| 00:01.25 | mikelp | I refuse to fix bad code like that |
| 00:05.39 | mikelp | Hmmm, libtool-cross death |
| 00:07.27 | kergoth | gah, got bit by one bug in the silicon while implementing the workaround for another |
| 00:11.59 | treke | hehe |
| 00:14.26 | *** join/#oe hugojosefson (~hugojosef@1-1-3-41a.msp.mlm.bostream.se) |
| 00:18.12 | mikelp | cool, it died on the athlon too |
| 00:24.08 | hugojosefson | trying to compile modutils-cross [fails!], do i need to compile something else first? |
| 00:26.27 | kergoth | hugojosefson: are you using oemake or oebuild? |
| 00:26.47 | hugojosefson | oemake |
| 00:26.59 | kergoth | then oe is already obeying dependencies. |
| 00:27.03 | kergoth | it builds what needs to be built first, first. |
| 00:29.53 | mikelp | gar: checking for x86_64-linux-gcc... ccache x86_64-linux-gcc |
| 00:30.12 | mikelp | what's the env var to override this nonsense, (to save me hunting it down ) |
| 00:30.43 | *** join/#oe ljp (~lpotter@203.94.178.46) |
| 00:31.12 | hugojosefson | okay... do you know if there are any other problems with the modutils-cross package |
| 00:31.25 | hugojosefson | or is it supposed to compile? |
| 00:31.56 | kergoth | it works fine for everyone else. |
| 00:32.04 | kergoth | how about you actually give us some real information? |
| 00:32.09 | kergoth | like.. say.. an error message, perhaps. |
| 00:32.12 | kergoth | read the log. |
| 00:32.38 | *** join/#oe thatweasel (~thatwease@pool-68-237-251-88.ny325.east.verizon.net) |
| 00:32.40 | kergoth | mikelp: you want it to just run 'gcc'? TARGET_PREFIX is most likely what you're looking for, if thats the case |
| 00:32.59 | mikelp | just trying to get the new build system up and running |
| 00:33.22 | hugojosefson | kergoth: yeah, sorry. -- obj_kallsyms.c: In function `obj_kallsyms' |
| 00:33.32 | hugojosefson | obj_kallsyms.c:204: error: invalid lvalue in assignment |
| 00:37.22 | hugojosefson | same error for line 279 |
| 00:38.46 | hugojosefson | this is the command that resulted in the error: |
| 00:38.50 | hugojosefson | ccache gcc -I/var/code/oe/build/tmp/staging/i686-linux/include -O2 -I/var/code/oe/build/tmp/work/modutils-cross-2.4.27-r3/modutils-2.4.27/obj/../include -D_GNU_SOURCE -DPACKAGE_NAME=\"\" -DPACKAGE_TARNAME=\"\" -DPACKAGE_VERSION=\"\" -DPACKAGE_STRING=\"\" -DPACKAGE_BUGREPORT=\"\" -DCONFIG_ROOT_CHECK_OFF=0 -DELF_MACHINE_H='"elf_arm.h"' -DARCH_arm -c -o obj_kallsyms.o obj_kallsyms.c |
| 00:51.39 | kergoth | grr |
| 01:00.19 | hugojosefson | the offending line 204 of obj_kallsyms.c is this: |
| 01:00.20 | hugojosefson | <PROTECTED> |
| 01:01.45 | hugojosefson | when i put an extra pair parenthesis like this the errormessage disappears: |
| 01:01.46 | hugojosefson | <PROTECTED> |
| 01:02.17 | mikelp | that look scarily wrong |
| 01:02.45 | hugojosefson | thanks, that's the kind of comment i was looking for! |
| 01:03.17 | *** join/#oe theig (~kyle@adsl-64-164-37-209.dsl.scrm01.pacbell.net) |
| 01:03.31 | mikelp | (osec->contents) would scare me less |
| 01:03.49 | *** join/#oe LordVan (~lordvan@LordVan.developer.gentoo) |
| 01:04.35 | hugojosefson | my compiler disagrees |
| 01:05.19 | mikelp | although modutils is rusty code, could be deep magic |
| 01:05.36 | chouimat|afk | night all |
| 01:05.54 | hugojosefson | weird thing is it seems to work for everyone else |
| 01:06.04 | hugojosefson | night chouimat|afk |
| 01:06.46 | hugojosefson | mikelp: what version gcc are you using? |
| 01:07.41 | mikelp | which one would you like ;) 3.3.2, x86_64->arm-cross-3.4, csl-3.4, |
| 01:09.15 | mikelp | The build system is hurting me right now, I'll be able to test modutils once I get this working |
| 01:09.26 | hugojosefson | i was told all -cross packages and -native packages are compiled with your normal gcc (?) |
| 01:09.41 | mikelp | right, x86-64-linux-unknown-gcc here |
| 01:10.07 | mikelp | so all bets are off when trying to compare with i686, too many possible variations |
| 01:11.21 | hugojosefson | i have been trying to get my build system to work for a while now. i really want to be able to do an `oemake opie-image` or similar... |
| 01:12.44 | hugojosefson | quite a lot of problems were solved by upgrading to the very bleeding latest versions of the tools listed in RequiredSoftware |
| 01:13.51 | hugojosefson | (and some by building some -initial packages...) but do you have any other suggestions where i should look? |
| 01:14.46 | mikelp | take a look at the run.do_compile_NNNN file, that'll let you know exactly what the build system tried to do |
| 01:15.11 | mikelp | I then try to emulate that interactively to find the problem |
| 01:16.30 | hugojosefson | okay, here goes: |
| 01:16.32 | hugojosefson | make[1]: Entering directory `/var/code/oe/build/tmp/work/modutils-cross-2.4.27-r3/modutils-2.4.27/obj' |
| 01:17.12 | hugojosefson | then it ran the long ccache gcc command i posted before |
| 01:17.46 | hugojosefson | In file included from obj_kallsyms.c:26: |
| 01:17.46 | hugojosefson | /var/code/oe/build/tmp/work/modutils-cross-2.4.27-r3/modutils-2.4.27/obj/../include/util.h:42: warning: built-in function 'log' declared as non-function |
| 01:17.54 | mikelp | oooh |
| 01:17.58 | mikelp | .h file fubarage |
| 01:19.01 | hugojosefson | then the rest about obj_kallsyms.c:204: error: invalid lvalue in assignment... |
| 01:19.21 | *** join/#oe ljp_ (~lpotter@ppp189.adsl238.pacific.net.au) |
| 01:20.05 | *** join/#oe synth (~cb@pcp858839pcs.ptchar01.fl.comcast.net) |
| 01:20.24 | hugojosefson | util.h:42 is just this: |
| 01:20.27 | hugojosefson | <PROTECTED> |
| 01:21.05 | mikelp | which, reading the above, something thinks should be a function |
| 01:21.23 | hugojosefson | sounds like i should refactor it... |
| 01:21.39 | hugojosefson | but then again, it works for evry1 else! (right?) |
| 01:21.42 | mikelp | my build is trying to run right now |
| 01:21.53 | mikelp | 99/100 its something screwy with your build env |
| 01:22.12 | hugojosefson | tell me about it :) |
| 01:25.39 | hugojosefson | mikelp: i really appreciate you looking at this w/ me |
| 01:25.59 | mikelp | no worries, I understand how frustrating this can be |
| 01:26.17 | hugojosefson | ...it just HAS to be something or other w/ my build env |
| 01:26.52 | hugojosefson | you running linux? |
| 01:27.24 | *** join/#oe mithro (~tim@dsl1-83.gw1.adl1.airnet.com.au) |
| 01:27.24 | mikelp | yep, but my build stuff is on the opteron, so its a little weird |
| 01:27.47 | hugojosefson | what distro? |
| 01:27.51 | mikelp | suse 9.1 |
| 01:28.47 | *** join/#oe hufnus (~slonsiki@DSL135-071.labridge.com) |
| 01:29.47 | treke|home | kergoth: still alive? |
| 01:29.52 | hugojosefson | well, i'm using a rather straightforward debian setup with a little extra new packages for the build env. |
| 01:29.58 | kergoth | treke|home: yep. |
| 01:30.04 | kergoth | more so than i was before, thanks to starbucks |
| 01:30.14 | treke|home | kergoth: The scotches you recommended were a little out of my price range :p |
| 01:30.19 | kergoth | hehe |
| 01:30.22 | kergoth | yeah, that happens |
| 01:30.23 | treke|home | kergoth: Not looking to spend 100 bucks on a bottle :p |
| 01:30.34 | mikelp | scotch bof !! |
| 01:30.35 | kergoth | heh, i only spent like 60 or 70 on a 14yr oban |
| 01:30.48 | mikelp | that stuff wli bought at ols was $95 CAD / bottle |
| 01:31.07 | kergoth | guy behind the counter was all "are you sure you're enough of a connesseur to spend that?" .. i'm like shit, you've got far more expensive stuff than that back there |
| 01:31.10 | kergoth | stupid people |
| 01:31.20 | kergoth | the lagovulin(sp?)? |
| 01:31.22 | kergoth | that was pretty good |
| 01:31.26 | mikelp | yeah |
| 01:31.37 | mikelp | that's damn good if you like the peaty scotches |
| 01:32.11 | *** join/#oe synth (~cb@pcp858839pcs.ptchar01.fl.comcast.net) |
| 01:33.09 | treke|home | but I also like being able to pay my rent |
| 01:33.26 | kergoth | bah.. paying rent before your scotch? gotta fix those priorities man |
| 01:33.27 | kergoth | ;) |
| 01:33.29 | mikelp | rent is overrated |
| 01:33.41 | treke|home | gotta have some place to drink it :p |
| 01:34.03 | mikelp | park bench is all you need |
| 01:34.13 | kergoth | i was thinking cardboard box, but that works too |
| 01:34.32 | gb2 | get a cray |
| 01:34.36 | treke|home | maybe another time. They appeared to have a nice selection. should keep me busy for a while |
| 01:34.36 | gb2 | live inside it |
| 01:34.42 | gb2 | when it gets cold in the winter, turn it on |
| 01:36.08 | treke|home | hehe |
| 01:36.33 | mikelp | treke|home: don't get me started on Apple and my ordered two months ago G5 |
| 01:36.44 | treke|home | I hate having new equipment, but not having a cheap but necessary accessory |
| 01:38.06 | treke|home | new g5? |
| 01:38.09 | treke|home | yummy |
| 01:38.18 | mikelp | yeah, the water cooled beast |
| 01:43.57 | gb2 | kergoth: i think i might have solved the signal thing |
| 01:44.03 | gb2 | include <asm/sigcontext.h> |
| 01:44.14 | gb2 | instead of <signal.h> and <sys/ucontext.h> |
| 01:44.20 | mikelp | lalala, glibc building - slooooow |
| 01:44.26 | gb2 | untested |
| 01:44.29 | kergoth | ah, its able to build that block of code when you do that? |
| 01:44.30 | kergoth | ah |
| 01:44.30 | kergoth | heh |
| 01:44.37 | gb2 | it was the brilliant idea i had before leaving work though |
| 01:45.27 | gb2 | needs edits to gcc/config/*/linux.h though |
| 01:49.11 | hugojosefson | does anybody here know of anyone who has... successfully :)... set up a build env for oe on debian? |
| 01:49.16 | gb2 | hmm |
| 01:49.23 | kergoth | i use oe on debian boxes on a daily basis. |
| 01:50.10 | mikelp | which debian though, wooody/testing/sid ? |
| 01:50.20 | kergoth | all of mine are unstable. |
| 01:51.09 | hugojosefson | i use testing, |
| 01:51.13 | mikelp | same here (well, except the firewall, but no dev on that) |
| 01:51.14 | hugojosefson | but gcc and other related packages (everything from RequeiredSoftware) i have apt-pinned to unstable or experimental |
| 01:51.39 | gb2 | kergoth: doing a quick test on that fix now.. |
| 01:51.43 | kergoth | cool |
| 01:51.45 | gb2 | hm. |
| 01:51.56 | gb2 | ofcourse.. what headers is it going to even look in |
| 01:52.26 | mikelp | gcc-cross-3.4.1 |
| 01:52.29 | mikelp | now I'm scared |
| 01:52.39 | gb2 | i also need to move forward |
| 01:52.43 | gb2 | i'm testing with 3.4.0 |
| 01:53.45 | hugojosefson | i have gone all the way to gcc-3.5. you live only once, right? |
| 01:54.07 | *** join/#oe evanchsa (~evanchsa@h00095b6e0a5f.ne.client2.attbi.com) |
| 01:54.11 | ljp | wonder if zecke had the time to test that qte patch yet |
| 01:54.45 | mikelp | hugojosefson: until you find a gcc bug ;) |
| 01:54.57 | mikelp | which are *bastards* to trace and solve |
| 01:55.25 | hugojosefson | mikelp: solution is to DOWNGRADE ;) |
| 01:55.50 | mikelp | nah, solution is to get the baseball bat out for the gcc people |
| 01:56.09 | hugojosefson | that'll encourage them |
| 01:56.17 | kergoth | heh, if you use a dev branch of gcc, and it breaks, and you knowingly used it lacking the ability to diagnose, you deserve the hell you get |
| 01:56.52 | *** join/#oe theig (~kyle@adsl-64-164-37-209.dsl.scrm01.pacbell.net) |
| 01:58.13 | mikelp | Or you're the one person to trigger an obscure corner case, causing gcc to loop forever :D |
| 01:59.12 | kergoth | hehe |
| 02:00.00 | mikelp | grrr, bzflag tgz builds fine, cvs is borked |
| 02:01.30 | hugojosefson | kergoth: okay, i gave up. it's back to gcc-3.4 now... |
| 02:02.07 | mikelp | It works with gcc 3.4 ? |
| 02:02.25 | treke|home | kergoth: booted linux on my 2200 today |
| 02:02.48 | hugojosefson | trying a build as we speak^H^H^H^H^Htype... |
| 02:05.27 | *** join/#oe hclewz (~discrete7@fluffy.superlucidity.net) |
| 02:05.54 | mikelp | wtf, building libxml2 really early in the chain |
| 02:06.59 | gb2 | NOTE: package gcc-cross-initial-3.4.0-r1: task do_compile started |
| 02:06.59 | gb2 | ERROR: expansion of ${@autotools_set_crosscompiling} returned non-string <function autotools_set_crosscompiling at 0xeed4764c> |
| 02:07.06 | gb2 | lesse if this works |
| 02:07.40 | kergoth | heh, that whole force of autotools to crosscompile thing was great in theory, but we forgot about packages like binutils, for whom we cant call autoreconf, so its stuck with old configure scripts that dont support that syntax |
| 02:11.01 | gb2 | kergoth: lame. lose. |
| 02:11.13 | gb2 | rt_sigframe has a siginfo and a ucontext in it |
| 02:11.21 | gb2 | and those aren't defined in any of this linux-libc-headers |
| 02:12.16 | gb2 | so the inhibit libc thing may be the only solution |
| 02:13.01 | kergoth | shit. that sucks. maybe we should just keep the -initial libc's around |
| 02:13.07 | kergoth | as annoying as it is |
| 02:13.14 | gb2 | fails on a bunch of arches |
| 02:13.18 | gb2 | because glibc sucks ass |
| 02:13.37 | gb2 | glibc requires configure be run to generate headers |
| 02:13.37 | mikelp | gb2: different host or target arches ? |
| 02:13.47 | gb2 | which requires a working cross compiler |
| 02:14.00 | gb2 | chicken |
| 02:14.00 | gb2 | egg |
| 02:15.05 | kergoth | heh, we hack it |
| 02:15.09 | kergoth | sed out the exits in glibc's configure |
| 02:15.17 | kergoth | ugly, messy |
| 02:15.29 | gb2 | kergoth: try a mipsel glibc build someday |
| 02:16.01 | mikelp | mipsel ublibc works, can we steal shit from it ? |
| 02:16.22 | gb2 | well |
| 02:16.37 | gb2 | i guess this is a different bug |
| 02:16.46 | gb2 | it adds -m32 to CFLAGS |
| 02:17.03 | gb2 | and calls $(CC) $(CFLAGS) -E as the preprocessor |
| 02:17.14 | gb2 | which is a bit of a suck. |
| 02:17.26 | gb2 | and... you really want the cross preproc if you're preproc'ing something |
| 02:18.26 | hugojosefson | mikelp: that baseball bat of yours may be useful after all... the gcc people are gradually making the modutils-cross package incompilable. i think i found the specific problem... |
| 02:19.40 | kergoth | hugojosefson: the gcc folks break compiles on a regular basis. in their defense, they're usually breaking people that were doing bad things anyway |
| 02:19.56 | gb2 | 3-stage cross toolchain is probably the right move |
| 02:20.03 | hugojosefson | type casted lvalues are deprecated and warned about in gcc-3.4, and in gcc-3.5 they are treated as flat out errors |
| 02:20.17 | mikelp | make sense |
| 02:20.25 | kergoth | yep, nothing wrong with that |
| 02:20.27 | mikelp | that double assign is just ugly |
| 02:20.38 | kergoth | gb2: heh, are you talking 3 passes of gcc? |
| 02:20.57 | kergoth | jesus, for a glibc nptl build, you'd have binutils-cross, then 3 passes of glibc and 3 passes of gcc |
| 02:21.01 | mikelp | hugojosefson, oooo, that's why the ( ... ) |
| 02:21.03 | mikelp | worked |
| 02:21.11 | mikelp | made gcc treat the cast as an rvalue |
| 02:21.51 | gb2 | gcc-initial(inhibit_libc), glibc-first-pass, gcc-firstpass, glibc-second-pass, gcc-cross, (glibc-nptl) |
| 02:22.10 | gb2 | more compiling! |
| 02:22.28 | mikelp | Bah, let's add gcc-native too |
| 02:22.39 | gb2 | heh |
| 02:22.40 | mikelp | bootstrap from /bin/bash and wget |
| 02:22.42 | hugojosefson | mikelp: now i see how right you were :D |
| 02:22.48 | kergoth | heh, make oe's only deps wget and python |
| 02:23.00 | kergoth | let it construct a rootfs, boot into it with uml or chroot, then start a build there |
| 02:23.05 | gb2 | can't you do a getter in python? |
| 02:23.10 | gb2 | why rely on wget? |
| 02:23.12 | gb2 | ;) |
| 02:23.34 | kergoth | heh, well, we were going to use libcurl eventually anyway.. somebody write python bindings for it :P |
| 02:23.36 | mikelp | hugojosefson: on an encouraging note, that's going to piss Keith off to no end |
| 02:24.16 | hugojosefson | keith? |
| 02:24.39 | mikelp | owens, modutils guy |
| 02:25.28 | hugojosefson | yea, that's what i thought about him too... |
| 02:35.05 | *** join/#oe rnoh-htat (~rnoh-htat@69-162-196-251.bflony.adelphia.net) |
| 02:35.57 | hugojosefson | i have started a fresh new oemake opie-image w/ gcc-3.4, and with any luck it will be finished in the morning... |
| 02:39.12 | mikelp | hmmm, wtf is unik-olsrd, and do I care ? |
| 02:39.51 | kergoth | iirc its a mesh thing. isnt there a DESCRIPTION in the .oe? |
| 02:40.18 | mikelp | indeed |
| 02:40.34 | mikelp | I think its building way, way too early in the chain |
| 02:40.38 | *** part/#oe hugojosefson (~hugojosef@1-1-3-41a.msp.mlm.bostream.se) |
| 02:40.48 | kergoth | what are you building? |
| 02:40.52 | kergoth | with oemake i mean |
| 02:40.53 | mikelp | oemake world :) |
| 02:40.56 | kergoth | ah |
| 02:41.04 | kergoth | that takes a seriously long fucking time |
| 02:41.24 | mikelp | Well, I need to test the cooling on the opteron anyway |
| 02:41.28 | kergoth | hehe |
| 02:42.13 | kergoth | gb2: mipsel glibc build is now building glibc.. no problems so far. |
| 02:43.37 | gb2 | kergoth: hm |
| 02:43.38 | gb2 | interesting |
| 02:43.41 | gb2 | glibc_cvs ? |
| 02:43.47 | gb2 | they might have fixed it |
| 02:43.56 | kergoth | glibc-2.3.2+cvs20040726 |
| 02:44.01 | kergoth | pb_'s arbitrary locked down cvs date |
| 02:44.08 | kergoth | which is our preferred version nowadays |
| 02:44.16 | gb2 | using a glibc-initial ? |
| 02:44.19 | kergoth | yep |
| 02:44.30 | gb2 | ok, i've been using late june |
| 02:44.47 | gb2 | mark: no stifmeister i guess |
| 02:44.52 | gb2 | oops, wrong window |
| 02:44.56 | kergoth | err, actually, it didnt use a libc initial |
| 02:44.59 | kergoth | i forgot about that |
| 02:45.09 | kergoth | i excluded arm and mipsel from enabling the initial dependency |
| 02:45.16 | gb2 | heh |
| 02:45.24 | kergoth | since they're known to build fine without it |
| 02:46.16 | gb2 | heh |
| 02:46.33 | kergoth | course if the initial package builds after the full one, due to, say, a world build |
| 02:46.34 | gb2 | kergoth: the inhibit_libc define patch trick will work |
| 02:46.38 | kergoth | the shit will hit the fan rather spectacularly |
| 02:46.57 | gb2 | to make it build |
| 02:47.12 | gb2 | it might misbuild glibc though |
| 02:47.17 | kergoth | gb2: well, that email seems to imply that would be Bad due to the implications in the build of the libc |
| 02:47.22 | kergoth | heh |
| 02:47.35 | gb2 | which is why i suggested a few extra passes |
| 02:48.24 | kergoth | yeah, that would work. but shit, i'd rather leave it as is, than add extra passes to the build, slowing us down even more |
| 02:48.58 | gb2 | kergoth: that is one interesting thing though.. it's a userspace/kernel ABI which isn't documented by the userspace exposed kernel headers |
| 02:50.56 | mikelp | out of date userspace headers ? |
| 02:51.01 | gb2 | no |
| 02:51.05 | gb2 | kernel headers |
| 02:51.08 | gb2 | lack the bits |
| 02:52.56 | mikelp | wild |
| 02:54.27 | mikelp | <PROTECTED> |
| 02:54.56 | kergoth | heh, ow |
| 02:55.06 | mikelp | libpng |
| 02:56.19 | mikelp | In staging/arm-linux/lib/libc.so.6 |
| 02:56.48 | mikelp | Oh crap, and a lot depends on libpng too |
| 02:59.13 | *** join/#oe hufnus (~slonsiki@DSL135-071.labridge.com) |
| 03:04.21 | *** join/#oe eck (foobar@p5487D0C3.dip.t-dialin.net) |
| 03:06.48 | *** join/#oe LordVan (~lordvan@LordVan.developer.gentoo) |
| 03:08.22 | mikelp | Grrr, that's annoying |
| 03:08.35 | mikelp | Using the wrong arm-linux-gcc |
| 03:09.38 | mikelp | It's finding the one in /opt/cross rather than in /tmp/cross |
| 03:09.53 | kergoth | thats odd... it should always prepend its own paths to the system path, not append.. |
| 03:09.58 | kergoth | unless your CROSS_DIR is /opt/cross |
| 03:09.59 | kergoth | or something |
| 03:11.18 | mikelp | Hmmm, I may have done something very stupid too |
| 03:11.30 | kergoth | heh, thats always a possibility. hate that |
| 03:11.47 | mikelp | oh well, rm -Rf ./tmp |
| 03:12.44 | mikelp | Right, do not take your anger out in local.conf: |
| 03:12.45 | mikelp | <PROTECTED> |
| 03:12.45 | mikelp | arm-linux-gcc-3.4.1 bolloxc++filt bolloxld bolloxreadelf |
| 03:12.45 | mikelp | bolloxaddr2line bolloxcpp bolloxnm bolloxsize |
| 03:13.16 | kergoth | hehe |
| 03:15.32 | orpheus | makelp: tell me if that fouls things up for you next time you run 'bk pull' |
| 03:16.01 | orpheus | also: you may want to toss in -k so it won't error out of the entire build as soon as something doesn't work |
| 03:16.18 | kergoth | orpheus: works fine here. |
| 03:16.40 | kergoth | course i have a script that checks what .oe files were changed upstream, then i use oemake to clean those packages for the next time i run oemake world |
| 03:16.45 | mikelp | shouldn't affect bk pull - totally different directories |
| 03:16.45 | kergoth | not full proof, but works well enough |
| 03:16.53 | kergoth | yeah |
| 03:17.08 | orpheus | it shouldn't... but i seem to be the one to have problems |
| 03:17.28 | orpheus | is that script sitting around somewhere? |
| 03:17.34 | *** join/#oe joshin (~joshin@VDSL-130-13-185-112.PHNX.QWEST.NET) |
| 03:17.40 | mikelp | bk gets very pissy if you start changing code on it and don't tell it about it |
| 03:17.51 | kergoth | not really. its only necessary if you want to *clean* the packages that were changed upstream. |
| 03:18.02 | kergoth | and you only need to do that, if you're continuing an existing oemake world after doing a pull |
| 03:18.10 | kergoth | and want to see the changes. |
| 03:18.28 | orpheus | that makes sense |
| 03:18.59 | orpheus | kergoth: was it you i was asking about aqpkg and symlinks? |
| 03:19.05 | kergoth | yep |
| 03:19.25 | *** join/#oe trthatguy (~thereal_t@CARPENOCTEM.MIT.EDU) |
| 03:19.27 | orpheus | i finally got it to tell me what it was doing wrong |
| 03:19.46 | trthatguy | anyone here happen to work with PICs at all? I got a warp 13a programmer a little while ago which appears to be DOA. I'm checking the voltages mentioned in the help file and it looks as though all power supplies are fscked and I can't even get the firmware version(get a comms error). earlier I could at least get the firmware version and ping the programmer, but now it seems to be completely foobar |
| 03:19.49 | orpheus | trying to make symlinks under /mnt/card/mnt/card instead of / |
| 03:20.03 | trthatguy | if anyone good with transistor circuits can divine from the schematic here: http://www.newfoundelectronics.com/warp-13%20repair.htm what I should check for problems, I'd be most grateful |
| 03:20.23 | *** join/#oe joshin (~joshin@VDSL-130-13-185-112.PHNX.QWEST.NET) |
| 03:20.40 | kergoth | trthatguy: i think the guys over in #elinux are more adept with hardware than those of us here. |
| 03:20.48 | orpheus | i tried to make /mnt/card/mnt/card link to /, and now aqpkg won't run |
| 03:20.56 | kergoth | this channel is for the OE project, not random electronics discussion |
| 03:20.57 | kergoth | heh |
| 03:21.13 | orpheus | so i guess i need to go bother the people in #opie some more |
| 03:21.15 | kergoth | orpheus: sounds like you had an opie library or app out there that aqpkg needed. |
| 03:21.24 | kergoth | so linking that back interfered with its operation |
| 03:21.38 | trthatguy | kergoth: sorry, took wild guess someone in this channel might be familiar. |
| 03:22.12 | kergoth | will probably have better luck in #elinux. thats more random embedded linux discussion, with some hardware/electronics stuff thrown in for good measure |
| 03:22.58 | orpheus | i still wanna know where the heck it got /mnt/card/mnt/card from tho... it's too bad it doesn't exist |
| 03:23.16 | kergoth | heh |
| 03:23.18 | kergoth | that is odd |
| 03:24.48 | orpheus | ya know how there's that quote about how if something seems idiot-proof, you just need a better idiot? i think i'm that idiot, 'cause i seem to have the wierdest problems |
| 03:25.10 | orpheus | do you install software to a destination other than root? |
| 03:25.20 | kergoth | i dont *use* my Z |
| 03:25.25 | kergoth | and this discussion belongs in #openzaurus |
| 03:26.02 | orpheus | i'll go ask and see if i'm alone with my odd issues then |
| 03:34.51 | orpheus | i guess this is the only place i ask 'cause #opie and #openzaurus are both dead all the time |
| 03:34.57 | *** join/#oe _Bala_ (~Bala@hsdbsk142-165-143-145.sasknet.sk.ca) |
| 03:40.39 | *** join/#oe synth (~cb@pcp858839pcs.ptchar01.fl.comcast.net) |
| 03:45.59 | *** join/#oe synth (~cb@pcp858839pcs.ptchar01.fl.comcast.net) |
| 03:48.00 | *** join/#oe synth (~cb@pcp858839pcs.ptchar01.fl.comcast.net) |
| 04:01.46 | CIA-4 | 03kergoth 07 * r1.2554 10packages/glibc/glibc-initial_2.3.2.oe: Unbork the glibc-initial set of FILESPATH. |
| 04:01.47 | CIA-4 | 03kergoth 07 * r1.2553 10packages/classes/base.oeclass: Updates to base_set_filespath syntax, so that .oe's that include others can call it, thereby ensuring that they're able to locate overrides specific files. |
| 04:01.47 | CIA-4 | 03kergoth 07 * r1.2552 10packages/ (8 files in 4 dirs): Resurrect initial libc pass, for the moment. |
| 04:05.21 | *** join/#oe Matthew_I (~matthew@pcp01931171pcs.huntsv01.al.comcast.net) |
| 04:05.38 | *** part/#oe Matthew_I (~matthew@pcp01931171pcs.huntsv01.al.comcast.net) |
| 04:09.35 | *** join/#oe Matthew_I (~matthew@pcp01931171pcs.huntsv01.al.comcast.net) |
| 04:09.44 | Matthew_I | how can I build my own oe images |
| 04:09.58 | kergoth | you start by reading. |
| 04:10.40 | Matthew_I | I have been reading the wiki |
| 04:10.44 | Matthew_I | it's not very clear |
| 04:10.48 | kergoth | GettingStarted. |
| 04:10.54 | kergoth | whats unclear about step by step instructions? |
| 04:10.57 | Matthew_I | I have the source |
| 04:11.14 | Matthew_I | man you have always been kind of non friendly... |
| 04:11.19 | *** join/#oe synth (~cb@pcp858839pcs.ptchar01.fl.comcast.net) |
| 04:11.45 | Matthew_I | I don't blame you though |
| 04:11.47 | kergoth | its a valid question. if there's a problem with the instructions, i need to know what it is. |
| 04:11.50 | kergoth | heh |
| 04:12.11 | kergoth | i just dont put up a nice front. i tell it like it is, like it or not |
| 04:12.11 | kergoth | heh |
| 04:12.25 | Matthew_I | well for one I don't know how to get the build and sources directories, I only found oe and packages |
| 04:12.36 | Matthew_I | I don't know bk well enough to get the others |
| 04:12.38 | kergoth | set DL_DIR. |
| 04:12.42 | kergoth | oe creates it when it needs it. |
| 04:12.43 | kergoth | (sources) |
| 04:12.48 | kergoth | *you* create the build directory |
| 04:12.53 | kergoth | just as you create the conf directory inside |
| 04:12.59 | kergoth | and the local.conf, based on local.conf.sample |
| 04:13.08 | Matthew_I | I have oe and packages |
| 04:13.15 | kergoth | yes, i heard you the first time |
| 04:13.26 | Matthew_I | :) |
| 04:13.33 | kergoth | mkdir build; cd build; mkdir conf; cp ../packages/conf/local.conf.sample conf/local.conf; .. edit .. |
| 04:13.44 | kergoth | its that simple. |
| 04:13.47 | kergoth | then run the tools inside build. |
| 04:13.58 | kergoth | i have 6 or 7 build directories, for different target platforms |
| 04:14.02 | Matthew_I | ok, ill try the other stuff on this getting started page |
| 04:14.11 | Matthew_I | this oe stuff is pretty sweet |
| 04:14.27 | Matthew_I | u have any time frame for when the next oz is comming out? |
| 04:14.31 | kergoth | there's a bit of a learning curve, mostly because its so flexible, so there are few hard & fast rules you *must* follow |
| 04:14.41 | kergoth | no timeframe. whenever the blocking issues are solved. see the page in the wiki |
| 04:15.25 | W8TVI | hey kergoth, do you happen to know if the fuji np60 works in a 5x00? |
| 04:15.33 | kergoth | dunno |
| 04:15.40 | Matthew_I | hmm, anything I can do to help? |
| 04:15.49 | Matthew_I | I have some coding skills |
| 04:15.56 | kergoth | look at the wiki, at the pagea bout the blocking issues :) |
| 04:15.59 | kergoth | pick one, and have at it |
| 04:16.09 | kergoth | otherwise, we could always use people to maintain packages |
| 04:16.23 | kergoth | (just involves making sure it always builds, and updating to new versions, that sort of thing) |
| 04:16.55 | kergoth | heh, you and me both |
| 04:16.56 | W8TVI | I've seen sevral people posting about the np60 battery being a "Cheap Replacement battery" |
| 04:17.07 | kergoth | we're really short on people for web work as well. though thats changing now |
| 04:17.27 | Matthew_I | hmm, well I have some experience with that kind of stuff |
| 04:17.27 | W8TVI | but I have yet to get a reply from some one that has used it... :( |
| 04:17.53 | W8TVI | I think my two 5x00 batteries are dieing |
| 04:18.09 | Matthew_I | do I need to look at the OzTodo page? |
| 04:18.36 | kergoth | Matthew_I: there are over a thousand packages oe can build, and only like 4 of us maintaining it all. |
| 04:18.46 | Matthew_I | right right |
| 04:18.46 | kergoth | you can imagine the amount of core developer time thats being wasted there |
| 04:18.54 | Matthew_I | I have read the news letters |
| 04:19.02 | Matthew_I | lets get the word out |
| 04:19.06 | kergoth | play around with it, make sure you can build things, go from there |
| 04:19.16 | kergoth | as far as visability is concerned, i expect the OE core going 1.0 to help that |
| 04:19.32 | kergoth | which will happen as soon as cterry gets the web site sucking less, and pb_ and i get some docs done |
| 04:19.54 | Matthew_I | nice |
| 04:20.31 | kergoth | the way things have been up until now, i've actually wanted things to stay relatively low profile, because things were changing so rapidly |
| 04:20.36 | W8TVI | kergoth: this photo show markings that make it look like its the wrong polarity: http://www.ebcbattery.com/data/PL60B(1).jpg |
| 04:21.33 | Matthew_I | kergoth: would you reccommend 3.3.5 or 3.3.6-pre1 ? |
| 04:21.43 | kergoth | 3.3.6-pre1 is far more stable than 3.3.5. |
| 04:21.49 | Matthew_I | kergoth: really? |
| 04:21.55 | kergoth | but you lose binary compatibility, so you'd need oz-compat to run old binary packages like hancom |
| 04:21.59 | kergoth | yep |
| 04:22.02 | kergoth | 3.3.5 has known issues |
| 04:22.05 | Matthew_I | kergoth: hmm... all this compat stuff if getting on my nerves, it does not always work |
| 04:22.12 | W8TVI | there are a bunch of np60s on ebay that go for cheap, but I dont want to waste money on something that wont work and might kill my zaurus |
| 04:22.20 | kergoth | well, we're not going to stay compatible with sharp's crap forever |
| 04:22.29 | Matthew_I | kergoth: I think I am gonna focus on this oe stuff |
| 04:22.34 | kergoth | there comes a time when we should take advantage of the advances being made |
| 04:22.36 | kergoth | thats what i'd recommend |
| 04:22.41 | Matthew_I | kergoth: so this crap will work |
| 04:22.46 | W8TVI | opera isn't crap |
| 04:23.03 | Matthew_I | W8TVI: and opera is from, opera soft... |
| 04:23.09 | Matthew_I | kergoth: ok, thanks for your time |
| 04:23.16 | orpheus | and can't you get it fro free from ibm? |
| 04:23.17 | W8TVI | now if only konq wasn't crap... |
| 04:23.19 | kergoth | np. |
| 04:23.27 | kergoth | W8TVI: itd help if we could move to qt3. |
| 04:23.29 | kergoth | W8TVI: or 4.. |
| 04:23.34 | kergoth | W8TVI: to use the new khtml in recent kde |
| 04:24.24 | W8TVI | orpheus: the opera from IBM wont run with out the compat stuff |
| 04:25.46 | W8TVI | kergoth: that would be nice |
| 04:26.06 | kergoth | course with gpe, you could use something gecko driven if your device has enough ram |
| 04:26.13 | kergoth | firefox, whatever |
| 04:26.35 | W8TVI | heh |
| 04:26.52 | W8TVI | firefox would be cool on the zaurus |
| 04:27.03 | W8TVI | isn't the yoppy gpe? |
| 04:27.37 | Matthew_I | minimoz or something |
| 04:27.51 | kergoth | yea, theres minimo too |
| 04:27.52 | Matthew_I | is what it's called, it's supposed to be better than konq and opera |
| 04:28.11 | Matthew_I | how do I check out oe-doc? |
| 04:28.38 | kergoth | dont bother. its bitrotted (or whatever the doc equivalent is) |
| 04:29.32 | W8TVI | "typo" |
| 04:30.28 | kergoth | heh |
| 04:33.38 | Matthew_I | he he , had the directory I was checking out to with the wrong permissions |
| 04:41.32 | treke|home | hrmm,. I think I've commiteded the cardinal sin. |
| 04:41.36 | treke|home | drinking and coding. |
| 04:41.41 | kergoth | ack. |
| 04:41.52 | kergoth | you're supposed to drink & *design*, code later |
| 04:41.53 | kergoth | heh |
| 04:43.23 | Matthew_I | who do I get oemake? |
| 04:43.29 | kergoth | ls oe/bin/ |
| 04:43.30 | kergoth | heh |
| 04:43.41 | kergoth | the instructions told you to add that to your PATH for a reason.. |
| 04:43.58 | Matthew_I | ok |
.jpg){kind=link}