**** BEGIN LOGGING AT Sun Oct 21 02:59:58 2012 Oct 21 05:31:48 Any advice please. Nitdroid is installed using u boot and no wireless networks show up at all Oct 21 07:18:36 DocScrutinizer05: what about that thermistor? Oct 21 09:51:15 my n900 rebooted and asked for a pin code while I was sleeping. How can I see what was the reason? Oct 21 09:51:55 akls: /proc/bootreason is a start Oct 21 09:53:16 sw_rst Oct 21 09:53:47 something rebooted Oct 21 09:53:59 do you have syslog installed? Oct 21 09:55:03 I did not install it Oct 21 09:55:08 so I guess that no Oct 21 09:55:37 anything weird going on? Oct 21 09:56:12 no, everything is fine. But last night I've opened my easy debian chroot and started upgrading it Oct 21 09:56:16 but the upgrade has finished Oct 21 09:56:22 successfully, I guess Oct 21 09:56:29 because everything is fine with apt-get Oct 21 09:56:43 ever had any problems with the phone? Oct 21 09:56:57 nope, I have it just for 4 days :) Oct 21 09:57:36 just stock fremantle then? Oct 21 09:57:42 was the battery charged? Oct 21 09:57:45 DocScrutinizer05: *poke* Oct 21 09:57:50 no, community ssu testing Oct 21 09:58:02 it was connected to my pc via usb charger Oct 21 09:58:05 and the battery is ok Oct 21 09:59:33 kerio, should I install syslog and then wait for the next time it happens? Oct 21 10:13:27 i suppose Oct 21 10:14:09 so, about "apt-get upgrade" in easy debian... It makes the device hang and reboot forcing you to reinstall easy debian Oct 21 10:14:24 so I've made a little script that upgrades the packages one by one Oct 21 10:14:27 https://dl.dropbox.com/u/71236259/update1by1 Oct 21 10:14:40 oooh, i think i know what went wrong then Oct 21 10:14:43 it requires apt-show-versions to be installed Oct 21 10:14:56 ShadowJK can probably back me up on this Oct 21 10:15:10 akls: IO intensive things make everything slow to a crawl Oct 21 10:15:24 and there's a watchdog that reboots the system if it isn't tickled every $something seconds Oct 21 10:15:48 kerio, I was upgrading with my script. And upgrade has finished successfully Oct 21 10:15:57 so it's not the reason Oct 21 10:15:59 then idk Oct 21 10:17:12 anyway, i've got http://pastebin.mozilla.org/1873316 this at the end of /etc/init.d/rcS Oct 21 10:17:53 what's that? Oct 21 10:18:03 I mean what is it going to do? Oct 21 10:18:16 speed up I/O quite dramatically Oct 21 10:18:26 at the cost of a bit more ram used for I/O Oct 21 10:18:50 ShadowJK: is that right? Oct 21 10:28:14 oh my fucking god what the hell is wrong with the easydebian developer Oct 21 10:29:16 qole? Oct 21 10:30:36 putting files in /root/ and /home/user, cycling a swap partition without even checking that it's there, overwriting /sys/block/mmcblk*/queue/nr_requests Oct 21 10:32:44 it's almost worse than easy-chroot's "user ALL = NOPASSWD: ALL" Oct 21 10:39:19 http://imagebin.org/232698 -> Hey, how can I insert a sim card safely into my new n900? Oct 21 10:39:40 I recall that previously it was simple to remove the relevant cover, but it does not seem so now. Oct 21 10:40:20 i don't get the meaning of that picture Oct 21 10:40:33 the sim is below the battery Oct 21 10:40:51 ah yes, I am confusing this with the N950. Oct 21 10:41:04 thank you! Oct 21 10:41:22 those are all buttons! D: Oct 21 10:41:27 yep Oct 21 10:41:40 long time I have used N900 :D Oct 21 10:42:04 but I have no other fallback phone here now ^^ Oct 21 10:43:00 Anyone knows is there a script that is triggered by opening/closing the back cover? Oct 21 10:43:12 so how can I get the back off Oct 21 10:43:42 djszapi: idk, a thumbnail Oct 21 10:44:36 djszapi: on the right side of the phone (in landscape) you've got a place to stick your nail in. Place it there and pull it Oct 21 10:44:40 kerio: http://pastebin.mozilla.org/1873364 Oct 21 10:45:37 MrPingu: oh, neat Oct 21 10:45:40 what's "service" there? Oct 21 10:45:56 also, why the stop directive? Oct 21 10:46:05 and why don't you run it earlier? Oct 21 10:46:22 http://wiki.maemo.org/Startup_script Oct 21 10:46:28 took just the example script :P Oct 21 10:46:39 Psotnick: ok, if that is safe ok then. Oct 21 10:47:03 djszapi: it should be safe I do it like that ;) Oct 21 10:47:05 I thought there would have been a button (or two) to press to get that automatically off. Oct 21 10:47:17 kerio: What would you propose then? Oct 21 10:47:20 djszapi: it's safe, although it 'sounds' dangerous :p Oct 21 10:47:27 the end of rcS :) Oct 21 10:47:44 unless those parameters get overwritten afterwards Oct 21 10:48:19 Sicelo: Psotnick thank you. Oct 21 10:49:08 djszapi: no problem Oct 21 10:50:07 MrPingu: where have you taken those parameters? Oct 21 10:50:13 page-cluster 0 seems weird Oct 21 10:50:51 FMG, on a post in the easy-debian thread Oct 21 10:51:06 oh freemangordon <3 Oct 21 10:51:55 http://www.forums.internettablettalk.com/showthread.php?p=1200669#post1200669 Oct 21 10:52:38 Increased nr_requests acording to ShadowJK Oct 21 10:56:30 is the n900 sim lock free? Oct 21 10:57:49 my vodafone sim card does not get recognized after the boot, apparently. Oct 21 10:59:37 I got this instance at the meego summit after the meegathon competition. Oct 21 10:59:54 so probably not from the operators, but directly from Nokia. Oct 21 11:01:04 hm, it was ok for the second boot. Oct 21 11:18:42 How can I obtain omap_hsmmc.c from N900 kerel power? Oct 21 11:35:43 I think N900 has the capability to do simlock Oct 21 11:36:31 Now whether you bought N900 from an operator or from a normal shop without operator contracts, and if operator locked it or not, is the question :P Oct 21 12:39:52 https://dl.dropbox.com/u/71236259/2012-10-21-153943_160x29_scrot.png hahahahha Oct 21 12:44:32 freemangordon: ((what about that thermistor?)) nice table with values, from -40 to +120 Oct 21 12:48:23 (( so I've made a little script that upgrades the packages one by one kerio, I was upgrading with my script. And upgrade has finished successfully)) how do you even know? does your script some logging, or rise a flag file when done? Oct 21 12:51:10 DocScrutinizer05, what do you mean? O_o Oct 21 12:51:36 DocScrutinizer05, it just gets the list of upgradable packages and then upgrades them one by one Oct 21 12:51:42 and sleeps for 20 seconds between each other Oct 21 12:51:47 akls: when your N900 reboots at arbitrary time, how would you tell your script had finished at that time? Oct 21 12:52:14 DocScrutinizer05, oh, apt-get upgrade says that there are no packages to upgrade Oct 21 12:52:22 aaah Oct 21 12:52:24 and that dpkg was not interrupted Oct 21 12:53:26 my personal suspicion is that your PC stopped charging Oct 21 12:53:47 maybe I'm a psycho, but I'm trying to load eclipse in easy debian Oct 21 12:54:40 holy shit Oct 21 12:54:41 it works Oct 21 12:54:54 sure, why not Oct 21 12:55:54 you got a pretty standard PC with a debian and a ~250MHz "real" CPU, with 200MB RAM and 700MB swap Oct 21 12:56:09 it crashed O_o Oct 21 12:56:46 yep, sounds like eclipse Oct 21 12:56:53 well, that's not exactly a surprise for any eclipse, no? ;-) Oct 21 12:57:00 no Oct 21 12:57:12 It crashed only once in my 5 years experience Oct 21 12:57:21 maybe twice Oct 21 12:57:21 wow Oct 21 12:57:40 maybe I'm just too lucky Oct 21 12:59:18 or you used eclipse onle once, maybe twice ? ;-P Oct 21 12:59:31 no, I'm using it a lot Oct 21 12:59:34 Anyone knows how can I make my N900 think that the back cover is always on? My sensor's probably broken and I can't use microsd cards. Oct 21 12:59:36 mostly for java development Oct 21 12:59:42 sometimes c++ or python Oct 21 13:00:08 Psotnick, what do you mean by "back cover is always on" ? Oct 21 13:00:13 Psotnick: not your sensor is broken, your cover is. missing magnet Oct 21 13:00:18 I thought there's no sensor for back cover? Oct 21 13:00:25 only for the camera thing Oct 21 13:00:35 there is a hall switch Oct 21 13:00:39 DocScrutinizer05: the magnet is on Oct 21 13:00:43 detecting stand magnet Oct 21 13:00:47 hmm Oct 21 13:01:07 try with a better magnet, for a start Oct 21 13:01:10 it crashed again.. hmm.. Oct 21 13:01:36 Psotnick: it's actually hard to fake that sensor/hallswitch Oct 21 13:02:03 afaik it's embedded deeply in mmc driver Oct 21 13:02:15 DocScrutinizer05: tried with a magnet from a case, they're pretty strong Oct 21 13:02:24 no luck? Oct 21 13:02:31 are you sure you placed it correctly? Oct 21 13:02:40 DocScrutinizer05: no Oct 21 13:02:55 you know where's that sensor? Oct 21 13:03:05 kerio: I tried with a lot of combinations Oct 21 13:03:44 DocScrutinizer05: under the place where the magnet is, between battery and slot Oct 21 13:04:19 i thought it was between usd and camera Oct 21 13:05:03 DocScrutinizer05: isn't it a hardware thing to completely disable the slot when the sensor doesn't detect the backcover? Oct 21 13:05:44 kerio: probably not, there's a driver for that omap_hsmmc Oct 21 13:06:32 cat /sys/class/mmc_host/mmc0/cover_switch Oct 21 13:07:16 Nokia-N900:~# cat /sys/class/mmc_host/mmc0/cover_switch Oct 21 13:07:16 open Oct 21 13:07:32 stupid thing: try echoing "closed" to it :3 Oct 21 13:07:51 watch -n 1 /sys/class/mmc_host/mmc0/cover_switch Oct 21 13:08:11 then try with a proper magnet Oct 21 13:08:38 after all it *might* really be defect Oct 21 13:08:53 the magnet in cover is the little one holding the stand Oct 21 13:09:18 in the center under the silver frame around cam slider Oct 21 13:09:36 DocScrutinizer05: weren't you saying something about reversing that GPIO pin and turning it on? Oct 21 13:09:38 watch says permission denied Oct 21 13:09:52 Psotnick: watch -n 1 cat /sys/class/mmc_host/mmc0/cover_switch Oct 21 13:09:55 eclipse is extremely slow :D no surprise Oct 21 13:10:13 kerio: thanks Oct 21 13:10:17 kerio: oh, thanks. I missed that Oct 21 13:10:32 * DocScrutinizer05 is missing coffee Oct 21 13:10:52 it says open all the time Oct 21 13:11:03 Psotnick: :( Oct 21 13:11:07 kerio: I once suggested something along that line, yes Oct 21 13:11:36 oh wait.. not that slow Oct 21 13:13:10 Psotnick: I guess mmc_host is the driver, and cover_switch is where it deals with this hall-sensor very directly Oct 21 13:13:26 you need to build your own mmc_host driver I guess Oct 21 13:14:03 is that driver open? Oct 21 13:14:08 why hasn't anybody fixed it, then? Oct 21 13:14:13 it's kernel, so yes? Oct 21 13:14:18 It's open, I've got the sources Oct 21 13:15:00 But it's too hard for me to understand the code Oct 21 13:15:04 DocScrutinizer05: are you sure it's not also a purely hardware thing? Oct 21 13:15:20 (note the "also") Oct 21 13:16:10 I thought about replacing all 'mmc_slot(host).get_cover_state' with 1, but I don't know what 'mmc_slot(host).get_cover_state(host->dev, host->slot_id)' stands for Oct 21 13:17:23 kerio: yep, I'm quite sure Oct 21 13:21:47 kerio: N5820 "Main Camera Cover Switch" goes to SoC pin C26 "CAM_D11" GPIO. No relation whatsoever to mmc-interface hw block Oct 21 13:22:10 k Oct 21 13:23:15 DocScrutinizer05: the sensor can be fixed only with new motherboard? Oct 21 13:25:56 ooops sorry, I told BS Oct 21 13:27:29 BS? Oct 21 13:27:46 N5210 "uSD card hotswap sensor" connects to SoC GPIO160 (pin T21) "McBSP1_CLKS". So the basic rationale stays the same: not related to uSD hw interface Oct 21 13:29:48 Psotnick: (new MoBo) that depends what's defect Oct 21 13:33:43 Ok, I see, thank you for your time Oct 21 13:34:31 if sensor N5210 or ESD-EMI-filter Z5210 are defect, or maybe R5210 or C5212, then you could fix the component Oct 21 13:35:18 anyway I *always* suggest reflashing a clean pure maemo fremantle pr1.3 and then retest Oct 21 13:35:25 Psotnick: ^^^ Oct 21 13:35:36 DocScrutinizer05: I tried that already Oct 21 13:35:41 ok Oct 21 13:36:40 Psotnick: try the gpio thing! Oct 21 13:37:10 the brute force "fix" (if SoC still ok): remove R5210, short "lower end" of R5210 to GND Oct 21 13:37:35 or what kerio said Oct 21 13:37:57 hey, anybody know how to make fmms stop auto rotating? Oct 21 13:37:57 though I'd ask somebody else to first test it on their device if it actually works Oct 21 13:38:00 ;-P Oct 21 13:38:09 I tried blacklisting but it's python Oct 21 13:38:13 I'm too dumb to understand you guys Oct 21 13:38:58 Psotnick: there's a generic way to handle GPIO in linux, via sysfs Oct 21 13:39:18 the sensor is connected to GPIO160 Oct 21 13:41:12 you could define GPIO160 as output, and then set it to 1/on/high, which would - in theory - signal any process watching this GPIO (mmc_host) that the cover is closed Oct 21 13:42:31 (([2012-10-21 15:16:10] I thought about replacing all 'mmc_slot(host).get_cover_state' with 1, but I don't know what 'mmc_slot(host).get_cover_state(host->dev, host->slot_id)' stands for)) sounds like a pretty good plan as well Oct 21 13:43:14 Okay, I think I understand Oct 21 13:44:21 I'm afraid that I can't even compile it Oct 21 13:47:24 Also if that fails, would simple kernel reflash fix that? Oct 21 13:47:33 jon_y: sure Oct 21 13:47:36 Psotnick: isn't it a module? Oct 21 13:47:48 kerio: prolly monolithic Oct 21 13:47:56 kerio: yes, It is Oct 21 13:48:21 err, it's a *.ko?? Oct 21 13:48:31 omap_hsmmc, mmc_block, mmc_core Oct 21 13:48:33 one of those Oct 21 13:49:10 omap_hsmmc.ko is responsible for back cover open/closed status Oct 21 13:49:54 modinfo: could not find module map_hsmmc Oct 21 13:50:17 duh, my bad Oct 21 13:51:35 DocScrutinizer05: put python in blacklist? Oct 21 13:52:46 jon_y: how about starting fmms not via "python fmms.py" but simply via "fmms.py" like any other proper script with proper shebang line? Oct 21 13:53:45 I guess CSSU rotation blacklist goes for $0 Oct 21 13:54:07 or *maybe* it even goes for window title - sorry I have NFC Oct 21 13:54:12 DocScrutinizer05: I don't know, it was started from the app menu Oct 21 13:54:21 I guess I can start it manually Oct 21 13:54:47 which user does it run as, I hope I don't need root Oct 21 13:55:04 I guess pasting the *full* ps line here (from htop if you have no proper ps) will tell us a lot Oct 21 13:55:35 also looking into the .desktop will enlighten us a bit Oct 21 13:55:52 sorry, no fmms here, so I can't help on that Oct 21 13:55:55 oh ok, just a normal user Oct 21 13:56:41 8825 user 2215 python /opt/fmms/fmmsd.py Oct 21 13:57:00 and another .... user .... python /opt/fmms/fmms_gui.py Oct 21 13:57:09 head -n 2 /opt/fmms/fmmsd.py Oct 21 13:57:43 head -n 2 /opt/fmms/fmms_gui.py Oct 21 13:58:04 is there a proper python shebang? Oct 21 13:58:29 #!/usr/bin/env python2.5 Oct 21 13:58:38 I guess it sort of has Oct 21 13:58:41 frals: why is fmms started as "python $fmms" ? Oct 21 13:58:58 sorry, ps says python2.5 too Oct 21 13:59:24 where are the .desktop files to edit? Oct 21 13:59:42 /usr/share/apps/hildon/ Oct 21 13:59:45 or sth Oct 21 14:01:02 hmm, Exec=/opt/fmms/fmms_gui.py Oct 21 14:01:22 hmmm indeed Oct 21 14:03:11 btw, x terminal -> new opens a new console Oct 21 14:03:17 how do I get back the old one? Oct 21 14:03:26 hm? Oct 21 14:03:43 switch to the other window? Oct 21 14:04:00 ugh, I feel stupid :( Oct 21 14:04:59 jon_y: please quit fmms Oct 21 14:05:09 I think /usr/bin/env is starting the python Oct 21 14:05:13 ok Oct 21 14:05:22 YESSSSSSSSSSSSSSSS Oct 21 14:05:22 jon_y: then start it from console, like `/opt/fmms/fmms_gui.py` Oct 21 14:05:36 finally my favorite music player moc on my phone!!! Oct 21 14:05:37 yeah, I did that, python2.5 is still there Oct 21 14:05:45 dang! Oct 21 14:05:57 friggin python Oct 21 14:05:58 running in easy debian, but still Oct 21 14:06:13 I wonder if it drains more or less battery than native n900 music player... Oct 21 14:06:22 jon_y: we'll fix that in CSSU eventually Oct 21 14:06:29 DocScrutinizer05: thanks Oct 21 14:06:35 you guys are awesome Oct 21 14:06:42 ^ agreed Oct 21 14:06:44 I already thought blacklist needs regex Oct 21 14:07:07 does hildon-desktop need restarting if I change the blacklist? Oct 21 14:07:37 regex can get kind of funky if everything is in one line Oct 21 14:07:57 make it m-m-m-multiline? :) Oct 21 14:09:57 sure it's still multiline :-P Oct 21 14:10:30 won't hurt to get simple shell patterns in the names Oct 21 14:10:38 DocScrutinizer05: it is? the example had everything in one line Oct 21 14:10:49 hmmm? Oct 21 14:11:03 sorry I have nfc, never looked into a real blacklist Oct 21 14:11:07 How can I get sources for kernel power? Oct 21 14:11:36 nah, I prefer regex, glob can get crazynfc? Oct 21 14:11:41 oops Oct 21 14:11:44 nfc? Oct 21 14:11:49 ooh Oct 21 14:12:02 it's a single friggin line in transitions.ini Oct 21 14:12:06 * jon_y no like glob Oct 21 14:12:27 ~nfc Oct 21 14:12:28 [nfc] No Fucking Clue, or near field communications Oct 21 14:12:35 oh ok Oct 21 14:13:08 new blacklist.ini? one entry per line? Oct 21 14:13:27 or blacklist.conf since it has little to no ini structure to it Oct 21 14:13:55 jon_y: that's been my requirement specs for blacklist, somebody ignored it :-S Oct 21 14:14:20 actually I specified a dir hildon-blacklist.d/ Oct 21 14:14:44 yes, that would be easier for apps to tell hildon they can't rotate right Oct 21 14:14:53 or even whitelist.d Oct 21 14:14:55 since we all know the PITA that comes from managing one-line edits in post-install post-remove scripts Oct 21 14:15:25 auto rotate can now be ternary, yes, no, auto Oct 21 14:15:38 yes means rotate all except blacklisted Oct 21 14:15:44 I guess I can prod somebody to add that silly little patch to cssu H-D Oct 21 14:15:49 no means no rotate at all Oct 21 14:15:58 auto means rotate only whitelist Oct 21 14:16:20 well, ternary looks good on paper Oct 21 14:23:53 or maybe I have to implement those patches myself :-S Oct 21 14:25:24 1) allow whitespace in blacklist items, for exact matching against full cmdline: blacklist= "appname1 parmeter1 parameter2" Oct 21 14:26:04 2) allow shell patterns: blacklist= "appname1 *" Oct 21 14:26:47 3) allow multiple lines: blacklist= "appname1" \n blacklist= "appname2" Oct 21 14:28:45 4) allow new parameter "blacklistinclude = /path/to/file/with/shellpatterns*", to 'source' all files that result from pathname expansion of the shell patterns in path parameter Oct 21 14:29:23 in those files, same blacklist related syntax applies as in original transitions.ini Oct 21 14:30:10 I don't think it makes sense to forbid blacklistinclude in included files, when stuff gets coded properly Oct 21 14:32:42 so 4a) it's user's responsibility to not define circular blacklistinclude. Nevertheless we may define a max nesting depth of maybe 10, after which H-D aborts evaluation of transitions.ini and throws an error to logging but doesn't exit Oct 21 14:34:22 DocScrutinizer05: or just /etc/hildondesktop.conf and /etc/hildondesktop.conf.d/* Oct 21 14:35:31 nope, I prefer to define arbitrary include patterns, so we *could* for example run over all blacklistinclude=/usr/share/apps/hildon/*.desktop Oct 21 14:37:23 since this windows-configfile syntax allows for arbitrary crap parameters in a file that simply get ignored when unknown, it would be pretty perfect to scan all .desktop for a parameter "blacklist=$me" Oct 21 14:38:18 I hope you appreciate that idea Oct 21 14:38:27 sounds like a mess Oct 21 14:38:35 heh? Oct 21 14:38:48 why? where? which mess? Oct 21 14:40:09 there's no "blacklist" parameter in .desktop files Oct 21 14:40:17 so what? Oct 21 14:40:34 doesn't mean we can't add one Oct 21 14:40:49 if we feel like Oct 21 14:41:45 see freedesktop.org specs of .desktop format Oct 21 14:42:14 it explicitly allows arbitrary parameters that simply get ignored by all the processes that don't know them Oct 21 14:42:28 iirc Oct 21 14:43:02 in fact H-D acts exactly according to that Oct 21 14:43:49 I accidentally tested it just so often, by adding something that had a typo Oct 21 14:46:23 ooh, I forgot to explicitly mention it Oct 21 14:46:46 4b) of course multiple lines "blacklistinclude=" are allowed Oct 21 14:48:06 4c) each include is executed ad hoc, and later definitions override earlier ones, if conflicting Oct 21 14:50:00 (technically all blacklist= parameters are kept in an internal list that gets scanned from end to start for first entry defining about rotation of the app under evaluation) Oct 21 14:52:44 ((footnote: while this last definition is meaningless for current blacklist, it starts to become relevant when we also get whitelist entries that are added to same list)) Oct 21 14:53:51 What should I write to /etc/apt/sources.list to use extras/devel? Oct 21 14:54:45 Oh, it worked Oct 21 14:54:47 4c1) TBD when this list gets refreshed, possible events to trigger refresh are: start of H-D (obviously), opening of taskswitcher screen, inotify evens (TBD), SIGHUP/SIGUSR1 Oct 21 14:54:57 moc eats 100% cpu on easy debian. Is there anything I can do with it? Oct 21 14:56:05 Psotnick: iirc HAM overwrites that each time. i think /etc/apt/sources.list.d/hildon-application-manager.list is more permant Oct 21 14:56:11 *permanent Oct 21 14:56:38 yep, exactly Oct 21 14:56:43 Sicelo: I know, just noticed that ;) Oct 21 14:57:19 even better: add your own files to /etc/apt/sources.list.d/ Oct 21 14:58:21 well, 'better' is relative here Oct 21 15:04:24 errrr /usr/lib/testserver/modules/handlers/sos_switch.so.0 o.O .oO(???) Oct 21 15:06:55 wtf is this testserver? Oct 21 15:10:01 "Report bugs to Paul Mundt " Oct 21 15:18:35 * DocScrutinizer05 is tempted to do a `start test` or `start local` Oct 21 15:18:57 less /etc/event.d/testserver Oct 21 15:22:47 oh LOL: preinit: >> STATE=`getbootstate` echo_g "Entering state '$STATE'." case $STATE in ... LOCAL|TEST) def_runlevel=3 << Oct 21 15:23:34 so what makes getbootstate respond "TEST" ? Oct 21 15:23:39 test-battery? Oct 21 15:24:52 BWAHAHA the notorious >> echo_g "Houston, we have a problem, powering off..." << X-P Oct 21 15:28:44 KTNXBYE! Oct 21 15:28:47 IroN900:~# getbootstate --help Oct 21 15:28:49 getbootstate: Unexpected reset occured. Using previous bootstate 'USER' Oct 21 15:28:50 USER Oct 21 15:34:04 are there any cycling ppl here? Oct 21 15:34:09 #cycling seems empty Oct 21 15:34:21 I want to ask a question about cycling in the alps Oct 21 15:34:34 guessing some ppl here are near the alps Oct 21 15:34:37 honestly, wrong channel Oct 21 15:35:03 which then? Oct 21 15:35:17 maybe try another network completely Oct 21 15:36:24 reading preinit is a real thrill, each single time Oct 21 15:38:45 hm. I'm trying to use my favorite cli music player moc on n900. It works fine with easy debian, but for some reason eats 100% cpu. I've tried opening the same song with mpg123 and when running from user it works fine, but from root it eats 100% cpu too. Oct 21 15:39:18 something is wrong :\ Oct 21 15:40:19 epic: preinit seems to offer booting from SDx "HDD" attached to USB host mode Oct 21 15:40:43 in addition to NAND flash, and MMC Oct 21 15:41:43 less "+/show_roots()" /sbin/preinit Oct 21 15:44:31 gawd Oct 21 16:27:52 anybody? Oct 21 16:36:00 akls, i used to use moc Oct 21 16:36:03 akls: well, mediaplayers shouldn't run with root permissions Oct 21 16:36:19 i think it might not be using hardware acceleration for playing mp3's, although i doubt even then it would eat 100% cpu Oct 21 16:36:24 i'm not sure if that even makes sense Oct 21 16:36:28 that's no reason why it eats 100% CPU, but anyway Oct 21 16:36:31 but it works fine for me Oct 21 16:36:37 so i dont know Oct 21 16:36:55 i ran mine in maemo and not debian Oct 21 16:37:00 i had compiled it Oct 21 16:38:17 akls: anyway if your mediaplayer is run under root account, I guess it will need run-standalone.sh at very least, since it might need access to USER's session bus etc Oct 21 16:38:49 Jade had any cpu problems? Oct 21 16:38:57 nope Oct 21 16:39:00 showed maybe 10-15% at most Oct 21 16:39:03 wtf Oct 21 16:39:03 why Oct 21 16:39:08 Jade was it estel or what? Oct 21 16:39:09 akls: anyway if your mediaplayer is run under root account, I guess it will need run-standalone.sh at very least, since it might need access to USER's session bus etc Oct 21 16:39:16 what's that Oct 21 16:39:38 Jade you were running it under easy debian? Oct 21 16:39:41 oh Oct 21 16:39:41 no Oct 21 16:39:43 OH Oct 21 16:39:45 i compiled it on maemo Oct 21 16:39:48 so installed from source, right? Oct 21 16:39:49 OK Oct 21 16:39:52 got it Oct 21 16:39:58 Jade thank you very much, my savior! Oct 21 16:39:59 mocp doesnt have that many dependencies, it compiles pretty fast Oct 21 16:40:01 now I have to go :) Oct 21 16:40:04 so you should be good if you really want to compile it Oct 21 16:40:04 ok Oct 21 16:40:09 yea-yea Oct 21 16:40:11 thank you! Oct 21 16:40:19 np Oct 21 16:40:29 I was not even thinking about running it without easy debian :D Oct 21 16:40:33 thanks Oct 21 19:57:52 DocScrutinizer05: setting all the "mmc_slot(host).get_cover_state" caused a reboot loop Oct 21 19:58:38 maybe changing anything on that module caused the reboot loop Oct 21 19:59:22 Psotnick: have you tried just loading it from a booted system? Oct 21 19:59:35 if vermagic doen't match, kernel won't load the module, and thus can't mount eMMC and /home and /opt Oct 21 19:59:51 ergo: bootloop Oct 21 20:00:13 well known effect when kernel modules don't match the kernel Oct 21 20:00:14 Ok, I'll try just to load it Oct 21 20:01:22 can someone with KP installed pastebin the content of his /etc/default/mount-opts? Oct 21 20:01:26 btw I'd just patch that function to always return state "closed" rather than edit all the ölocations where it gets called Oct 21 20:03:25 Psotnick: and I'm afraid there might as well be some IRQ handler that deals with state changes of GPIO160 on a hardware level Oct 21 20:04:20 since it doesn't make sense when the kernel module would _poll_ mmc_slot(host).get_cover_state 5 times a second, to catch wehn somebody dares to open battery cover Oct 21 20:05:07 I modprobed that module and it loaded, but still can't see /dev/mmcblk1 Oct 21 20:06:18 I think I'll just give up Oct 21 20:20:46 aww :( Oct 21 20:20:52 try the gpio thing now! Oct 21 20:30:42 hacking kernel modules isn't exactly trivial, mainly due to the poor debugging opportunities Oct 21 20:31:14 tbh I doubt my idea about GPIO reprogramming will fly Oct 21 20:31:50 DocScrutinizer05: sure, with *that* attitude... Oct 21 20:31:57 lol Oct 21 20:32:56 find each reference to GPIO160 in that kernel module, weed out the result of an changing input at lowest possible level Oct 21 20:33:30 as mentioned there might as well be a IRQ handler waiting for an IRQ triggered from that GPIO Oct 21 20:41:58 dafaq! my t900 is definitely messed up Oct 21 20:42:25 which gpio? Oct 21 20:45:00 DocScrutinizer05: D: Oct 21 20:45:11 the billion-year-uptime one? Oct 21 20:45:13 rootfs filled up I guess Oct 21 20:45:18 yep Oct 21 20:45:26 vi__: 160 Oct 21 20:45:27 this shit right here is the reason you don't install sysklogd! Oct 21 20:45:50 DocScrutinizer05: are you going to reflash it? Oct 21 20:47:26 kerio: I'm going to reboot it :-O Oct 21 20:47:37 DocScrutinizer05: nooooooooooooooooo Oct 21 20:47:42 don't do it Oct 21 20:47:47 think of the children! Oct 21 20:47:48 t900:~# uptime Oct 21 20:47:50 23:00:07 up 204 days, 23:55, load average: 0.00, 0.02, 0.00 Oct 21 20:47:54 WAAAAAAAH! Oct 21 20:47:54 fix it! Oct 21 20:48:01 no more mounts Oct 21 20:48:07 what's fucked about it? Oct 21 20:48:13 what does dmesg say? Oct 21 20:48:20 dmesg is ok Oct 21 20:48:24 syslog isn't Oct 21 20:48:28 mount isn't Oct 21 20:48:31 stop it then Oct 21 20:48:33 and clean the rootfs Oct 21 20:48:37 cat /etc7mtab isn't Oct 21 20:48:41 maybe mount can't write to /etc/mtab Oct 21 20:48:47 and mtab can be restored from /proc/mounts Oct 21 20:48:53 hmm Oct 21 20:49:22 t900:~# cat /proc/mounts >/etc/mtab Oct 21 20:49:23 -sh: cannot create /etc/mtab: No space left on device Oct 21 20:49:29 indeed Oct 21 20:49:37 just rm /var/log/syslog Oct 21 20:49:42 you don't need it :P Oct 21 20:49:52 -rw-r--r-- 1 root root 29316716 Oct 14 14:02 syslog Oct 21 20:49:54 t900:~# rm /var/log/syslog Oct 21 20:49:57 already did Oct 21 20:50:03 sync Oct 21 20:50:07 oooh Oct 21 20:50:10 ubifs is weird like that Oct 21 20:50:25 t900:~# sync Oct 21 20:50:26 t900:~# cat /proc/mounts >/etc/mtab Oct 21 20:50:28 -sh: cannot create /etc/mtab: No space left on device Oct 21 20:50:35 what does df say? Oct 21 20:50:50 t900:~# df -h Oct 21 20:50:51 Filesystem Size Used Available Use% Mounted on Oct 21 20:50:56 SIC! Oct 21 20:51:02 nice Oct 21 20:51:06 df / Oct 21 20:51:14 there must be a way to force df to work Oct 21 20:51:25 df: /: can't find mount point Oct 21 20:51:31 oh lawds Oct 21 20:51:36 can you make a symlink? Oct 21 20:51:41 o.O Oct 21 20:51:42 ln -s /proc/mounts /etc/mtab Oct 21 20:51:52 I doubt Oct 21 20:52:03 have you stopped sysklogd and klogd, btw? Oct 21 20:52:11 t900:~# ln -s /proc/mounts /etc/mtab Oct 21 20:52:12 ln: /etc/mtab: File exists Oct 21 20:52:21 ok then, delete it Oct 21 20:52:35 if syslog is still open then maybe it's keeping /var/log/syslog alive Oct 21 20:52:41 t900:~# rm /etc/mtab Oct 21 20:52:42 t900:~# ln -s /proc/mounts /etc/mtab Oct 21 20:52:44 ln: /etc/mtab: No space left on device Oct 21 20:52:50 dammit! Oct 21 20:53:11 DocScrutinizer05: what have you done? D: Oct 21 20:53:16 lol Oct 21 20:53:24 killed my best uptime Oct 21 20:53:28 srsly tho Oct 21 20:53:30 it's fixable Oct 21 20:53:47 syslog ended Sep 29 Oct 21 20:53:57 DocScrutinizer05: mount a new /etc via nfs Oct 21 20:54:07 o.O Oct 21 20:54:18 IT CAN'T GO WRONG Oct 21 20:54:20 I'll just reboot Oct 21 20:54:51 well that's just weak :S Oct 21 20:55:16 i wonder if it'll boot Oct 21 20:56:00 we'll see Oct 21 20:58:17 booting to ACT_DEAD? Friggin KP46 Oct 21 20:59:51 hahaha Oct 21 21:00:12 t900:~# uptime Oct 21 21:00:13 23:12:29 up 1 min, load average: 5.03, 1.56, 0.54 Oct 21 21:00:17 DocScrutinizer05: i sometimes get that, but only after doing something with charge*.sh Oct 21 21:00:28 t900:~# df -h / Oct 21 21:00:30 Filesystem Size Used Available Use% Mounted on Oct 21 21:00:30 (in backupmenu, mostly) Oct 21 21:00:31 rootfs 227.9M 220.1M 3.6M 98% / Oct 21 21:00:33 :-o Oct 21 21:00:39 DocScrutinizer05: see Oct 21 21:00:50 * RST38h always suspected Maemo and possums had something in common Oct 21 21:01:05 -rw-r--r-- 1 root root 29316716 Oct 14 14:02 syslog Oct 21 21:01:05 (aside from being terminally cute, I mean) Oct 21 21:01:14 well, du -mx | sort -n Oct 21 21:02:16 Oh lol, 80 friggin new mails Oct 21 21:02:37 wait, modest caused this? Oct 21 21:05:18 kerio: are you surprised or what? Oct 21 21:05:40 220mb used? hell yeah i'm surprised Oct 21 21:05:51 DocScrutinizer05: isn't 228mb a bit small anyway? Oct 21 21:05:54 how old is this n900? Oct 21 21:06:28 hmm? Oct 21 21:06:47 Doc: So, did you have a chance to look at the log before delering it? Oct 21 21:06:56 Doc: Who shat that much? Oct 21 21:07:00 yeah, ended Sep 29 Oct 21 21:07:24 DocScrutinizer05: i thought the rootfs was bigger than that Oct 21 21:07:31 never been Oct 21 21:07:34 DocScrutinizer05: ubifs avoids bad blocks, though Oct 21 21:08:27 the real mess: there's no single huge file to delete Oct 21 21:08:45 DocScrutinizer05: install cssu-thumb Oct 21 21:08:47 :D Oct 21 21:08:52 though I wonder where's the 29MB eaten by now deleted syslog gone Oct 21 21:09:50 29316716 Oct 14 14:02 syslog Oct 21 21:10:19 I wonder if it's still eating blocks or sth Oct 21 21:10:31 Ah, this part is simple Oct 21 21:10:56 Doc: Maemo rootfs reports full file sizes but compresses them in flash Oct 21 21:11:15 Doc: So, when you delete a 29MB text file, the freed space may well only account for 3MB Oct 21 21:11:25 oh yeah, and df -h reports the on-nand sizes Oct 21 21:11:43 This behaviour is only true for / though, MMC-based stuff measures normally Oct 21 21:12:28 RST38h: thought as much, but compression ratio 10:1 ? Oct 21 21:13:02 DocScrutinizer05: it's not unheard of Oct 21 21:13:11 especially for something with very little entropy like a log Oct 21 21:13:15 du vs du --apparent-size? :) Oct 21 21:13:28 * ShadowJK guesses busybox du doesn't have it Oct 21 21:13:51 ShadowJK: bingo Oct 21 21:14:28 DocScrutinizer05: gnu du has it Oct 21 21:16:20 http://pastebin.com/3ZtdGpV0 :-/ Oct 21 21:17:36 WTF /usr/share/nokia-maps Oct 21 21:19:12 i've got a 1mb "rdat" file there Oct 21 21:19:32 ubifs \o/ Oct 21 21:19:36 Filesystem Size Used Available Use% Mounted on Oct 21 21:19:37 rootfs 227.9M 188.4M 35.3M 84% / Oct 21 21:19:40 \o/ Oct 21 21:20:06 F U ubifs Oct 21 21:20:52 I actually did NUTTIN to regain that space - no boot, no deleting any file Oct 21 21:21:26 DocScrutinizer05: did you sync? Oct 21 21:21:33 just moved from 3.6M 98% / to 35.3M 84% / Oct 21 21:21:35 nope Oct 21 21:21:44 DocScrutinizer, reboot to make ubifs recount Oct 21 21:21:45 i mean, did you sync after deleting syslog? Oct 21 21:22:00 yep, but that been before boot Oct 21 21:22:06 hm, weird Oct 21 21:22:11 oh well, must've autorepaired something Oct 21 21:22:14 3.6M 98% / to 35.3M 84% / been *after* reboot Oct 21 21:22:20 hooray for ubifs Oct 21 21:22:28 ubifs \o/ Oct 21 21:22:53 I think the reboot thing is a remnant from jffs2 days. To mount jffs2, the entire filesystem is scanned, and all metadata is read into ram, where it keeps hogging ram. Oct 21 21:23:41 One of the big points about ubifs was that mounting it is fast, so it probably resumes after a reboot from the same state it had before reboot. Oct 21 21:23:43 ShadowJK: but how do you explain free space 3.6M 98% / to 35.3M 84% /, *without* any reboot or anything in between? Oct 21 21:24:03 It probably has garbage collection running when idle :-) Oct 21 21:24:07 DocScrutinizer05: the free space is an estimate Oct 21 21:24:08 yeah Oct 21 21:24:23 Doc: 10:1 is not unusual, not for syslog anyway Oct 21 21:24:24 something like fsckng garbage collection Oct 21 21:24:37 Doc: You have got a text file with a dictionary of what? 200 words? Oct 21 21:24:50 RST38h: quite possible, yes Oct 21 21:24:57 On my sheevaplug where I have a nilfs2 partition, the delay between deleting files and free space appearing can be hours :-) Oct 21 21:25:07 nevertheless see new miracle Oct 21 21:25:26 DocScrutinizer05: http://www.linux-mtd.infradead.org/doc/ubifs.html#L_spaceacc Oct 21 21:25:35 Doc: The short answer is "ubifs is weird" Oct 21 21:25:45 DocScrutinizer05: especially "Dirty space" Oct 21 21:29:30 :nod: >>Another possibility would be to implement background GC in UBIFS (just like in JFFS2), which would lessen effect of dirty space with time.<< seems somebody did Oct 21 21:32:10 kerio: thanks for link Oct 21 21:32:16 very useful Oct 21 21:33:01 ubifs is *really* weird Oct 21 21:33:04 but also neat Oct 21 21:33:11 so, stack cleanup, back tp Psotnick Oct 21 21:34:16 DocScrutinizer05: is it safe to play around with gpio? Oct 21 21:34:57 basically yes Oct 21 21:35:05 though that depends Oct 21 21:35:33 in theory you could create a short Oct 21 21:35:47 in particular, is it safe to set pin 160 to write mode and turn it on? Oct 21 21:35:48 but I guess the chip won't care too much Oct 21 21:36:44 yes, should be safe, except maybe you create an output of high level to a sensor with open collector output shorting that to GND Oct 21 21:37:33 the result might not help in the end, since reading back from outputs usually still represents real level seen at pin Oct 21 21:37:41 which probably still is low Oct 21 21:37:50 Oct 21 23:46:26 t900 kernel: [ 2222.749023] mmc0: cover is open, card is now inaccessible Oct 21 21:37:51 Oct 21 23:46:26 t900 kernel: [ 2223.059051] mmc0: card e624 removed Oct 21 21:38:04 first thing in syslog on opening back cover Oct 21 21:39:27 so no msg about [GPIO 160] Oct 21 21:40:25 oooh tracker STFU! Oct 21 21:40:50 Oct 21 23:53:09 t900 tracker-indexer[1655]: GLIB WARNING ** Tracker - No data was allocated for row 0, this means the database contained NULL or non-managed values for all the data in it Oct 21 21:40:58 * 10000 Oct 21 21:42:42 here's what I hoped to see for gpio160 as well: Oct 21 23:54:35 t900 kernel: [ 2711.715667] slide (GPIO 71) is now open Oct 21 21:44:30 anyway if you wanna patch kernel module, search for "mmc0: cover is open, card is now inaccessible" Oct 21 21:45:08 NB the mmc number = 0 Oct 21 21:45:28 for kernel it's still mmc0 on uSD and mmc1 on eMMC Oct 21 21:46:15 rationale: the romboot will check on mmc0 for xloader, but not on mmc1 Oct 21 21:46:34 aaaah, the old insta-dismount mmc on door open problem. Oct 21 21:46:40 so it's kinda logical to have a removable media on mmc0 Oct 21 21:46:53 vi__: yep Oct 21 21:47:01 You would think they would at lest make it unmount gracefully. Oct 21 21:47:06 Psotnick has problem with his hall sensor broken Oct 21 21:47:29 some people stick a tiny magnet inside the case. Oct 21 21:47:39 his sensor is defect Oct 21 21:48:11 that sucks, it is not like yu can just chmod 444 a /sys entry for the backdoor sensor. Oct 21 21:48:18 which is a tad unfortunate since you can't use a perfectly working uSD slot anymore Oct 21 21:48:25 it works at an annoyingly low level. Oct 21 21:49:01 yup Oct 21 21:49:17 remove hall sensor component and short the pads? Oct 21 21:49:27 do NOT short the pads Oct 21 21:49:37 or leave open as appropriate. Oct 21 21:49:53 no? oh well. Oct 21 21:49:54 errr Oct 21 21:50:23 well I dunno, my logic says it should usually not consume power via the pullup resistor Oct 21 21:50:41 so magnet-detected should _open_ the 'switch' Oct 21 21:51:17 as I said 'or leave open as appropriate.' Oct 21 21:51:22 yeah Oct 21 21:51:56 seems Psotnick isn't that eager to do hw meddling Oct 21 21:51:57 I have not look at the schematic so I have NFC how the hall sensor is hooked up. Oct 21 21:52:13 DocScrutinizer05: then his n900 is fucked. Oct 21 21:52:38 the plan been to patch the kernel driver to ignore the sensor Oct 21 21:53:00 I have wanted that for 1+ year. Oct 21 21:53:08 even better than that... Oct 21 21:53:12 a great idea actually Oct 21 21:53:40 allow kernel parameter to enable or disable the sensor Oct 21 21:54:26 DocScrutinizer05: if it should be off when the backcover is closed, it should be fairly easy, right? Oct 21 21:54:36 right Oct 21 21:54:37 just physically remove the sensor Oct 21 21:54:38 make it check for some setting that either obeys sensor, plays loud noise+vibrates on actuation by sensor or just outright ignore it. Oct 21 21:55:51 well, there's still that ESD/EMI-filter Oct 21 21:56:08 so I got my sip to work. Oct 21 21:56:13 it has two tranzorbs/Zener to GND, and a series resistor Oct 21 21:56:40 I have even hooked up an old e65 to be the house sip phone. Oct 21 21:57:03 final piece was to setup mrs vi__ with sip on her 5800. Oct 21 21:57:12 It has the sip stack. Oct 21 21:57:16 kerio: which means you can't create that bad a short when plaing with GPIO Oct 21 21:57:32 It has settings to define a sip provider. Oct 21 21:57:55 it has no function to create a sip call. Oct 21 21:58:13 vi__: Oct 21 21:58:14 Nokia REMOVED the sip call function on purpose. Oct 21 21:58:15 Oct 22 00:04:08 t900 telepathy-sofiasip[1449]: nta_agent: received garbage from udp/217.10.79.9:54680/sip Oct 21 21:58:17 Oct 22 00:04:36 t900 last message repeated 2 times Oct 21 21:58:37 vi__: sipgate Oct 21 21:58:43 keepalive Oct 21 21:59:28 what are you trying to tell me? Oct 21 21:59:50 just telling you sipgate keep alive mechanism is flooding my syslog Oct 21 22:00:03 aaah Oct 21 22:00:22 you can set the keepalive method/timeout though. Oct 21 22:01:01 which I probably should do, but afaik you can't extend keep-alive interval as suggested by peer, only shorten it Oct 21 22:01:04 is battery use with sip terrible? Oct 21 22:01:12 idle Oct 21 22:01:21 not that bad Oct 21 22:01:39 I seem to get a dayish idle Oct 21 22:01:58 on WLAN I get way more Oct 21 22:02:06 vi__, at one point in time, Nokia considered SIP an "Enterprise" feature, and it was only enabled in their E-series phones. Oct 21 22:02:06 on Alan Oct 21 22:02:24 on wlan Oct 21 22:02:50 I have left the e65 connected to wifi with sip for 27 hours now. it has not lost any 'bars' of battery from the meter. Oct 21 22:03:16 ShadowJK: it costs them nothing to include an executable. Oct 21 22:03:26 indeed Oct 21 22:03:27 what the fuck where they thinking? Oct 21 22:03:41 They were thinking some people would pay extra for e-series to get the enterprisey features Oct 21 22:04:18 I had forgotten how much of a pain in the testes symbian is. Oct 21 22:04:34 n900 for life. Oct 21 22:04:51 YOH! Oct 21 22:05:39 I did't choose the unix life, the unix life chose me. Oct 21 22:05:56 I fell for it Oct 21 22:06:08 in the late 70s Oct 21 22:06:33 My Nokia E75, for example, had a non-e-series equivalent. The E75 is pretty cumbersome for playing music or multimedia. The non-E E75 has volume controls and music controls on the hw keys. E75 on the other hand, has SIP, vpn, provisioning, office suite, and less plastic. Oct 21 22:07:49 now we got N9 and hooker ;-P Oct 21 22:07:55 all nice software features that should be rolled out on all symbian devices as standard. Oct 21 22:08:01 yes Oct 21 22:08:10 n9 is a stinker. Oct 21 22:08:23 hooker is a bitch ;-P Oct 21 22:08:52 I must get another n900...or 2. Oct 21 22:08:55 still hard to tell them apart Oct 21 22:09:12 what n9 and whores? Oct 21 22:09:14 vi__: hurry before I snatch them all away Oct 21 22:09:36 spanish whore Oct 21 22:09:41 lumia Oct 21 22:09:50 * ShadowJK has 78.1M free on / Oct 21 22:09:53 It is just so risky to buy them 2nd hand. Oct 21 22:10:07 hm, I've gotten 2 good ones so far Oct 21 22:10:14 that's why I buy those "new2 ones Oct 21 22:10:17 half of the n900s on ebay are fucked. Oct 21 22:10:43 used ones a kinda ... so used Oct 21 22:10:50 even ones with broken usb sell for a nice price. Oct 21 22:11:14 * DocScrutinizer05 should get #5 and #6 Oct 21 22:11:25 although I did see a brand new one, still with scrreen foil for only £190. Oct 21 22:11:28 I was looking recently Oct 21 22:11:48 SpeedEvil: uk? Oct 21 22:11:51 I missed one. eBay app notification needs to be loader Oct 21 22:11:56 yeah Oct 21 22:12:00 Seems declared-to-work scratch-free N900s are going for 100€ here Oct 21 22:12:10 50 quid for no USB Oct 21 22:12:12 !!!! Oct 21 22:12:23 ShadowJK: where do you liive? Oct 21 22:12:27 germany? Oct 21 22:12:31 Finland Oct 21 22:12:34 ah Oct 21 22:12:59 n900 not black metal enough? Oct 21 22:13:14 hm? Oct 21 22:14:45 so who's going to finally fix that backcover hallsensor idiocy for good? patch mmc kernel module Oct 21 22:15:50 DocScrutinizer05: you? Oct 21 22:15:53 sounds like a job for KP Oct 21 22:16:12 paaaaaaaaali! Oct 21 22:16:38 March 2012, bought my current devN900, for 120€ Oct 21 22:16:45 Everything working on it, scratchfree screen Oct 21 22:16:59 trying to install moc. ./configure says that "no acceptable C compiler found in $PATH" . I have installed gcc-4.6 but it didn't help Oct 21 22:18:52 try build-essential Oct 21 22:19:30 http://www.ebay.co.uk/itm/Nokia-N900-Not-working-/150927609849?pt=UK_Mobile_Phones&hash=item2323fc8ff9 Oct 21 22:19:37 worth a shot? Oct 21 22:20:01 ShadowJK, package build-essential has no installation candidate Oct 21 22:20:18 vi__, battery is fucked, usb port us fucked, or entire thing is fucked Oct 21 22:20:49 11€, go for it Oct 21 22:20:55 akls: trying to build on target? will give a much PITA Oct 21 22:20:57 uh, £ Oct 21 22:21:19 n900 is expensive here. Oct 21 22:21:44 DocScrutinizer05, somebody said here that he was able to install moc Oct 21 22:21:50 so I'm trying to do the same Oct 21 22:21:56 Jade it was Oct 21 22:22:07 vi__: get it from the continent then, you islander Oct 21 22:22:10 Oct 21 22:22:19 yes Oct 21 22:22:33 Jade how did you compile moc? Oct 21 22:22:41 i forgot what deps i needed Oct 21 22:22:43 it was a while ago Oct 21 22:22:51 Jade for now I'm getting error "no acceptable C compiler found in $PATH" Oct 21 22:22:55 I guess you used SB nevertheless Oct 21 22:23:15 akls, add the nokia repo with dev tools then you can install gcc, binutils and other stuff Oct 21 22:23:19 hold on Oct 21 22:23:27 o.O Oct 21 22:23:43 you built a huge project on N900 ? Oct 21 22:23:47 * ShadowJK used to do that in Maemo4 days Oct 21 22:23:48 http://tablets-dev.nokia.com/eula/index.php Oct 21 22:23:52 then add "deb http://repository.maemo.org/ fremantle/ nokia-binaries Oct 21 22:23:52 " Oct 21 22:23:54 in sources.list Oct 21 22:24:04 it should have all the deps you need for moc, if my memory serves me right Oct 21 22:24:26 oh Oct 21 22:24:27 just needs gcc, binutils and some other misc stuff i forgot now Oct 21 22:24:28 Jade: on N900?? Oct 21 22:24:29 you know what would be good? Oct 21 22:24:30 yeah Oct 21 22:24:35 moc isnt too heavy to compile directly on n900 Oct 21 22:24:38 ok! thanks Oct 21 22:24:40 status menu in task switcher. Oct 21 22:25:10 vi__: really? why? Oct 21 22:25:48 I admit I found it a little bit annoying sometimes to have to switch back to desktop Oct 21 22:26:06 but usually my taskswitcher is opened for <3s Oct 21 22:26:41 vi__: I find my n900 perfectly black metal enough to host my Horna and Behexen albums. Can't still figure out that fscking modifier key thingy though. Oct 21 22:27:01 Jade, what is scratchbox? Oct 21 22:27:29 akls: the crosscompiling environment on PC Oct 21 22:27:54 sounds like something I don't really need Oct 21 22:27:56 or what.. Oct 21 22:28:09 yeah, nobody likes it Oct 21 22:28:42 but it's the only true development environment for maemo Oct 21 22:29:50 Skry: but sweden is home of melodict death metal, not black metal! Oct 21 22:29:54 while for on-device compiling you sooner or later run into obstacles like missing free space on rootfs and non-optified c libs etc Oct 21 22:30:05 did you try changing rx-51? Oct 21 22:30:54 sweden is the home of göteborg metal Oct 21 22:30:59 boing boing boing Oct 21 22:31:01 is it ok that my /etc/apt/sources.list is empty? Oct 21 22:31:02 akls: there are several approaches to overcome those obstacles, like optification, chroot, bindmounts, whatelse Oct 21 22:31:32 * ShadowJK remembers having gcc in /opt/ on his PC Oct 21 22:31:58 skope: dont you mean gotheburg? Oct 21 22:32:16 yeah, dismember rule like fuck! Oct 21 22:32:37 english: gothenburg swedish: göteborg Oct 21 22:32:41 ^gothenburg? Oct 21 22:33:01 i am but a swednoob Oct 21 22:33:02 Goth e Burg Oct 21 22:33:31 Many swedish places and provinces sound much cooler in english/latin than in swedish Oct 21 22:33:34 Like "Gothia". Oct 21 22:34:41 Random weird fact: "The Swedish Empire" is included in wikipedia's list of longest lasting empires. Oct 21 22:34:51 (ended in 19th century) Oct 21 22:34:53 isn't Goth something different than Metal? Oct 21 22:34:56 is it ok that my /etc/apt/sources.list is empty? Oct 21 22:35:00 vi__: göteborg is the scandinavian way of saying it ;) Oct 21 22:35:24 akls, that sounds normal, there's a sources.d directory probably? Oct 21 22:35:47 akls: yes Oct 21 22:36:46 ls -l /etc/apt/sources* Oct 21 22:37:15 vi__: did edit it, got nothing useful done. configured ctrl level 3 as super, and tried to modifier_map it but it still shows up as control in modifier map though xev reports it to be super. Unholy. Oct 21 22:37:57 /etc/apt/sources.list.d/hildon-application-manager.list 317 Oct 21 22:39:27 Skry: there is another possibility. Oct 21 22:39:48 you could maybe use xbindkeys and xdotool to send super. Oct 21 22:39:59 just like right click hack. Oct 21 22:41:19 Jade: why don't you share your moc binary? Oct 21 22:42:13 Jade: is the malware part not yet finalized? ;-D Oct 21 22:43:52 hold on just making sure the rootkit actually works on 2.6.28 Oct 21 22:43:54 then ill upload Oct 21 22:44:11 lol wtf Oct 21 22:44:23 so I have added a source from here http://tablets-dev.nokia.com/eula/index.php to /etc/apt/source.list but I get same error on "apt-get install gcc" or build-essential - no installation candidate Oct 21 22:44:27 and yes, I did apt-get update Oct 21 22:45:59 holy f Oct 21 22:46:11 have you guys seen tmo? Oct 21 22:46:28 speciffically apkenv? Oct 21 22:46:47 they said alien dalvik was a myth! Oct 21 22:50:24 Jade :\ Oct 21 22:52:22 what is alien dalvik? Oct 21 23:01:59 then add "deb http://repository.maemo.org/ fremantle/ nokia-binaries Oct 21 23:02:04 maybe you meant deb http://repository.maemo.org fremantle/sdk free non-free ? Oct 21 23:35:17 anytime I try to compile something myself.. Oct 21 23:35:30 I'm thinking that shooting my foot is going to be less painful Oct 21 23:35:53 shooting foot vs compiling on-device? certainly Oct 21 23:36:13 Especially if it's something with more dependencies than libc Oct 21 23:36:19 currently I'm getting this error: configure: no Berkeley DB version 4 or higher found configure: error: BerkeleyDB (libdb) not found Oct 21 23:36:44 and I have installed libdb1 libdb4.2 libdb4.2++c2 Oct 21 23:38:08 you want the -devel varieties for compiling Oct 21 23:38:16 or -dev Oct 21 23:38:24 yeah Oct 21 23:38:27 but... Oct 21 23:39:27 but there's no libdb-dev Oct 21 23:39:31 or db-dev Oct 21 23:40:10 should be db4-devel Oct 21 23:41:10 AHA Oct 21 23:41:32 libdb4.2-devel Oct 21 23:41:37 dev* Oct 21 23:41:47 it's in fremantle sdk Oct 21 23:47:39 ok Oct 21 23:47:45 I hope it was the last error Oct 22 00:52:41 I missed the drama on Talk last week. Oct 22 00:52:43 Too much fun. Oct 22 00:53:04 what drama? Oct 22 00:55:03 http://talk.maemo.org/showthread.php?p=1282399#post1282399 Oct 22 01:47:54 How can there be drama on a web forum Oct 22 01:48:51 In my mind, there's so much effort involved in getting a web forum to display, that once you manage it, there'd be no energy left for drama Oct 22 01:49:21 The effort part is why I don't follow tmo anymore **** ENDING LOGGING AT Mon Oct 22 03:00:00 2012