**** BEGIN LOGGING AT Mon Oct 12 02:59:56 2009 Oct 12 03:03:26 might try putting the card into a camera an formatting it Oct 12 03:03:48 may be the simple solution Oct 12 06:49:14 Settings on shr-u is not saved:/ Oct 12 06:49:46 what's the behaviour if bluetooth is set to automatic? Oct 12 06:55:29 I could've swear there's stuff on shr that was fixed and came back, like tapping on the screen when it's blank will perform an action:/ Oct 12 06:57:06 jonkristian: read FSO_resources wikipage please Oct 12 06:57:14 I am on it Oct 12 06:57:24 right now in fact Oct 12 06:58:00 jonkristian: (tapping on blank screen) isn't it exactly the way it works on any other X-based system? Oct 12 06:58:51 jonkristian: it's just that you want some screen-locking program to be run prior to blanking. Oct 12 06:59:14 I c Oct 12 06:59:50 Well, It was a problem a long time ago with shr, and then it was fixed Oct 12 07:00:05 Probably my settings then Oct 12 07:00:08 jonkristian: resource policy settings can't be saved because FSO doesn't provide an API for that. Also policy is a special thing and shouldn't be generally used by end-users so it's intended too. Oct 12 07:03:56 fsoraw looks very useful Oct 12 07:13:04 jonkristian: blueman should be in the SHR feeds already. I wonder if runtime deps were finally met. Oct 12 07:17:37 PaulFertser: Cool! Oct 12 07:36:51 moo Oct 12 07:38:02 meow Oct 12 07:42:41 PaulFertser, good morning Oct 12 07:44:54 panicking: ! Oct 12 07:45:14 you are hiding from the and-fr chans Oct 12 07:45:23 Not sure Oct 12 07:45:31 I'm working on android rightnow Oct 12 07:45:53 :) Oct 12 07:45:57 I have just logged Oct 12 07:55:27 strange Oct 12 07:56:15 can someone confirm that blueman is in the shr-u feed as stated here: http://wiki.shr-project.org/trac/ticket/622 Oct 12 07:58:17 panicking: hey :D Oct 12 07:58:31 sorry for the delay Oct 12 07:58:44 jonkristian: probably it's in the mrmoku experimental feeds ;) Oct 12 07:59:12 ok Oct 12 07:59:14 panicking: sorry for everything you might not like about the process Oct 12 07:59:21 i think i found more missing deps Oct 12 07:59:28 for vagalume Oct 12 07:59:50 btw, The hdq goes in timeout Oct 12 08:00:04 and with my patch I just retain the oldvalue Oct 12 08:00:09 panicking: and without? Oct 12 08:00:28 without you can fill with timeout Oct 12 08:00:30 panicking: you told me you had some bogus strings from sysfs nodes. But i can't reproduce it here. Oct 12 08:00:46 -ETIMEOUT some fields Oct 12 08:01:10 There are some not terminated data Oct 12 08:01:39 It crashes android Oct 12 08:02:20 panicking: i remember you told that, yes. But i couldn't reproduce. Which field exactly gave you unterminated data? Oct 12 08:04:27 let me check Oct 12 08:07:47 jonkristian: (fsoraw) thanks :-) Oct 12 08:08:15 PaulFertser, I think that is an android issue that try to read attribute if the battery is not present or Oct 12 08:08:23 It's not detected by the system Oct 12 08:08:27 the big problem Oct 12 08:08:37 are the garbage value when you have -ETIMEOUT Oct 12 08:08:44 panicking: i actually tried reading attributes from sysfs exactly when it's not present Oct 12 08:08:45 on the battery read operation Oct 12 08:08:55 Technology Oct 12 08:08:57 ? Oct 12 08:08:58 field Oct 12 08:09:06 too? Oct 12 08:09:14 I read these files Oct 12 08:09:27 #define AC_ONLINE_PATH "/sys/class/power_supply/ac/online" Oct 12 08:09:27 #define USB_ONLINE_PATH "/sys/class/power_supply/usb/online" Oct 12 08:09:27 #define BATTERY_STATUS_PATH "/sys/class/power_supply/battery/status" Oct 12 08:09:27 #define BATTERY_HEALTH_PATH "/sys/class/power_supply/battery/health" Oct 12 08:09:27 #define BATTERY_PRESENT_PATH "/sys/class/power_supply/battery/present" Oct 12 08:09:28 #define BATTERY_CAPACITY_PATH "/sys/class/power_supply/battery/capacity" Oct 12 08:09:30 #define BATTERY_VOLTAGE_PATH "/sys/class/power_supply/battery/voltage_now" Oct 12 08:09:32 #define BATTERY_TEMPERATURE_PATH "/sys/class/power_supply/battery/temp" Oct 12 08:09:34 #define BATTERY_TECHNOLOGY_PATH "/sys/class/power_supply/battery/technology" Oct 12 08:11:55 panicking: please see http://pastebin.com/m1365671c Oct 12 08:11:58 panicking: you are aware of the bq27k-detail.sh script? Oct 12 08:12:52 panicking: might find it inspiring Oct 12 08:12:53 no Oct 12 08:13:34 panicking: it's to decode raw values from bq27k, gives all the info from sysfs + much more. Oct 12 08:13:47 panicking: but let's see what's the problem with sysfs first. Oct 12 08:16:13 panicking: i can do more tests to reproduce the problem, i'm ready to compile simple c programs on device. Oct 12 08:17:37 But i hope "cat" is convincing enough. Oct 12 08:18:44 cat /dev/urandom? Oct 12 08:21:08 ColdFyre: especially that, yes Oct 12 08:21:10 cat /dev/mem >/dev/mem Oct 12 08:21:35 It's simpy Oct 12 08:21:49 what happen Oct 12 08:23:58 icons seems to be gone in vagalume, also it doesn't want to stream anything, anybody got vagalume working? Oct 12 08:24:10 it's ETIME btw Oct 12 08:24:50 jonkristian: never ever Oct 12 08:25:18 joerg_42: that didn't sound promising at all:D Oct 12 08:25:49 I suppose but I don't find the reason Oct 12 08:26:00 ret = psy->get_property(psy, off, &value); Oct 12 08:26:00 if (ret < 0) { Oct 12 08:26:00 if (ret != -ENODEV) Oct 12 08:26:00 dev_err(dev, "driver failed to report `%s' property\n", Oct 12 08:26:00 attr->attr.name); Oct 12 08:26:01 return ret; Oct 12 08:26:03 } Oct 12 08:26:05 if (off == POWER_SUPPLY_PROP_STATUS) Oct 12 08:26:09 return sprintf(buf, "%s\n", status_text[value.intval]); Oct 12 08:26:11 else if (off == POWER_SUPPLY_PROP_HEALTH) Oct 12 08:26:12 jonkristian: well I rarely ever tried :-P Oct 12 08:26:13 return sprintf(buf, "%s\n", health_text[value.intval]); Oct 12 08:26:16 else if (off == POWER_SUPPLY_PROP_TECHNOLOGY) Oct 12 08:26:17 return sprintf(buf, "%s\n", technology_text[value.intval]); Oct 12 08:26:19 else if (off >= POWER_SUPPLY_PROP_MODEL_NAME) Oct 12 08:26:21 return sprintf(buf, "%s\n", value.strval); Oct 12 08:26:23 return sprintf(buf, "%d\n", value.intval) Oct 12 08:26:25 that sometimes ret is > 0 Oct 12 08:26:27 and the index on these array are out of the possible interval valuie Oct 12 08:26:37 joerg_42: yeah, i need something that streams Oct 12 08:26:38 the best think is to chech the interval Oct 12 08:26:42 panicking: beware to get kicked. use pastebin Oct 12 08:26:49 ok sorry Oct 12 08:27:26 panicking: he's not joking, he's got the rights finally ;) Oct 12 08:27:45 panicking: not to nag;) have serdar asked you about this? http://code.google.com/p/android-on-freerunner/issues/detail?id=31 Oct 12 08:28:08 uhm, please don't kick me for bein off-topic Oct 12 08:28:27 jonkristian: no, offtopic is not among the reasons to kick :) Oct 12 08:28:43 jonkristian: in fact i don't think anybody got ever kicked from here. Oct 12 08:28:56 panicking: do you mean you can't reproduce it too? Oct 12 08:29:41 panicking: where's this code you pasted from? Oct 12 08:30:42 jonkristian: why not use mplayer? Oct 12 08:31:35 PaulFertser: (rights) not really, as I'm even worse sometimes. but freenode will do autokick anyway Oct 12 08:31:36 PaulFertser: Well, i wanted scrobbling functionality, not only that, but streaming from last.fm or other service. Oct 12 08:33:41 power_supply_show_property Oct 12 08:33:46 function Oct 12 08:33:55 drivers/power/power_supply_sysfs.c Oct 12 08:34:12 hm Oct 12 08:35:22 PaulFertser: my dyndns account controls paste.dyndns.org Oct 12 08:35:43 panicking: yes but it looks correct and works here all right. Oct 12 08:36:49 lindi-: http://pastebin.ca/1614753 Oct 12 08:38:05 PaulFertser: odd. works here: Oct 12 08:38:05 $ echo test | nc paste.dyndns.org 1234 Oct 12 08:38:05 http://paste.debian.net/48839/ Oct 12 08:38:32 PaulFertser: maybe paste.debian.net was having downtime? Oct 12 08:39:09 lindi-: i've retried just right now Oct 12 08:40:11 panicking: probably it's android code that simply doesn't check what open() or read() returns? Oct 12 08:41:31 PaulFertser: can you retry now that I have strace running? Oct 12 08:41:42 What is returned from that thing are always string Oct 12 08:42:00 lindi-: just did Oct 12 08:42:00 but if for some reason that index are out of range (look at the code seems impossible) Oct 12 08:42:05 PaulFertser: ok. Oct 12 08:42:12 they return invalid data Oct 12 08:42:41 panicking: do you check what open() and read() return in your code? Oct 12 08:43:27 Is there a file that disappear Oct 12 08:43:27 ? Oct 12 08:44:35 The code check Oct 12 08:44:40 iff the file exitst Oct 12 08:44:44 and the return code of the read Oct 12 08:45:07 heh, is there really a coder out there not to check return code of *all* calls? ;-P Oct 12 08:45:44 PaulFertser: ah, stupid bug. can you retry now to make sure it works? Oct 12 08:45:57 (in alsa even close() may throw error ;-D ) Oct 12 08:46:58 lindi-: still doesn't :) Oct 12 08:49:49 PaulFertser: only thing different in the request is the username (= IP address) Oct 12 08:50:57 lindi-: well, it still doesn't work for me for whatever reason. Oct 12 08:50:58 aha, and one extra "'" Oct 12 08:51:31 your IP address was too short ;-) Oct 12 08:54:05 PaulFertser: final retry now? ;) Oct 12 08:54:15 lindi-: fixed :D Oct 12 08:54:38 I always found it higly annoying an IP of 192.169.002.001 doesn't work Oct 12 08:55:38 PaulFertser: ok, thanks for reporting Oct 12 09:02:25 lindi-: thanks for providing this awesome service :) Oct 12 09:03:29 panicking: here you go, a simple test program: http://paste.debian.net/48847/ Oct 12 09:04:32 panicking: in my test the battery wasn't there, then i plugged it in, then unplugged again, then plugged in. Oct 12 09:06:20 ok Oct 12 09:17:14 PaulFertser: no problem, now code an AJAX vnc viewer so that developers can share their desktop to others without them needing to install any extra software :-) Oct 12 09:18:00 lindi-: i know how i "love" ajax, don't you ;) Oct 12 09:19:09 lindi-: i don't think there's much demand for it anyway. SHR folks use screen -m on buildhost e.g. Oct 12 09:19:16 PaulFertser: heh. i just feel that's the next step in remote debugging a bug on somebody else's system or teaching somebody remotely how do something complicated Oct 12 09:19:29 PaulFertser: but screen -m requires that you grant the other party ssh access Oct 12 09:19:52 with paste.debian.net you don't need to trust the other party in any way Oct 12 09:20:36 lindi-: remote debugging is pita anyway. But i see what you're after. Probably it'd be enough to have "script" streaming live to somewhere? Oct 12 09:21:25 PaulFertser: maybe but then the other party would need to constantly refresh the web page? Oct 12 09:22:29 lindi-: the browser is not an ideal solution for all problems. Why not just run nc to somewhere and get all the bytes live? Oct 12 09:22:57 PaulFertser: the problem is that the other party might not have netcat :) Oct 12 09:23:38 PaulFertser, the get_property does\t acquire any lock Oct 12 09:23:57 is it possible that it is deschedule, return to the worker Oct 12 09:24:05 change the value after the if Oct 12 09:24:07 lindi-: if a person can't install netcat, he is probably too hard to remote debug with anyway. Oct 12 09:24:11 and the return a positive value Oct 12 09:24:20 PaulFertser: not necessarily. he just might not have root access Oct 12 09:24:42 panicking: you say you reproduce the issue almost everytime but now you're talking about weird races... Oct 12 09:24:53 PaulFertser: we might be dealing with electronics for example and the other party might not know much about free operating systems Oct 12 09:25:56 Ok I'm going to test on android Oct 12 09:26:42 panicking: (your race) if it returned control, it can't change *value anymore, i don't quite see what you envision. Oct 12 09:28:50 } Oct 12 09:28:50 /* power is actually going in or out... */ Oct 12 09:28:50 if (di->regs.flags < 0) Oct 12 09:28:50 return di->regs.flags; Oct 12 09:28:50 if (di->regs.flags & BQ27000_STATUS_CHGS) Oct 12 09:28:51 val->intval = POWER_SUPPLY_STATUS_CHARGING; Oct 12 09:28:53 else Oct 12 09:28:55 val->intval = POWER_SUPPLY_STATUS_DISCHARGING; Oct 12 09:28:57 here for exmaple Oct 12 09:28:59 sorry the paste bin Oct 12 09:29:16 so read the condition about flags and then return a positive value Oct 12 09:29:37 But I'm not sure if the preemption is disable on this path Oct 12 09:33:12 panicking: hm, i think i see it now Oct 12 09:33:46 The best think Oct 12 09:33:56 is to acquire a lock in the get_propery path Oct 12 09:34:04 and the worker path Oct 12 09:34:46 hello all.. i just want to build my own image.. like it's written in the wiki, vala makes a failure which crashes the buildin Oct 12 09:34:55 there's also a workaround.. Oct 12 09:34:57 panicking: i see, probably yes. Oct 12 09:35:07 but i've got a error with bitbake Oct 12 09:35:17 ERROR: Please set the 'PERSISTENT_DIR' or 'CACHE' variable. Oct 12 09:35:25 I'm going to verify this issue on android and check if it is related to an invalid value Oct 12 09:35:30 so i couldn't use that workaround Oct 12 09:35:39 could someone help me? Oct 12 09:35:41 panicking: i do not think you should spent time on than right atm because probably we'll have to use bq27x00 driver instead. Oct 12 09:35:51 i define the cache-variable in bitbake/conf/bitbake.conf Oct 12 09:36:00 but it doesn't help Oct 12 09:36:20 skamster_: are you sourcing ./env-setup or something like that? Oct 12 09:36:36 setup-env Oct 12 09:37:01 i don't have this command Oct 12 09:37:35 ah, i see it Oct 12 09:37:39 skamster_: should have been created by the SHRMakefile: [ -e shr-testing/setup-env ] || ( cd shr-testing ; ln -sf ../common/setup-env . ) Oct 12 09:37:39 just a moment Oct 12 09:39:06 it doesn't work Oct 12 09:39:07 :s Oct 12 09:39:18 i didn't make changes on this file.. Oct 12 09:40:27 now i tried to make export CACHE=/tmp Oct 12 09:40:34 in the setup-env-file Oct 12 09:40:42 but it also doesn't work Oct 12 09:41:31 skamster_: can't you just trace the execution, read bitbake for a moment and understand the problem? Googling also usually helps. Oct 12 09:46:30 panicking: possibly your race is not actually possible because bq27000_battery_work is always scheduled on global workqueue and i think all property requests are going through that too. Oct 12 09:46:56 ok Oct 12 09:47:34 I'm not really sure in the second statement though. But at least power_supply changes are scheduled to run on global workqueue. And since we have only one cpu, there's only one global workqueue. Oct 12 09:47:49 maybe i search the wrong way, but i didn't found a reason in google.. Oct 12 09:50:29 But if you read from the sysfs Oct 12 09:50:39 the get_property is on process context Oct 12 09:51:05 is it correct Oct 12 09:51:17 woah, http://www.youtube.com/watch?v=erdl6Lz6fGo Oct 12 09:52:24 panicking: i'm not sure get_property doesn't go through global workqueue. Anyway, the probability of the race you suggest is quite low. Oct 12 10:14:23 usb transfer to phone is terrible, 130 KB/s any ideas? usb2 port on laptop. Oct 12 10:38:28 jonkristian: usual Oct 12 10:58:11 hello Oct 12 10:58:26 hola Oct 12 10:58:52 how is openmoko different from ,say, android or maemo Oct 12 11:00:01 Openmoko is an open system, all the components are open-source. Oct 12 11:00:06 The user has root Oct 12 11:00:19 And can run binary programs of their own design. Oct 12 11:00:38 This differs from android. Oct 12 11:01:03 Maemo5 has closed elements, though again at this time, the user gets root and can run binary programs. Oct 12 11:01:30 hey all, according to the wikipedia entry there was a 3G freerunner planned at somepoint, I'm assuming that's been canned now? Oct 12 11:01:50 what about from a end user point of view, if i buy a openmoko phone, why do i need to be root? Oct 12 11:02:03 There are unclear intimations of what nokia is going to do. - some statements from them indicate the next version may be more locked-down. Oct 12 11:02:14 indus: At the moment - frankly - you probably shouldn't. Oct 12 11:03:04 SpeedEvil: i need to research a project like openmoko or moblin or maemo, what would be a good suggestion Oct 12 11:03:09 While some images may be getting there in terms of functionality, and the hardware is OK, it's only real advantage at the moment is that you can develop on it. Oct 12 11:03:23 Inlcuding very low-level stuff if you choose to. Oct 12 11:03:39 domiel: well - openmoko - corporate - is not workin g on any more new phones. Oct 12 11:04:06 domiel: and to make a 3G freerunner would take a large slice of 100K at least to make a first tiny run of phones. Oct 12 11:04:45 The hardware is considerably less shiny than the n900 - but it's also more open and cheaper. Oct 12 11:05:00 ok got the major point, even the hardware is open source Oct 12 11:05:03 cool Oct 12 11:06:20 indus: research for what? Oct 12 11:06:52 SpeedEvil: well, i need to give some training to some people on mobile frameworks Oct 12 11:07:02 speedevil: so what choices do we have at the moment for an open 3G phone? Oct 12 11:07:05 SpeedEvil: how to develop apps on it etc Oct 12 11:07:18 domiel: none, pretty much. Oct 12 11:07:27 domiel: Wait 3 weeksish for the n900 Oct 12 11:07:29 indus: you can connect 3G dongle to FR because it features usb host. Oct 12 11:07:42 is the n900 going to be an EU or US release? Oct 12 11:07:49 Or go android and slightly less open - with a dev version. Oct 12 11:07:50 indus: n900 won't have an open telephony daemon for at least several months after the release. Oct 12 11:07:53 paul_bipless: wrong nick :) Oct 12 11:08:16 domiel: worldwide. But it only works on tmobile-3g in the US. Oct 12 11:08:17 to be fair i would be happy with 2g if there would be realworld-usable hw for me Oct 12 11:08:18 indus: hm, i just like yours too much Oct 12 11:08:20 ;) Oct 12 11:08:29 heh Oct 12 11:08:30 thanks Oct 12 11:08:39 Me too. Oct 12 11:08:49 I'm in AU so it should work here then I guess Oct 12 11:08:51 * indus is reading openmoko wiki and liking it Oct 12 11:08:52 roh: Especially when I get this web compressy thing sorted. Oct 12 11:09:00 roh: Currently benchmarking approaches. Oct 12 11:09:03 expensive powerhungy touchscreen 3g phones don't solve my 'where i currently still use my razrV3i' usecase. Oct 12 11:09:28 shame though I really like the fact that the whole platform was open with the freerunner Oct 12 11:09:30 oh well! Oct 12 11:09:32 indus: be aware that much of it was written when it seemed that openmoko was going places rapidly. Oct 12 11:09:33 means atleast 3 days without external power, WITH moderate usage and the usual low-density gsm networks like here in europe. Oct 12 11:10:02 SpeedEvil: i had heard of openmoko 2 years ago, but then android seems to have stolen the show Oct 12 11:10:10 in taiwan it made a week or so.. but there its all just glass and paper/steel, not concrete (buildings) and ive not seen less than 5 bars. Oct 12 11:10:15 indus: don't take offense but we usually call some "less experienced Indian coders" exactly the same as your nick ;) Oct 12 11:10:46 paul_bipless: please explain that Oct 12 11:10:55 indus: Personally I believe that if the software had been handled right, there could have been a mass geek market phone by xmas 2007. Oct 12 11:11:38 SpeedEvil for me everything with > qvga and or screen of >2.5" cannot fit the 'small enough to have with me allways' category Oct 12 11:11:40 indus: your nick is used in russian to mean "somebody from India" and it's a common belief that India has plenty of coders who write extremely weird code :) Oct 12 11:11:40 indus: It would have had a slightly ugly UI, but would have had a functional kernel, suspend, reasonable battery life, and have been usable for your average linux geek to hack on. Oct 12 11:11:52 paul_bipless: yeah iam indian Oct 12 11:11:57 paul_bipless: greetings Oct 12 11:12:26 People just doing it for a job with no skill at it are worldwide alas. Oct 12 11:12:28 paul_bipless: i think all code is weird anyway :) but dont worry, no offence taken Oct 12 11:13:02 indus: greetings :D Oct 12 11:13:36 paul_bipless: but why would you say indians produce weird code? Oct 12 11:14:26 paul_bipless: but i agree with SpeedEvil all that IT money glamour boom recruited all kinds of people Oct 12 11:14:51 anyways Oct 12 11:15:11 any phones with openmoko?? Oct 12 11:15:28 indus: not really other than the neo1973 and freerunner Oct 12 11:15:48 Some people have sort-of-got-it-running on other hardware, but I'm not aware of any ongoing projects. Oct 12 11:15:56 SpeedEvil: so you say that lack of applications slowed down the project Oct 12 11:16:24 indus: no. Oct 12 11:16:29 indus: Lack of a stable platform. Oct 12 11:16:58 what would you say is the current status of openmoko Oct 12 11:17:14 indus: There were - for example - several complete changes of software stack dropped from on-high, with little notice other than several weeks before an announcement that 'something really cool is coming up'. Oct 12 11:17:35 indus: openmoko as hardware platform is there: you can buy freerunner and enjoy coding (and even everyday usage). Oct 12 11:17:36 indus: And not enough work on core stability - getting the kernel actually stable ans bug-free. Oct 12 11:18:01 hmm Oct 12 11:18:10 indus: freesmartphone (FSO) is continuing to evolve. It'll most probably be available for palm pre soon. Oct 12 11:18:48 too many projects seem to be spurting now Oct 12 11:19:32 paul_bipless: looks like this is funded by open moko Oct 12 11:20:14 damn it, i need to learn c now probably Oct 12 11:20:43 ok thanks for your suggestions, not really sure what to do now Oct 12 11:21:27 paul_bipless: SpeedEvil i am beginning to get interested in the maemo project , what do u think Oct 12 11:21:29 indus: no, openmoko-the-company doesn't seem to do anything for about 2 months already. Oct 12 11:21:59 indus: i don't like maemo -- it's too closed. Even for N770 and N8x0 there's still no free software OS. Oct 12 11:22:31 mer Oct 12 11:22:32 paul_bipless:closed? says open on sitet Oct 12 11:22:43 There are large closed components in maemo Oct 12 11:22:47 paul_bipless: aah no says open in large parts Oct 12 11:22:48 oops Oct 12 11:22:49 mer is the open alternative Oct 12 11:23:06 ok popular question., what about android Oct 12 11:23:27 And interestingly, there are rumors of a free n810 graphics driver in the next couple of weeks Oct 12 11:23:33 paul_bipless: seen any hindi movies recently? its popular in roos i hear Oct 12 11:23:43 :) Oct 12 11:24:24 indus: they were popular 30 years ago among those who likes dances and indian songs. Oct 12 11:24:33 ;) Oct 12 11:24:50 ok Oct 12 11:25:08 i need some android comments Oct 12 11:25:15 what do you thinks Oct 12 11:25:34 indus: not really openly developed. Not targeted for hackers anyway. Oct 12 11:25:41 ok Oct 12 11:25:47 SpeedEvil: ill check mer Oct 12 11:26:01 indus: not targeted for development of native apps. Oct 12 11:26:21 paul_bipless: native apps? could you explain a bit Oct 12 11:26:37 SpeedEvil: hey cool , open maemo Oct 12 11:26:37 indus: android folks like java apps instead of C/Vala/C++ etc Oct 12 11:26:55 paul_bipless: which is heavy correct? Oct 12 11:27:32 indus: java is generally heavier compared to properly written native apps, yes. Oct 12 11:28:10 so are you people devs for open moko Oct 12 11:28:11 indus: to sum up: android is semi-open platfrom designed to make selling apps from Android Market easy and profitable for both google and devs. Oct 12 11:28:33 indus: most of us are random users/sometime devs Oct 12 11:28:40 indus: maemo is semi-open platform of Nokia and nobody knows if it'll become more open or less in the future. Oct 12 11:28:47 i think i will study mer a little Oct 12 11:28:51 indus: FSO is already there and works. Oct 12 11:29:10 indus: mer guys told me mer is not ready for "prime-time" (probably yet). And mer lacks telephony support. Oct 12 11:29:15 paul_bipless: i dont think once you releaase code under gpl, you can close it again/ Oct 12 11:29:25 Yes, you can. Oct 12 11:29:32 hmm interesting Oct 12 11:29:37 But you cannot close the GPL version if anyone has a copy Oct 12 11:29:52 ya that was what i thought Oct 12 11:30:03 so derivatives can exist and grow Oct 12 11:30:32 indus: real life shows it's usually a dead-end anyway :-/ Oct 12 11:30:40 ok Oct 12 11:30:58 basically, i work for a company which supports open projects Oct 12 11:31:22 so just scanning for info Oct 12 11:31:42 if indus likes it, he supports it :) Oct 12 11:31:48 indus: (dead-end) look at Qtopia/OPIE history. Or at any free android forks etc. Oct 12 11:31:50 anyways thanks for all the suggestions/comments Oct 12 11:32:10 bye , ill probably drop in again some time Oct 12 11:32:29 indus: feel free. Oct 12 11:32:38 indus: at least OM community is somewhat active Oct 12 11:32:42 And i hope friendly ;) Oct 12 11:32:49 :) yes of course Oct 12 11:33:04 thank you once again, i appreciate it Oct 12 11:33:08 byeee Oct 12 11:33:25 indus: the kernel work is ongoing btw. And some lowlevel stuff. You can get here soldering advices even ;) Oct 12 11:39:56 ta Oct 12 12:16:56 freesmartphone.org: 03mickey 07phonesim * reec68ecc2564 10/lib/simfilesystem.cpp: Oct 12 12:16:56 freesmartphone.org: simfilesystem: first attempt at fixing READ BINARY implementation to something that vaguely resembles a real GSM Oct 12 12:16:56 freesmartphone.org: device Oct 12 12:16:56 freesmartphone.org: 03mickey 07phonesim * rb6cc7dbc485d 10/modems/generic.xml: generic.xml: define SIM elementary file EFspn (SIM Provider Name) Oct 12 12:45:53 Any downsides mounting sdcard directly to /home/root in fstab? Oct 12 12:53:27 DocScrutinizer: (shorting R3004 R3005) had to dismount both caps (they were indeed 1uF) altogether and solder a small wire short to the pcb. Oct 12 13:11:49 PaulFertser: why? Oct 12 13:13:21 joerg_42: too accedentally made a bridge between those and i lack desoldering wick or something equivalent. Oct 12 13:13:41 joerg_42: apparently my soldering skills are suboptimal to say the least Oct 12 13:14:53 jonkristian: the old school is /-fs should contain ~root/ (as /root, not /home/root). So my recommendation was to switch to sane user management, have /root (to boot a operable system without uSD) and /home mounted from uSD-home/ Oct 12 13:16:11 jonkristian: in /home (on uSD) then create ~standarduser/ Oct 12 13:16:53 and start x-session under user standarduser, rather than root Oct 12 13:16:55 PaulFertser, I have some issue on accelrometer Oct 12 13:17:06 I have removed the activate/deactivte funciton Oct 12 13:17:12 but after resume suspend Oct 12 13:17:16 I have a lot of interrupts Oct 12 13:17:20 from lis driver Oct 12 13:18:14 panicking: more saladoil? Oct 12 13:19:50 panicking: are you using the the driver from openwrt svn or openmoko git? Oct 12 13:19:57 openmoko git Oct 12 13:20:12 try openwrt svn Oct 12 13:20:14 I'm working on a new android release Oct 12 13:20:20 i had the same problems Oct 12 13:20:30 Ok Oct 12 13:20:39 That's is great Oct 12 13:20:53 or did you already change the irq handler? Oct 12 13:21:09 No, the problem is that the device is powe off Oct 12 13:21:13 but it sends interrupt Oct 12 13:21:17 hu? Oct 12 13:21:19 joerg_42: thanks Oct 12 13:21:23 It can be related to the incorrect power off Oct 12 13:21:52 And the great thing is that I recevie interrupt from both Oct 12 13:21:54 joerg_42: I just went back to regular usage, and did a mount bind in fstab for Music, Photos and Videos folders. Oct 12 13:22:08 And I was using just one Oct 12 13:22:13 this is after resume Oct 12 13:22:27 strange Oct 12 13:22:35 the problem is related to the suspend resume path Oct 12 13:22:42 so the lis-1 is close Oct 12 13:22:47 and the lis-2 is close Oct 12 13:22:55 I'm sure because Oct 12 13:23:03 I put a debugging stuff in the driver Oct 12 13:23:12 and the close is called when nothing use them Oct 12 13:23:31 And it's correct because I change android to release file descriptor Oct 12 13:23:35 jonkristian: the problem with roothome being on a non-/ fs is you may find some cmds to fail until that fs is mounted. imagine e.g. how would you run a manual fsck on uSD if you can't login as root Oct 12 13:23:50 Strange i get a EXT3-fs: invalid journal inode. when booting shr, and i'm using reiserfs. Oct 12 13:24:14 joerg_42: yeah, whole problem was, intone didn't work with symbolic linking. Oct 12 13:24:52 joerg_42: but i think i will do it like this, just mount up the dirs i want Oct 12 13:28:28 ouch Oct 12 13:28:55 a2dp is waay too sluggish when played from sdcard Oct 12 13:34:07 panicking: what's your verdict about bq27k driver finally? Oct 12 13:38:33 I think that it needs to protect with a mutex, I can't reproduce the crash with the mutex Oct 12 13:38:38 and read invalid data from sysfs Oct 12 14:06:46 mutex? what for? Oct 12 14:10:23 panicking: of course you see irw from lis when powering down. IRQ is active-low. obviously you have to disable irq before powering down the gmeter chips o.O Oct 12 14:10:47 s/irw/IRQ/ Oct 12 14:10:47 joerg_42 meant: panicking: of course you see IRQ from lis when powering down. IRQ is active-low. obviously you have to disable irq before powering down the gmeter chips o.O Oct 12 14:15:48 joerg_42: panicking is right, there should be a mutex protecting regs, since it is accessed from different threads Oct 12 14:16:11 hmm Oct 12 14:18:30 if it's for the values read via HDQ and written to a data struct, then (possibly concurrently) read out by another thread... if it's that one I strongly recommend another approach: double buffering and an atom (means writeable in one machine OP cycle) pointer to the bouble buffer Oct 12 14:19:09 s/bouble/double Oct 12 14:20:46 if you can't have a atom pointer, then use a static array of pointers [2] and a byte index of 0/1 to the pointer Oct 12 14:20:59 aka array index Oct 12 14:21:22 the mutex is the right way Oct 12 14:21:29 it is already double buffered Oct 12 14:21:48 so why do we need any mutex then? Oct 12 14:22:16 cause you don't want to switch buffers when you access them Oct 12 14:22:31 huh? Oct 12 14:22:54 why? Oct 12 14:23:02 the code does di->regs.bla Oct 12 14:23:12 if it did regs = di->regs Oct 12 14:23:36 and the regs.bla Oct 12 14:23:51 you indeed wouldn't need the mutex Oct 12 14:28:26 the readout CC part writes to valuearray[1 - ptr].foo, and then when done does ptr=1-ptr Oct 12 14:28:51 what? Oct 12 14:29:11 the read part reads from valuearray[ptr] Oct 12 14:30:28 the concept is the same, yes Oct 12 14:31:25 and then the read part does things like if (valuearray[ptr].foo < 0) return valuearry[ptr].foo; Oct 12 14:31:59 and if ptr is updated in between you'll get incorrect data Oct 12 14:35:14 fix is either use a mutex protecting ptr. or do local_ptr = ptr at the begining of reading part and use local_ptr then instead of ptr Oct 12 14:40:17 agree Oct 12 15:21:14 larsc: don't use mutex. of course you need to safe the index/ptr to a local var. Oct 12 15:24:38 hi Oct 12 15:24:46 larsc, we have problem in the resume path Oct 12 15:24:50 of the accelerometer too Oct 12 15:25:21 joerg_42, the mutex is a clean solution Oct 12 15:25:34 we don't have here performance problem Oct 12 15:25:36 larsc: btw to be absolutely safe you need a second global index/ptr which the reading code sored the array-of-buffers element-"name" to of the buffer that currently is read. Oct 12 15:26:54 panicking: mutex is always a clean solution (if used correctly), but often introduces rather difficult to spot deadlocks or performance hogs Oct 12 15:27:53 there is not performance hog Oct 12 15:28:09 we need to use to protect userspace Oct 12 15:28:19 and worker Oct 12 15:28:27 the only performance hog in the freerunnner Oct 12 15:28:30 are the accelerometer Oct 12 15:28:36 that disable the interrupts Oct 12 15:29:02 the retry in the hdq_read is 3 Oct 12 15:29:12 but you can receive interrupt Oct 12 15:29:17 from other subsystem Oct 12 15:29:24 you don't block anything Oct 12 15:30:02 panicking: I'm not interested in random rant. If FR is too slow for you, search for different device. If you don't understand hw-irq issues, then get a good book about electronics Oct 12 15:31:07 joerg_42, I just said the the design of the accelerometer is wrong Oct 12 15:31:15 (driver) Oct 12 15:31:30 I don't undestand hw-irq issue Oct 12 15:31:32 aha Oct 12 15:31:43 but I understand that the code is wrong Oct 12 15:31:52 so we must fix Oct 12 15:32:08 so comment out my patch Oct 12 15:32:12 and provide a better one Oct 12 15:33:03 first of all I think we must prove you assumption. Andy Green rarely published wrong design Oct 12 15:33:52 This part is not written by Andy Oct 12 15:33:59 looking on the git Oct 12 15:34:09 and very obviously Andi has a better mandatory knowledge of the hw implications than you Oct 12 15:34:23 yes of course Oct 12 15:35:18 I don't discuss the andy work Oct 12 15:35:29 I just said tha doing the read of the accelerometer Oct 12 15:35:31 in interrrupt Oct 12 15:35:32 I don't see any problem with the gmeters driver Oct 12 15:35:54 is not a good think Oct 12 15:38:55 aha. I'd suggest you elaborate on this in a mail to [kernel] on OM-ml Oct 12 15:39:49 after eading all relating posts from ml-archives Oct 12 15:43:14 panicking: maybe you're suggesting that the actual reading should be done in a bottom half for the interrupt? Oct 12 15:44:18 panicking: what Weiss is sugessiting is what is done in the openwrt version of the driver Oct 12 15:45:08 Yes I know Oct 12 15:45:16 but you must move on a tasklet Oct 12 15:45:25 not a deferreblw work Oct 12 15:45:36 is a performance issue when you play with the freerunnner Oct 12 15:46:13 Check my patch on the android-kernel Oct 12 15:46:17 I just provide two Oct 12 15:46:32 of for the mutex and the other for the resume path of accelerometer Oct 12 15:58:02 I really hate the idea of patches pushed into kernel without any common discussion. Noone will read the patch to understand the reasoning, and noone will open a ticket after the patch has been supplied, to give other more savvy devels a chance to review the whole issue. Oct 12 15:58:51 I'm agree Oct 12 15:58:59 I don't push anything in the kernel Oct 12 15:59:04 I just sent an RFC Oct 12 15:59:13 I need to push in my android version of the kernel Oct 12 15:59:18 because fix an open ticket Oct 12 15:59:18 maybe you could suggest this procedure of wild random patching to linus torvalds Oct 12 16:00:28 joerg_42, Is there some bugs, if you don't see them it's not my problem Oct 12 16:00:34 there are some bugs Oct 12 16:00:36 sorry Oct 12 16:00:51 but maybe rewiever problem Oct 12 16:01:37 panicking: there are widely accepted best practices, if you don't know or understand them, that's not my problem Oct 12 16:02:40 Sorry :) can you exaplain how to proceed? Oct 12 16:02:40 exaclty Oct 12 16:02:46 I find a problem and I suggest a fix Oct 12 16:02:52 and send an RFC patch Oct 12 16:02:56 to the openmoko-kernel Oct 12 16:24:24 larsc, can you check your resume path of the accelerometer Oct 12 16:24:25 ? Oct 12 16:26:03 panicking: sure. I'll do when I'm home Oct 12 16:26:54 now it works like a charm the return path Oct 12 16:27:00 resume path Oct 12 16:27:09 and android go fast after resume Oct 12 17:03:58 panicking: so were you able to reproduce the race? Oct 12 17:04:45 PaulFert1er, with my patches with the mutex Oct 12 17:04:47 I can't reproduce Oct 12 17:05:11 the problem Oct 12 17:05:14 in android Oct 12 17:05:14 larsc: when i tried to solve gta01_battery problems on gta02 i faced the whole way of obtaining values from the driver. I suspect even if you simply request any value from sysfs then your request goes through main workqueue, the same bq27k worker thread runs on. Oct 12 17:05:25 panicking: can you reproduce it without your patch with mutex? Oct 12 17:05:42 if you read throw the sysfs Oct 12 17:05:46 you call the fuction Oct 12 17:05:49 directly Oct 12 17:05:56 so you don't use the workqueue Oct 12 17:06:05 if you use the uevent Oct 12 17:06:10 you don't have any problem Oct 12 17:06:48 panicking: ah, ok. I wonder why i don't see the problem with my test program i sent to you. Oct 12 17:13:14 * PaulFertser has to use gprs because his adsl somehow doesn't work so any e-mail related work is postponed. Oct 12 17:13:48 My neo1973 is 1/3 the speed of my broadband. Oct 12 17:13:52 gprs Oct 12 17:14:13 Unfortunately, this doesn't mean it's unusually fast. Oct 12 17:14:40 I'd use gprs without much difficulties if only it wasn't inadequately expensive here. Oct 12 17:15:43 adsl completely down, routing issue, or large packets aren't going across right? Oct 12 17:15:51 joerg_42: btw, with "R3004" and "R3005" shorted the sound was louder, that's no surprise. Oct 12 17:16:37 yup Oct 12 17:16:50 Blu3: adsl is up, PADO discovery (PPPoE) doesn't work, ISP says they have some problems they're fixing (no time estimate given). Oct 12 17:16:51 * joerg_42 away now. Too much noise today Oct 12 17:17:02 nod Oct 12 17:19:32 the latest is the a7 correct? Oct 12 17:20:56 SpeedEvil, lindi- ? Oct 12 17:21:27 objorn: correct Oct 12 17:21:32 objorn: a8 was designed but never MPed Oct 12 17:23:22 MPed? Oct 12 17:24:12 objorn: mass produced Oct 12 17:35:36 freesmartphone.org: 03mickey 07cornucopia * r96bd628bc3ea 10/fsogsmd/src/ (7 files in 2 dirs): fsogsmd: org.freesmartphone.GSM.SIM.GetSimInfo() partly done (imsi and issuer) Oct 12 17:37:38 PaulFertser: do you have a link to the a8? Oct 12 17:38:35 objorn: not really, why do you ask? Basically it was a7 immune to gsm interference, i.e. no chance for GSM to ruin internal or headset mic. Oct 12 17:39:18 PaulFertser: because i can probably make it myself... Oct 12 17:40:48 objorn: nope, it's really easier to make it work properly with bluetooth headset. Oct 12 17:43:57 okay Oct 12 17:44:07 well, here i go, purchasing an a7 right now Oct 12 17:44:34 objorn: if you're diyer you can buy a6 and just implement buzzfix. Oct 12 17:44:46 objorn: you'll probably want to open the device anyway for #1024 fix. Oct 12 17:46:15 android must be constantly reading the battery Oct 12 17:46:19 info Oct 12 17:46:44 hi all Oct 12 17:47:37 dropbear sshd runs on openmoko. How the keys will be generated (public + private) and also the ssh/authorized_keys? Oct 12 17:48:03 who will generate it and when? where are the tools dropbearkey and ssh-keygen for doing it on target? Oct 12 17:49:25 anyone please help? Oct 12 17:54:29 hackbot: can't you just read /etc/init.d script for dropbear? Oct 12 17:55:00 in the openmoko rootfs? Oct 12 17:55:22 Paul: read in /etc/init.d of openmoko rootfs? Oct 12 17:57:28 hackbot: yes Oct 12 17:57:45 Sure Paul Oct 12 17:57:47 Thanks Oct 12 17:58:22 hackbot: and i think authorized_keys are always modified only by hand Oct 12 18:01:36 if so, then how i am able to login to the OM device when i type ssh root@ip? Oct 12 18:01:47 sorry if i miss something Oct 12 18:01:49 becasue there is no root password Oct 12 18:02:39 the confusion dropbear is not running without the following 2 keys: /etc/dropbear/dss_key and /rsa_key Oct 12 18:02:57 so if i put those 2 keys Oct 12 18:03:21 and then do ssh root@ip, it says Permission denied, publickey, etc Oct 12 18:03:34 hackbot: those keys should be generated on the first boot after you flash the device Oct 12 18:04:02 auto generated on startup? Oct 12 18:04:42 what r generated? rsa_key, dss_key and .ssh/authorized_keys? Oct 12 18:05:18 hackbot: oh, you seem to have quite an approximate knowledge of how ssh is configured usually. Oct 12 18:05:33 yes Oct 12 18:07:02 has anyone entertained the idea of implementing some animation for incoming calls/sms'? Oct 12 18:08:14 I'd kill to have the 'badgers' flash vid to go along with the ringtone I have Oct 12 18:08:28 tig| - please reply Oct 12 18:09:23 hackbot: usually the rsa and dsa keys nothing is put in authorized keys as by default most of the distros don't ship with key based auth Oct 12 18:09:46 openmoko has dropbear in it Oct 12 18:09:54 and the keys are not there? Oct 12 18:10:02 so all you just do ssh root@192.168.0.202 then when it asks you for a password just hit enter Oct 12 18:10:20 thats working Oct 12 18:10:21 as there is no root password set on the default image you are supposed to log in and set one Oct 12 18:10:43 but on phone side, i m using my own rootfs Oct 12 18:10:54 by compiling dropbear myself Oct 12 18:11:08 I can't help you there and I have to go now as my car is here Oct 12 18:11:10 bye Oct 12 18:11:17 sure Oct 12 18:13:32 Paul Froster: any idea of how to start dropbear without these 2 keys? Oct 12 18:13:34 :( Oct 12 18:13:50 when i dont give this 2 keys, its not starting at all? Oct 12 18:16:15 Paul - any ideas? Oct 12 18:18:35 PaulFertser: i don't think there's an option to just buy an a7 anyway in the states Oct 12 18:18:46 http://sdgsystems.com/estore/cart.php?target=category&category_id=17 has the phone but in orders of 10 now Oct 12 18:18:52 hackbot: Not many people highlight on partial versions of their nicks. :P Oct 12 18:18:59 PaulFertser_ ^ Oct 12 18:19:10 PaulFertser: i don't think there's an option to just buy an a7 anyway in the states Oct 12 18:19:18 http://sdgsystems.com/estore/cart.php?target=category&category_id=17 has the phone but in orders of 10 Oct 12 18:19:49 :P Oct 12 18:19:56 i guess if it's just a matter of opening it up and replacing a a part or two, i can do that Oct 12 18:23:41 Thanks Paul Oct 12 18:23:48 Everything in init.d scripts Oct 12 18:24:14 objorn: for #1024 fix you'll also need to open the can but it's not hard. Oct 12 18:24:19 hackbot: welcome Oct 12 18:24:42 objorn: for buzzfix you have to just replace one 0402 R and add a big 100uF nearby. Oct 12 18:25:47 objorn: so if you have a chance to buy non-buzzfixed a6 for a reasonable price probably you might want to do buzzfix yourself. Oct 12 18:26:10 PaulFertser: so far the cheapest price is $250 from openmoko.com Oct 12 18:26:47 objorn - Any idea where to get in Taiwan Oct 12 18:27:17 I have my friend over there to buy for me Oct 12 18:28:49 freesmartphone.org: 03mickey 07cornucopia * r1026537fb9be 10/fsogsmd/src/ (5 files in 2 dirs): fsogsmd: implement org.freesmartphone.GSM.SIM.{Get|Set}ServiceCenterNumber() Oct 12 18:29:11 You can't buy it in taiwan - if you mean at the factory gates Oct 12 18:29:21 hackbot: http://www.dmatek.com.tw/ Oct 12 18:29:33 linked from the openmoko website Oct 12 18:29:55 I mean - physically from openmoko Oct 12 18:31:12 :( Oct 12 18:31:22 i thought in factory i can get cheaper :( Oct 12 18:31:41 objorn - thank u Oct 12 18:33:21 np hackbot Oct 12 18:47:59 objorn: i could have missed something, still some problems with my isp Oct 12 18:58:36 objorn: have you considered http://pulster.eu/ Oct 12 19:06:23 PaulFertser: not until now, but they charge 80usd for shipping Oct 12 19:06:40 i'm just going to purchase the a6 from openmoko's site and apply the fixes myself Oct 12 19:19:31 hi guys Oct 12 19:20:14 has there really been no update for shr unstable in a long time now? Oct 12 19:24:21 objorn: feel free to ping DocScrutinizer or me if you have any questions. Do you also consider the bass fix? Oct 12 19:25:21 http://cgi.ebay.com/Big-Mouth-Billy-Bass-Singing-Fish_W0QQitemZ150378062683QQcmdZViewItemQQptZLH_DefaultDomain_0?hash=item23033b235b&_trksid=p3286.c0.m14 Oct 12 19:29:59 SpeedEvil: wtf is that? I honestly can't understand why it is called bass singing. Oct 12 19:30:16 The fish is a bass Oct 12 19:30:19 it's a species Oct 12 19:31:11 SpeedEvil: i thought it's related to the bass guitars :D Oct 12 19:34:37 PaulFertser: i might look into it when i get the phone, thanks for offering to help, i'll probably ask a few more questions before breaking it open Oct 12 19:34:53 objorn: wiki has quite some info Oct 12 19:35:17 yeah, i just got through reading over the bass fix on the wiki Oct 12 19:35:24 and here's a listing of reworks: http://wiki.openmoko.org/wiki/Category:Reworks_and_mods Oct 12 19:35:49 note: i won't be ordering the phone until tomorrow night or wednesday morning unfortunately Oct 12 19:39:24 objorn: misses #1024 fix Oct 12 19:46:41 see history - might miss one or two bits Oct 12 19:47:16 PaulFertser: how's your flu? mine's about to start Oct 12 19:47:46 DocScrutinizer: luckily it wasn't flu, just something flu-like but light enough, i'm recovering currently. Oct 12 19:48:02 DocScrutinizer: (about to start) doesn't sound good :( Oct 12 19:48:10 so for me there's still hope then :-D Oct 12 19:48:22 DocScrutinizer: probably some immune-boosting preventive measures can help? Oct 12 19:48:53 pondering that. wpwrak usually recommends Jaegermeister Oct 12 19:52:51 DocScrutinizer: i've tried adding some brendy to my tea in the beginning of illness. Not sure it affected anything :) Oct 12 19:54:30 DocScrutinizer: i don't know about Germany but here real flu usually comes considerably later. And it can be mostly prevented or at least be made milder by vaccination. Oct 12 19:55:42 DocScrutinizer: (miss one or two bits) did i miss anything really important? Oct 12 19:56:10 not you - the age Oct 12 19:56:12 page Oct 12 19:56:28 quite new un non-mature Oct 12 19:56:53 bah! /me away. Don't find the keys anymore Oct 12 19:57:03 DocScrutinizer: take care Oct 12 19:58:41 Took me a minute to get it (which page misses what) :) Oct 12 20:21:05 I have added a short note about 1024 to the reworks category Oct 12 20:21:33 and the c1767 thing has now a picture Oct 12 21:03:27 Haven't used mt gta02 in months, just plugged it in to try android on it, and not seeing any evidence it still works or is charging Oct 12 21:03:36 fso-illume-image is maintained in org.oe.dev ? Oct 12 21:03:47 the red aux led isn't on for example, which I thought it used to do when charging the battery? Oct 12 21:05:22 nope, 'tis the blue LED (orange when charging) at the power button Oct 12 21:05:40 (in shr-u at least) Oct 12 21:05:51 redargless of colour, I'm not getting any led on with it plugged in Oct 12 21:06:07 regardless too Oct 12 21:06:25 PaulFertser, three days of heavy writing to the SD card with gsm radio turned off. zero problems Oct 12 21:06:52 yashy, power adapter or usb cable? Oct 12 21:07:04 is it possible if the battery is completely dead it'll take some time for the LED to come on? Oct 12 21:07:24 BLu3: power adapter (tied USB first with no luck) Oct 12 21:08:01 try pulling the battery out, plug in the power adapter and try turning it on (mine turns on/resumes automatically) Oct 12 21:08:04 then put the battery in Oct 12 21:08:06 s/tied/tried/ Oct 12 21:08:06 Yashy meant: BLu3: power adapter (tried USB first with no luck) Oct 12 21:08:45 if it hasn't been used in months, the battery is dead, you'll have to play some tricks to get it to wake back up. what i just mentioned is one trick Oct 12 21:09:02 Yashy: yes, search the wiki for 'flat battery' Oct 12 21:10:26 Blu3: do you think the majority of gta02s would experience similar problems? Can you give some easy way to reproduce so anybody can try? Oct 12 21:10:48 Blu3: probably it's your device that has some oddity. Oct 12 21:10:55 being that i'm mostly the only one i've seen talking about it, i doubt it Oct 12 21:11:11 but it's 100% reproducible by enabling gsm and doing sdio Oct 12 21:11:31 anytime both of them are doing something, the sdcard loses Oct 12 21:11:44 mickey|bbl, I want to show off with fso on kovsky but illume-image is broken Oct 12 21:12:58 after i get gentoo installed, i'm going to see if both reading and writing are 100% breakable or if it's writing that is most susceptible and reading is more transient Oct 12 21:13:19 Blu3: that's, well.. you know eh like uhm sad. Probably DocScrutinizer can suggest some shielding strategy. Oct 12 21:13:45 aye Oct 12 21:14:00 *wraps phone in tin foil hat* Oct 12 21:14:02 ;-) Oct 12 21:14:36 damn, no love with Blu3's idea of the wiki suggestion Oct 12 21:14:42 of = or Oct 12 21:14:53 keep trying Oct 12 21:14:56 it'll come alive Oct 12 21:14:58 ok Oct 12 21:15:03 Yashy: do you have a wallcharger? Oct 12 21:15:07 or feel free to donate it to my toybox :D Oct 12 21:15:11 Yashy: and A6 device? Oct 12 21:15:16 yeah, using the wall charger right now Oct 12 21:15:55 Where would it say A6? Oct 12 21:15:58 Yashy: take out the bat, boot to NOR, should work. Then boot the kernel, insert the bat. Oct 12 21:16:35 nothing happens when I take out the battery, plug it into the wall charger, hold down AUX + hit power button Oct 12 21:16:45 (with no battery in it) Oct 12 21:16:54 Yashy: you're supposed to hold AUX before plugging in the wall charger. Oct 12 21:17:39 Yashy: plugging in is considered to be one of those events that turn the device on. Oct 12 21:17:47 aux button->charger->power button, then release buttons Oct 12 21:17:50 omg i should really go sleep Oct 12 21:17:56 heheheh Oct 12 21:18:41 Yashy: insert bat, hold aux, insert chager plug, if nothing happens press powerbutton while still holding aux. Then release aux while holding powerbutton Oct 12 21:19:42 Competition yay Oct 12 21:20:40 hmm, i need to get some actual tin foil. lay a strip below and above the sd card and see if that'll help blank enough interference Oct 12 21:21:18 Blu3: i doubt it's sd card itself that needs shielding. Because the same cards work in others' gta02s. Oct 12 21:21:42 ok removed battery, held down AUX, plugged into wall, pressed power. Nothing. Just tried this again but added battery before holding aux and plugging into wall. Still no signs of life Oct 12 21:22:13 yes, and my cards work perfectly elsewhere and when gsm is turned off Oct 12 21:23:07 have patience. how long've you tried holding in the power button? >8sec? Oct 12 21:23:38 less than 3 seconds Oct 12 21:24:01 didn't realize I should just hold aux + power until it works Oct 12 21:25:35 no luck with ~20 seconds Oct 12 21:26:54 yashy, http://wiki.openmoko.org/wiki/Neo_FreeRunner_Battery Oct 12 21:29:37 yeah I'm going through it now. Just tried with USB instead of power, no luck Oct 12 21:29:45 ok Oct 12 21:36:36 I'm thinking I have A5 or early A6, as it won't start without the battery Oct 12 21:37:03 well. you can always jumpstart the battery Oct 12 21:37:50 it's 3.7v, LI, so you want to apply a low current charge Oct 12 21:38:15 its interesting for me, what is ealy A6? Oct 12 21:38:26 it won't need to be charging for too long to have enough juice to kick over the phone with a wall charger then start charging while remaining powered on Oct 12 21:38:40 I bought mine on Sep, year ago in poland, so it was manufactoried at least in aug Oct 12 21:38:52 and mine can boot without battery Oct 12 21:39:23 Yashy: if the way I recommended above won't show the NOR uboot menu, then you're out of luck, no matter which method. jumpstart or recharge bat externally Oct 12 21:39:35 the only way to recover Oct 12 21:39:45 paroli won't start errors with ImportError: No module named dbus. any ideas? Oct 12 21:40:36 max_posedon: A6 (with 47uF) is random work/don't without bat Oct 12 21:41:15 47uf == buzz fix? Oct 12 21:41:21 max_posedon: 10uF devices never, 100uF devices (A7) always work Oct 12 21:41:32 ok thanks joerg_42 Oct 12 21:41:40 max_posedon: c1767 Oct 12 21:41:48 Vsys buffer C Oct 12 21:41:51 hm, ok, but I don't understand this numbers) Oct 12 21:41:54 2*47 will of course work Oct 12 21:42:00 oh Oct 12 21:42:02 nvm Oct 12 21:43:51 max_posedon: grep ml for c1767, and refer to schematics and the notes on hw rev I placed there. then you'll understand Oct 12 23:07:55 ooops, maybe not. s/10uF/4.7uF/ Oct 13 01:29:00 Hi guys, I'm using my FR (with SHR) as main phone. I experienced sometimes that when I start SHR, GSM functions doesn't start Oct 13 01:30:12 I don't know how can I check any log or how can I start manually GSM services, I type /etc/init.d/frameworkd restart, but I don't know if it's supossed to solve this issue Oct 13 01:34:37 can you give me some advices to take in count when I have this problem? Oct 13 01:37:30 Cristianpark1, I can't, though I wish I could This IRC channel is not very active anymore. You might get better results by posting a question to the cumminity mailing list. It is pretty active. Oct 13 01:45:13 KenYoung: I'm sorry to hear that. I think that we have a nice way to communicate by IRC Oct 13 01:45:32 we can't let die this IRC channel Oct 13 01:46:09 I agree. I hang out here a bit, but sadly I don't know enough to answer anything except the very easiest questions. Oct 13 01:47:20 I resonded to your question in part just to let you know someone is here besides CIA-42 ! Oct 13 01:48:30 thanks for the answer, always is good to know that at least you're talking to some human being Oct 13 01:50:11 I would have to say I have never see the problem you are reporting, and I've been running SHR for months. Maybe you've just got a buggy version, and reloading with an older or newer release of SHR might help. Did your problem begin after an SHR upgrade? Oct 13 01:51:19 it happens to me few hours ago, I was renaming icons of the desktop so I can have them in the order I want Oct 13 01:52:21 when I was editing pidgin.desktop I save the file (I was connected via SSH) and got an error message that give me two choices: Restart - Stop Oct 13 01:52:52 I press Stop and it goes to the terminal and hangs on a message: restarting xserver Oct 13 01:54:07 so I power down by removing the battery (I know, the ugly way) and wait a few seconds to put it back again. When I power on the FR again, when SHR starts, I don't have GSM services Oct 13 01:54:56 I let the FR without any intervention a few minutes and it never get GSM active, I tried restarting frameworkd and it doesn't do anything Oct 13 01:55:55 note that I have last SHR Oct 13 01:56:10 with the ERROR: Cannot Satisfy issu Oct 13 01:56:11 e Oct 13 01:57:32 Well, I have my FR running SHR, so if you want me to check any files etc on a working phone, I can do that for you. But probably not much else. Oct 13 02:12:27 KenYoung, I didn't see your answer, hehe. Maybe you can check if your /usr/share/applications/pidgin.desktop have weird characters in the app name in other languages different than english Oct 13 02:14:06 Cristianpark1, Yes, there are lots of non-latin characters in that file. Arabic, some derivitive of Sanscrit, Greek, etc. Oct 13 02:15:01 Japanese, Korean - lots of names. Oct 13 02:15:33 Sorry about my poor spelling. Oct 13 02:16:17 thanks, right now I have GSM working but I'll restart soon with the sim card with wich I'm having the problem and let you know Oct 13 02:33:44 have you tried Android? Oct 13 02:34:24 Cristianpark1, Yes, I had Android on one of my Freerunners for a while, and I have a G1 with Android. Oct 13 02:35:27 And how good is it on the FR¿? Oct 13 02:36:00 Not as good as SHR, IMO. Oct 13 02:36:25 But it does work. Oct 13 02:36:59 I never tried WiFI or GPRS with it, though. Oct 13 02:38:00 For me SHR works well enough that I only need to reboot about once every three days. Oct 13 02:38:34 WiFi is OK, at home at least. GPRS is reliable, but of course very slow. Oct 13 02:40:38 I like SHR too Oct 13 02:40:45 Its very user friendly Oct 13 02:41:57 but I had some issues with wifi on my previous installation Oct 13 02:42:11 Cristianpark1, You probably saw the poll conducted a few weeks ago surveying which distribution FR users ran. SHR had an order of magnitude more users than any other distribution. Basically nearly everyone runs SHR. Oct 13 02:42:16 how do you make to connect with SHR? I did it with wpa_supplicant Oct 13 02:42:49 but I couldn't get Wifi working every time Oct 13 02:42:50 I use Mofi, though most people do not. Mofi is losing "Market Share", I think. Oct 13 02:43:15 Cristianpark1, WiFI does not work everywhere for me either. Oct 13 02:43:28 but does it works well? Oct 13 02:44:01 When it works at all, it works well. COnnections last for many hours. As long as I want, basically. Oct 13 02:44:36 "When it works at all, it works well." <-- my thoughts exactly. Oct 13 02:45:28 mostly_useless, Yes, sadly for me, it doesn't work at all with my workplace WiFi. Oct 13 02:45:54 But at home, no problems. Oct 13 02:46:27 KenYoung: what encryption uses your home wifi? Oct 13 02:48:02 Cristianpark1, WPA Oct 13 02:48:25 Mofi is on the SHR repo? Oct 13 02:49:00 KenYoung: I get about 50/50. sometimes it works beautifully, and sometimes it a frustrated battle of boy vs. machine. Oct 13 02:49:11 Cristianpark1, Yes, at least it was when I last loaded a new SHR, which was a couple of months ago. Oct 13 02:49:58 KenYoung: seems to be every distro with me. haven't yet found a solution. Oct 13 02:50:02 KenYoung: Thank you Oct 13 02:52:23 mostly_useless, For what it's worth, on my home WiFi, FR WiFi is flawless. Oct 13 02:53:14 KenYoung: I think I need to play with tx power. haven't gotten around to it yet. I've been busy playing with android. Oct 13 02:56:10 mostly_useless, I'm curious, why do you use Android rather than SHR? **** ENDING LOGGING AT Tue Oct 13 02:59:56 2009