| 00:00.42 | killefiz | ahh - i've finally finished updating all servers to new ssh packages |
| 00:02.28 | nightfire | does the buildroot currently contain the new c7x0 keymap for opie? |
| 00:02.31 | nightfire | <PROTECTED> |
| 00:02.34 | kergoth | 'new' |
| 00:02.40 | nightfire | hmm |
| 00:02.42 | nightfire | 'newer' |
| 00:02.50 | kergoth | mickey made some slight alterations, i dont know what changes you're referring to |
| 00:02.56 | nightfire | closer to the sharp keymap |
| 00:02.56 | kergoth | the keymap still isnt fully populated |
| 00:03.00 | nightfire | ah |
| 00:03.06 | kergoth | closer, yes |
| 00:03.09 | nightfire | thought i saw something on the dev list about it |
| 00:03.15 | nightfire | cool |
| 00:03.46 | treke | what was the content of the email? |
| 00:04.11 | killefiz | kergoth: i guess that all people that could help out hang out here anyway |
| 00:04.14 | kergoth | just a notification that the transition has begun and a request for volunteers to assist on it |
| 00:04.22 | kergoth | killefiz: yeah |
| 00:04.49 | killefiz | how many packages are needed for a working image that includes opie? |
| 00:05.07 | kergoth | well, look at the defconfigs |
| 00:05.10 | gb2 | hrm |
| 00:05.13 | kergoth | they're the best list of that |
| 00:05.31 | kergoth | the minimal packageset is nearly done, minus kernels, kernel modules, and other packages that need the kernels |
| 00:05.56 | killefiz | what about dropping everything that is not strictly neccessary from buildroot now and adding it later |
| 00:06.09 | kergoth | heh, no 'drop' is necessary |
| 00:06.15 | kergoth | its just a matter of only adding whats needed as we go |
| 00:06.19 | kergoth | which is what i've been doing |
| 00:06.19 | killefiz | doing that the transition could be at 40% already |
| 00:06.24 | kergoth | high priority items first. |
| 00:06.33 | killefiz | kergoth: it's more a psychological thing |
| 00:06.37 | kergoth | which is why minimal is nearly done |
| 00:06.56 | kergoth | i could take the minimal & opie defconfigs and update our todo with all the packages needed at each point |
| 00:07.03 | kergoth | and then in a seperate section list those that are low priority |
| 00:07.09 | kergoth | so we have a better picture of what needs doing |
| 00:13.07 | *** join/#openembedded qip|Virus|da (~Virus@pD9E654FD.dip.t-dialin.net) |
| 00:20.10 | kergoth | ljp: bastard. |
| 00:20.32 | kergoth | libfetish.a |
| 00:20.33 | kergoth | hmm |
| 00:21.13 | chouimat | kergoth I ditched gentoo today I'm on suse 8.2 so far it's working well |
| 00:21.25 | ljp | dont forget about the other free stuff I can;t tell you about |
| 00:21.53 | chouimat | ljp: ok you go a free QT commercial license and so what :) |
| 00:22.00 | chouimat | s/go/got |
| 00:22.20 | ljp | no I dont. I can't sell comemrcial software |
| 00:22.53 | chouimat | ljp :( |
| 00:24.15 | ljp | small price to pay to get free espresso |
| 00:24.25 | chouimat | ljp hehe |
| 00:25.14 | ljp | chouimat: theres a new TT mailing list you might be interested in called marketplace. occasional Qt job listings from TT and/or other companies and such |
| 00:25.38 | ljp | not much activity, though |
| 00:25.45 | chouimat | ljp ok thanks btw I really need to talk to harryF tomorrow |
| 00:26.30 | chouimat | ljp advertised on the site? |
| 00:26.43 | ljp | its on the mailing lists page |
| 00:29.57 | chouimat | later kergy |
| 00:30.41 | chouimat | ljp got some news about your visa? |
| 00:31.49 | ljp | not yet. just that they will decide by the 30th for sure |
| 00:32.09 | chouimat | ljp ok |
| 00:35.28 | gb2 | grrr |
| 00:38.34 | *** join/#openembedded bisho (~bisho@136.Red-213-97-191.pooles.rima-tde.net) |
| 00:38.45 | *** join/#openembedded effy-kun (~dontemail@merx.co.jp) |
| 00:38.51 | chouimat | "and the australian asscatcher is here feeding its young" |
| 00:39.09 | effy-kun | "asscatcher"? |
| 00:39.39 | chouimat | effy-kun nm someone is creating name for tropical bird on tv |
| 00:42.51 | chouimat | ljp http://ktown.kde.org/~coolo/contributors.html |
| 00:44.49 | gb2 | kergoth: if i set arbitrary variables in a .oe, do they endup in the environment? |
| 00:45.38 | chouimat | http://techcentralstation.com/081803C.html |
| 00:47.31 | gb2 | facinating |
| 00:47.34 | gb2 | i got OE into a loop |
| 00:49.28 | ljp | ok, choui. who is who there?! ;) |
| 00:51.02 | ljp | where's the naked women? |
| 00:51.08 | chouimat | ljp: I'm on the pict somewhere |
| 00:53.00 | ljp | your the bald guy in the front row? ;) |
| 00:53.06 | chouimat | ljp yes |
| 00:54.15 | ljp | hmm, that looks like Don in the back row |
| 00:54.33 | chouimat | ljp I share the room with don |
| 00:54.41 | chouimat | s/share/shared |
| 00:56.31 | ljp | really? Don from TT? |
| 00:57.46 | chouimat | ljp Don Sander? |
| 00:57.51 | kergoth`bbl | gb2: define environment. |
| 00:58.02 | kergoth`bbl | gb2: they end up in the generated shell script, and so are usable from your functions |
| 00:58.13 | kergoth`bbl | gb2: but they dont get shell 'export' unless you mark them as such |
| 00:58.22 | gb2 | ok |
| 00:58.42 | ljp | I think thats his name |
| 00:58.43 | gb2 | can i just put an export in top level? |
| 00:58.53 | ljp | mac guy |
| 00:59.00 | kergoth`bbl | nah, right now you must place export as part of the variable definition |
| 00:59.08 | kergoth`bbl | see doc/TODO, the item regarding export syntax |
| 00:59.13 | chouimat | ljp: I shared the room with a don who spoke aussie english so maybe it him |
| 00:59.15 | kergoth`bbl | so you cant just 'export var' |
| 00:59.17 | gb2 | ok |
| 00:59.24 | gb2 | export var=.. |
| 00:59.24 | gb2 | ? |
| 00:59.33 | kergoth`bbl | you could export var := ${var} if it was already defined elsewhere |
| 00:59.34 | kergoth`bbl | yeah |
| 00:59.36 | kergoth`bbl | exactly |
| 00:59.37 | ljp | ahh. hmm. cool. I knew he was gone, didnt think he went toEurope |
| 00:59.48 | ljp | he's actually from New Zealand |
| 00:59.56 | gb2 | hm... need support for separated build/source |
| 01:00.07 | chouimat | ljp ok ... |
| 01:00.11 | kergoth`bbl | enhance autotools.oeclass |
| 01:00.50 | gb2 | that's what i'm doing |
| 01:01.37 | kergoth | spiffy |
| 01:02.06 | gb2 | hmm.. needs a better do_compile |
| 01:02.34 | ljp | is it lunch yet? |
| 01:04.53 | mewyn | hey, anyone know if there's been any OSX sync for the Z? |
| 01:05.33 | kergoth | mewyn: qtopia desktop 1.7 for osx released today |
| 01:05.57 | bisho | http://www.linuxdevices.com/news/NS9214394349.html <--nice!!! |
| 01:06.10 | bisho | $300 handheld with linux/Qtopia |
| 01:08.15 | ljp | I even tried to test that qtopiadesktop, but I am not much of a mac user. I can confirm that yes, it does in fact install and run on OS X |
| 01:10.13 | gb2 | test -z "i686-i686-i686-i686-i686-i686-i686-i686-i686-i686-i686-... |
| 01:10.14 | gb2 | neat |
| 01:10.27 | kergoth | hah |
| 01:10.34 | kergoth | use := if you want something resolved now |
| 01:10.40 | kergoth | same as gmake |
| 01:10.57 | kergoth | hmm |
| 01:11.11 | kergoth | what the hell, my xmms decided to play all my mp3s at the wrong speed all of a sudden |
| 01:13.07 | gb2 | hm |
| 01:13.34 | chouimat | ljp which version of knoppix |
| 01:13.48 | gb2 | kergoth: use of ${FOO} in a .oeclass .. those get expanded before writing out the run.* script |
| 01:14.13 | kergoth | yep, all oe style vars will get expanded prior to output |
| 01:14.18 | gb2 | which makes it hard to modify one |
| 01:14.29 | kergoth | we should probably allow $${FOO} or \${FOO} |
| 01:14.36 | kergoth | just use $FOO for vars at shell runtime |
| 01:14.40 | gb2 | yeah |
| 01:15.24 | chouimat | ljp I will soon have a triple head development box ... |
| 01:15.39 | ljp | hmm |
| 01:17.36 | gb2 | hrm. |
| 01:17.42 | ljp | knoppix 3.2 |
| 01:22.24 | gb2 | embedding quotes into strings is dodgy |
| 01:23.33 | kergoth | yep |
| 01:23.40 | kergoth | cant easily escape them |
| 01:23.49 | kergoth | part of why i wanted to rewrite the parser |
| 01:24.03 | *** join/#openembedded Pendalar (Pendalar@68.114.156.106) |
| 01:24.28 | gb2 | fog% ls staging/bin |
| 01:24.28 | gb2 | i386-uclibc-addr2line* i386-uclibc-ld* i386-uclibc-readelf* |
| 01:24.29 | gb2 | ... |
| 01:24.30 | gb2 | score |
| 01:24.42 | gb2 | now.. for more than just binutils |
| 01:25.36 | gb2 | WOAH |
| 01:25.40 | gb2 | my whole apartment building just shook |
| 01:27.10 | gb2 | oh, it's construction outside, ok. |
| 01:27.35 | gb2 | kergoth: if i want to nullify say.. do_install, i seem to need to do: do_install () { true } |
| 01:27.47 | gb2 | do_install () { } doesn't work. |
| 01:28.15 | kergoth | yeah, thats an annoying little issue |
| 01:28.31 | kergoth | been planning on making the exec function check for a null or only whitespace function ,and bypass its execution |
| 01:28.38 | kergoth | should stuff that into doc/TODO |
| 01:28.50 | kergoth | there are a ton of these little low priority things that are in my head but not in TODO |
| 01:29.03 | ljp | ahh, a todo todo |
| 01:29.11 | kergoth | hehe |
| 01:30.07 | mewyn | kergoth: i'd rather have it sync with the mac stuff |
| 01:31.58 | *** join/#openembedded victorvdl (~victorvdl@ip135.124-173-207.eli-du.nwlink.com) |
| 01:34.06 | chouimat | SCO Claims $15,300,000 From SCOsource <--- lol |
| 01:35.24 | *** join/#openembedded Soopaman (~Soopaman@h24-66-54-32.wp.shawcable.net) |
| 01:39.56 | *** join/#openembedded chouimat (~root@r2351064.cidc.net) |
| 01:41.00 | ljp | ROOOOOOTTTT! has landed |
| 01:41.27 | aman | ljp die |
| 01:41.28 | aman | ibot die |
| 01:41.29 | | ACTION takes two shots to the head and crumples to the ground, lifeless. |
| 01:49.37 | gb2 | crap |
| 01:49.40 | gb2 | why is include not working |
| 01:50.13 | *** join/#openembedded caffeine (~kevin@clt74-76-015.carolina.rr.com) |
| 01:51.05 | gb2 | i keep trying to do %{foo} |
| 01:51.07 | gb2 | ala rpms |
| 02:08.40 | kergoth | added that bootmenu to OE |
| 02:09.00 | gb2 | hrm. these scriptlets sure are restrictive |
| 02:09.06 | kergoth | what scriptlets? |
| 02:09.13 | gb2 | in .oe files |
| 02:09.18 | kergoth | what do you mean? |
| 02:09.48 | gb2 | i think i'm running into parser quirks |
| 02:10.05 | kergoth | that wouldnt surprise me. avoid ${} syntax and you should be fine. |
| 02:10.16 | gb2 | great :P |
| 02:10.30 | gb2 | hrm. no base_do_patch |
| 02:10.48 | kergoth | its python |
| 02:10.53 | kergoth | you just ran into what we discussed earlier |
| 02:11.03 | gb2 | yeah |
| 02:11.07 | kergoth | inability to call the python version from a shell oe func |
| 02:11.12 | kergoth | see what i did for glibc to work around it |
| 02:11.15 | gb2 | didn't error out at the command not found htough |
| 02:11.24 | kergoth | thats odd |
| 02:11.34 | kergoth | if you put 'false' on a single line in the function, does it fail out? |
| 02:11.38 | masticator | man theres a lot of star trek themed porno sites. |
| 02:12.12 | gb2 | lemme clean up and check |
| 02:12.15 | kergoth | k |
| 02:12.27 | kergoth | just do that to confirm the -e mode of the shell is behaving |
| 02:12.52 | gb2 | k.. it's unpacking.. |
| 02:13.15 | gb2 | do_patch() { set -e; false; ... } |
| 02:13.18 | gb2 | is roughly what's in ther enow |
| 02:13.31 | kergoth | k |
| 02:13.34 | gb2 | NOTE: package uclibc-gcc-3.3.1-r0: task do_patch succeeded |
| 02:13.43 | kergoth | try removing the set -e |
| 02:13.46 | kergoth | -e is default now |
| 02:13.54 | gb2 | ok. |
| 02:14.09 | kergoth | figured -e is the common case, not failing is the rarity |
| 02:14.32 | kergoth | brb, food |
| 02:14.44 | gb2 | + do_patch |
| 02:14.45 | gb2 | + set -e |
| 02:14.45 | gb2 | + false |
| 02:14.46 | gb2 | ... |
| 02:19.06 | *** join/#openembedded treke_ (~ggilbert@12.107.12.130) |
| 02:20.14 | *** join/#openembedded hcarty (~hcarty@dsl093-100-214.wdc2.dsl.speakeasy.net) |
| 02:20.15 | chouimat | hi treke_ |
| 02:20.25 | treke_ | hey |
| 02:20.30 | *** part/#openembedded hcarty (~hcarty@dsl093-100-214.wdc2.dsl.speakeasy.net) |
| 02:25.18 | gb2 | doh. |
| 02:25.20 | gb2 | import != inherit |
| 02:29.23 | kergoth | i should implement the fix for file permissions & ownership in our packaging |
| 02:30.56 | *** join/#openembedded W8TVI_ (~root@1114104758.mi.dial.hexcom.net) |
| 02:31.09 | gb2 | kergoth: doing the python do_patch fixed my bug btw |
| 02:31.19 | kergoth | k, cool |
| 02:31.57 | gb2 | interdependance between libc and gcc |
| 02:32.00 | gb2 | bigass sucks |
| 02:32.01 | gb2 | grrr |
| 02:32.10 | kergoth | sure does |
| 02:32.30 | kergoth | maybe we should make it use the libc from the toolchain and stuff those files into the package |
| 02:32.34 | kergoth | hehe |
| 02:33.41 | chouimat | kergoth bah |
| 02:34.22 | gb2 | kergoth: this is part of why i liked my gb-toolchain-1.0.oe idea |
| 02:34.23 | gb2 | ;) |
| 02:34.32 | kergoth | yeah, simplifies things alright |
| 02:34.47 | kergoth | problem is, how the hell do you pull off something like our 'use gcc2 for some things, gcc3 for others' deal |
| 02:34.56 | kergoth | woudl be nearly impossible to manage |
| 02:35.13 | kergoth | brb |
| 02:35.14 | gb2 | fix the kernel |
| 02:35.17 | gb2 | ;) |
| 02:35.32 | gb2 | move away from gcc2 entirely |
| 02:36.03 | kergoth | gb2: i'm not talking about hte kernel, the kernel is easily patched |
| 02:36.13 | gb2 | what else requires gcc2? |
| 02:36.13 | kergoth | i'm referring to qt/qpe userspace due to c++ ABI |
| 02:36.27 | gb2 | oh. desire for binary compatibility? |
| 02:36.31 | kergoth | yep |
| 02:36.40 | chouimat | kergoth a "build-dependency " command? |
| 02:36.45 | kergoth | we have build dependencies |
| 02:36.53 | kergoth | but how would you tell it to build two toolchains? |
| 02:36.58 | kergoth | and not stomp on one another in staging? |
| 02:37.03 | kergoth | sounds like a pain in the ass. |
| 02:37.12 | chouimat | kergoth suffix or prefix |
| 02:37.14 | gb2 | kergoth: i have uclibc-binutils-cross-2.14.90.0.6.oe, which has "include uclibc-binutils-${PV}.oe" in it. |
| 02:37.38 | gb2 | why does it print out uclibc-binutils-2.14.90.0.6-r0 in all the messages? |
| 02:37.41 | gb2 | rather than -cross |
| 02:38.05 | chouimat | kergoth or different path for each toolchain |
| 02:38.06 | kergoth | hmmm |
| 02:38.14 | kergoth | i'm betting the set of PN/PV/PR from filename is messed |
| 02:38.27 | kergoth | the include is making it set them for the included file |
| 02:38.39 | kergoth | so its setting them from the fn prior to parsing, not after |
| 02:38.57 | kergoth | which makes sense, since we want to allow a .oe to override them itself |
| 02:39.02 | kergoth | but doesnt, since it breaks including |
| 02:39.05 | kergoth | ideas? |
| 02:39.05 | kergoth | heh |
| 02:39.07 | gb2 | heh |
| 02:39.12 | gb2 | special case include? |
| 02:39.26 | kergoth | right now include just passes the fn to the same parser that handled the toplevel file |
| 02:39.29 | kergoth | no special case |
| 02:39.32 | kergoth | nice & clean |
| 02:39.37 | kergoth | i'd lke to keep it that way if we can |
| 02:39.38 | kergoth | heh |
| 02:39.43 | gb2 | hm |
| 02:40.13 | gb2 | we need depth first parsing |
| 02:40.33 | gb2 | wait.. |
| 02:40.39 | gb2 | hrm |
| 02:40.42 | kergoth | right now its completely linear. it loads it at the exact point in the file you specify |
| 02:40.47 | gb2 | yeah |
| 02:40.50 | gb2 | i was thinking that |
| 02:40.51 | kergoth | which is nice since it gives you nice control |
| 02:40.54 | kergoth | over what you can override |
| 02:40.58 | kergoth | specify before or after |
| 02:41.09 | gb2 | yeah |
| 02:41.24 | kergoth | we could do it this way.. |
| 02:41.34 | kergoth | set PN/PV/PR from fn both before and after parsing |
| 02:41.42 | kergoth | but the before one sets them unconditionally |
| 02:41.49 | kergoth | after sets them only if each var is not already set |
| 02:41.59 | kergoth | bah, that wont work either |
| 02:42.07 | kergoth | maybe just make it completely conditional |
| 02:42.14 | kergoth | only set htem if they arent already specified |
| 02:42.18 | kergoth | s/specified/set/ |
| 02:42.36 | kergoth | then since it was already set at toplevel, it wont get set from the fn of the higher depths |
| 02:42.52 | kergoth | bah, that doesnt work either |
| 02:42.59 | kergoth | since the defaults are set at oe.conf level |
| 02:43.05 | kergoth | bleh. |
| 02:43.42 | gb2 | i actually think it should be an error if they're not set at the end of .oe parsing |
| 02:44.14 | kergoth | you think we shouldnt allow the default r0 for PR? |
| 02:44.22 | gb2 | hrm |
| 02:44.24 | kergoth | we'll need to add its set to every .oe we have right now :) |
| 02:44.30 | gb2 | PN and PV certainly |
| 02:44.44 | gb2 | well. |
| 02:44.46 | gb2 | hrm |
| 02:44.48 | gb2 | eh |
| 02:45.17 | gb2 | oh. |
| 02:45.18 | kergoth | guess tracking depth is the cleanest way to handle it |
| 02:45.21 | gb2 | how about this. |
| 02:45.23 | kergoth | only set from fn at toplevel |
| 02:45.52 | gb2 | no.. oe.conf screws my idea |
| 02:46.11 | kergoth | hehe. we either have to track depth or track source |
| 02:46.24 | kergoth | ie. store whether it came from a config file or a .oe in the metadata flags |
| 02:46.28 | kergoth | the way make does |
| 02:46.36 | kergoth | 'command line', 'environment', 'file' |
| 02:46.41 | kergoth | hmm. |
| 02:47.25 | nightfire | hmm.. |
| 02:47.33 | kergoth | heh, there's an easy solution to this |
| 02:47.37 | gb2 | i had the thought of conditionally assign at file load, and if unassigned at the end of file load, apply the oe.conf ones |
| 02:47.49 | kergoth | set the default versions of the variables in the function that sets them from the fn |
| 02:47.57 | kergoth | then oe.conf isnt an issue |
| 02:48.11 | kergoth | but i prefer having all metadata contents in .oe & .confs |
| 02:48.11 | kergoth | heh |
| 02:48.26 | nightfire | ok, what's up with double suspending on button press (but not apm -s) |
| 02:48.28 | gb2 | default_PV=... |
| 02:48.35 | nightfire | is this just me? |
| 02:49.47 | kergoth | hmm. |
| 02:49.58 | kergoth | okay, ?= to set a var only if it isnt already set |
| 02:50.13 | kergoth | then, use ${@} python substitution to call the function to get the var from the fn for the default |
| 02:50.26 | kergoth | _or_ if that returns None, to the actual var default |
| 02:50.31 | kergoth | then, you put that in the base oeclass |
| 02:50.35 | kergoth | not oe.conf |
| 02:50.59 | kergoth | then we have only set if need to be set, plus a proper fallback to a default |
| 02:59.19 | gb2 | kergoth: hrm.. cascading inheritance would be nice |
| 02:59.22 | *** join/#openembedded TimRiker (~timr@proxyle02.ext.ti.com) |
| 02:59.36 | gb2 | evening timriker |
| 02:59.43 | chouimat | hi tim |
| 02:59.45 | kergoth | hey TimRiker |
| 02:59.50 | ljp | timmy too! |
| 03:00.49 | chouimat | ljp go back to work |
| 03:01.02 | ljp | I'm at work |
| 03:01.38 | gb2 | kergoth: do you have anything like a formal grammar for .oe files? |
| 03:02.01 | kergoth | gb2: nope. it used to be exactly what the format was of an ebuild, but its deviated so far that its something else entirely now |
| 03:02.09 | kergoth | should probably document it in bnf in doc/ |
| 03:02.42 | gb2 | i think we're going to want to make this a little less shell scriptish |
| 03:03.04 | kergoth | i assume you mean the format itself, not the shell/python function support |
| 03:03.09 | gb2 | yes |
| 03:03.10 | kergoth | in which case i agree wholeheartedly |
| 03:03.18 | gb2 | something like: |
| 03:03.38 | ljp | yes, have a document day soon |
| 03:03.40 | gb2 | something closer to how rpm has %post %pre %build, etc. |
| 03:03.49 | gb2 | not necessarily the same format |
| 03:03.57 | gb2 | but more formal |
| 03:04.03 | kergoth | i dont want to remove the ability to add any arbitrary metadata item. |
| 03:04.12 | kergoth | so i dont want a defined set of "valid" functions |
| 03:04.22 | gb2 | yeah, i agree. |
| 03:04.38 | kergoth | need to retain our flexibility. thats one of hte biggest advantages of the system |
| 03:04.47 | kergoth | also need to keep the format as simple as possible |
| 03:04.51 | kergoth | i.e. no XML .oe files |
| 03:05.04 | kergoth | s/simple/simple and editable by an end user/ |
| 03:05.31 | gb2 | currently.. it's too free form, so errors aren't easily detected |
| 03:05.41 | gb2 | i keep trying to do "import autotools" for example |
| 03:05.44 | kergoth | agreed |
| 03:05.59 | kergoth | well, the parser doesnt respond to errors well (.. at all) |
| 03:06.22 | kergoth | have you looked at A-A-P? |
| 03:06.22 | gb2 | function [python|sh] <function> () {<eol> |
| 03:06.31 | gb2 | <func_body> |
| 03:06.33 | gb2 | }<eol> |
| 03:06.35 | gb2 | for example |
| 03:06.36 | kergoth | i kinda like their makefile-like function definition format |
| 03:07.07 | gb2 | got a quick example file? |
| 03:09.00 | gb2 | in my example.. <func_body> would be whatever format the function type required. this has the advantage of making it easy to write the vim syntax files for it |
| 03:09.07 | gb2 | since the body is nice and contained line-wise. |
| 03:09.07 | kergoth | gb2: http://www.a-a-p.org/examples.html#port |
| 03:09.15 | kergoth | take a look at do-checksum |
| 03:09.36 | kergoth | the lines beginning with colons within the "function" are calls to a-a-p functions, not shell/local code |
| 03:09.55 | gb2 | ok |
| 03:10.51 | kergoth | i kinda like the : method of doing internal calls, since you can do it etiher at toplevel or inline in the function |
| 03:11.35 | kergoth | you can see that inter-function dependency is represented the same way make does it, which is clearer than our 'addtask' cruft |
| 03:11.56 | gb2 | for functions.. it needs to be freeform within the line bounds though |
| 03:12.00 | kergoth | anyway, it has some interesting ideas i think need considering |
| 03:12.02 | kergoth | yeah |
| 03:12.49 | kergoth | they have :variant, which is a basic switch like conditional |
| 03:12.57 | kergoth | not as flexible as i'd like, but .. |
| 03:14.01 | gb2 | yeah |
| 03:14.50 | gb2 | what's aap written in? |
| 03:16.03 | kergoth | python |
| 03:16.13 | kergoth | its nowhere near as flexible as what we have wrt metadata handling |
| 03:16.22 | kergoth | but it has far more extensive internal functions for doing things |
| 03:18.05 | gb2 | python is starting to grow on me |
| 03:21.27 | *** join/#openembedded _Psycho (~Bob@HSE-Montreal-ppp3472455.sympatico.ca) |
| 03:27.07 | *** join/#openembedded _Bala_ (~Bala@hsdbsk142-165-117-29.sasknet.sk.ca) |
| 03:42.40 | kergoth`bbl | hmm |
| 03:43.14 | gb2 | kergoth: where is the parsing done in the code? |
| 03:43.28 | kergoth`bbl | bin/oe/parse/ |
| 03:43.42 | gb2 | oh duh |
| 03:44.01 | gb2 | i was looking at those .py files, going "which one would be parsing.. none of them is named parse... hmm..." |
| 03:44.07 | gb2 | totally ignoring the dir |
| 03:44.13 | kergoth`bbl | hehe |
| 03:45.05 | kergoth`bbl | that parser code is a mess |
| 03:45.15 | kergoth`bbl | i restructured it somewhat, but its just a restructured mess |
| 03:45.23 | ljp | c7xx needs to detect when the screen is flipped |
| 03:45.50 | kergoth`bbl | its a keycode at runtime |
| 03:46.04 | kergoth`bbl | and a binary that checks the status at resume/bootup time |
| 03:46.11 | kergoth`bbl | talk to treke about it |
| 03:46.19 | ljp | ok. cool |
| 03:47.33 | gb2 | i was thinking about these files.. |
| 03:47.43 | gb2 | do we really need more than variables and functions? |
| 03:48.03 | gb2 | hm. |
| 03:48.06 | gb2 | conditionals i guess |
| 03:48.08 | kergoth`bbl | not if we have a way to express function dependency |
| 03:48.11 | kergoth`bbl | yeah, and conditionals |
| 03:48.59 | gb2 | should be relatively easy parsing then |
| 03:49.02 | kergoth`bbl | yep |
| 03:49.18 | kergoth`bbl | well, there are a few other cases |
| 03:49.21 | kergoth`bbl | i.e. event handlers |
| 03:49.32 | kergoth`bbl | which you can insert into the event handler list |
| 03:49.32 | gb2 | what are those? |
| 03:49.42 | kergoth`bbl | events are triggered at various points in the build |
| 03:49.45 | gb2 | sounds like a function |
| 03:49.48 | kergoth`bbl | i.e. task failure, task success, etc |
| 03:49.48 | kergoth`bbl | it is |
| 03:49.53 | kergoth`bbl | but its a function + associated data |
| 03:49.58 | gb2 | ok. |
| 03:49.59 | kergoth`bbl | you need to flag it as an event handler |
| 03:50.12 | kergoth`bbl | kinda lke in a-a-p they have function (param=value): |
| 03:50.20 | kergoth`bbl | functions + associated extra metadata |
| 03:50.28 | gb2 | ah |
| 03:50.39 | gb2 | oh. and thinking about the overrides thing. |
| 03:50.41 | *** join/#openembedded Soopaman (~Soopaman@h24-66-54-32.wp.shawcable.net) |
| 03:50.50 | gb2 | syntax could be a little better |
| 03:51.31 | gb2 | since _ is a perfectly valid character to have in a variable name |
| 03:51.58 | kergoth`bbl | yeah. proper conditionals could eliminate that, but i like that our current conditionals arent evaluated until the end of the parsing |
| 03:52.07 | kergoth`bbl | the conditional stays associated with the variable contents until the end |
| 03:55.22 | kergoth`bbl | course that may not be necessary... *shrugs* |
| 03:56.01 | gb2 | so, we have variables(with sub variables), functions(with attributes), variable assignment operators, and conditional definitions. |
| 03:56.18 | chouimat | borg:~ # acroread |
| 03:56.18 | chouimat | Xlib: connection to ":0.0" refused by server |
| 03:56.18 | chouimat | Xlib: No protocol specified |
| 03:56.19 | chouimat | Error: Can't open display: :0.0 |
| 03:56.19 | chouimat | borg:~ # |
| 03:56.23 | *** join/#openembedded stigger (stigger@p065-n1-gos.nlc.net.au) |
| 03:56.40 | kergoth | gb2: yeah. and of course interally a 'function' is just a variable flagged as executable. |
| 03:56.43 | kergoth | internally rather |
| 03:56.49 | gb2 | yeah |
| 03:57.22 | kergoth | oh |
| 03:57.29 | kergoth | hmm |
| 03:57.43 | gb2 | you might call all these things.. objects. |
| 03:57.45 | gb2 | ;) |
| 03:58.00 | kergoth | indeed. |
| 03:59.35 | gb2 | i'm envisioning inherit for .oe files rather than just classes rather than include. |
| 04:00.03 | kergoth | hehe. there's no difference between include an inherit anymore |
| 04:00.08 | kergoth | not internally |
| 04:00.34 | kergoth | except i think inherit looks in classes/ off of the path instead of its root |
| 04:00.37 | kergoth | thats it though |
| 04:01.53 | gb2 | i'm thinking the resolution of values would be somewhat different than it is today and it would be different and include would literally be inserting a file linearly in the parsing. |
| 04:02.05 | gb2 | my mind is sorta just wandering i think |
| 04:02.19 | Soopaman | BONER! |
| 04:03.01 | chouimat | wtf I can't start any X apps |
| 04:05.25 | gb2 | kergoth: i'm also thinking of proper ways to do cross-language multi inheritance of functions |
| 04:05.47 | kergoth | could use something like a-a-p's : syntax inline |
| 04:05.55 | kergoth | then our code does the resolution |
| 04:06.46 | gb2 | it's not so hard if you restrict it |
| 04:07.43 | *** join/#openembedded chouimat (~root@r2351064.cidc.net) |
| 04:12.16 | *** part/#openembedded chouimat (~root@r2351064.cidc.net) |
| 04:13.12 | gb2 | the only sane way i can think of to properly do functions is basically doing somethilng like |
| 04:13.20 | gb2 | \begin{verbatim} |
| 04:13.24 | gb2 | in LaTeX |
| 04:14.16 | gb2 | :function ...func definition... |
| 04:14.27 | gb2 | <all this goes into the func> |
| 04:14.30 | gb2 | :endfunction |
| 04:15.22 | gb2 | calling the superclass function would be somewhat tricky there, unless you do it entirely before or entirely after. |
| 04:15.52 | gb2 | or you could split it in the middle perhaps |
| 04:16.15 | gb2 | and put a :superfunc .. call in there |
| 04:16.41 | gb2 | or... |
| 04:17.01 | gb2 | the linkage infrastructure provides a super() in the native scripting language |
| 04:18.02 | kergoth | thatd be the cleanest method, since it could easily fit within the structure of the native code |
| 04:18.23 | kergoth | the :superfunc style wouldnt easily fit into native conditionals or loop structures |
| 04:18.46 | kergoth | course, you could provide such things ourselves the way a-a-p does, but thats a bit beyond what we should be tackling |
| 04:19.41 | W8TVI | :) |
| 04:22.55 | ljp | superfunc sounds like a bad album of 70's music, you'd buy from tv |
| 04:23.05 | kergoth | hehe |
| 04:33.36 | ljp | get yer Ronco Superfunc album with all your favorites from the movie Car Wash |
| 05:10.12 | *** join/#openembedded TimRiker (~timr@proxyle01.ext.ti.com) |
| 05:13.56 | cyn_ | haha what'd buildroot just grab from sony.net - netbase? |
| 05:14.15 | cyn_ | ah well, its still chugging thats the important part - lets hope she runs all night |
| 05:17.40 | kergoth | i'd go nuts if i dreamt about computing |
| 05:17.46 | kergoth | gotta get away from it _sometime_ |
| 05:24.26 | ljp | :) |
| 05:28.43 | kergoth | yeah, i think its best to lay out the metadata manipulation portion of the library before moving on to the parser. |
| 05:39.01 | TimRiker | who's SDG? http://sdgsystems.com/ ? they the best fbvncserver location at present? |
| 05:39.56 | TimRiker | ~fbvncserver |
| 05:41.33 | TimRiker | ~fbvncserver is a VNC server used on the Zaurus and iPAQ. Was http://opensource.lineo.com/fbvncserver/ but now http://sdgsystems.com/download.html |
| 05:41.34 | | okay, TimRiker |
| 05:43.34 | kergoth | TimRiker: schurig's started cleaning it up and making it less device specific, last i heard |
| 05:43.41 | kergoth | sdg still maintains the original code |
| 05:47.19 | *** join/#openembedded SuKoSh| (~sukoshi@dyn-81-166-18-32.ppp.tiscali.fr) |
| 06:12.01 | *** join/#openembedded exastra (~go@l5.org) |
| 06:14.59 | *** join/#openembedded heatxsink (~jack@12-234-106-86.client.attbi.com) |
| 06:20.22 | heatxsink | ibot seen TCMAN |
| 06:20.23 | | tcman <~dbehr@adsl-64-170-122-35.dsl.snfc21.pacbell.net> was last seen on IRC in channel #zaurus, 4d 3h 40m 25s ago, saying: 'cu guys'. |
| 06:59.44 | *** join/#openembedded lsmith (~vandal@p213.54.24.108.tisdip.tiscali.de) |
| 07:06.00 | *** join/#openembedded imk|z_ (pda@68.170.56.122) |
| 07:22.30 | *** join/#openembedded cmartin (~cmartin@gw.krypnet.de) |
| 08:13.25 | Twiun | Good morning! |
| 08:13.38 | cmartin | moin Twiun |
| 08:28.50 | *** join/#openembedded stief (~Stiefum@a194-109-160-120.adsl.xs4all.nl) |
| 08:28.51 | stief | hi |
| 08:46.41 | cmartin | hi mickeyl |
| 08:46.47 | mickeyl | morning cmartin |
| 08:48.00 | cmartin | just working in the backend of the HP so that it's up to date as soon as the next release is available |
| 08:49.04 | cmartin | i also downloaded serload-1.1 from CVS and tar'ed that version and uploaded it to sf.net, so that we don't have to deal with that version with the broken Makefile |
| 08:49.46 | mickeyl | good |
| 08:50.35 | mickeyl | i will take 30 minutes (more i don't have) to update the release notes now. I'll mail them for you afterwards. |
| 08:51.40 | cmartin | k, i already started that on the HP - with the current one as a base, but i can add those things i forgot from your version |
| 08:51.48 | *** join/#openembedded W8TVI (~root@1114105203.mi.dial.hexcom.net) |
| 08:52.01 | mickeyl | ah well - do you have a link for your latest version? no need to reproduce work |
| 08:52.23 | cmartin | a guy got onto my case yesterday about the Disclaimer i had on the HP - so i changed it to this: |
| 08:52.43 | cmartin | "Disclaimer: Well, according to German law thier gonna get you anyway. So don't bother disclaiming anything ;-)" |
| 08:53.22 | cmartin | http://opensimpad.org/index.php?id=148 |
| 08:57.00 | *** join/#openembedded lsmith (~vandal@port-212-202-39-108.reverse.qsc.de) |
| 09:00.22 | mickeyl | hehe |
| 09:00.25 | mickeyl | that's nice :) |
| 09:02.19 | cmartin | mickeyl: should i set up a "Weiterleitung" for you at opensimpad.org, like mickeyl@opensimpad.org and add that to the bottom of the Release Notes also? |
| 09:03.14 | cmartin | mickeyl: or should i just add your hh.org address? |
| 09:04.16 | cmartin | mickeyl: have anything that should be added (or deleted) from my Release Notes? |
| 09:04.47 | mickeyl | cmartin: I'm collecting Email-forwardings :) |
| 09:05.04 | mickeyl | cmartin: mickeyl@opensimpad.org --> mickey@tm.informatik.uni-frankfurt.de please |
| 09:05.11 | cmartin | k |
| 09:05.18 | mickeyl | cmartin: i will add a bit more detail to the changelog. will mail to you soon |
| 09:08.31 | cmartin | k, just sent a mail to mickeyl@opensimpad.org to test it |
| 09:15.23 | mickeyl | nothing arrived |
| 09:15.24 | mickeyl | yet |
| 09:15.32 | cmartin | hmm |
| 09:15.33 | mickeyl | cmartin: just sent you a few corrections/additions |
| 09:15.39 | cmartin | k |
| 09:15.45 | mickeyl | cmartin: be sure that it is mickey@tm, not mickeyl@tm ;) |
| 09:16.12 | mickeyl | aaah it's there |
| 09:16.12 | cmartin | i just copied it from irc, so it should be right |
| 09:16.15 | cmartin | ok |
| 09:16.16 | mickeyl | it wandered in my SIMpad folder |
| 09:16.17 | mickeyl | hehe |
| 09:16.22 | mickeyl | ok, working. |
| 09:21.57 | cmartin | mickeyl: 1.0.2-pre1? - has the version number changed in CVS? - mine says 1.0.1.1-snapshot at the time |
| 09:22.54 | cmartin | mickeyl: and i have no opie-rdesktop |
| 09:27.42 | mickeyl | cmartin: 1.0.1.1 is 1.0.2-pre1, we just have it 1.0.1.1 because of easier upgrade from snapshots to releases (you know ipkg and its lexical comparison). so if we're talking about releases we should use the pre-notation. |
| 09:28.03 | mickeyl | but if you think that may confuse users, then drop it. |
| 09:28.12 | mickeyl | opie-rdesktop and the others are in HEAD. |
| 09:28.21 | mickeyl | we have to build them later and put them into the feed seperatly. |
| 09:29.05 | cmartin | well, when they open SystemInfo it displays 1.0.1.1-snapshot - this may confuse users |
| 09:29.40 | cmartin | are HEAD and BRANCH bc? |
| 09:31.04 | mickeyl | yes, thankfully. |
| 09:31.42 | mickeyl | hmm... i could hack sysinfo to display 1.0.1.1 as 1.0.2-pre1 :D |
| 09:31.45 | cmartin | k, so those packages can be added to the feeds as-is - no need for an extra feed for those |
| 09:31.51 | mickeyl | yes |
| 09:32.17 | mickeyl | okay lets stick with 1.0.1.1 but please put 1.0.2-pre1 in brackets afterwards |
| 09:32.40 | cmartin | k |
| 09:32.59 | mickeyl | versioning would be much easier if ipkg would suck less :) |
| 09:33.11 | cmartin | 1.0.1.1-snapshot (=1.0.2-pre1) - that o.k. |
| 09:33.19 | Twiun | when's the ipkg redesign due? |
| 09:33.20 | mickeyl | yeah |
| 09:33.46 | mickeyl | Twiun: no idea. the hh.org folks are considering this for quite a while |
| 09:34.12 | Twiun | mickeyl: that's what I thought |
| 09:34.39 | mickeyl | but they have a lot to do and that even without a buildroot :D |
| 09:35.08 | cmartin | what's the prob anyway - is it that ipkg will not install 1.0.2 over 1.0.2-pre1, because it think 1.0.2-pre1 is newer? |
| 09:35.10 | Twiun | yup - no criticism from me |
| 09:35.20 | Twiun | mickeyl: it's not as if I'm actively developing atm anyway |
| 09:44.11 | mickeyl | cmartin: yes, that exactly is the problem |
| 09:46.53 | *** join/#openembedded bisho (~bisho@138.100.21.100) |
| 09:47.56 | cmartin | mickeyl: Release Notes updated |
| 09:49.38 | *** join/#openembedded bisho (~bisho@138.100.21.100) |
| 09:50.27 | mickeyl | cmartin: looks good. one minor thing, it is the "O-Menu", not the "Q-Menu" ;) |
| 09:50.38 | cmartin | mickeyl: do you have an URL for me to get those HEAD apps - i'm too lazy to build a completely new HEAD |
| 09:50.39 | cmartin | k |
| 09:51.12 | cmartin | changed ;) |
| 09:51.36 | mickeyl | cmartin: unfortunately not. HEAD compiles but doesn't package at the moment. this has to be investigated. those applications have to arrive later |
| 09:52.39 | cmartin | mickeyl: k - i'll mention that in the news on the HP when we release it |
| 09:53.20 | *** join/#openembedded bisho (~bisho@138.100.21.100) |
| 09:53.21 | mickeyl | good. |
| 09:53.29 | cmartin | mickeyl: have you heard any complaints recently about the kernel-image ipk for other devices? |
| 09:53.59 | mickeyl | cmartin: no complaints (but then again, I fear no one is doing images at the moment...) |
| 09:54.49 | cmartin | k, killefiz said it worked fine for the iPAQ - i was just wanting to know if i broke anything for the other devices since that last change |
| 09:58.16 | mickeyl | OpenSIMpad 0.8.1 will probably be the last official release based on the current buildroot. |
| 09:58.31 | mickeyl | as major efforts have started in migrating to OE |
| 09:58.44 | cmartin | is OE that far already? |
| 09:58.46 | *** join/#openembedded bisho (~bisho@138.100.21.100) |
| 09:59.25 | cmartin | mickeyl: should we try to have the next release >0.8.1 built uclibc? |
| 09:59.25 | mickeyl | cmartin: can't comment on that. It seems to, but when I started in doing something I got only errors... ;) |
| 09:59.45 | mickeyl | cmartin: yes, that would be important. uclibc really saves a lot. |
| 09:59.46 | cmartin | s/built with |
| 09:59.51 | mickeyl | especially for the t-sinus folks |
| 09:59.56 | mickeyl | and it is faster, also |
| 10:00.05 | mickeyl | so my long-term plan would be |
| 10:00.21 | killefiz | mickeyl: does opie start with uclibc already? |
| 10:00.26 | mickeyl | 0.9 -> uclibc + Opie 1.2 |
| 10:00.33 | mickeyl | 1.0 -> Kernel 2.6 |
| 10:00.48 | mickeyl | killefiz: no idea. it compiles though :) |
| 10:00.48 | cmartin | sounds good to me |
| 10:01.38 | cmartin | will uclibc be used for other devices also? or are we gonna be BIC to all other devices then? |
| 10:03.27 | *** join/#openembedded bisho (~bisho@138.100.21.100) |
| 10:04.26 | mickeyl | can't comment on all packagers using buildroot, but I suspect that, once we have the possibility of building OZ-images based on uclibc, all packagers will take the opportunity to save space and improve efficiency. |
| 10:04.55 | mickeyl | also because of gcc 3.x |
| 10:05.08 | mickeyl | which i'm looking forward to |
| 10:05.22 | mickeyl | because on xscale-devices 2.9x just plain sucks |
| 10:09.24 | *** join/#openembedded bisho (~bisho@138.100.21.100) |
| 10:13.23 | *** join/#openembedded bisho (~bisho@138.100.21.100) |
| 10:19.24 | *** join/#openembedded bisho (~bisho@138.100.21.100) |
| 10:20.22 | *** join/#openembedded gruemel|laptop (~Jojo@p50802D75.dip0.t-ipconnect.de) |
| 10:25.11 | cmartin | time for lunch |
| 10:27.27 | *** join/#openembedded bisho (~bisho@138.100.21.100) |
| 10:28.21 | *** join/#openembedded bisho (~bisho@138.100.21.100) |
| 10:34.19 | *** join/#openembedded bisho (~bisho@138.100.21.100) |
| 10:38.50 | *** join/#openembedded bisho (~bisho@138.100.21.100) |
| 10:40.42 | *** join/#openembedded LordVan (~lordvan@62.218.218.165) |
| 10:56.02 | *** join/#openembedded Shdwdrgn (Picard@sourpuss.net) |
| 11:05.46 | *** join/#openembedded bisho (~bisho@138.100.21.101) |
| 11:30.25 | cmartin | sh... treke.net is down! |
| 11:31.26 | mickeyl | ? |
| 11:31.30 | mickeyl | doesn't matter or does it? |
| 11:31.55 | cmartin | kernel-header-2.4.18 are there |
| 11:32.20 | mickeyl | that file is mirrored on hundreds of places in the web |
| 11:32.32 | mickeyl | also here: |
| 11:32.36 | mickeyl | http://opie.net.wox.org/sources |
| 11:32.43 | cmartin | aahh |
| 11:33.40 | cmartin | thanx |
| 11:33.55 | cmartin | just doing my first uclibc build |
| 11:34.13 | mickeyl | cool. |
| 11:34.54 | cmartin | but that change to $STAGINGDIR is only possible after starting to build, right? |
| 11:35.05 | mickeyl | yeah |
| 11:35.23 | cmartin | when exactly? |
| 11:35.42 | cmartin | after the first error comes up? |
| 11:36.59 | cmartin | is it me or do things compile much faster with uclibc? |
| 11:37.33 | mickeyl | after the first error, yeah |
| 11:37.42 | mickeyl | of course, uclibc has much more lean headers |
| 11:38.15 | cmartin | k, first error |
| 11:38.25 | cmartin | now i go edit some things |
| 11:38.30 | mickeyl | yeah |
| 11:38.35 | mickeyl | read my STATUS and edit it if needed |
| 11:38.45 | cmartin | k |
| 11:39.40 | cmartin | oh, no $STAGING_DIR/target/include there yet - just lib and bin |
| 11:40.40 | cmartin | blkpg.c:252: error: asm-specifier for variable `__r1' conflicts with asm clobber list |
| 11:40.40 | cmartin | blkpg.c:252: confused by earlier errors, bailing out |
| 11:40.48 | cmartin | mickeyl: did you get that too? |
| 11:41.10 | mickeyl | what are you compiling ? what toolchain? |
| 11:42.33 | cmartin | kernel - what is ccache doing there - i comented out that CC_FOR_BUILD line in the Vars.local |
| 11:44.20 | mickeyl | ah |
| 11:44.34 | mickeyl | you need to do a combined build |
| 11:44.39 | mickeyl | gcc 2.x for kernel |
| 11:44.42 | mickeyl | gcc 3.x for apps |
| 11:44.58 | mickeyl | take the def-config/simpad-opie-gcc3 as a start and edit to build with uclibc |
| 11:44.59 | cmartin | huh |
| 11:45.27 | cmartin | k |
| 11:45.51 | prh | are they still making simpads? |
| 11:45.55 | mickeyl | while you're there, you can add us a def-config/simpad-opie-uclibc |
| 11:46.00 | mickeyl | prh: no |
| 11:46.01 | cmartin | prh: nope |
| 11:46.27 | prh | aww - they seem like a nice machine. updated with a pxa255 and more ram/flash... or a tosh 1.8" drive... |
| 11:46.29 | cmartin | mickeyl: yeah as soon as i have a working config |
| 11:46.48 | prh | actually a Z with a tosh 1.8" (ie the same drive in the iPOD...) would be fab! |
| 11:47.06 | cmartin | mickeyl: no def-config/simpad-opie-gcc3 there |
| 11:47.18 | mickeyl | ups |
| 11:47.29 | mickeyl | then do it by hand |
| 11:47.36 | mickeyl | enable gcc 2 and gcc 3 toolchain |
| 11:47.45 | mickeyl | and check [x] use gcc 3 for application builds |
| 11:47.56 | mickeyl | and _uncheck_ [] use gcc 3 for kernel builds |
| 11:48.02 | mickeyl | that's about all |
| 11:48.10 | mickeyl | buildroot will choose automatically the right compiler :) |
| 11:48.18 | cmartin | k, done |
| 11:48.33 | prh | hehe |
| 11:48.33 | Twiun | hehe |
| 11:48.50 | prh | I don't understand teh current one and don't see any point in trying to understand it |
| 11:48.59 | prh | I'd love to migrate the new OZ build to arch though |
| 11:49.12 | mickeyl | do it, but you're pretty much on your own then :) |
| 11:49.32 | prh | the point being to demonstrate that it can work really well with free tools... |
| 11:49.44 | prh | kergot might then be persuaded to switch too |
| 11:50.44 | Twiun | prh: migrate to arch, what's arch in this context? |
| 11:53.05 | cmartin | are there many differences between the current and the coming buildroot - how big will the learning kurve be? |
| 11:53.19 | cmartin | s/kurve /curve |
| 11:53.52 | cmartin | mc |
| 11:54.00 | cmartin | wrong term |
| 11:54.19 | mickeyl | no ide |
| 11:54.20 | mickeyl | a |
| 11:54.37 | mickeyl | from what i've seen yet, the new buildroot is |
| 11:54.40 | mickeyl | * more flexible |
| 11:54.49 | mickeyl | * easier to handle (because of dedicated application build classes) |
| 11:54.59 | mickeyl | * more comprehensible (due to python) |
| 11:55.55 | prh | Twiun: an RCS |
| 11:56.11 | prh | Twiun: look for gnu-arch |
| 11:56.15 | Twiun | prh: ahh |
| 11:56.31 | prh | Twiun: it's fab ;O) and particularly good for distributed dev |
| 11:56.58 | Twiun | prh: what's the merge/compare cycle like? |
| 11:57.03 | *** join/#openembedded BBMatthew (~mas01r@arm.ecs.soton.ac.uk) |
| 11:57.04 | Twiun | prh: any good guis? |
| 11:57.10 | Twiun | prh: windows port? |
| 11:58.03 | BBMatthew | hehe just got a boot to the command prompt with the 2.4.18, built gcc3.3.1 for armv5te xscale and lovely to see |
| 11:58.17 | Twiun | Woohoo! :) |
| 11:58.30 | BBMatthew | indeed |
| 11:58.32 | BBMatthew | :) |
| 11:58.38 | BBMatthew | just building opie now |
| 11:59.11 | BBMatthew | is march=armv5te mtune xscale == mcpu xscale? |
| 11:59.23 | prh | Twiun: there is a port to windows but I believe there are issues with filename lengths |
| 11:59.26 | mickeyl | BBMatthew: yo |
| 11:59.46 | prh | Twiun: there is a viewARCH and someone is working on a gui too. |
| 11:59.58 | Twiun | prh: sounds very green ;) |
| 12:00.01 | prh | Twiun: the merging is amazing |
| 12:00.19 | BBMatthew | hmm will get an extra large coffee and then to getting it operational |
| 12:00.34 | prh | Twiun: it's been around for 3 years - although the rate of adoption and development has accelerated in the last 6 months |
| 12:12.14 | cmartin | mickeyl: k, still getting this error - |
| 12:12.16 | cmartin | blkpg.c: In function `blk_ioctl': |
| 12:12.16 | cmartin | blkpg.c:252: error: asm-specifier for variable `__r1' conflicts with asm clobber list |
| 12:12.16 | cmartin | blkpg.c:252: confused by earlier errors, bailing out |
| 12:12.16 | cmartin | distcc[5119] ERROR: compile on localhost failed with exit code 1 |
| 12:12.29 | cmartin | ccache is no longer involved |
| 12:13.34 | cmartin | it comes after diving into this DIR - build/linux-2.4.19/drivers/block |
| 12:14.38 | cmartin | this looks wrong: |
| 12:14.48 | cmartin | -I /usr/local/arm/3.3.1uc2/lib/gcc-lib/arm-linux/3.3.1/include |
| 12:15.57 | cmartin | i have the following in make menuconfig: |
| 12:17.33 | cmartin | (/usr/local/arm) Path to existing gcc toolchain;[*] selected toolchain is uclibc toolchain;() User defined crosscompilation prefix (override default);[*] Use gcc3 toolchain for application builds;[ ] Use gcc3 toolchain for kernel builds (NOT recommended);[*] Gcc2 toolchain available for use;[*] Gcc3 toolchain available for use;(3.3.1uc2) version of existing gcc3 toolchain;(2.95.3) version of existing gcc2 toolchain;() User field. If you dont know |
| 12:17.33 | cmartin | what it is, leave it blank. |
| 12:19.49 | mickeyl | hmm - looks good |
| 12:20.28 | mickeyl | did a full rebuild? |
| 12:20.37 | mickeyl | make mrproper recommended |
| 12:20.46 | mickeyl | because of output/kernel* |
| 12:20.47 | cmartin | k, i'll try |
| 12:20.48 | mickeyl | etc |
| 12:22.12 | cmartin | BTW: news from skeye |
| 12:22.31 | cmartin | my original contact is no longer in the company :( |
| 12:22.54 | cmartin | i'm in contact with the person who took over his tasks |
| 12:22.59 | BBMatthew | hehe someone (ie me) has messed with the ipkg generation but doesn't want to do a full clean |
| 12:23.10 | BBMatthew | thank goodness for ccache |
| 12:23.42 | prh | mickeyl: I did an mrproper on buildroot the other day and didn't realise that it would rm -rf sources/*... so I've been stuck waiting for treke to come back... until I saw your note of a mirror of sources earlier |
| 12:23.56 | cmartin | after writing about the status and what the first guy had promised us i haven't heard anything else |
| 12:24.06 | mickeyl | prh: it is always good to have a Vars.local |
| 12:24.14 | mickeyl | prh: make mrproper is not dangerous then :) |
| 12:24.17 | mickeyl | mine looks like: |
| 12:24.28 | mickeyl | export DEBIAN_MIRROR=ftp://ftp.de.debian.org/debian/pool |
| 12:24.28 | mickeyl | export SOURCEFORGE_MIRROR=http://belnet.dl.sourceforge.net/sourceforge |
| 12:24.28 | mickeyl | export SOURCE_DIR=/usr/local/projects/sources |
| 12:24.28 | mickeyl | export SOURCEFORGE_MIRROR=http://umn.dl.sourceforge.net/sourceforge |
| 12:24.39 | prh | ooh cunning |
| 12:24.46 | mickeyl | so i have the sources dir seperated from buildroot |
| 12:24.54 | mickeyl | because i have five trees which are sharing the sources dir |
| 12:25.14 | BBMatthew | don't have the room for that unfortunately |
| 12:25.27 | prh | yea absoluately |
| 12:25.38 | BBMatthew | should do it though as I have a number of hand mods to the sources so they compile |
| 12:25.44 | prh | didn't know you could do that, but then I'm a buildroot novice |
| 12:25.56 | cmartin | mickeyl: mrproper didn't help i still have a "-I /usr/local/arm/3.3.1uc2/lib/gcc-lib/arm-linux/3.3.1/include" in the kernel building |
| 12:27.02 | mickeyl | cmartin: odd. i did a combined simpad build and it worked like that. |
| 12:27.04 | cmartin | BBMatthew: but remember mrproper still removes the sources under build/ |
| 12:27.28 | BBMatthew | I just thought of that -- the whole brain in gear before speaking thing |
| 12:27.44 | cmartin | hehe |
| 12:28.02 | mickeyl | cmartin: could always do it manually. Or just copy the kernel + module packages from an older simpad build into the output/ipks and touch stamps/.kernel.binary |
| 12:28.52 | cmartin | mickeyl: i think i' change to gcc2 and after the kernel's built interrupt the process and change to uclibc - maybe that'll work |
| 12:29.14 | cmartin | that would work too |
| 12:32.39 | cmartin | mickeyl: i think i have to touch some more files in stamps - the kernel is still wanting to build |
| 12:32.51 | mickeyl | cmartin: let me check my stamps dir |
| 12:33.08 | mickeyl | mickeyl@gandalf:/usr/local/projects/opensimpad$ ls -l stamps/.kernel* |
| 12:33.08 | mickeyl | -rw-rw-r-- 1 mickeyl mickeyl 0 Sep 12 00:02 stamps/.kernel-headers.source |
| 12:33.08 | mickeyl | -rw-rw-r-- 1 mickeyl mickeyl 0 Sep 12 00:02 stamps/.kernel-headers.sourcetree |
| 12:33.08 | mickeyl | -rw-r--r-- 1 mickeyl mickeyl 0 Sep 12 00:05 stamps/.kernel.ipk |
| 12:33.08 | mickeyl | -rw-rw-r-- 1 mickeyl mickeyl 0 Sep 12 00:02 stamps/.kernel.source |
| 12:33.09 | mickeyl | -rw-rw-r-- 1 mickeyl mickeyl 0 Sep 12 00:02 stamps/.kernel.sourcetree |
| 12:33.11 | mickeyl | -rw-rw-r-- 1 mickeyl mickeyl 0 Sep 12 00:04 stamps/.kernel.staginginstall |
| 12:33.53 | cmartin | well i would probably have to touch all those modules too, right? |
| 12:34.08 | mickeyl | no |
| 12:34.18 | mickeyl | all modules are generated by the kernel package |
| 12:34.44 | mickeyl | except the external modules, of course |
| 12:34.50 | mickeyl | (orinoco,hostap,pcmcia-cs) |
| 12:44.08 | BBMatthew | heheh doom and doom2 here I come... |
| 12:44.16 | cmartin | mickeyl: i should copy those over too, since they should also be built with gcc2, right? |
| 12:44.31 | BBMatthew | on a side note how much fun are they compared to some of the modern games |
| 12:47.11 | mickeyl | cmartin: right. all modules must be built with gcc2 |
| 12:47.41 | cmartin | mickeyl: i also need the kernel images in output ;) |
| 12:47.50 | Twiun | BBMatthew: difficult to play without a mouse ;) |
| 12:47.55 | Twiun | BBMatthew: but still very playable |
| 12:49.40 | Twiun | BBMatthew: depends if you're one of those 600 frag-an-hour people or not |
| 12:50.01 | cmartin | output/kernel-version is also necessary when doing that stuff manually, mickeyl |
| 12:50.20 | mickeyl | aww right |
| 12:50.35 | mickeyl | so may be changing compilers in menuconfig seems to be easier when not having multiple trees :D |
| 12:53.29 | cmartin | mickeyl: are you sure that is no problem having the kernel compiled with gcc2 and the rest with uclibc? |
| 12:53.51 | cmartin | BTW: did you ever try to boot your uclibc build? |
| 12:54.24 | mickeyl | cmartin: we are doing this always for corgi/shepherd/husky and also for collie sometimes |
| 12:54.36 | cmartin | k |
| 12:54.54 | mickeyl | cmartin: i tried to boot it (over nfs-over-usb) but that didn't work due to usb problems. |
| 12:55.09 | mickeyl | cmartin: strangely booting the glibc image worked - as if the kernel would be somehow affected |
| 12:55.15 | mickeyl | i didn't test more |
| 12:55.28 | cmartin | ah, so it may work after all when i install the image onto the device |
| 12:55.37 | mickeyl | it may _just_ work |
| 12:55.45 | mickeyl | (but given may experience in CS, i doubt ;) |
| 12:56.25 | mickeyl | s/may/my |
| 12:57.59 | cmartin | mickeyl: what's "CS" |
| 12:58.06 | mickeyl | computer science |
| 12:58.10 | cmartin | ah |
| 13:03.15 | BBMatthew | when it fails on build image needing opf-vera how do I make it build the ipk as its not in the ipks dir? |
| 13:03.23 | BBMatthew | qpf even ? |
| 13:04.37 | BBMatthew | cant seem to find the package they owuld be part of to build |
| 13:04.55 | mickeyl | qpf is done by qte |
| 13:04.58 | mickeyl | no |
| 13:05.05 | mickeyl | vera is done by vera |
| 13:05.10 | mickeyl | qpf-bitstream-vera or so |
| 13:05.30 | BBMatthew | cheers |
| 13:05.30 | mickeyl | ozmake vera-fonts-qpf |
| 13:10.01 | prh | bah qte is failing to build again |
| 13:11.14 | prh | gzip -dc /home/paul/Z/d/buildroot-oz/sources/qt-embedded-2.3.7.tar.gz | tar x |
| 13:11.18 | prh | /home/paul/Z/d/buildroot-oz/scripts/patcher -p 0 -i /home/paul/Z/d/buildroot-oz/build/opie/qt/qte234-for-opie091-gfxraster.patch |
| 13:11.21 | prh | NOTE: import patch '/home/paul/Z/d/buildroot-oz/build/opie/qt/qte234-for-opie091-gfxraster.patch' as 'qte234-for-opie091-gfxraster' |
| 13:11.24 | prh | NOTE: patch with name 'qte234-for-opie091-gfxraster' already applied |
| 13:11.26 | prh | ERROR: patch did not apply, try again with force (-f -i ...) |
| 13:11.29 | prh | Patch had errors |
| 13:12.26 | prh | any ideas? |
| 13:14.58 | mickeyl | oznew qte |
| 13:15.37 | Twiun | mickeyl: loving those bash functions eh? |
| 13:15.51 | mickeyl | hehe, yeah |
| 13:16.08 | mickeyl | prh: your build dir may be polluted from earlier builds |
| 13:16.11 | cmartin | makes things a lot easier |
| 13:16.54 | prh | mickeyl: I did a make mrproper |
| 13:17.12 | *** join/#openembedded chowells (pp0u20d9@pc232032.pctc.liv.ac.uk) |
| 13:17.44 | mickeyl | odd |
| 13:19.22 | prh | and oznew doesnt fix it |
| 13:20.14 | mickeyl | out of ideas then |
| 13:20.18 | mickeyl | afaik nothing changed in qt-2.3.7 |
| 13:20.21 | mickeyl | and opie |
| 13:20.32 | mickeyl | since the last time i build |
| 13:20.56 | prh | should I have a qt-2.3.2 in build ? |
| 13:21.04 | cmartin | prh : remove the build manually - rm -rf build/qt-2.3.7 - and run make again |
| 13:21.15 | cmartin | prh: yes |
| 13:21.16 | prh | cmartin: done that |
| 13:22.28 | cmartin | prh: and - rm -rf stamps/.qt-embedded.* |
| 13:23.38 | cmartin | prh: wait when did you pull the last time - i had exactly the same problem a couple of days ago |
| 13:23.57 | cmartin | it was a problem with patcher |
| 13:24.34 | cmartin | mickeyl: remember? |
| 13:27.03 | *** join/#openembedded schurig (~schurig@pD958AC2C.dip.t-dialin.net) |
| 13:27.03 | BBMatthew | where does opie put itself in the tree |
| 13:27.44 | cmartin | BBMatthew: sources/opie , build/opie |
| 13:27.59 | prh | cmartin: I've done a fresh pull - no change (nothing to pull |
| 13:28.29 | BBMatthew | on the zaurus fs I should have said sorry, and if tiny login is installed shouldn't it have started opie already? |
| 13:28.34 | prh | paul@wacka:~/Z/d/buildroot-oz$ bk pull; bk -r co -q |
| 13:28.34 | prh | Pull http://openzaurus.bkbits.net/buildroot |
| 13:28.34 | prh | <PROTECTED> |
| 13:28.34 | prh | Nothing to pull. |
| 13:29.06 | cmartin | k |
| 13:31.07 | cmartin | prh: make SUBDIRS=packages/opie mrproper;make SUBDIRS=packages/qte mrproper;make SUBDIRS=packages/qte |
| 13:33.08 | BBMatthew | no opie login DOH! |
| 13:35.12 | prh | cmartin: same error :O0 |
| 13:35.12 | mickeyl | cmartin: i thought you had a problem with "can't patch file twice", not "already patched" |
| 13:35.30 | prh | /home/paul/Z/d/buildroot-oz/scripts/patcher -p 0 -i /home/paul/Z/d/buildroot-oz/build/opie/qt/qte234-for-opie091-gfxraster.patch |
| 13:35.33 | prh | NOTE: made /home/paul/Z/d/buildroot-oz/build/qt-2.3.7 to projectdir by creating .patches/ |
| 13:35.36 | prh | NOTE: import patch '/home/paul/Z/d/buildroot-oz/build/opie/qt/qte234-for-opie091-gfxraster.patch' as 'qte234-for-opie091-gfxraster' |
| 13:35.39 | prh | /home/paul/Z/d/buildroot-oz/scripts/patcher -p 0 -i /home/paul/Z/d/buildroot-oz/build/opie/qt/qte234-for-opie091-listview.patch |
| 13:35.42 | prh | NOTE: import patch '/home/paul/Z/d/buildroot-oz/build/opie/qt/qte234-for-opie091-listview.patch' as 'qte234-for-opie091-listview' |
| 13:35.46 | prh | /home/paul/Z/d/buildroot-oz/scripts/patcher -p 0 -i /home/paul/Z/d/buildroot-oz/build/opie/qt/qte234-for-opie091-makefile.patch |
| 13:35.48 | prh | NOTE: import patch '/home/paul/Z/d/buildroot-oz/build/opie/qt/qte234-for-opie091-makefile.patch' as 'qte234-for-opie091-makefile' |
| 13:35.51 | prh | ERROR: patch did not apply, try again with force (-f -i ...) |
| 13:35.54 | prh | Patch had errors |
| 13:35.56 | mickeyl | uhm |
| 13:36.08 | mickeyl | build opie BRANCH_1_0, not v1_0 |
| 13:36.23 | mickeyl | makefile-patch is obsolete and has been removed in current cvs |
| 13:38.26 | prh | I refetch the whole buildroot |
| 13:38.50 | mickeyl | ? |
| 13:38.53 | mickeyl | unneccessary |
| 13:38.59 | cmartin | prh: you just need to cvs up opie |
| 13:39.02 | mickeyl | cvs update -rBRANCH_1_0 |
| 13:39.10 | mickeyl | -Pd of course |
| 13:39.20 | prh | too late - it's all gone |
| 13:39.27 | *** join/#openembedded chouimat (~root@r2351064.cidc.net) |
| 13:40.51 | chouimat | morning |
| 13:41.54 | mickeyl | hi ChanServ |
| 13:41.56 | mickeyl | argh |
| 13:41.59 | mickeyl | ~lart Tab Completion |
| 13:42.03 | mickeyl | hi chouimat |
| 13:45.26 | BBMatthew | for opie on the zaurus you just need opie? not qtopia? |
| 13:46.02 | cmartin | BBMatthew: qtopia is the desktop progi |
| 13:46.11 | cmartin | BBMatthew: for your PC |
| 13:46.40 | cmartin | BBMatthew: so yes only opie and qte |
| 13:46.47 | mickeyl | actually qtopia is not qtopia desktop :) |
| 13:46.49 | BBMatthew | thats what I though |
| 13:46.50 | mickeyl | qtopia is qtopia |
| 13:46.56 | mickeyl | the TT PDA environment |
| 13:47.20 | BBMatthew | I am using the collie image defs instead as I recon I am missing some of the packages |
| 13:47.27 | cmartin | oh |
| 13:47.38 | mickeyl | BBMatthew: you're building for poodle? |
| 13:47.42 | BBMatthew | well I am missing opie |
| 13:47.49 | BBMatthew | in the initrd. |
| 13:48.58 | BBMatthew | which explains the boot prompt |
| 13:51.04 | mickeyl | ipkg install task-opie-minimal |
| 13:54.11 | BBMatthew | oh dear the inird is missing libgcc |
| 13:54.26 | cmartin | mickeyl: i'm still not getting why the 3.3.1uc2 toolchain is getting included while building the kernel in a combined configuration |
| 13:55.19 | cmartin | /usr/bin/distcc arm-linux-gcc -D__KERNEL__ -I/home/simpad/opensimpad-uclibc/build/linux-2.4.19/include -Wall -Wstrict-prototypes -Wno-trigraphs -Os -fno-strict-aliasing -fno-common -Uarm -fno-common -pipe -mapcs-32 -D__LINUX_ARM_ARCH__=4 -march=armv4 -mtune=strongarm1100 -mshort-load-bytes -msoft-float -Uarm -nostdinc -I /usr/local/arm/3.3.1uc2/lib/gcc-lib/arm-linux/3.3.1/include -DKBUILD_BASENAME=blkpg -DEXPORT_SYMTAB -c blkpg.c |
| 13:57.48 | mickeyl | cmartin: hold on, i start a build here |
| 14:00.10 | *** join/#openembedded prpplague (~billybob@12.148.134.14) |
| 14:01.18 | mickeyl | cmartin: works perfect here |
| 14:01.25 | mickeyl | cmartin: I'll send 'ya my config |
| 14:01.40 | cmartin | k |
| 14:02.36 | mickeyl | underway |
| 14:03.08 | cmartin | got it |
| 14:04.28 | BBMatthew | can you tell me what package libgcc would be in? |
| 14:04.34 | cmartin | k, one difference - you have "[ ] selected toolchain is uclibc toolchain" - and i have it selected |
| 14:05.09 | *** join/#openembedded sieve_work (~klu5ter@12.148.134.14) |
| 14:05.54 | mickeyl | cmartin: that shouldn't make a difference - that setting has no effect yet ;) |
| 14:05.59 | cmartin | hi kergoth |
| 14:06.01 | BBMatthew | I am so doing a make mrproper make tonight |
| 14:06.03 | mickeyl | moin kergoth |
| 14:06.13 | cmartin | hmm |
| 14:06.41 | mickeyl | cmartin: D'oh! |
| 14:06.51 | mickeyl | cmartin: it includes the 3.x toolchain path here also |
| 14:06.55 | kergoth | hey |
| 14:06.56 | mickeyl | cmartin: but it doesn't matter |
| 14:07.01 | chouimat | hi kergoth |
| 14:07.44 | cmartin | still get: |
| 14:07.47 | cmartin | blkpg.c: In function `blk_ioctl': |
| 14:07.47 | cmartin | blkpg.c:252: error: asm-specifier for variable `__r1' conflicts with asm clobber list |
| 14:07.47 | cmartin | blkpg.c:252: confused by earlier errors, bailing out |
| 14:07.47 | cmartin | distcc[17503] ERROR: compile on localhost failed with exit code 1 |
| 14:07.47 | cmartin | make[5]: *** [blkpg.o] Fehler 1 |
| 14:08.14 | mickeyl | cmartin: it is not taking the 2.9x toolchain on your system |
| 14:08.19 | kergoth | cmartin: that kernel needs to be patched to be able to build with gcc3. |
| 14:08.20 | mickeyl | cmartin: don't ask me why. it is using it here |
| 14:08.57 | mickeyl | kergoth: cmartin attempted to do a combined build, but it seems it always takes the 3 toolchain on his system |
| 14:09.06 | mickeyl | (also for kernel, that is) |
| 14:09.28 | kergoth | ahh yeah. |
| 14:09.42 | cmartin | could it be a PATH thing? |
| 14:09.57 | mickeyl | cmartin: urghs. don't tell me you have a toolchain path in your PATH? |
| 14:10.23 | cmartin | yep - always have before, without probs |
| 14:10.26 | mickeyl | ibot: lart cmartin |
| 14:10.46 | mickeyl | that wasted my precious brain cycles.... ;) |
| 14:10.47 | cmartin | but now with a combined config that may be the prob now |
| 14:11.05 | mickeyl | may? of course it is. |
| 14:11.10 | cmartin | sorry |
| 14:11.13 | kergoth | technically that shouldnt matter |
| 14:11.25 | BBMatthew | hmm to get the dual compile to work you need a compiler in the path at several points :) |
| 14:11.27 | mickeyl | the toolchain path _should_ be prepended |
| 14:11.33 | BBMatthew | the kernel / apps |
| 14:11.37 | mickeyl | but that may don't work always |
| 14:11.44 | kergoth | err, it is prepended. but this is a system PATH, which finds the last version of a thing, not the first, mickeyl |
| 14:11.46 | mickeyl | doesn't even |
| 14:11.50 | kergoth | i.e. /bin:/usr/bin:/usr/local/bin |
| 14:11.57 | schurig | http://www.linuxdevices.com/news/NS5890766843.html |
| 14:12.04 | schurig | kergoth: hi! |
| 14:12.04 | kergoth | hey schurig |
| 14:12.04 | mickeyl | ah ye |
| 14:12.09 | schurig | kergoth: heard you're busy with OE :-) |
| 14:12.16 | kergoth | mickeyl: so thats the problem, our PATH mangling is incorrect |
| 14:12.19 | kergoth | mickeyl: in Vars.make |
| 14:12.29 | kergoth | schurig: yup, builds quite a bit of stuff nowadays |
| 14:12.36 | mickeyl | kergoth: i see. cmartin: feel free to correct |
| 14:13.18 | schurig | hmm, Intel's PXA team did not really work well with the OpenSource community so far, see the hickhack with RedBoot on Lubbock ... |
| 14:14.30 | cmartin | mickeyl: i just removed the X-compiler from the system PATH - let's see if that helps |
| 14:15.14 | kergoth | s/s$// |
| 14:15.16 | cmartin | mickeyl: nope - now it even bails out earlier |
| 14:15.33 | cmartin | make[4]: arm-linux-gcc: Kommando nicht gefunden |
| 14:15.33 | cmartin | hehe |
| 14:16.00 | mickeyl | hmm |
| 14:16.03 | kergoth | sounds like your TOOLCHAINPATH isnt set right |
| 14:16.04 | mickeyl | that shouldn't happen |
| 14:16.09 | mickeyl | check your TOOLCHAINPATH |
| 14:16.14 | cmartin | for those not speaking german that says - make[4]: arm-linux-gcc: Command not found |
| 14:16.28 | schurig | kergoth: it build the 58 or so *.oe files? |
| 14:17.02 | kergoth | schurig: yep. excepting perhaps the ppc and alou chains, which have collected a bit of cruft, and ncurses-4.2 which needs some TLC |
| 14:17.03 | cmartin | mickeyl: i thought that was set at build time |
| 14:17.09 | kergoth | schurig: the rest build fine |
| 14:17.12 | prh | kergoth: why you leave d-handhelds? |
| 14:17.31 | kergoth | prh: because its completely inactive and not likely to go anywhere anytime soon |
| 14:17.34 | schurig | alou suffered from bit rotting desease |
| 14:17.35 | mickeyl | cmartin: is your toolchain in /usr/local/arm ? |
| 14:17.43 | schurig | kergoth: what's TLC? |
| 14:17.45 | kergoth | schurig: yeh. gb2 and I are working on the toolchain builds using OE |
| 14:17.51 | kergoth | schurig: he's working on the uclibc one |
| 14:17.53 | cmartin | yep |
| 14:17.55 | kergoth | schurig: tlc = tender loving care |
| 14:18.04 | cmartin | both of them |
| 14:18.06 | mickeyl | cmartin: then i'm out of ideas |
| 14:18.14 | schurig | kergoth: I had good success with Dan Kegel's script. The resulting gcc3.3 compiler is my compiler of choice for Linux 2.4 and 2.6 now |
| 14:18.31 | cmartin | i don't have a TOOLCHAINPATH |
| 14:18.35 | kergoth | cmartin: eh? |
| 14:18.38 | kergoth | cmartin: make menuconfig. |
| 14:18.45 | kergoth | cmartin: build options |
| 14:19.06 | cmartin | never have, but that wasn't a prob before since i had the toolchain in my system path |
| 14:19.08 | gb2 | kergoth: i was thinking about how scripting is handled now. and thinking about how metadata should be handled. |
| 14:19.16 | prh | kergoth: fair enough |
| 14:19.29 | cmartin | (/usr/local/arm) Path to existing gcc toolchain |
| 14:19.32 | mickeyl | cmartin: then it has been wrong for you since ever |
| 14:19.42 | mickeyl | cmartin: if it doesn't work without local path hacks, it is broken somehow |
| 14:19.54 | kergoth | gb2: come up with anything so far? I gave some brief thought to metadata handling and algos for its access in a new core lib |
| 14:19.59 | prh | kergoth: know if the buildroot snapshots are working again? (I want to setup an arch repository of buildroot for myself and anyone else who wants to use buildroot but can't use bk) |
| 14:20.08 | gb2 | seems like what we need is a big list of "exported" vars, and those get put as variables in the native script environment. and they are recovered in the "normal termination" of a script to pickup any changes. |
| 14:20.21 | gb2 | this helps to compartmentalize scripts somewhat |
| 14:20.48 | kergoth | gb2: recovered.. so allow functions to manipulate metadata at the higher level? interesting idea. |
| 14:21.10 | kergoth | gb2: the former "big list of exported vars" we have now, in the form of an 'export' function metadata parameter |
| 14:21.26 | kergoth | prh: snaps are still toast due to navi being down. i'll throw up ones done manually today |
| 14:21.44 | prh | kergoth: could you snap oe too please? |
| 14:21.46 | kergoth | yeah |
| 14:21.52 | gb2 | i'm thinking you'd emit a setup and an oe_exit() function (and you'd by default call that if you fall through the bottom of a script). |
| 14:22.07 | prh | if I got a good arch setup running - whould you at least have a look at it? |
| 14:22.08 | gb2 | that oe_exit would walk over the list of "exported" vars and pull their current values up into the higher levels |
| 14:22.26 | kergoth | course right now we use 'set -e' to manage automatic failure.. we'd need some other method to ensure fallthrough occurs |
| 14:23.12 | kergoth | do_install () { ( set -e; stuff ) || return 1 } would probably be sufficient |
| 14:23.20 | gb2 | you could just put a call to the "normal" exit function at the bottom of the executed script |
| 14:23.41 | kergoth | you're missing the point. set -e results in _immediate_ shell execution if any call returns nonzero |
| 14:23.45 | kergoth | not fallthrough. exit. |
| 14:23.52 | kergoth | s/execution/exit/ |
| 14:23.57 | gb2 | yeah. non-normal exit |
| 14:24.00 | gb2 | don't recover metadata changes |
| 14:24.19 | kergoth | just saying you'd need that subshell to retain that functionality. no biggie. |
| 14:24.44 | Twiun | kergoth: toplevel oebuild says 'cannot stat /usr/share/...' any idea? |
| 14:25.05 | kergoth | Twiun: you didnt set OEDIR to the oe repos, or you installed the OE rpm |
| 14:25.11 | gb2 | i was actually thinking that you'd pass the script something like $OE_VARS and in normal termination it would write out $OE_VARS_FILE with all of $OE_VARS in it. |
| 14:25.11 | kergoth | our default OEPATH ordering isnt quite correct |
| 14:25.22 | gb2 | which makes it easy to tell if you terminated via oe_exit or not |
| 14:25.27 | Twiun | ahh, so OEDIR has to be set by hand? |
| 14:25.44 | kergoth | Twiun: unless you're building from within the oe repos, yes. |
| 14:25.47 | kergoth | Twiun: default OEDIR=`pwd` |
| 14:26.05 | Twiun | kergoth: I was |
| 14:26.11 | kergoth | i have no problems doing so here. |
| 14:26.12 | cmartin | mickeyl: what does your TOOLCHAINPATH look like? |
| 14:26.14 | Twiun | kergoth: /home/twiun/oe |
| 14:26.56 | Twiun | kergoth: just worked fine with OEDIR explicitly set |
| 14:27.04 | Twiun | kergoth: now to try oemake again |
| 14:27.31 | Twiun | hmmm, dependency patcher unsatisfied |
| 14:28.00 | kergoth | heh, good call, it is setting the default OEDIR to /usr/share/oe, which isnt problematic unless you try to build oe.oe, which references it directly |
| 14:28.08 | kergoth | its time to kill OEDIR in favor of OEPATH |
| 14:28.12 | gb2 | enable_bypass() { set_bit(reg, DISABLE_BYPASS_BIT); } |
| 14:28.22 | kergoth | Twiun: oemake content/patcher*.oe content/blah.oe |
| 14:28.35 | kergoth | Twiun: it obeys deps only on things you tell it where to find |
| 14:28.42 | kergoth | Twiun: it doesnt have the ability to locate oe files on its own |
| 14:29.37 | Twiun | aha! |
| 14:29.51 | Twiun | kergoth: I was trying oemake without args |
| 14:29.58 | kergoth | silly thing, i need to seperate its knowledge of what exists from its knowledge of what you want. |
| 14:31.37 | *** join/#openembedded jumpkick (~chatzilla@CPE0080c6f0e7ba-CM014500002511.cpe.net.cable.rogers.com) |
| 14:31.37 | *** join/#openembedded prpplague (~billybob@12.148.134.14) |
| 14:31.37 | *** join/#openembedded exastra (~go@l5.org) [NETSPLIT VICTIM] |
| 14:31.37 | *** join/#openembedded aman (~aman@h-69-3-87-233.MCLNVA23.covad.net) [NETSPLIT VICTIM] |
| 14:31.37 | *** join/#openembedded mickeyl (mickeyl@gandalf.tm.informatik.uni-frankfurt.de) [NETSPLIT VICTIM] |
| 14:31.37 | *** join/#openembedded mewyn (~mike@dsl081-228-056.chi1.dsl.speakeasy.net) |
| 14:31.37 | *** join/#openembedded killefiz (~sven@ssv-niederkruechten.de) [NETSPLIT VICTIM] |
| 14:31.37 | *** mode/#openembedded [+bbb *!*@12-227-17-69.client.attbi.com *!*@*12-227-27-36.client.attbi.com *!*@12-227-27-36.client.attbi.com]] by calvino.freenode.net |
| 14:31.56 | Twiun | wasn't me! |
| 14:32.09 | chouimat | hi bart |
| 14:32.38 | cyn_ | Twiun: your idiot stressing of OE spewed over to IRC |
| 14:33.06 | Twiun | cyn_: now that's what I call a pervasive test :) |
| 14:34.16 | kergoth | heh, tom gall's toolchain .oe's dont have a inhibit_glibc first pass. ah, he's doing ppc to ppc.. that makes sense |
| 14:34.46 | kergoth | gb2: you have gcc .oe's handy? |
| 14:35.34 | gb2 | kergoth: didn't get that finished last night |
| 14:35.40 | kergoth | k |
| 14:35.57 | kergoth | figured i'd test a i686/cygwin->arm/linux gcc run |
| 14:36.01 | kergoth | s/run/build/ |
| 14:36.12 | gb2 | fun |
| 14:36.27 | kergoth | may as well, before i boot back into linux |
| 14:37.45 | cmartin | mickeyl: if i don't arm-linux-gcc in my PATH it's not found at all |
| 14:37.53 | cmartin | s/don't put |
| 14:38.04 | kergoth | cmartin: what is the gcc3ver set to in build options? |
| 14:38.22 | cmartin | 3.3.1uc2 |
| 14:38.34 | kergoth | does /usr/local/arm/3.3.1uc2 exist? |
| 14:38.35 | kergoth | i think not. |
| 14:38.40 | kergoth | change that to 3.3 |
| 14:38.48 | cmartin | found under /usr/local/arm/3.3.1uc2 |
| 14:38.59 | kergoth | and /usr/local/arm/3.3.1uc2/bin/arm-linux-gcc exists? |
| 14:39.02 | chouimat | 3.3.1uc2? |
| 14:39.04 | kergoth | thats hte path it was _built_ into? |
| 14:39.08 | kergoth | you cant relocate a toolchain. |
| 14:39.22 | cmartin | yep |
| 14:39.27 | mickeyl | 3.3.1uc2 was the path it has been built into |
| 14:39.29 | kergoth | then your system is just fucked. |
| 14:39.34 | kergoth | i'd say. |
| 14:39.40 | kergoth | echo $PATH in the binary target |
| 14:39.52 | mickeyl | + export |
| 14:40.03 | kergoth | ? |
| 14:40.16 | mickeyl | let's see all env. variables |
| 14:40.22 | kergoth | ah |
| 14:40.38 | cmartin | at build time? |
| 14:40.46 | kergoth | eh? |
| 14:40.49 | kergoth | i said echo $PATH |
| 14:40.56 | kergoth | in the binary target |
| 14:40.59 | kergoth | was i ambiguous? |
| 14:42.25 | chouimat | wtf is 3.3.1uc2? |
| 14:42.36 | kergoth | chouimat: uclibc, most likely |
| 14:42.43 | cmartin | " in the binary target" - sorry don't understand what you want |
| 14:42.54 | cmartin | chouimat: uclibc |
| 14:43.00 | chouimat | kergoth ok ... hmmm me need to reinstall bk since I rebuild this box |
| 14:43.08 | kergoth | $(stamp-binary) ~= binary |
| 14:43.16 | kergoth | target = [something]:<newline> |
| 14:43.23 | kergoth | therefore, binary target is... |
| 14:43.25 | kergoth | $(stamp-binary): |
| 14:43.42 | kergoth | vi packages/what_you're_trying_to_fucking_build/Makefile |
| 14:43.48 | cmartin | k |
| 14:44.05 | cmartin | i understand |
| 14:48.48 | mickeyl | code red? sounds like an army expression |
| 14:49.03 | mickeyl | is that one of the new designer drinks? |
| 14:49.09 | kergoth | mt dew code red |
| 14:49.15 | kergoth | cherry flavored highly caffinated |
| 14:49.24 | kergoth | why do people always insist on emailing me directly |
| 14:49.30 | kergoth | mailing lists exist for a fucking reason |
| 14:49.30 | mickeyl | ha! |
| 14:49.35 | mickeyl | can image you hating that |
| 14:50.27 | chouimat | kergoth don't give you home phone number :) |
| 14:50.30 | BBMatthew | hmm a hidden broken qte |
| 14:50.38 | mickeyl | ? |
| 14:50.56 | kergoth | mickeyl: added bootmenu to OE.. someone should test it out |
| 14:51.00 | BBMatthew | in *my* buildroot - qte was broken but I didn't know |
| 14:51.08 | mickeyl | kergoth: the one we talked about yesterday? |
| 14:51.12 | BBMatthew | thank god for ccache |
| 14:51.15 | kergoth | mickeyl: yes |
| 14:51.16 | mickeyl | kergoth: cool. i'll give it a try |
| 14:52.31 | BBMatthew | must learn the command for irc emotes at somep point |
| 14:52.46 | kergoth | <PROTECTED> |
| 14:53.04 | BBMatthew | cheers |
| 14:53.51 | BBMatthew | I rmed the staginginstall stamps to force a rebuild of the ipks and discovered that qte didn't exists |
| 14:54.02 | BBMatthew | which explains why it wont boot opie |
| 14:54.33 | BBMatthew | and also explains why I should leave buildroot well alone |
| 14:54.47 | kergoth | hehe. its a quirky beast |
| 14:54.48 | cmartin | mickeyl: ok you wanted it: |
| 14:55.02 | cmartin | mickeyl: here's my PATH first: |
| 14:55.08 | cmartin | /home/simpad/opensimpad-uclibc/scripts:/home/simpad/opensimpad-uclibc/output/staging/build/bin:/home/simpad/opensimpad-uclibc/scripts:/home/simpad/opensimpad-uclibc/output/staging/build/bin:/bin:/sbin:/usr/bin:/usr/sbin:/usr/local/bin:/usr/local/sbin:/usr/X11R6/bin:/opt/bitkeeper-3.0 |
| 14:55.28 | cmartin | env is pretty much which part would you prefer? |
| 14:55.34 | mickeyl | none |
| 14:55.45 | mickeyl | that is enough |
| 14:55.56 | mickeyl | toolchain paths are missing at all |
| 14:56.04 | cmartin | as we see there's no CC in that PATH |
| 14:56.53 | mickeyl | cmartin: Vars.make doesn't work for you |
| 14:57.07 | mickeyl | you must check out why it doesn't. i never seen it failing so obviously |
| 14:57.52 | mickeyl | did you have to add your local path also when using other distributions? |
| 14:59.00 | BBMatthew | I should have done the rm staging install after I found buzybox was missing |
| 14:59.03 | chouimat | hmm ccache doesn't seem to work here |
| 14:59.07 | cmartin | yes |
| 15:02.18 | *** join/#openembedded TimRiker (~timr@proxyle02.ext.ti.com) |
| 15:03.45 | cmartin | mickeyl: can you send me your Vars.make so i can compair if somethings broke in mine |
| 15:05.16 | kergoth | bk diffs Vars.make |
| 15:05.17 | kergoth | heh |
| 15:08.49 | cyn_ | well, I think todays gunna be a 'dick around with php' day instead of poke at buildroot - I gots an interview tomorrow for a little 4mo stint and would hate to lose it for lack of familiarity |
| 15:08.56 | mickeyl | cmartin: i have rev 1.124 |
| 15:09.07 | mickeyl | cmartin: you can get that with bk get |
| 15:10.05 | *** join/#openembedded Twiun_ (~twiun@adorphuye.com) |
| 15:13.50 | BBMatthew | YAY opie started |
| 15:14.19 | kergoth | congrats |
| 15:14.42 | BBMatthew | thanks go to you i think :) |
| 15:29.42 | kergoth | s/$/ to OE/ |
| 15:34.32 | cmartin | mickeyl: Vars.make 1.125: 373 lines |
| 15:35.37 | mickeyl | cmartin: try using 1.124 |
| 15:35.57 | cmartin | how do i get it |
| 15:36.07 | mickeyl | bk helptool --> bk get |
| 15:39.19 | cmartin | mickeyl: want to see the diff between the two? (it's not much) |
| 15:39.32 | cmartin | [opensimpad-DEV-uclibc] opensimpad-uclibc # diff Vars.make /tmp/Vars.make |
| 15:39.35 | cmartin | 154,156d153 |
| 15:39.35 | cmartin | < ifdef CONFIG_GCC-3.4 |
| 15:39.35 | cmartin | < GCC3VER=3.4 |
| 15:39.35 | cmartin | < endif |
| 15:39.35 | cmartin | 193c190 |
| 15:39.35 | cmartin | < export PATH:=$(if $(CONFIG_TOOLCHAIN),$(CROSS_DIR)/bin:)$(SCRIPTS_DIR):$(STAGING_BINDIR):$(PATH) |
| 15:39.39 | cmartin | --- |
| 15:39.41 | cmartin | > export PATH:=$(if $(CROSS),$(CROSS_DIR)/bin:)$(SCRIPTS_DIR):$(STAGING_BINDIR):$(PATH) |
| 15:39.55 | cmartin | /tmp/Vars.make is 1.124 |
| 15:40.00 | kergoth | heh, that explains it |
| 15:40.04 | kergoth | my bad. |
| 15:40.05 | mickeyl | kergoth: did you break it? |
| 15:40.08 | kergoth | yep |
| 15:40.11 | mickeyl | hehe |
| 15:40.19 | kergoth | we need it to append if CONFIG_TOOLCHAIN _or_ CROSS |
| 15:41.28 | BBMatthew | Is the opie quicklauncher working? |
| 15:41.35 | kergoth | yes |
| 15:42.06 | BBMatthew | something is wrong with either aqpkg or quicklauncher |
| 15:42.14 | BBMatthew | am working on it |
| 15:42.19 | kergoth | aqpkg |
| 15:42.22 | kergoth | is your opie tree current? |
| 15:42.29 | cmartin | yes, now we're rockin' |
| 15:42.46 | BBMatthew | not sure |
| 15:42.48 | cmartin | kernel is compiling with 2.95.3 |
| 15:43.07 | BBMatthew | I could do a make mrproper |
| 15:43.11 | BBMatthew | and pulll it again |
| 15:43.26 | kergoth | export PATH:=$(PATH):$(SCRIPTS_DIR):$(STAGING_BINDIR):$(if $(CONFIG_TOOLCHAIN)$(CROSS),:$(CROSS_DIR)/bin) |
| 15:43.31 | kergoth | would be the correct fix |
| 15:43.52 | cmartin | BBMatthew: just update the cvs tree in source/opie |
| 15:44.07 | BBMatthew | no problem |
| 15:44.08 | kergoth | and build/opie, or clean and re"extract" after doing so |
| 15:44.36 | BBMatthew | am tempted by a make mrproper make |
| 15:44.44 | BBMatthew | esp on *my* buildroot |
| 15:44.52 | kergoth | may as well |
| 15:44.55 | kergoth | as long as you have ccache of course |
| 15:45.00 | BBMatthew | indeed :) |
| 15:45.28 | cmartin | BBMatthew: well it will only work if you don't have SOURCESDIR set in Vars.local |
| 15:45.48 | *** join/#openembedded TimRiker (~timr@proxyle01.ext.ti.com) |
| 15:45.53 | BBMatthew | i dont |
| 15:46.38 | cmartin | s/ SOURCE_DIR |
| 15:46.50 | *** join/#openembedded DanWilliams (~DanWillia@proxyle01.ext.ti.com) |
| 15:47.01 | cmartin | bbl |
| 15:47.04 | *** part/#openembedded cmartin (~cmartin@gw.krypnet.de) |
| 15:47.22 | BBMatthew | coming down the pipe 5 by 5 |
| 15:49.01 | BBMatthew | the root jffs on a poodle is 32MB in size? |
| 15:49.09 | kergoth | yep |
| 15:49.15 | kergoth | it has 64mb of flash, split into two 32mb partitions |
| 15:49.49 | BBMatthew | which would explain how I buggered the initrd with konqueror, xmms and I think sdl |
| 15:50.02 | kergoth | too much stuff? |
| 15:50.31 | BBMatthew | yep did that a while ago, no space left even to delete anything |
| 15:50.37 | kergoth | hehe |
| 15:50.40 | kergoth | hate that |
| 15:50.50 | BBMatthew | I am using a minimal image and a feed on my departments server to put the rest on |
| 15:51.09 | kergoth | remember minimal doesnt include opie at all |
| 15:51.28 | BBMatthew | that was with the 2.95.3 xscale build this is the 3.3.1 armv5te xscale feed |
| 15:51.40 | BBMatthew | sorry wrong choice of syntax |
| 15:51.42 | BBMatthew | :) |
| 15:51.49 | BBMatthew | smallest with opie |
| 15:52.02 | kergoth | gotcha |
| 15:56.08 | treke | mornin k-dog |
| 15:56.23 | kergoth | dont remind me |
| 15:56.27 | kergoth | sup treke |
| 16:00.12 | mickeyl | BBMatthew: if you have a working poodle image with opie, post it anywhere so we can upload it to unofficial/experimental/poodle |
| 16:00.23 | mickeyl | s/anywhere/somewhere |
| 16:02.28 | BBMatthew | ok, well I find a lot of stuff isn't prefixing the binaries, opie didn't just there |
| 16:02.51 | BBMatthew | it seems to be a feature of using gcc 3.3.1 and 2.95.3 at the same time |
| 16:02.52 | *** join/#openembedded qip|Virus|here (~Virus@pD9E65165.dip.t-dialin.net) |
| 16:03.15 | kergoth | BBMatthew: scroll up, read the discussion with cmartin and subsequent fix |
| 16:07.26 | BBMatthew | kergoth: cheers must have been playing with my lovely new OZ install |
| 16:07.30 | BBMatthew | kergoth: and missed it |
| 16:09.16 | prh | gzip -dc /home/paul/Z/sources/qt-embedded-2.3.7.tar.gz | tar x |
| 16:09.16 | prh | /home/paul/Z/d/buildroot-oz/scripts/patcher -p 0 -i /home/paul/Z/d/buildroot-oz/build/opie/qt/qte234-for-opie091-gfxraster.patch |
| 16:09.19 | prh | NOTE: import patch '/home/paul/Z/d/buildroot-oz/build/opie/qt/qte234-for-opie091-gfxraster.patch' as 'qte234-for-opie091-gfxraster' |
| 16:09.22 | prh | NOTE: patch with name 'qte234-for-opie091-gfxraster' already applied |
| 16:09.25 | prh | ERROR: patch did not apply, try again with force (-f -i ...) |
| 16:09.28 | prh | Patch had errors |
| 16:09.30 | prh | make[1]: *** [/home/paul/Z/d/buildroot-oz/stamps/.qt-embedded.sourcetree] Error 1 |
| 16:09.33 | prh | any ideas? |
| 16:09.56 | BBMatthew | yep |
| 16:10.07 | BBMatthew | make SUBDIRS=clean |
| 16:10.51 | BBMatthew | or ozclean |
| 16:12.28 | BBMatthew | and if you must a make mrproper |
| 16:12.28 | BBMatthew | it happens when buildroot gets confused |
| 16:16.09 | mickeyl | yeah right |
| 16:16.27 | mickeyl | if somehow the buildprocess stops during one package you're best off 'oznew'ing' that package |
| 16:16.34 | mickeyl | bbl |
| 16:18.29 | kergoth | i should reimplement that again |
| 16:18.41 | BBMatthew | mickeyl: I hope that was agreement :) |
| 16:19.28 | BBMatthew | kergoth: How long has the buildroot been in existance? |
| 16:20.02 | kergoth | hmm |
| 16:20.27 | kergoth | since before i got my 5500 i think, which was in may of the year after the 5000d dev program existed |
| 16:20.34 | kergoth | what was that, 2 or 3 years now? |
| 16:20.50 | *** join/#openembedded Twiun_ (~twiun@adorphuye.com) |
| 16:23.01 | BBMatthew | and how far has it come |
| 16:23.33 | kergoth | heh, you shouldve seen the first pass. the fetch functions were recursing shell functions that heavily used bash arrays with declare -a |
| 16:23.36 | kergoth | :) |
| 16:26.12 | BBMatthew | kergoth: you really hated yourself, didn't you? |
| 16:27.25 | kergoth | heh, it did the job |
| 16:28.56 | *** join/#openembedded pb_ (~pb@2002:5160:45ef:0:2e0:7dff:fe74:8b87) |
| 16:29.10 | kergoth | hey pb_ |
| 16:29.33 | BBMatthew | is the fix for Vars.make being pushed into bk? |
| 16:29.55 | pb_ | hi kergoth |
| 16:32.21 | pb_ | kergoth: btw, did you get the email I sent you yesterday? |
| 16:33.33 | kergoth | yeah, havent had a chance to reply to any of the mail i've gotten this week |
| 16:33.41 | pb_ | righto |
| 16:34.58 | chouimat | hmmm ccache don't work with teambuilder |
| 16:35.08 | kergoth | gcc 3.3.1 pass 2 is in OE, adding pass 1 |
| 16:35.18 | treke | probably works with distcc though |
| 16:35.43 | chouimat | treke work with distcc ... but I'm doing a comparaison between the two |
| 16:52.21 | BBMatthew | interesting opierec just died on compile with linker errors like : undefined reference to `std::ios_base::Init::~Init [in-charge]()' |
| 16:52.26 | BBMatthew | .obj/sharp-linux/main.o(.text+0x604): In function `__static_initialization_and_d |
| 16:52.38 | BBMatthew | lots of them - in fact they are terribly familiar to me after too much C++ template hell |
| 16:55.04 | kergoth | templates. bleh. |
| 16:55.30 | BBMatthew | indeed, a gcc3.3 issue? are you still building with 2.95.3? |
| 16:56.14 | kergoth | i havent used 2.95.3 for a while |
| 16:56.33 | BBMatthew | and opie builds for you? |
| 16:56.50 | kergoth | i havent built opie in weeks |
| 16:59.47 | BBMatthew | wait a sec ios_base? thats a string funcition |
| 17:02.35 | BBMatthew | I don't trust my buildroot, I am doing a mrproper clean make def config |
| 17:17.48 | *** join/#openembedded Soopaman (~Soopaman@h24-77-52-90.wp.shawcable.net) |
| 17:26.20 | gb2 | 2 |
| 17:26.23 | gb2 | oops |
| 17:27.22 | BBMatthew | can you tell me where I can get a replacement set of kernel headers as trekes site is still down? |
| 17:27.31 | treke | google |
| 17:28.11 | pb_ | treke: sounds like you need to get a better site |
| 17:28.23 | *** join/#openembedded mallum (~mallum@82-35-112-58.cable.ubr07.dals.blueyonder.co.uk) |
| 17:28.40 | treke | pb_: Yup. their getting flooded with virus support requests. No time to turn the server on :p |
| 17:28.44 | mallum | treke: have you used oz buildroot recently ? |
| 17:28.48 | treke | mallum: yes |
| 17:29.05 | mallum | treke: had any problems with the haxord libtool ? |
| 17:29.24 | treke | not that I can think of |
| 17:29.30 | treke | did a full build on saturday |
| 17:29.38 | treke | everything but X |
| 17:29.51 | pb_ | treke: hah |
| 17:30.20 | treke | pb_: I was supposed to call last night after hours and the admin was going to take a look, but I got stuck fixing phones :) |
| 17:30.43 | pb_ | treke: hate it when that happens. |
| 17:30.50 | mallum | treke: yeah me too ... but I updated the matchbox package to 0.7 and for some reason '/usr/lib/libfontconfig.so' was getting passed to the linker |
| 17:32.10 | mallum | treke: so I made mrproper libtool after trying various bits, and now libtool wont build with its confiugure trying to run arm-linux-gcc ( it should be the host gcc right ) |
| 17:32.22 | treke | not seen much in the way of introductory phone stuff :) |
| 17:37.27 | treke | and a three foot thick stack of manuals is a bit intimidating |
| 17:40.42 | pb_ | sounds nasty. what kind of phones? |
| 17:40.53 | treke | nortel meridian stuff |
| 17:42.46 | BBMatthew | tiome to go home and attempt this 'life' stuff people talk about |
| 17:42.58 | treke | life? |
| 17:43.01 | treke | never heard of it |
| 17:44.07 | BBMatthew | well I shall be carrying my laptop turned on... |
| 17:44.16 | BBMatthew | mobile build |
| 17:44.18 | BBMatthew | :) |
| 17:44.29 | BBMatthew | bye |
| 17:44.30 | *** part/#openembedded BBMatthew (~mas01r@arm.ecs.soton.ac.uk) |
| 18:19.44 | kergoth | gb2: k, cross-gcc-initial-3.3.1.oe builds properly. we're just lacking a cross-gcc do_stage target. once thats in place we can build binutils/gcc/glibc chains |
| 18:19.53 | kergoth | and you can make the uclibc changes where necessary |
| 18:29.04 | kergoth | hmm |
| 18:29.22 | kergoth | guess i should do a toolchain-glibc that includes the glibc and overrides its do_stage |
| 18:35.03 | kergoth | *cough*OE*cough* |
| 18:35.27 | prh | kergoth: does it have enough of what I need converted to it yet? |
| 18:35.28 | prh | make[1]: Entering directory `/home/paul/Z/d/buildroot-oz/packages/opie' |
| 18:35.28 | prh | #chmod 755 /home/paul/Z/d/buildroot-oz/build/opie/bin/osudo /home/paul/Z/d/buildroot-oz/build/opie/bin/osudo_helper |
| 18:35.31 | prh | cp -f /home/paul/Z/d/buildroot-oz/packages/opie/defconfig-zaurus-"v1_0" /home/paul/Z/d/buildroot-oz/build/opie/.config |
| 18:35.34 | prh | cp: cannot stat `/home/paul/Z/d/buildroot-oz/packages/opie/defconfig-zaurus-v1_0': No such file or directory |
| 18:35.34 | kergoth | depends on what you need |
| 18:35.37 | prh | make[1]: *** [/home/paul/Z/d/buildroot-oz/stamps/.opie.binary] Error 1 |
| 18:35.40 | prh | hehe true |
| 18:35.57 | prh | why is is looking for defconfig-zaurus-v1_0 <- which does not exist... |
| 18:36.17 | treke | there is you problem |
| 18:36.23 | prh | there is defconfig-zaurus-BRANCH_1_0 or defconfig-zaurus-HEAD |
| 18:36.31 | treke | there is no config for tag v1_0 |
| 18:37.08 | prh | so should I copy one of the others? |
| 18:37.32 | treke | use BRANCH_1_0 |
| 18:37.58 | prh | shoudld I copy it or tweak config.in ? |
| 18:38.13 | prh | <PROTECTED> |
| 18:38.17 | prh | bah missed |
| 18:47.40 | gb2 | kergoth: heh |
| 18:53.52 | gb2 | :P |
| 18:54.06 | gb2 | did you push your .oe's ? |
| 18:54.51 | kergoth | the cross-gcc ones? not yet |
| 18:54.53 | kergoth | brb, meeting |
| 18:55.19 | Pendalar | kick their asses kergoth, kick their asses |
| 18:55.30 | kergoth | er nm, i did push those |
| 18:55.32 | kergoth | k brb |
| 19:08.30 | Twiun | re |
| 19:15.54 | *** join/#openembedded bipolar (~bflong@dhcp-103-630-1.msns.sm.ptd.net) |
| 19:16.03 | bipolar | ,,0.o,,, |
| 19:16.11 | kergoth | ? |
| 19:16.12 | kergoth | heh |
| 19:16.20 | bipolar | just peeking in :) |
| 19:16.28 | bipolar | busy as hell |
| 19:18.52 | bipolar | god... I hate being the only IT guy |
| 19:19.39 | bipolar | It's like everyone thinks that when anything with a cpu or is written in binary breaks , it must be my fuckin' fault |
| 19:19.56 | kergoth | hmm, in order to properly implement oerootfs and oeimage, i need oemake to split its knowledge of available OEFILES from packages we want built. |
| 19:20.11 | *** join/#openembedded cmartin (~cmartin@p5087CCF4.dip0.t-ipconnect.de) |
| 19:30.16 | treke | herasy |
| 19:30.45 | Twiun | s/herasy/heresy |
| 19:31.28 | chouimat | treke: not really ... can you tell distcc the maximum number of jobs *AND* the maximum load ? |
| 19:31.55 | *** join/#openembedded BBMatthew (~mas01r@mattaw.plus.com) |
| 19:31.56 | treke | Twiun: bah. You don't know how to spell in proper American |
| 19:32.17 | Twiun | treke: never claimed to ;) |
| 19:32.32 | Twiun | treke: must be all this british influence around me |
| 19:32.58 | treke | yeah. we had that problem also a while back. |
| 19:33.11 | BBMatthew | z's everywhere? that really screwed me with ssh as its authorized_keys instead of authorised_keys |
| 19:35.51 | Twiun | or color instead of colour, etc |
| 19:38.11 | kergoth | gb2: heh, just ran into a case where I _want_ the included .oe to override the file P vars |
| 19:38.28 | kergoth | gb2: toolchain-glibc should share the glibc build area, just override stage and force a rerun of that task |
| 19:40.09 | *** join/#openembedded prpplague (~billybob@12.148.134.14) |
| 19:41.06 | chouimat | hi plague |
| 19:41.26 | cmartin | why do i get this? - kernel/qpngio.cpp:47:17: png.h: File not found - png.h is in staging/include and under 3rdparty in qte |
| 19:41.44 | treke | you dont have libpng3-dev installed on your workstation |
| 19:41.51 | treke | same goes for libjpg6-dev |
| 19:41.53 | kergoth | cmartin: its building an x86 binary |
| 19:41.56 | kergoth | yep, as treke says |
| 19:43.13 | *** join/#openembedded Bernardo (~jbbrs@213.58.80.173) |
| 19:43.25 | Bernardo | hi |
| 19:44.22 | kergoth | Bernardo: use the channel |
| 19:52.53 | *** join/#openembedded Virusmaster (~Virus@pD9E65165.dip.t-dialin.net) |
| 20:01.59 | kergoth | then i'll implement --force/-f and --short-circuit/-s for oebuild |
| 20:02.08 | kergoth | then oerootfs/oeinstall & oeimage |
| 20:04.21 | kergoth | hehe, should we make oeinstall be able to bypass our packaging? |
| 20:04.22 | kergoth | i.e. |
| 20:04.45 | kergoth | oeinstall -d /path/to/rootfs /path/to/blah.oe could have it set D to rootfs and force an execution of do_install to put the files there directly |
| 20:05.03 | kergoth | basically, oeinstall -d / /path/to/blah.oe would behave like portage, only without tracking/uninstallation |
| 20:05.42 | BBMatthew | kergoth: the buildroot adds path_to_arm_gcc to the enviroment a program is run in correct |
| 20:05.53 | BBMatthew | sorry built in |
| 20:06.01 | kergoth | BBMatthew: it add TOOLCHAINPATH/GCCVER to the path |
| 20:06.09 | kergoth | where toolchainpath and gcc version are set in build options |
| 20:06.17 | BBMatthew | fair enough |
| 20:08.42 | BBMatthew | is the strategy to set paths via buildroot or patch makefiles, as I have the issue where on a few pkgs the enviroment is wrong so it can't find a gcc |
| 20:09.01 | BBMatthew | I am using the gcc2 kernel gcc3 app setup |
| 20:09.49 | kergoth | are you unable to read? |
| 20:09.55 | kergoth | scroll up and read the fix we gave to cmartin |
| 20:14.49 | cmartin | BBMatthew: run 'bk get -p -r1.124 Vars.make >/tmp/Vars.make' and copy /tmp/Vars.make to BUILDROOT_DIR/Vars.make |
| 20:22.36 | *** join/#openembedded bisho (~bisho@136.Red-213-97-191.pooles.rima-tde.net) |
| 20:26.27 | kergoth | <PROTECTED> |
| 20:26.42 | kergoth | now you can just export OEFILES=/path/to/packages/content/*.oe |
| 20:26.45 | kergoth | then oemake oe |
| 20:26.47 | kergoth | or oemake glibc |
| 20:26.50 | kergoth | or oemake ncurses-5.3 |
| 20:27.11 | mickeyl | coolness |
| 20:27.25 | kergoth | much simpler |
| 20:27.48 | mickeyl | cmartin: got anywhere further with uclibc? |
| 20:28.09 | kergoth | man does it take a while to load up all the .oe's metadata |
| 20:28.17 | kergoth | we need to switch to an sqlite data cache soon. |
| 20:28.45 | prpplague | kergoth: sqlite rocks |
| 20:28.52 | kergoth | yep |
| 20:34.02 | cmartin | mickeyl: further yes - but now the next error (and your comments in STATUS aren't exact): |
| 20:34.13 | cmartin | checking for pthread_create in -lpthread... no |
| 20:34.13 | cmartin | configure: error: pthread needed |
| 20:34.21 | cmartin | building libxine |
| 20:34.21 | kergoth | mickeyl: hows the vt handling stuff going? |
| 20:34.38 | kergoth | hehe, libxine's autotools use is so extensive its easy to make it exlpode |
| 20:34.43 | kergoth | i should add it to OE and fix its issues |
| 20:34.54 | mickeyl | kergoth: finished. apart from the keyboard handling |
| 20:35.07 | mickeyl | kergoth: i added a vtapplet to Opie HEAD to switch from Opie to console |
| 20:35.23 | mickeyl | kergoth: console to Opie is done by Alt-arrow. |
| 20:35.32 | mickeyl | kergoth: on c7x0, that is. on collie it is pending |
| 20:35.55 | mickeyl | cmartin: easy fix. go into configure.ac and make that AC_ERROR a AC_WARN |
| 20:36.24 | cmartin | isn't it critical? |
| 20:36.34 | mickeyl | interestingly not :) |
| 20:36.41 | mickeyl | it is only @ configuring |
| 20:37.11 | kergoth | heh, cmartin, ar eyou building with gcc3? |
| 20:37.20 | kergoth | and if so, ar eyou using my most recent 3.3.1 final chain or the pre? |
| 20:38.19 | aman | hello |
| 20:38.28 | mickeyl | kergoth: he is using my 3.3.1-final-uclibc |
| 20:39.00 | aman | who woulda known, you need to give the kernel "noinitrd load_ramdisk=1 prompt_ramdisk=1 root=/dev/fd0 ramdisk_blocksize=4096" to get it to use a cramfs rootfs floppy |
| 20:39.05 | cmartin | kergoth: uclibc 3.3.1 |
| 20:40.09 | kergoth | ah, k |
| 20:40.12 | cmartin | mickeyl: in STATUS you speak of _MSC_DEV, it's not there - you probably meant _MSC_VER |
| 20:40.16 | kergoth | was gonna say.. hte glibc gcc3.x chains we used to have |
| 20:40.28 | kergoth | the libpthread.so is a linker script that references /lib/libpthread* directly |
| 20:40.33 | kergoth | breaks the pthread tests |
| 20:40.38 | kergoth | heh |
| 20:40.45 | mickeyl | kergoth: maybe that is the same problem here |
| 20:40.50 | mickeyl | cmartin: right. please correct |
| 20:40.50 | kergoth | worth checking into anyway |
| 20:40.59 | kergoth | its due to the fact that pthread needs the _nonshared.a and the .so both |
| 20:41.06 | kergoth | same deal as /usr/lib/libc.so on most machines |
| 20:41.10 | cmartin | mickeyl: and you mention a video_linux_fb and probably mean video_out_fb.c |
| 20:41.25 | mickeyl | cmartin: hehe, yeah. i wrote that from memory, not directly after fixing :D |
| 20:41.41 | mickeyl | kergoth: i see. that thing probably hasn't been thought of by andersee yet |
| 20:42.26 | mickeyl | however, since 80% of buildroot compiles against uclibc i feel we are not that much away from migrating |
| 20:42.42 | mickeyl | can't wait to dance on glibc's grave... how you would say ;) |
| 20:42.47 | aman | heh |
| 20:43.07 | aman | uclibc breaks BC too right? |
| 20:43.11 | mickeyl | yeah |
| 20:43.12 | mickeyl | finally |
| 20:43.16 | mickeyl | we will break it |
| 20:43.26 | mickeyl | thus taking the tance to go to a recent compiler |
| 20:43.30 | mickeyl | and maybe to a recent Qt/E :D |
| 20:43.34 | aman | yea |
| 20:43.35 | mickeyl | s/tance/chance |
| 20:43.46 | aman | i'm all for a gcc3+qt3+uclibc bc breaking party |
| 20:44.06 | mickeyl | yeah, when we break it then lets break it (hopefully) just once |
| 20:44.12 | kergoth | heh |
| 20:44.20 | Soopaman | are there any 17" wuxga laptops out yet? |
| 20:44.20 | kergoth | assuming uclibc doesnt break bc with itself in future releases |
| 20:44.27 | mickeyl | hah. right |
| 20:44.28 | aman | heh |
| 20:44.34 | kergoth | mickeyl: which it has before :) |
| 20:44.44 | aman | Soopaman whats wuxga? |
| 20:44.49 | kergoth | and assuming we dont decide to change the options we use to build it, since that could break it as well |
| 20:44.57 | kergoth | so better get our uclibc config right the first time |
| 20:45.03 | mickeyl | kergoth: yeah, right. however andersee has promised to slow down :) |
| 20:45.07 | Soopaman | aman, wide-screen uxga |
| 20:45.09 | Soopaman | 16:9 |
| 20:45.11 | Soopaman | res |
| 20:45.46 | kergoth | you know, i really dont like buildroot anymore |
| 20:45.48 | kergoth | i'm spoiled by oe |
| 20:45.50 | aman | cool |
| 20:45.50 | aman | whats uxga? |
| 20:46.26 | aman | lol |
| 20:46.31 | mickeyl | kergoth: really? well... another situation here. I think i really understand the old one now (finally... took a while!) which makes it hard start with a new one... |
| 20:46.41 | kergoth | hehe |
| 20:46.48 | aman | heh |
| 20:46.56 | aman | anyone have a copy of the one that came before that one? |
| 20:47.01 | kergoth | since i built most of buildroot there was no learning curve, so i dont mind |
| 20:47.16 | kergoth | aman: nah, unless sf's cvs repos for buildroot is still there |
| 20:47.25 | kergoth | if it is, the old stuff might be there |
| 20:47.36 | aman | that was a longass time ago |
| 20:47.39 | aman | i remember using that |
| 20:47.51 | BBMatthew | aman: the dell literature for this laptop says 1600x1200 UXGA+ screen |
| 20:47.52 | aman | yea |
| 20:47.54 | aman | crazy stuff |
| 20:47.56 | kergoth | and remember our dependencies? |
| 20:47.57 | kergoth | ahah |
| 20:47.58 | kergoth | funny shit |
| 20:48.01 | kergoth | they'd just keep calling one another |
| 20:48.03 | kergoth | over, and over |
| 20:48.12 | kergoth | until we added a cache of what we were done with |
| 20:48.18 | aman | haha |
| 20:48.20 | aman | i remember that |
| 20:48.31 | kergoth | entering directory .. leaving directory... like 800 lines of that |
| 20:48.33 | kergoth | hehe |
| 20:49.20 | aman | mickeyl build another rc release so i can use it on my 5500 =P |
| 20:49.22 | kergoth | man, oe is alot easier to use with that new syntax |
| 20:49.34 | aman | mickeyl oh, btw. whats up with orinococs/spectrum. did we get it figured out? |
| 20:49.36 | kergoth | oemake apache; |
| 20:49.38 | kergoth | heh |
| 20:49.40 | mickeyl | aman: yeah |
| 20:49.53 | mickeyl | aman: we have to downgrade to the pcmcia-modules from oz 3.2 |
| 20:49.56 | mickeyl | aman: then it works |
| 20:49.59 | aman | ok cool. the image noda built last week didn't have it so i was wondering |
| 20:50.11 | mickeyl | aman: you have a fairly recent image? |
| 20:50.13 | aman | he prolly just left it out |
| 20:50.23 | aman | mickeyl yea, noda built one last week. its in unofficial/unstable/noda/ |
| 20:50.38 | mickeyl | aman: k. i also built one which has all the necessary downgrades |
| 20:50.48 | aman | ok cool |
| 20:50.49 | aman | i want that |
| 20:50.54 | *** part/#openembedded chouimat (~root@r2351064.cidc.net) |
| 20:50.56 | aman | put that on sf as the next rc |
| 20:50.58 | kergoth | heh, theres no way we're gonna end up retaining a sane upgrade path when we move to OE |
| 20:51.05 | mickeyl | for sure |
| 20:51.14 | kergoth | i havent even added the packaging shit to these at all yet |
| 20:51.19 | aman | did you build mplayer by any chance? i have this urge to watch movies on my Z |
| 20:51.24 | kergoth | much less with correct versioning and conflicts and replaces and crap |
| 20:51.39 | mickeyl | aman: woudln't want to compile for me last time i tried |
| 20:51.45 | aman | damn |
| 20:51.48 | aman | it keeps breaking |
| 20:51.49 | *** join/#openembedded chouimat (~root@r2351064.cidc.net) |
| 20:52.03 | aman | prolly cuz cvs head is now their new mplayer g2 or whatever |
| 20:52.08 | Bernardo | anyone's using kismet successfully? |
| 20:52.27 | aman | ibot botmail for pnis whats up with the mplayer in buildroot. do we wanna use cvs head or the 0_90 or whatever branch? could you fix that? |
| 20:52.44 | kergoth | I dont want anyone using cvs head directly anymore. we can do convenience classes that switch something to head, confirm the build sanity, then on successful completion, updates the cvsdate |
| 20:52.48 | kergoth | (for OE that is) |
| 20:52.54 | mickeyl | aman: http://opie.net.wox.org/openzaurus/oz-snapshot-030909-1225/ is the best one since OZ 3.2. |
| 20:52.56 | aman | k |
| 20:53.24 | Bernardo | mickeyl: that's the one missing tinykate? |
| 20:53.25 | mickeyl | kergoth: yes. things like that ensure a bit more upstream reliabilty |
| 20:53.29 | kergoth | mickeyl: yep |
| 20:53.29 | aman | cool. i didn't know they had stuff on wox.org |
| 20:53.52 | aman | i need to setup a dev environment again and get in on some of this OE action |
| 20:53.52 | kergoth | heh |
| 20:53.57 | mickeyl | Bernardo: no idea - it is a fairly recent BRANCH build and not all apps are converted to quicklaunch yet |
| 20:53.58 | kergoth | yeah you do |
| 20:54.01 | kergoth | we need a transition guide |
| 20:54.04 | kergoth | intro doc |
| 20:54.08 | kergoth | and 247 packages converted |
| 20:54.12 | mickeyl | transition guide would be really nice |
| 20:54.17 | mickeyl | would help the convertion process ;) |
| 20:54.21 | kergoth | yeah, the guide could outline some basic things |
| 20:54.27 | kergoth | like Name is no longer necessary due to filename |
| 20:54.30 | kergoth | Version either |
| 20:54.35 | kergoth | Sources -> SRC_URI |
| 20:54.41 | kergoth | source format differences |
| 20:54.50 | kergoth | $(stamp-binary) -> do_compile () { } |
| 20:54.56 | kergoth | etc |
| 20:54.57 | cmartin | mickeyl: no string.h in libpcap - which header file did you mean? |
| 20:55.04 | kergoth | and recommend things |
| 20:55.11 | kergoth | like if your app uses autotools, dont define the functions |
| 20:55.14 | kergoth | just inherit autotools |
| 20:55.32 | mickeyl | cmartin: string.h in the toolchain |
| 20:55.40 | aman | mickeyl oh, this is awesome.. you should link to this on oz news or something |
| 20:55.40 | kergoth | and if your app is in portage for gentoo, you can swipe the ebuild with slight alteration |
| 20:55.43 | kergoth | do the cross changes |
| 20:55.47 | kergoth | and be done. |
| 20:56.00 | mickeyl | kergoth: and while you're there... guidelines! e.g. what to prefer: "libreadline-4.3.oe", "libreadline4-4.3.oe", "readline-4.3.oe" or even "readline4-4.3.oe" |
| 20:56.17 | kergoth | libreadline is bad, since it attempts to enforce distribution policy |
| 20:56.24 | kergoth | we name based on upstream source naming |
| 20:56.33 | kergoth | package naming for output is distro specific, build naming is not |
| 20:56.34 | cmartin | aahh |
| 20:56.41 | *** join/#openembedded SuKoShi (~sukoshi@dyn-81-166-18-32.ppp.tiscali.fr) |
| 20:56.51 | mickeyl | aman: yeah i had quite a lot of success stories with that image. however at the moment, I'd like to whine a bit e.g. see the opie.net.wox.org top page :) |
| 20:57.21 | Bernardo | mickeyl: I think I am using that one. only blows when I try to install opie-dvorak |
| 20:57.38 | BBMatthew | lying on the couch glass in hand music playing buildroot compiling before my eyes. heaven.. |
| 20:57.49 | mickeyl | kergoth: right. so readline is to prefer. but what about that 4? readline4-4.3 or readline-4.3 ? |
| 20:57.50 | BBMatthew | ;) |
| 20:58.00 | mickeyl | BBMatthew: lol |
| 20:58.01 | kergoth | libreadline4 is debianism |
| 20:58.03 | kergoth | libreadline.so.4 |
| 20:58.17 | kergoth | upstream is readline-4.3, so are we at build level |
| 20:58.24 | mickeyl | ok, so we won't append that thing. good. |
| 20:58.29 | kergoth | yep |
| 20:58.37 | kergoth | i want th is to stay distro agnostic |
| 20:58.52 | kergoth | PACKAGES_debianish = libreadline4 |
| 20:58.58 | kergoth | PACKAGES_redhatish = readline |
| 20:59.01 | kergoth | hehe |
| 20:59.20 | aman | lol |
| 21:00.25 | kergoth | so with that, OE could build redhat or debian. with oeinstall, OE could build/act like gentoo/portage |
| 21:00.37 | kergoth | hehe |
| 21:00.41 | cmartin | mickeyl: is it correct that the __USE_BSD ifdef is in string.h 3 times? |
| 21:00.58 | *** join/#openembedded prpplague (~billybob@12.148.134.14) |
| 21:01.07 | mickeyl | cmartin: maybe, but IIRC the only offending place is at the bottom |
| 21:01.21 | mickeyl | cmartin: where the two additional functions are defined. just erase them |
| 21:02.11 | kergoth | mickeyl: the thing with OE is the amount of flexibility it give syou |
| 21:02.18 | kergoth | mickeyl: i cant implement a Policy Manual for OE as a whole |
| 21:02.26 | killefiz | mickeyl: do you mind if i post a news item to zaurus.help4free.de pointing to opie.net.wox.org? |
| 21:02.29 | kergoth | mickeyl: I can, and will, implement such a manual for the primary OE/OZ metadata repository |
| 21:02.49 | mickeyl | killefiz: yes, that would be important. i'm really serious |
| 21:03.15 | mickeyl | kergoth: what I like the most is the inheritance stuff it gives us |
| 21:03.29 | mickeyl | killefiz: right |
| 21:03.29 | aman | hehe |
| 21:03.32 | kergoth | mickeyl: me too, it makes it not only easy, but recommended, to abstract out commonalities |
| 21:03.33 | mickeyl | killefiz: it's late :) |
| 21:03.42 | kergoth | mickeyl: consolodates |
| 21:03.46 | mickeyl | yup |
| 21:03.49 | mickeyl | repeat is evil |
| 21:03.55 | kergoth | autotools.oeclass is the perfect case of this |
| 21:04.02 | kergoth | i mean how many of our packages use autoconf? |
| 21:04.07 | kergoth | 80%? |
| 21:04.08 | mickeyl | 75% |
| 21:04.11 | kergoth | :) |
| 21:04.11 | mickeyl | ya |
| 21:04.27 | aman | mickeyl are the downgrades to pcmcia/orinoco-whatever pushed to buildroot? |
| 21:04.45 | mickeyl | aman: no, because they are only as long as we have to support collie-2.4.6 |
| 21:05.02 | aman | heh. you say that like its only gonna be a couple days |
| 21:05.12 | mickeyl | hehe |
| 21:05.21 | mickeyl | i like to think about it that way :) |
| 21:05.33 | kergoth | I wont be back on the kernel stuff until the transition is either complete or out of my hands. |
| 21:05.40 | kergoth | so the more people converting packages, the more likely it is. |
| 21:06.33 | cmartin | mickeyl: BTW: libpcap errors on string.h in staging/target..... and not in the toolchain - or is this file just from the toolchain? |
| 21:06.51 | mickeyl | libpcap includes string.h |
| 21:06.53 | mickeyl | and it is... |
| 21:06.55 | mickeyl | in the toolchain |
| 21:07.02 | mickeyl | or in STAGING |
| 21:07.11 | mickeyl | hard to remember... |
| 21:07.22 | aman | masticator you have a new box yet i can do oe devel on? |
| 21:07.31 | cmartin | i just edited STAGING and i'm now building opie :) |
| 21:07.38 | mickeyl | good |
| 21:07.47 | mickeyl | opie should ran through |
| 21:07.49 | kergoth | aman: hehe. miss the beast? i certainly do |
| 21:07.58 | mickeyl | the only thing was in rsync and i fixed that |
| 21:08.01 | mickeyl | glibc-specialty |
| 21:08.01 | aman | kergoth oh yea. wasn't he supposed to get it back? |
| 21:08.18 | kergoth | aman: was gonna get a new one. quad proc or something |
| 21:08.22 | kergoth | aman: *shrugs* |
| 21:08.41 | kergoth | hmm. |
| 21:08.56 | kergoth | oemake's handling of the command to use for its deps isnt proper |
| 21:09.05 | kergoth | oemake -c clean ncurses doesnt work as a result |
| 21:09.06 | kergoth | hm |
| 21:09.15 | kergoth | would have to oemake -c clean base/ncurses-5.3-r0 |
| 21:10.30 | cmartin | gotta go - i'll let my uclibc build continue while i sleep ;-) |
| 21:10.56 | mickeyl | n8 |
| 21:11.20 | cmartin | bye |
| 21:11.26 | *** part/#openembedded cmartin (~cmartin@p5087CCF4.dip0.t-ipconnect.de) |
| 21:11.54 | kergoth | actually started work at a reasonable time for once |
| 21:12.00 | gb2 | heh |
| 21:12.01 | gb2 | same |
| 21:12.01 | cyn_ | wow |
| 21:12.08 | gb2 | i'm home already ;) |
| 21:12.15 | mickeyl | i love starting early in the morning.... theoretically |
| 21:12.20 | kergoth | mickeyl: hehe. exactly. |
| 21:12.30 | mickeyl | but hell i can't get up earlier than 9 usually ;) |
| 21:12.33 | mickeyl | flesh is weak |
| 21:12.54 | kergoth | actually, i think i'll stay and finish what i was working on first |
| 21:13.22 | mickeyl | do they control how long you work? like using a clock to stick your card or whatever in? |
| 21:14.16 | *** part/#openembedded Pendalar (Pendalar@68.114.156.106) |
| 21:14.16 | kergoth | eh, somebody remind me |
| 21:14.26 | kergoth | how do i do a static var in python? |
| 21:14.33 | kergoth | mickeyl: nope |
| 21:14.40 | mickeyl | kergoth: good i hate things like that |
| 21:14.42 | mickeyl | kergoth: class-static ? |
| 21:15.07 | kergoth | nah, just a static var inside a function |
| 21:15.11 | mickeyl | ah |
| 21:15.15 | kergoth | can probably do without, but would make it easier |
| 21:15.16 | mickeyl | def func(): |
| 21:15.26 | mickeyl | func.variable = 10 |
| 21:15.29 | kergoth | ah |
| 21:15.37 | kergoth | shouldve known. cool |
| 21:16.16 | mickeyl | that's a pretty new feature |
| 21:16.21 | mickeyl | since 2.1 or 2.2 at earliest |
| 21:16.46 | killefiz | do the buildroot snapshots work again? |
| 21:20.02 | BBMatthew | a quick glass of glayva methinks |
| 21:21.47 | kergoth | mickeyl: ah. well, we're already dependent on really new stuff |
| 21:21.56 | aman | i need a devel machine |
| 21:21.58 | kergoth | i wonder if the implementation still uses the static class member functions |
| 21:22.05 | kergoth | like it did a while back |
| 21:22.48 | kergoth | ugh |
| 21:22.56 | kergoth | this digraph class we're using from portage is sorely deficient |
| 21:23.03 | kergoth | only one type of relationship between items |
| 21:23.24 | kergoth | and no methods to obtain the path between two of them, or to check the relative distance through the graph between two items, or .. |
| 21:26.52 | *** join/#openembedded chouimat (~dieu@r2351064.cidc.net) |
| 21:26.53 | *** mode/#openembedded [+o chouimat] by ChanServ |
| 21:26.57 | *** join/#openembedded CAMCAM (~camrein@AMontsouris-108-1-31-42.w81-53.abo.wanadoo.fr) |
| 21:38.43 | aman | anyone have a fast box i can devel on? |
| 21:39.01 | kergoth | not i |
| 21:39.44 | pb_ | oh, didn't you have some kind of multiple opteron machine? |
| 21:39.50 | *** join/#openembedded W8TVI (~root@1114105291.mi.dial.hexcom.net) |
| 21:39.56 | pb_ | heh |
| 21:39.57 | kergoth | ~lart mickeyl |
| 21:40.09 | mickeyl | ouch |
| 21:40.15 | kergoth | pb_: yeah but it wasnt a permanent thing, knew someone who had one temporarily |
| 21:40.25 | kergoth | was fun while it lasted *sniff* |
| 21:40.27 | pb_ | kergoth: oh, right. darn. |
| 21:40.58 | chouimat | win 5 |
| 21:46.58 | BBMatthew | hah opie just built so it was the buildroot that had gone south |
| 21:47.38 | pb_ | hm, I wonder whether building a separate libc6 package for pxa is worthwhile. |
| 21:48.00 | kergoth | with what changes, just tuned for xscale? |
| 21:48.16 | pb_ | well, we use -mtune=xscale anyway; this would be using armv5 instructions. |
| 21:48.30 | aman | life would be so much easier if i didn't keep losing my CF card |
| 21:48.56 | pb_ | possibly more aggressive xscale tuning, though we already tune for xscale to the almost complete exclusion of other cpus. |
| 21:49.16 | ljp | ahhh, espresso... nectar of the gods |
| 21:49.21 | kergoth | ljp: true that. |
| 21:49.25 | pb_ | morning ljp |
| 21:49.32 | ljp | hello |
| 21:51.05 | kergoth | hmm, the cross-binutils package just tarred up /usr/local/arm/3.3 of its own accord |
| 21:51.26 | pb_ | crikey. hope it isn't becoming sentient. |
| 21:51.37 | kergoth | ITS ALIVE! |
| 21:52.37 | cyn_ | mmm c700 |
| 21:53.17 | ljp | I guess Oslo is using c700's to show off Qtopia |
| 21:54.00 | aman | kergoth you know.. if you finish 2.4.21 people will be more motivated to convert all the packages from old buildroot to OE |
| 21:54.12 | treke|home | <PROTECTED> |
| 21:54.23 | pb_ | aman: if _you_ finish 2.4.21, it would probably be just as effective |
| 21:54.30 | treke|home | I think most people who have touched oe so far are 2.4.18 users :p |
| 21:55.13 | kergoth | oe wont affect users |
| 21:55.17 | kergoth | so its irrelevent |
| 21:55.21 | kergoth | its primarily a developer concern |
| 21:55.28 | kergoth | and that piece affects everyone regardless |
| 21:56.00 | ljp | treke: kergoth said you knew how the c7xxx does the auto rotate when the physical screen gets rotated |
| 21:56.05 | aman | damn. it was worth a shot anyway |
| 21:56.37 | treke|home | ljp: key 102 sis sent when the screen is rotated |
| 21:56.56 | treke|home | ljp: there is an apm_bios ioctl you can use when you want to check the hinge at arbitrary times |
| 21:57.13 | kergoth | of 60 packages, 39 are outputting tarballs |
| 21:57.22 | pb_ | is that good? |
| 21:57.40 | treke|home | pb_: Nah we want to get that number down to 0 |
| 21:57.52 | kergoth | some of those are native, or crosscompiler builds, so those can be deducted |
| 21:57.58 | kergoth | so yeah, not too bad |
| 21:58.02 | kergoth | only a couple failed compiles |
| 21:58.13 | mewyn | kergoth: prolly gonna need to reload sshd again. another hole |
| 21:58.15 | ljp | cool, thnx |
| 21:58.28 | aman | oe can build ipks now? |
| 21:58.29 | kergoth | mewyn: another? is it fixed in 3.7p1? |
| 21:58.35 | kergoth | aman: it can buildi tarballs, rpms, or ipks |
| 21:58.38 | kergoth | s/buildi/build/ |
| 21:58.48 | kergoth | the rpm build is pretty spiffy for building an rpm of OE itself |
| 21:58.53 | mewyn | nope, fixed in 3.7.1 |
| 21:59.00 | aman | hmm |
| 21:59.03 | kergoth | mewyn: k |
| 21:59.18 | aman | if school lets out tommorow cuz of the hurricane, i'll go on a crazy converting spree |
| 21:59.51 | kergoth | those fuckers |
| 21:59.56 | kergoth | 3.7.1 breaks my cross patch again |
| 22:00.02 | treke|home | kergoth: ha |
| 22:00.56 | aman | kergoth lol. the one you made yesterday? |
| 22:01.08 | kergoth | aman: yes! |
| 22:01.15 | treke|home | ljp: You going to add support in qtopia? |
| 22:01.46 | treke|home | ljp: http://ent.brooks.edu/~ggilbert/chkhinge.c |
| 22:02.42 | kergoth | son of a bitch |
| 22:02.45 | kergoth | hmm |
| 22:03.05 | kergoth | dont ya hate it when you accidnetally install arm binaries into x86 system dirs? |
| 22:03.09 | mickeyl | kergoth: no chance to get anything of that upstream? |
| 22:03.11 | kergoth | god, it feels like a monday |
| 22:03.50 | kergoth | mickeyl: i did this one properly, cleanly. so yes, theres a good chance. just havent had time to submit it |
| 22:06.45 | ljp | in the very least, support for it in this buildroot qtopia. Not sure if they would want a device specific thing as this in the qtopia tree |
| 22:07.19 | treke|home | you guys need something like odevice |
| 22:07.24 | ljp | but seeing as patching in buildroot is easy, and I have a few patches already.. |
| 22:07.51 | ljp | ya, maybe I can suggest that for Qtopia 2.0 |
| 22:08.01 | aman | ljp you moved to AUS yet? |
| 22:08.14 | ljp | yes, I am here currently |
| 22:08.20 | aman | ah awesome |
| 22:08.21 | aman | congrats |
| 22:08.25 | ljp | :) |
| 22:08.33 | ljp | but still waiting my extended visa ok |
| 22:09.01 | kergoth`bbl | pushed the oemake syntax changes |
| 22:09.09 | kergoth`bbl | k im out |
| 22:09.51 | aman | alright, so.. the faster i covert packages to oe, the faster kergoth starts working on kernel, the faster we have a 2.4.21 |
| 22:10.01 | mickeyl | right analysis |
| 22:10.21 | aman | i better get working then |
| 22:13.48 | *** join/#openembedded drw (~drw@12-241-65-192.client.attbi.com) |
| 22:18.40 | pb_ | wow, gcc 3.3 is putting all manner of crap in the GOT. |
| 22:18.46 | pb_ | I wonder how long that's been going on. |
| 22:19.58 | treke|home | GOT? |
| 22:20.07 | pb_ | global offset table. |
| 22:20.14 | treke|home | GOT it |
| 22:20.48 | pb_ | it's meant to contain runtime pointers to variables whose addresses aren't known at static link time, but gcc is blithely putting all its little bits of static internal data in there as well. |
| 22:21.06 | aman | ick |
| 22:21.11 | pb_ | which means that ld.so needs to do about ten times the amount of work it ought to during program startup. |
| 22:21.17 | aman | ouch |
| 22:21.46 | pb_ | (and the resulting code is slower as well, because every variable has to be doubly indirected) |
| 22:22.06 | pb_ | so, all in all, suck. |
| 22:22.20 | treke|home | cool |
| 22:22.32 | pb_ | heh, that's one way to look at it. |
| 22:23.10 | mickeyl | "nice" - so we lose the optimization gained by 3.x through to that stuff? |
| 22:23.34 | pb_ | some of it, yeah. |
| 22:23.55 | pb_ | try compiling int f() { g("hello"); } with -fPIC and look at the output. |
| 22:24.18 | mickeyl | there must be a reason for it |
| 22:24.21 | mickeyl | the gcc team is not dumb |
| 22:24.26 | treke|home | pb_: No one but you is sick enough to do that |
| 22:24.38 | *** join/#openembedded TheMasterMind1 (~aman@h-69-3-87-233.MCLNVA23.covad.net) |
| 22:24.55 | pb_ | mickeyl: yeah. I'm sure this used to work better, though. |
| 22:25.14 | pb_ | treke|home: hrm |
| 22:25.41 | mickeyl | pb_: btw., did get anyway in doing gpe for buildroot? have some patches handy? |
| 22:26.00 | pb_ | mickeyl: I was scuppered by xfree86 lossage. |
| 22:26.07 | mickeyl | ah |
| 22:26.17 | treke|home | pb_: mallum has it building again, I think |
| 22:26.19 | pb_ | to be more specific, I didn't have the strength to try to fix its staginginstall rule. |
| 22:26.23 | pb_ | treke|home: aha |
| 22:26.25 | treke|home | pb_: He just posted c760 ipks |
| 22:26.32 | pb_ | oh right, cool |
| 22:26.36 | pb_ | maybe I should try again. |
| 22:26.45 | pb_ | is the magic snapshot-updating thing working yet? |
| 22:26.48 | mickeyl | mallum said to me there is just one link or file necessary to get it building |
| 22:26.52 | mickeyl | pb_: seems not |
| 22:27.01 | treke|home | pb_: nope. I'll do a new snapshot for you |
| 22:27.03 | mickeyl | pb_: the snapdate is current, but the files in there are darn old |
| 22:27.36 | treke|home | pb_: actually you probably should hold off |
| 22:27.53 | treke|home | pb_: wait until oe builds X instead of wasting time on buildroot |
| 22:27.56 | pb_ | treke|home: ok |
| 22:28.04 | mickeyl | hmmm |
| 22:28.17 | treke|home | of course you are free to try puting the gtk stuff in buildroot :p |
| 22:28.28 | treke|home | err oe |
| 22:28.29 | mickeyl | you all seem to have the feeling that OE will be useful for doing releases in just a couple of days... |
| 22:29.06 | treke|home | mickeyl: I do have the fealing that gpe is going to take a bit of work, so he might as well not do it twice |
| 22:29.27 | mickeyl | however the real work is in the patches |
| 22:29.34 | mickeyl | and these can be used |
| 22:29.59 | pb_ | aha, gcc seems to be fixed. |
| 22:30.05 | mickeyl | ? |
| 22:30.15 | pb_ | needed SYMBOL_REF_LOCAL_P() adding in two places. |
| 22:30.28 | mickeyl | pb_: in which version? |
| 22:30.36 | pb_ | trunk, but I think the same problem is in 3.2 and 3.3. |
| 22:30.42 | mickeyl | ah |
| 22:30.43 | pb_ | probably all the way back to 3.0 in fact |
| 22:32.30 | BBMatthew | nuts the opie recorder is broken under gcc3.3.1 |
| 22:32.41 | mickeyl | BBMatthew: ==> ljp |
| 22:33.19 | BBMatthew | did he do it? :) |
| 22:33.23 | mickeyl | yea |
| 22:34.03 | BBMatthew | looks like similar probs that I had when changing to gcc3 with my c++ apps and template instatiation |
| 22:38.30 | mickeyl | bed time |
| 22:38.32 | mickeyl | g'night all |
| 22:46.59 | BBMatthew | interesting opie doesnt use buildroot vars folr optimizations |
| 22:49.38 | ljp | what version. whats the error? |
| 22:50.10 | BBMatthew | its the latest buildroot pull off cvs |
| 22:50.12 | ljp | of opierec, that is |
| 22:50.25 | BBMatthew | lots of linking errors in opierec |
| 22:50.35 | ljp | pthread |
| 22:50.36 | ljp | ? |
| 22:50.48 | BBMatthew | the errors are coming up now, as I have modded the makefile to compile for xscale |
| 22:51.33 | ljp | I have been working on opierec the past few days, so I can check on gcc3 |
| 22:51.51 | ljp | its now got a waveform display |
| 22:52.01 | BBMatthew | cool |
| 22:52.24 | BBMatthew | well buildroot is running opie now so I shall report as soon as I know whats going on |
| 22:54.50 | *** join/#openembedded Supacat (aaloo@user-2746.wfd13.dsl.pol.co.uk) |
| 22:55.10 | Supacat | hi folks - if i have an old pc i want to convert to an mp3 player.. shold i be looking at openembedded |
| 22:55.13 | BBMatthew | the tag it checks out (I think) is BRANCH_1_0 and the errors are .obj/sharp-linux/main.o(.text+0x600): In function `__static_initialization_and_destruction_0(int, int)': |
| 22:55.13 | BBMatthew | : undefined reference to `std::ios_base::Init::~Init [in-charge]()' |
| 22:55.13 | BBMatthew | .obj/sharp-linux/main.o(.text+0x604): In function `__static_initialization_and_destruction_0(int, int)': |
| 22:55.35 | BBMatthew | Supacat: no, not in any way really. unless you are good |
| 22:56.11 | treke | Supacat: just install a standard linux distibution |
| 22:56.47 | BBMatthew | version 20030914 |
| 22:57.04 | ljp | hmm |
| 22:57.25 | BBMatthew | I was going to tweak the build so it didn't make it :) |
| 22:57.35 | ljp | odd that it needs to link to libstdc++ |
| 22:57.48 | Supacat | thanks |
| 22:57.51 | *** part/#openembedded Supacat (aaloo@user-2746.wfd13.dsl.pol.co.uk) |
| 22:58.35 | ljp | wonder where that is coming from |
| 22:59.06 | ljp | cause it links fine with gcc2 |
| 22:59.41 | pb_ | welcome to c++ |
| 22:59.57 | BBMatthew | hehe |
| 23:00.01 | treke | yippe c :p |
| 23:00.19 | BBMatthew | I hate c++ portability its not even possible across gcc vers, let alone to M$ tools |
| 23:00.37 | BBMatthew | like the declarations of typename in template declerations |
| 23:00.42 | TheMasterMind1 | haha sweet |
| 23:00.44 | TheMasterMind1 | schools are closed |
| 23:00.53 | BBMatthew | y? |
| 23:00.56 | ljp | opierec doesnt use any stdc++ |
| 23:01.05 | treke | TheMasterMind1: that would be cool if your house wasnt going to be flattened |
| 23:02.04 | pb_ | treke: the glass is always half empty for you. |
| 23:02.14 | kergoth`bbl | mickeyl|away: um, i'll be producing minimal collie images from OE within days. |
| 23:02.27 | kergoth`bbl | mickeyl|away: qte is in, getting opie building wont take much longer. |
| 23:02.32 | treke | pb_: Yeah, the waiter here sucks |
| 23:03.18 | ljp | treke: it sends key 102 both times the screen is rotated? |
| 23:03.34 | treke | ljp: looks like it |
| 23:04.04 | BBMatthew | ljp: looks like it must be coming in from a header |
| 23:04.15 | ljp | hmm |
| 23:04.26 | gb2 | wish FOO_append added a space for you |
| 23:04.49 | kergoth`bbl | gb2: yeah, it gets annoying. but i didnt want to limit it, in case you dont want the space |
| 23:05.07 | gb2 | why not just have a += ? |
| 23:05.11 | ljp | ibot: FOO_append is <reply> space for you |
| 23:05.12 | | okay, ljp |
| 23:05.34 | treke | ibot ljp is reply ibot forget ljp |
| 23:05.35 | | ...but ljp is already something else... |
| 23:05.43 | treke | ibot ljp |
| 23:05.43 | | i heard ljp is currently an earthling from the planet Mergatroid or http://llornkcor.com |
| 23:05.50 | kergoth`bbl | gb2: i'd expect += to behave the same as append.. |
| 23:06.01 | gb2 | kergoth: yeah. hrm |
| 23:06.11 | gb2 | ++= ;) |
| 23:06.22 | treke | +nospace= |
| 23:06.39 | *** join/#openembedded W8TVI|z (~zic@1114105291.mi.dial.hexcom.net) |
| 23:06.40 | gb2 | +dwim= |
| 23:06.42 | gb2 | ;) |
| 23:06.59 | kergoth`bbl | or, could not strip initial and ending spaces |
| 23:06.59 | kergoth`bbl | so that |
| 23:06.59 | pb_ | kergoth`bbl: do you have any mechanism for running the gcc testsuite in a cross build? |
| 23:07.04 | kergoth`bbl | THIS += that |
| 23:07.08 | kergoth`bbl | keeps the space |
| 23:07.24 | treke | kergoth`bbl: that doesnt seem intuitive |
| 23:07.46 | kergoth`bbl | treke: agreed, i like spaces just to make it seem cleaner. |
| 23:07.47 | treke | THIS += foo and THIS+=foo should be the same imo |
| 23:07.51 | kergoth`bbl | yeah |
| 23:08.06 | gb2 | &= perhaps |
| 23:08.37 | treke | how about THIS += " FOO " |
| 23:08.38 | treke | :p |
| 23:08.46 | gb2 | looks uglier |
| 23:08.52 | gb2 | foo = bar |
| 23:08.56 | gb2 | foo += " baz" |
| 23:09.12 | gb2 | foo &= baz |
| 23:09.12 | kergoth`bbl | treke: thats what we do to handle it now |
| 23:09.35 | gb2 | kergoth: will this work: |
| 23:09.38 | gb2 | foo = bar \ |
| 23:09.39 | gb2 | baz |
| 23:09.58 | kergoth`bbl | yep |
| 23:10.27 | treke | foo[]= baz? |
| 23:10.52 | kergoth`bbl | that makes it sound like you're initializing a damn array |
| 23:10.54 | kergoth`bbl | no thanks |
| 23:10.55 | kergoth`bbl | heh |
| 23:10.58 | treke | hehe :p |
| 23:11.48 | gb2 | foo <= baz |
| 23:11.53 | BBMatthew | ljp: in the header you have iostream.h but in the main cpp you dont |
| 23:12.05 | kergoth`bbl | gb2: um, no comparison operators. hehe. |
| 23:12.13 | treke | foo .= ? |
| 23:12.13 | BBMatthew | would compile but working on code in th ebuildroot is really dangerous |
| 23:12.26 | gb2 | treke: that's confusing with the perl .= |
| 23:12.29 | kergoth | .= wouldnt be bad. like perl's string concatenation |
| 23:12.31 | treke | foo += appends a space, .= just does a straight concat? |
| 23:12.31 | kergoth | yep |
| 23:12.36 | kergoth | but we _are_ concatenating |
| 23:12.41 | kergoth | treke: that i wouldnt be opposed to |
| 23:12.46 | gb2 | i like that |
| 23:12.50 | kergoth | same |
| 23:12.51 | gb2 | treke's suggestion |
| 23:13.36 | kergoth | can kill prepend in favor of THIS := ${THIS} stuff ala make |
| 23:13.41 | ljp | ahh |
| 23:13.59 | ljp | hang on.. |
| 23:17.05 | ljp | cvs up |
| 23:17.23 | gb2 | kergoth: that's probably more intuitive |
| 23:17.38 | gb2 | _specials might lead to confusion |
| 23:18.16 | BBMatthew | ljp: fixed it yet :P |
| 23:18.29 | kergoth | gb2: yep. |
| 23:18.39 | kergoth | gb2: we still need a way to handle overrides, just because its a damn cool feature. |
| 23:18.49 | kergoth | but the specials, definately gone |
| 23:22.11 | TheMasterMind1 | time to watch orgazmo |
| 23:22.29 | gb2 | kergoth: yeah. |
| 23:23.01 | gb2 | DEPENDS is build depends, and RDEPENDS is package depends, right? |
| 23:23.10 | kergoth | yeah |
| 23:23.11 | kergoth | R is runtime |
| 23:23.38 | *** join/#openembedded TheMasterMind1 (~aman@h-69-3-87-233.MCLNVA23.covad.net) |
| 23:23.41 | kergoth | a bunch of our .oe's are missing correct deps |
| 23:23.48 | ljp | does kitchensync sync with Sharps new pim? |
| 23:24.01 | kergoth | we really need a helper to check the actual elf shared lib deps on the binaries and populate RDEPENDS |
| 23:24.34 | gb2 | yeah |
| 23:24.50 | gb2 | rpm has some scripts we might be able to easily steal |
| 23:24.55 | treke | ljp: just xml afaik |
| 23:24.57 | kergoth | yeah, so does debian |
| 23:25.01 | kergoth | the debhelper stuff |
| 23:25.44 | kergoth | thing is, how doe it associate the soname of a dep to the runtime package name |
| 23:26.05 | kergoth | either we make the runtime lib package Provides the lib filename, or we find a way to make that association |
| 23:26.26 | kergoth | or we have it establish package name based on a lib package naming policy variable |
| 23:29.13 | *** join/#openembedded fdask (~fdask@CPE00e0292582a8-CM024020107339.cpe.net.cable.rogers.com) |
| 23:29.13 | gb2 | S has a default value, right? |
| 23:29.23 | gb2 | hrm |
| 23:29.29 | gb2 | i need to override it anyway |
| 23:29.49 | kergoth | not sure |
| 23:29.54 | kergoth | if it doesnt, we should define one |
| 23:30.00 | kergoth | default should be ${WORKDIR}/${P} |
| 23:30.18 | pb_ | kergoth: I guess you can just construct a map of which files are shipped in which package, and then use that to map sonames backwards onto the providing package name. |
| 23:30.49 | kergoth | pb_: yeah, thought about that too. but that implies that the packages that will provides those names were built in _this_ build run |
| 23:31.05 | kergoth | guess thats okay |
| 23:31.15 | kergoth | but what if it digs something out of the toolchain lib path? |
| 23:31.15 | pb_ | kergoth: well, they probably need to either have been built, or be available as binaries. |
| 23:31.17 | kergoth | hehe |
| 23:31.57 | pb_ | kergoth: that doesn't sound like the kind of thing one would really want to encourage. |
| 23:32.28 | kergoth | agreed |
| 23:32.58 | kergoth | guess this fits into our need to be able to access the metadata of other packages |
| 23:33.14 | kergoth | or at least, push vars up from the package into a shared metadata store |
| 23:33.17 | pb_ | fwiw, I'm fairly sure that's how dpkg-shlibdeps works. |
| 23:33.26 | kergoth | cool |
| 23:33.32 | pb_ | though of course it has the advantage of being able to poke in /var/lib/dpkg for the metadata. |
| 23:34.34 | kergoth | yeah, exactly |
| 23:36.23 | BBMatthew | looks like a clean build at last |
| 23:37.30 | BBMatthew | ahh that would be until I did a slightly foolish thing - ie a make clean in the wrong dir |
| 23:38.16 | *** join/#openembedded andersee (~andersee@codepoet.org) |
| 23:40.52 | BBMatthew | but the ipks are still there so I should be able to make a initrd! |
| 23:50.16 | gb2 | hrm |
| 23:50.19 | gb2 | NOTE: package base/uclibc-0.9.21-r0: dependency patcher unsatisfied |
| 23:50.40 | treke | make sure patcher is in OEFILES |
| 23:51.13 | BBMatthew | gah its missing some important libs AGAIN |
| 23:51.30 | BBMatthew | what ipk would libgcc be in? |
| 23:52.12 | gb2 | odd |
| 23:54.00 | kergoth | heh, we need a WARN: |
| 23:54.12 | gb2 | how is oemake supposed to be used? |
| 23:54.13 | TheMasterMind1 | no schooool!! |
| 23:54.18 | kergoth | gb2: read --help |
| 23:54.21 | kergoth | example |
| 23:54.25 | gb2 | i want to specify a single package and have it build all its deps then it |
| 23:54.33 | kergoth | OEFILES=/path/to/packages/*.oe oemake packagename |
| 23:54.36 | kergoth | that will do what you want |
| 23:54.40 | gb2 | ah, ok. |
| 23:54.45 | kergoth | it'll then have all the .oe files available, just only build what you want |
| 23:55.30 | *** join/#openembedded stigger (stigger@p065-n1-gos.nlc.net.au) |
| 23:55.57 | kergoth | damnit, went to assemble my tv stand and realized it uses dowels and i have no wood glue |
| 23:56.04 | kergoth | time to go back to sears |
| 23:56.09 | treke | sucker |
| 23:56.23 | kergoth | heh. buying furniture doesnt help much as it sits unassemble |
| 23:56.24 | kergoth | d |
| 23:56.39 | BBMatthew | is that a popular chain? |
| 23:56.39 | gb2 | it started off building a package not in th dependency tree |
| 23:56.45 | treke | BBMatthew: Sears? |
| 23:56.51 | kergoth | gb2: ? |
| 23:56.52 | treke | I guess you could call it popular |
| 23:56.56 | kergoth | gb2: such as? |
| 23:57.01 | kergoth | patcher is always in the dependency chain |
| 23:57.09 | kergoth | its in base.oeclass's DEPENDS_append or something |
| 23:57.13 | BBMatthew | none in the UK |
| 23:57.15 | kergoth | so that the base do_patch works |
| 23:57.18 | gb2 | yeah, i have that. |
| 23:57.31 | kergoth | seems to work fine here |
| 23:57.35 | treke | BBMatthew: It's a us based company. Pretty big. |
| 23:57.41 | gb2 | maybe i'm missing a PN := in here |
| 23:57.46 | kergoth | keep in mind that oemake will build _all_ satisfiers of a dependency |
| 23:57.54 | kergoth | so in multiple provides situations, it builds them all |
| 23:57.55 | kergoth | hehe |
| 23:57.58 | kergoth | need to fix that |
| 23:58.23 | gb2 | nope. hrm. |
| 23:58.37 | gb2 | i have uclibc-gcc which is included by other packages |
| 23:58.47 | gb2 | but isn't itself in the deps tree of what i tried to do |
| 23:58.50 | ljp | I'm hungry |
| 23:59.08 | kergoth | gb2: you're setting PN after the include? |
| 23:59.15 | gb2 | yes |
| 23:59.27 | chouimat | re |
| 23:59.28 | kergoth | odd |
| 23:59.36 | kergoth | does building that package alone behave properly? |
| 23:59.45 | gb2 | well.. it does nothing |
| 23:59.52 | gb2 | because i didn't fill in the building part |