00:00.24 | zecke | kergoth: I'll need to look into this |
00:00.27 | zecke | kergoth: first I need to get my qwertz keyboard going |
00:00.41 | kergoth | most of the zauruses need proper kernel keymaps still |
00:00.50 | kergoth | a lot of the combinations arent mapped |
00:01.15 | zecke | kergoth: theoretically you could map any key to be the Alt modifier right? |
00:01.24 | kergoth | afaik thats correct. |
00:01.30 | *** join/#oe hufnus (~slonsiki@DSL135-071.labridge.com) |
00:01.43 | zecke | I wouldn't know how but that would be scancode ->keycode conversion |
00:01.48 | zecke | and we use K_MEDIUMRAW |
00:01.52 | kergoth | heh, cant load library: libqpe.so.1 |
00:01.54 | kergoth | damnit |
00:02.24 | mikelp | kergoth: /opt/cross/depmod_x86_64/arm-linux-depmod |
00:08.18 | kergoth | ah ha, poodle was setup to use the old modutils-collatoral stuff |
00:08.20 | kergoth | need to fix that |
00:09.07 | mikelp | kergoth: doing a binary fix is gonna blow up on non-x86 compat build arches |
00:09.59 | kergoth | thats why i said itd be a modutils-cross in a seperate package repo from upstream. doesnt need to get pushed, but itd be handy locally on the opteron |
00:10.13 | kergoth | setup a packages.local repo with the modified modutils-cross that swipes the binary |
00:10.21 | kergoth | just to avoid teh manual crap |
00:10.25 | kergoth | heh |
00:10.29 | mikelp | conditional patching in oe would be nice too |
00:11.04 | kergoth | you can do conditional patching. |
00:11.07 | mikelp | if (barfed_upstream) |
00:11.08 | kergoth | qte does, as do other builds |
00:11.15 | mikelp | Hmmm, me and my big mouth |
00:11.18 | kergoth | oh, you mean not just on overrides? |
00:11.32 | kergoth | SRC_URI_append_x86_64 += "file://heh" |
00:11.37 | kergoth | works |
00:11.43 | mikelp | the patch is true for everything |
00:11.54 | mikelp | and that line will blow up when upstream fixes, the patch will reject |
00:12.08 | kergoth | then what do you mean by conditional patching? oh, poke into its tree and check? |
00:12.14 | mikelp | yeah, evil shit |
00:12.28 | mikelp | or don't fail on already applied patch |
00:12.35 | kergoth | nod |
00:12.36 | mikelp | but that's nasty, we'll get lazy and leave patches around forever |
00:12.38 | kergoth | yeah |
00:12.52 | kergoth | 'could be done easily via another parameter on the url, but yeah, evil |
00:13.14 | mikelp | we could implement the SRC_URI_append_pb's_fault ;) |
00:15.31 | *** join/#oe CoreDump|afk (~mhentges@xdsl-213-196-240-25.netcologne.de) |
00:19.11 | kergoth | goddamnit |
00:19.19 | *** join/#oe CoreDump|af- (~mhentges@xdsl-213-196-240-25.netcologne.de) |
00:19.36 | kergoth | aha, had a .patches dir in my build dir |
00:19.39 | kergoth | no wonder patches werent applying |
00:19.51 | kergoth | i wish calling patcher with no arguments or with --help would stop creating .patches |
00:19.54 | kergoth | thats really fucking annoying |
00:20.46 | *** join/#oe Virusmaster (~Thomas@p5091BED5.dip.t-dialin.net) |
00:21.40 | mikelp | gar, the two build are bouncing processors like a bitch, can't be good for cache coherency |
00:22.14 | *** join/#oe noclouds (~mhfan@61.191.205.190) |
00:27.17 | zecke | kergoth: http://handhelds.org/~zecke/downloads/python_keytable_creator.py for my script |
00:27.43 | *** join/#oe tigrux (~tigrux@201.135.236.13) |
00:29.21 | *** join/#oe sunny (sunny@sunny.user) |
00:44.17 | *** join/#oe Cwiiis (~cwiiis@82-43-43-168.cable.ubr02.croy.blueyonder.co.uk) |
00:49.08 | *** join/#oe tigrux (~tigrux@201.135.236.13) |
00:51.18 | *** join/#oe CoreDump|af- (~mhentges@xdsl-213-196-243-217.netcologne.de) |
00:59.26 | *** join/#oe glm2k (~glm2k@rrcs-24-199-11-44.west.biz.rr.com) |
01:18.35 | *** join/#oe qfh (~qfh@ppp-62-245-160-69.mnet-online.de) |
01:36.54 | *** join/#oe Nevarrie_sl6000 (~zaurus@CPE-69-76-210-201.kc.rr.com) |
01:42.53 | kergoth | hmm, cmdline mtd partition parsing isnt working |
01:59.58 | *** join/#oe tigrux (~tigrux@201.135.236.13) |
02:15.05 | *** join/#oe hufnus (~slonsiki@DSL135-071.labridge.com) |
02:22.50 | mikelp | Nice |
02:22.58 | mikelp | <PROTECTED> |
02:40.18 | *** join/#oe mithro (~tim@dsl1-83.gw1.adl1.airnet.com.au) |
02:43.34 | kergoth | woot |
02:43.44 | kergoth | fixed commandline partition parsing for the sharpsl nand in 2.6 |
02:43.45 | kergoth | lets see |
02:44.27 | mikelp | Nope, doesn't exist |
02:50.08 | mikelp | aha a mickeyl |
02:50.10 | *** join/#oe noobie (~matsavhal@81-178-44-43.dsl.pipex.com) |
02:50.21 | *** part/#oe Nevarrie_sl6000 (~zaurus@CPE-69-76-210-201.kc.rr.com) |
02:50.28 | noobie | could someone help me with my local.conf, cant seem to get an env variable for OEFILES that works |
02:50.42 | kergoth | whats there to understand? |
02:50.46 | kergoth | point it at your packages directory |
02:50.51 | kergoth | wherever it lies in your filesystem |
02:51.11 | mikelp | the set env scripts isn't doing things it shouldn't ? |
02:51.43 | noobie | /home/user/src/packages is where i cloned the packages and ive tried that |
02:51.56 | noobie | bit of a noob, trying to dive in |
02:52.00 | kergoth | you set it to /home/user/src/packages/*/*.oe |
02:52.05 | kergoth | *read* the examples and comments |
02:52.11 | kergoth | we put them in there for a reason |
02:53.41 | noobie | sorry typo mistake, seems to be working now |
02:53.54 | noobie | i read a lot b4 i asked tho.... :-) |
02:54.03 | kergoth | heh, good |
02:54.10 | noobie | if i can get it going u definitely have an idiot proof system |
02:54.12 | noobie | :-) |
02:54.59 | mikelp | kergoth: mickeyl in charge of the opie cvs tree ? |
02:55.47 | noobie | to compile for 860, all i do is change the model from collie right? (sorry if this is another stupid quiestion) |
02:56.06 | *** join/#oe CoreDump|afk (~mhentges@xdsl-213-196-243-217.netcologne.de) |
02:56.22 | kergoth | mikelp: mickeyl is one of the opie devs. |
02:56.37 | mikelp | Hmmm, wtf |
02:56.44 | mikelp | Its in the cvs tree, now why didn't it get pulled |
03:00.25 | treke|home | CVSDATE is hardcoded to a few days ago |
03:00.42 | mikelp | fucks it up when the file was only added to the tree 6 hours ago |
03:00.52 | treke|home | yup |
03:01.04 | treke|home | feel free to bump cvs date for that opie package |
03:01.24 | treke|home | opie is going to be releasing with a new opie release |
03:01.29 | mikelp | I'll get more fun out of ragging on mickey about it tomorrow |
03:02.08 | treke|home | we're trying to avoid updating packages unnecessarily before release |
03:02.13 | mikelp | If I'm reading this right though the CVSDATE for opie is set to today |
03:05.04 | *** join/#oe chouimat (~dieu@r2351064.cidc.net) |
03:05.30 | kergoth | fuck fuck fuck |
03:05.32 | mikelp | Hmmm, germany is 6 hours ahead |
03:05.38 | kergoth | my sl5600's leds are blinking |
03:05.39 | kergoth | both |
03:05.43 | kergoth | immediately after hitting fufll reset |
03:05.48 | kergoth | cant even go into the updater |
03:05.51 | kergoth | this cant be good |
03:05.59 | mikelp | fubar'ed nand ? |
03:06.27 | kergoth | must be |
03:06.35 | kergoth | i'm fucking around with mtd mapping drivers |
03:06.40 | kergoth | nothing shouldve done this though |
03:06.46 | kergoth | unless jffs2 tried to mount mtd1 or something |
03:06.59 | kergoth | and fucked it up in the process |
03:07.16 | kergoth | mikelp: any idea wher ei can get a full nand backup image from a 5600? |
03:09.09 | *** join/#oe hufnus (~slonsiki@DSL135-071.labridge.com) |
03:09.12 | mikelp | gotta be googlable |
03:09.21 | kergoth | yeah, that involves opening a browser |
03:09.23 | kergoth | damn laziness |
03:09.50 | mikelp | sec |
03:11.09 | ljp | take the battery out |
03:12.52 | kergoth | fuck |
03:13.02 | kergoth | i found recovery images for c700 c750 c760 and c860 |
03:13.03 | kergoth | no 5600 |
03:13.07 | kergoth | ljp: you have a 5600, right? |
03:13.45 | ljp | ya |
03:13.53 | ljp | has qtopia 2.1 on it |
03:14.56 | mikelp | I've found a link, but the dude has taken the image down, have to ask for it |
03:15.23 | mikelp | gar, fresh cvs pull get the file |
03:15.50 | ljp | soon to have qtopia phone on it |
03:16.19 | kergoth | mikelp: yeah, i found that too |
03:16.29 | kergoth | ljp: can you go into the diags menu and dump a full nand backup image? |
03:16.35 | kergoth | ljp: do you have a large card to put it on? |
03:16.43 | kergoth | 64mb, for the full nand contents |
03:16.43 | ljp | biggest here is 64 |
03:16.50 | kergoth | hmm, thatd cut it close |
03:16.52 | kergoth | damnit |
03:19.48 | treke|home | mikelp: What is this fixed opie do anyway? |
03:19.52 | treke|home | s/is/does |
03:20.45 | mikelp | treke|home: provide distinct 640x480 images |
03:20.49 | treke|home | ah |
03:21.06 | mikelp | gar |
03:21.13 | mikelp | Its cvs fuckage |
03:21.24 | mikelp | It'll work tomorrow |
03:22.24 | ljp | how do I get to the maint window? |
03:22.58 | kergoth | you mean the diags? |
03:23.03 | kergoth | diags is how you do the full nadn backup/restore |
03:23.21 | mikelp | So, CVSDATE is set to 20040913, and the cvs server is making that HEAD cut off before mickey added the files |
03:23.32 | ljp | ya that |
03:23.35 | kergoth | maint menu is how you flash. just hold down ok as you press power, after a full reset |
03:23.40 | kergoth | let me look up the combo for diags |
03:23.42 | ljp | diag |
03:23.45 | mikelp | D+M ? |
03:24.07 | ljp | yup |
03:24.15 | kergoth | yep, d+m |
03:24.23 | kergoth | heh |
03:24.33 | kergoth | Press and hold two main keys [D] and [M], and press the reset switch at the same time. |
03:24.45 | kergoth | Caution: Be careful not to press any other key. Be sure to press the proper keys to enter the service diag. |
03:25.17 | kergoth | alright, emailed that guy about the backup image |
03:25.52 | kergoth | this pisses me off, finally got the cmdline partitioning working properly for nand, and the ROM chip was being picked up via CFI + physmap |
03:25.59 | kergoth | was ready to get down to some other drivers |
03:26.36 | treke|home | I guess you will have to do it on the c700 |
03:26.39 | treke|home | such a shame |
03:27.00 | kergoth | the c700 is in my crossfire's trunk |
03:27.04 | kergoth | at the body shop |
03:27.12 | treke|home | wtf |
03:27.16 | kergoth | my sd card is in the slot of the c700 |
03:27.18 | treke|home | the thing isn't done yet? |
03:27.19 | kergoth | heh |
03:27.27 | kergoth | dood, its been in the shop for over a month now |
03:27.33 | kergoth | they say i can definately get it back tomorrow |
03:27.35 | treke|home | it probably doesn't take this long to build one of the fuckers |
03:27.45 | kergoth | but i take tht with a grain of salt, they've been giving me dates and missing them for weeks |
03:27.49 | kergoth | so will see |
03:28.02 | mikelp | So, I'm betting CVDATE=20040913 really means 20040913 00:00 |
03:28.11 | treke|home | yes it does |
03:28.36 | gb2 | mike: UTC |
03:28.39 | treke|home | someone was adding a way to specify the absolute tip of head iirc |
03:28.45 | mikelp | gb2: Nope, local to repository |
03:28.56 | treke|home | mikelp: should be UTC |
03:28.58 | mikelp | unless you actually spec a timezone |
03:29.05 | gb2 | mike: CVS uses UTC dates |
03:29.05 | kergoth | we pass the date as UTC into cvs |
03:29.16 | kergoth | no idea how it makes use of it |
03:29.52 | mikelp | <PROTECTED> |
03:29.52 | mikelp | <PROTECTED> |
03:30.05 | mikelp | so, unless was say its UTC, its even worse |
03:30.11 | gb2 | uh |
03:30.12 | mikelp | s/was/we/ |
03:30.41 | mikelp | anyway, immaterial |
03:30.48 | mikelp | 20040913 00:00 UTC would still not get the file |
03:30.58 | gb2 | hrm. |
03:30.59 | gb2 | odd |
03:31.12 | treke|home | set the exact time |
03:31.14 | gb2 | i could have sworn there was some force as UTC thing that cvs did by default |
03:31.18 | treke|home | or try HEAD as the cvsdate |
03:31.47 | mikelp | Wait, we can't be setting it to UTC |
03:31.53 | mikelp | DATE -> UTC would give 0914 |
03:31.57 | gb2 | date: OE's ${DATE} is utc |
03:32.02 | gb2 | s/date/mike/ |
03:32.11 | mikelp | I beg to differ |
03:32.14 | gb2 | uh |
03:32.19 | gb2 | i changed it to be UTC |
03:32.25 | gb2 | unless someone went and fucked it up again |
03:33.45 | gb2 | DATE := "${@time.strftime('%Y%m%d',time.gmtime())}" |
03:33.49 | gb2 | that's what's in bk |
03:33.53 | mikelp | Right, that's what I've got |
03:33.55 | gb2 | that should be UTC |
03:34.13 | mikelp | Right, which should give us 20040914 now |
03:34.38 | mikelp | gar |
03:34.53 | mikelp | Its frigging overwritten in openzaurus-3.5.1.conf |
03:34.57 | gb2 | heh |
03:35.09 | mikelp | openzaurus-3.5.1.conf:CVSDATE = 20040913 |
03:35.16 | gb2 | that sounds correct |
03:35.19 | gb2 | pick a date |
03:35.22 | gb2 | for a versioned release |
03:36.05 | mikelp | and then change the .oe to expect a file added after the date |
03:36.13 | mikelp | whee - boom |
03:37.08 | treke|home | so you override it in local.conf |
03:37.23 | mikelp | I'm not too pissed, I'm trying to hunt down oz 3.5.1 fuck ups |
03:37.26 | mikelp | this is one of them |
03:37.53 | treke|home | how is this a fuckup in oz? |
03:38.31 | mikelp | because distro is set to openzaurus-3.5.1 |
03:38.41 | treke|home | ? |
03:38.45 | treke|home | What is the fuckup? |
03:38.48 | mikelp | I shouldn't have to override in local.conf to build |
03:38.59 | treke|home | I built with it this morning |
03:39.04 | kergoth | you cant override in local.conf anyway. DISTRO .conf is included after local.conf is |
03:39.13 | mikelp | mickey made the change only 7 hrs ago |
03:40.42 | mikelp | I hesitate to change the oz-3.5.1 cvdate to 0914, could break other shit |
03:41.44 | ljp | ok, backing up to 64 mb cf |
03:41.52 | kergoth | cool, hopefully it fits |
03:42.04 | ljp | it didnt complain about no enough space |
03:42.10 | treke|home | mikelp: only change the cvs date on the necessary packages |
03:42.30 | mikelp | urgh, ugly |
03:42.52 | treke|home | changing the dates on everything for one package is the ugly part :) |
03:43.07 | mikelp | what's the syntax for that anyway CVSDATE_opie-taskbar = "20040914" ? |
03:43.11 | gb2 | cvs under unix uses utc dates i think |
03:43.18 | treke|home | yup that's right |
03:43.21 | gb2 | based on the description of the CVS/Entries file |
03:44.02 | mikelp | treke|home: best place for the override, oz-3.5.1.conf ? |
03:44.08 | treke|home | yes |
03:44.46 | gb2 | > The timezone on the timestamp in CVS/Entries (local or universal) should be the same as the operating system stores for the timestamp of the file itself. |
03:45.07 | gb2 | and unix uses universal dates |
03:45.12 | mikelp | So you're at the mercy of the cvs server ? |
03:45.33 | gb2 | yes |
03:45.43 | gb2 | assuming the server is not vms or windows |
03:45.43 | gb2 | you're fine |
03:46.34 | mikelp | gb2: so, OE's DATE, specific down to the minute ? |
03:47.03 | gb2 | no |
03:47.06 | gb2 | DATETIME is though |
03:47.18 | gb2 | in cvs, if you don't specify a time, it's 00:00 |
03:47.40 | mikelp | OK, so we'd still get screwed |
03:49.22 | gb2 | poor design |
03:50.02 | mikelp | Its a small corner case |
03:50.26 | mikelp | We set a cvsdate for a distro which is the right thing to do, file get updated, forget to bump date |
03:51.56 | gb2 | explicit overrides for the user of oz-3.5.1 |
03:52.06 | gb2 | sounds like the right solution |
03:52.21 | ljp | hmm |
03:52.31 | mikelp | Yeah, just my luck to hit it. But that's what trying completely clean builds is all about |
03:52.34 | ljp | still apparently backing up... |
03:55.24 | mikelp | treke|home: setting that CVSDATE worked, but its the wrong answer |
03:55.45 | mikelp | we grab too much from opie |
03:56.21 | kergoth | we have too many builds from cvs HEAD |
03:56.23 | mikelp | And you end up with two .tar.gz of the same thing |
03:56.32 | mikelp | with diff cvsdates |
03:56.33 | kergoth | need to lock things down and move them forward when someone has actually tested current |
03:56.57 | kergoth | mikelp: someone has a patch to make our cvs fetcher do a cvs update to get the new tarballs |
03:57.04 | kergoth | never got accepted and checked in |
03:57.22 | mikelp | and untar, cvs update, retar ? |
03:57.31 | kergoth | iirc thats how it worked |
03:57.32 | mikelp | nice |
03:57.38 | kergoth | check teh oe ml archives |
03:57.55 | mikelp | bandwidth is cheap though... |
03:58.08 | kergoth | yep |
03:59.59 | mikelp | I hate to think how many copies of the same stuff we've got on this drive now |
04:00.44 | kergoth | heh, yeah, should create a global oe sources dir for downloads, except that oe doesnt use file locking, so if we both try to download the same file at the same time, boom |
04:00.51 | kergoth | :) |