**** BEGIN LOGGING AT Sun Jun 16 02:59:58 2013 Jun 16 06:03:18 what am I going to do about this GSM modem shield **** BEGIN LOGGING AT Sun Jun 16 07:13:03 2013 Jun 16 08:40:42 hello guys Jun 16 08:43:28 I tried (twice) lastest update for cssu-testing. N900 works fine for few to seweral hours, then phone app stops to work. After switching it off it never boots again - it neven actually hits those flashing dots at booting - only nokia logo for longer, then usual time and dies death after minute or so - every time. Any suggestions? Jun 16 08:54:51 is battery level normal? Jun 16 09:12:44 <_rd> What is the difference between Jun 16 09:12:46 <_rd> deb http://repository.maemo.org/extras/ fremantle-1.3 free non-free Jun 16 09:12:48 <_rd> and Jun 16 09:12:53 <_rd> deb http://repository.maemo.org/extras/ fremantle free non-free Jun 16 09:12:55 <_rd> ? Jun 16 09:18:09 should be none Jun 16 09:18:30 until we roll out PR1.4 ;-P Jun 16 09:22:33 <_rd> Hmm....but there are different files in there, e.g. check Jun 16 09:23:29 <_rd> http://repository.maemo.org/extras-devel/dists/fremantle-1.3/free/binary-armel/ and http://repository.maemo.org/extras-devel/dists/fremantle/ Jun 16 09:23:45 if there are different files, then we have a problem that needs fixing Jun 16 09:23:51 <_rd> fremantle seems to have newer stuff than fremantle-1.3 Jun 16 09:23:59 <_rd> (check file size and date) Jun 16 09:27:36 well, then we have a good way to deduce when it got broken Jun 16 09:28:21 you just experience live the daily nightmare that is maemo administration Jun 16 09:29:21 hey could you do me a favor and send a rant to support@nemein.com? Jun 16 09:29:37 CC techstaff@maemo.org Jun 16 09:32:15 <_rd> Sure, will do. Jun 16 09:35:27 <_rd> done Jun 16 09:38:45 <_rd> I get GPG warning from the nokia server, when I run apt-get update Jun 16 09:38:46 many thanks (not that I hope for much) Jun 16 09:38:47 <_rd> W: GPG error: https://downloads.maemo.nokia.com ./ Release: The following signatures were invalid: KEYEXPIRED 1349249546 KEYEXPIRED 1349249546 KEYEXPIRED 1349249546 Jun 16 09:38:56 <_rd> is that expected? Jun 16 09:39:07 you already have a clue when they started to differ? Jun 16 09:39:20 <_rd> I hope that I get a reply at least ;-) Jun 16 09:39:33 yes, that's another known fuckup, this time courtesy Nokia Jun 16 09:39:58 <_rd> what do you mean with "when *they* start to differ?" Jun 16 09:40:13 fremantle and fremantle1-3 Jun 16 09:40:51 or let me put it this way: what are the latest dates in fremantle Jun 16 09:41:24 err. fremantle-1.3 Jun 16 09:41:27 <_rd> fremantle is June-12, fremantle-1.3 is Dec-16 Jun 16 09:41:38 yeah, thought as much Jun 16 09:41:53 that's just before Nemein started migration#1 Jun 16 09:42:46 re "W: GPG error:" W: should mean "Warning!" Jun 16 09:42:58 it _should_ not cause much trouble Jun 16 09:43:11 usually Jun 16 09:43:39 if it does, you can try setting your local (N900) date to 2011 Jun 16 09:43:57 or 2012 Jun 16 09:44:41 (reply) you already get reply from techstaff ;-) Jun 16 09:44:51 <_rd> no it does not cause trouble, I just saw it. Jun 16 09:44:56 don't hope for reply from nemein though Jun 16 16:34:35 * DocScrutinizer05 treats channel to a free ABSOLUT vodka, for 4444 thanks on tmo Jun 16 16:35:46 :D Jun 16 16:35:49 Congrats. Jun 16 16:35:55 Now where's my Drink? Jun 16 16:54:35 * DocScrutinizer05 hands schnaps to Raimu Jun 16 16:54:51 wtf is wrong with DocScrutinizer51? Jun 16 16:56:08 ugh, hetzner network issue Jun 16 16:56:19 hardly ever seen that Jun 16 16:57:00 \o/ Jun 16 16:59:09 ~bupbat Jun 16 16:59:11 well, bupbat is http://www.mouser.com/catalog/632/2009.pdf http://il.farnell.com/taiyo-yuden/pas414hr-va5r/cap-0-06f-3-3v-80ohm-4-8mm-coin/dp/1853000?Ntt=PAS414HR-VA5R, or use the capacitive type, LiIon are breaking during 12 months Jun 16 17:02:39 ~bupbat is also http://www.digikey.de/product-detail/de/PAS414HR-VG1/587-2157-1-ND/1959153 Jun 16 17:02:40 DocScrutinizer05: okay Jun 16 17:07:34 ~bupbat is also https://hbe-shop.de/KONDENSATOR006F-33V-STAKED-COIN Jun 16 17:07:34 okay, DocScrutinizer05 Jun 16 17:08:04 ~bupbat Jun 16 17:08:06 it has been said that bupbat is http://www.mouser.com/catalog/632/2009.pdf http://il.farnell.com/taiyo-yuden/pas414hr-va5r/cap-0-06f-3-3v-80ohm-4-8mm-coin/dp/1853000?Ntt=PAS414HR-VA5R, or use the capacitive type, LiIon are breaking during 12 months, or http://www.digikey.de/product-detail/de/PAS414HR-VG1/587-2157-1-ND/1959153, or https://hbe-shop.de/KONDENSATOR006F-33V-STAKED-COIN Jun 16 17:09:59 infobot: no, bupbat is use the capacitive type, LiIon are breaking during 12 months, or http://www.digikey.de/product-detail/de/PAS414HR-VG1/587-2157-1-ND/1959153, or https://hbe-shop.de/KONDENSATOR006F-33V-STAKED-COIN Jun 16 17:10:00 DocScrutinizer05: okay Jun 16 17:15:22 * DocScrutinizer05 feels like spending some 20..30bucks for getting a dozen bupbat, so they can get sent out to Europe for recipients donating 3 bucks to maemo in return Jun 16 17:16:46 farnell not selling to consumers at all, HBE asking for 10 bucks shipping and package, no matter if you order one or 20 of those 1.50EUR critters Jun 16 17:17:58 freemangordon_ once offered to get some of them with his company's next huge order at digikey or farnell, but I bet he forgot about it Jun 16 17:19:29 hell, I probably even could offer rework service Jun 16 17:20:00 but nah, that's just too much hassle Jun 16 17:30:06 done :-) Jun 16 17:30:33 if you are in need of a replacement bupbat and live in Europe, contact me! Jun 16 17:32:04 and now, time for a nice cool beer in a nice comfortable beergarden Jun 16 19:23:07 web frontend, ohmy Jun 16 19:23:33 o/ Jun 16 19:46:31 Is the n900 supposed to reboot automatically? Jun 16 19:47:04 Not generally Jun 16 19:52:01 SpeedEvil: It keeps rebooting when I do an upgrade in the debian chroot, or when cp hangs on the mydocs filesystem :/ Jun 16 19:52:26 This breaks the filesystem, and I'd like to know why before I reformat it (so it doesn't happen again) Jun 16 19:52:28 Any ideas? Jun 16 19:56:05 I would run badblocks while watching dmesg Jun 16 19:57:29 badblocks? Jun 16 19:57:48 Root command, I presume? Or is it an application to be installed? Jun 16 19:59:02 It's a console application, that is designed to read (or write) a block device to see if it has issues Jun 16 19:59:08 afk Jun 16 20:16:37 swaphell Jun 16 20:17:52 lansiir: I guess you tried copying huge amounts of data when reboot happens Jun 16 20:18:27 It was the debian image (as the debian image was currupted by the crash during apt-get upgrade). Jun 16 20:18:34 So 3.4GB IIRC. Jun 16 20:19:54 DocScrutinizer05: ^ Jun 16 20:21:45 :nod: Jun 16 20:23:01 lansiir: known issue, watchdog kicks in when IO hogs CPU. You can cure that by tweaking some parameters regarding IO-handling, like number of queue entries, and swappiness and whatnot Jun 16 20:23:25 Figured as such, my android phones did that. Jun 16 20:23:53 googling for swappolube might bring up some helpful links Jun 16 20:24:00 Then I'll just be sure not to do that then :P Jun 16 20:24:17 Any way to disable that entirely on a temporary basis (for copying images, etc.) Jun 16 20:28:16 DocScrutinizer05: ^ Jun 16 20:44:44 lansiir: known issue, watchdog kicks in when IO hogs CPU. You can cure that by tweaking some parameters regarding IO-handling, like number of queue entries, and swappiness and whatnot Jun 16 20:45:17 you hardly can disable a bug Jun 16 20:45:40 I hail from mac-land: "It's not a bug, it's a feature" Jun 16 20:46:07 But thanks, I'll research that and take a look. Jun 16 20:46:36 just a minute... Jun 16 20:46:52 DocScrutinizer05: Can I just umount and reformat the mydocs repartition to cure my corrupted partition woes, or is there something more complex? Jun 16 20:50:28 MyDocs is easily umount'able and you can mkfs.vfat any type when umounted Jun 16 20:50:38 Great :D Jun 16 20:50:39 echo 100 > /proc/sys/vm/swappiness Jun 16 20:50:40 echo 5 > /proc/sys/vm/page-cluster Jun 16 20:50:42 echo 500 > /proc/sys/vm/dirty_expire_centisecs Jun 16 20:50:43 echo 4 >/sys/block/mmcblk0/queue/nr_requests || echo "ohnoes, hope it boots1" Jun 16 20:50:45 echo 4 >/sys/block/mmcblk1/queue/nr_requests || echo "ohnoes, hope it boots2" Jun 16 20:51:03 is close to end of my /etc/init.d/rcS Jun 16 20:52:09 there might be better values, other parameters, and better ways to mess around with them, but generally that's whete you start Jun 16 20:52:57 ShadowJK claims nr_requests is best set high Jun 16 20:53:27 DocScrutinizer05: Looks standard (except for those nr requests). Jun 16 20:53:44 * lansiir has it saved for future reference. Jun 16 20:54:04 Since I'm reformatting MyDocs, should I be reformatting as ext2? Jun 16 20:54:09 no you shouldn't Jun 16 20:54:24 maemo will shit its pants if the first partition on the emmc isn't vfat Jun 16 20:55:01 Right, so, vfat is is. Jun 16 20:55:31 i don't think there's specific options you must set when doing mkfs Jun 16 21:14:15 well, setting the nr_requests high creates a long queue with buffers pending to write out, all those buffers are allocated from RAM. But eventually they need to get written to MMC and then IO is busy for a looong time Jun 16 21:15:59 setting queue short creates lots of writes to eMMC for a small chunk of real new data each time, but the short queue doesn't hog RAM and when the queue got emptied, user space has another opportunity to swap in some program text and actually execute some code Jun 16 21:18:43 Nice sky tonight - picture courtesy of n900 http://i.imgur.com/IdLrFbJ.jpg Jun 16 21:18:44 go figure how cp fills the queue with zillions of buffers it previously malloc()ed, and while queue grows, the swapd frees up more and more RAM by swapping out program text to MMC instead of writing buffers from queue to MMC Jun 16 21:20:24 heh linux and its crazy buffering... reminds me how copying stuff to say an usb key finishes much later than the actual cp command as well Jun 16 21:20:48 actually s/program text/program data sections/ Jun 16 21:21:35 nox-: cp && sync Jun 16 21:21:40 program text just gets free'd from RAM and needs to get read in from the executable file on MMC later on when it is needed again Jun 16 21:21:46 lansiir, sure Jun 16 21:22:38 DocScrutinizer05: Well, I'll give low buffer counts a shot - what are the bricking odds when screwing around with swappolube? Jun 16 21:23:56 reducing swappiness may result in program text getting free()d later or not at all, so programs don't block kernel that much via swapd loading same program text again. echo 100 > /proc/sys/vm/swappiness might be a pretty poor idea after all Jun 16 21:24:35 swappolube does exactly the same, it readjusts swappiness and other related settings Jun 16 21:25:14 I haven't heard of swappolube bricking a device, it just migght make stuff slower instead faster Jun 16 21:25:32 Good enough for me >:D Jun 16 21:26:15 since I also have installed and activated swappolube, the above quoted settings might well not nbe in effect since ages on my device Jun 16 21:26:47 hmm Jun 16 21:26:50 IroN900:/home/user# cat /proc/sys/vm/swappiness Jun 16 21:26:51 100 Jun 16 21:27:01 That's default IIRC Jun 16 21:27:16 yes Jun 16 21:27:18 cat /sys/block/mmcblk1/queue/nr_requests Jun 16 21:27:29 and also the nr_requests is 4 Jun 16 21:27:51 Default is 128 there. Jun 16 21:28:01 I might have to check what's with swappolube Jun 16 21:28:24 anyway I hardly ever see those nasty freezes Jun 16 21:29:14 and I think I copied a few GB from eMMC to uSD a year or so ago, and it took ages but eventually finished, without freezing or rebooting my device Jun 16 21:30:08 yeah, 128. ShadowJK suggests 2048 or even 4096 iirc Jun 16 21:30:25 Also, since some knowledgable people are around, I have some questions about "sleeping". Jun 16 21:30:38 this might be ultra-fast and nice for files up to ~100MB or so Jun 16 21:30:43 then BOOOM Jun 16 21:31:00 Coming from an android phone - does the n900 slow down and pause everything when the screen goes off? Jun 16 21:31:08 (or similar) Jun 16 21:31:08 no Jun 16 21:31:22 absolutely nothing like that Jun 16 21:31:45 all apps are supposed to be 'clean' in the sense that they don't do busy loops Jun 16 21:32:09 or use usleep in a loop just to poll shit Jun 16 21:32:42 normat state of maemo is: all processes wait on IO or a long term timer event Jun 16 21:33:43 So if I won't have to keep the screen on when I'm doing stuff? (finally) Jun 16 21:33:49 if there's any process not waiting wor IO or a timer that is at least 1s (rather 10 or 60s), then that app is considered broken and shall get nuked from repositories Jun 16 21:34:35 heh Jun 16 21:35:10 lansiir: no, for example xchat is running on my N900 24/7 and never missing any highlight to my nick Jun 16 21:35:38 DocScrutinizer05: We have n900 xchat? Jun 16 21:36:05 * lansiir needs to look through the repos some more Jun 16 21:36:08 apps shall stop refreshing their display as soon as it's not visible either for not having focus or screen completely blanked Jun 16 21:36:37 this isn't particularly new for any X11 app Jun 16 21:36:52 just on maemo it's absolutely mandatory Jun 16 21:38:48 on android the OS / the kernel enforces non-rogue behaviour even of crappy processes. On maemo the processes themselves are supposed to behave Jun 16 21:40:24 the rest is unix-kernel scheduler's standard job which it is pretty good in Jun 16 21:40:58 even the kernel is no_ticks Jun 16 21:41:13 or NO_HZ or whatever the name Jun 16 21:42:34 let's but it this way: when you look at a maemo system's screen, the CPU wakes up once per minute to update the clock display. When you lock the screen then even the clock sleeps Jun 16 21:43:33 when you unlock, the clock gets an X-event to redraw the display and on that event the clock re-schedules the timer on every full minute, to keep the display updated Jun 16 21:43:45 ...that's epic. Jun 16 21:43:48 let's put... Jun 16 21:44:04 If I could just get google voice working, I'd never look back XD Jun 16 21:45:11 meh, I never looked into it, but aiui google voice is a glorified SIP-VoIP Jun 16 21:45:42 I rather use true SIP Jun 16 21:46:08 It's a lot of things, but I use it as a phone call forwarder. Jun 16 21:46:10 fwd.com, or sipgate.de, or whatever meets your taste Jun 16 21:47:05 The n900 wasn't what I was thinking about when I signed up :P Jun 16 21:47:13 it seems SIP doesn't work that great, audio-quality wise, for some N900 users. For me it works like a charm Jun 16 21:47:19 even via 3G Jun 16 21:47:50 lansiir: I'm sorry to hear that Jun 16 21:48:21 DocScrutinizer05: No data plan. Jun 16 21:48:25 <= le cheapass Jun 16 21:51:31 my dataplan is prepaid 10 bucks / month, for 500MB (more than enough usually), and when getting a 30.-- charge I get a 20.-- on top worth in free minutes and text, which usually suffices for the 3 months till next 30.-- charge Jun 16 21:52:12 (we got no inbound minutes fees here) Jun 16 21:54:10 I got this "plan" and this SIM since ~2000 (last 13 years) now Jun 16 21:55:00 of course 10 years ago data been more expensive if available at all on prepaid Jun 16 21:57:30 but those subsidized contracts with "free" phone on top and supercheap minutes that turn out to be the most expensive option available after 6 months when everybody else reduces tariffs, that never appealed to me Jun 16 21:57:49 I have $0.06/mb, or $25/mo for 250Mb "unlimited" Jun 16 21:58:04 holy shit Jun 16 21:58:11 indeed Jun 16 21:58:11 i pay 5€/mo for 100mb/day Jun 16 21:58:46 As you can tell, I did not choose it for the data rates - $0.02/min with no expiration for voice. Jun 16 21:59:06 well, "$" makes me think "no surprise, third world of mobile communication" Jun 16 21:59:15 You are correct. Jun 16 21:59:22 got a prepaid, limitless, about 10$/month but capped at 384kbps Jun 16 21:59:44 384kbps? Jun 16 21:59:52 if you pay bit more, you can max out your 3g, and almost 4g Jun 16 22:00:02 I was at 150 until just last year ._. Jun 16 22:00:50 non-prepaid data plans are bit better, but I prefer prepaid with no contract Jun 16 22:01:09 non-prepaid is about $45/mo here. Jun 16 22:01:20 Where do you people live? Jun 16 22:01:26 finland here Jun 16 22:01:40 data is cheap here, but nothing else is Jun 16 22:01:43 it's called the old Europe Jun 16 22:02:11 USA has always been bit anal with their wire/wireless internet plans Jun 16 22:02:17 :/ Jun 16 22:02:21 * lansiir is stuck heree Jun 16 22:02:22 from what I've observed Jun 16 22:02:23 we even have a 6 months dealer warranty here, by law Jun 16 22:02:26 ;-) Jun 16 22:02:39 Any chance you can mail me a sim card that would work here? Jun 16 22:02:47 in addition to the 2 years manufacturer warranty? Jun 16 22:04:17 kerio: actually all is dealer warranty, just the dealer bitches at manufacturer, and the 24 months are with you proving that the bug been since manufacturing the good Jun 16 22:04:46 first six months it's assumed it's been by manufacturer and you don't need toprove anything Jun 16 22:05:00 No? Nobody's willing to mail me a international sim? D: Jun 16 22:05:19 after 6 months the duty of proof goes from manuf to customer Jun 16 22:05:48 lansiir: alas the european SIMs all are pretty pathetic when it comes to international roaming Jun 16 22:06:17 hell they are even mere robbery when roaming intra-EU Jun 16 22:07:03 darn :/ Jun 16 22:07:15 when I travel from Germany to e.g. italy then suddenly I *do* have to pay for inbound minutes. And data is like cents/byte Jun 16 22:07:37 But this is good news - I am going to travel next year. Jun 16 22:07:46 * lansiir will be sure to pick up an international sim Jun 16 22:08:16 in fact US-SIMs are way better when it comes to international roaming Jun 16 22:08:23 orly? Jun 16 22:08:34 I heard some of your plans are cheaper abroad that in USA Jun 16 22:08:42 than* Jun 16 22:09:25 Lies? Jun 16 22:09:43 It's pretty hard to leave the country, so it doesn't surprise me. Jun 16 22:10:36 and of course there are providers like suraja, who seem to have roaming contracts with 99% of carriers worldwide. And when you're at the dead end of sahara, you use their sat Jun 16 22:11:45 calls worldwide for as little as .49$/min Jun 16 22:12:21 I think they even offer 14kBd data, for checking your email etc Jun 16 22:12:29 * Win7Mac points to http://talk.maemo.org/showthread.php?p=1352563#post1352563 Jun 16 22:12:47 spelled Thuraya i suppose? Jun 16 22:13:53 satellite? Wikipedia didn't mention satradio protocols on the n900 >_> Jun 16 22:14:53 nox-: yep Jun 16 22:15:25 well, that's N945I then, to get rolled out 2015 ;-) Jun 16 22:15:51 So, what's up with nokia these days? Jun 16 22:15:59 winphones :( Jun 16 22:16:10 This is actually a pretty good phone, but the failphones are bad. Jun 16 22:16:27 or let's wait and see what becomes of the proposed N900i-Phoenux-gta04 project Jun 16 22:16:45 Are you still buying them, or are we waiting for a n1000 kickstarter, or what :P Jun 16 22:17:17 nobody else can buy N900, since I BUY THEM ALL ;-P Jun 16 22:17:47 more than a few not enough? Jun 16 22:17:49 :P Jun 16 22:17:53 (I actually throttled down a bit when I had my fifth) Jun 16 22:17:59 haha Jun 16 22:17:59 DocScrutinizer05: there's a point where you should stop, and you clearly passed it Jun 16 22:18:04 DocScrutinizer05: Are you kidding me? D: Jun 16 22:18:42 lansiir: because they're not enough? Jun 16 22:19:44 wtf Jun 16 22:19:45 http://negrielectronics.com/nokia-n900-media-tablet-unlocked-black-p-2709.html Jun 16 22:20:15 my n900 just started making random noise Jun 16 22:20:24 haha Jun 16 22:20:28 reboot it Jun 16 22:20:30 entitled: Time to go find your bible. Jun 16 22:20:39 this also happened to me *once* Jun 16 22:20:42 yeah! Jun 16 22:21:05 started clicking from speakers - lart PA Jun 16 22:21:12 kind of like the interface sound, when you press the display, but repeated very very fast Jun 16 22:21:23 clicking yes Jun 16 22:21:25 entitled: DocScrutinizer05 will buy it from you :P Jun 16 22:21:25 DocScrutinizer05: will the N904 run a free-ish modem? Jun 16 22:21:38 It's the noise from the ghost of OSSO. Beware. Jun 16 22:21:55 kerio: sure, check out gta04.org Jun 16 22:22:02 GeneralAntilles, shall I put my answers to your "Qs for HiFo candidates" thread? Jun 16 22:22:19 DocScrutinizer05: also, we're going to call it N904, right? Jun 16 22:22:22 more free'ish than BB5 anyway Jun 16 22:22:27 I'll eBay this and charge 10 000$, "warning: haunted!" Jun 16 22:22:36 N904 sounds good Jun 16 22:23:13 entitled: you mixed that up with N950, which is ~2000 bucks Jun 16 22:23:27 yet I won't sell mine soonish Jun 16 22:24:11 kerio: "RX-52 board?" Jun 16 22:24:39 o.O Jun 16 22:24:57 GTA04-54 board Jun 16 22:25:47 probably the domesheet will be one of the most difficult to source parts Jun 16 22:25:53 Win7Mac, going to start a new thread with the finalized answers. Jun 16 22:26:03 why? Jun 16 22:26:15 already there... Jun 16 22:26:16 * DocScrutinizer05 moos at GeneralAntilles Jun 16 22:26:20 Hey-oh Jun 16 22:26:34 Win7Mac, because that thread was for discussing the questions to be asked. Jun 16 22:26:56 and now now the main part... Jun 16 22:27:07 rabulism? Jun 16 22:27:28 ...probably introduce new "phase" of thread Jun 16 22:29:33 I think the finalized answers belong on wiki, while the thread on forum is for *dialog* Jun 16 22:30:22 well, both, to reach people Jun 16 22:30:45 which reminds me I been to lazy to point out there that it's council's job to represent community on official events, not HiFo BoD Jun 16 22:30:46 Finalized /questions/ are going in a new thread. Jun 16 22:31:33 Korintenkacker! Jun 16 22:31:49 otoh, new thread are fresh threads are good threads Jun 16 22:32:06 *Korinthen~ even Jun 16 22:32:25 i don't mind Jun 16 22:33:07 at least point from old thread Jun 16 22:33:08 GeneralAntilles: start a new thread, then wait for chem|st or thedead1440 to join it with old one ;-) Jun 16 22:34:19 GeneralAntilles: I'm eagerly waiting for next MWKN :-D Jun 16 22:34:35 It'll be a largely boring issue. Jun 16 22:34:39 I think there's a number of good stories for that one Jun 16 22:34:43 yeah, it's monday here already... Jun 16 22:34:54 Don't have the energy for a great deal of editorializing tonight. Jun 16 22:35:38 largely boring unification of councils, extremely boring fact that we have 6 candidates for BoD election, and the negligance that BoD has a meeting on thursday Jun 16 22:36:44 wait, some fool suggested to get a new life to N900 case and screen by reshaping GTA04 circuit board Jun 16 22:36:45 ...and the last thing being crucial to elections Jun 16 22:37:05 DocScrutinizer05, is there anything I can actually /link/ to for the BoD meeting announcement? Jun 16 22:37:25 harr, good question Jun 16 22:37:41 Yes, the GTA04 thing I missed. Link for that? Jun 16 22:37:43 chanlog of last meeting maybe, where woody confirmed it? Jun 16 22:37:52 community ML Jun 16 22:38:02 Ah, I missed that meeting, too. Jun 16 22:38:03 bleck Jun 16 22:38:08 * GeneralAntilles expands his todo Jun 16 22:38:36 DocScrutinizer05: Seems to be working - it hasn't crashed yet :P Jun 16 22:38:41 well, this time we had a party, not a meeting ;-P Jun 16 22:38:54 Though - at 180Kb/s, it would have been faster to download the image again :/ Jun 16 22:39:48 lansiir: what you're doing? Jun 16 22:40:14 DocScrutinizer05: Moving a (non broken) debian image from my sd card to mydocs Jun 16 22:40:28 aaah Jun 16 22:40:47 I have 2Mb/s here, so I should have just downloaded it again :/ Jun 16 22:41:08 yeah, toldya it will get damn slow since it needs a flsh erase page rewrite for every 4 IO requests Jun 16 22:41:42 erase page rewrite can take up to 100ms or sth Jun 16 22:42:22 so depending on the block size of your queued IO requests, that determines your bandwidth you see Jun 16 22:43:23 DocScrutinizer05: I take it that it wouldn't be much faster if I was to download it then? :/ Jun 16 22:43:38 Where can the board meeting be attended? Jun 16 22:43:43 I assume it's another Concerto thing. Jun 16 22:44:24 if you use dd bs=50M to copy stuff from eMMC to uSD, then even 4 requests in IO queue will fsck your RAM usage and swapping to death Jun 16 22:44:50 most likely Jun 16 22:45:03 Rob supposed to announce it on hifo.com Jun 16 22:45:36 http://hildonfoundation.org/news/ zilch new news Jun 16 22:45:57 * GeneralAntilles whistles quietly. Jun 16 22:46:04 DocScrutinizer05: I just use cp. dd would probably be better then - what bs is the most appropiate, do you think? Jun 16 22:46:06 * GeneralAntilles makes excuses about not being kept in the look. Jun 16 22:46:16 s/look/loop/ Jun 16 22:47:55 you will be soon Jun 16 22:49:07 Don't remind me. :o Jun 16 22:49:08 lansiir: you could try dd if=/dev/zero of=/home/user/MyDocs/erase-me bs=2M count=50, and play a bit with bs and count to keep the bs*count size, I'd guess 2MB blocksize should yield best results Jun 16 22:49:14 maybe 4MB Jun 16 22:49:28 (with nr_requests=4) Jun 16 22:50:09 nr_requests is the amount of queued requests, right? Jun 16 22:50:34 yep Jun 16 22:50:58 What is this about fscking the RAM usage? Jun 16 22:51:22 so 4 requests * 2MB/request = 8MB ram buffer usage Jun 16 22:51:58 which means that 8MB of program RAM get swapped out to make space for the IO buffers Jun 16 22:52:59 while the system will probably work just fine with 8MB buffers, it would fail epically with 128 * 512kB Jun 16 22:53:13 for nr_requests=128 Jun 16 22:54:48 it's all a bit of gambling since those dials are basically not the correct ones to tweak Jun 16 22:56:02 if you have a damn stupid program that writes each singe byte with its own IO request, then you probably are really better off with nr_requests =8192 Jun 16 22:57:29 unless those 8192 IO requests try to write to virtually every single erase page on MMC, since then flushing the queue would mean you have to rewrite *all* erase pages Jun 16 22:58:23 if otoh they are all to same file, then odds are you get all 8192 requests/buffers flsuhed by a single erase page rewrite, that's finished in 0.1s Jun 16 23:00:28 and the RAM usage for 8192 1 byte IO requests = 8192 * (1 byte + overhead for queue and system-call), so maybe 800k to 1 or 2 MB at max Jun 16 23:01:47 there are other nice parameters to tweak, like milliseconds-till-flush-dirty or whatever it's called. Maybe those are worth tweaking as well, and you could go higher with nr_requests Jun 16 23:02:14 Did minutes from the meeting Friday-before-last get posted? Jun 16 23:02:30 err, no clue Jun 16 23:02:37 ask chem|st Jun 16 23:02:40 no Jun 16 23:03:07 you see I start getting kinda sloppy with council even Jun 16 23:03:36 got a lot of other stuff on my plate Jun 16 23:03:58 and still hanging out here Jun 16 23:04:02 ;-P Jun 16 23:05:07 I just feel like I achieved a good chunk of work got done during last 6 months, and I have to throttle everything a bit or you have to miss me completely soon Jun 16 23:06:02 Like the eye of god imploding in upon itself to infinite. Jun 16 23:08:12 hmm, maybe I missed a major class in my college degree, maybe english literature or theology? I can't really parse that Jun 16 23:08:32 It's a near-quote from a Squidbillies episode. Jun 16 23:09:01 When the sheriff gets his hair shampooed at the local salon with meth-doped shampoo. Jun 16 23:10:06 hmm, I start to get the ... mood it transports. and I think it fits Jun 16 23:11:28 like a dragster rubber-burning for weeks, instead of finally reaching the finish line Jun 16 23:15:24 One last story. Jun 16 23:15:25 while on maemo.org bug fixing and maintenance battlefront everything turned into stalingrad Jun 16 23:15:43 DocS, what's your finish line? Jun 16 23:17:27 the point where maemo.org runs like it did 2 years ago, and there are alert maintainers for every imaginable hickup, so I could have 6 weeks of holiday and every single day of them I had the cosy feeling that maemo.org definitely is running fine today just like it did yesterday and will do tomorrow Jun 16 23:18:16 DocS, how long are you maemo-admin? Jun 16 23:19:07 *for ~ Jun 16 23:19:41 Win7Mac: you should ask that in a more detailed way.. Are you talking about common people days or Doc's days? ;) Jun 16 23:20:20 hmh? Jun 16 23:20:55 common peoples days I guess Jun 16 23:21:31 ...how many years Jun 16 23:23:10 *since even Jun 16 23:28:43 err, I guess since 5 or 6 months Jun 16 23:31:23 oh, assumed you're doing this for ages already... Jun 16 23:31:51 then I really oughta get bashed for creating such a mess Jun 16 23:34:16 you're doing very well, afaik :) Jun 16 23:39:32 I think I already know the answer to this one, but do I *really* have to open up my n900 and resolder the USB port :/ Jun 16 23:39:53 Or is it all a bunch of people who worry way too much about a rare device? Jun 16 23:51:11 good night Jun 16 23:57:13 lansiir: depends Jun 16 23:58:30 if you feel with your hardware, you probably will manage to handle it in a way so either mothing bad ever happens to it or at least you notice immediately the first signs of doom pending Jun 17 00:00:53 when however once in your life you already broke a screw while fastening it, or you had a car accident no matter if your fault or somebody else's, then you might want to make sure nothing breaks with USB port of your N900. At least when you care enough about this little device to bother doing the hassle of reinforcing the USB Jun 17 00:01:53 I for one didn't reinforce mine so far Jun 17 00:02:52 buth I checked the soldering on both N900 that I actually use regularly Jun 17 00:02:58 but* Jun 17 00:03:20 and I found the soldering looked decent Jun 17 00:03:36 there might be some devices with flawed soldering of USB port Jun 17 00:04:03 and for sure there are users that simply lack the motoric skills to handle the USB plug Jun 17 00:04:24 I'm not one of them, so I decided to take the risk Jun 17 00:04:45 Alright. Jun 17 00:05:31 I'm pretty good with electronics, but if it ain't broke, don't fix it, knowwhatimean? Jun 17 00:05:47 Important to pull out the usb plug straight Jun 17 00:06:08 lansiir except usb, always fix usb Jun 17 00:06:15 dont make the same mistake i did Jun 17 00:06:39 RiD: Which was? :o Jun 17 00:06:56 ShadowJK: Already filed down most of my USB ports. Jun 17 00:07:01 not soldering down the microusb port before it was too late Jun 17 00:07:26 I'm on my 4th N900 and never broken a usb port so fa Jun 17 00:07:50 maybe because you've been switching around? :P Jun 17 00:08:22 But then I've also seen people with q month old Lumias with broken usb ports Jun 17 00:09:15 thats nothing I have seen month old computer case with broken usb port Jun 17 00:09:33 don't call me uncareful :z Jun 17 00:09:41 but breaking any port in 1 month is really bad achievement Jun 17 00:09:59 it broke after 2 years and 2 or 3 months Jun 17 00:10:45 mine was used by my father, who knew nothing about those ports and I use it already for rawly 7 mopnth Jun 17 00:11:17 for 2 months using same battery and charging from computer ONLY Jun 17 00:11:31 I even don't know where is my second charger Jun 17 00:12:52 never have been more happy with any other "smartphone" Jun 17 00:13:38 I even used to have to have SG3, but after 1 day of use I said it is crap Jun 17 00:13:58 wanna trade your sg3 for my n900? Jun 17 00:14:20 I expect smartphone to wake me up even if it is dead and battery is at edge of being of any use Jun 17 00:14:29 this SG3 never been mine Jun 17 00:14:33 I'll sell it and buy 2 n900s Jun 17 00:14:38 or 3 Jun 17 00:14:43 I borough it after my n95 have died Jun 17 00:15:00 hahaha - I would do same Jun 17 00:15:31 nah i'd just buy 1 n900 and save the money for a pc upgrade Jun 17 00:15:45 I think I would get contact with SG3 or SG4 just to sell it an buy proper phone (n900) Jun 17 00:16:14 N900 is far from a proper phone Jun 17 00:16:20 RiD: are you trying to make joke? Jun 17 00:16:26 it's one of the things it does worse :L Jun 17 00:16:47 how the hell SG3 can buy you both n900 and computer upgrade? Jun 17 00:17:16 it can't, therefore "save the money" Jun 17 00:17:23 I am thinking about upgrade, but it would be at least 300gbp Jun 17 00:17:24 therefor*? Jun 17 00:17:32 hahaha Jun 17 00:17:54