**** BEGIN LOGGING AT Tue Sep 03 02:59:58 2013 Sep 03 03:00:08 yeah it seems pretty unknown? Sep 03 03:00:18 like... the thing where you walk around some world doing nothing? Sep 03 03:02:08 Macer, I thought you made money off of it by designing in-game items and selling them Sep 03 03:02:32 like... real money? Sep 03 03:02:36 Macer: yep, the thing with linton dollar Sep 03 03:02:38 who buys shit like that? Sep 03 03:03:04 Bored people Sep 03 03:03:10 if it had a porn theme it would be a mega hit Sep 03 03:03:24 hehe... Are you thinking it /doesn/t? Sep 03 03:03:24 ooh it FOR SURE has Sep 03 03:03:26 world of whackcraft Sep 03 03:03:39 it does? Sep 03 03:03:40 lol Sep 03 03:04:19 there already been trials in real world, trying to prosecute real porsons for doing prostittution and worse in SL Sep 03 03:04:21 ...I once ran into a tech support story(wrong number etc)... The user calling in had a bluetooth-enabled, SL-connected dildo. Sep 03 03:04:31 persons* Sep 03 03:04:56 exhibitionism in SL Sep 03 03:05:08 jail in RL Sep 03 03:05:18 ;-P Sep 03 03:07:51 but recently I heard nuttin about SL anymore.Thought it might have vanished in a supernova somebody injected to the main servers via some malware Sep 03 03:12:36 people moved to twitter Sep 03 03:12:49 and facebook Sep 03 03:20:46 lol http://allthingsd.com/20130902/microsoft-to-buy-nokias-device-business-for-7-17-billion-euros/?utm_source=twitterfeed&utm_medium=twitter Sep 03 03:22:32 ...Yup Sep 03 03:22:51 Elop was a shill, and everyone knew it! Sep 03 03:24:13 brought in to give MS a phone company, and kill a major competetor(Maemo & symbian) who could have threatened things. Sep 03 03:24:13 Oh, and guess what, Mr. Elop might well get to take over MS, after gutting Nokia Sep 03 03:24:17 I can only hope what's left of Nokia gets some balls back Sep 03 03:29:14 fare thee well Nokia, exactly what I expected to happen has Sep 03 03:29:55 MS-Nokia? old news, everybody expected err *knew* this when MrFlop started Sep 03 03:32:44 Yeah Sep 03 03:33:20 MrFlop the new Uncle Fester though is *really* funny Sep 03 03:35:47 sounds like the mafia boss asking his worst most sadistic killer to now babysitting the boss' children Sep 03 03:37:06 well, probably MS actually is the first and maybe one of only 10 companies worldwide that's really just too big to fail Sep 03 03:38:15 Are they? Sep 03 03:38:26 What would happen if MS went out of business? Sep 03 03:38:45 The rights to Windows would be sold off, someone else would take on supporting current versions Sep 03 03:38:45 I'd think yes, they are. Otherwise they wouldn't exist anymore since years Sep 03 03:39:48 Even if MS just disappeared, windows licences stopped being sold... Sep 03 03:39:59 the question "what would happen if..." is basically meaningless, since they simply CANNOT go out of business Sep 03 03:40:09 things would go on. There'd be a lot of porting of software; probably a massive econemy boom Sep 03 03:40:55 Well, to be fair, they /do/ have enough money in the bank that if they stopped selling any new licenses, they'd not go out of business for years Sep 03 03:41:10 they could be losing money slowly for decades and still be OK Sep 03 03:41:38 exactly Sep 03 03:42:43 That being said, I do think that if they aren't careful, this NSA stuff will quickly render them obsolete anywhere but the US Sep 03 03:42:44 and even form a market perspective they can't go out of business, just like exxon or BP can't stop producing oil Sep 03 03:44:52 NSA? LOL! in 12 months MS comes up with WinSecure, now absolutely safe Sep 03 03:44:53 and they will earn even more money Sep 03 03:45:18 I rather guess NSA will get bought by MS. That's called outsourcing Sep 03 03:46:08 Yeah. And will any government outside the US(and perhaps UK) trust them? Sep 03 03:46:18 NSA fires 99% of their employees, and MS does all the jobs of NSA Sep 03 03:46:44 sure, since they are not the NSA, they are "the good ones" Sep 03 03:47:36 I sure hope not Sep 03 03:47:44 and the heck, prolly they are. After all they won't be that much intrested in spying any other industries except their direct competitors Sep 03 03:48:08 The NSA, however, will be interested in spying on everything Sep 03 03:48:18 and so, they will have orders to find such things Sep 03 03:49:10 nonono, you don't get it. NSA gets bashed by US public, and the US residents rather trust in MS than in NSA. They don't like thaat spying either Sep 03 03:49:42 Well, US residents, I don't hold much stock in them knowing much of anything Sep 03 03:49:49 I'm thinking everyone outside Sep 03 03:49:51 sure Sep 03 03:50:01 EU -- whom we've been spying on Sep 03 03:50:14 China, russia, who really don't want to be spied on Sep 03 03:50:33 but don't hope on NSA and prism/tempora having any negative impact on MS business. Rather the opposite, if any Sep 03 03:51:03 It might help in the US, but I'm thinking their global numbers will fall Sep 03 03:51:29 hardly noticeable. Sep 03 03:51:47 what might kill their international business is win8 crap Sep 03 03:52:10 And, I'm still waiting to see what happens with XP Sep 03 03:52:19 Considering 35% marketshare, still Sep 03 03:52:33 We're still running XP in various places at work, no plans to upgrade Sep 03 03:53:03 when really the last noob shouts in the coputer shop "WHATS *THAT*? I don't want that please give me some alternative" Sep 03 03:53:34 [2013-09-03 04:43:07] Industry is still on winXP, since everything since been useless for professional use, even at industry level (I consider industry requirements rather undemanding here) Sep 03 03:54:12 Well, here, it's half Win 7, half XP Sep 03 03:54:40 ...Though, while Win 7 is generally nice and up-to-par with Debian KDE, I do get frustrated with it some times Sep 03 03:54:52 at my last 2 companies it been 99% XP, and 70% XP Sep 03 03:54:57 I open my 'downloads' folder, and it takes like 20s just to display the contents. Sep 03 03:55:37 I've opened much larger folders on my various Debian installs, and it doesn't take more than 1s, if that! Sep 03 03:56:10 the 99% XP company struggled hard to find a way to migrate all their tools and stuff to FOSS Sep 03 03:57:11 and they had a *lot* of tools and proprietary apps to port to a new OS. That's probably why Sep 03 03:58:00 if you already have to port all that cruft to a new OS, then pretty please an OS where you got full control over updates and lifespan Sep 03 03:58:44 any win-only apps? so what, we got VMware Sep 03 03:59:15 Yeah Sep 03 03:59:23 and Lauterbach will notice eventually, and offer a linux version of their tools Sep 03 04:00:03 as will all those ITIL-system companies Sep 03 04:00:52 hah, lauterbach. Haven't checked for a year what they're doing Sep 03 04:01:26 when Autodesk and Catia offer a Linux version Sep 03 04:01:55 yeah, now it's official: MS buys Nokia core business for 5.4 billion USD Sep 03 04:02:29 today I'll actually watch NBC news Sep 03 04:02:47 particularly the NYSE Sep 03 04:03:30 farewell Nokia Sep 03 04:07:57 *lol* call me a fool - I just registered ##maemo Sep 03 04:11:33 just in case you wonder what's "lauterbach": http://www.lauterbach.com/frames.html?country=de%3fhome_d.html Sep 03 04:14:07 http://www.lauterbach.com/frames.html?country=de%3fhome_d.html Sep 03 04:15:39 meh, of course they already have linux support Sep 03 04:16:48 Qt and Motif Sep 03 04:30:44 Hello msn-nokia :s Sep 03 04:32:02 ooops, EUR, not USD Sep 03 04:33:13 well 5.4 billion EUR or USD, not much of a difference Sep 03 04:34:06 * DocScrutinizer05 wonders if we will again be able to buy NOKIA rubberboots finally Sep 03 04:34:19 waders maybe :) Sep 03 04:35:10 yeah, probably Nokia will again be really big business in 30 years, with all that global warming ;-) Sep 03 04:36:27 ruberboats, boots, waders, tents, waterbags... I see a bright future for Nokia Sep 03 04:37:54 ;) Sep 03 04:43:16 DocScrutinizer05: how many n900's do you have now ? ;) Sep 03 04:43:40 they'll have antique value soon :) Sep 03 04:44:27 well, not enough. But honestly, the today's hot news won't have any impact on that. Sep 03 04:44:59 today's hot news? Sep 03 04:45:07 n900s cure cancer? Sep 03 04:45:10 farewel Nokia Sep 03 04:45:16 heh Sep 03 04:45:24 farewell nokia? Sep 03 04:45:31 did i miss something? Sep 03 04:45:32 bought by MS Sep 03 04:45:38 oh Sep 03 04:45:48 5.5E9 EUR Sep 03 04:45:48 kinda saw that coming didnt we? ;) Sep 03 04:45:58 sure Sep 03 04:45:59 http://press.nokia.com/2013/09/03/nokia-to-sell-devices-services-business-to-microsoft-in-eur-5-44-billion-all-cash-transaction/ Sep 03 04:46:06 wow thats it? Sep 03 04:46:15 cheap buy for ms :) Sep 03 04:46:37 what is a € worth us nowadays? Sep 03 04:46:49 $2.50? Sep 03 04:46:53 nah Sep 03 04:46:59 1.20? Sep 03 04:47:07 wow fell that low? Sep 03 04:47:10 nice Sep 03 04:47:27 well... maybe ms is going to make nokia their iphone ;) Sep 03 04:47:30 maybe 1.40, I honestly don't know Sep 03 04:47:35 win8 phones arent too bad Sep 03 04:47:43 lumia 1020 is a beast Sep 03 04:48:14 then again tho.. google bought motorola and havent done anything with it Sep 03 04:48:33 probably just to pocket patents Sep 03 04:48:37 sony have arguably the best phones atm Sep 03 04:48:47 but that's a moving finishing line Sep 03 04:49:32 sony? Sep 03 04:49:47 i cant say i see too many sony phones out and about tbh. not here at least Sep 03 04:50:09 doesn't mean they are crap Sep 03 04:50:21 i didnt say they were Sep 03 04:50:33 just saying Sep 03 04:50:38 it is rare to see one Sep 03 04:50:59 Sony Xperia Z Ultra 6.4" phablet Sep 03 04:51:07 well, you get xperias in evry shop here Sep 03 04:51:09 very cool big phone Sep 03 04:51:11 if i take a route tho i think i will be going the ms route next Sep 03 04:51:29 is that the sony answer to a note? Sep 03 04:51:52 depends on whether the not was a question :p Sep 03 04:51:53 if my n900 ever dies. heh. i still have a brand new e7 on standby too Sep 03 04:52:05 haha Sep 03 04:52:38 Macer: and take a look at samsung mega 6.3" Sep 03 04:52:39 looks like 3 types of phones from here on out Sep 03 04:52:53 ad-droid, windows, and ios Sep 03 04:53:07 * jpinx will go for a phablet next Sep 03 04:53:10 maybe ms will use symbian code to create true multitasking Sep 03 04:53:13 :) Sep 03 04:53:32 i can noly imagine a windows phone with a symbian base on today's hardware Sep 03 04:53:46 using a windows ui Sep 03 04:53:51 what about the ubuntu phone ? Sep 03 04:53:54 that would be insanely good heh Sep 03 04:54:03 ubuntu fell short 19M :-P Sep 03 04:54:19 * jpinx is not writing them off just because of money Sep 03 04:54:34 i would Sep 03 04:54:42 unless they got something out from a major manufacturer Sep 03 04:54:59 unity is an ugly interface tho Sep 03 04:55:18 they have a good chance and the market is begging for anopen source phone Sep 03 04:55:19 on the phone it seems to be more usable than on a laptop though Sep 03 04:56:06 yah Sep 03 04:56:17 werent they going to use wayland too? Sep 03 04:56:24 like. no more xorg? Sep 03 04:57:09 im sure that may improve performance. but realistically the ubuntu phone is far from reality Sep 03 04:57:27 ms is actually starting to shape up to take a good position in the market Sep 03 04:57:28 Macer is a negative thinker ? Sep 03 04:57:30 Mir, no wayland and no X Sep 03 04:57:45 jpinx: no. i am a realist ;) Sep 03 04:57:57 they'll use Mir on desktop too, I think the next release will include it Sep 03 04:58:23 wow Sep 03 05:00:16 Macer: reality is relative . it's a question of perception :) Sep 03 05:04:07 good! one linux distro less I have to service on any of my friends' computers. If they want me to service their junk, they better stay away from buntkuh with mir. Sep 03 05:12:31 hmm Sep 03 05:12:45 Microsoft buys Nokia's devices & services so what's left ? Sep 03 05:14:11 welly boots :) Sep 03 05:15:21 I don't see the difference between "Microsoft buys Nokia's devices & services" and "Microsoft buys Nokia and bye bye for good" Sep 03 05:16:43 don't forget the wellies :) Sep 03 05:16:54 MS didn't buy them :) Sep 03 05:17:22 hehe Sep 03 05:18:06 what is left still has revenue of over 15billion so there is something left :P Sep 03 05:19:51 what Nokia has besides "devices & services" ? Sep 03 05:20:12 the building and employees? so that's left, they are not part of the deal Sep 03 05:21:09 networks is a separate org from devices and services, at least Sep 03 05:22:05 about 32,000 Nokia employees will transfer to Microsoft Sep 03 05:22:43 the deal will be finalised in early 2014 Sep 03 05:23:01 subject to approval by Nokia shareholders and regulators Sep 03 05:23:24 nokia will keep the HERE map bussiness, microsoft just paid to use it in the phones Sep 03 05:24:11 the whole board of directors of Nokia will accept the deal, they have to Sep 03 05:24:51 are you sure of that beford? Sep 03 05:24:58 that looks like it is part of the "services" deal Sep 03 05:25:01 that's how I read it, not sure Sep 03 05:25:24 Sc0rpius: yep - the patents and mapping ae going to be good earners for what is left of Nokia Sep 03 05:25:59 Nokia will also license its patents and mapping services to Microsoft Sep 03 05:26:07 not "sell" Sep 03 05:26:21 and who uses Nokia maps anyway... Sep 03 05:26:36 I thought everybody used Google maps for navigation in their phoens Sep 03 05:26:38 phones Sep 03 05:26:40 MS will :) Sep 03 05:26:59 MS will be glad of a way to cut googlemaps out Sep 03 05:28:52 at least Elop is stepping down for what's left of Nokia Sep 03 05:29:02 that's like a tiny bit of light in all these darkness hehe :) Sep 03 05:29:25 this was all a dirty plot Sep 03 05:29:37 Elop will go back to Microsoft after the acquisition closes Sep 03 05:29:51 this was all a plot from the very beginning Elop came to Nokia Sep 03 05:30:58 Sc0rpius: everyone has seen this coming for a year or so now Sep 03 05:31:06 not me! Sep 03 05:31:54 ever since Elop was appointed ;) Sep 03 05:36:48 well nokia it was nice knowing you Sep 03 05:36:56 heh Sep 03 05:40:27 as a maemo 4 user this all means absolutely nothing to me, Nokia died several years ago for us. Sep 03 05:41:17 * jpinx has a spare working n900 for sale :) Sep 03 05:41:30 how much? Sep 03 05:42:27 I'm still happy with mine *knock on wood* Sep 03 05:42:51 and my backups include sailfish, firefoxos and tizien(or whatever you spell it like) Sep 03 05:43:14 Wulfe: $150 US ? Sep 03 05:43:43 meh, i am good. i have found them for a bit cheaper. Sep 03 05:43:52 but decent price all the same Sep 03 05:44:57 yea and postage from here would make it more Sep 03 05:45:07 Wulfe: make me an offer ;) Sep 03 05:45:13 yeah S&H really gets ya Sep 03 05:45:41 unless you can beat $110 free S&H i wont be interested. Sep 03 05:45:54 where are you ? Sep 03 05:45:59 place i am looking at has a very limited supply so i may jump on it. Sep 03 05:46:11 midwest usa Sep 03 05:46:46 gimmie a city I can google S&H to ;) Sep 03 05:46:59 kansas city, mo Sep 03 05:47:15 ok Sep 03 05:47:47 what's the favourite S&H people there ? DHL or who ? Sep 03 05:47:49 dont rely on me as a buyer though only reason i have not bought one yet is i am waiting to sell off a tablet for the money first Sep 03 05:48:02 UPS or USPS Sep 03 05:48:07 Wulfe: what tablet you selling ? Sep 03 05:48:12 Wulfe: pm me ;) Sep 03 05:48:20 deal streak 7 with modified charging port Sep 03 05:48:40 i removed the propriatary 30 pin for a mini usb one Sep 03 05:48:56 dell* Sep 03 05:49:44 never heard of dell streak 7 -- googlng nw Sep 03 05:49:48 main reason i am looking to get a n900 is for developing on it Sep 03 05:50:01 its an okay tablet but it has its downsides Sep 03 05:50:07 mainly being a dell Sep 03 05:50:32 dell is like the apple of pc's..... Sep 03 05:50:46 so damn propriatary with everything they do Sep 03 05:51:01 what OS Sep 03 05:51:10 android 3.2 Sep 03 05:51:19 possible to upgrade ? Sep 03 05:51:34 yeah XDA has custom roms to take it to jellybean Sep 03 06:00:56 Wulfe: ok - no sale :) I can buy new tablets with jellybean for $100 here and not have to put up with Dell's oddities (non-removeable battery - wtf is that?) Sep 03 06:01:25 i completely understand lol Sep 03 06:01:46 although the battery can be changed out you just have to open the thing up. Sep 03 06:01:56 but yeah dell blows Sep 03 06:02:19 no biggie though i am sure someone on craigslist will want one Sep 03 06:02:20 Samsung Mega 6.3" here is about $75 Sep 03 06:02:28 christmas coming up and all Sep 03 06:02:59 yeah samsung makes a decent tablet Sep 03 06:03:08 good price to Sep 03 06:04:05 :) Sep 03 06:22:45 DocScrutinizer05: Yes, 1983 in Christian Date Sep 03 06:24:04 Apic: I just think it's extremely funny that ddate binary is bart of core tools and thus on *every* linux ;-D Sep 03 06:24:09 part* Sep 03 06:24:29 Yes, thanks to Slackware, the very first Distro, and the Church of the Subgenius. Praise Bob! Sep 03 06:25:09 love it :-) Sep 03 06:25:25 * Apic 2 Sep 03 06:28:35 who didn't see this coming: http://www.usatoday.com/story/money/business/2013/09/03/microsoft-buys-nokia-handset/2756653/ ? Sep 03 06:29:43 nobody ever did :P Sep 03 06:29:56 it was not a question of 'if' but rather 'when' Sep 03 06:30:09 exactly Sep 03 06:30:22 he was always seen as a trojan Sep 03 06:30:26 horse Sep 03 06:30:31 heh Sep 03 06:40:28 seems he did the same thing with Macromedia, joined as ceo three months before they were bought by Adobe Sep 03 07:05:21 Hi, i still need help to solve some dependency issue Sep 03 07:05:21 HAM + CSSU package update Sep 03 07:05:31 where can i get HAM logs ? Sep 03 07:05:42 Does it write some file ? Sep 03 07:07:49 DocScrutinizer05: hi Sep 03 07:24:50 heh Sep 03 07:24:58 i knew this the day they announced elop Sep 03 07:25:30 yeah Sep 03 07:25:57 B-( Sep 03 07:26:00 so many deals with microsoft ends up with the other party getting screwed Sep 03 07:26:41 I wonder if this will get any regulatory bodies interested Sep 03 07:27:03 For Microsoft, I think it's too late actually. Sep 03 07:27:11 hmm? Sep 03 07:27:37 ...though they have been very good at turning around in earlier days. Not so the last 10+ years. Sep 03 07:28:22 I dunno, look at the xbox Sep 03 07:28:31 and Ballmer on his way out may help Sep 03 07:55:15 Embrace, Extend, Extinguish Sep 03 08:20:36 apparently Elop is now a favorite as the monkeys replacement Sep 03 08:23:58 developers! Sep 03 08:24:43 but yeah interested if any of the regulators will stick their nose in this as well Sep 03 08:25:21 embrace, extend, extinguish, elop Sep 03 08:25:39 jaska: LOL Sep 03 09:07:54 So Microsoft-Nokia is now reality :) Sep 03 09:09:59 And Finnish news sites have amusing front pages as a result. It's all Nokia this and Nokia that as far as the eye can see. Sep 03 09:12:01 well. there arent a lot of things in finland to talk about is there? Sep 03 09:12:15 at least in the US we murder each other all day long Sep 03 09:15:12 150 years of Nokia undone in 3 years Sep 03 09:15:21 good job guys Sep 03 09:16:04 so are they releasing th N950 owners from their obligation now? Sep 03 09:18:21 this is the kind of thing that will eventually happen to all shareheld companies sooner or later, if they are not part of a state sponsored industrial complex Sep 03 09:18:58 MS buys Nokia Sep 03 09:19:09 * FIQ saw this coming 2 years ago already Sep 03 09:20:59 so now, there are basically 2 choices for consumer phones, Samsung and Apple Sep 03 09:21:14 they both look the same that I can't tell them apart Sep 03 09:23:46 the difference is that one is made in china and the other is made in korea and chine, one has samsung chips and screens, the other has samsung screens and chips Sep 03 09:25:03 haha Sep 03 09:25:14 the only thing that actually matter is your OS of choice Sep 03 09:25:23 and then you choose device based on that Sep 03 09:26:19 [11:18:22] this is the kind of thing that will eventually happen to all shareheld companies sooner or later, if they are not part of a state sponsored industrial complex Sep 03 09:26:34 yes but usually it dies slowly or gets obsoleted by something else Sep 03 09:26:41 nokia basically went suicidal Sep 03 09:27:01 for a second time FIQ Sep 03 09:28:27 LauRoman: ? Sep 03 09:28:43 fiq first is when they signed the WP deal Sep 03 09:28:53 exclusivity Sep 03 09:28:56 oh I counted this all as a single action Sep 03 09:29:32 as it was bound to be absorbed by MS the moment elop ruined nokia back in feb 2011 Sep 03 09:30:17 nokia shareholders wanted money now and promises of more coming later, that's what they got from MS, and what they got from google was zero money now and aassurances of money later, they just chose "money now" Sep 03 09:31:01 how much are all the shares worth? Sep 03 09:31:18 MS has 86 billion in the bank iirc Sep 03 09:31:44 if this bites MS hard, so much the better :) Sep 03 09:32:05 http://talk.maemo.org/showthread.php?t=91142 Sep 03 09:32:12 8 bil is nothing compared to the motorola deal or the verizon vodafone one Sep 03 09:38:32 hi Sep 03 09:56:03 I wonder what the merger laws are in Finland and whether there is a government agency that will need to give blessing to the Nokia-MS deal in the way it is here in Australia Sep 03 10:00:24 Good job, Elop, mission accomplished :/ Sep 03 10:00:57 jonwil: very likely. Sep 03 10:01:49 however, since Google and Android are the two largest player, and MS share will not be that large at all, I doubt anyone will deny this. Sep 03 10:02:06 eh, Google and Apple, I meant. Sep 03 10:07:16 jonwil, there is no merger, one company is selling a divison to another company. Theoretically it should only be regulated if MS already had a hardware division (or a homonym to whatever other part is included in the deal) in Finland, or if the State held a big enough stake in Nokia that allwoed it to make some decisions based on how many shares it had or if there was an aditional agreement Sep 03 10:08:02 since navteq and many other divisons are not being sold, thre shouldn't be regulation backlash Sep 03 10:10:15 the only thing i fear is that the Lumia division could end up a writeoff in 3-4-5 years Sep 03 10:17:59 if Microsoft buying the Nokia phone division hurts Microsoft, that's a good thing (I have no great love for Microsoft) Sep 03 10:19:02 LauRoman: and why is that bad? :) Sep 03 10:20:25 divVerent, people being layed off because of corporate machinations is what will be attached to that writeoff Sep 03 10:20:40 About the only things I use from Microsoft are Windows (because only Windows can run things I want/need to run like games) and Visual Studio (because I write code that will only be usable in Visual C++ due to ABI issues) Sep 03 10:20:50 well... and where is the difference? Lumia wasn't exactly a success Sep 03 10:21:04 Nokia was struggling anyway, I don't really think it gets WORSE from this for the employees Sep 03 10:21:09 maybe not better either, though Sep 03 10:21:35 jonwil: have to admit MSVS is the best IDE there is... if only it had decent compilers :) Sep 03 10:21:58 What would you consider a better compiler than MS? Sep 03 10:22:05 ANY other compiler Sep 03 10:22:12 e.g. for starters, a C compiler that doesn't STILL ignore C99 Sep 03 10:22:30 The new Visual C++ 2013 compiler supports quite a few parts of C99 Sep 03 10:22:39 does it FINALLY allow variables inside blocks? :P Sep 03 10:23:25 also, a compiler that comes with a runtime that doesn't warn because code is portable and not Microsoft vendor-lock-in code Sep 03 10:23:53 e.g. fopen() is forbidden apparently, need to use the MUCH MORE SECURE fopen_s(). Why is it more secure? Because it DETECTS WHEN YOU PASS NULL AS FILENAME. Sep 03 10:25:22 I wouldn't complain if it were actually more secure, but SOME of the functions in the _s set are just ridiculous Sep 03 10:25:58 divVerent: those are C++ template tricks Sep 03 10:26:18 I know because I worked on adding those _s to mingw-w64 Sep 03 10:26:27 You can turn off all those warnings Sep 03 10:26:35 I do it all the time my in C++ code Sep 03 10:26:41 it really needs to be automated, added half of it in only Sep 03 10:26:54 jon_y: SOME are Sep 03 10:26:56 fopen_s is not Sep 03 10:27:13 many of these actually do some sizeof() tricks to find array sizes, and add a compile time check for buffer overflows Sep 03 10:27:20 kinda like _FORTIFY_SOURCE does in gcc/glibc Sep 03 10:27:32 actually, gcc already has non_null attribute that warns on possible null Sep 03 10:27:59 gcc still best for C99 and the proposed C11 Sep 03 10:28:04 and via fortify (also some nasty template and macro tricks) it adds an extra function arg of the array size if you pass an array buffer Sep 03 10:28:20 yup, I know those Sep 03 10:28:21 if it cannot detect an overrun at compile time already Sep 03 10:28:35 just, MS insisted on breaking a standard just so they can lock-in Sep 03 10:28:41 it directs unsafe to "safe" Sep 03 10:28:58 while glibc managed to do it ALMOST without issues without any compatibility breakage Sep 03 10:29:06 also, RT apparently has nothing to do with security Sep 03 10:29:12 but anyway Sep 03 10:29:17 according to the VLC guys Sep 03 10:29:18 so MS finally has a major part of C99 done? Nice. Sep 03 10:29:23 Took them only... 15 years Sep 03 10:29:27 it does? Sep 03 10:29:27 so I know when they will have C++11 done Sep 03 10:29:38 I thought it does stdint.h only? Sep 03 10:29:42 oh, still? Sep 03 10:29:46 was just told they added more :P Sep 03 10:29:54 Visual C++ 2013 does implement a bunch of the C99 libs Sep 03 10:30:01 afaik, it is just stdint.h Sep 03 10:30:04 does it also add the C99 syntax enhancements? Sep 03 10:30:11 jonwil: what libs? Sep 03 10:30:14 like being allowed to declare variables in the middle of a block? Sep 03 10:30:29 (yes, that enhancement C++ had from the start) Sep 03 10:30:45 I remember my C lecturer keeps insisting that you CANNOT DECLARE variables in the middle of the block Sep 03 10:30:59 then again I've been using gnu89 since forever Sep 03 10:31:00 regarding C11, that is mostly pointless... it adds very little. But also is very easy to implement then. Sep 03 10:31:05 it adds complex.h, stdbool.h, fenv.h and inttypes.h Sep 03 10:31:13 I mainly like that there is finally a standard way for thread local variables Sep 03 10:31:20 that is the whole point of C :) Sep 03 10:31:30 along with new functions in math.h, ctype.h, wctype.h, stdio.h, stdlib.h and wchar.h Sep 03 10:31:41 does ms finally have long double? Sep 03 10:31:44 oh, and gets() FINALLY gone :P Sep 03 10:31:46 in C11 Sep 03 10:31:52 jon_y: huh, it did not? Sep 03 10:32:04 fun fact about long double: it'll soon be gone from the CPUs Sep 03 10:32:05 Visual C++ current version does not have that Sep 03 10:32:12 AMD and Intel decided to get rid of it Sep 03 10:32:14 no, it did not, sizeof(double) == sizeof(long double) Sep 03 10:32:22 of the 80bit float type Sep 03 10:32:23 The new one that has been released as a beta has this new header support Sep 03 10:32:27 because SSE doesn't support it Sep 03 10:32:30 mingw* had mingwex extensions to do 80bit float Sep 03 10:32:39 yes, via x87 FPU instructions Sep 03 10:32:45 which are deprecated and announced to break "soonish" Sep 03 10:32:53 so that OS kernels will have to SW emulate 80bit floating point Sep 03 10:33:00 is it normal that I can't edit the settings for maemo.org repo in HAM? Sep 03 10:33:06 ehm I don't think SW is going to break Sep 03 10:33:15 ok, found the URL from the MS blog with the details of C99 support in Visual C++ 2013 Sep 03 10:33:15 jon_y: not break per se Sep 03 10:33:16 http://blogs.msdn.com/b/vcblog/archive/2013/07/19/c99-library-support-in-visual-studio-2013.aspx Sep 03 10:33:23 OS kernels will have to implement "assists" for it Sep 03 10:33:27 basically, emulate them Sep 03 10:33:29 well, Intel isn't going to break CPU support that is Sep 03 10:33:37 they will raise illegal instruction, then the kernel steps in and the application won't notice Sep 03 10:33:42 is how they want it to be Sep 03 10:33:52 yes, I know about trapping errors Sep 03 10:34:07 just... a precondition for that is breaking support for old OSes :P Sep 03 10:34:09 like DOS Sep 03 10:34:16 normally, MSVC long double functions are exactly the same as double Sep 03 10:34:18 so I would suppose a CPU that will do this will also be UEFI-only Sep 03 10:34:23 while at it, let's kill real mode plz? Sep 03 10:34:43 so mingw* has a different ABI than msvc with floating points Sep 03 10:34:53 Does GCC support static thread-local-storage on Windows (i.e. __declspec(thread) or whatever GCC has decided to call it) Sep 03 10:34:59 divVerent: it will only execute signed code :) Sep 03 10:35:02 yes it does Sep 03 10:35:05 __thread Sep 03 10:35:30 __thread int aaa; Sep 03 10:35:36 ok, great, I remember many years ago I got a copyright assignment for GCC expressly with the intent of trying to add that support Sep 03 10:35:45 yes Sep 03 10:35:46 BTW, I only recently had to use long double in new code Sep 03 10:35:46 I wanted to do some float math on 64bit integers Sep 03 10:35:46 they JUST fit into 80bit long doubles Sep 03 10:35:49 it relies on libgcc and windows tls slots Sep 03 10:35:59 but, that code won't complain if it were to be SW emulated Sep 03 10:36:20 on MSVC, long double is just double Sep 03 10:36:26 really just 64bit afaik Sep 03 10:36:47 fun, C11 allows conforming implementatiosn to have no variable length arrays? interesting Sep 03 10:36:58 that would be MSVC Sep 03 10:37:08 does this also imply no alloca()? Sep 03 10:37:12 as otherwise it's pointless to do that Sep 03 10:37:14 probably Sep 03 10:37:29 if you want C, use gcc :) Sep 03 10:37:33 admittedly, a C++ compiler can emulate alloca() by other means :P Sep 03 10:37:45 oh no, new[] Sep 03 10:37:51 i.e. on the heap, and with an automatic destructor Sep 03 10:38:06 auto_ptr? Sep 03 10:38:10 i.e. turn char *a = alloca(1234) into unique_ptr a = new char[1234] Sep 03 10:38:20 basically, yes Sep 03 10:38:46 yeah, C++ is still looks pretty scary to me Sep 03 10:39:02 implementation sucks Sep 03 10:39:03 one other thing I really don't get in C11 Sep 03 10:39:05 is quick_exit() Sep 03 10:39:10 do we really need a third exit path? Sep 03 10:39:13 hmm, what does it do? Sep 03 10:39:26 like exit() but with a second atexit() function chain Sep 03 10:39:28 freemangordon, is it normal that I can't edit the settings for maemo.org repo in HAM? Sep 03 10:39:56 so it doesn't run the regular atexits, but a normally shorter chain of at_quick_exit functions Sep 03 10:40:01 probably added by some embedded manufacturer Sep 03 10:40:45 there is also the C11 decimal floating point fun Sep 03 10:40:57 decoding 128bit decimals is fun Sep 03 10:41:10 *binaries Sep 03 10:41:38 I burnt out a bit after writing the decoder for printf* Sep 03 10:42:02 %x was the hardest, 3bits at a time, really!? Sep 03 10:43:07 I don't even get how float %A even works Sep 03 10:45:40 bennypr0fane: yes Sep 03 10:45:42 is it normal for the N900's bundled headphones to sound wrong when connected to other stuff (stuff that doesn't use the microphone contacts at least)? and strangely, it sounds right again if i keep the headphones' button pressed Sep 03 10:47:03 Hmmm, it sounds like the TLS support is not done the way Visual C++ does it Sep 03 10:47:26 how does tls do it? Sep 03 10:47:35 Visual C++ does it using the TLS directory in the PE file Sep 03 10:47:40 using .tls section yes? Sep 03 10:47:41 and a field in the TEB Sep 03 10:47:51 oh ok so it does do it the same way as Visual C++ then? Sep 03 10:47:52 ah ok, mingw* does it too Sep 03 10:48:11 except just using it for ctor/dtor to run libgcc TlsAlloc Sep 03 10:48:11 what does it need from libgcc then? Sep 03 10:48:29 could soembody help me out with the printing on the SoC of OMAP3630 !GB RAM boards? Sep 03 10:48:36 1GB even Sep 03 10:48:44 hmmm ok Sep 03 10:48:58 libgcc will be responsible for the actual initialization, mingw* just provides the ctor/dtor to do it Sep 03 10:49:17 I mean, all that code was already in libgcc, no sense to rewrite it Sep 03 10:50:58 The way I was trying to do it way back when was to basically copy what Visual C++ does and store each thread local variable in the .tls segment and access each variable via TEB->0x2C Sep 03 10:51:22 Where I got stuck was finding a way to make GCC output the correct ASM (which uses the FS or GS segment IIRC) Sep 03 10:53:05 oh, goody! I installed kernel-power 52, no I'm unable to boot! looks like another reflash is in order... so close... Sep 03 10:53:11 BTW Sep 03 10:53:30 is there a "trick" to make a ThinkPad/Nexus4/iPhone headphone set work with the N900? Sep 03 10:53:53 by default, it sounds AWFUL because apparently it confuses mic/gnd pins and puts voltage on the gnd, which even changes a lot (about 10hz tone) Sep 03 10:53:53 divVerent, you mean a wired one? Sep 03 10:53:56 yes Sep 03 10:54:01 I want only the headphone part to work... Sep 03 10:54:10 so is there a way to turn off mic voltage feed (which I suppose is the cause)? Sep 03 10:54:37 other than by using duct tape :P that actually MIGHT work Sep 03 10:55:02 afaik wired headsets can't remote control N900, but I've never heard about the headphone function being off Sep 03 10:55:17 I only ever use bluetooth ones with n900 Sep 03 10:55:31 I do note though that MingW still doesn't support SEH (as implemented by the kernel on Win32) without extra patches or libs Sep 03 10:55:31 I only want to hear audio :) Sep 03 10:55:40 Or has that changed too anyway? :) Sep 03 10:55:49 jonwil: SEH isn't :P Sep 03 10:56:12 or rather, never understood what is "structured" in it Sep 03 10:56:13 divVerent: did you see my question above? it looks like the complement of what you asked Sep 03 10:56:26 divVerent, what is gnd? Sep 03 10:56:55 bennypr0fane: trying to figure out Sep 03 10:57:03 which pin is SUPPOSED to be ground, and which IS, on these headphones vs n900 Sep 03 10:57:21 because I suspect the issue is that there are two ways to assign Ground and Mic to the pins :P Sep 03 10:57:35 so if I can make the N900 ground the Mic pin, it should work Sep 03 10:57:49 so you guys say, n900 has the pins set up, so only headphones with the same confusion can work right with it? :-D Sep 03 10:58:17 *n900 has the pins set up wrong Sep 03 10:58:44 aha, "gnd" stands for ground Sep 03 10:58:49 ah, yes Sep 03 10:59:00 N900 has left right mic ground order Sep 03 10:59:04 US standard is left right ground mic Sep 03 10:59:21 N900 isn't "special" here, Nokia did use the same order on 2.5mm plugs for ages Sep 03 10:59:40 so basically I wonder if there is a software way to turn off the mic voltage Sep 03 11:00:00 so is that like standard for europe, or just standard for Nokia? Sep 03 11:00:03 jonwil: SEH as in exceptions? Sep 03 11:00:16 or do you mean calling __try explicitly? Sep 03 11:00:56 as for the former, like C++ exceptions, is supported in 64bit Sep 03 11:01:07 as for 32bit, just waiting for patents to expire Sep 03 11:02:11 bennypr0fane: I doubt a standard even existed before Sep 03 11:02:27 (for those who didn't notice: microsoft bought nokia, it seems, today.) Sep 03 11:04:42 HAHA, just MAYBE answered my own question Sep 03 11:05:10 alsamixer has a "jack function" item Sep 03 11:05:20 it was TV-Out when I plugged the headset in Sep 03 11:05:33 any other setting doesn't have that buzzing Sep 03 11:07:20 don't know if it'll actually play though Sep 03 11:08:03 Yet another thing to add to the list of things being held back by software patents... Sep 03 11:08:11 (win32 SEH in GCC) Sep 03 11:08:31 it expires in 2014 Sep 03 11:08:38 so, soon Sep 03 11:11:42 wait, really? Win95 had SEH? Sep 03 11:11:49 or did they patent it then not use it for ages? Sep 03 11:11:58 Microsoft doesn't hold the patent IIRC Sep 03 11:12:10 IIRC its Borland that does Sep 03 11:12:35 or whoever inherited it from the corpse of Borland Sep 03 11:12:57 As a software developer I think software patents are stupid period for all sorts of reasons. Sep 03 11:13:06 "depends" but usually this is the case Sep 03 11:13:18 especially patents on file format reading just shouldn't exist, ever Sep 03 11:13:19 go NZ Sep 03 11:13:36 gry MS didn't buy all of Nokia, just their "devices&services" Sep 03 11:14:00 I wonder if rubber boots are "devices" or not Sep 03 11:14:08 wading device Sep 03 11:14:20 but tires and such were spun off much earlier Sep 03 11:14:26 bennypr0fane: yes, that's what I meant. sorry for being vague. (that was the main part of it that I knew of.) Sep 03 11:14:42 The patents on things like MPEG or Elliptic Curve Cryptography are definatly one of the kinds of patents that really need to go away Sep 03 11:15:42 divVerent, jonwil , SEH is this? http://msdn.microsoft.com/en-us/library/swezty51%28v=vs.90%29.aspx Sep 03 11:16:08 jonwil: sort of, indeed Sep 03 11:16:27 SEH in this case is this Sep 03 11:16:28 I understand that patents on encoding algorithms MIGHT be useful. Decompression of videos however MUST be free. Sep 03 11:16:29 http://www.microsoft.com/msj/0197/Exception/Exception.aspx Sep 03 11:16:38 jonwil: yes, borland patents Sep 03 11:16:40 But even if they are useful, 20 years is INSANE in this field. Sep 03 11:16:48 though they promised not to sue Sep 03 11:18:45 divVerent, I'm interested in that "patents on file format reading" area. It seems to me that Linux (-applications) cannot read all the metadata of MS office files, specifically author, keywords,... the classic set also attached to all PDFs Sep 03 11:19:02 you guys know anything about that? Sep 03 11:20:01 I keep wonderin if that's anything to do with the closed file MS formats, or rather with the file systems Sep 03 11:20:20 *closed MS file formats Sep 03 11:21:14 divVerent, I think the way patents are handled by the US jurisdiction is just generally insane overall Sep 03 11:21:16 bennypr0fane: no, that isn't a patent issue Sep 03 11:21:21 more of a lack of useful documentation issue Sep 03 11:21:27 hmmm Sep 03 11:21:37 bennypr0fane: especially bad are "unavoidable patents" for supporting a file format Sep 03 11:21:39 e.g. S3TC Sep 03 11:21:49 I wonder if the final one of the S3TC patents is finally killed? Sep 03 11:21:56 three out of four were down'd by Apple Sep 03 11:22:11 Patents on file SYSTEMS are the lowest of the low and really need to go Sep 03 11:22:23 FAT is still patented Sep 03 11:22:29 MS trying to push exFAT Sep 03 11:22:29 you mean VFAT :P Sep 03 11:22:33 i.e. all the various FAT type patents Microsoft is so fond of using to kill their competitors Sep 03 11:22:37 original FAT is expired Sep 03 11:22:44 but the long file name extensions aren't Sep 03 11:22:50 well, exFAT going to go another 20 years Sep 03 11:22:55 WUT? FAT is patented? Sep 03 11:22:55 yes Sep 03 11:23:00 FAT was, but is too old now Sep 03 11:23:09 only the Win95-introduced LFN extension still has SOME validity time left Sep 03 11:23:18 yeah, original DOS FAT is no longer under patent Sep 03 11:23:26 VFAT aka long file names is Sep 03 11:23:37 FAT32 also, probably Sep 03 11:23:45 who the hell would *want* a patent for vfat Sep 03 11:23:48 no, FAT32 did not add anything vs FAT Sep 03 11:23:52 nothing to patent there Sep 03 11:23:54 i'd be ashamed of it Sep 03 11:23:55 VFAT is the issue, though :P Sep 03 11:23:57 kerio: hehe Sep 03 11:24:04 It seems like Microsoft will continue to have device makers and memory card makers over a barrel because they will never support any filesystems on Windows that they dont own and control Sep 03 11:24:14 therefore device makers have to use the MS file systems (i.e. fat and now exfat) Sep 03 11:24:17 and pay MS the royalties Sep 03 11:24:26 jonwil: that's the idea Sep 03 11:24:38 the only ones who would have a chance to achieve something else is Apple... Sep 03 11:24:42 but then it's THEIR patented crap Sep 03 11:24:49 not only that Sep 03 11:24:53 apple sucks at file systems too Sep 03 11:25:04 HFS+ is such a bad filesystem Sep 03 11:25:09 how long have all you guys been in programming? I'm starting IT school right now, yesterday was my first day Sep 03 11:25:10 that too Sep 03 11:25:18 kerio: fun fact: you can make working .dmg files on Linux Sep 03 11:25:21 tool: mkisofs :P Sep 03 11:25:24 bennypr0fane: 4 years? :) Sep 03 11:25:28 OS X mounts ISOs perfectly fine if renamed to .dmg Sep 03 11:25:36 even WITHOUT the HFS extensions Sep 03 11:25:37 I'm really excited Sep 03 11:25:40 OS X mounts ISOs perfectly fine, period Sep 03 11:25:47 the default action for .iso files is to mount them Sep 03 11:25:58 kerio: oh right... still .dmg is needed so the apple fanboys are happy Sep 03 11:25:59 bennypr0fane: make sure you learn C, don't give a crap about those hipsters that say C is outdated Sep 03 11:26:16 it is the very foundation to understand what is going on behind the JVM Sep 03 11:26:20 C is outdated, but so? Doesn't make it not useful. Sep 03 11:26:21 jon_y you mean 4 years work, or 4 years ago you started your education? Sep 03 11:26:22 jon_y: he's starting IT school, he's not going to learn how to program Sep 03 11:26:22 Exactly. Sep 03 11:26:36 4 years since I learned programming Sep 03 11:26:38 You won't write a whole web forum in C, for example :P Sep 03 11:26:41 he'll probably learn how to program in spite of his IT school Sep 03 11:26:43 you COULD, but it'd be stupid Sep 03 11:26:47 * bennypr0fane heard real men code in C :-D Sep 03 11:26:50 divVerent: and by stupid, you mean AWESOME Sep 03 11:26:56 that too Sep 03 11:26:56 of course you won't do it in C, you get the basics in C Sep 03 11:26:59 but total waste of time Sep 03 11:27:00 first time I programmed some 23 years ago or something, but never really worked as a programmer (I'm concerned I would be too stressful for me) ;p Sep 03 11:27:14 and for goodness sake, read the linkers and loaders pdf too Sep 03 11:27:18 yes Sep 03 11:27:23 the knowledge is important Sep 03 11:27:27 of how stuff actually works Sep 03 11:27:31 understand wtf is going on when you run programs Sep 03 11:27:34 I am a programmer by profession although I have not been able to find a job in my area for a while Sep 03 11:27:35 or you e.g. use C++ then don't understand why something weird is happening Sep 03 11:27:36 linkers and loaders pdf? Sep 03 11:27:54 it is free, find it on google Sep 03 11:27:58 like "why can't I just return a reference to a local variable?" Sep 03 11:28:02 it's good to know low-level stuff Sep 03 11:28:09 teach you about executable file formats Sep 03 11:28:18 if only to remind you there's something else under all that crap Sep 03 11:28:23 jonwil, that is unusual, no? Sep 03 11:28:31 so far it covers linux a.out, elf, win32 PE, and DOS exe and COM Sep 03 11:29:07 oh, if you have the time, learn about CPU architecture design too Sep 03 11:29:21 know it when sales people are bullshitting you on hardware Sep 03 11:29:22 It's always seemed to me (from the outside, so far) that low-level programming is the most highbrow studd to understand Sep 03 11:29:36 highbrow how? Sep 03 11:29:39 The problem I am finding is that all the people offering programming jobs in my area are asking for 2 or 3 or 5 years experience in technologies I dont really understand Sep 03 11:29:44 *stuff Sep 03 11:29:59 jonwil: opensource doesn't count? :( Sep 03 11:30:07 jonwil, which technologies? Sep 03 11:31:03 is it this one? http://norfs.sourceforge.net/linkers_and_loaders.pdf Sep 03 11:31:06 Stuff like .NET C# SSIS CRM Oracle J2EE MVC ASP and some others I cant remember off the top of my head Sep 03 11:31:07 yes Sep 03 11:31:20 ew, J2ME Sep 03 11:31:27 *J2EE Sep 03 11:31:33 oh and Sharepoint too Sep 03 11:31:45 13:31:06 jonwil | Stuff like .NET C# SSIS CRM Oracle J2EE MVC ASP and some Sep 03 11:31:45 you can't code anything other than Windows programs in .NET, can you? Sep 03 11:31:47 hehe Sep 03 11:31:48 I worked on this Java codebase before, it was so enterprisy Sep 03 11:31:59 of these, I would probably not admit the J2EE part :P Sep 03 11:32:16 class inheriting classes Sep 03 11:32:20 I never list Java related stuff on my resume because I hate the Java runtime library Sep 03 11:32:25 base class did not have any prublic methods Sep 03 11:32:36 don't want to code Java all day\ Sep 03 11:32:38 and using instanceof() in derived classes Sep 03 11:32:43 egads Sep 03 11:32:49 why do you even do that Sep 03 11:32:52 using instanceof explicitly? Sep 03 11:32:56 yes Sep 03 11:33:00 well, first... SOME uses are there Sep 03 11:33:06 this code should be on thedailywtf Sep 03 11:33:08 e.g. if you have a derived method taking another object Sep 03 11:33:20 divVerent: it was doing if-else instanceof Sep 03 11:33:21 you may want to know if the other object is of your derived class or if you only can assume the base class Sep 03 11:33:26 hehe Sep 03 11:33:32 instanceof of ANOTHER than your current class is weird Sep 03 11:33:45 but e.g. derived Equals methods tend to do check for the other object to be instanceof your own class Sep 03 11:33:52 if(instanceof) dothis else if(instanceof) dosomething ..... Sep 03 11:34:04 I hope they didn't do this in the base class, even Sep 03 11:34:08 long long chains for each branch of the inheritance Sep 03 11:34:16 base class didn't have anything Sep 03 11:34:20 the base class containing an instanceof chain mention all subclasses... Sep 03 11:34:23 no public methods Sep 03 11:34:23 that would be DailyWTF worthy Sep 03 11:34:33 In your opinion, what is the best channel to hang out in for getting help with newbie issues? And the kind of questions I asked above? Sep 03 11:34:40 more DailyWTF score if the subclasses are all empty Sep 03 11:34:44 all this was just a tree structure implementation Sep 03 11:34:45 and all logic is done by instanceof in the base class :P Sep 03 11:35:01 this is almost as good as the for-case paradigm then :P Sep 03 11:35:13 also, it had add() but no remove() in the derived classes Sep 03 11:35:19 haha, really? Sep 03 11:35:22 yes Sep 03 11:35:31 well, remove() not defined in the parent then, I hope Sep 03 11:35:33 oh we're agile, we don't need it yet Sep 03 11:35:44 as I said, base class had nothing Sep 03 11:35:50 then it's actually SOMEWHAT ok... Sep 03 11:35:56 except for the base class has nothing part Sep 03 11:35:59 why have that class then Sep 03 11:36:02 just a child array of objects Sep 03 11:36:10 I asked the same thing Sep 03 11:36:26 that also means the abstract base class didn't implement an interface defining the methods, right? Sep 03 11:36:30 got no response other than because it was there already Sep 03 11:36:32 not that I think you guys can't help me, but it's obviously more for experienced people in here :-) (besides the fact that we've been off topic for hours) Sep 03 11:36:58 it only had a private array of object called "child" Sep 03 11:37:08 or was it protected I can't remember Sep 03 11:37:12 jon_y: haha... maybe anonymize that in a few months and post to dailywtf? Sep 03 11:37:28 yeah, I don't even want to dig it up Sep 03 11:37:38 thankfully I don't work on it anymore Sep 03 11:38:14 the program was a monstermash of TCSH, BASH, Java, Perl and C Sep 03 11:38:14 speaking of dailywtf: recently sent this to them: view-source:http://www.led24.ch Sep 03 11:38:20 but probably won't make it to the frong page Sep 03 11:38:40 the program was converted from C++, so I heard Sep 03 11:38:53 but half way through they found out Java had no fork() Sep 03 11:39:10 they apparently wanted to make sure that the newline at the bottom of the HTML source is in Arial 20, and not in Times 24 like the rest Sep 03 11:39:17 nor did it have any PID/process management calls Sep 03 11:39:19 jon_y: haha, Java and fork()... Sep 03 11:39:28 reminds me of something :P Android-related... Sep 03 11:39:30 and OpenSSL-related Sep 03 11:39:30 you know what they did? :) Sep 03 11:39:51 it's public, the android openssl thing that led to bitcoins being stolen Sep 03 11:39:59 ran a process starter on itself and restarted itself again Sep 03 11:40:01 long story short: Android loads OpenSSL, then uses fork(). Sep 03 11:40:05 And OpenSSL is not fork-safe Sep 03 11:40:21 as in, the random number generator state is just cloned too Sep 03 11:40:26 so all apps had the same RNG state Sep 03 11:40:45 well duh, static structures are cloned Sep 03 11:40:50 yes Sep 03 11:40:53 and the best part Sep 03 11:41:00 that would be the whole point of fork() Sep 03 11:41:00 this very same bug also hit PostgreSQL and Ruby Sep 03 11:41:16 yes, there was in all of them lack of special handling for invalidating the state on fork Sep 03 11:41:21 e.g. by adding more randomness from /dev/urandom Sep 03 11:41:43 >counting on urandom for anything serious Sep 03 11:41:45 what. Sep 03 11:41:50 that's kinda ok in this case Sep 03 11:41:56 it only matters that the mixed in state is different :P Sep 03 11:42:03 in this specific case, that is, not in general Sep 03 11:42:19 even using the PID would be secure enough, assuming a cryptographically secure PRNG Sep 03 11:42:31 well... and what did OpenSSL do? It DID use the PID... Sep 03 11:42:40 however, PIDs can repeat, which is why exactly this does not suffice Sep 03 11:42:45 an ever increasing counter would have done then Sep 03 11:42:49 or using the PID at the right time Sep 03 11:42:57 e.g. mixing in the child process's PID when the parent forked Sep 03 11:43:14 oh, there is another wtf from the java program, well on the current maintainer anyway Sep 03 11:43:19 thing is, if a RNG is cryptographically secure, and you fork the state Sep 03 11:43:25 and one process mixes in 1, the other mixes in 0 Sep 03 11:43:26 for-loops and parallelism, oh boy Sep 03 11:43:36 then an attacker still can't infer anything about the other process's RNG from its own one Sep 03 11:44:05 unlooping the for loops so each elements can be fed into a parallel scheduler is apparently bad Sep 03 11:44:09 for-loops and parallelism, how is that a wtf? Sep 03 11:44:19 oh, wait... a tight for loop handld THAT way? Sep 03 11:44:35 because multithreaded code is hard to read and we need the UI to block to show the user it is doing something Sep 03 11:44:50 why!!!!?? you call yourself a phd and you say that!!! Sep 03 11:44:55 why not just #pragma omp for Sep 03 11:45:04 it is a java program Sep 03 11:45:07 I know Sep 03 11:45:11 just saying... one CAN do it right Sep 03 11:45:20 more like looping down a tree structure Sep 03 11:45:28 oh wait, parallel tree traversal? Sep 03 11:45:30 THAT sounds... insane Sep 03 11:45:37 unlooping involves flattening it first Sep 03 11:45:48 which is already a full tree traversal Sep 03 11:45:51 and the feeding all the elements into a job scheduler Sep 03 11:45:54 I wish more people would take security (including cryptography, SQL injection etc) more seriously Sep 03 11:46:01 jonwil: indeed Sep 03 11:46:01 that was my idea anyway Sep 03 11:46:07 but in this case, I have a hard time blaming anyone Sep 03 11:46:10 regarding OpenSSL vs fork Sep 03 11:46:18 nope, lead developer insists on using for loops and recursive calls Sep 03 11:46:27 OpenSSL guys say they want portable code, no unixisms like doing something odd to catch fork use (and be it by pthread_atfork) Sep 03 11:46:54 and Android/Postgres/Ruby guys... well Sep 03 11:47:02 OpenSSL kinda HID the issue by mixing in the PID on each random call Sep 03 11:47:13 granted I'm not majoring in compsci, but guys, why you hate multithread so much? Sep 03 11:47:14 so they quickly tested, saw "ah, the state differed", and were happy Sep 03 11:47:24 jon_y: I don't hate it Sep 03 11:47:27 I hate ugly code from it Sep 03 11:47:32 but one can write it well too Sep 03 11:47:38 divVerent: I mean the lead developer Sep 03 11:47:43 I know Sep 03 11:47:44 openssh abuses every cute little feature of openbsd, on the other hand Sep 03 11:47:51 he has once seen really bad threaded code Sep 03 11:47:58 and then hated threads in general Sep 03 11:48:12 how hard can flattening a tree structure be? Sep 03 11:48:18 jon_y: what's wrong with urandom? :o Sep 03 11:48:22 JUST the flattening? Easy. Sep 03 11:48:25 isn't it guaranteed to be crypto-secure? Sep 03 11:48:32 kerio: "sort of" Sep 03 11:48:34 it's still a PRNG Sep 03 11:48:41 kerio: urandom is not guaranteed to have enough randomness Sep 03 11:48:43 but I would still trust it over whatever some library provides - and be it OpenSSL Sep 03 11:48:50 jon_y: neither is any other PRNG Sep 03 11:48:52 like OpenSSL's Sep 03 11:48:56 P stands for pseudo after all Sep 03 11:49:01 random would block if there isn't enough randomness Sep 03 11:49:04 yes Sep 03 11:49:07 if you want that, do that Sep 03 11:49:17 yeah Sep 03 11:49:34 in this specific case, I probably would rather have read 16 bytes of /dev/random on fork Sep 03 11:49:38 urandom is "give me something, I don't care how random it is" Sep 03 11:49:42 128 bits of extra entropy per process is probably enough Sep 03 11:49:45 and usually doesn't block Sep 03 11:49:46 even if it is not Sep 03 11:49:49 jon_y: but it's crypto-secure! Sep 03 11:49:53 :c Sep 03 11:49:57 it says so Sep 03 11:50:02 urandom is about as crypto secure as OpenSSL's :P Sep 03 11:50:08 both shouldn't claim it Sep 03 11:50:45 oh, btw, the java program calls a small C program that read timestamps etc Sep 03 11:50:57 the lead programmer said the java code did it too slowly Sep 03 11:51:04 really, you wonder why Sep 03 11:51:25 unfortunately, the C code was written by an intern that has long left Sep 03 11:51:55 using glib because he didn't know how to do malloc/free memory management Sep 03 11:51:57 what video formats/codecs are supported on maemo? Sep 03 11:52:07 nor know how to concaternate strings in C Sep 03 11:52:21 ccxCZ: i think only h2xx is supported video hw decoding Sep 03 11:52:47 but you can always install mplayer for everything albeit at the expense of cpu Sep 03 11:53:16 kthnx Sep 03 11:53:22 I have also seen sudo-exec wrappers that join argc together and pass it to system() Sep 03 11:53:27 ccxCZ: there is a site about it Sep 03 11:53:28 so I guess avi for the container Sep 03 11:53:29 uhm... Sep 03 11:53:35 ~video Sep 03 11:53:35 extra, extra, read all about it, video is interrupt 10h for the video BIOS Sep 03 11:53:40 ok that's not it Sep 03 11:53:56 https://wiki.maemo.org/Video_encoding Sep 03 11:53:58 there it is Sep 03 11:59:57 thanks Sep 03 12:29:14 13:51:55 jon_y | using glib because he didn't know how to do Sep 03 12:29:16 | malloc/free memory management Sep 03 12:29:23 jon_y: wait Sep 03 12:29:26 glib doesn't solve this either Sep 03 12:29:31 so you SURE have a leak there :P Sep 03 12:30:36 13:53:22 jon_y | I have also seen sudo-exec wrappers that join argc together and pass it to Sep 03 12:30:38 | system() Sep 03 12:30:46 jon_y: ARGH... TOTALLY reminds me of an old project Sep 03 12:30:58 task: they are in PHP, and want to send a command to a serial port Sep 03 12:31:28 problem: they don't know that fopen() works on devices too, and can't access it from PHP anyway Sep 03 12:31:41 their solution: run apache process as root, and PHP code like this: Sep 03 12:31:52 open /tmp/temp via fopen Sep 03 12:32:02 write to it: "echo $stuff > /dev/ttyS1" Sep 03 12:32:09 system("chmod +x /tmp/temp") Sep 03 12:32:12 system("/tmp/temp") Sep 03 12:32:26 do I have to mention that $stuff also came from the user and was not sanitized... Sep 03 12:33:13 also, obviously there was nothing to prevent the obvious race condition(s) Sep 03 12:33:43 the only GOOD part: this stuff isn't connected to the internet, usually. Only to a home automation network. Sep 03 12:34:20 but it does beat the "internet cafe" I r00ted via netscape (to get rid of the restrictions and timer of free use) Sep 03 12:34:56 they had restrictions via a proxy setting in netscape... but there was ONE trick. ONE link on the mainpage opened in new window. That window had a menu bar, so I could turn off the proxy. Voila, free access. Sep 03 12:35:28 then used the window's address bar to look around... essentially associated *.sh with the shell in netscape settings, and used JavaScript:'some text' then File/save as to create scripts I could run Sep 03 12:35:47 the timer daemon ran as another user... not bad Sep 03 12:35:50 but as root... a bit worse Sep 03 12:36:03 but the kiosk user had kiosk ALL=(ALL) ALL in /etc/sudoers... THAT was bad Sep 03 12:36:21 so having found that, kill -STOP timerpid and the thing was free :P Sep 03 12:36:27 (with sudo) Sep 03 12:37:12 oh, I forgot the NOPASSWD: part :P Sep 03 12:39:46 divVerent: oh now I remember Sep 03 12:39:51 am i supposed to confirm somewhere (and if so, where) that the cssu-devel fix for SUPL works? Sep 03 12:40:01 wait, glib does do some momory management stuff Sep 03 12:40:16 a long with some utility code Sep 03 12:41:14 LjL-Alps: oncssu-devel thread? Sep 03 12:41:21 ~cssu-devel Sep 03 12:41:22 extra, extra, read all about it, cssu-devel is http://talk.maemo.org/showthread.php?t=84292, or extremely deprecated for any non-developer Sep 03 12:42:09 freemangordon: ok, i didn't know whether an influx of "me too" would be welcome there Sep 03 12:42:23 divVerent: well, I did not have your level of exploit :) Sep 03 12:43:09 most of this stuff was written by guys used to writing shell scripts Sep 03 12:43:32 I'm tempted to give "script kiddie" another meaning Sep 03 12:44:28 alt meaning is "guys who write shell scripts, should not be trusted with anything compiled" Sep 03 12:44:47 LjL-Alps: hmm? wouldn't be wecomed by whom? Sep 03 12:45:26 freemangordon: by TMO at large. anyway i guess maybe if i just say "it works" here in your presence, that'll be enough feedback? :P Sep 03 12:45:29 the whole point of that thread is to report cssu-devel related stuff Sep 03 12:45:40 fair enough Sep 03 12:45:50 i'll try supl.google.com and then give it a post Sep 03 12:46:06 no change for google supl Sep 03 12:46:14 it still doesn;t work Sep 03 12:46:20 ah, no change? some people on the thread said it started working too Sep 03 12:46:36 NFC, it has never stopped working for me :D Sep 03 12:46:44 oh :P Sep 03 12:47:06 but that might by just my MNO supl working, not google Sep 03 12:47:24 your what, freemangordon? Sep 03 12:47:38 mobile network operator Sep 03 12:47:42 oh Sep 03 12:47:59 i thought Maemo only used the one specified in the settings Sep 03 12:48:05 no Sep 03 12:48:32 it uses network supplied supl server too, if any Sep 03 12:48:46 but first tries the one in the settings Sep 03 12:49:25 i wonder if carriers charge for data usage (assuming you don't have a flat plan) when connecting to their own supl Sep 03 12:50:41 nfc, i have data plan Sep 03 12:50:54 me too, just idly wondering Sep 03 12:51:13 I guess it is similar to MMS Sep 03 13:14:24 i wish i had brought a usb hub with me, or whatever would work to connect to the RTLSDR Sep 03 13:16:10 except, i have!!! Sep 03 13:16:28 host mode here i come Sep 03 13:18:21 yayayay it works Sep 03 13:22:18 news about my trying to make an unsupported headset work via alsamixer: Sep 03 13:22:23 the mic just can't be used, as expected Sep 03 13:22:34 the headphones work, but the mic port can only be set to float, not ground Sep 03 13:22:47 by that alsamixer setting that decides what type of device is plugged in Sep 03 13:23:00 basically means no stere, but I hear difference of L and R Sep 03 13:23:15 LjL-Alps: do you use DVB-T dongle with it? Sep 03 13:23:17 still... not bad, works as "emergency headphones" that way Sep 03 13:24:41 BTW, hint to those who don't know: alsamixer -c 0 is the trick to get useful settings and not pulseaudio crap :P Sep 03 13:24:43 edheldil: that's the idea, but so far i've only checked that the DVB-T dongle powers up, and that some MP3 player goes into mass storage mode. still not tried anything else Sep 03 13:26:13 ah, I thought you were interested in the RTLSDR part :) Sep 03 13:29:10 edheldil: i am Sep 03 13:29:24 edheldil: i'm just saying, so far i didn't get RTLSDR working. only hostmode. but that's a start1 Sep 03 13:29:37 especially since i didn't think i had a cable that could do USB-OTG Sep 03 15:30:01 ugh.. Sep 03 15:30:15 so it's happened :( Sep 03 15:31:02 Micro-Nokia-soft Sep 03 15:35:12 Me, I will continue to use (and hack on and play with) my N900 until it either dies and becomes unusable, until something happens at my carrier that renders it unusable (e.g. if my carrier stops using frequencies the N900 supports) or until a better phone comes along (which at this point is highly unlikely) Sep 03 15:36:11 me too jonwil. my N900 is my daily device. i still have a working n-gage too, and a 9300i communicator which i gave away, but it's found it's way back to me. Sep 03 15:36:30 * Sicelo feels like removing the nokia logo from the bezel Sep 03 15:36:36 My N900 is also my daily (and only) phone. Sep 03 15:37:18 and good to see you jonwil :) Sep 03 15:38:17 I fear that no company will have the willingness to make a high-end smartphone with a full physical QWERTY again anytime soon Sep 03 15:38:48 which means if my N900 dies, my only options are to buy a used N900 as a replacement or give up my love of physical keyboards :( Sep 03 15:38:55 even if they did, might just be so closed that you can't do anything on such a phone Sep 03 15:39:22 yeah that would suck Sep 03 15:42:58 Seems like all the interest in reverse engineering this phone and figuring out how it works seems to have died off, even I cant get inspired to get back into reverse engineering if no-one out there cares about my results Sep 03 15:43:25 what makes you feel that way? Sep 03 15:43:41 just general observations from what I see on tmo etc Sep 03 15:43:50 as far as i can tell, from user POV, interest is still high Sep 03 15:45:00 I have noticed that in the past there was a sort of "hey, if we can replace with something better, that's a good thing" (e.g. ICD or MCE) but now the attitude I observe from the technical people is "forget about it, if it works, why bother trying to replace it" Sep 03 15:45:16 hmm. :( Sep 03 15:46:03 jonwil: probably you means me: I wrote that about closed source ipv6 module for icd2 Sep 03 15:46:10 well, i know for sure that some (Pali, FMG) are still pushing to have stuff replaced, incl. the kernel Sep 03 15:46:22 ah, there's Pali to speak for himself :) Sep 03 15:46:49 I wrote that because it really does not make sense to RE only one closed plugin in big ICD2 Sep 03 15:47:23 and it really does not make sense to RE ICD2 Sep 03 15:47:46 better network daemon should be written from scratch and only needs to have icd2 i2c interface Sep 03 15:47:51 which is of cource public Sep 03 15:48:24 It matters to reverse engineer ICD2 and its plugins as other things talk to it via not just dbus Sep 03 15:48:33 but also via other methods that are not documented Sep 03 15:48:47 so only RE API Sep 03 15:49:17 (documented/undocumented dbus/IPC) Sep 03 15:49:30 Thats what I have been doing over time Sep 03 15:49:42 reverse engineering the interfaces to ICD and its plugins Sep 03 15:49:44 but its hard Sep 03 15:49:49 \o/ jonwil Sep 03 15:49:50 especially when dealing with things like gconf Sep 03 15:50:03 jonwil: freemangordon and ade wanted to RE some hildon code Sep 03 15:50:09 maybe you can help Sep 03 15:50:19 what hildon code specifically? Sep 03 15:50:31 hildon code for notification or something similar Sep 03 15:50:45 bad that he is offline now :-( Sep 03 15:50:52 will try to look at chanlog Sep 03 15:51:40 sounds like hildon-plugins-notify-sv package Sep 03 15:51:53 I think yes Sep 03 15:52:25 I have tried to re that but its too hard :( Sep 03 15:52:34 libhildon-plugins-notify-sv.so Sep 03 15:53:53 jonwil: it is really hard to RE this? ^ Sep 03 15:54:21 last I checked it was annoying to RE, yes Sep 03 15:54:36 ok Sep 03 15:54:54 and systemui or clockd daemon? Sep 03 15:55:00 I guess Microsoft-Nokia news has been already mentioned here gasilion times Sep 03 15:55:02 both also hard to RE Sep 03 15:55:10 but have you seen the true story behind it? https://fbcdn-sphotos-f-a.akamaihd.net/hphotos-ak-ash3/q71/1170764_623665574332364_1926310099_n.jpg Sep 03 15:57:04 yes dos1. i'm late to the party. but the news is still scary enough. anyways Sep 03 15:58:34 what about using connman as icd2 replacement? Sep 03 15:59:27 jonwil, Pali: while on one hand I don't mind to get meamo more open if it's RE'd but on the other hand didn't you already do a lot and upstream pretty much everything? shouldn't you be at the point where you could simply compile a new os without maemo closed bits? Sep 03 15:59:50 We are nowhere near that level Sep 03 15:59:59 Plenty of key parts that have not been REd or replaced Sep 03 16:01:12 jonwil: if you want to stay compatible with maemo or generaly spoken? Sep 03 16:02:00 If you just want "an OS" on the N900, there are ways to do it with minimal binary blobs (GPU driver being one of the few that still remain) Sep 03 16:02:05 Orgul: also N900 hardware itself.. afaik, we don't yet have good power-savings out of Maemo for example Sep 03 16:02:18 that's just a matter of correct configuration though Sep 03 16:02:24 it's not dependent on any closed blob Sep 03 16:02:43 ok. Sep 03 16:05:51 okay, let's set GPU aside. Is it still impossible to do a meamo like os for n900? based on upstream software? I mean I still don't get modem to work in debian but you guys are probably got this. Sep 03 16:06:42 i did do GPRS/D3G in Debian.. that part is easy Sep 03 16:07:14 s/D3/3/ Sep 03 16:07:14 Sicelo meant: i did do GPRS/3G in Debian.. that part is easy Sep 03 16:08:26 so from my point of view all that missing is a nice gui for phone calls. okay, I nice gui in general in debian but I keep wondering why still maemo. Sep 03 16:08:43 someone with greater RE skills than me could probably reverse engineer the x86 versions of hildon-plugins-notify-sv, systemui or clockd Sep 03 16:08:51 but I lack the skills to RE those binaries Sep 03 16:09:21 Orgul: you're actually missing programs that don't drain your battery by using the cpu when backgrounded, for instance Sep 03 16:10:41 not to disturb your train of thought jonwil .. question: what uses the x86 version of the binaries? surely in scratchbox arm binaries are used.. Sep 03 16:10:59 no, x86 is used in scratchbox Sep 03 16:11:45 kerio: ok, yes. debian wise. in case you guys do something new, and use a recent kernel you could implement upstreamed android bits, like wakelocks and stuff. I mean it's there, it could as well be used. Sep 03 16:11:59 yeah, shit like that is the reason android sucks Sep 03 16:12:16 interesting. 10 times wanted to setup scratchbox already.. but each time stopped because i don't really know 'what then?' Sep 03 16:12:24 programs might have a valid reason to do stuff in background Sep 03 16:12:36 i'd expect them to do that stuff, were i using one of those programs Sep 03 16:14:26 kerio: I don't know much about it, but I don't think it impossible to have background application running. It's a matter of configuration. other then that it's a power saving feature. Sorry if I am wrong or imposing. Sep 03 16:14:57 I do wonder what libhildon-plugins-notify-sv is doing though and why it needed to be closed-source Sep 03 16:14:59 not imposing at all :) Sep 03 16:15:25 the thing is, (most) current maemo programs already know how to behave Sep 03 16:15:35 but Orgul, android's way is clearly 'feature-phone' like :p Sep 03 16:15:46 programs installed straight from the debian repos are much less guaranteed Sep 03 16:16:59 Seems mostly to be related to playing sounds and things (e.g. vibra) for events like incoming phone call, incoming SMS, alarm, incoming IM etc. Sep 03 16:19:42 Sicelo: true, but it does save power, doesn't it? it was just an Idea. I am not competent enough to give help anyway. Just thinking. Sep 03 16:20:17 Orgul: the thing is, it implies a lack of trust in the programmer Sep 03 16:21:21 FYI: first hand info about smartreflex aka SR: >>Again, TI and we [Nokia] couldn't fix SmartReflex - we say memory corruption in front of our own eyes with that enabled, so we had to ship that disabled.<< Sep 03 16:22:29 ~sr Sep 03 16:22:29 * infobot grabs a SteamRoller flag and cuddles up to docscrutinizer05 Sep 03 16:22:42 kerio: you mean trust in programmer to design their application to use little power ? Sep 03 16:22:47 eyup Sep 03 16:23:14 ah, ok. that was a bit difficult. glad I got it. Sep 03 16:23:34 ~#maemo sr is [SmartReflex] >>Again, TI and we [Nokia] couldn't fix SmartReflex - we say memory corruption in front of our own eyes with that enabled, so we had to ship that disabled.<< Sep 03 16:23:35 DocScrutinizer05: okay Sep 03 16:25:44 jon_y: possible Sep 03 16:26:01 a nice new puzzle for you ;-) Sep 03 16:27:53 kerio: how about giving infraction points to a process for each call to usleep() ? Sep 03 16:28:30 on 666 infraction points the kernel sends a mail to debian repo to nuke the package off the planet Sep 03 16:30:35 lol Sep 03 16:30:57 also lol regarding the sr quote Sep 03 16:47:36 [/clear Sep 03 16:53:41 what is the practical download speed limit in n900 Sep 03 16:53:57 how blue is the sky? Sep 03 16:54:55 kerio was that a reply to me? Sep 03 16:55:02 yep Sep 03 16:55:17 sky is kinda blue.. and now its black here Sep 03 16:56:11 I think there is a limit at somewhere around 500kbps? Sep 03 16:56:25 in practice.. Sep 03 16:56:32 i don't see why that would be Sep 03 16:57:10 Well, I may try local now..\ Sep 03 17:08:12 Kerio, Well I got it, 1350kbps right. Is there something wrong now on my device? This time its from a local PC via Wlan, Sep 03 17:08:24 The other was my downloads speeds.. Sep 03 17:10:28 actually KB/s Sep 03 17:22:22 qwazix: the front camera is actually VGA, I won;t call that crappy :P Sep 03 17:22:35 it is the way it is used that sucks Sep 03 17:23:13 both skype and gtalk are not using VGA, but 320x240 (QVGA?) Sep 03 17:23:15 it seems to have a number of dead pixels for me Sep 03 17:24:58 qwazix: nikocam can use front cam, try it someday and play a bit with brigtness, contrast, etc. You'll see it is far from "that bad" :) Sep 03 17:29:50 wait what is nikocam now, other name for camera-ui2? Sep 03 18:23:46 freemangordon, I admit i haven't used nikocam with the front camera, but i played a bit with mirror and I could barely see myself Sep 03 18:24:45 VGA is not bad, not even QVGA, it was the brightness that I thought crappy, but maybe that's software too Sep 03 18:33:08 freemangordon: tracker-cfg has enough votes for promotion but no maintainer.... I thought I'd mention this as you pushed it to testing IIRC. http://maemo.org/packages/package_instance/view/fremantle_extras-testing_free_armel/tracker-cfg/0.3.2-4/ :) Sep 03 18:36:06 hey, is it allowed to request for used n900 in here? just got mine stolen... Sep 03 18:41:34 raccoon_ oh no :( Sep 03 18:42:13 I've been looking for some backup or two or three of them too...but should've bought them when they were still around in auctions, not so much anymore Sep 03 18:42:37 same here... i managed to find it on a local web based flea market Sep 03 18:42:38 fortunately my gf also has one, maybe she'll get something else eventually ;p Sep 03 18:42:53 for an awesome price, previous owner thought it was useless and sub-par technology wise Sep 03 18:43:13 yeah, I got the second one essentially free Sep 03 18:44:00 not saying i assume to score one like that again though, i'll happily pay $whatever for one as long as the usb port is ok Sep 03 18:44:18 hmm...seems like one n950 on auction heh Sep 03 18:44:28 210 euros upwards :P Sep 03 18:46:24 yeah, figures... :-P Sep 03 18:47:00 hmm... so there's something called Neo900 ._o Sep 03 18:47:26 * joga squees Sep 03 18:49:02 yeah it's the new open source prototype codename Sep 03 18:49:10 it's been discussed in here quite a bit Sep 03 18:49:17 or one of the codenames, i guess Sep 03 18:50:51 qwazix: nope, that's evidently frontcam that needs _much_ light to be barely usable at all Sep 03 18:54:05 no sw issue Sep 03 18:54:27 terribly noisy even with good light Sep 03 18:54:32 that's what I thought too Sep 03 18:54:40 fmg suggests otherwise Sep 03 18:55:13 btw much the same like on N810 Sep 03 18:57:12 never even tried the N810 cam Sep 03 18:57:29 I don't know if my unit's camera even operates Sep 03 18:57:32 :) Sep 03 19:18:11 DocScrutinizer05: is http://maemo.org/packages/repository/qa/fremantle_extras-testing/ updated automatically? Sep 03 19:19:52 DocScrutinizer05: or rephrased: i just promoted marble-maps and voted for marble, but neither change appears there. should i worry about that? Sep 03 19:38:55 it can take several hours I guess. Sep 03 19:40:09 please recheck tomorrow and if nothing happened (or sth bad happened) then ask freemangordon or merlin1991 about it. I'm not experienced in repository management and package interface and autobuilder Sep 03 19:41:08 it seems the guys yesterday found some flaws/bugs introduced be the mess that came with fremantle-1.3 repo which basically never should have existed at all. Sep 03 19:43:13 freemangordon yesterday said he'll go to fix stuff today, but iirc he later on realized that this might not be exactly easy to fix, or at least needs further evaluation and investigation to assure we don't break 80% of users' auto-update when we replace fremantle-1.3 true repository by a symlink to fremantle repository Sep 03 19:45:10 when fremantle-1.3 ever had some sound rationale, then as a symlink to fremantle (or maybe fremantle a symlink to fremantle-1.3) until eventually we move on to fremantle-1.4 Sep 03 19:49:01 the concept probably being: devices update from fremantle-1.3 while packages go from builder to (symlink) fremantle. When fremantle-1.4 gets released, we create a new repository fremantle-1.4, link fremantle symlink to it and copy the "old" stuff from fremantle-1.3 to fremantle. Then upgrade the fremantle-1.4 repo with new packages and particularly a new catalogs-setting that changes users' catalogs list from fremantle-1.3 to fremantle-1. Sep 03 19:49:02 4 during the system update. Sep 03 19:50:35 thos users who DON'T update to fremantle-1.4 can then still install (app) package versions - e.g. during osso-backup resore process - that match their OS, since they would fetch them from fremantle-1.4 repo, not from fremantle-1.4 Sep 03 19:52:31 alas during migration#1/consolidation (or even earlier than that) somehow the symlink mutated into a real directory aka separate repository, so we now have a fremantle-1.33 and a fremantle repo and both are not in sync and nobody has a clue what gets done where Sep 03 19:52:41 s/33/3/ Sep 03 19:52:43 DocScrutinizer05 meant: alas during migration#1/consolidation (or even earlier than that) somehow the symlink mutated into a real directory aka separate repository, so we now have a fremantle-1.3 and a fremantle repo and both are not in sync and nobody has a clue what gets done ... Sep 03 19:54:14 current concerns raised are: will fremantle repo work with fremantle-1.3 Packages.gz? Sep 03 20:08:28 DocScrutinizer05: no Sep 03 20:09:41 DocScrutinizer05: fremantle repo contains many more packages than fremantle-1.3 (so these packages are not listed into fremantle-1.3's Packages.gz) Sep 03 20:11:23 xes: it's not primarily about the differences in content between the two repos. It's about if a Package.gz built for fremantleA/ repo can work when user uses the symlink fremantleB/ to access the package Sep 03 20:12:03 the differneces need to get merged I'd guess, so the final merged repo contains all the newest freshest packages of both Sep 03 20:12:58 DocScrutinizer05: common packages are inside the same structure Sep 03 20:13:46 xes: actually that "many more packages in fremantle than in fremantle-1.3" supports my above made assumtion that autobuilder builds into fremantle but devices update from fremantle-1.3 Sep 03 20:14:56 if i'm not wrong, latest package in fremantle-1.3 reports well know date of 20121228 Sep 03 20:15:29 sounds about conclusive Sep 03 20:15:35 hold on a second, is the fact basically everything i have comes from extras-devel and little to none from extras because i have extras enabled as fremantle-1.3? ;( Sep 03 20:16:47 for extras-devel there should no such thing like fremantle-1.3 or fremantle-1.2 at all Sep 03 20:16:58 should be no* Sep 03 20:17:27 yeah i have fremantle on extras-devel, but i had fremantle-1.3 on extras Sep 03 20:17:27 one more thing: fremantle-1.2 == fremantle-1.3 Sep 03 20:17:28 since extras-devel is considered rolling release Sep 03 20:17:51 that's fine, xes Sep 03 20:19:15 we need to keep fremantle-1.2 as is. We need to rename fremantle to fremantle-1.3. We need to create a symlink from fremantle to fremantle-1.3 Sep 03 20:19:39 (tha latter mainly for autobuilder) Sep 03 20:20:43 DocScrutinizer05: sorry you lost me.. why do we need to preserve fremantle-1.2? Sep 03 20:25:25 well, actually we probably don't, since it got corrupted a long time ago when it is identical to fremantle-1.3 as of end of last year Sep 03 20:26:13 so it's useless for users still running pr1.2 on their devices anyway, and we as well can nuke it all together Sep 03 20:27:02 but not before we checked for any packages in there that by bad luck are fresher than the corresponding pkg in fremantle Sep 03 20:30:27 it's actualy no surprise that we see extremely low load on our uplink after initial burst for a week. 98% of devices check maemo-extras_fremantle-1.3 and that aiui hasn't seen *any* updates or changed Packages.gz since last new year Sep 03 20:31:52 * DocScrutinizer05 prepares for complete congestion of network uplink after we fixed that fremantle-1.3 issue Sep 03 20:32:18 DocScrutinizer05: i would expose here the sqlite db about fremantle/xx comparison so everyone could give his opinion (it could also be a nice tool to search for packages) Sep 03 20:32:50 sqlite db? Sep 03 20:33:02 you mean the files on cloud-7? Sep 03 20:33:10 yep Sep 03 20:33:23 go ahead, that's why I uploaded them to this location Sep 03 20:34:28 so, this is the db containing the fremantle/fremantle-1.2/fremantle-1.3 comparison until 20130617: http://maemo.cloud-7.de/maemo_repository_lists/pkg.db.bz2 Sep 03 20:38:25 bunzip pkg.db.bz2, and for example: sqlite3 pkg.db "select * from fremantle where fremantle.name not in (select fremantle13.name from fremantle13);" Will list the packages present in fremantle but not in fremantle-1.3 Sep 03 20:39:48 xes: could you create a symlink from fremantletest -> fremantle and then find a few users willing to test the fremantletest catalog/repo in apt and HAM? Sep 03 20:42:08 if that works, and we don't find any pkgs in fremantle-1.[23] which are newer than same pkg in fremantle, then I think we are good to rename fremantle to fremantle-1.3 and create a symlink fremantle->fremantle-1.3 (and nuke resp move away for now the fremantle1.2 repo) Sep 03 20:43:40 i think that to make a test we don't need fremantletest. It's enough that the betatesters remove every repo containing fremantle-1.2 or fremantle-1.3 Sep 03 20:44:46 and they should try to install one of the latest deb just built from autobuilder Sep 03 20:46:52 ..at the end, my opinion is that renaming and not linking fremantle-1.2 and fremantle-1.3 would force everyone to use only fremantle (doing so we obtain also a bandwidth usage of 1/3 compared to actual situation and a faster apt-get update into the devices) Sep 03 20:47:13 are those official repos? Sep 03 20:47:17 if not, fuck em Sep 03 20:47:45 xes: no! we need to make sure that a renaming of the repo doesn't conflict with any "hardcoded" repo names inside Packages.gz or other signature files etc Sep 03 20:48:52 and we can't "force" users to do anything like that, since we don't reach them to tell them what to do Sep 03 20:50:20 fact is that we have a mixed set of clients: 90+% have fremantle-1.3 for extras repo. Some 10% maybe, have fremantle for extras repo. We can't break stuff for any of them Sep 03 20:51:08 DocScrutinizer05: ok. But, in this moment, everyone without /fremantle/ repos can't pick new packages Sep 03 20:51:17 yes Sep 03 20:51:29 that's why we need to fix this Sep 03 20:52:18 ln -s fremantle fremantle-1.3 Sep 03 20:52:59 sure. Does freemangordon has some thought about all this? Sep 03 20:54:53 thus, again: after testing with fremantletest repo, and after checking for no newer packages in fremantle-1.3, we do: mv fremantle1.2 fremantle-1.2-obsolete; rm fremantle-1.3; mv fremantle fremantle-1.3; ln -s fremantle-1.3 fremantle Sep 03 20:55:21 xes: ( freemangordon ) I don't know Sep 03 20:55:50 his last concern been "will renaming of a repo break apt resp Packages.gz?" Sep 03 20:56:13 thus we do a test with fremantletest repo symlinking to fremantle Sep 03 20:57:28 talking again about the sqlite db, sqlite3 pkg.db "select count(distinct name) from fremantle13;" will count fremantle-1.3 packages, while: sqlite3 pkg.db "select count(distinct name) from fremantle;" will count the fremantle packages. Results: fremantle-1.3= 72843 fremantle-1.2=72843 fremantle=83579 Sep 03 20:59:25 sounds about right. Now we just need to check if any of the packages in fremantle12 or fremantle13 is newer than the pkg with same basename in fremantle Sep 03 21:00:47 we could actually do this by conparing fremantle-1.2-obsolete (the ex fremantle-1.3) to fremantle or fremantle-1.3 after we fixed stuff Sep 03 21:01:06 would probably be easier since we can do it on server Sep 03 21:01:30 with up-to-date data, even exploiting find or rsync Sep 03 21:07:28 I think an updating rsync from (ex)fremantle-1.3 to fremantle shall do a fine job to make sure we nuke no new packages? Sep 03 21:09:45 I have no clue though how this will impact on package interface aka midgard(-db), but since all packages are still available under unchanged pathnames, either fremantle/*/*/packagename or fremantle-13/*/*/packagename, since after the fix fremantle-1.3 is a symlink to fremantle, I don't expext any side effects Sep 03 21:10:23 anyway, step one: test with a fremantletest repo which is a symlink to fremantle Sep 03 21:11:31 xes: could you please do that? at least create the needed symlinks so fremantletest is available on repository.m.o, similar to our real repos Sep 03 21:11:41 DocScrutinizer05: fremantletest is ready Sep 03 21:11:53 wow, that was fast ;-D Sep 03 21:12:40 kerio: how about a new repo on your device? Sep 03 21:12:54 i have no wifi :c Sep 03 21:12:56 what do i have to test? Sep 03 21:14:29 missing packages? Sep 03 21:18:06 see if you can update and install cleanly from extras fremantletest free nonfree Sep 03 21:19:19 simply change extras data from fremantle-1.3 to fremantletest Sep 03 21:20:17 then same step disable -testing and -devel, and see if HAM behaves and if you can install resp upgrade a package Sep 03 21:20:36 from extras (via fremantletest) Sep 03 21:33:54 anyone here got the N950 manual as a .pdf or a working link to one? the two links (FCC.gov one and another one found via a Google search) I've found so far are both defunct :-/ Sep 03 22:06:32 umm Sep 03 22:10:44 hello Sep 03 22:29:15 err, is there a reasonably handy way of making the rtlsdr tarballs work, or would i have to chase and symlink each lib/ directory to /usr/lib? Sep 03 22:33:38 very interesting post about the little Lion cell battery Sep 03 22:33:52 but as long as I remember my N900 has always lost the clock settings even brand new Sep 03 22:34:29 LjL-Alps: which is the problem? Sep 03 22:36:22 xes, i just don't know how to go about this. i've only had an N900 for a few weeks. i extracted everything to /opt (like, /opt/rtlsdr, /opt/gnuradio, etc), now things complain about missing libraries (obviously, i guess), so is the idea that i should "ln -s /opt/rtlsdr/lib /usr/lib/rtlsdr", etc etc for every directory? Sep 03 22:38:07 gqrx probably needs to link gnuradio libs into a standard path. All the bin into /opt/rtlsdr/bin/* don't need any link Sep 03 22:38:26 ah Sep 03 22:39:12 or fix ld.so.conf or something Sep 03 22:39:13 if you want to avoid to specify the full path when running: /opt/rtl-sdr/bin/rtl_fm -f 96600000 -W -s 200000 -r 48000 - | aplay -r 48k -f S16_LE -t raw -c 1 Sep 03 22:39:15 xes: uhm, /opt/rtlsdr/bin/rtl_fm does however say it can't load librtlsdr.so.0 Sep 03 22:41:58 LjL-Alps: you are right, you need also those links Sep 03 22:42:38 xes: why not just fix ld.so.conf? Sep 03 22:43:04 add the search path to it Sep 03 22:43:10 yeah, i'm going to try ld.so.conf first. but i imagined since my ld.so.conf is virtually empty (just two entries), the "proper" way to do it was another Sep 03 22:43:18 Sc0rpius: what do you think does 'brand new' mean, production time wise Sep 03 22:43:26 jon_y: sure Sep 03 22:44:16 Sc0rpius: also, some of those critters break even faster than others. It's just that after 12months it seems 99% are dead Sep 03 22:44:24 good question, I wonder if I can find the production date Sep 03 22:44:37 ok now rtl_fm gives me a "no supported device found", which sounds good (i don't have the dongle connected). now... here goes nothing! Sep 03 22:53:59 * nox- wonders if the poor arm cpu is even fast enough for sdr Sep 03 22:56:00 hmm, i guess this USB hub isn't really doing a proper job of pretending to be an OTG cable... Sep 03 22:56:12 since "lsusb" hangs up when in host mode Sep 03 22:56:40 swith host mode off/on and try high/full speed Sep 03 22:57:00 ..with h-e-n Sep 03 23:03:03 nope, doesn't seem to work. on "full", lsusb comes up empty, on "high", lsusb freezes until i disable hostmode Sep 03 23:03:14 but, really, a USB hub is not made for this Sep 03 23:03:16 i should get a proper OTG cable Sep 03 23:05:10 LjL-Alps: this sometimes happens also with a proper OTG cable. Try connecting the key, with h-e-n, swithc on full host mode, boost, enumarate, then try the lsusb Sep 03 23:17:41 no luck, lsusb keeps only giving the root if on "full" mode, and locks up if on "high" Sep 03 23:20:56 LjL-Alps: are you using h-e-n? Sep 03 23:21:29 yes xes, i did it as you said, full, boost on, enumerate, lsusb Sep 03 23:23:56 probaly your device wants too much current... in my test i saw that the device with the E4000 fails more often to be detected if compared to the one with r820t. I think that the current absorption is really near the limit for the host mode Sep 03 23:25:36 you could also try to connect the device after the host mode switch on and boost Sep 03 23:25:37 i thought that too. i do have an E4000. although i tried using the "booston" script that was linked, and also connecting the USB hub to a charger (but i think the USB system may get confused by that) Sep 03 23:28:39 xes: still nope, although if i do that, "full" makes lsusb freeze too (like "high"). also, i noticed now that "lsusb" only freezes if i do "enumerate"; if i don't do "enumerate", lsusb just returns empty ("Bus 001 Device 001: 1d6b:0002") Sep 03 23:32:12 while lsusb is hanged, stop boot and host mode, then swith both on again and enumerate a few times... I have just done this steps and lsusb has passed the "hang" reporting the deviced listed Sep 03 23:32:29 Bus 001 Device 013: ID 0bda:2838 Realtek Semiconductor Corp. Sep 03 23:32:43 Bus 001 Device 001: ID 1d6b:0002 Sep 03 23:33:34 ...i'm just repeating the tests with the E4000 Sep 03 23:34:19 i mean, the previous steps has been verified with the E4000 Sep 03 23:35:23 i think it's the hub. it's a piece of junk, i think i remember it not working well with flash drives either (stuff got corrupted). Sep 03 23:36:13 * LjL-Alps wonders... could i possibly use my computer instead of the hub? *tries* Sep 03 23:36:29 i can't say that these steps are always easy, but, after 2/3 times the device appears Sep 03 23:37:16 LjL-Alps: ahem... Sep 03 23:37:40 guess not Sep 04 00:28:22 https://mikepohjola.wordpress.com/2013/09/03/dear-mr-nokia/ **** ENDING LOGGING AT Wed Sep 04 02:59:58 2013