**** BEGIN LOGGING AT Wed Sep 14 02:59:57 2011 Sep 14 06:39:16 DocScrutinizer: infobot joined! Sep 14 06:51:00 oh ffs I fell into the trap of newbs resurrecting old threads Sep 14 07:33:58 hah Sep 14 08:26:42 DocScrutinizer: i'm looking at the possibility to make myself a cable for hostmode on the N900. http://wiki.openmoko.org/wiki/Specialized_USB_cables#How_to_make_a_USB_cable_for_powering_attached_USB_devices_and_charging_the_Neo_while_in_host_mode_.28aka_Y-Cable.29 Sep 14 08:27:00 yeah Sep 14 08:27:38 unfortunately the schematic isn't there anymore.. http://om.quatrox.org/om/ Sep 14 08:27:53 I think it's not really worth the effort, as simple good cheap Y-cables are available everywhere, and all it needs in addition to this is a F-F-adapter Sep 14 08:29:09 DocScrutinizer: can show a pic? Sep 14 08:29:09 i already have F-F, but Y-cables proving hard to find here :P Sep 14 08:29:18 will keep trying though Sep 14 08:35:13 external HDD usually come with a Y-cable, when they are 3.5" type Sep 14 08:35:54 most have a mini-B-USB plug, so you'll also need an adapter mini->micro Sep 14 08:37:18 interesting, reuters classifies news about political turmoil in China as "us-china-internet" Sep 14 08:39:16 Sicelo: modding a cheap USB hub is simple Sep 14 08:48:03 DocScrutinizer crashanddie Sicelo moo Sep 14 08:48:15 moo Sep 14 08:48:27 ltns Sep 14 08:53:05 ~cssu Sep 14 08:53:05 somebody said cssu was http://wiki.maemo.org/Community_SSU Sep 14 08:55:14 does anyone know who hacks/maintains the camUI (lazy to ask... I know) Sep 14 08:55:39 nope... Sep 14 08:57:00 wassa camui? Sep 14 08:57:07 can't find anything Sep 14 08:57:56 osso-camera's user interface :p Sep 14 08:58:06 ÁÒÒ Sep 14 08:58:10 ahhh Sep 14 08:58:13 IC Sep 14 08:59:15 I'm having an issue with the headset and/or connector. it works when playing music but when making a phone call it doesn't switch over to using the headset Sep 14 08:59:28 anyone experienced something similar? Sep 14 08:59:38 d'oh... and it's Nokia N900 Sep 14 08:59:54 and firmware? Sep 14 09:00:13 I use headphones without any trouble Sep 14 09:00:19 I used to as well Sep 14 09:00:20 same here Sep 14 09:00:22 but no longer. Sep 14 09:00:35 well normal headphone, not headset Sep 14 09:00:51 it could be an issue with the audio connector port. I think there were some t.m.o. threads about it when N900 can sometimes detect the plugged in device is either a headset or just a headphone Sep 14 09:00:54 that may make a difference (microphone on headset) maybe its misdetected Sep 14 09:00:57 mfridh: headset or earplugs? Sep 14 09:01:08 ah ok as both work for me Sep 14 09:01:18 the standard earplug headset which came with my N900 Sep 14 09:01:36 i have sometimes problems that it misdetects the thing plugged in when you plug it in too slowly Sep 14 09:01:39 it's a headset in terms of it having a mic Sep 14 09:01:56 rarely it things that it wants to activate tvout then which gives some ugly noise Sep 14 09:01:57 cehteh: when playing music it's a bit wobbly... it can miss the plug-in if I plug too fast Sep 14 09:02:11 Moo, gentlemen. What is cooking today? Sep 14 09:02:20 cehteh: +1 but then just the mic is not working and it uses the earphones for output and the bulk-mic for input Sep 14 09:02:30 mfridh, have you read this thread? http://talk.maemo.org/showthread.php?t=53399 Sep 14 09:02:36 RST38h: moo Sep 14 09:02:43 RST38h, moo Sep 14 09:02:46 yes thats what i do very often, just headphones and device mic Sep 14 09:02:54 psycho_oreos: no, will do. Sep 14 09:03:02 mfridh: we got it that it is the retail nokia-headset Sep 14 09:03:16 mfridh, that's one prime example, but I'm sure there's plenty others floating around Sep 14 09:03:47 curious to know, what are both of yours (mfridh && chem|st) hw revision? Sep 14 09:03:49 mfridh: one chance is dirty connectors, both onboard or plug itself Sep 14 09:04:21 psycho_oreos: I think mine was the second Sep 14 09:04:49 I really have to ease it in to get it working Sep 14 09:04:52 chem|st, I meant board revision :) i.e. cat /proc/cpuinfo Sep 14 09:05:05 mostly it fails to detect it (playing music) Sep 14 09:05:21 3 Sep 14 09:05:26 so second shipped Sep 14 09:05:39 gah, my phone crashed Sep 14 09:05:47 but I think I just glanced at "rev 3" from cpuinfpo Sep 14 09:05:50 ah the 2101 you mean Sep 14 09:06:07 chem|st, yes.. I was about to ignore you lol Sep 14 09:06:08 mfridh: he meant the board Sep 14 09:06:12 the bottom number Sep 14 09:06:18 yeah the last line.. bah Sep 14 09:06:27 i couldn't read that far before the phone crashed :) but i'm rebooting Sep 14 09:06:34 ouch... Sep 14 09:06:55 you could also see the info when you use flasher and when you enable R&D mode, at nolo screen Sep 14 09:07:27 seems like more issues with lower revision numbers.. dunno if mine suffers the same fate Sep 14 09:07:42 Revision 2104 Sep 14 09:07:44 psycho_oreos: what are the different revisions? as 2101 is the one from spring 2009 afair Sep 14 09:07:57 psycho_oreos: I mean the time frames Sep 14 09:08:26 chem|st, I wouldn't have a clue :) I'm just personally surveying hw revision numbers from various hw faults Sep 14 09:08:34 psycho_oreos: true Sep 14 09:08:48 both of my N900 here have hw revision 2204 Sep 14 09:09:18 dmesg said "headphone connected" Sep 14 09:09:26 still, it doesn't output in there when i call! Sep 14 09:09:29 would be nice to know which digit means what as there seriously were no 100 baord revisions Sep 14 09:09:30 seems like a soft error Sep 14 09:09:48 mfridh: do you see the icon? Sep 14 09:09:53 also I hardly use wired audio cable. Almost all of the time I use bluetooth headset Sep 14 09:11:06 chem|st: no icon in the activity field Sep 14 09:11:14 chem|st, heh only nokia has that info, I think someone asked about it on one t.m.o. thread (notably I think the thread was about the hw revision numbers that nokia had produced for N900). Unfortunately that question fell on deaf ears Sep 14 09:11:24 mfridh: then pulse is not switched to it Sep 14 09:12:33 and flasher would say how many hw revisions for N900, definitely not 100 or 50 either.. maybe 15 I'd say at most Sep 14 09:12:39 chem|st: anything I can do about it? Sep 14 09:12:57 perhaps it's further down this maemo thread... Sep 14 09:13:00 I know that >2000 is boards for wholesale Sep 14 09:13:08 mfridh: disconnect reconnect Sep 14 09:13:27 mfridh: but a crashing device within xterm... Sep 14 09:13:31 maybe clean the audio port with q-tip that's been lightly dabbed in metholated spirits? *shrugs* Sep 14 09:13:38 mfridh: what is your firmware version? Sep 14 09:13:56 there's a headphone icon when I connect a standard pair of earplugs Sep 14 09:14:14 mfridh: maybe the headset is borked Sep 14 09:14:20 and not your device Sep 14 09:14:34 have you tried another headset Sep 14 09:15:57 chem|st, you mean hw revision number above 2000? if so I think all N900 had hw version above 2000.. I also think that each `groups' of hw revision numbers are reserved for each NIT versions (e,g, rev 1101 maybe for N800).. hmm I should double-check this Sep 14 09:16:18 chem|st: the only headset I have around is an apple one, there's just alot of static when I connect it Sep 14 09:16:33 it detects connection in dmesg, but no icon Sep 14 09:16:59 the icon iinm is only shown if the detected audio device is actually a headset Sep 14 09:17:07 chem|st: Firmware Version 20.2010.36-2 Sep 14 09:17:44 ~iinm Sep 14 09:18:22 chem|st: it crashed while playing music via Media Player and doing cat /proc/cpuinfo in xterm Sep 14 09:18:30 regarding the crash earlier Sep 14 09:19:21 Sicelo, if I'm not mistaken Sep 14 09:19:29 thanks :) Sep 14 09:19:36 ~iirc Sep 14 09:19:36 well, iirc is "if I recall correctly" Sep 14 09:19:37 mfridh, interesting Sep 14 09:19:43 Sicelo, nw Sep 14 09:19:50 psycho_oreos: there's mention in that headphone thread you linked about a "white line" being cut inside the headphone outlet Sep 14 09:20:12 infobot, iinm is if I'm not mistaken Sep 14 09:20:12 okay, psycho_oreos Sep 14 09:20:17 ~iinm Sep 14 09:20:17 extra, extra, read all about it, iinm is if I'm not mistaken Sep 14 09:20:26 :) Sep 14 09:21:01 ... Well for one thing, the white line was cut when I opened it up. I think that is the only issue, and the reason is because the white line was given absolutely no give. So any tug or pull on the cable will result in the whiteline dislodging from its connection point... Sep 14 09:21:06 "It was exactly same for me - white line was cut. After little soldering job my headset is working like new :-) " Sep 14 09:21:18 http://talk.maemo.org/showpost.php?p=748212&postcount=36 Sep 14 09:21:24 wonder if that's it.... Sep 14 09:21:42 if that could make a standard headphone work but a headset not work Sep 14 09:21:53 headset (combined mic/headphone) Sep 14 09:22:11 * cehteh thinks the first thing he done when receiving the n900 was dumping the earplugs into the trashbin :P Sep 14 09:22:16 static when connecting earphones suggests a plug not fitting correctly. my original headset has static when i use it on laptop, but none on any others. of course, other earphones work fine on the laptop Sep 14 09:22:28 mfridh, I can't vouch for the correct-ness of tht thread :) but you did say before that you tried plugging in ipod/iphone audio cable (headset?), if that made lots of static noise but did produce an icon in N900 then that might not be the issue Sep 14 09:23:18 static could also mean poor connectivity :) Sep 14 09:24:11 yeah, actually meant it that way psycho_oreos :p Sep 14 09:24:27 psycho_oreos: no it did NOT produce an icon either Sep 14 09:24:39 cehteh: I got some sennheiser... Sep 14 09:24:55 so conclusion right now: headset is detected in dmesg but there's no icon. headphone plug works perfect. Sep 14 09:25:19 Sicelo, lol my bad Sep 14 09:25:37 mfridh, ahh ok, weird Sep 14 09:27:19 mfridh: try the headset on your ipod Sep 14 09:27:37 chem|st, my bad I counted all the hw revisions for N900 according to the info dumped from flasher when it came to reading stuff from the bin file. According to all those numbers, it adds up to 76 hw revisions.. realistically I do not know if those numbers are truly hw revision numbers as I have yet to hear owners with hw rev: 2610 (which appears to be the last set of number for hw rev I think). Sep 14 09:29:00 keen suggestion, hw-rev is nothing but the production series Sep 14 09:29:37 second digit is the series, last is the change in solder Sep 14 09:29:49 I'm guessing they have hw rev up to 2610 could be placeholder for N900 fixes going up to that number and I personally believe there would hardly be anyone would have anywhere near 2217 Sep 14 09:30:41 so in the case of 2101, series is 1 and that solder part is also 1? :) Sep 14 09:30:42 or another dome sheet... Sep 14 09:30:51 *shrugs* Sep 14 09:30:58 was there 20xx? Sep 14 09:31:18 they could be the pre-production series.. i.e. the test devices Sep 14 09:31:19 18 or 19 was the preproduction devices on summit Sep 14 09:32:31 o.O so there were 18xx and 19xx for N900? interesting.. they don't have their hw rev entries for the latest fw (area code 203, probably NMEA release) Sep 14 09:32:33 could be board revision 2 - series 2 for the first two digits Sep 14 09:32:41 ahh Sep 14 09:33:43 likely to use 4 digits for series tracking Sep 14 09:34:17 mayor series minor series Sep 14 09:35:24 speculations all over the place again Sep 14 09:35:48 you can thank nokia for that ;) Sep 14 09:36:36 leaving the community in a disarray of confusion, no further commitments to community, no detailed explanation of how their maemo series actually works, etc Sep 14 09:36:48 fyi: http://pastebin.com/9nJ6n2FM Sep 14 09:37:11 all those `valid for:' entries I'm assuming they are hw rev. Sep 14 09:38:46 that would mean they had some serious changes Sep 14 09:41:46 N800 fw info: http://pastebin.com/XSYBizC3; N810 fw info: http://pastebin.com/ADVpGBse Sep 14 09:43:32 770 never had hw rev it seems.. and with the information from N800, N810 combined.. there's a gap in between 18xx,19xx hw rev which nokia used on N900 at the summit.. interesting Sep 14 09:49:52 mfridh: when your headset's pushbutton is stuck or broken, it will short the 3rd contact of plug to 4th/sleeve always, thus the headset "looks like" a headphone. Phonecalls aren't supposed to route to headphones I guess (though I never tried it) Sep 14 09:50:59 DocScrutinizer: nope Sep 14 09:51:11 normal headphones work for calls Sep 14 09:51:27 but a shortcut might be one reason Sep 14 09:51:54 it gets detected but sys not changed to headset-available Sep 14 09:54:41 I recently acquired a cheap broken n900 and fixed it (new screen) all looks ok but it refuses to dial out *121# and similar numbers (my network balance query, credit top-up, etc use that style on number) anyone got any clues how to fix that? Sep 14 09:54:46 broken wire on mic (3rd pin) would lead to unpleasant effects as well, it either detects the hs as AV cable or not at all Sep 14 09:55:29 jpinx: reflash Sep 14 09:55:49 bummer :( Sep 14 09:56:47 err - starhash? Sep 14 09:57:16 SpeedEvil: don't... we like people to reflash^^ Sep 14 09:57:36 SpeedEvil: USSD supposed to work - even without *# - on PR1.2+ Sep 14 09:58:19 ah Sep 14 09:58:19 even in earlier versions you could activate it via config edit Sep 14 09:58:34 http://www.youtube.com/watch?v=eAoC818Mxy4&feature=feedu - on low power mono displays Sep 14 09:58:43 (like the metawatch) Sep 14 09:59:04 also -- what's the trick to stop auto-rotate of the phone application? Sep 14 09:59:42 it is an option in the phone app itself. Sep 14 09:59:52 to lock it to portrait Sep 14 09:59:56 hello, I get the basic theme.xml at /usr/share/themes/Default/matchbox/ , but cannot find a way to modify the left-corner button into a circle :D Sep 14 10:00:19 scoobertron: I want to lock to landscape ;) Sep 14 10:01:17 jpinx: don't reflash Sep 14 10:01:22 it wont help Sep 14 10:01:26 use Sep 14 10:02:05 jpinx: interesting option, then I am not sure. Assuming the turning control is a gconf setting, it may accept more values - i.e. portrait, landscape and off - but that is just speculation Sep 14 10:02:24 apt-get install ussd-widget Sep 14 10:02:24 well CSSU has an option to be able to turn it off Sep 14 10:03:14 ildar: stop advising obsolete BS please Sep 14 10:03:19 * jpinx ssh's into the device to see if gconf does anything Sep 14 10:03:56 DocScrutinizer: I say reflasing wont fix USSD Sep 14 10:04:10 I say stop caliming nonsense Sep 14 10:05:00 USSD is officially fixed since PR1.3 latest, iirc even since PR1.2 Sep 14 10:05:32 so yes, reflashing to PR1.0 for sure won't fix USSD Sep 14 10:05:59 RX-51_2009SE_20.2010.36-2_PR_COMBINED_MR0_ARM.bin . USSD doesn't work Sep 14 10:06:14 oops Sep 14 10:06:15 USSD works for me Sep 14 10:06:23 hey, it worked! Sep 14 10:06:29 first time! Sep 14 10:06:31 :) Sep 14 10:06:38 sorry for the mess Sep 14 10:06:45 jpinx: I got 3 options in dialer menu "screen rotation": auto | portrait | landscape Sep 14 10:06:49 DocScrutinizer: is there not an up/date/grade rather than reflash? Sep 14 10:07:15 not really recommended Sep 14 10:07:26 I think this n900 must have an old system Sep 14 10:07:30 jpinx: there are, but reflash is the best Sep 14 10:07:40 most stable etc. Sep 14 10:08:14 jpinx: ok, now I feel bad sending you in search of gconf keys - I havn't looked at that menu for so long I didn't realise it was in the menu Sep 14 10:08:15 it's been way too long since PR1.3 and you seem to have a pre-PR1.2 system. SSU Upgrade across several revisions doesn't usually work great Sep 14 10:09:31 ok DocScrutinizer - I hear you Sep 14 10:09:35 scoobertron: in PR1.2 (or pre-CSSU, dunno) it only has auto and portrait, no landscape option Sep 14 10:16:13 (I don't think though it got tweaked in CSSU, so probably is a new feature as of PR1.3) Sep 14 10:16:57 ( i can't check as I have no plain clean Pr1.3 system here) Sep 14 10:18:02 jpinx: you could check (and post here) your system's version as of settings->about_product Sep 14 10:18:43 the "Version: xxxxx" there Sep 14 10:20:07 st like 5.2012.42-3 Sep 14 10:20:44 Version: 3.2010.02-8.003 ;) Sep 14 10:20:57 :nod: Sep 14 10:21:49 looks like this one has never been maintained -- just used as a phone till the guy broke the screen Sep 14 10:21:59 I'm not quite sure if that's PR1.1.1 or PR1.2 Sep 14 10:22:25 and it's locale version 003 Sep 14 10:22:40 where's that -- I bought it in NZ Sep 14 10:22:44 I suggest you usually go for MR0 global version Sep 14 10:23:26 hmm Sep 14 10:23:27 ~flashing Sep 14 10:23:27 i guess maemo-flashing is http://wiki.maemo.org/Updating_the_tablet_firmware Sep 14 10:24:15 * jpinx has to go out for a few hours - maybe try reflashing tonight Sep 14 10:24:23 Trying to upgrade CSSU but it says that it depends on hildon-desktop 2.2.142-16 but *-13 is going to be installed...any idea how to trace the problem? Sep 14 10:24:56 My camera starting failing to start, and am trying to debug. camera-ui.launch fails immediately with a sigsegv, strace on camera-ui gets as far as reading from a socket, nothing obvious there. Sep 14 10:25:01 Any one got any suggestions Sep 14 10:25:41 zutesmog: did you install power-kernel? Sep 14 10:25:52 I am running power kernel but have been for 8 months or so without problems, Sep 14 10:26:13 I used the camera only a few days ago. Sep 14 10:26:23 zutesmog: the old bug was fixed, but I can not get the cam to work with power-kernel - I gave up ;) Sep 14 10:27:45 I have been using fcam and normal camera all the time, when I launch the normal camera, it comes the control display comes up, then disappears and a message "operation failed" Sep 14 10:27:50 nothing in dmesg Sep 14 10:28:45 I am not sure I am launching camera from the command line correctly, if I could work out how to do that, I think strace would help a lot. Sep 14 10:34:12 strace -f Sep 14 10:34:23 (followw forks) Sep 14 10:34:40 DocScrutinizer: yep tried that on camera-ui Sep 14 10:34:58 camera-ui is preloaded on boot time afaik Sep 14 10:35:11 no UI is shown and it's sitting reading on a domain socket Sep 14 10:35:27 you need to ps|grep camera, then attach strace -f -p Sep 14 10:35:45 if I watch porcess when I start the camera app, I see /usr/bin/camera-ui started Sep 14 10:36:08 but it doesn't hang around long enough ;-( Sep 14 10:36:15 IroN900:~# ps ax|grep camer Sep 14 10:36:17 1618 user 51416 S /usr/bin/camera-ui Sep 14 10:36:39 oops your right Sep 14 10:36:39 no camera "started" on that device, the process gets started on boottime Sep 14 10:36:41 will try that Sep 14 10:38:46 IroN900:~# strace -f -p `pidof camera-ui` Sep 14 10:38:47 Process 1688 attached with 3 threads - interrupt to quit Sep 14 10:38:49 [pid 1688] restart_syscall(<... resuming interrupted call ...> Sep 14 10:38:50 [pid 1649] restart_syscall(<... resuming interrupted call ...> Sep 14 10:39:25 DocScrutinizer: aha, that's a valid theory I guess, about the pushbutton being stuck or broken. Thanks, will look into it. Sep 14 10:39:36 yw Sep 14 10:39:43 come back reporting please Sep 14 10:40:10 sulx: apt-get or ham Sep 14 10:43:29 to disturb a little: i am using h-e-n with the latest booston script. connecting an externally powered hub (not for charging) i always get VBUS OVERVOLTAGE - You just fried your N900. at least the port is still fine :) any ideas? Sep 14 10:49:00 i am connecting the hub to N900 soon after i turn VBoost On Sep 14 10:52:45 Sicelo: this indicates your externally powered hub feeds back a somewhat too high voltage to the N900 USB port Sep 14 10:54:02 Sicelo: I had to look up in bq24150 datasheet what exactla VBUS overvoltage means, I guess it gets triggered on >6V or sth Sep 14 10:54:52 yay! thanks. good thing hub isn't mine. Sep 14 10:56:47 Sicelo: keep polling the booston file on cloud-7, I'm currently about to edit and upload new versions on a >once-per-day basis Sep 14 10:58:24 VBUS overvoltage and undervoltage thresholds vary some per chip. Sep 14 10:59:27 It's a pity in a way that undervoltage isn't reliably 3V or so, as you could run it from a dumb li-ion cell. But overvoltage indication does not mean it will have damaged the phone - there is a range of overvoltage which is 'safe' Sep 14 10:59:40 IIRC ~6.5V will not damage the chip. Sep 14 10:59:55 But it may display overvoltage at 5.8 Sep 14 11:00:16 But it's been a while since I looked at it, and of course, you may get lucky, and your chip is fine to 7.5 Sep 14 11:01:41 SpeedEvil: I think ABS MAX ratings are way higher than "recommended operating conditions" which VBUS-OV is based on Sep 14 11:02:00 ABS max is ~7 I think Sep 14 11:02:13 But it's been a while since I read it Sep 14 11:04:11 VBUS ABS MAX -0.3 .. 20 V Sep 14 11:04:44 (3) The bq24150/1 family can withstand up to 10.6 V continuously and 20 V for a maximum of 432 hours. Sep 14 11:05:23 ABS MAX for other chips' pins connected to VBUS is way lower though Sep 14 11:06:44 recomended operating conditins: VBUS Supply voltage, VBUS min:4 max:6 Volt Sep 14 11:07:31 (1) The inherent switching noise voltage spikes should not exceed the absolute maximum rating on either the BOOST or SW pins. A tight layout minimizes switching noise. Sep 14 11:11:23 VBUS(OVP) Overvoltage protection threshold for boost (VBUS pin) Threshold over VBUS to turn off converter during boost min:5.8 nominal:6 max:6.2 V Sep 14 11:12:18 seems my 6V nominal for VBUS-OV Error was correct Sep 14 11:21:26 DocScrutinizer: Ah Sep 14 11:21:47 Oooh. Sep 14 11:21:59 * SpeedEvil wonders if he's seen a n900 which got overvoltaged lots Sep 14 11:22:08 and then the USB port ripped off Sep 14 11:22:18 GAIA was not noticing vbus Sep 14 11:22:39 btw: V(OVP-IN) Input | VBUS OVP threshold voltage | Threshold over VBUS to turn off converter during charge: 6.3 6.5 6.7 V Sep 14 11:23:40 SpeedEvil: iirc GAIA vbus detect pin ABS MAX was something insane like 7V or sth Sep 14 11:23:42 yep Sep 14 11:24:43 I'm not going to open the twl4030 user manual now, not enough free RAM ;-P Sep 14 11:25:14 :) Sep 14 11:25:19 so I leave it up to you to check that friggin GAIA chip's ABS MAX for vbus detect pin Sep 14 11:28:47 :nod: N4200:R8 GAIA twl4030 VBUS pin directly connected to VBUS bus, no series R, no tranzorbs, no dedicated clamp diodes... nuttin Sep 14 11:30:47 DocScrutinizer: been looking at strace output for camera-ui. Every things looks to be ok (ie no obvious failed opens, etc...) the last interesting thing it does, query ioctl on /dev/video0, ioctl(25, VIDIOC_QUERYCAP or VT_OPENQRY, 0x228210) = 0 then creates a domain socket /var/tmp/sapwood-:0.0 does some reading and writing to a couple of sockets then shuts the whole group down. Sep 14 11:31:28 hmmm Sep 14 11:31:28 I do remember the last time I used the camera it was in video mode. Sep 14 11:32:49 any idea where it stores the state of the camera choice (video or still) ? I might try resetting it to still and see if it behaves differently. Sep 14 11:33:04 Though bless and fcam also both fail. Sep 14 11:33:21 unless the ioctl on the /dev/video is for the viewfinder. Sep 14 11:38:54 ok, something more useful. Tried mbarcode from the command line and got this Sep 14 11:38:55 GST error: v4l2camsrc_calls.c(694): gst_v4l2camsrc_open (): /GstPipeline:test-camera/GstV4l2CamSrc:camera_src: Sep 14 11:38:55 Capabilities: 0x0 Sep 14 11:38:55 Aborted Sep 14 11:40:13 and going back to the strace the ioctl on /dev/video0 was ioctl(25, VIDIOC_QUERYCAP or VT_OPENQRY, 0x228210) = 0 Sep 14 11:41:05 need to go hunting to see what valid ioctl call on /dev/video0 might return. I hope this doesn't mean my camera is fried ;-( Sep 14 11:43:15 I suggest reinstalling the matching kernel drivers for camera Sep 14 11:43:41 anyone know how to monitor dbus or something to find out if a process crashes starts? Sep 14 11:43:51 I need to start recovery-terminal when hildon-desktop crashes Sep 14 11:43:55 but not with script methods Sep 14 11:43:59 crashes starts? Sep 14 11:44:23 after hildon-desktop crashes 3 times, recovery-terminal would start Sep 14 11:44:25 MohammadAG: that's difficult, I mentioned that yesterday(?) Sep 14 11:44:38 DocScrutinizer: yeah I think I will, fcamera gives me FCam Error: VIDIOC_S_CTRL: Invalid argument Sep 14 11:44:39 FCam Error: V4L2Sensor: Can't acquire a frame when not streaming Sep 14 11:44:39 you did? Sep 14 11:44:47 yes, I did Sep 14 11:45:09 I said sth like "moh: we might need to patch dsmetool for that" Sep 14 11:46:51 MohammadAG: I'm sure dbus monitoring is pointless here Sep 14 11:47:10 need to check if this happens http://talk.maemo.org/showpost.php?p=1037567&postcount=8 Sep 14 11:47:13 h-d won't tell from grave "I died" Sep 14 11:47:47 I doubt dsmetool is telling anything except maybe "prepare for doom everybody, I'm going to reboot the system now" Sep 14 11:48:00 which obviously is a bit late ;-D Sep 14 11:48:31 DocScrutinizer, 16.8 has your suggestion Sep 14 11:48:38 ? Sep 14 11:48:40 dsmetool gives up rather than rebooting Sep 14 11:48:47 yeah, seen it Sep 14 11:49:10 but cranking up a recue-shell in this case it non-trivial - alas Sep 14 11:49:12 :-S Sep 14 11:49:42 but cranking up a rescue-shell in this case is non-trivial - alas Sep 14 11:51:17 MohammadAG: you could write a daemon that polls for h-d being alive every 5 seconds (maybe only when screen unlocked) and start the rescue-shell when h-d seems to be gone for 2 consecutive polls Sep 14 11:52:01 I actually see no better way (except from augmenting dsmetool) Sep 14 11:52:54 or... hmm... you could "poll" only on kbd-slider-opens Sep 14 11:53:42 you can tell about that even easily from dbus, then check if h-d is around and if it is just abort; else: crank up rescue-shell Sep 14 11:53:59 s/even /event / Sep 14 11:53:59 DocScrutinizer meant: you can tell about that event easily from dbus, then check if h-d is around and if it is just abort; else: crank up rescue-shell Sep 14 11:54:36 let me check if dbus sends something about hildon-desktop Sep 14 11:54:47 if it starts 3 times consecutively, obviously it's not staying up Sep 14 11:54:48 what shall it send? Sep 14 11:54:57 NameOwnerChanged Sep 14 11:55:05 see http://talk.maemo.org/showpost.php?p=1037567&postcount=8 Sep 14 11:56:11 hmm, fair enough, start rescue-shell poll on third start of h-d, and keep polling for err 10s? finish with quit when h-d still there after 10s, else start rescue-shell Sep 14 11:56:37 sounds sane Sep 14 11:56:42 needs a daemon Sep 14 11:56:51 anyone up for writing a C one? or should I use Qt? Sep 14 11:56:55 meh, dbus-scripting Sep 14 11:58:17 I'd love the idea of just starting the rescue-shell when there's no h-d on slide-open the hwkbd Sep 14 11:58:34 WAY simpler, WAY cleaner Sep 14 11:59:20 I guess I could tweak my playswoosh dbus-script to do that, in like 15min Sep 14 12:00:13 pidof h-d || start-rescue-shell Sep 14 12:00:18 done :-P Sep 14 12:00:41 one additional line in my playswoosh Sep 14 12:02:37 so s/15min/30s/ ;-D Sep 14 12:04:04 plus additional benefit: the rescue shell doesn't open inadvertedly, but rather gives me an opportunity to watch the plain desaster prior to covering everything under an xterm screen Sep 14 12:05:49 http://talk.maemo.org/showpost.php?p=897443&postcount=14 Sep 14 12:11:38 DocScrutinizer: went back to standard kernel (not reflashed) and still no go. Full reflash in order I think Sep 14 12:16:41 yes, a full reflash (combined, not eMMC/VANILLA) is a good thing to get clear situation regarding this. If you do a full backup of your rootfs+/opt via backupmenu prior to that, you can revert to current state anytime later Sep 14 12:17:36 you might need to flash matching kernel via flasher though, if the rootfs+opt backup is working on a custom kernel Sep 14 12:18:34 of course this is a PITA when there's no proper kernel image for flasher available, only HAM packages Sep 14 12:19:16 robbiethe1st: how's about adding nandread and nandwrite to backup / restore kernel partition? Sep 14 12:20:12 ooops nanddump & nandwrite Sep 14 12:26:32 Maby later. Before that, when I actually get around to it, I need to add kernel warnings and copy-over-modules Sep 14 12:30:33 so, my n900 boots, the web shortcuts on the desktop and status bar show up, but no widgets. I can open applications from the application choose window, but eventually it reboots. Sort of a slow boot looping. Sep 14 12:34:37 <_trine> does anyone have freeswitch running on their N900? Sep 14 12:35:47 http://www.engadget.com/2011/09/14/netgears-n900-dual-band-router-hits-900mbps-top-speed-pats-its/ Sep 14 12:35:55 that model is kinda annoying Sep 14 12:38:02 <_trine> netgear are annoying period Sep 14 13:03:06 robbiethe1st: anyway http://paste.debian.net/129969/ Sep 14 13:05:22 That's the kernel? Sep 14 13:05:40 flash_erase /dev/mtd1 && nandwrite -m Sep 14 13:05:57 should restore the partition (CAL here) Sep 14 13:08:38 dang, CAL is size 60000, not 20000 :-o Sep 14 13:09:00 hey guys :) is git available for the N900? Sep 14 13:09:54 robbiethe1st: http://paste.debian.net/129970/ Sep 14 13:10:43 MoonTiger: Never tried it, but it's in the repos Sep 14 13:10:59 really? doh thnx :) Sep 14 13:11:30 Extras-devel I think Sep 14 13:11:42 So maybe not stable Sep 14 13:12:02 ahhhhhhhh Sep 14 13:12:16 http://maemo.org/packages/ http://maemo.org/packages/search/?org_maemo_packages_search[1][property]=name&org_maemo_packages_search[1][constraint]=LIKE&org_maemo_packages_search[1][value]=git&org_maemo_packages_search[2][property]=title&org_maemo_packages_search[2][constraint]=LIKE&org_maemo_packages_search[2][value]=same Sep 14 13:12:23 i try to avoid that repo as it managed to bork my sip stack somehow before Sep 14 13:12:50 thnx DocScrutinizer Sep 14 13:12:54 extras-devel doesn't bork sip-stack, apt-get autoremove does though Sep 14 13:13:06 really? Sep 14 13:13:48 http://talk.maemo.org/showthread.php?t=70875&highlight=sip Sep 14 13:13:53 ive done autoremove a few times and it strill works Sep 14 13:14:55 ahhhhhhh i never uninstalled the voip thing tho Sep 14 13:15:04 ok cool for the info thnx Sep 14 13:16:47 on some systems (probably those that did upgrade to PR1.3 via SSU) those "following packages were automatically installed and are no longer required:", not so on systems that flashed to a fresh PR1.3 Sep 14 13:17:47 anyway, my recommendation to NEVER DO AUTOREMOVE seems sensible Sep 14 13:18:42 DocScrutinizer, i did a fresh install of 1.3 so that would explain ... also if i ever do a autoremove i look at the stuff its proposing to remove and if anything looks dodge (ie, not something that was just installed) i dont do it Sep 14 13:19:10 fair enough Sep 14 13:19:36 good advice tho for sure Sep 14 13:19:46 good to know to solution to losing sip too Sep 14 13:19:48 thnx for that Sep 14 13:21:11 there's a thanks button on TMO ;-) Sep 14 13:21:50 tmo? Sep 14 13:21:55 ~tmo Sep 14 13:21:56 i heard tmo is http://talk.maemo.org, or trolls, morons, oxes Sep 14 13:22:10 oxes? Sep 14 13:22:17 trolls, morons, oxymorons?:) Sep 14 13:22:30 ~factinfo tmo Sep 14 13:22:30 tmo -- created by Mece <~mece@dsl-tkubrasgw1-fe3fdc00-44.dhcp.inet.fi> at Mon Jun 14 16:09:21 2010 (456 days); last modified at Wed Sep 8 12:31:25 2010 by mohammadag51; it has been requested 8 times, last by DocScrutinizer, 35s ago. Sep 14 13:25:33 f**k! still get segfault running svn Sep 14 13:25:42 so fustrating Sep 14 13:26:08 i dnloaded the replacement debs but still no go Sep 14 13:26:11