**** BEGIN LOGGING AT Fri Mar 18 02:59:57 2011 Mar 18 03:00:13 source of failure Mar 18 03:00:25 if mounted Mar 18 03:01:00 maybe .. but i dont want to imagine more failure than what currently happens Mar 18 03:01:09 ...and I bet a heli can be operated with less than 10 servos as well Mar 18 03:01:47 prolly not a chinook Mar 18 03:02:00 these are huge .. did you ever seen one in real? Mar 18 03:02:16 well, according to prophecies of experts the big bang will happen in now+12h or somesuch Mar 18 03:02:23 I hope they're wrong Mar 18 03:05:32 OTOH AIUI all those experts were thinking and talking about the reactors only. It seems though as if the SFPs are the worse problem Mar 18 03:06:26 never knew you could get a core melt where's no core at all Mar 18 03:10:00 (chinook) sure, my granny lived next to a US mini airbase, Coburg (15km to DDR) Mar 18 03:11:10 15km to DDR? mighty long traces to memory :D Mar 18 03:11:14 or whatever the facility was. Every once in a week there landed some of those dual rotor helis Mar 18 03:12:01 +++ Polizeigewerkschaft: Deutsche AKW nicht terrorsicher +++ Mar 18 03:12:01 [02.55 Uhr] Die Deutsche Polizeigewerkschaft hält die Atomkraftwerke in der Bundesrepublik für nicht sicher vor Terrorangriffen. Mar 18 03:12:10 back when those helis landed there, they prolly did because it was close to the iron curtain back when Mar 18 03:12:19 muhaha .. as if that would be the most important concern Mar 18 03:12:37 actually it might Mar 18 03:13:26 I seem to remember we had a near-GAU when some "terrorists" cut down a power pylon Mar 18 03:14:08 no more connectio to grid. SCRAM. No more power. Failing generators... Mar 18 03:14:55 also next stuxnet might blackout half Europe Mar 18 03:15:08 cyberterrorism Mar 18 03:15:48 if cars would be constructed like nuklear plants then they would be extremely dangerous ... Mar 18 03:16:23 imagine you hit the brake at 100kmh .. and at 30kmh it just stops braking and goes over rolling Mar 18 03:17:12 :nod: Mar 18 03:18:08 also losing steer servo, and never come up again. Nice when you're on a downhill road - which for a NPP you always are Mar 18 03:20:06 ah something n900 related .. the device lock button (slider?) .. is that a mechanic switch or magnetic or optical? Mar 18 03:24:39 mech Mar 18 03:25:36 and (dunno where) I seem to remember I noticed recently it bounces like hell Mar 18 03:25:53 mhm .. sometimes i wonder if mine gets weared out ... i know that the device sometimes reacts slowly on it under load .. but rarely it doesnt react at all without any load Mar 18 03:26:25 same here, but I'm not sure about the load Mar 18 03:27:36 i didnt disassembled my n900 yet .. but maybe some kontact-spray will help .. but needs disassembling Mar 18 03:27:54 OTOH mce/bme/whatnot is notorious to lose an occasional charger-plugged signal, though /sys knows there's a charger Mar 18 03:28:10 nah, do *not* do this Mar 18 03:28:10 yes Mar 18 03:28:28 contact spray ruins switches usually Mar 18 03:28:46 believe me, I ruined several :-P Mar 18 03:28:55 heh works for me .. Mar 18 03:29:06 i revived a lot mouses with it Mar 18 03:29:18 it works like for 4 weeks, then back to original and getting worse Mar 18 03:29:34 well "Kontakt 61" .. thats a special fluid for switches Mar 18 03:29:49 nah permanently (well years at least i can tell) Mar 18 03:29:54 hell, I *know* all sorts of contact sprays Mar 18 03:29:57 ~joerg Mar 18 03:29:58 i heard joerg is a HW-developer and engineer of Openmoko, usually known as DocScrutinizer Mar 18 03:30:36 yes hence i saied Kontakt 61 (from Kontakt chemie) no other :) Mar 18 03:30:55 i know that a lot others types are not good or even worse Mar 18 03:31:21 * DocScrutinizer makes mental note to get new canned air Mar 18 03:31:27 especially for switches which have some load and then using silicone based oils .. thats perfect destruction :) Mar 18 03:32:08 the problem is a lot of switches already have that silicone oil Mar 18 03:32:21 you can wash that out Mar 18 03:32:32 no you can't really Mar 18 03:32:55 you'll wash it on contact surface Mar 18 03:32:57 sufficently .. maybe not perfect Mar 18 03:33:31 well .. wash it out .. dry it and then apply the fluid i mentioned before Mar 18 03:33:37 anyway, for N900 it won't help at all Mar 18 03:33:42 that really worked for me .. always Mar 18 03:34:15 dunno .. i prolly wait until the switch fails completely before repairing it Mar 18 03:34:27 does it get worse? Mar 18 03:34:35 exactly :) Mar 18 03:34:54 hmm Mar 18 03:35:31 well for some contacts i apply it for new equipment ... Mar 18 03:35:55 basically anything which has some danger to get humid but not accumulate dust Mar 18 05:06:26 wtf... local news must be run by pure morons Mar 18 05:54:53 just got simple mobile for my nokia n900, but can't get internet to work. can anyone help? Mar 18 05:56:05 yep, your mobile provider manual Mar 18 05:57:18 you're such a wit. bet all the girls luv u Mar 18 05:58:31 You are such a nice person to say that someoen who helps you for free. You are lucky i am not an op.... Mar 18 05:59:38 we're all lucky you're not an op Mar 18 05:59:59 and that wasn't very much help at all Mar 18 06:01:11 all your questions were answered. Mar 18 06:02:20 yeah, ok. thanx Mar 18 06:19:19 hi Mar 18 06:44:47 * timeless_w7ip sighs Mar 18 06:44:52 one-tree != onetree Mar 18 06:44:53 http://www.linuxfordevices.com/c/a/News/legal-experts-address-Android-GPL-issues/ <-- hmmm Mar 18 06:48:59 How do I charge a li-ion battery manually? Mar 18 06:49:00 ruskie: some of us read /. yesterday Mar 18 06:52:33 ruskie: do note that "void main (int argc, char* argv);" should not be copyrightable Mar 18 06:52:43 agreed Mar 18 06:52:55 * timeless_w7ip frowns Mar 18 06:53:02 so i woke up this morning to discover 4 emails from cron Mar 18 06:53:13 telling me i couldn't spell the name of my script correctly Mar 18 06:53:17 rofl Mar 18 06:53:32 hahaha Mar 18 06:53:32 to be fair, i didn't write or name the script, someone else somewhere else did Mar 18 06:56:46 :) Mar 18 06:59:51 timeless_w7ip, but you do agree that all that work required someones creativity and effort to write up so they do have a copyright on it... Mar 18 07:01:56 no Mar 18 07:02:06 not like that anyway Mar 18 07:02:25 for instance, i claim that a list of words that are spelled correctly in English is not copyrightable Mar 18 07:02:51 so basically you don't consider books copyrightable? Mar 18 07:02:56 which is a list of words Mar 18 07:02:59 i don't think 'effort' is a criterion for copyright Mar 18 07:03:24 timeless_w7ip, U.S.Code Title 17 Section 117(a)(1) covers this Android case Mar 18 07:03:34 it could take someone a lot of effort to write "1, 2, 3, 4, 5, 6 ...." Mar 18 07:03:40 but that doesn't make it copyrightable Mar 18 07:06:55 ruskie: by "list of words that are spelled correctly", i mean literally a list of correctly spelled words, no repeats Mar 18 07:07:10 roughly the thing that would enable a spellchecker to decide that 'teh' is not correctly spelled Mar 18 07:11:56 you mean /usr/share/something/dict? Mar 18 07:13:20 all i have is /usr/share/bash-compeletion/dict Mar 18 07:13:23 so not that :) Mar 18 07:13:43 .. /usr/share/dict/words Mar 18 07:13:56 yeah, a file like that Mar 18 07:13:56 whatevs Mar 18 07:14:30 well you can probably find that file at some point in pi Mar 18 07:14:48 similarly, i would hope that a sorted list of books would also fail copyright Mar 18 07:15:13 the reason that both should fail is that they are not 'creative' Mar 18 07:15:15 i'd say that anything that can be turned into a number can't be subject to a copyright Mar 18 07:15:22 there is a database copyright... but I believe it has specific requirements for a work to be considered a copyrightable database Mar 18 07:15:39 ruskie: database copyright is the EU's way of making everything copyrightable Mar 18 07:15:42 and that's absolutely insane Mar 18 07:16:01 the EU deserves to be wiped off the face of the earth just for that Mar 18 07:16:32 timeless_w7ip: how are software patents working for you? Mar 18 07:17:07 kerio: much less disturbing than the eu database copyright Mar 18 07:17:13 i could take them or leave them Mar 18 07:17:28 heck, i would take them if that's what was required to get rid of the eu database copyright Mar 18 07:17:37 i'd rather *not* take them Mar 18 07:17:39 but... Mar 18 07:17:41 timeless_w7ip, russia also has it ;) Mar 18 07:18:02 on the scale of evil, they are insignificant when compared to the db c Mar 18 07:18:20 ruskie: when i get out of here, you can have some of my sympathies Mar 18 07:18:34 erm? Mar 18 07:19:10 i'm in .eu Mar 18 07:19:33 ahh Mar 18 07:19:43 if there has not been substantial investment (which need not be financial), the database will not be protected <-- so technically not everything is copyrightable Mar 18 07:19:54 database rights specifically protect the "qualitatively and/or quantitatively [a] substantial investment in either the obtaining, verification or presentation of the contents": Mar 18 07:21:19 ruskie: db c is designed to allow corporations to sue everyone else Mar 18 07:21:30 and to prevent smaller groups from claiming it Mar 18 07:21:45 if you're a single person, perhaps you can't claim "substantial investment" Mar 18 07:21:50 and thus your work is inelligible :) Mar 18 07:22:16 sure you can... inventoring all the books in my hous is a substantial investment of my time Mar 18 07:22:32 but basically between all of the or's in the clause pretty much everything is db c Mar 18 07:22:43 ruskie: and if a second person comes in and makes an inventory Mar 18 07:22:50 they're infringing on your db? Mar 18 07:22:56 that's the stuff stupidity is made of Mar 18 07:23:09 no they are infringing on my property ;) Mar 18 07:23:22 but I see what you mean Mar 18 07:23:27 they made some stupid calls there Mar 18 07:23:30 some? Mar 18 07:23:34 the entire concept is stupid Mar 18 07:23:35 sorry Mar 18 07:23:43 a work can be creative Mar 18 07:23:47 an index of a work is not creative Mar 18 07:23:58 what do you call openstreetmap then? Mar 18 07:23:59 it can be done by an indexer, whicih is not a creative process Mar 18 07:24:07 openstreetmap is a collection Mar 18 07:24:12 aka a database Mar 18 07:24:24 individual elements in the collection may be creative and thus should perhaps be elligible for copyright Mar 18 07:24:29 the collection itself should not Mar 18 07:24:34 roughly speaking... Mar 18 07:24:57 the complete works of william shakespeare should not be entitled to c Mar 18 07:25:14 but each individual work of william shakespeare should be ellligible Mar 18 07:25:20 s/lll/ll/ Mar 18 07:25:20 timeless_w7ip meant: but each individual work of william shakespeare should be elligible Mar 18 07:25:43 ruskie: the other problem w/ db c is that it has a provision for ensuring infinite copyright duration Mar 18 07:25:51 true Mar 18 07:25:56 you can touch the database every 16 years or so Mar 18 07:26:01 and the copyright will never expire Mar 18 07:26:06 imho corporations shouldn't be allowed to have copyrights or patents Mar 18 07:26:12 like a cron job? Mar 18 07:26:19 kerio: pretty much Mar 18 07:26:24 cool Mar 18 07:26:35 ^^ Mar 18 07:26:36 although on the scale of 16 years, you'll probably need to migrate the cron job more often than it runs! Mar 18 07:26:48 0 0 0 0 0 0 1 1 touch mycopyrights Mar 18 07:26:48 lol Mar 18 07:32:13 oops, time for work Mar 18 07:32:20 have fun Mar 18 07:47:54 Mar 18 08:08:47 hmm, jabber is broken on my phone for some reason Mar 18 08:09:20 it works on desktop Mar 18 08:20:50 broken how? Mar 18 08:46:05 wish i could tilt the cam on n900 Mar 18 08:46:46 take pictures while holding the device normal angle Mar 18 08:52:25 http://www.guardian.co.uk/technology/2011/mar/17/us-spy-operation-social-networks <-- fun Mar 18 08:58:11 Hm, that'll actually be good. It means that we can now blame -anyone- disagreeing with us(especially a majority) as astroturfing shills! Mar 18 08:58:16 lol Mar 18 08:58:39 heh Mar 18 08:58:55 People may also start watching IPs - There's only so many VPS hosts Mar 18 09:00:15 But, either way, it should be good that this news is out there - Take -any- message, no matter how many people agree with it, with a grain of salt. Mar 18 09:01:24 always do Mar 18 09:01:26 always have Mar 18 09:01:35 Yea, but not everyone does Mar 18 09:01:48 note: tor can do something similar though you need to do a lot of manual footwork... Mar 18 09:02:36 as for watching ips... with ipv6... I wonder how fun that will be... Mar 18 09:03:05 Mar 18 09:03:12 morning Doc Mar 18 09:03:19 moo Mar 18 09:03:33 * ruskie hands Doc some grass to chew on... Mar 18 09:03:44 ~moo Mar 18 09:03:45 * infobot mooooooooo! I am cow, hear me moo, I weigh twice as much as you. I am cow, eating grass, methane gas comes out my ass Mar 18 09:05:48 mooooooooo! I am cow, hear me moo, I weigh 15 times you do... Mar 18 09:07:32 I am cow, I eat my son, this is when BSE begun Mar 18 09:07:53 you seem bored Mar 18 09:07:56 heh Mar 18 09:08:07 Morning, all Mar 18 09:08:26 Oh, that reminds me: can anyone explain quickly why we can't use a newer kernel on the n900? Mar 18 09:08:34 I know there's a good reason, but I forget Mar 18 09:08:44 I thought it was some of the drivers Mar 18 09:08:59 Aren't they userspace drivers or something? Mar 18 09:09:24 robbiethe1st: /sys changed Mar 18 09:09:34 Oh Mar 18 09:09:58 also some libs might not feel happy with new kernel Mar 18 09:10:08 would be nice if that actually got fixed somehow Mar 18 09:10:20 *wishful thinking* Mar 18 09:13:05 well there looking at just barriing the nuke plant in sand and concrete Mar 18 09:25:44 ~spell barriing Mar 18 09:25:46 possible spellings for barriing: barring baring burring bearing Barron Bering barrings bring barraging barreling barren boring barbing barfing barging barking earring jarring marring parring tarring warring Bahrain Behring berrying Mar 18 09:27:59 burry ing Mar 18 09:28:21 bbbury Mar 18 09:28:24 bury Mar 18 09:28:36 totaly forggot how to spell that oone Mar 18 09:29:09 burying Mar 18 09:33:01 rofl Mar 18 09:33:50 ~spank Mar 18 09:34:01 ~moo Mar 18 09:34:02 * infobot mooooooooo! I am cow, hear me moo, I weigh twice as much as you. I am cow, eating grass, methane gas comes out my ass Mar 18 09:34:54 hello all Mar 18 09:34:55 \o Mar 18 09:37:43 heya Mar 18 09:54:44 ~spank tybollt Mar 18 09:54:45 * infobot bends tybollt over his knee and tatoos 'ibot' on tybollt's pasty white buttocks. Mar 18 10:05:53 is there a maemo zoneminder client Mar 18 10:08:02 lol Mar 18 10:10:07 DocScrutinizer: :D Mar 18 10:25:54 find . -iname "*.mp3" -mtime -4 -exec mplayer \-shuffle \-quiet {} + Mar 18 10:26:04 all you need is $ Mar 18 10:26:14 all you need is $, $, $. $ is all you need Mar 18 10:26:27 erm... Mar 18 10:26:40 [beatles] Mar 18 10:27:34 all you need is ❤ Mar 18 10:38:27 Out of curiosity, where does the n900 store all my IM accounts? Mar 18 10:38:42 And the history database of communication events? Mar 18 10:45:52 Anyone know if QtMobility contacts works on maemo? Mar 18 10:46:12 yacc: history is in .rtcomm Mar 18 10:46:20 Uhm was there a way to extend ovimaps to say where to go along? Mar 18 10:46:34 nope Mar 18 10:46:47 yacc: accounts: ~/.rtcomm-accounts Mar 18 10:47:03 yacc: history: ~/.rtcomm-eventlogger Mar 18 10:47:04 hm shit is there anyway to use it like a navigfational device? Mar 18 10:47:32 marble Mar 18 10:47:36 maep Mar 18 10:47:40 There's a thread on t.m.o about modding ovi maps to do turn based navigation actually. Mar 18 10:47:52 that too Mar 18 10:47:58 alterego: link? Mar 18 10:48:04 \o morning Mar 18 10:48:10 moo lcuk Mar 18 10:48:19 xxtjaxx: search? Mar 18 10:48:26 Ik don't have a link .. Mar 18 10:48:45 search in HAM for map is easy Mar 18 10:51:18 err s/maep/mappero/ Mar 18 10:51:33 :P Mar 18 10:53:03 all this talk of maps Mar 18 10:53:17 DocScrutinizer, NokiaCareers have just posted an opening Mar 18 10:53:26 * lcuk gets it Mar 18 10:53:32 RES00000109 #jobs at #nokia Visual/Animation Designer (web: maps.ovi.com) http://bit.ly/gk4OuQ Mar 18 10:53:32 http://twitter.com/NokiaCareers/status/48687476373598208 Mar 18 10:53:32 @NokiaCareers Re: RES00000109 Connect with Damian Waradzyn, author of Maemo CloudGPS application. http://bit.ly/eSBVjO Mar 18 10:53:32 http://twitter.com/lcuk/status/48689865394626560 Mar 18 10:53:48 err crap, meant to only paste the top line Mar 18 10:54:09 hehe Mar 18 10:54:11 well, anim/visual is your domain :-D Mar 18 10:54:26 DocScrutinizer, see the link I posted to Damians work Mar 18 10:54:41 he has really nice slick ux :) Mar 18 10:54:54 especially regarding rotation Mar 18 10:55:09 the map gently rotates and the UI slides out and back in with ease Mar 18 10:55:17 it is amazingly simple and effective Mar 18 10:55:41 about 1:48 in the youtube video Mar 18 11:04:46 I was implementing something similar for columbus 2 :) Mar 18 11:06:23 alterego, sure, get cloudgps and see if you can reproduce it as a standard component Mar 18 11:11:52 hmmm marble aint bad reminds me of what was that app on the 770 open street? Mar 18 11:25:03 xxtjaxx: on extending maps with voice: http://talk.maemo.org/showthread.php?t=66985 Mar 18 11:29:23 jhb: thanks! Mar 18 12:29:03 Anyone knows application for drawing tracks with waypoints for N900? It would be usable to have something like Garmin eTrex in N900 Mar 18 12:36:28 OkropNick: Columbus draws live tracks, but I've yet to publish the next version that can import. Mar 18 12:37:04 alterego: thank you for reply Mar 18 12:38:04 np Mar 18 12:46:44 alterego: if you add there import/export functionality - it will be perfect! Mar 18 12:47:31 Well it already exports, nmea data. Mar 18 12:47:45 OkropNick: I'll make a release in the next week :) Mar 18 12:49:15 alterego: yes, you're right. It has one of best interfaces from all applications that I've seen for N900 - realy good job :) Mar 18 12:50:12 It's sad that I'm just poor PHP coder - I have to change it finally :| Mar 18 12:50:14 OkropNick: thank you very much :) Mar 18 12:50:31 Heh, web apps are the future apparently Mar 18 12:50:45 good for me :) Mar 18 12:50:58 Well, I do web dev too :P Mar 18 12:51:10 :) Mar 18 13:14:49 ruskie, ping Mar 18 13:25:58 has anyone tried to set up mutt @maemo (on the n900)? i don't find getmail+procmail pkgs which i need for my pop3-mailaccount (but i think about changing the mailservice) Mar 18 13:54:08 * lcuk tickled by this thread: http://talk.maemo.org/showthread.php?t=71158 Mar 18 13:57:51 Yeah, me too :d Mar 18 13:58:26 Why is Apple seemingly the only company capable of putting computers together with USB ports that don't look like they came out of the cheapest factory in China. . . . Mar 18 13:59:09 lol lcuk Mar 18 13:59:22 lcuk, some of them are going all CSI in proving "NOT SHOOPED!!!!11" Mar 18 13:59:35 yeah, hence the amusement Mar 18 14:02:04 tmo is seemingly becoming a new 4chan board. Mar 18 14:02:15 Hello Mar 18 14:02:16 (meaning: tmo is improving) Mar 18 14:03:33 someone know how notification in small red icon with a small number to notify the user that new mail is incoming in mail.app ? Mar 18 14:05:03 (same principe on some iphone app icons to signal new things is available) Mar 18 14:05:37 khertan, overlay text onto icon and generate image Mar 18 14:05:39 then update icon Mar 18 14:06:41 ApplicationBadge Mar 18 14:07:10 crashanddie: i know how to do ... i was looking for the name :) Mar 18 14:07:50 khertan, Adium should have the code in their source. Mar 18 14:07:52 thx anywas Mar 18 14:07:55 thx anyway Mar 18 14:13:38 Can someone explain what Harmattan actually is? Is it just meego-on-n900? Is it bits of Maemo and bits of MeeGo merged into one? Mar 18 14:14:11 can anyone recommend me a cheap or free+secure(no googlish)+imap webmailservice?did use inbox.com+gmail Mar 18 14:15:01 jonwil: who said harmattan would run on n900? Mar 18 14:15:18 Corsac: Who said harmattan would run ? Mar 18 14:15:29 I guess I must have been reading too much into comments on blogs etc Mar 18 14:15:37 that too :) Mar 18 14:15:41 :) Mar 18 14:15:56 Trusting Engadget for reliable news about anything in the noila-linux ecosystem is probably not a good idea Mar 18 14:16:04 nokia Mar 18 14:17:39 engadget is "zomgapplereleasedsomethingzomg" Mar 18 14:17:40 jonwil: Trusting Engadget for reliable news about anything is not a good idea Mar 18 14:19:50 jonwil, Harmattan is Maemo 6 base OS plus a Qt-based UI similar to MeeGo Handset, also based on MTF (aka. DUI) Mar 18 14:20:50 What makes maemo6 different from maemo5 then (other than what device it may run on)? Mar 18 14:21:13 jonwil: transitional to meego 1 Mar 18 14:22:07 jonwil, not much. it still uses Debian packaging with updated stuff. ask timeless or thiago about details. they may tell you more Mar 18 14:22:41 in any case it sounds like none of Maemo6/Harmattan/MeeGo/etc work is going to help those of us with N900s running Maemo who just want better software and a few features that other phones have had for years but the N900 does not Mar 18 14:22:53 jonwil, we don't really know more Mar 18 14:22:56 jonwil: some of the backend structure is changed, as venemo said still .deb but core parts of meego integrated Mar 18 14:23:04 different kind of trackerd Mar 18 14:23:29 jonwil: reading what features? Mar 18 14:24:06 Cell Broadcast SMS support would be nice for those of us with carriers who send such SMSs (mine does) Mar 18 14:24:08 and n900 is the development platform Mar 18 14:24:25 my previous phone couldn't do that... Mar 18 14:24:56 My previous Motorola could do it. As could my mums old B&W screen Nokia dumbphone Mar 18 14:25:04 all other phones I know cannot do alot of stuff you can do with n900s Mar 18 14:25:21 chem|st ++ Mar 18 14:25:26 kerio ++ Mar 18 14:25:34 yeah the N900 is a great phone Mar 18 14:25:51 jonwil: actualy it is exactly what I want Mar 18 14:26:32 chem|st ++ Mar 18 14:26:38 the only thing I would add is a compass (for hardware) Mar 18 14:26:46 for me it s ll be a mix between the n900 and the n810 Mar 18 14:26:47 :) Mar 18 14:27:17 khertan: what does the n810 have the n900 doesn't? (had no 810...) Mar 18 14:27:38 chem|st, bigger screen, bigger keyboard Mar 18 14:28:34 Be nice if I had more control over which networks (WiFi and cellular) that my N900 connected to Mar 18 14:28:38 chem|st, better quality build materials, some hw buttons that the N900 doesn't have, and handwriting recognition (also not in N900) Mar 18 14:29:19 Venemo: sure a n910 would do great... Mar 18 14:29:25 * jonwil wonders if M6/Harmattan is going to be using ofono or the old crap cell stack Mar 18 14:29:50 jonwil, my guess would be any random closed-source shit Mar 18 14:29:59 chem|st: larger screen !!! Mar 18 14:30:09 chem|st: hardware button ! Mar 18 14:30:20 and a pad Mar 18 14:30:32 khertan: if any nokian wants my design I could make many CMs happy^^ Mar 18 14:30:49 CMs ? Mar 18 14:30:57 communitymembers Mar 18 14:30:59 :) Mar 18 14:31:34 khertan: the n900 is a waste of space Mar 18 14:31:51 chem|st, waste of space? Mar 18 14:32:15 it looks like they used the prebuild devel PCB and just put another shell around.... Mar 18 14:32:40 PCB? Mar 18 14:32:58 the hardware... Mar 18 14:33:25 mhm Mar 18 14:33:39 where did you see waste of space Mar 18 14:33:50 what will you change ? Mar 18 14:34:24 there is at least space for another 64gb memory and a compass Mar 18 14:34:47 where too ? Mar 18 14:34:51 s/too/to ? Mar 18 14:35:03 just talking sizes... Mar 18 14:35:33 has someone pictures handy`? Mar 18 14:36:03 http://people.openmoko.org/joerg/n900/ Mar 18 14:36:37 yeah but where will you put the compass ? Mar 18 14:36:46 i didn't see any freespace Mar 18 14:38:15 there is a spacer component in n900 (structural probably, but it is certainly removable and not actually part of the chassis Mar 18 14:38:34 There is huge amounts of free space. Mar 18 14:38:45 Compass is ~4mm*4mm*1mm Mar 18 14:39:20 khertan: frontside I fit 2-3 backside 1-2 Mar 18 14:39:34 'waste' of space - there needs to be room to allow stuff to flex Mar 18 14:39:51 SpeedEvil: hum ... i think it was larger ... Mar 18 14:40:20 SpeedEvil: sure but you don't need space like 2x4cm for flexibility Mar 18 14:40:57 http://www.sparkfun.com/products/9441 Mar 18 14:41:01 khertan: Mar 18 14:41:01 khertan, it is not a Jack Sparrow Compass! http://images.wikia.com/pirates/images/f/fa/Piratecompass600.jpg Mar 18 14:41:04 Yarrr! Mar 18 14:41:23 and replace the cheap material and you safe a lot more Mar 18 14:41:27 oh ... dude ... i think it was ! Mar 18 14:42:20 this is a reliable : http://www.google.fr/imgres?imgurl=http://img.alibaba.com/photo/249621099/Marine_Magnetic_Compass.jpg&imgrefurl=http://french.alibaba.com/product-gs/marine-magnetic-compass-249621099.html&usg=__MBmDk0lx3lcDMNOryY9qbU9XgDU=&h=486&w=498&sz=16&hl=fr&start=0&sig2=S_4ZFwz79XQeJbRStMif-Q&zoom=1&tbnid=0ZlorPKYLtdkeM:&tbnh=143&tbnw=147&ei=kW6DTaGWDciyhAfzn6C-BA&prev=/images%3Fq%3Dmagnetic%2Bcompass%26um%3D1%26hl% Mar 18 14:42:21 3Dfr%26biw%3D1680%26bih%3D917%26tbs%3Disch:10%2C426&um=1&itbs=1&iact=rc&dur=542&oei=kW6DTaGWDciyhAfzn6C-BA&page=1&ndsp=45&ved=1t:429,r:42,s:0&tx=118&ty=24&biw=1680&bih=917 Mar 18 14:42:28 ouche sorry for the link Mar 18 14:42:39 http://img.alibaba.com/photo/249621099/Marine_Magnetic_Compass.jpg Mar 18 14:42:46 ^^ a real reliable compass Mar 18 14:43:31 khertan, is that a compass in your pocket or are you just pleased to see me? Mar 18 14:43:56 a design where you use needed parts as connection between housing materials is not common yet but used as antenna for iphone4 for example Mar 18 14:44:32 lcuk: :) Mar 18 14:45:16 about that ... chem|st where is the antenna on this picture ? there isn't ? Mar 18 14:45:23 and the wifi antenna ? Mar 18 14:46:50 the antennas are the bottom and the top edge of your back Mar 18 14:48:21 well side bottom and top of the devices back, on the pics only the top one is disected Mar 18 14:49:29 thx Mar 18 14:50:17 gtg bye gents Mar 18 15:02:55 "We have recently received a Forum Nokia Launchpad membership application from you and would like to thank you for your interest towards our developer programs. Since Nokia’s new strategy and the alliance with Microsoft was announced in February, we have been flooded with membership applications from developers interested in Launchpad membership. Due to the vast number of applications, the review times are considerably longer than usually. Mar 18 15:06:05 lol lolcat Mar 18 15:15:13 lolcat, wtf is that? Mar 18 15:21:31 a quote. Mar 18 15:23:38 lolcat, bullshit. Mar 18 15:25:01 GAN900: Bullshit? Mar 18 15:25:50 Venemo: Nokia sent me an answer for my developer account or whatever, I read an article about them giving away phones Mar 18 15:26:30 lolcat: did they say no? :) Mar 18 15:26:37 They said that Mar 18 15:26:46 So I guess I will have to wait Mar 18 15:28:25 lolcat, I smell marketing at work. Mar 18 15:32:00 I applied for that ages ago Mar 18 15:32:04 I never got anything :/ Mar 18 15:35:57 Oh hang on Mar 18 15:36:06 I just got an email from 'forumnokialaunchpad' Mar 18 15:37:08 Pfft, I appleid for that months ago Mar 18 15:37:22 I'd actually completely forgotten until now :D Mar 18 15:42:14 Did you get approved? Mar 18 16:36:32 lol many developper ... mouarf Mar 18 17:03:21 http://www.youtube.com/watch?v=eaScyfSHc-Y&NR=1 All that is old is topical again. Mar 18 17:05:56 whaaaat Mar 18 17:06:09 nokia will release maemo 6 phone? :o Mar 18 17:07:37 Jartza, IMO Nokia calls it "MeeGo-based", some Intel folks call it "Maemo 6", and we will all probably see in the end how meegoish it is Mar 18 17:13:53 wth is error: void value not ignored as it ought to be Mar 18 17:15:29 MeeGo N Mar 18 17:15:51 andre__, Intel folks calling it Maemo 6 is just cruel. Mar 18 17:16:21 oh nvm Mar 18 17:16:26 * MohammadAG facepalms Mar 18 18:35:41 anyone knows where N900 stores sugestions for word autocompletion? (using onscreen keyboard) Mar 18 18:45:13 * slonopotamus wonders who are all these 409 people Mar 18 18:45:49 and why they hang here Mar 18 18:46:57 slonopotamus: i have an n900, and i program for it! Mar 18 18:47:10 seems like a good enough reason to me Mar 18 18:47:26 * slonopotamus programs ON his :P Mar 18 18:48:06 slonopotamus, what do you write though! Mar 18 18:48:49 was hacking sflphone recently Mar 18 18:49:38 slonopotamus, was build/compile time an issue? Mar 18 18:49:50 nope Mar 18 18:50:06 n900 cpu is nice, especially @1ghz Mar 18 18:50:17 hah Mar 18 18:50:34 * lcuk just doing MeeGo comparisons of build time Mar 18 18:50:52 and don't start that shit about lifetime shorteting Mar 18 18:50:54 slonopotamus: :< Mar 18 18:50:54 and so far Maemo is taking a whole extra minute to build stuff Mar 18 18:50:57 on same machine Mar 18 18:51:13 s/ting/ning/ Mar 18 18:51:14 slonopotamus meant: and don't start that shit about lifetime shortening Mar 18 18:51:17 slonopotamus, I am just trying to compare/contrast the write speed things Mar 18 18:52:15 dunno about n900, n8x0 kernel had explicit delays to slow down mmc accesses Mar 18 18:54:42 those were optimisations in waiting. Mar 18 18:55:40 grr, can't figure out why a window's causing my mediaplayer to segfault Mar 18 19:02:15 segfault :D Mar 18 19:02:23 the most informative error that exists Mar 18 19:02:38 "blablaprogramoutputSegmentation fault." Mar 18 19:07:32 gdb MohammadAG Mar 18 19:07:34 r Mar 18 19:07:40 lcuk, not that helpful Mar 18 19:07:45 bt Mar 18 19:07:51 why? Mar 18 19:08:30 lcuk, http://pastebin.com/raw.php?i=H8NciNYb Mar 18 19:08:50 I'm not creating any layouts outside designer Mar 18 19:09:21 MohammadAG, what was the app doing at the time? Mar 18 19:09:32 nothing, the window is just there Mar 18 19:09:38 it's not connected to any signals Mar 18 19:10:06 actually, it is, but they don't have to do with layouts Mar 18 19:11:35 http://gitorious.org/qt-mediaplayer/mediaplayer/blobs/master/radionowplayingwindow.cpp#line19 Mar 18 19:14:42 MohammadAG, it is happening when you close it? Mar 18 19:17:09 lcuk, yes Mar 18 19:17:18 in the destructor, obviously :) Mar 18 19:18:11 so it is not just sitting there Mar 18 19:18:51 the destructor's only deleting the UI Mar 18 19:18:59 which is done with Qt Designer Mar 18 19:19:09 so I'm guessing this is a bug in it Mar 18 19:20:21 what about the timer you started? Mar 18 19:22:42 that's only fired when you click the volume button Mar 18 19:23:04 I'm not touching that, the window is 3 stacked windows deep and the mediaplayer segfaults right when you open -> close it Mar 18 19:25:25 MohammadAG, then don't close it. Mar 18 19:27:18 lcuk, lol Mar 18 19:34:57 slonopotamus, Mar 18 19:35:17 from what I can gather, MeeGo filesystem is quite substantially faster than same machine running Maemo (at least for the tests I just ran) Mar 18 19:35:41 running "time make" on same codebase in Maemo and MeeGo o_O: Mar 18 19:35:42 MeeGo: Mar 18 19:35:42 make[1]: Leaving directory `/home/meego/svn_tab/libliqbase/src' Mar 18 19:35:42 real 2m43.337s Mar 18 19:35:42 Maemo: Mar 18 19:35:42 make[1]: Leaving directory `/home/user/svn_tab/libliqbase/src' Mar 18 19:35:44 real 4m 28.92s Mar 18 19:42:12 lcuk: is that on flash or harddrive? Mar 18 19:42:43 (looks like same magnitude of difference as I saw in comparing btrfs vs ext3 on flash) Mar 18 19:43:00 ShadowJK, that is on my n900 Mar 18 19:43:13 lcuk, same media? Mar 18 19:43:16 and using different places on the filesystem Mar 18 19:43:28 MohammadAG, MeeGo is on a C2 4gb external card Mar 18 19:43:38 the maemo results are on different places on its own filesystem Mar 18 19:44:00 but location within maemo did not appear to make a difference Mar 18 19:44:15 Yeah it makes sense. iirc there was only 2X difference between btrfs and ext3 when I benchmarked it a year ago Mar 18 19:44:49 what brand of card? Mar 18 19:46:26 can tell you in a minute Mar 18 19:46:29 just checking something else Mar 18 19:48:34 btw if Maemo is running from internal Nand, with swap and your working directory on emmc, it would have an advantage vs MeeGo on microsd. However, if MeeGo on microsd is using swap on emmc, I would say MeeGo has a benefit especially if compiling C++ stuff Mar 18 19:49:35 Just saying that the spreading of IO load across the 3 possible devices can have bigger effect than the speeds of the underlying media or speed of filesystem Mar 18 19:49:48 hi Mar 18 19:50:03 ShadowJK: MMC raid array? =) Mar 18 19:50:31 lol Mar 18 19:51:04 does anyone know where openssh puts logs on the N900? /var/log/ is pretty much empty Mar 18 19:52:05 RST38h, probably harmful to performance :-) Mar 18 19:52:17 puzzled, install syslogd (or searcg for syslog) from sdk-tools repo, reboot Mar 18 19:52:47 ShadowJK: thanks Mar 18 19:56:04 fixed the bastard Mar 18 20:03:30 ShadowJK: Ok, Raid0 Mar 18 20:03:59 ShadowJK, MeeGo uses the external MMC card for everything Mar 18 20:05:48 RST38h, probably harmful to performance :-) Mar 18 20:06:00 lcuk, I have a hard time construdtiong a summary of whether IO loads were equally distributed in both test cases :-) Mar 18 20:07:47 ShadowJK, do you know any real benchmark tools which can give real results? Mar 18 20:07:58 because ali1234 has similar questions Mar 18 20:08:09 * lcuk was just curious to see this case Mar 18 20:10:44 compiling isn't a very bad test Mar 18 20:12:57 but running a benchmark on a filesystem+drive while the OS has its files and/or swap there will make the results different Mar 18 20:13:47 sometimes much different :-) Mar 18 20:16:16 GOD DAMMIT Mar 18 20:16:20 i need some help Mar 18 20:16:45 backupmenu Mar 18 20:16:56 has an SSH and a serial console function Mar 18 20:17:03 how the fuck do I use them Mar 18 20:17:19 whenever I use serial console I get the error: Mar 18 20:18:01 11 Mar 18 20:19:03 does the ovi map loader from symbian also load maps onto the n900? Mar 18 20:19:46 login: cannot run /bin/bash: No such file or directory Mar 18 20:20:06 whenever I try USB SSH it says password incorrect Mar 18 20:20:13 WTF am I missing Mar 18 20:20:19 How can I access rootfs? Mar 18 20:20:28 please for the love of god somone help me Mar 18 20:21:28 somone Mar 18 20:21:31 anyone Mar 18 20:21:32 btw, while my own tests showed btrfs as twice as fast as ext3, nilfs2 was twice as fast as btrfs :) Mar 18 20:21:41 I am at the edge here Mar 18 20:22:07 can somone please answer me? Mar 18 20:22:12 is nilfs2 write-only? Mar 18 20:22:12 but nilfs2 is a bit too batteryhungry. Mar 18 20:22:24 RST38h, no? Mar 18 20:23:02 ok, the name is weird =) Mar 18 20:24:06 so no one knows huh? Mar 18 20:25:48 please dont feel pressured into answering a bro in distress here Mar 18 20:26:11 vi____, no idea, but if this thing worked at some point, but no longer works, I'd assume something happened in between, and you telling us what that was could help Mar 18 20:27:09 if this "backupmenu" thing relies on sh and sshd from rootfs and merely provides a convenient way to bypass the usual startup procedure and directly launch sh on serial port, or sshd on network, then I'd say messing up the rootfs would make those things stop work Mar 18 20:27:21 but I don't really know how backupmenu is built Mar 18 20:29:53 preinit script Mar 18 20:32:01 right so, damaged rootfs can disable functionality of backupmenu Mar 18 20:33:01 backupmenu is not broken Mar 18 20:33:19 there is a corrupt file that is preventing the phone from booting Mar 18 20:33:46 I should be able to just ssh into the phone but neither console or SSH FUCKINGWELL WORK Mar 18 20:34:08 and what file is this? Mar 18 20:34:25 etc/systemui.xml Mar 18 20:34:58 right, save yourself the time and kill the watchdog Mar 18 20:35:00 hmm disconnected the n900 @06:30 in the morning... and have been playing music non-stop until 21:00 and of that atleast 4 hours was running through the fm transmiter... and it was in the red when I plugged it back in... Mar 18 20:35:20 why? Mar 18 20:35:24 what will this do? Mar 18 20:35:29 flasher-3.5 --enable-rd-mode --set-rd-flags=no-lifeguard-reset Mar 18 20:35:54 fix the file, then flasher-3.5 --disable-rd-mode Mar 18 20:36:08 kill the watchdog Mar 18 20:36:17 but the phone wont even boot Mar 18 20:36:28 phone will no longer reboot if a critical process dies Mar 18 20:36:48 (you know, how all linux systems handle dead processes) Mar 18 20:37:08 It doesnt matter now Mar 18 20:37:15 well, if it's only systemui, then it will if you kill the watchdog Mar 18 20:37:33 I made a backup with BUmenu Mar 18 20:37:39 stop being stubborn and do what I said :P Mar 18 20:37:41 and edited the offending file Mar 18 20:37:48 then tried to restore Mar 18 20:38:00 it freaked out with a tar 2 error Mar 18 20:38:12 now the thing is fucked and wont even go to BUmenu Mar 18 20:38:17 FFS Mar 18 20:38:54 now i have to reflash Mar 18 20:38:55 FFS Mar 18 20:38:57 FFS Mar 18 20:38:58 FFS Mar 18 20:39:17 no need to spam .. Mar 18 20:39:27 dude Mar 18 20:39:48 stop spamming Mar 18 20:39:52 meh Mar 18 20:39:56 it's always funny when people mess around with their phones kernel and then get all worked up when the need to reflash Mar 18 20:40:07 STFU man Mar 18 20:40:16 twho even mentioned the kernel Mar 18 20:40:31 either your phone is a plaything to experiment on or it's something you want to stay working Mar 18 20:40:55 well, I hardly thought an XML error would cripple the thing Mar 18 20:41:00 Remember the full tar error? Mar 18 20:41:03 well, I gave you a solution, you decided to ignore it and fuck with the filesystem Mar 18 20:41:06 good luck Mar 18 20:41:09 heh Mar 18 20:41:14 I appreciate your help Mar 18 20:41:21 I did it before you suggested it Mar 18 20:41:27 thank you anyway Mar 18 20:42:59 vi____, instead of writing all caps, can you try remember what the tar error was? Mar 18 20:45:18 hey Mar 18 20:45:40 moo. Mar 18 20:47:57 RST38h: have u seen impad 2 ? =) Mar 18 20:47:58 lol Mar 18 20:48:42 nah Mar 18 20:50:03 dsme is retarded tbh Mar 18 20:50:21 it shouldn't reboot if it already did 3+ times Mar 18 20:50:36 dsme? Mar 18 20:50:40 what is dsme? Mar 18 20:51:49 software watchdog Mar 18 20:52:20 the one that announce u about updates? Mar 18 20:52:43 while true; do killall hildon-desktop && sleep 1; done, watch the device reboot in less than 10 secs Mar 18 20:52:47 MohammadAG: is icons for python apps bug was reported? Mar 18 20:52:52 (unless you killed the watchdog) Mar 18 20:53:17 um, no Mar 18 20:53:21 MohammadAG: for cssu Mar 18 20:53:27 the one that reboots the device when a system app dies Mar 18 20:53:32 nope Mar 18 20:53:34 what bug Mar 18 20:53:44 err, start over Mar 18 20:53:46 the one that apps don't start from desktop icons Mar 18 20:53:59 after last updates thye stop working Mar 18 20:54:05 thought it was app releated Mar 18 20:54:19 but as ther are several apps Mar 18 20:54:21 DrLaunch or normal shortcuts? Mar 18 20:54:26 normal Mar 18 20:54:30 even from dash Mar 18 20:54:39 and from terminal it works fine Mar 18 20:54:45 any python app? Mar 18 20:55:04 papertoss, bouncebounce Mar 18 20:55:25 and now i remember somebody had trouble with heroes port Mar 18 20:55:26 menu launches bouncebounce fine for me Mar 18 20:55:40 it doesn't for me and some other guys Mar 18 20:55:51 it launches and crashes Mar 18 20:56:11 can't get info as from teminal it does work Mar 18 20:56:12 so does the desktop icon Mar 18 20:56:17 hmm Mar 18 20:56:28 wait i'll link you Mar 18 20:56:47 basically, it seems to be user specific Mar 18 21:05:54 hi all, what do people mean here by 'reflashing' they mean the kernel? or wiping out the system? http://talk.maemo.org/showthread.php?t=32408&page=2 Mar 18 21:08:21 oh N810, how awesome you are: After my reflash, it is fine now. 10 hours later, it still reads 10 days standby, 7 hours of use. Mar 18 21:08:34 GuySoft, I would take that as an OS reflash Mar 18 21:08:39 not just the kernel Mar 18 21:09:05 lcuk, and how would i restore it all back? it would be a pain Mar 18 21:09:26 that is not the question you asked Mar 18 21:09:42 so far you have not indicated you need to do anything Mar 18 21:10:55 lcuk, i have problems with power drain Mar 18 21:11:17 lasts around 5 hours Mar 18 21:11:28 then check what is running Mar 18 21:14:00 finally Mar 18 21:14:17 MohammadAG: i was on phone and than webchat lagged Mar 18 21:22:42 MohammadAG: http://talk.maemo.org/showthread.php?t=70962&highlight=bouncebounce&page=3 from second post and on Mar 18 21:22:48 lcuk, i did, nothing takes the cpu, except maemo-launcher, and sometimes browserd (even though i didnt open a browser since the devices was started) Mar 18 21:22:56 also as i understand papertoss has same issues Mar 18 21:23:13 lcukn900: hey Mar 18 21:23:53 lcuk, you can seehttp://nopaste.dk/p3012 Mar 18 21:24:10 ZogG_w__, #30 doesn't have the CSSU, and is having the problem Mar 18 21:24:19 this is right after boot Mar 18 21:24:22 so please, don't blame the CSSU :P Mar 18 21:24:50 MohammadAG: i don't blame, where did i. i just thought it was cssu releated =) Mar 18 21:25:00 as i saw several people having same thing Mar 18 21:25:09 probably somthing broke it during updates Mar 18 21:25:22 some python package Mar 18 21:25:46 MohammadAG: do u have profileX thing? Mar 18 21:31:02 what the F Mar 18 21:31:13 just connected second monitor to laptop Mar 18 21:31:39 and using mplayer on fullscreen just get half picture on one and other half on other =) Mar 18 21:32:49 ZogG_w__, no, I'm quite happy with 2 profiles :P Mar 18 21:33:24 ok Mar 18 21:58:49 GuySoft, do you use gps and stuff? Mar 18 22:01:06 ZogG_w__, N900 DE is the open source aim at the moment: http://wiki.meego.com/ARM/N900/DeveloperEdition Mar 18 22:01:14 monitor its progress and get involved :) Mar 18 22:01:16 lcuk, no Mar 18 22:01:38 lcuk, i think something is draining it Mar 18 22:20:25 GuySoft: then reflash, and DO NOT restore all those fancy little applets and whatnot, that some of them are known to keep system busy. Deactivate skype and IM as well Mar 18 22:23:42 GuySoft: worst case you have a perfectly working power-humble system that 'just' is logged in to a rogue fubar 3G data carrier that forwards random pings originating around the world to the IP addr of your GSM interface. This will keep even a completely idle system in a state of permanent TX similar to a never ending phonecall Mar 18 22:26:13 GuySoft: Another option would be to install "powertop" and see which application is waking up the CPU Mar 18 22:26:37 DocScrutinizer, its a n810, its like a phone without the evil Mar 18 22:27:04 ohh, so no idea then Mar 18 22:27:26 MohammadAG: did u think of submiting maemo-cssu to gcos '11 ? =) Mar 18 22:27:26 powertop is your best bet then Mar 18 22:27:44 if that's even available for diablo Mar 18 22:28:14 DocScrutinizer, i am not sure, let me look Mar 18 22:28:35 DocScrutinizer, GuySoft: I think it's only available for fremantle / N900 Mar 18 22:30:00 E: Couldn't find package powertop :( Mar 18 22:31:05 well, for N810 I think there are hidden and not so hidden power hogs as well: CPU freq governor policy (keep CPU at high clock all the time, profile "power" or "performance" or sth), vboost providing 5V to USB for nuttin, frontcam enabled, ...??? Mar 18 22:32:30 GuySoft, do you have it on wifi? is powersaving at max? Mar 18 22:32:41 i found a package of powertop Mar 18 22:32:53 It doesn't work on N810's kernel Mar 18 22:32:55 ShadowJK, i have wifi on yes Mar 18 22:34:21 check wifi advanced settings, powersaving Mar 18 22:34:40 it's a HUGE hog when not at max Mar 18 22:35:05 even at max it may still fail on particular APs Mar 18 22:37:42 Though if this is a sudden problem, and the AP hasn't been changed or touched, it's probably not that it's crap at max :) Mar 18 22:55:42 yeah IF Mar 18 22:56:24 debugging powerdrain in a system is basically partially dirt simple, partially it's mere voodoo Mar 18 22:57:20 and for sure it's virtually impossible with a "did nuttin, now borked, HALLP!" report Mar 18 22:58:20 I.E "more info needed" Mar 18 23:01:43 If it's truly a "did nuttin, now borked" case, then I'd blame the battery Mar 18 23:06:58 sure, *usually* it's hw then (except for filled up root due to growing syslog, automatic updates fsckng up sth, etc) Mar 18 23:09:59 but 99.5% are of the class "*I* did nothing! ... !!! ...! ... ... ... Sure, HAM installed that update of foobar, but I don't even use foobar" Mar 18 23:14:05 or "I didn't even touch the device! Definitely! ... aaah, yes, my neighbour got a new DSL. Why is that relevant? ... Yes I use my neighbour's AP, but still I didn't do any changes to my device" Mar 18 23:25:16 wow Mar 18 23:25:28 skype/telepathy is such a battery whore Mar 18 23:25:56 it's skype Mar 18 23:26:05 but it has dbus! Mar 18 23:26:22 telepathy (eg XMPP, and SIP) hooks into the device-wide heartbeat stuff, to suppress keep-alives Mar 18 23:26:49 yeah Mar 18 23:27:02 skype doesnt seem to do that ;) Mar 18 23:27:16 it really sucks the battery dry Mar 18 23:27:27 Robot101: suppress keep-alives? how's that? Mar 18 23:27:56 DocScrutinizer: there's a heartbeat daemon, and an iptables module which queues up TCP keepalives, and releases them when the daemon says so Mar 18 23:27:58 or are you talking about that bundling of activities? Mar 18 23:28:08 yeah Mar 18 23:28:17 tcp keepalives are pretty insignificant compared to IM protocol keepalives :-) Mar 18 23:28:19 other stuff that sends connection-level keepalives including SIP and XMPP connect to the same daemon Mar 18 23:28:32 and only send them out when the timeouts go, or something else is using the connection Mar 18 23:28:57 so they all skip to the same beat, and only wake up the device & antenna if the connections need to be kept alive Mar 18 23:29:13 yes, I know about that Mar 18 23:29:28 skype has no such thing, it makes 10s of TCP connections and talks crap over them all all the time Mar 18 23:29:34 and wakes up almost constantly Mar 18 23:29:37 kiss your battery goodbye :P Mar 18 23:29:57 "suppressing keep-alives" somewhat sounded strange to me though, I probably misunderstood it Mar 18 23:30:06 I've only seen one tcp connection from skype, and yeah it does use power :-) Mar 18 23:30:41 also all that doesn't really help for curing bad efects of random inbound pings Mar 18 23:30:58 which you get a lot for skype/IM aiui Mar 18 23:32:33 for inbound you'd need a proxy doing that buffering/bundling for you Mar 18 23:33:13 iirc ShadowJK implemented sth along that line for IRC Mar 18 23:33:46 time for 12648430 ☕ Mar 18 23:41:13 Robot101: there's a problem with telepathy-sofiaSIP, contacts, and dialer. I got an account at sipgate.de, a german SIP provider. It has an associated landline number. Inbound call from e.g landline "+49 911 123456" are signalled and logged as origin:"SIP:0911123456@sipgate.de". This doesn't match to any contact that has tel:"0911123456" nor to "+49911123456". My question: is there a concept in telepathy to deal with this? Mar 18 23:42:50 alas no, the address book code in the N900 has special-cases for doing suffix lookups on incoming PSTN calls, but the same isn't true of SIP URIs Mar 18 23:43:10 it would need to have some heuristic to decompose them to pull out a telephone number Mar 18 23:43:19 but it's always going to be a kinda lame hack Mar 18 23:58:04 well, in twinklephone there's an option to check for "...;call=tel" parameter and in that case discard the domain part and display the user part as a plain simple phonenumber. Service "tel:" in any SIP addr should get handled same way. But yeah, basically I agree there has to be heuristics to deal with that, as there seems to be no consistent standard how carriers handle signalling of inbound calls via PSTN gateway Mar 18 23:58:51 aha, call=tel sounds useful Mar 18 23:59:44 want to send a wishlist bug to telepathy-rakia and say "expose tel: and ;call=tel numbers through a different namespace so that UIs can apply PSTN heuristics"? Mar 19 00:00:41 so maybe sofiaSIP might learn to callback a hook to arbitrary scripts that act as a (unix) filter converting stdin sip:0911123456 to stdout 0911123456, and then sofiaSIP would use that cooked addr for further processing of the inbound call Mar 19 00:01:07 that seems... er... unlikely :P Mar 19 00:02:39 (want to send...) nah, sorry. I got nfi how telepathy works, and what to implement where to fix/augment things and who's the dude to talk to about it. Mar 19 00:03:17 i.e. I don't even know rakia, never heard of it/him/her(?) Mar 19 00:04:22 #telepathy is a good starting point Mar 19 00:04:28 tp-rakia is the new name for tp-sofiasip Mar 19 00:04:37 mikhailz is the main dev there Mar 19 00:04:43 alsuren is also working on some SIP stuff on and off Mar 19 00:05:31 thanks, but unlikely I want to spend weeks to get knowledge up to a level so I could participate there Mar 19 00:05:54 well if you're going to be so obstructive about it, I'll file the (*&!098 bug :P Mar 19 00:06:25 :-D Mar 19 00:07:32 * Arkenoi thinks telephony providers that present caller id in stupid "local" format should be burned. Mar 19 00:08:21 DocScrutinizer: https://bugs.freedesktop.org/show_bug.cgi?id=35437 Mar 19 00:08:22 Bug 35437: was not found. Mar 19 00:08:26 DocScrutinizer: feel free to add yourself on CC :P Mar 19 00:08:45 thanks Mar 19 00:16:57 hi all Mar 19 00:17:25 anyone familiar with tvout/overlay around here? Mar 19 00:40:17 e-yes: shoot Mar 19 00:40:21 ~ask Mar 19 00:40:21 Questions in the channel should be specific, informative, complete, concise, and on-topic. Don't ask if you can ask a question first. Don't ask if a person is there; just ask what you intended to ask them. Better questions more frequently yield better answers. We are all here voluntarily or against our will. Mar 19 00:45:43 DocScrutinizer, is it possible switch output lcd->tv (and back) without X server? (I mean something simple like shellscript) Mar 19 00:46:40 Robot101: ( address book code in the N900 has special-cases for... ) There's a checkmark box on SIP accounts in fremantle "[X] use this account for PSTN" (OWTTE). Seems to me it should work both directions, i.e. handle inbound addr from that account in a f'() way of what it does for dialing PSTN number outbound via that account Mar 19 00:46:47 Without X server, with the default kernel, you get a reboot Mar 19 00:47:04 e-yes: what are you trying to do? Mar 19 00:47:56 DocScrutinizer: yeah - it doesn't :P Mar 19 00:48:09 e-yes: TVout is via a dedicated framebuffer that just happens to (usually) have same content like the one used for LCD, afaik Mar 19 00:48:31 i have process that open framebuffer Mar 19 00:48:59 trying to switch (the same) framebuffer to tvout Mar 19 00:49:10 i'd assumed it was able to capture video overlays like some desktop graphics cards Mar 19 00:50:03 e-yes: i am no expert for that, but would assume you can't simply switch the LCD framebuffer to output to AVout Mar 19 00:50:58 afaik they have significantly differing properties, regarding resolution, framerate, whatnot else Mar 19 00:52:30 57.2Hz? wtf? Mar 19 00:53:41 (from xrandr, run using ssh -X to a normal linux machine that actually has the xrandr command) Mar 19 00:53:55 is that really the refresh rate? why? Mar 19 00:55:45 BCMM, why not? Mar 19 00:55:55 it's a rather unusual choice... Mar 19 00:56:04 enought for nonflicker on lcd Mar 19 00:56:26 i'm wondering if there's a specific reason for that choice is all Mar 19 00:57:18 some DMA limitations? LVDS bandwidth limit? Mar 19 00:59:06 Baseball's back. Mar 19 00:59:13 I can be happy again. Mar 19 01:02:14 Is there any good method for imaging both the rootfs and the emmc to clone a device? Mar 19 01:05:22 DocScrutinizer, e-yes : observation: tv out normally outputs the translated x11 RGB buffer (how it does it, IDK - call it magic) but if you open a YUV overlay, that is output instead. Mar 19 01:05:41 for as long as the YUV overlay is in effect, that is what is sent to the tvout Mar 19 01:06:16 normally, maemo video player stops the YUV overlay when you pan to dashboard, and tvout resumes showing the x11 content Mar 19 01:06:18 YUV overlay == using xv? Mar 19 01:06:24 yes Mar 19 01:06:27 ah Mar 19 01:06:41 GeneralAntilles: (clone) I think .tgz images created by backupmenu should suffice Mar 19 01:07:10 but it is entirely feasible (and played with quite a lot) that you can continue playing YUV overlay content whilst you have entirely different data on the X11 RGB Mar 19 01:07:30 lcuk, thx. looks like so, I'm googling OMAP/DSS now:) Mar 19 01:07:43 lcuk: (normally video player stops...) yes, a very annoying property Mar 19 01:08:02 DocScrutinizer, stops for a different reason Mar 19 01:08:09 the YUV layer is not composited ;) Mar 19 01:08:21 so you end up seeing a postage stamp window into the whole movie Mar 19 01:09:15 lcuk: nah, I'd prefer media player to contine playback movie via AVout, even while I'm checking email on N900 Mar 19 01:09:49 (you can continue playing YUV overlay content whilst you have entirely different data on the X11 RGB) Mar 19 01:09:56 yes Mar 19 01:10:26 I know however the majority use case is not tvout Mar 19 01:10:32 so the YUV layer gets stopped for that Mar 19 01:10:35 tbh that's a design flaw - there had to be a config feature or menu setup entry Mar 19 01:10:49 I think it is overlooked Mar 19 01:10:56 of course Mar 19 01:11:01 like so many others Mar 19 01:11:18 I thought about using a BT keyboard and mouse to be honest Mar 19 01:11:36 e.g audio playback via AV to home stereo, where you for sure don't want ts clicks and ringtones via AV Mar 19 01:11:56 and btw did i get that right, n900 does something like vaapi/vdpau too? Mar 19 01:11:56 * jonwil is about ready to give up hope of seeing anything from Nokia that will benefit Maemo, despite https://bugs.maemo.org/show_bug.cgi?id=11128#c8 Mar 19 01:11:57 Bug 11128: Opening the source for Sharing services framework Mar 19 01:12:35 jonwil: poke Jaffa about state of negotiations Mar 19 01:12:41 * lcuk recognises that bug Mar 19 01:12:49 *grin* Mar 19 01:13:06 nox-, well... if you push something through gstreamer it might go through the hardware decoded path, or fail Mar 19 01:13:16 nox-, it's something like, um, openmax.. Mar 19 01:13:18 or openil Mar 19 01:13:21 whatever Mar 19 01:13:26 at the end of gstreamer Mar 19 01:13:46 but its also gpu offloading? Mar 19 01:13:55 or is that the dsp? Mar 19 01:14:10 GPU is worthless for video. Mar 19 01:14:36 * nox- just curious cause i had some `fun' with vdpau/vaapi and libxine/vdr... Mar 19 01:15:11 so its the dsp that does it? Mar 19 01:15:11 (Yeah on nvidia desktops chips more capable than the low-end ones, there are enough shader units that they actually decode video with sheer bruteforce) Mar 19 01:15:15 Yes. Mar 19 01:15:29 ok Mar 19 01:15:57 And the result is fed to DSS2 which does the colourspace transform and scaling Mar 19 01:16:11 DSS2? Mar 19 01:16:26 there is also the IVA(2) which as far as I know is a whole core like the DSP but essential unused Mar 19 01:16:34 heh Mar 19 01:16:35 IVA is used Mar 19 01:16:39 ShadowJK, the IVA is meant to do that work Mar 19 01:16:41 ahh Mar 19 01:16:47 But it's undocumented Mar 19 01:16:55 it does YUV<->RGB too Mar 19 01:17:12 ShadowJK, used by what? Mar 19 01:17:23 When you play h264 movie Mar 19 01:17:35 When is the best time to catch Jaffa? Mar 19 01:17:39 how would a mere mortal use it Mar 19 01:17:45 lcuk, one wouldn't Mar 19 01:17:50 jonwil, just after it falls from the tree. Mar 19 01:17:56 for camera resizer iirc Mar 19 01:17:59 Unless you go through mafw or gstreamer Mar 19 01:18:12 ShadowJK, how does gstreamer use it? Mar 19 01:18:30 the closed dsp blobs Mar 19 01:18:39 is it just a coprocessor (like old amiga blitter) or does it have its own kernel and extra stuff (like the DSP) Mar 19 01:19:02 * SpeedEvil tries to remember how many cores he counted in the n900. Mar 19 01:19:07 Basically, IVA is this unit full of useful algorithms common to most video formats, and it's accessible from the DSP. The "DSP Codecs" in fact access IVA for the heavy lifting Mar 19 01:19:07 I think I got up to a dozen once. Mar 19 01:19:08 4 Mar 19 01:19:11 11 Mar 19 01:19:15 1 cpu, 1 powervr, 1 dsp, 1 iva Mar 19 01:19:29 ;-D Mar 19 01:19:30 Bluetooth, wlan, modem (2), ... Mar 19 01:19:31 also, neon runs almost like a coprocessor too :) Mar 19 01:19:48 ShadowJK, those are not compute cores in the same sense Mar 19 01:20:11 my understanding was the others mentioned are some kind of mutated arm Mar 19 01:20:14 over9000 cores Mar 19 01:20:27 The DSP is nothing like an ARM. Mar 19 01:20:35 I would guess that DSS2 is Display SubSystem version 2. It would do scaling and colourspace transform, I'd think. Mar 19 01:20:36 * lcuk is wrong then Mar 19 01:20:38 leg is not like an arm Mar 19 01:20:42 and you would know derf, thanks \o Mar 19 01:20:54 the DSP is, well, a DSP Mar 19 01:21:14 * ZogG_w is bored Mar 19 01:21:25 if DSP codecs was outputting RGB, I think you could composite it instead of going through Xv... If GPU/SGX has enough bandwidth for 800x480 RGB textures at 30fps, that is Mar 19 01:21:50 It's VLIW capable of dispatching 8 instructions/clock (of various types, not arbitrarily), with 64 32-bit registers. Mar 19 01:22:21 So like a Transmeta P4 :-( Mar 19 01:22:29 (yeah I intentionally mixed metaphors) Mar 19 01:22:30 Yeah, somewhat. Mar 19 01:22:47 Except it runs off 2W of power going full bore. Mar 19 01:23:03 (and is considered a power hog in embedded circles for using that much) Mar 19 01:23:12 I doubt the DSP uses 2W full-bore. Mar 19 01:23:14 (alone) Mar 19 01:23:21 C64x+™ DSP and Video Accelerators (3525/3530 only) Mar 19 01:23:33 http://focus.ti.com/lit/sg/sprt537a/sprt537a.pdf Mar 19 01:23:37 ShadowJK, problem is not explicit to textures Mar 19 01:23:47 the XV mode runs in arbritary resolution Mar 19 01:24:06 so you send video through at 400*240 and it does not waste bandwidth using all 800*480 Mar 19 01:24:13 as would be currently required in RGB Mar 19 01:24:51 OMAP35x devices integrate the superscalar Cortex-A8 core and TI’s TMS320C64x+™ DSP into four applications processors: OMAP3503, OMAP3515, OMAP3525 and OMAP3530. Mar 19 01:25:42 ShadowJK, it *might* be feasible to do it of course, but it would be such a waste of power and bandwidth Mar 19 01:26:05 lcuk, sure, but considering SGX is a 3D GPU, you'd assume it could stretch at 400x240 texture to 800x480 too Mar 19 01:26:09 Yeah definitely Mar 19 01:26:35 I installed the telnet client today and suddenly, the SMS application does not display contact information. Anyone know what might be going on here? I've rebooted a couple of times and that has not helped. Mar 19 01:27:05 DSS is purpose built, SGX/3D/GPU can theoretically do the same thing but will be much more inefficient at it Mar 19 01:30:08 is the LCD controller chip inside the omap or like the old n8x0 with an external one Mar 19 01:30:12 (was it epson) Mar 19 01:30:19 Yeah it was epson Mar 19 01:30:25 And I think DSS is out LCD controller Mar 19 01:30:27 SpeedEvil: Yeah, I think the DSP by itself may only be 1W. But that's still huge. Mar 19 01:30:29 our* Mar 19 01:31:02 ... and more than an order of magnitude less than a P4 (or a Transmeta CPU). Mar 19 01:31:05 ShadowJK, also, by converting to 16but RGB you lose more definition Mar 19 01:31:12 and people would complain of banding in movies :P Mar 19 01:31:18 derf, get some real world power consumption readings of playing video, it's about 1W *total* :-) Mar 19 01:31:20 bit Mar 19 01:31:29 lcuk: That's what dithering is for. Mar 19 01:31:39 derf, YUV does not require dithering Mar 19 01:31:56 Sure it does. It's just dithering in hardware. Mar 19 01:32:00 and it is lower ba ndwith Mar 19 01:32:08 Or if it isn't, it's still got banding. Mar 19 01:32:29 derf: yeah Mar 19 01:32:37 it sends data down the wire as far as it can in 24b Mar 19 01:32:50 the lcd itself might only be 18bit or whatever they are Mar 19 01:32:57 Jaffa: ping Mar 19 01:33:06 You mean in 12b. Mar 19 01:33:17 Or does it run at 4:2:2? Mar 19 01:33:49 can someone remind me of the right way to reboot the DSP if it's doing the 'I can't play any video at all' thingy? Mar 19 01:33:51 422 Mar 19 01:34:09 Okay, then 16b. Mar 19 01:34:11 SpeedEvil, power button Mar 19 01:34:22 lcuk: I was assuming that was the improper way. Mar 19 01:34:46 derf, no 8 bit grey, + 8 bit U + 8 bit V Mar 19 01:35:12 SpeedEvil: stop dsp/start dsp? Mar 19 01:35:34 lcuk: But you don't have 8 bits of U and V per pixel. Mar 19 01:35:41 sure you do Mar 19 01:35:49 Argh. Mar 19 01:35:53 Not if you're using 4:2:2 subsampling, you don't. Mar 19 01:36:03 It's not that - it's that BBC have changed something. Mar 19 01:36:07 420*240 is true colour Mar 19 01:36:24 ... Mar 19 01:36:24 it is full colour, low resolution Mar 19 01:37:48 derf that is the normal mode of videos since forever, and jpegs too Mar 19 01:38:13 No, most vide is 4:2:0. But no one calls it 24bpp, either. Mar 19 01:38:16 *video Mar 19 01:40:24 YUV also has about 1/4 the gamut of RGB. Mar 19 01:40:35 800mW - playing video at low volume over speakers, with LCD at minimum brightrness Mar 19 01:40:52 (well, more accurately, you lose about 2 bits in the conversion process) Mar 19 01:42:31 (1.2W with mplayer) Mar 19 01:43:17 Also - meh. I now need to work ourt how to transcode videos that used to play fine. Mar 19 01:43:18 derf, since the lcd cannot display 24bit anyway, that is not too bad Mar 19 01:44:14 Sure. I'm just telling you that claiming it's "TrueColor" is a lie. Mar 19 01:44:15 lcd is pretty bad at BPP Mar 19 01:44:34 They make 30-bit panels. Mar 19 01:45:08 they make. yeah. The small embedded LCDs though are 12..16bpp Mar 19 01:45:16 derf, roger Mar 19 01:45:29 Yeah. 30-bit panels aren't cheap. Mar 19 01:45:35 * lcuk just never notices banding Mar 19 01:45:54 seems N900 LCD is 15 or 16, while I suspect AV fb is like 12 Mar 19 01:47:17 I guess this LCD is a 5/6/5 Mar 19 01:50:19 The N950 maybe has it. Mar 19 01:53:43 user: ??? Mar 19 01:54:26 A bether screen and hdmi out. Mar 19 01:54:56 I'venever had a real issue with LCD colour gamut. Mar 19 01:55:04 And I want VGA out! Mar 19 01:55:05 :) Mar 19 01:55:16 HDMI is quite a large plug Mar 19 01:55:39 component on RCA ;-P Mar 19 01:55:58 SCART Mar 19 01:56:15 SCART would just about fit along the long edge. Mar 19 01:56:20 DocScrutinizer, sod that, can we have a demodulator that plugs in the back Mar 19 01:56:24 Amiga A500 style Mar 19 01:56:46 I was actually looking to make a TV modulator for n900. Mar 19 01:56:54 But all the nice chips for it were EOL'd Mar 19 01:57:14 Also analog TV is getting turned off. Here it's going away in June/July. Mar 19 01:58:00 DVBT-TX it's then Mar 19 01:58:16 no plug \o/ Mar 19 01:59:11 if we can do FMTX so why we can't do DVBT-TX Mar 19 02:00:32 that's actually a good use made of DSP ;-P Mar 19 02:01:36 dvb-tx would be fun. Mar 19 02:02:05 yeah, but please with >>5nW Mar 19 02:02:57 I don't want to use a remote to control the N9N+ sitting next to TV's antenna receptacle Mar 19 02:03:32 at very least a 5m should be design goal Mar 19 02:04:27 given the digital nature of DVBT there's little to no concern about long range interference with public broadcast DVBT Mar 19 02:04:44 Much less. Mar 19 02:04:52 The rationale for interference on FM is valid. Mar 19 02:05:02 yeah, somewhat Mar 19 02:05:14 It takes a really teeny amount of power to disrupt a signal that's at the edge of coverage, and make it much worse. Mar 19 02:05:19 though I don't give a f...art Mar 19 02:05:22 DVB-T - less so. Mar 19 02:07:27 when it starts that my torch flashlight has stronger EMI than the FM-TX of my N900, then it's at edge of trifle Mar 19 02:07:43 yeah Mar 19 02:17:06 Anyone with a video encoding clue? ffmpeg says of the input stream - Mar 19 02:17:09 Stream #0.0(und): Video: h264 (Main), yuv420p, 640x360 [PAR 90:90 DAR 16:9], 698 kb/s, 25 fps, 25 tbr, 25 tbn, 50 tbc Mar 19 02:17:21 I thought the n900 could do h264 Mar 19 02:27:34 DocScrutinizer, did you happen to see this earlier: http://pastebin.com/RTP1bUDT Mar 19 02:30:26 lcuk: no. I admit I'm scratching my head trying to figure what that means Mar 19 02:30:35 DocScrutinizer, me too Mar 19 02:30:43 it surprised me immensely Mar 19 02:30:57 since world + dog says that file system is slower on meego Mar 19 02:31:07 and that is the main difference I think Mar 19 02:31:21 I doubt the compiler itself gained 50% optimisations Mar 19 02:33:02 and the variations on maemo side are me trying different combinations of partitions to store onto Mar 19 02:33:09 I neither understand completely what's written there, nor do I think it's giving a good proof of anything, without proper explanation of the whole setup and tests done Mar 19 02:33:30 sure, it is certianly not scientific Mar 19 02:33:48 I just got build-essential + the -dev libs required and ran make Mar 19 02:33:53 on both maemo and meego Mar 19 02:35:24 * lcuk should make a set of liq* specific tests :) Mar 19 02:35:28 hmm, seems to me a more standardized testbed would help a lot. Like e.g. use the absolutely identical source that uses the identical includes etc Mar 19 02:35:39 it was identical source Mar 19 02:35:58 but obviously using (maemo|meego) compiler and headers Mar 19 02:36:36 yet way too complex to assure the same process taking place, regarding all the includes which might differ significantly between OS Mar 19 02:37:22 yes DocScrutinizer Mar 19 02:37:53 Uh, who the hell says MeeGo filesystem is slower? Mar 19 02:38:18 same who tell eMMC is slower than NAND I guess Mar 19 02:38:52 I think Maemo had the advantage here anyway, because default Maemo setup has stuff on NAND, offloading emmc somewhat. Mar 19 02:39:47 For comparative testing you'd have to have both Maemo and MeeGo running from the same devices Mar 19 02:40:11 this comparision has way too many unknowns you'd need to define to make the whole thing a somewhat meaningful statement about anything Mar 19 02:40:19 But still, I assume his setup was slanted in favour of Maemo, which makes brtfs' win all the more impressive :-) Mar 19 02:41:08 ShadowJK, it was from the same device Mar 19 02:41:21 I booted meego, and just happened to time a build Mar 19 02:41:32 then got curious and booted same device to maemo Mar 19 02:41:37 and repeated Mar 19 02:41:41 So you've got Maemo running on MicroSD then? Mar 19 02:41:52 ShadowJK, no Mar 19 02:41:54 normal system Mar 19 02:41:59 This is what I meant Mar 19 02:42:06 I didn't mean device as in "N900" Mar 19 02:42:12 but device as in NAND, emmc, microsd Mar 19 02:42:27 I'd use a 10liner shellscript to mkdir;cp $fn N$fn; fn=$Nfn Mar 19 02:43:15 ShadowJK, the problem is that meego slow bootup speed has been attributed to the filesystem Mar 19 02:43:16 for {fn=2000; fn>0; fn--) Mar 19 02:43:22 sth along that line Mar 19 02:43:30 lcuk, so does it run faster with ext3? Mar 19 02:43:32 DocScrutinizer: why 10? Mar 19 02:43:40 what confused me was just that building this library on meego I was expecting it to be slower Mar 19 02:43:50 well, you can pack everything into one line, always Mar 19 02:43:52 for i in {10..1}; let j=i+1; mv $i $j; done Mar 19 02:43:56 I don't know why Mar 19 02:44:05 but on meego it ran like shit off a shovel Mar 19 02:44:18 and meego has cpu freq at 500mhz Mar 19 02:44:23 it SHOULD be slower Mar 19 02:45:07 (I am really pleased it is not, but that is beside the point) Mar 19 02:45:09 lcuk, in reality I think the problem is that neither ext3 nor btrfs are tuned for flash based storage in any way (well btrfs somewhat, but it's mostly for SSD type storage, which is probably more harmful than useful for flash), and Maemo only accidentally rather than by design avoids driving the emmc/microsd into utter slowass operation.. Actually under Maemo you can hit the slow-as-molasses situations too by trying to extract easy-debian, for example.. Mar 19 02:45:49 ShadowJK, yes, I use n900 itself to extract meego images using dd Mar 19 02:45:50 * SpeedEvil wonders if the SD committe will do something useful and define TRIM Mar 19 02:45:53 and that operation takes just under an hour Mar 19 02:46:28 SpeedEvil, trim is kinda useless when most controllers (intra-SD ones) are so braindead they wouldn't have one fucking clue what to do with trim Mar 19 02:47:00 ShadowJK: I am unsure. I've read conflicting things about SD controllers. Mar 19 02:47:20 SpeedEvil, well it's the software on them that makes the difference, and most of it is braindead Mar 19 02:47:37 They seem to have at least arm cortex class CPU on them according to reports I've seen. Mar 19 02:48:22 except maybe sandisk, but sandisk has been saying "Our ExtremeFFS technology kicks ass" for years, it's just that they're trying to monetize it so hard that they've actually sold about 0 copies of it so far, when it really should be sitting in every SD card they're selling... Mar 19 02:49:11 In practice - for many apps - flash smartness is quite unimportant. Mar 19 02:49:13 Not that I'm saying sandisk cards are shit, they're about 5 magnitudes of performance better than Kingston, but it could be so much better if the industry tried to get their shit together. Mar 19 02:49:36 It doesn't help noticably for recording large contiguous files Mar 19 02:49:46 SpeedEvil, if you consider one single app in isolation, sure, it's extremely unimportant Mar 19 02:50:03 For OS use, it's not so great. Mar 19 02:50:21 But you've got like 100 processes running, and each one creates IO load :-) Mar 19 02:50:25 whether they intend to or not Mar 19 02:50:42 * SpeedEvil mehs. Mar 19 02:50:55 I should really have gotten on with that swap lineariser idea. Mar 19 02:51:02 :-) Mar 19 02:51:13 nbd client, that fakes a block device on a file. Mar 19 02:51:52 https://wiki.linaro.org/WorkingGroups/KernelConsolidation/Projects/FlashDeviceMapper :-) Mar 19 02:52:42 I have no clue who linaro is, what they do, but they seem to know more about flash, and trying to do something about it, than companies that have been trying to use flash for years... Mar 19 02:53:17 Sort-of. Mar 19 02:53:21 But not quite. Mar 19 02:53:34 That nokia hacked N900 kernel is somewhat swap-linear actually :-) Mar 19 02:53:38 Basically, for swap, for obvious reasons, you can drop the whole persistant block mapping shit. Mar 19 02:53:48 Yeah Mar 19 02:53:56 Which does wonders for complexity. Mar 19 02:54:01 lcuk: see, meego might use different fs tuning settings (see swappolube et al) Mar 19 02:54:34 Also 'Finally, the superblock records its own age as an increasing 32 bit number, and a checksum. ' Mar 19 02:54:44 N900 kernel keeps track of "biggest free area", and writes sequentially there, and when exhausted, jumps to next "biggest free are". As time goes on, these free areas shrink to stupidly small in size, and you hit the classic "my device suddenly feels slower after 2-3 days uptime, what's going on?" Mar 19 02:55:27 Ideally, you don't want to do that, if you can avoid it, as you want to not ensure 100% of the time (well - 50%) incrementing the superblock counter will require a block erase Mar 19 02:55:46 Swappolube people are kinda funny. They're on the extremes of "maximum interactivity, minimum throughput", and "maximum throughput, minimum interactivity", and randomly flip parameters between the two extremes :D Mar 19 02:56:25 I have wondered about truly stupid hacks. Mar 19 02:56:40 500 8M swapfiles Mar 19 02:56:50 and you swapoff as each one fills. Mar 19 02:56:58 DocScrutinizer, possibly, either way, pleasantly surprised and given urge to find out more Mar 19 02:57:09 Hi! Can /usr/lib/microb-engine/libxul.so be safely moved (on a running system) to somewhere, then have that path symlink to the new path? Mar 19 02:57:23 SpeedEvil, from a performance point of view, the swap storage becomes slow long before it's "filled" Mar 19 02:57:25 stop microb Mar 19 02:57:26 IIRC Mar 19 02:57:40 "The superblock is stored in first allocation group on the medium, but may be at changing addresses within the allocation group. Only the superblock with the highest version number and a valid checksum gets used when mounting the medium. " Mar 19 02:57:46 stop it first? Then move it then start it again? Mar 19 02:57:51 ShadowJK: By filled, I mean that it has completed a linear write Mar 19 02:57:53 ebzzry: yes Mar 19 02:57:58 This is kinda clever actually. Some cards have extra optimization for the first allocation group :-) Mar 19 02:58:08 Thanks. Since we're on that, what other things are stoppable? Mar 19 02:59:17 ebzzry: risk of bootloop Mar 19 02:59:31 yeah Mar 19 02:59:36 DocScrutinizer: Uh oh. Mar 19 02:59:49 oh - and killall browser 3 times, seperated by 20s should kill browser **** ENDING LOGGING AT Sat Mar 19 02:59:56 2011