someone who does guis. ;) |
| 06:18.15 | kergoth | well yeah |
| 06:18.19 | devnull_cat | kergoth: Message sent to openembedded-devel, with subject line "arm-linux-gcc error" - will wait for reply on the list. |
| 06:18.57 | Disconnect | I think I understand. at least enough to do the array. |
| 06:19.18 | Disconnect | instead of the array I gave you (keycode -> Qt) you want scancode -> Qt.. one step farther "back" on the chain.. |
| 06:19.36 | Disconnect | and if its automated, all the better ;) |
| 06:19.37 | kergoth | you gave me pre map keycode to qt |
| 06:19.39 | kergoth | not scancode |
| 06:19.44 | *** part/#openembedded devnull_cat (~devnull_c@pcp01823855pcs.howard01.md.comcast.net) |
| 06:19.53 | kergoth | and we want kernel post-map keysym (K() value) -> qt keysym |
| 06:20.10 | Disconnect | ok. pretend I haven't done this before. ;) |
| 06:20.18 | kergoth | okay, lets see |
| 06:20.31 | kergoth | kernel keyboard driver reads a value from the hardware |
| 06:20.35 | kergoth | that value is a scancode |
| 06:21.07 | kergoth | that gets sent through the kernel's kbd_translate() function to produce a "keycode", which is still just an int without an associated type. just a low level keycode |
| 06:21.14 | Disconnect | ok. |
| 06:21.23 | kergoth | (in our example, scancode is irrelevlent, but it gets a raw keycode of 01) |
| 06:21.46 | kergoth | qt then gets the event at that point, since we're in mediumraw |
| 06:22.01 | kergoth | (raw gets scancodes, mediumraw gets keycodes.. in 5500 context they're one and the same) |
| 06:22.24 | kergoth | okay, qt reads the kernel keymaps, which are just an array of this "keycode" value that includes a type. i call this the post-map keycode |
| 06:22.52 | kergoth | kernel_keymap[modifierstate][raw keycode] is the post-map keycode |
| 06:23.08 | kergoth | the *zaurus* keymap maps raw keycode 01 to 'a' |
| 06:23.12 | ljp | you sure about that? |
| 06:23.13 | Disconnect | ah. eg kernel_keymap[shift][01] for A..? |
| 06:23.20 | kergoth | a, not A |
| 06:23.20 | kergoth | but yes |
| 06:23.32 | Disconnect | shift-01 is still 'a'? |
| 06:23.43 | kergoth | oh, didnt notice your 'shift' |
| 06:23.45 | kergoth | yeah, thats correct |
| 06:23.55 | kergoth | kernel_keymap[shift][01] is A |
| 06:24.07 | kergoth | which is shiftstate KG_SHIFT |
| 06:24.10 | kergoth | type KT_LETTER |
| 06:24.22 | kergoth | value K_VAL('a'), which ends up being just 0x61, the ascii value |
| 06:24.47 | kergoth | K([type],[value]) creates a post-map keycode |
| 06:25.02 | kergoth | so, we need to map K(KT_LETTER, K_VAL('a')) to Qt::Key_A |
| 06:25.09 | kergoth | instead of 01 to Qt::Key_A |
| 06:25.28 | kergoth | basically, we want kernel post-map keycode -> qt keysym |
| 06:25.35 | kergoth | instead of the existing scancode -> qt keysym |
| 06:26.13 | Disconnect | ok. and **kernel_keymap is the missing piece..? |
| 06:26.29 | kergoth | no, kernel_keymap is what maps our raw keycode to kernel post-map keycode |
| 06:26.34 | kergoth | here's the path |
| 06:27.03 | kergoth | read raw keycode -> kernel_keymap array as read from the kernel -> kernel post-map keycode -> [our new array] -> qt keysym |
| 06:27.07 | kergoth | previous |
| 06:27.16 | kergoth | read scancode -> old array -> qt keysym |
| 06:27.40 | *** join/#openembedded Grev (~dbasinge@h-68-164-43-14.DNVTCO56.dynamic.covad.net) |
| 06:29.11 | Disconnect | ok. I think I have it now. KQ[K(KT_LETTER,K_VAL('a')]=Qt::Key_A.. hrm. is this gonna get a modifier state (eg a/A) |
| 06:29.19 | Disconnect | (KQ being the new array) |
| 06:29.41 | kergoth | look at the function that submits it upstream, i translate the kernel shiftstate into a qt shiftstate that gets passed up |
| 06:29.51 | kergoth | (hardcoded atm, should be a table of state translations later) |
| 06:30.54 | kergoth | what we do is, watch for a certain keycode. when we see it, we switch keymaps (shiftstates). we store that, and use it later when submitting upstream as an event |
| 06:31.01 | Disconnect | lemme get my sources unpacked again |
| 06:31.12 | kergoth | (note that the keycode -> modifier state change is also hardcoded atm) |
| 06:31.15 | kergoth | but, that neednt stay the case |
| 06:31.23 | kergoth | theres a SHIFT key type |
| 06:31.35 | Grev | kergoth, thanks for 3.3.5, loving it |
| 06:31.39 | kergoth | if we make use of hte kernel's existing mechanisms, we can use its idea of whats mapped to a modifier as well |
| 06:31.54 | kergoth | Grev: no problem. thank mickeyl|away, he did the work getting the release out the door |
| 06:32.04 | kergoth | most of what i'm focused on is the new system, you'll see the effects of that in 3.5.1 :) |
| 06:32.09 | Grev | thanks mickey:) |
| 06:32.20 | Disconnect | (ugh buildroot is squirrely gimme a sec) |
| 06:32.40 | kergoth | np |
| 06:33.44 | W8TVI | kergoth, are these images different than the rc1 images? |
| 06:33.52 | kergoth | ask the person who did the release. |
| 06:33.54 | Disconnect | ah. def-config for minimal doesn't include opie or qte. which blows up the build. fun. so its got a bunch of extra stuff to unpack, but in theory it'll be done in a sec. |
| 06:34.06 | *** join/#openembedded bisho (~bisho@136.Red-213-97-191.pooles.rima-tde.net) |
| 06:34.10 | kergoth | Disconnect: fyi, make SUBDIRS=nodeps_packages/qte sourcetrees; |
| 06:34.12 | kergoth | Disconnect: ;) |
| 06:34.22 | Disconnect | ok. I think I'm understanding how it processes. I'm not understanding which piece is missing. |
| 06:34.25 | Disconnect | yah thats what blew up :( |
| 06:34.29 | kergoth | ah |
| 06:34.30 | W8TVI | kergoth, I though you might have talked to him about it and would know. |
| 06:34.38 | treke|home | W8TVI: Just ipkg fixes afaik |
| 06:34.39 | kergoth | ah, nope, dunno |
| 06:35.22 | treke|home | kergoth: You need to chooise either opie or qtopia or the qte build blows up |
| 06:35.29 | kergoth | ah. |
| 06:35.38 | treke|home | you can't build qte without specifiying what you will use it for |
| 06:35.38 | W8TVI | treke|home: sence I have the rc1 I dont really need to download the new ones yet? |
| 06:35.40 | kergoth | how silly. if netiher are selected it should not apply either's patches |
| 06:35.52 | Disconnect | ..which is reasonable, unless you want just, say, a kernel... |
| 06:35.54 | Disconnect | or (in this case) sources |
| 06:35.57 | treke|home | buildroot used to work fine :) |
| 06:36.07 | ljp | hey |
| 06:36.18 | ljp | dont blame me for your ills |
| 06:36.24 | mithro | anyone here use outlook or mozilla mail? |
| 06:36.35 | kergoth | ljp: qte shouldnt fail to build if neither qtopia or opie are selected, it should $(warning) about it and continue applying neither's patches |
| 06:36.44 | Grev | mithro, i do |
| 06:36.54 | kergoth | but anyway, its buildroot, who cares |
| 06:36.59 | kergoth | the last release that used it is out the door |
| 06:37.00 | ljp | not my fault |
| 06:37.14 | mithro | could you send me some nasty html email to "timadsl AT airnet.com.au" |
| 06:37.28 | kergoth | heh, testing your procmail rules? |
| 06:37.32 | kergoth | <PROTECTED> |
| 06:37.41 | treke|home | actually ljp, I think it is your fault :) |
| 06:37.49 | treke|home | mithro: Outlook web access good enough? |
| 06:38.07 | ljp | prove it |
| 06:38.12 | mithro | treke|home: yeah should be... |
| 06:38.22 | mithro | i'm testing my html -> plaintext convertor |
| 06:38.23 | treke|home | ljp: Ok. It worked, then you added qtopia to buildroot. It broke. |
| 06:38.29 | treke|home | mithro: Ah. |
| 06:38.39 | kergoth | Disconnect: let me know when you have the sources extracted, we'll go through it with the code in front of us |
| 06:38.59 | mithro | it does evolutions fine, so i need to test outlook and mozilla |
| 06:39.14 | Grev | mithro, give me a sec |
| 06:39.38 | Disconnect | ok. qte is extracted, its working on redownloading the kernel stuff (ugh. mrproper wipes sources. I'm of two minds as to whether thats "right".. on the one hand, it makes it really clean .. on the other hand, thats a LOT of download it wiped out) |
| 06:40.13 | treke|home | ok sent |
| 06:40.16 | kergoth | clean wipes everything but sources, mrproper wipes everything including sources. docs/BUILD documents that |
| 06:40.19 | kergoth | :P |
| 06:40.23 | Disconnect | kernel sources at sustained 314KB/sec .. that makes this usable. |
| 06:40.31 | Disconnect | yah |
| 06:40.44 | Disconnect | it ate my new kernel keymap, I think its still saved on the Z tho |
| 06:41.01 | Disconnect | (thats the one that took about 5 minutes to make ;) ..) |
| 06:41.16 | Grev | mithro, on its way |
| 06:41.27 | treke|home | save time. set $(SourceTree) to package/qte |
| 06:41.40 | Disconnect | ok its all extracted. |
| 06:41.41 | kergoth | heh |
| 06:41.51 | kergoth | Disconnect: okay, open up build/qt-2.3.*/src/kernel/qkeyboard_qws.cpp |
| 06:42.11 | Disconnect | 2.3.7 |
| 06:42.12 | Disconnect | got it |
| 06:42.13 | treke|home | 2.3.7 |
| 06:42.44 | kergoth | okay, lets take it one step at at ime, go to the QWSTtyKeyboardHandler constructor |
| 06:43.01 | Disconnect | fyi the licensing on any file modified by oz should be changed to reflect that some of the changes may not be available under the trolltech license.. |
| 06:43.18 | kergoth | yes, all our patches should have a copyright notice and probably a gpl blurb |
| 06:43.29 | kergoth | "someone" needs to go through it with a fine tooth comb |
| 06:43.39 | W8TVI | lol |
| 06:43.40 | W8TVI | Welcome to OpenZaurus "Collie on Steroids" 3.3.5 |
| 06:43.50 | kergoth | okay anyway, constructor opens the vt master |
| 06:43.52 | ljp | if you look at the makefile, it is not my fault |
| 06:43.54 | kergoth | sets up vt stuff |
| 06:43.59 | Disconnect | its not too bad to comb it - bk changes | awk | sed | etc | uniq | xargs gvim .. |
| 06:44.04 | Disconnect | ok |
| 06:44.30 | kergoth | then we have readKeyboardMap and readUnicodeMap |
| 06:44.30 | treke|home | if "someone" would get trolltech to commit your changes, there wouldnt be a problem. |
| 06:44.40 | kergoth | readKeyboardMap pulls in kernel_keymap |
| 06:44.51 | kergoth | which again, is our raw keycode -> transformed keycode translation table |
| 06:44.57 | *** join/#openembedded jdh41 (jdh41@student.cusu.cam.ac.uk) |
| 06:45.06 | kergoth | readUnicodeMap is just that, the kernel's idea of the unicode values for said keys |
| 06:45.16 | kergoth | 'acm' |
| 06:45.48 | kergoth | you'll note that the current kernel keymap as we read it only stores the value, not the type. that should probably be fixedc |
| 06:46.06 | kergoth | er no, thats not the case, i fixed that. good |
| 06:46.07 | kergoth | okay anyway |
| 06:46.13 | kergoth | go down to handleKey() |
| 06:46.29 | Disconnect | ok |
| 06:46.39 | kergoth | it gets passed in the value read from the kernel, a raw kernel keycode |
| 06:47.15 | kergoth | it gets directly used as the index into the qt keymap array |
| 06:47.20 | kergoth | to produce the keysym to pass up |
| 06:47.53 | Disconnect | ignore the special-casing for now? |
| 06:47.57 | mithro | btw you should make the jap keys control/alt :) |
| 06:48.01 | kergoth | you can see in the switch, is hardcoded info on the modifier mappings |
| 06:48.06 | Disconnect | mithro: thats my plan |
| 06:48.15 | kergoth | yeah, its special cased due to the fact that we dont handle the kenrel's modifier key types right now |
| 06:48.18 | Disconnect | right |
| 06:48.19 | kergoth | once we do, thatll just vanish |
| 06:48.38 | kergoth | you can also wee the map_to_modif function, which is just a kernel shiftstate to qt modifier translator |
| 06:49.18 | kergoth | okay, now what we *want*, is to take 'code', obtain the kernel transformed keycode for it including type (kernel_keymaps[current_map][code]) |
| 06:49.32 | kergoth | use that to index into the table that doesnt exist yet, to get the qt keysym for it |
| 06:49.43 | kergoth | also use that to get the kernel's unicode value for it |
| 06:49.46 | kergoth | then push it up |
| 06:50.10 | kergoth | the missing piece is transformed keycode -> qt keysym |
| 06:50.16 | kergoth | right now we only have raw keycode -> qt keysym |
| 06:50.21 | kergoth | the glue between those is the kernel keymap |
| 06:50.46 | kergoth | so like i said, to look at a .map which we know matches a qt keymap, look at sl5xxx vs drivers/char/collie_keymap.map |
| 06:51.08 | kergoth | thats sufficient information to get us from the former to the latter translation |
| 06:51.26 | kergoth | now obviously, collie doenst map everything, there *will* be holes |
| 06:51.40 | kergoth | but unless you want to go through it one sym at a time.. |
| 06:52.24 | Disconnect | hang on, lemme read that a few more times. it sounds right but when I try to parse it it just falls apart. its like trying to read someone else's perl code ;) |
| 06:52.40 | kergoth | yeah, i know.. it took me quite a while to wrap my head around the whole mess |
| 06:53.47 | Disconnect | ok. so code -> mCode as it exists now. then mCode -> ..? |
| 06:54.04 | kergoth | basically if you can get the raw -> transformed associations that match up with an existing raw -> qt keysym association, you can update the table |
| 06:54.21 | kergoth | mCode is the value of the key, blindly ignoring its type |
| 06:54.30 | kergoth | obviously itll only work if its a letter |
| 06:54.36 | kergoth | or certain other keys |
| 06:54.43 | kergoth | not exactly elegant, but .. |
| 06:54.54 | kergoth | we want to throw out mCode and use the entire keycode, not just the value |
| 06:55.32 | Disconnect | ahhhh. ok. so code holds type info? hang on, you covered that ;) lemme check.. |
| 06:55.37 | kergoth | yup |
| 06:56.16 | kergoth | its just a word. type is high byte, value is low byte |
| 06:56.31 | kergoth | K() assembles one, KTYP() rips off the high byte, KVAL() rips off the low |
| 06:56.40 | Disconnect | ahhh ok. |
| 06:57.26 | kergoth | so yeah, you had it right before. we want blaharray[K(KT_LETTER,KVAL('a'))] = Qt::Key_A; |
| 06:57.37 | kergoth | instead of blaharray[01] = Qt::Key_A; |
| 06:58.06 | kergoth | so if you have somewhere, the information that 01 is mapped to K(KT_LETTER,KVAL('a')) for this device, you have all the info you need |
| 06:58.16 | kergoth | and you *do*, thats exactly waht the collie kernel .map file is |
| 06:58.20 | Disconnect | ok. |
| 06:58.22 | *** join/#openembedded Kimchi__ (~kimchi@dyn-195-242-104-189.ppp.tiscali.fr) |
| 06:59.20 | kergoth | so it could be scripted, though you'd probably have to use a tool from console-tools to print the transformed keycode for each raw keycode |
| 06:59.47 | kergoth | then your script takes that and the qt keyMap and populates the new one |
| 07:00.10 | Disconnect | I thought we were only seeing the transformed keycode (eg '01' not 0x23..) |
| 07:00.35 | Disconnect | well '01 down' '01 release' but.. |
| 07:00.38 | kergoth | 01 isnt transformed, its raw, but yeah, i know what you're saying |
| 07:00.55 | kergoth | you thought we got handed K(KT_LETTER,KVAL('a')) |
| 07:01.01 | Disconnect | ugh. ok. yah. my bad - keycode = scancode again |
| 07:01.06 | kergoth | hehe |
| 07:01.15 | Disconnect | er keycode != that is |
| 07:01.22 | kergoth | its confusing. we have scancode, raw keycode, transformed keycode, then keysym |
| 07:01.29 | kergoth | not to mention the associated unicode value |
| 07:01.57 | Disconnect | ok. scancode = random hardware crap, we don't care ATM. keycode = 01. transformed = 'a' ("Console-Key" in the map comments) and keysym = Qt::Key_A .. right? |
| 07:02.11 | kergoth | right |
| 07:02.35 | kergoth | where 'a' becomes K(KT_LETTER, KVAL('a')) by the app that loads the .map into the kenrel |
| 07:02.42 | kergoth | (read loadkeys.y in the console-tools sources) |
| 07:02.47 | Disconnect | ok. the transformed keycode is in the map, although it needs a few more added (for the outside buttons, etc) |
| 07:02.55 | kergoth | right |
| 07:03.17 | kergoth | whew, i think you've got it |
| 07:03.20 | kergoth | :) |
| 07:03.26 | Disconnect | whee. my head hurts. ;) ;) but yah. |
| 07:03.42 | gb2 | heh |
| 07:03.44 | gb2 | keymaps again? |
| 07:03.46 | kergoth | there are only 2 decent kernel keyboard/console docs on the net |
| 07:03.48 | Disconnect | yep |
| 07:03.49 | kergoth | and they arent very good |
| 07:03.52 | gb2 | finally solving it? |
| 07:04.18 | Disconnect | if I don't stroke out we are ;) |
| 07:04.21 | kergoth | gb2: yep, will be nice to finally [x] out that todo item from the comments at the top of that file |
| 07:04.23 | cyn | wait, a keymap for c7x0 is almost there |
| 07:04.27 | kergoth | hehe, how long ago was that |
| 07:04.36 | kergoth | cyn: we arent talking about the keymap itself |
| 07:04.41 | kergoth | but making qte obey it better than it does now |
| 07:04.41 | cyn | right |
| 07:04.44 | cyn | the qte using it |
| 07:04.48 | kergoth | which will benefit every last device we've got |
| 07:04.53 | gb2 | kergoth: proper kernel keysym -> qt keysym map, then device specific kernel hardware key -> kernel keysym maps? |
| 07:04.54 | mithro | we need the back keys working as well |
| 07:04.55 | Disconnect | cyn: the kernel side is easy. I didn't check J1/J2 but I only had to add 4 missing ones. |
| 07:05.03 | mithro | ie the rocker etc |
| 07:05.11 | kergoth | gb2: yes, we manage remapping using kernel keymaps |
| 07:05.15 | kergoth | gb2: qte just obeys it |
| 07:05.19 | Disconnect | mithro: thats the easy part. you can even do that on the fly. but qte doesn't listen. |
| 07:05.20 | gb2 | right |
| 07:05.21 | gb2 | good good |
| 07:05.25 | cyn | goddamn that will be so tasty |
| 07:05.30 | kergoth | yep |
| 07:05.33 | cyn | I love you all, I just want to say that |
| 07:05.35 | kergoth | want to write an opie keymap manipulator? |
| 07:05.39 | kergoth | we could use one |
| 07:06.20 | gb2 | users need somewhere they can enter their credit card numbers and have beers pop out at developers' consoles. |
| 07:06.29 | cyn | yea |
| 07:06.37 | cyn | or other preferred beverage |
| 07:06.45 | kergoth | yeah |
| 07:06.47 | gb2 | espresso perhaps |
| 07:06.55 | kergoth | or at least gift certificates to a liquor store |
| 07:06.57 | kergoth | :) |
| 07:07.00 | gb2 | hehe |
| 07:07.03 | kergoth | someone donated a thinkgeek gift cert to me one time |
| 07:07.20 | gb2 | i've occasionally gotten amazon ones |
| 07:07.30 | Disconnect | kergoth: where do I find K() KVAL() etc? unless you want to populate this on the fly (it'll need code cleanup - been a while - but its doable) |
| 07:07.30 | kergoth | caffeine's a great donation for a dev |
| 07:07.41 | kergoth | they're happy, and the user benefits from their productivity resulting from it |
| 07:07.41 | Disconnect | lol |
| 07:07.42 | gb2 | and i have managed to collect on some "i owe you a beer" bits at OLS |
| 07:07.50 | kergoth | Disconnect: include/linux/kernel.map |
| 07:07.52 | kergoth | er |
| 07:07.53 | kergoth | .h |
| 07:07.58 | kergoth | jeeze, i've got keymaps on the brain |
| 07:08.04 | kergoth | ugh |
| 07:08.08 | kergoth | Disconnect: lets try that again |
| 07:08.12 | kergoth | Disconnect: include/linux/keyboard.h |
| 07:08.17 | Disconnect | flypiper owes me a tractor ride for doing up kiss a while back (ipaq dist, more as a political 'push' than as a dist. but he loves it.) |
| 07:08.39 | gb2 | wow OE is big. 4gigs in tmp |
| 07:08.44 | gb2 | 264M downloads |
| 07:08.59 | kergoth | gb2: yeah, we're collecting quite a bit of package metadata |
| 07:09.01 | Disconnect | ah ok. |
| 07:09.10 | kergoth | look at how many packages are sitting unfinished in nonworking/ nowadays |
| 07:09.16 | Disconnect | hrm. thought. |
| 07:09.40 | Disconnect | maybe not. gimme a sec. ;) |
| 07:09.52 | kergoth | Disconnect: applicable sources: qt-2.3.7/src/kernel/qkeyboard_qws.cpp, linux/include/linux/keyboard.h, linux/drivers/char/keyboard.c, linux/drivers/char/collie_keymap.map, console-tools/kbtools/loadkeys.{c,y} |
| 07:11.30 | Disconnect | how fast is switch/case on arm, using static values? not sure that a huge array is the best way to do it, and if switch/case is fast it should be relatively simple to maintain. (or - more importantly - generate) |
| 07:11.36 | gb2 | ok.. 2 do_install failures, 34 do_compile failures, 59 do_configure failures |
| 07:11.44 | gb2 | the rest are trickier |
| 07:12.09 | kergoth | Disconnect: well, this is a one time thing. if they change the transformed keycodes it could break things badly. code _can_ rely on that. code _shouldnt_ rely on scancodes, which it does now |
| 07:12.10 | gb2 | actually |
| 07:12.15 | Disconnect | that'd give us KQ(K(KT_LETTER,KVAL('a'))) == Qt::Key_A |
| 07:12.22 | gb2 | i think the gcc builds that it managed to do in the middle screwed it up |
| 07:12.48 | Disconnect | instead of a 65k array |
| 07:12.55 | Disconnect | (a mostly-empty 65k array) |
| 07:13.32 | kergoth | yeah, i wouldnt be opposed to that.. itd only be an init time performance hit anyway |
| 07:14.23 | Disconnect | ..? the switch/case would get hit for every keystroke. (although with any luck gcc would do something pretty with it since its all static values) |
| 07:14.39 | kergoth | oh, right |
| 07:15.02 | kergoth | heh, so how much do we trust gcc |
| 07:15.31 | Disconnect | the 65k array would also have to be machine-maintained, I'm thinking. whereas switch/case is human-parsable (although with any luck I'll be able to script it - the last one was 80% scripted) |
| 07:16.58 | kergoth | i dunno, the array is a direct index, a switch we need to compare each and every item until we find what we're looking for |
| 07:17.27 | Disconnect | I'm thinking something like "switch KTYP(keycode): case(KT_LETTER): int key=KVAL(keycode); switch(key):" .. |
| 07:17.40 | kergoth | why not something like.. |
| 07:17.41 | Disconnect | yah. we could do it iether way, but -I'm- not gonna poke through a 65k array to map the new key on whatever new device shows up ;) |
| 07:17.47 | kergoth | nononononono |
| 07:17.49 | Disconnect | or to change the mapping of J1/J2 |
| 07:17.57 | kergoth | the whole point of this is to NOT have to ever touch qt sources to map |
| 07:17.58 | kergoth | ever |
| 07:18.05 | Disconnect | ok. then its an array. easy enough ;) |
| 07:18.11 | kergoth | by mapping the kernel keymap output to a qt keysym, *everything* can be done within a kernel .map |
| 07:18.14 | kergoth | its static data |
| 07:18.30 | kergoth | now instead of doing a 64k array, why dont we do something like.. |
| 07:18.37 | Disconnect | ..I can load kernel keymaps on the fly, so qt will need some way to reread them (qcop presumably) |
| 07:18.53 | Disconnect | hrm. how many KTYP values are there, in reality? |
| 07:19.16 | Disconnect | oh. I'm looking at it. 12. |
| 07:19.20 | kergoth | struct keymap { int keycode; qt::keymap keysym } in a bigass array? |
| 07:19.25 | kergoth | then we can binary search the array |
| 07:19.45 | kergoth | it stays small, not bound as index == keycode, but we can search quicker than a switch/case |
| 07:19.50 | Disconnect | hrm. thats a thought. |
| 07:19.57 | kergoth | as long as its sorted of course |
| 07:20.07 | kergoth | by keycode |
| 07:20.15 | kergoth | i think thatd be cleaner, personally |
| 07:20.19 | kergoth | brb |
| 07:21.32 | Disconnect | although with only 12 values in the front, it becomes a 4k array |
| 07:22.06 | kergoth | Disconnect: hows that? |
| 07:22.13 | Disconnect | which gets you back to a static jump.. |
| 07:22.33 | Disconnect | kergoth: 2^4 + 2^8 |
| 07:22.57 | Disconnect | up to 16 types (13 currently) and 8 bits for value |
| 07:24.03 | Disconnect | the shifts to get into it get a little weirder (but not bad, although I'd have to think about it a bit to be sure it was right) but its still usable just as fast |
| 07:24.31 | kergoth | Disconnect: that doesnt gain you much, this is arm, itll pad the structure to the nearest word anyway. the keycode should be a normal 16 bit type, same as kernel |
| 07:24.39 | kergoth | (regarding your 2^4 + 2^8) |
| 07:24.54 | kergoth | bleh, headache |
| 07:25.46 | Disconnect | talking about an array, not the structure. no bsearch needed.. just KQ[Q(KTYP,KVAL)] where the shifts are different |
| 07:25.57 | Disconnect | (different from the existing K() ..) |
| 07:26.50 | kergoth | oh, you're trying to reduce the mem footprint of the array |
| 07:27.22 | Disconnect | something in me cringes when I look at a 64k array thats 63k of zeros. 4k with 3k of zeros isn't so bad ;) |
| 07:28.26 | kergoth | hehe |
| 07:28.55 | Disconnect | esp when its in a source file that someone is gonna end up sitting up late on irc trying to understand ;) |
| 07:29.06 | kergoth | wonder how much a bsearch would cost us in real world performance |
| 07:30.19 | Disconnect | in reality I think all 3 solutions would be "fast enough".. but 64k voids, huge switch blocks and binary searches aren't saying 'elegent' .. (neither is the 4k array, but its at least not screaming "life is pain please just let me DIEEE" like that 64k array was ;) ..) |
| 07:30.29 | Disconnect | well all 4 solutions that is |
| 07:30.36 | kergoth | yeah, i agree |
| 07:31.14 | kergoth | i dunno, i think as far as readability is concerned, an array of struct keymap's would be easiest to persue |
| 07:31.17 | kergoth | peruse |
| 07:31.18 | Disconnect | lets toss out the switch/case. that makes it big array, small array, struct.. |
| 07:31.25 | kergoth | yeah, i agree |
| 07:31.54 | Disconnect | possibly. although the existing array like that made my head spin ;) |
| 07:32.26 | kergoth | well, the main thing i dislike about the existing one is the readability regarding hte use of the index as the keycode |
| 07:32.38 | kergoth | you cant tell at a glance at line X is keycode X |
| 07:32.47 | Disconnect | but that gives us { KT_LETTER, Qt::Key_A }, { KT_Letter, Qt::Key_B }, etc .. right? |
| 07:32.56 | Disconnect | (and thats fixable with comments, same as I did on the one I sent in) |
| 07:33.20 | kergoth | { K(KT_LETTER, KVAL('a')), Qt::Key_A }, |
| 07:33.24 | kergoth | { K(KT_LETTER, KVAL('b')), Qt::Key_B }, |
| 07:33.29 | Disconnect | ah. my bad.. |
| 07:33.30 | Disconnect | yah. |
| 07:33.36 | kergoth | helps to include the value |
| 07:33.37 | kergoth | hehe |
| 07:33.38 | Disconnect | heh |
| 07:35.07 | Disconnect | all keys are now to be capitol A. that'd just make this so much simpler. ;) |
| 07:35.51 | Disconnect | and its all static values, so in theory gcc will parse it out properly. (in about a week, from what I've heard of 3.3's so-called speed..) |
| 07:36.04 | kergoth | hehe |
| 07:41.49 | Disconnect | ugh. awk hates me. |
| 07:41.57 | Disconnect | whats wrong with this? echo $line | awk " { print \"{ \$2, Qt::Key_$KEY },\" } " |
| 07:42.10 | Disconnect | bash is handling $KEY as expected ... oh. |
| 07:42.12 | Disconnect | its the comma. |
| 07:42.35 | Disconnect | no, its not. damnit. |
| 07:44.07 | kergoth | not escaping your { inside the print? |
| 07:44.14 | Disconnect | nah, doing that just causes warnings |
| 07:44.15 | kergoth | could be interpreting that as a block |
| 07:44.16 | kergoth | hm |
| 07:44.19 | kergoth | odd |
| 07:44.23 | Disconnect | "warning, \{ treated as normal {" |
| 07:44.59 | kergoth | what error is it giving? |
| 07:46.31 | Disconnect | none. its printing literal $2 |
| 07:46.43 | kergoth | ah |
| 07:46.45 | Disconnect | I'm just gonna set $line and use bash to do it |
| 07:48.06 | Disconnect | hrm. I can just use K_F1,Qt::Key_F1 since K_F1 == K(FT_FN,0) .. or did I confuse myself again..? |
| 07:48.30 | Disconnect | that'll handle most of them, then the keymap will get the rest.. |
| 07:48.55 | kergoth | yep, thats fine |
| 07:50.04 | kergoth | will be nice to have the modifiers handled cleanly. K_SHIFT,Qt::Key_Shift |
| 07:52.09 | Disconnect | yah gonna special-case that. basically a script that does the conversion with the kernel names, then translates them |
| 07:54.26 | kergoth | wonder how qt handles compose keys, for accents |
| 07:58.02 | Disconnect | ok got the first-pass converter for the "special" keys (from keyboard.h) |
| 07:58.09 | Disconnect | its way cleaner without that awk anyway |
| 07:58.21 | Disconnect | { K_F3 , Qt::Key_F3 }, |
| 07:58.32 | kergoth | nice |
| 07:59.01 | Disconnect | hrm. does bc do bitshifts? |
| 07:59.02 | kergoth | course at some point we'll want to sort it by keycode, so we'll need to pump it through the preprocessor |
| 07:59.10 | kergoth | i dont think so, i'm always wishing it does |
| 07:59.24 | kergoth | end up compiling a stupid test app just to get the value of a shift |
| 08:00.09 | Disconnect | hum. thats a pita. although { (KT_LATIN & (1>>8)),Qt::Key_A }, will work I think. if I'm remembering my binary properly. |
| 08:00.37 | Disconnect | ignoring the alt keycodes for now |
| 08:01.19 | kergoth | um |
| 08:01.24 | kergoth | 1>>8 is 0 |
| 08:01.29 | kergoth | last i checked |
| 08:01.31 | Disconnect | er. heh. |
| 08:02.38 | Disconnect | hey. duh. we talked about this earlier in fact. I can just spit K(KT_LATIN,1) .. or 'a'.. lemme go read my notes. |
| 08:02.42 | Disconnect | either way its no biggie. |
| 08:03.12 | kergoth | going from the loadkeys source, 'a' becomes K(KT_LETTER (or KT_LATIN), KVAL('a')) |
| 08:03.19 | Disconnect | yah 'a' (and is it KT_LETTER or KT_LATIN?) |
| 08:03.28 | kergoth | ones defined to the other in the loadkeys source |
| 08:03.33 | kergoth | so who knows |
| 08:03.53 | Disconnect | ah. kernel sees them differently. there was a comment, lemme find it.. |
| 08:04.01 | Disconnect | ah. |
| 08:04.19 | Disconnect | latin is latin, letter is affected by capslock. so I suspect we don't care either way. |
| 08:04.35 | kergoth | ah, shiftstate is irrelevlent at this point |
| 08:04.36 | kergoth | yeah |
| 08:04.57 | kergoth | wonder how qt handles the modif locks |
| 08:05.19 | Disconnect | grab me one, I'm out ;) |
| 08:06.18 | Disconnect | wow. that one is easy. unless I screwed it up.. |
| 08:06.25 | Disconnect | { (K(KT_LATIN,'Control')), Qt::Key_Control }, |
| 08:06.35 | Disconnect | it needs the translation still, but thats not bad |
| 08:06.48 | Disconnect | oh. damn. double translation. |
| 08:06.52 | Disconnect | { (K(KT_LATIN,'equal')), Qt::Key_equal }, .. |
| 08:07.29 | kergoth | thats not right. KVAL() is only a byte. a byte isnt sufficient to hold more than one character, much less a string for the word equal |
| 08:07.40 | kergoth | unless i'm missing something |
| 08:07.41 | Disconnect | yah thats what I mean by the double translation |
| 08:07.52 | kergoth | ah, k |
| 08:08.12 | kergoth | look at the loadkeys source, it is what takes the .map and makes a K() out of each one |
| 08:08.13 | Disconnect | hrm. and its got 'Shift' mapped. gonna have to filter and translate. fun. |
| 08:09.02 | kergoth | you ever look at the kernel drivers/char Makefile? |
| 08:09.07 | kergoth | it calls loadkeys --mktable on the .map |
| 08:09.23 | kergoth | might find that output useful |
| 08:23.40 | Disconnect | ugh. sed is turning "s/asterisk/\*/" into "ChangeLog" ?!? sed hates me. |
| 08:23.58 | kergoth | heh, need a \\\* perhaps? |
| 08:24.09 | kergoth | i hate figuring out wht needs escaping and what doesnt with sed |
| 08:24.21 | Disconnect | that gives me '\*' |
| 08:24.26 | kergoth | heh |
| 08:25.05 | Disconnect | 40 of these substitutions, it only gets one wrong.. ugh. well. moving right along.. I'll just pick it up in pass 2 ;) |
| 08:25.28 | kergoth | i like to use '' unless i damn well need bash to do a substitution. anything that cuts down on the number of \'s is good |
| 08:25.32 | kergoth | cool |
| 08:25.47 | Disconnect | um. ok. now here's the million dollar question. |
| 08:25.59 | Disconnect | what do I do with (eg) keycode 88 = Control .. |
| 08:26.48 | kergoth | hmmm |
| 08:27.58 | kergoth | example? the whole line? |
| 08:28.20 | Disconnect | keycode 88 = Control |
| 08:28.49 | Disconnect | and the comments indicate that 89 is contacts, but also Alt.. |
| 08:30.21 | kergoth | remember, we arent trying to be zaurus specific here. we're just trying to map the kernel's keycodes to the qt keysyms. how the qte keyboard handler makes use of it to ensure things happen as they should is outside the scope of the data you're trying to work on, imho |
| 08:30.30 | kergoth | s/keycodes/transformed keycodes/ |
| 08:30.54 | kergoth | it just happens that the 5500 is an example of a .map to associate with an existing qt keymap |
| 08:32.59 | Disconnect | lemme run out what I have and send it to you real quick |
| 08:33.04 | kergoth | k |
| 08:37.03 | Disconnect | first pass, hasn't been filtered for blanks or converted to Qt names.. I'm suspecting that (eg) that ctrl mapping will be handled by the K_CTRL mapping but I want another set of eyes on it before I finish it up |
| 08:37.06 | Disconnect | sent |
| 08:37.31 | kergoth | k, ill take a look |
| 08:39.21 | kergoth | yeah, that looks right |
| 08:39.55 | kergoth | and yeah, the shift/control stuff should be handled by the earlier mappings. the latter just control what raw keycode gets mapped to that K_SHIFT or what have you |
| 08:40.17 | Disconnect | cool. then its mostly done :) |
| 08:40.45 | kergoth | looks that way |
| 08:40.49 | kergoth | excellent. |
| 08:44.49 | *** join/#openembedded lsmith (~vandal@p213.54.232.133.tisdip.tiscali.de) |
| 08:46.44 | Disconnect | ugh what a PITA. test man pages (and googled docs, examples, etc) don't match the test on this system. |
| 08:46.54 | kergoth | doh |
| 08:47.55 | Disconnect | ah. gotta hardcode the path. damned sucky bash test. |
| 09:00.07 | Disconnect | I suspect there will be lots of keys missing from QT entirely. for now I'll just dump them (with a third filter) and if/when they start getting added it'll be easy enough to pull them from that one and add them to the translation filter |
| 09:00.44 | kergoth | yup |
| 09:00.53 | kergoth | i doubt qt has that many damn F keys for example |
| 09:01.01 | kergoth | much less handling for VT switch keys and shit |
| 09:01.07 | Disconnect | and 8 different ctrl buttons ;) |
| 09:01.10 | kergoth | hehe |
| 09:01.28 | ljp | ya, frickin Qt, you should use Gtk |
| 09:01.37 | Disconnect | ..or - heh - a numeric keypad |
| 09:01.48 | Disconnect | (which is what I suspect ASC0-9 is) |
| 09:02.10 | kergoth | no kidding, damn trolls |
| 09:02.11 | kergoth | :P |
| 09:05.21 | ljp | ya, TT is switching to Gtk for Qt 4 anyway |
| 09:05.50 | kergoth | hah |
| 09:06.52 | ljp | so you guys might as well give it up and start develpoing gpe |
| 09:07.22 | ljp | the reason for the switch is that gtk is LGPL |
| 09:07.40 | kergoth | yeah, who needs profit anywya |
| 09:15.00 | mithro | anyone here know much about bgp? |
| 09:17.53 | ljp | bgp? |
| 09:19.40 | *** join/#openembedded XavierXeon (~xx@p50877491.dip.t-dialin.net) |
| 09:25.33 | Disconnect | kergoth: done, I think. I didn't put in the third filter pass, but its easy enough to do. wanna check it one last time? (this time I'm sending the conversion script and hitting the sack, I'll do the filter tomorrow if nobody beats me to it. validate.sh | grep MISSING is a good start ;) ..) |
| 09:26.28 | kergoth | k, sure |
| 09:27.10 | Disconnect | here ya go.. enjoy :) and now tis time for bed, it being 4:30 am and all |
| 09:27.47 | kergoth | thanks alot, i appreciate the effort on this. |
| 09:27.48 | kergoth | night |
| 09:28.30 | Disconnect | who gets the keymap patch to map the 760 rear buttons? (shouldn't change this section, since they bind to preexisting keys) |
| 09:28.32 | ljp | yur clock is wrong, its 19:26 |
| 09:28.54 | Disconnect | not here it isn't.. |
| 09:29.30 | kergoth | Disconnect: i'd send it to openzaurus-devel |
| 09:29.50 | Disconnect | good enough. ttyl. and lemme know when you have a patch for me to test out ;) |
| 09:29.50 | kergoth | (lists.sf.net) |
| 09:29.55 | kergoth | will do |
| 09:30.07 | kergoth | i'll make use of it some time .. later today |
| 09:39.43 | ljp | bleh |
| 10:29.17 | *** join/#openembedded Kimchi__ (~kimchi@dyn-195-242-104-189.ppp.tiscali.fr) |
| 11:29.48 | *** join/#openembedded BadCode (~sumwid@dsl-213-023-250-169.arcor-ip.net) |
| 11:42.14 | *** join/#openembedded maxik (~maxik@82.140.72.252) |
| 11:53.59 | *** join/#openembedded dan2003 (~dan@cpc3-ware1-4-0-cust206.lutn.cable.ntl.com) |
| 12:05.49 | *** join/#openembedded stigger (~jeff@p017-n1-gos.nlc.net.au) |
| 12:33.59 | *** join/#openembedded DaveEL (~dave@dikleatherdale.plus.com) |
| 12:53.19 | dan2003 | ive been meryly hacking away at makefiles and patch files uptill this pint but this one has me stumped! |
| 12:53.19 | dan2003 | arm-linux-strip: ././CONTROL/preinst: File format not recognized |
| 12:53.19 | dan2003 | arm-linux-strip: ././CONTROL/postrm: File format not recognized |
| 12:53.19 | dan2003 | arm-linux-strip: ././CONTROL/postinst: File format not recognized |
| 12:53.19 | dan2003 | arm-linux-strip: ././CONTROL/prerm: File format not recognized |
| 12:59.10 | dan2003 | why is it trying to strip symbols form the config files? |
| 13:05.16 | mickeyl | we obviusly call arm-linux-strip for every file. don't worry, that's cosmetics |
| 13:09.44 | dan2003 | ok, so the build is complete :) |
| 13:10.00 | dan2003 | can i just check, c760 is husky or shepeard? |
| 13:11.29 | mickeyl | ibot: husky |
| 13:11.31 | | extra, extra, read all about it, husky is Sharp Zaurus SL-C760 or a dog |
| 13:11.47 | dan2003 | ahh, noce1! |
| 13:24.47 | *** join/#openembedded tomasito (~tomasito@p508C29E1.dip.t-dialin.net) |
| 13:34.39 | *** join/#openembedded darkschneider (~gab@62-101-126-192.fastres.net) |
| 13:39.26 | *** join/#openembedded XavierXeon (~xx@pD951FA08.dip.t-dialin.net) |
| 13:44.00 | *** join/#openembedded CAMCAM (~camrein@AMontsouris-108-1-35-207.w81-49.abo.wanadoo.fr) |
| 13:46.50 | darkschneider | mm the updater.sh 3.3.5 unstable/poodle look like not working |
| 13:47.26 | darkschneider | it can't flash those correctly, going to try with with the old updater.sh in the experimental dir |
| 13:48.58 | mickeyl | good |
| 13:49.03 | mickeyl | i can't do anything about that, i have no poodle |
| 13:49.09 | mickeyl | if you got a solution, tell me |
| 13:49.11 | mickeyl | bbl |
| 13:49.16 | darkschneider | ok i will be your testing suit |
| 13:49.17 | darkschneider | later |
| 13:54.04 | *** join/#openembedded pb_ (~pb@2002:5160:45ef:0:230:65ff:fe0a:e182) |
| 13:58.00 | darkschneider | mm i really wonder why kernel is not flashing... |
| 14:29.43 | Twiun | just found out the laptop I'm looking at is 21% cheaper in the states and comes bundled with extras... how typical |
| 14:30.07 | darkschneider | yeah |
| 14:30.52 | darkschneider | that's common... |
| 14:31.39 | darkschneider | mickeyl|advent: changing the updater.sh did not helped... it still iddles at the logo and apprentlt isn't flashing the kernel.. |
| 14:38.24 | Twiun | grrr hate not being able to ring direct to a local store |
| 14:38.36 | darkschneider | yeah that sucks |
| 14:48.55 | darkschneider | mickeyl|advent: i can azard that the initrd.bin is too big, because replaing flashing the old experiemntal rom jut works (it's 11.x MB) |
| 14:57.21 | *** join/#openembedded bisho (~bisho@136.Red-213-97-191.pooles.rima-tde.net) |
| 15:04.42 | pokryfka | what is the difference between gcc-cross, gcc-initial-cross and gcc-initial-uclibc-cross? |
| 15:30.15 | pokryfka | ? |
| 15:34.03 | *** part/#openembedded pokryfka (~pokryfka@b16m10.Zamkowa.virt.pl) |
| 16:04.53 | *** join/#openembedded leffe (~leffe@as5-6-1.h.s.bonet.se) |
| 16:17.00 | chouimat | morning |
| 16:59.47 | *** join/#openembedded Kimchi_ (~kimchi@dyn-81-166-42-30.ppp.tiscali.fr) |
| 17:19.42 | *** join/#openembedded pb___ (~pb@dsl-62-3-66-201.zen.co.uk) |
| 17:20.07 | *** join/#openembedded mattaw (~mas01r@arm.ecs.soton.ac.uk) |
| 17:20.17 | mattaw | ibot seen mickeyl |
| 17:20.22 | | mickeyl is currently on #openembedded (16h 30m 29s) #handhelds.org (16h 30m 29s) #openzaurus (16h 30m 29s). Has said a total of 577 messages. Is idling for 3h 31m 11s |
| 17:20.35 | darkschneider | lo AndersB |
| 17:20.43 | darkschneider | mickeyl: left soem time abgo |
| 17:20.57 | darkschneider | <darkschneider> it can't flash those correctly, going to try with with the old updater.sh in the experimental dir |
| 17:20.58 | darkschneider | <mickeyl> good |
| 17:21.00 | darkschneider | <mickeyl> good |
| 17:21.00 | darkschneider | <mickeyl> i can't do anything about that, i have no poodle |
| 17:21.00 | darkschneider | <mickeyl> if you got a solution, tell me |
| 17:21.00 | darkschneider | <mickeyl> bbl |
| 17:21.22 | mattaw | ah ok then so the 3.3.5 doesn't fly for the sl5600 |
| 17:21.31 | mattaw | will revert to my experimental 2.0 |
| 17:21.31 | darkschneider | lok lijk nope |
| 17:21.34 | darkschneider | i have tested |
| 17:21.47 | darkschneider | to mix never kenel image with older initrd.bin |
| 17:21.54 | darkschneider | and vice versa |
| 17:22.01 | mattaw | did that too but no luck |
| 17:22.10 | darkschneider | and tested the other updater.sh around |
| 17:22.17 | darkschneider | ywah :) no luck |
| 17:22.34 | darkschneider | look like it does nto flash the kernel imho |
| 17:22.43 | darkschneider | maybe it's a size problem |
| 17:23.00 | darkschneider | it's the only obivios difference of the 2 roms |
| 17:23.40 | mattaw | I think it does flash the kernel and some things work |
| 17:24.00 | darkschneider | mmm |
| 17:24.02 | darkschneider | intresting |
| 17:24.07 | mattaw | I have been using a freshly formatted fat16 cf card and a Documents DIR mysteriously appears |
| 17:24.15 | darkschneider | yeah! |
| 17:24.17 | darkschneider | that's true |
| 17:24.52 | darkschneider | but normaly it output package isntalltion |
| 17:24.55 | mattaw | never mind, the next release would be good. I am just annoyed with the hardware bugged xscale 250 |
| 17:24.58 | darkschneider | so it halts before packages |
| 17:25.12 | mattaw | darkschneider, true |
| 17:25.27 | darkschneider | we need debugging rom.. |
| 17:25.33 | darkschneider | that output something.. |
| 17:25.49 | darkschneider | i hope next relase come from eobuild |
| 17:25.53 | mattaw | trouble is I was able to get it up and running with buildroot so I am curious what mickey used that I didn't |
| 17:26.24 | darkschneider | yeah, same here for last version... i never sucseded a full compialtion.. |
| 17:27.14 | leffe | mattaw: i can confirm your problem, just tried with a 128MB cf... update went ok |
| 17:27.17 | mattaw | perhaps we should wait for that lovely new zaurus with the 640x480 screen and built in bluetooth and wifi |
| 17:27.31 | mattaw | leffe, drat, and I hoped it was just me :( |
| 17:27.41 | mattaw | I shall mail the list that we three get the same result |
| 17:27.56 | darkschneider | leffe: so same problem... mmm |
| 17:28.01 | leffe | sure |
| 17:28.10 | darkschneider | mattaw: yeah sweet.. i hope to have money for it |
| 17:28.19 | leffe | im very sad now ;-( |
| 17:28.21 | darkschneider | mattaw: ok, then mail |
| 17:28.29 | darkschneider | anwyay i wll report to ker this nigth |
| 17:29.15 | mattaw | good plan |
| 17:29.48 | mattaw | there mailed |
| 17:29.54 | mattaw | and I have to get back to work |
| 17:29.57 | mattaw | all the best people |
| 17:32.24 | treke|home | mickeyl: ping. |
| 17:33.30 | darkschneider | treke he is away |
| 17:34.18 | treke|home | oh well. I've got a pretty much complete pxa feed done when he comes back :) |
| 17:34.53 | darkschneider | mm |
| 17:34.55 | darkschneider | intresting |
| 17:35.12 | darkschneider | we have reports on sl5600 |
| 17:42.46 | mickeyl | pong |
| 17:42.56 | mickeyl | treke: got that ipaq build? |
| 17:43.22 | chouimat | hi mickeyl treke |
| 17:44.00 | darkschneider | lo mickeyl |
| 17:44.06 | mickeyl | yo |
| 17:45.15 | treke|home | no ipaq build |
| 17:45.16 | treke|home | yet |
| 17:45.48 | mickeyl | k |
| 17:45.55 | mickeyl | got around to a gcc2-pxa build? |
| 17:46.10 | mickeyl | there's something I didn't realize |
| 17:46.21 | mickeyl | we unfortunately need a seperate feed and image for poddle |
| 17:46.24 | treke|home | gcc2-pxa build is pretty much done |
| 17:46.27 | treke|home | mickeyl: why? |
| 17:46.31 | mickeyl | treke: two reasons |
| 17:46.51 | mickeyl | 1.) the c7x0 feeds contain the big images for high resolution |
| 17:46.53 | treke|home | I think i'm just waiting for konq and a couple misceleanous apps. |
| 17:47.00 | mickeyl | 2.) hmm... can't remember right now |
| 17:47.43 | darkschneider | i can add that's why the actual one does not work :) |
| 17:47.57 | mickeyl | no that must be another problem |
| 17:48.02 | mickeyl | it should work |
| 17:48.04 | mickeyl | i don't get it |
| 17:48.12 | mickeyl | how can the flash image be too large? |
| 17:48.19 | mickeyl | the 5600 has 32mb rootfs hasn't it? |
| 17:48.37 | darkschneider | yeah yeah the problem reveled to be another |
| 17:48.43 | darkschneider | and unkown |
| 17:48.58 | darkschneider | we are 3 with 5600 and all with the same problems |
| 17:49.14 | mickeyl | hmm |
| 17:49.16 | darkschneider | mattaw mailed the lsit with the report |
| 17:49.27 | darkschneider | we tried varuois things and all got the same results |
| 17:49.28 | mickeyl | you need to do a build with debug output on the console |
| 17:49.35 | darkschneider | and the same "documents" dir appearing |
| 17:49.38 | darkschneider | yeah |
| 17:50.04 | darkschneider | but mattaw was nto able to build the buildroot |
| 17:50.19 | darkschneider | he was lookign for you from some suggestion |
| 17:50.26 | mickeyl | mmmh |
| 17:50.33 | mickeyl | i'm not in the mood for suggestions |
| 17:50.37 | mickeyl | pretty much outburned |
| 17:50.41 | darkschneider | <mattaw> trouble is I was able to get it up and running with buildroot so I am curious what mickey used that I didn't |
| 17:50.41 | mickeyl | doing such a big release is hard |
| 17:50.49 | darkschneider | yeah :) |
| 17:50.53 | darkschneider | but you good ): |
| 17:51.02 | mickeyl | i even coordinated opie 1.0.3 *sigh* |
| 17:51.14 | darkschneider | i have read your work |
| 17:51.37 | darkschneider | i'm allways astonished by many of you |
| 17:52.50 | mickeyl | i will build you a kernel with debug output on the console |
| 17:52.55 | mickeyl | but that's about all i can do |
| 17:53.10 | darkschneider | that's a lot! |
| 17:53.19 | darkschneider | so i can work on it! |
| 17:54.09 | mickeyl | underway... |
| 18:00.51 | mickeyl | uploaded. |
| 18:00.58 | mickeyl | check poodle/kernels/... |
| 18:06.07 | darkschneider | oki |
| 18:06.10 | *** join/#openembedded erf (~erf@81-86-164-132.dsl.pipex.com) |
| 18:12.24 | cyn | mickeyl: prolly icons different size too, no? [c7x0 feed vs poodle] |
| 18:13.48 | mickeyl | at the moment not |
| 18:13.53 | mickeyl | but it should be made so |
| 18:14.00 | mickeyl | atm, they're sharing the same feed |
| 18:15.52 | cyn | yeah you were talking about the reasons they needed a sep. feed |
| 18:15.52 | cyn | er wait feed |
| 18:16.14 | cyn | feed is just the installation ipkgs not the image/etc. - oh I guess each app comes with its own icon dur |
| 18:20.15 | mickeyl | not all |
| 18:20.20 | mickeyl | a lot of apps are sharing inline images |
| 18:20.22 | mickeyl | which are in libqpe |
| 18:20.30 | mickeyl | for better performance |
| 18:30.20 | kergoth | dan2003: it calls arm-linux-strip on every file thats executable. calling file or readelf on every binary would be a large performance hit |
| 18:30.26 | kergoth | morning people |
| 18:32.43 | chouimat | hi kergoth |
| 18:34.34 | chouimat | <PROTECTED> |
| 18:36.07 | mickeyl | hi kergoth |
| 18:43.35 | pb_ | yo kergoth |
| 18:43.49 | *** join/#openembedded vandal (~vandal@p213.54.232.70.tisdip.tiscali.de) |
| 19:08.22 | darkschneider | yo kergoth |
| 19:10.40 | dan2003 | ive built downloaded qt-embeddedand opie and am trying to follow the tutorial for hello world, it gives some trivial code but it doesnt say what to do with it, just says compile! am i suposed to just put it in file (if so what do i copile it with?) or do i put it in qt-designer somehwere or what? (sorry for being so dumb!) |
| 19:16.38 | kergoth | http://doc.trolltech.com/qtopia1.6/html/start.html |
| 19:16.41 | kergoth | try reading |
| 19:16.49 | kergoth | heh |
| 19:17.41 | darkschneider | here we go... stdout kernel |
| 19:17.52 | darkschneider | mmm nothing... |
| 19:18.05 | darkschneider | nothing... |
| 19:18.27 | darkschneider | and nothing... |
| 19:18.44 | *** join/#openembedded mattaw (~mas01r@arm.ecs.soton.ac.uk) |
| 19:18.50 | darkschneider | mickeyl: bad news is that it outputs nothing |
| 19:18.52 | darkschneider | mattaw: |
| 19:18.55 | mattaw | mickeyl, hiya |
| 19:18.57 | darkschneider | wellcome just in time |
| 19:19.01 | mattaw | just read your email |
| 19:19.09 | mattaw | hiya dark |
| 19:19.11 | darkschneider | i just tested the new kernel from mickeyl |
| 19:19.16 | mattaw | and? |
| 19:19.17 | darkschneider | same issure |
| 19:19.24 | darkschneider | and no output |
| 19:19.43 | mattaw | what did mickey have to say about it? |
| 19:19.49 | darkschneider | let me check if it generates the Documents dir |
| 19:20.16 | mickeyl | i have no idea folks |
| 19:20.19 | mickeyl | i never even seen a poddle |
| 19:20.25 | mickeyl | poodle even |
| 19:20.36 | darkschneider | a change! |
| 19:20.42 | mattaw | what happened? |
| 19:20.44 | darkschneider | cf/Documents/Install_Files |
| 19:20.49 | darkschneider | another dir created |
| 19:21.07 | mattaw | it must be the installer I suppose but why on earth would that have been affected |
| 19:21.09 | darkschneider | i really wonder what is the heck happening |
| 19:21.15 | darkschneider | no idea... |
| 19:21.28 | mattaw | mickeyl, is ipkg used to install the base system from the initrd? |
| 19:21.35 | darkschneider | i just changed the kernel image with the one jus provvided |
| 19:21.38 | mattaw | or is the initrd a complete fs |
| 19:21.42 | mickeyl | fs |
| 19:21.47 | mattaw | right |
| 19:22.09 | darkschneider | yeah it's image... it has no meaning of crating just 2 dirs |
| 19:22.15 | mattaw | well the kernel isn't the issue as I have tried known good kernels with this initrd with exactly the same result |
| 19:22.32 | darkschneider | indeed, same here |
| 19:22.54 | mattaw | err wait a sec, the initrd is independent of kernel assuming the kernel is not too old |
| 19:23.05 | darkschneider | i have switeched bith too |
| 19:23.15 | darkschneider | i used old kernel new init and vice versa |
| 19:23.40 | darkschneider | so this kernel does not boot with a working initrd |
| 19:23.44 | mattaw | I think I should just mail my sl5600 to mickey and let him have it :) oh well there is always the /unofficial/experimental stuff I did ages ago |
| 19:24.08 | darkschneider | mattaw: i'm still using it :) (was still) |
| 19:24.51 | mattaw | well thats too out of date for me now. I am in the last year of my PhD so spare time I prefer to waste elsewhere unfortunately |
| 19:25.06 | darkschneider | mattaw: sweet |
| 19:25.14 | darkschneider | what you study |
| 19:25.20 | darkschneider | me future engineer |
| 19:25.34 | mattaw | researching in the dept of Electronics and ComputerScience |
| 19:25.36 | darkschneider | lol |
| 19:25.44 | darkschneider | wow, phd in that is nice |
| 19:25.47 | mattaw | kergoth, hello how are things these days? |
| 19:25.51 | Twiun | kergoth: probably same place you put your spare time |
| 19:26.01 | darkschneider | kergoth: you said to have a working sl5600 :) |
| 19:26.07 | mickeyl | mattaw: no please not... i'm already buried in PDAs here |
| 19:26.09 | mickeyl | ;) |
| 19:26.26 | mickeyl | i need a prism2 cf card though if someone's in donation mode... |
| 19:26.27 | mattaw | hehe |
| 19:26.49 | mattaw | you want the host ap stuff going or just like the prism? |
| 19:27.49 | darkschneider | mickeyl: packing a minimal initrd.bin is a issure? |
| 19:27.50 | mickeyl | i want to work on AP functionality for the Z |
| 19:28.16 | mickeyl | darkschneider: uhmm.... i'm really tired |
| 19:28.39 | mickeyl | there has to be someone else |
| 19:28.46 | darkschneider | mickeyl: ok,nm then, sorry to be so insistent.. i feel useless like allways ): |
| 19:29.39 | darkschneider | i think i can give another try to buildroot if some of you provvide some fast link :) |
| 19:29.40 | mattaw | mickeyl, just a quick one are you still using buildroot? and if so how heavily modified is it |
| 19:30.03 | mattaw | darkschneider, look, when I last did it it was a nightmare |
| 19:30.04 | mickeyl | mattaw: not at all. the only thing I have is a populated source directory. |
| 19:30.09 | darkschneider | mattaw: indeed following your idea, if we divide work we can get somethign out |
| 19:30.28 | mattaw | mickeyl, you building on debian? |
| 19:30.32 | mickeyl | eeek no |
| 19:30.38 | mickeyl | I'm a mandrake guy :) |
| 19:30.42 | darkschneider | eheh |
| 19:30.48 | mattaw | and the toolchain was from handhelds? or both from openzaurus? |
| 19:30.52 | mickeyl | handhelds.org |
| 19:30.55 | mattaw | I am using fedora core |
| 19:31.03 | mickeyl | the 3.3.2 from handhelds.org and the 2.95.3 from oz.org |
| 19:31.06 | darkschneider | me usong one from zaurus |
| 19:31.18 | mattaw | mickeyl, doesn't that one give rise to a missing stdlibc++ dep? |
| 19:31.42 | mickeyl | mattaw: yeah, but i was pragmatic and just added STL to the rootfs :D |
| 19:31.51 | darkschneider | lol |
| 19:32.07 | mickeyl | i don't feel like working around buggy toolchains today |
| 19:32.17 | mickeyl | so just going the easy road |
| 19:32.41 | mickeyl | s/working around/fixing |
| 19:32.52 | darkschneider | mattaw: we waiting ideas... |
| 19:33.04 | kergoth | nice, disconnect's keysym table is nearly complete |
| 19:33.07 | mattaw | hehe well thats the strategy I used |
| 19:33.11 | kergoth | just need to alter my keyboard handler to make use of it.. |
| 19:33.13 | mickeyl | ah cool |
| 19:33.25 | kergoth | then we can use the damn jogdial and shit |
| 19:33.26 | mattaw | mickeyl, you got xinelib building without modifying it? |
| 19:33.36 | mickeyl | mattaw: yes. builds out of the box for me |
| 19:33.49 | darkschneider | mickeyl: what about just packing the whole thing up and passign it to me ans mattaw? |
| 19:33.56 | kergoth | whether it builds or not out of hte box depends on what you have installed on your desktop |
| 19:34.01 | mickeyl | define "the whole thing" |
| 19:34.04 | kergoth | the ./configure script looks in /usr for aalib |
| 19:34.08 | kergoth | if you have it, things break badly |
| 19:34.25 | darkschneider | mickeyl: the buildroot confs yo used (no source) |
| 19:34.32 | kergoth | no worries, i'll fix it when i move xine into OE |
| 19:34.43 | kergoth | heh |
| 19:34.52 | mickeyl | didn't i already tell that i didn't change anything in buildroot? |
| 19:35.00 | mickeyl | all i have is a populated source directory |
| 19:35.10 | mickeyl | so no missing upstream locations |
| 19:35.12 | *** join/#openembedded pb_ (~pb@2002:5160:45ef:0:230:65ff:fe0a:e182) |
| 19:35.14 | darkschneider | mmm ok i missed that sorry |
| 19:35.34 | Twiun | wb pb_ |
| 19:35.39 | pb_ | Twiun: :-) |
| 19:35.50 | darkschneider | ok i have done, i'm out of stupid ideas |
| 19:35.59 | mickeyl | me too |
| 19:36.04 | mickeyl | need a rest |
| 19:36.04 | mickeyl | bbl |
| 19:36.08 | darkschneider | later |
| 19:37.17 | kergoth | Twiun: problems? |
| 19:39.23 | mattaw | kergoth, I love the fact you can think about things that make no sense to me whatsoever! |
| 19:39.49 | Twiun | kergoth: yeah, trying to get Swing working on fbAWT... there's a missing link between repainting on offscreen images and dumping the contents finally on the framebuffer |
| 19:40.05 | Twiun | kergoth: both bits seem to work great, but independently! |
| 19:40.23 | kergoth | ahh that sucks |
| 19:41.11 | Twiun | it's like, image(0) is the framebuffer, and I get images drawing 4->3, 3->2 and 4->1 but not a single one that does ->0 |
| 19:41.21 | Twiun | so the framebuffer ends up all black ;) |
| 19:41.38 | kergoth | hah, thats annoying |
| 19:41.56 | Twiun | and the levels of abstraction are... erm, distracting |
| 19:42.13 | Twiun | I'm trying to juggle the class hierarchy in my head ;) |
| 19:42.44 | Twiun | on the good side, I'm not getting any exceptions thrown, so nothing breaks badly |
| 19:44.02 | Twiun | I suppose I should be thankful I don't need to test on the actual hardware anymore ;) |
| 19:45.30 | kergoth | hehe. draw pictures |
| 19:45.45 | kergoth | ah ha, i can handle compose and string easy enough |
| 19:45.47 | kergoth | hmm |
| 19:46.36 | Twiun | I'd draw pictures but the execution path is what I *really* need a graph of, I need to know where to hack that missing step in |
| 19:47.00 | kergoth | ah |
| 19:47.04 | mattaw | just a quick question but any info on how bugged the xscale pa250 is compared to the pa255 shipped with the new 760's? |
| 19:47.14 | kergoth | mattaw: intel.com. |
| 19:47.24 | kergoth | look for pxa250 errata |
| 19:47.48 | Twiun | kergoth: even with only one component visible on screen, I get 35 graphics contexts created :( |
| 19:48.04 | Twiun | Swing's not exactly resource conscious |
| 19:48.17 | darkschneider | i dream the sl-6000 :)) |
| 19:48.33 | Disconnect | kergoth: anything important missing from those scripts? |
| 19:48.45 | Disconnect | darkschneider: nah. you dream a 760 w/ reflective screen ;) |
| 19:49.07 | darkschneider | eheheh |
| 19:49.27 | kergoth | Disconnect: nope, seems fine |
| 19:50.06 | Disconnect | cool |
| 19:50.30 | kergoth | the backlight thing is easily fixed, the ranges are set in ODevice, opie's abstraction for deviceisms |
| 19:51.08 | Disconnect | did you want that filter list, for keys that have no Qt bindings? (just gonna pipe the validator through sed/awk to make a sed filter) |
| 19:51.52 | Disconnect | someone needs to update def-configs to reflect whats actually built. in theory I should be able to copy def-configs/husky-3.3.5-release .config ; make oldconfig (NO MISSING QUESTIONS) and build exactly whats on the site... |
| 19:52.12 | kergoth | buildroot wont be used for any more releases, so its pretty low priority |
| 19:52.14 | Disconnect | in practice.. its a little (lot) messier. I realise everyone is focusing on the new build method, but still.. |
| 19:52.16 | kergoth | but yeah, thats right |
| 19:52.35 | Disconnect | its easy enough. just put the configs that were used up somewhere we can get to them ;) |
| 19:53.09 | Disconnect | screw checking them into bk or whatever. throw them in the feed dir or something |
| 19:53.27 | darkschneider | mm intresting the buildroot for poodle by default select kernel for c700.. shpuldn't it be b500 ? |
| 19:53.33 | kergoth | pester mickey when he gets back |
| 19:53.35 | kergoth | heh |
| 19:53.35 | Twiun | <PROTECTED> |
| 19:53.38 | Twiun | oops |
| 19:53.41 | kergoth | darkschneider: the 'kernel-shepherd' tree supports poodle |
| 19:53.43 | darkschneider | c700/c750 |
| 19:53.47 | darkschneider | oki |
| 19:53.55 | darkschneider | tx |
| 19:53.59 | ljp | I'm giving up opie development, and starting pocketpc devel |
| 19:54.15 | ljp | openpocket |
| 19:55.26 | Disconnect | ljp: isn't that what familiar is for? take the commercial bits and make them free.. (someone really needs to write Pwine to run ppc apps under fam/oz ;) ..) |
| 19:55.41 | ljp | no. that runs on linux |
| 19:55.50 | Disconnect | yah I was kidding |
| 19:56.08 | ljp | heh, me too |
| 19:56.21 | Disconnect | somehow I suspected that ;) |
| 19:56.32 | ljp | I think my boss might get a bit dismayed if I switched to pocketpc |
| 19:56.56 | pb_ | unusually perceptive for a boss if he notices. |
| 19:57.04 | *** join/#openembedded Virusmaster (~Virus@pD9E04843.dip.t-dialin.net) |
| 19:58.17 | ljp | I think he might notice right away |
| 19:58.38 | Disconnect | all the cursing? or the different devices..? |
| 19:58.49 | ljp | :) |
| 19:59.19 | ljp | I did write a pocketpc app... I changed an appwiz about notice to say copyright trolltech |
| 19:59.33 | ljp | I was checkin gout embeddedvc++ |
| 19:59.49 | ljp | checking out, even |
| 20:00.09 | ljp | my app is called freezer |
| 20:00.20 | pb_ | sounds ominous |
| 20:00.22 | ljp | I am going to sel it for $24.95 |
| 20:00.27 | ljp | hehe |
| 20:01.06 | ljp | prob'ly need to change the icon, though |
| 20:01.56 | ljp | heh make the app popup an dialig which says, 'this wouldnt happen if you ran linux' and then freeze the system |
| 20:02.19 | ljp | I need to drink more coffee |
| 20:08.01 | Disconnect | you and me both. |
| 20:08.08 | Disconnect | course, I was up till 5am this morning, so maybe not. |
| 20:10.12 | ljp | it is 6 am here |
| 20:10.39 | Disconnect | ah. well.. there ya go ;) you prolly need less coffee. |
| 20:12.40 | Disconnect | mickeyl|tv: you still afk? |
| 20:18.00 | CIA | repository=03buildroot user=11mickeyl@gandalf.tm.informatik.uni-frankfurt.de |
| 20:20.41 | Disconnect | mickeyl|tv: you can run but you can't hide from the checkin daemon ;) ;) wanted to get a copy of the config used for 3.3.5 on husky, so I can test some patches |
| 20:21.29 | mickeyl|tv | commercial break is over - must run |
| 20:21.34 | mickeyl|tv | yeah, that's what i commited |
| 20:21.38 | Disconnect | cool thanks :) |
| 20:40.42 | *** join/#openembedded LinuxDoc (~chrissy@p5080D3F3.dip.t-dialin.net) |
| 21:01.20 | *** join/#openembedded SubZerooo (sub@24-90-164-89.nj.rr.com) |
| 21:02.30 | SubZerooo | hey is open embeded gonna be the next generation(version) of openzaurus.. or is it just some developer tools? |
| 21:03.19 | kergoth | openzaurus isnt going away. |
| 21:03.29 | SubZerooo | i know |
| 21:03.37 | kergoth | the oe tools will be used to produce version 3.5.1 of OZ |
| 21:03.44 | kergoth | among other distributions and incarnations |
| 21:03.47 | SubZerooo | ahh ic ok thanks |
| 21:03.53 | SubZerooo | i was just always confused |
| 21:16.27 | gb2 | <PROTECTED> |
| 21:16.37 | gb2 | 13 packages failed to build properly with uclibc |
| 21:16.59 | kergoth | not bad |
| 21:17.12 | ljp | 13 out of how many? |
| 21:17.24 | *** join/#openembedded devnull_cat (~devnull_c@pcp01823855pcs.howard01.md.comcast.net) |
| 21:17.24 | gb2 | no wait.. 14, i missed one. |
| 21:17.43 | gb2 | i have 113 package tarballs |
| 21:18.03 | gb2 | hm, that sounds short |
| 21:18.39 | kergoth | we we have like 179 packages that are buildable atm |
| 21:18.50 | gb2 | some spit out multiple tarballs though |
| 21:19.13 | kergoth | find . -name \*.oe|wc -l |
| 21:19.14 | kergoth | heh |
| 21:19.36 | kergoth | bleh, set maxdepth so nonworking doesnt show |
| 21:20.33 | gb2 | 233 |
| 21:21.12 | gb2 | i wonder why it chose not to build a lot of stuff |
| 21:22.27 | gb2 | currently build order doesn't do the right thing since some things require libc but it doesn't get specified explicitly |
| 21:26.54 | Disconnect | wow. wget over the cablemodem is 95% as fast as scp over the wlan.. makes me feel silly for aborting 5M into the dl.. |
| 21:27.42 | Disconnect | does OE require the gcc2 env as well as 3? (its not listed either way in required-software) |
| 21:28.13 | kergoth | OE doesnt require anything. what you're building may |
| 21:28.26 | kergoth | s/anything/much/ |
| 21:28.27 | kergoth | heh |
| 21:28.51 | Disconnect | oe requires python (minor) and bitkeeper (major, to many people..) ;) |
| 21:29.20 | kergoth | oe doesnt require bitkeeper. |
| 21:29.45 | Disconnect | "To start using OE, first get the code from BitKeeper. You need both the repositores 'oe' and 'packages'." |
| 21:29.56 | kergoth | http://treke.net/~bk/ |
| 21:30.16 | kergoth | technically we dont have to make snaps available until you ask, for gpl compliance |
| 21:30.17 | Disconnect | yah but thats not in the wiki (yep, its a wiki. yep, I'm adding it..) |
| 21:30.55 | gb2 | heh |
| 21:31.10 | Disconnect | technically never until you ship bins from it |
| 21:31.16 | cyn | say who's this Disconnect bloke and why's he so useful knowledgable and helpful |
| 21:31.32 | gb2 | cyn: new infobot |
| 21:31.34 | cyn | afaik he comes outta nowhere gets on working on the qt key issue and is helping out |
| 21:32.20 | kergoth | heh, he's been around for a while, just not in the Z community, over in the ipaq stuffs |
| 21:32.24 | Disconnect | cyn: been working with pda-linux since the ipaq 3600 came out (and since fam was a bunch of read-only cramfs images ;) ..) |
| 21:32.47 | gb2 | kergoth: heh. i discovered some bugs in my sifting of files into packages. |
| 21:32.52 | gb2 | kergoth: figured out what it was |
| 21:32.52 | Disconnect | even rewrote a (working) installer for intimate, way back in the day. |
| 21:33.02 | CIA | repository=03packages user=11kergoth@direwolf.ppp.ti.com |
| 21:33.02 | CIA | 07ChangeSet@1.346 Add murasaki 0.7.10 (an alternative hotplug). |
| 21:33.11 | gb2 | i got halfway done moving them so that glibc and uclibc environments could coexist |
| 21:33.19 | Disconnect | um. how do I make the wiki leave the ~ in the url?? |
| 21:33.21 | gb2 | but didn't finish |
| 21:33.41 | gb2 | tried ~~ or \~ ? |
| 21:34.05 | Disconnect | ~~ worked, thanks |
| 21:34.07 | cyn | whoa, so is he becoming a full convert kergoth? |
| 21:34.14 | cyn | should we have like a birthday celebration |
| 21:34.27 | cyn | born again OZer |
| 21:35.22 | Disconnect | nah. just my new device. Once its usable I'll be in and out for a few years when issues get up my butt that I want fixed. ;) |
| 21:35.40 | cyn | did you get a c700 c750 or c760 |
| 21:35.46 | Disconnect | 760 of course |
| 21:35.50 | cyn | schweet |
| 21:35.57 | Disconnect | yah |
| 21:36.04 | cyn | I think I woulda gone 750 |
| 21:36.17 | gb2 | i like my 760 |
| 21:36.21 | gb2 | wish the battery lasted longer though |
| 21:36.23 | cyn | but I dunno coz the battery is like 119 or somethin |
| 21:36.26 | cyn | gb2: ! |
| 21:36.29 | Disconnect | we were expecting a bonus, so I started poking around (cuz the ipaqs are just annoying me these days) and didn't expect to have the cash for one.. |
| 21:36.41 | gb2 | cyn: with the wireless card it's dead in about 2-3 hours |
| 21:36.50 | cyn | yah well |
| 21:36.52 | cyn | :P |
| 21:36.55 | gb2 | i'm using a symbol card too |
| 21:37.01 | cyn | I wonder how well it'd do with a bluetooth card |
| 21:37.06 | gb2 | with the linksys one it's dead in about 90 minutes |
| 21:37.07 | Disconnect | gb2: which card? my symbol card lasts longer than that, unless I'm (eg) streaming mp3s with the backlight at full, etc.. |
| 21:37.11 | cyn | haha gb2 yeah |
| 21:37.32 | gb2 | disconnect: it's probably 3 hours. it's a socket branded one, blue sticker |
| 21:37.56 | Disconnect | ok this is annoying as hell. tar tfj cross-3.3.1.tar.bz2 works.. sudo tar xfj fails.. |
| 21:38.06 | gb2 | disconnect: it could probably be more agressive powersaving by doing active frequency scaling of the xscale |
| 21:38.11 | Disconnect | md5s match to the one I successfully unpacked. etc. I'll bbiab. |
| 21:38.20 | cyn | so shit I should prolly get a bluetooth card for xmas and help getting bluetooth happy usin either bluez or affix |
| 21:38.29 | Disconnect | gb2: blue sticker? this one is socket, orangeish.. |
| 21:38.33 | gb2 | hrm |
| 21:38.38 | gb2 | maybe this is symbol branded then |
| 21:38.40 | gb2 | it's in the other room |
| 21:38.43 | kergoth | i should get a bt cf card |
| 21:38.51 | gb2 | i should too |
| 21:38.52 | kergoth | have a bt usb dongle for my laptop and a bt cell |
| 21:38.55 | gb2 | my thinkpad has bt in it |
| 21:38.56 | kergoth | so may as well use it |
| 21:38.57 | cyn | same kergoth |
| 21:38.57 | kergoth | nice |
| 21:39.11 | Disconnect | kergoth: I'm still on the fence with bt. no bt devices, so nothing new I buy has bt, so I have no bt devices... etc |
| 21:39.20 | gb2 | i should also get one of those battery powered bt/rs232 dongles |
| 21:39.28 | cyn | haha when we had this power outage for days I setup bluetooth on my laptop using gprs with my c700 and t68i over infrared |
| 21:39.29 | gb2 | that'd be useful for debugging a few things |
| 21:39.40 | cyn | ... and a CF card to xfer to laptop |
| 21:40.03 | gb2 | cyn: where'd you get power for the electronics? ;) |
| 21:40.07 | cyn | local library |
| 21:40.10 | gb2 | heh |
| 21:40.22 | cyn | they got power back quick and were friggin PACKED every day |
| 21:40.25 | gb2 | my cell phone is great for wireless while traveling |
| 21:40.50 | cyn | I also had the options of: Barnes and Noble, wifes workplace (Hospital) |
| 21:41.05 | gb2 | disconnect: http://www.expansys.com/i/b/b101736.jpg |
| 21:41.08 | gb2 | this is the card i have |
| 21:41.18 | Disconnect | cyn: I just used the usb cable for my phone off a car inverter. also came in handy with the free b/w portable tv I'd gottten 2-3 weeks earlier (just for S&G I test-drove some car to get it) |
| 21:41.19 | cyn | wireless networker heh |
| 21:41.35 | Disconnect | ahh. thats a different card entirely. |
| 21:41.35 | cyn | yeah I should totally get a car inverter |
| 21:41.57 | cyn | Disconnect: does the usb cable charge || power the phone ? |
| 21:42.01 | gb2 | http://www.wireless-computer-networking.com/store/images/socket-CF-card.jpg |
| 21:42.09 | gb2 | it should be the same as this card |
| 21:42.15 | Disconnect | cyn: no, gotta disassemble it to do that. its just serial->usb/serial->usb |
| 21:42.24 | cyn | ah |
| 21:42.27 | gb2 | socket and symbol are the same company iirc |
| 21:42.33 | gb2 | or atleast, the same hardware |
| 21:43.07 | Disconnect | thats the card I have |
| 22:10.45 | Disconnect | does oe use distcc? |
| 22:12.45 | ljp | yes |
| 22:14.24 | Disconnect | hrm. cuz its not :( |
| 22:15.12 | *** join/#openembedded _stigger (~jeff@p027-n1-gos.nlc.net.au) |
| 22:16.58 | Disconnect | hrm. according to the log its only calling ccache |
| 22:17.21 | Disconnect | anything funny I need to do to it? |
| 22:18.21 | kergoth | oe or buildroot? |
| 22:18.29 | Disconnect | oe. |
| 22:18.37 | kergoth | ah, no it doesnt atm |
| 22:18.41 | Disconnect | ah ok. damn. |
| 22:18.51 | kergoth | echo TARGET_PREFIX_prepend="distcc " |
| 22:18.54 | kergoth | >> conf/local.conf |
| 22:19.26 | Disconnect | ah cool. |
| 22:22.19 | Disconnect | still only calling ccache. prepend or PREPEND? |
| 22:22.24 | kergoth | lowercase |
| 22:22.46 | kergoth | do a.. oebuild showdata blah.ow|grep _PREFIX |
| 22:22.50 | kergoth | make sure it took effect |
| 22:22.56 | kergoth | s/ow/oe/ |
| 22:23.28 | Disconnect | TARGET_PREFIX="distcc arm-linux-" .. maybe glibc just hates it |
| 22:23.41 | kergoth | did you clean and rerun it? |
| 22:23.47 | kergoth | it gets set at configure time, not compile time |
| 22:23.49 | Disconnect | yep |
| 22:23.51 | kergoth | for autoconf based builds |
| 22:23.52 | kergoth | k |
| 22:24.14 | kergoth | grep CC= in showdata, make sure distcc is there, ass it should be |
| 22:24.49 | Disconnect | rerunning configure real quick |
| 22:24.59 | kergoth | k |
| 22:25.37 | Pendalar | kergoth guy enters my linux help channel on efnet and : <eBUNNY> where is my files installed ? |
| 22:25.42 | Disconnect | export CC="ccache arm-linux-gcc" |
| 22:26.00 | kergoth | Pendalar: the appropriate answer is "up your ass" |
| 22:26.02 | Pendalar | I thought you might get a kick out of that |
| 22:26.04 | Pendalar | =-) |
| 22:26.19 | Disconnect | but its still showing the prefix properly. strange. |
| 22:26.32 | kergoth | hmmm that is odd. |
| 22:26.58 | kergoth | oh, i know what it is |
| 22:27.06 | kergoth | set HOST_PREFIX instead of TARGET_PREFIX |
| 22:27.44 | Disconnect | thats got it (according to showdata) |
| 22:27.49 | kergoth | cool |
| 22:27.57 | *** join/#openembedded ldspartan (~generp@alb-24-29-58-80.nycap.rr.com) |
| 22:28.21 | Disconnect | thanks :) now to rerun, see if it actually works (and if not, chances are its my problem) |
| 22:28.24 | *** part/#openembedded ldspartan (~generp@alb-24-29-58-80.nycap.rr.com) |
| 22:28.30 | kergoth | np |
| 22:28.33 | ljp | how about teambuilder? |
| 22:28.41 | Disconnect | hrm. buildroot is hanging on ncurses-term. been there a while now. |
| 22:28.50 | ljp | that takes a while |
| 22:29.05 | Disconnect | ok |
| 22:29.37 | kergoth | yeah, that ipk creation piece has a step where it iterates through the file list.. ncurses-term has a shitload of tiles |
| 22:29.40 | kergoth | s/tile/file/ |
| 22:30.41 | Disconnect | ah ok |
| 22:32.21 | Disconnect | strange. it seems to be calling it but the logs aren't showing it actually connecting. |
| 22:42.52 | ljp | somethings fucked up email here |
| 22:45.21 | ljp | cool, it's not just me |
| 22:48.57 | pb_ | not just you that fucked it up? I guess that's a relief. |
| 22:49.45 | ljp | yes. its way far away thats messed up, so we just have to wait |
| 22:54.33 | *** join/#openembedded Bernardo (~jbbrs@213.58.96.193) |
| 22:54.37 | Bernardo | hi |
| 22:56.16 | mattaw | mickeyl, hey could you mail me your def-config for the 5600 build? I am reconstructing my buildroot PROPERLY this time |
| 22:56.35 | mickeyl | already in buildroot |
| 22:56.46 | mickeyl | but hold on... |
| 22:56.53 | mickeyl | while we're there |
| 22:56.59 | mickeyl | the kernel configs need to be altered |
| 22:57.07 | mickeyl | not to include the sucky orinoco 0.12 |
| 22:57.14 | mickeyl | i'll push that soon |
| 22:58.18 | Bernardo | bye |
| 22:58.21 | *** part/#openembedded Bernardo (~jbbrs@213.58.96.193) |
| 22:59.13 | mattaw | you used the poodle-opie? |
| 22:59.20 | mickeyl | no |
| 22:59.25 | mickeyl | def-configs/3.3.5-PXA- |
| 23:00.08 | Twiun | g'night guys |
| 23:00.14 | mickeyl | n8 Twiun |
| 23:00.17 | mattaw | I am feeling at serious risk of being stupid but that isn't in my buildroot |
| 23:01.04 | mickeyl | maybe do a bk pull? |
| 23:01.12 | mattaw | this is fresh today |
| 23:01.25 | mickeyl | today is relative |
| 23:01.29 | mattaw | I destroyed my buildroot when I ditched gentoo for fedora |
| 23:02.10 | mattaw | some months ago - and its your fault I am looking at it again. |
| 23:02.20 | mattaw | Not sure if I should thank you yet :) |
| 23:04.38 | mickeyl | hehe |
| 23:04.40 | mickeyl | well... |
| 23:05.22 | mickeyl | I'll never do release for so many models togehter |
| 23:05.24 | mickeyl | it's killing me |
| 23:05.38 | mattaw | have you been looking at the lists? |
| 23:05.48 | mickeyl | email? |
| 23:05.52 | mattaw | yes |
| 23:05.55 | kergoth | heh, no worries. i'll manage the 3.5.1 release |
| 23:05.57 | mickeyl | mostly |
| 23:06.05 | mickeyl | kergoth: hehe, good. |
| 23:06.17 | mickeyl | however you can't really test all models, or can you? |
| 23:06.25 | mattaw | well like when I released the sl5600 stuff I did the first thing I got was a complaint |
| 23:06.30 | kergoth | i have a 5500, 5600, c700, and c750 |
| 23:06.38 | mattaw | which made me a little pissed as I recall |
| 23:06.39 | kergoth | which covers most our bases |
| 23:06.40 | mickeyl | okay, i take that back :D |
| 23:06.56 | mickeyl | mattaw: no kidding, can imagine. |
| 23:06.56 | mattaw | kergoth, your favourite? |
| 23:06.58 | mickeyl | but hey... that's normal |
| 23:07.00 | mickeyl | USERS |
| 23:07.03 | mickeyl | always complain |
| 23:07.14 | mattaw | indeed. I complained at first too. |
| 23:07.19 | kergoth | mattaw: any c7x0 is the obvious choice, c750 more so due to the pxa255 |
| 23:07.29 | kergoth | noticable performance difference compared tot he 5600 or c700 |
| 23:07.34 | mattaw | I was going to say those must me pretty speedy |
| 23:08.02 | mattaw | well a friend who works for arm UK says they were crippled with laughter when they found out what intel had done |
| 23:08.32 | kergoth | hehe, i bet |
| 23:08.49 | mattaw | since intel bought the design with a one off payment arm saw no need to let them know their mistake |
| 23:10.05 | mickeyl | hmm |
| 23:10.07 | mickeyl | bkbits.net ? |
| 23:10.42 | mattaw | I seem to be getting no response to my attempts to build this system |
| 23:10.52 | mattaw | sorry to my pull from bk |
| 23:11.09 | mickeyl | no connection here either |
| 23:11.51 | *** join/#openembedded maxik (~maxik@82.140.72.252) |
| 23:12.09 | *** part/#openembedded maxik (~maxik@82.140.72.252) |
| 23:13.28 | mattaw | oh well I should be writing a paper anyhow. |
| 23:13.44 | mattaw | goes out on 3pm monday for 11pm local time hongkong. |
| 23:16.53 | *** join/#openembedded cyn (~cyn@adsl-155-129-59.mem.bellsouth.net) |
| 23:19.31 | _Psycho | ljp you around ? |
| 23:21.25 | *** join/#openembedded Disconnect (~dis@arent.you.special.you.gotontheinter.net) |
| 23:24.23 | mickeyl | ibot: pester treke to commit c7x0 odevice rotation |
| 23:24.25 | | treke to commit c7x0 odevice rotation: Are we there yet? .. Are we there yet? .. Are we there yet? |
| 23:25.19 | _Psycho | ibot: pester mickeyl to pester treke to commit c7x0 odevice rotation |
| 23:25.25 | | mickeyl to pester treke to commit c7x0 odevice rotation: Are we there yet? .. Are we there yet? .. Are we there yet? |
| 23:25.29 | *** join/#openembedded maxik (~maxik@82.140.72.252) |
| 23:27.34 | *** join/#openembedded cyn (~cyn@adsl-155-129-59.mem.bellsouth.net) |
| 23:31.54 | mickeyl | g'night |
| 23:49.35 | *** join/#openembedded vvff (~victor@ip229.100-173-207.eli-du.nwlink.com) |
| 23:49.37 | *** join/#openembedded icefox (~ben@ool-18b87f53.dyn.optonline.net) |
