**** BEGIN LOGGING AT Sun Jul 28 02:59:58 2013 Jul 28 03:00:35 if anything is not safe, then that are packages of crappy (or even rogue?) apps Jul 28 03:01:02 and system components some joker uploaded to extras-devel Jul 28 03:01:20 how did it manage to get through? Jul 28 03:01:44 no idea Jul 28 03:01:58 so it was a security breech then? Jul 28 03:02:00 I never looked ibnto all that Jul 28 03:02:21 I dunno if there *is* any security on that level Jul 28 03:02:42 :| Jul 28 03:03:11 after all, what means "security" here? Jul 28 03:03:56 DocScrutinizer05: not letting some random joker upload stuff? Jul 28 03:04:00 generaly speaking, HAM will refuse to update system packages from extras repos, since Nokia base repos' trust-level is higher Jul 28 03:04:02 ...(truncated)... I have no battery gague anymore :< <-- are you by any chance using advanced power monitor daemon? Jul 28 03:04:10 yes Jul 28 03:04:33 I've got a nifty script that sorts it out, albeit not automatically. Jul 28 03:04:43 ? Jul 28 03:04:52 it worked fine until some update a while ago Jul 28 03:05:12 aaaah now I recall: advanced power monitor was exactly what the name suggested: monitor advanced power usage Jul 28 03:05:39 The script just kills the python which runs apmdaemon.py iirc. After the python process (for that one) gets killed, it will respawn and since respawn I have never seen the battery gauge meter failing to show. Jul 28 03:06:24 o.O. Though the applet that I have can control which BME to use, Nokia's BME or alternative BME. Jul 28 03:06:35 for extra fun, killall -9 python :) Jul 28 03:07:00 Why not make it into a while loop >:D Jul 28 03:07:25 ps -aux | grep apmdaemon.py | gawk '{print $1}' | xargs -r /bin/kill -9 Jul 28 03:07:41 kills the first thing that has apmdaemon.py in it Jul 28 03:08:04 I've got a better one, hang on lemme get it from my script out. I have it working with Queen BeeCon widget. Jul 28 03:08:11 I know there are better ps args, too lazy to dig it up :) Jul 28 03:09:39 kill `pgrep -f /opt/maemo/usr/lib/advanced-power-monitor/apmdaemon.py` Jul 28 03:10:15 I remember the satanic union of an app between Java and Perl at work Jul 28 03:10:23 the thing won't die after exiting Jul 28 03:10:31 it kept some files locked Jul 28 03:10:36 It gets zombified? o.O Jul 28 03:10:43 no idea Jul 28 03:10:46 killall -9 java fixed it Jul 28 03:11:04 Yeah I was thinking Java would most likely be the culprit :) Jul 28 03:11:05 stupid java coder can't write perl nor java Jul 28 03:11:42 the java script backgrounds itself somehow and launches java in the backround again Jul 28 03:11:43 Zombified process are the more horrid ones, no way to kill it without restarting the machine iinm. D: Jul 28 03:11:57 so you can't kill it from your termninal builtin kill Jul 28 03:12:07 nor do you see it in your terminal jobs list Jul 28 03:12:08 lolwut, sounds almost like a fork bomb. Jul 28 03:12:41 I know not what transpired such madness Jul 28 03:12:47 Yeah I guess the actual base Java program is probably ran by root or something and gets respawned probably by the likes of init. Jul 28 03:13:14 no, it definitely runs as the current user Jul 28 03:13:23 o.O Jul 28 03:13:28 it was just way over engineered Jul 28 03:13:36 like Rube Goldberg Jul 28 03:13:47 Rube's cartoon inventions Jul 28 03:14:19 ~seen robbiethe1st Jul 28 03:14:22 robbiethe1st <~robbiethe@50-37-148-53.mscw.id.frontiernet.net> was last seen on IRC in channel #maemo, 13d 47m 6s ago, saying: '... of course, this is the wrong window, but meh'. Jul 28 03:14:43 I have a sinking feeling most in-house apps my company does is because of the constant turn-over rate Jul 28 03:14:51 nobody knows what they are doing Jul 28 03:15:08 nor do they know how to look on google before starting something new Jul 28 03:15:49 there was another GUI "flow" manager, which more or less reinvents make Jul 28 03:15:52 I guess it boils down to... "If it ain't broken, don't fix it." sort of strategy. So things just gets left behind without improvements. Jul 28 03:16:01 depdenencies, parallelism and all that Jul 28 03:16:07 except with a GUI Jul 28 03:16:15 lol. Jul 28 03:16:30 and it's in Java Jul 28 03:16:44 I guess no one heard of makefiles when they started it Jul 28 03:16:54 Would sir like more memory usagfe on top of that? :D Jul 28 03:17:04 Java developers still don't have makefiles Jul 28 03:17:12 s/fe/e/ Jul 28 03:17:12 psycho_oreos meant: Would sir like more memory usage on top of that? :D Jul 28 03:17:27 why of course, they have entire suites of number crunching apps in tcl Jul 28 03:17:38 and write C in KnR style Jul 28 03:18:07 guys, KnR style is older than your mom when she started working Jul 28 03:18:45 foo_function(arg1, arg2, arg3) Jul 28 03:18:53 double arg1; Jul 28 03:18:57 double arg2; Jul 28 03:19:02 char * arg3; Jul 28 03:19:11 { ... begin function ...} Jul 28 03:19:17 that sort of madness, today Jul 28 03:19:37 while running gcc-4.5.x Jul 28 03:20:47 also, using %d for floats, breaking the app when it was ported to x86_64 Jul 28 03:21:51 I need to reach for some drinks to forget what I saw Jul 28 03:23:31 Maybe those devs should be thankful that you're not their boss :D *ducks* Jul 28 03:25:24 The ironic thing is that machine programming languages are hardly different from human foreign language. The lesser one uses those newly learnt skills generally equates to rusty knowledge and sloppy aftermath. Jul 28 03:25:58 I mean idea-wise. Jul 28 03:26:06 psycho_oreos: you're damn right, I'd be labeled as an obstructionist Jul 28 03:27:39 C shell script writers shouldn't be allowed anywhere near high level Java or low level C Jul 28 03:27:44 not even Perl Jul 28 03:28:34 Inevitably its funny to see how people are generally, I recall one of my ex boss was using a computer and he was tapping on the mouse really hard for each click like as if he was indirectly trying to smite the mice (and get rid of his frustration/anger). Jul 28 03:28:57 * psycho_oreos has never heard of C shell script :x Jul 28 03:29:16 psycho_oreos: BSD/HP mainframe era stuff Jul 28 03:29:24 Ahh Jul 28 03:29:26 CSH and TCSH Jul 28 03:29:56 Yup, those names I do recall now. The fancy programmer's shell prompts. Jul 28 03:29:58 it is a bitch to write for, all the quirks Jul 28 03:30:19 it sucks for scripting, it sucks for batch programming Jul 28 03:30:36 generally, even dash/ash/sh wipes the floor with it Jul 28 03:31:24 psycho_oreos: so it tries to be C like in presentation Jul 28 03:31:38 it was a lie, not C like, and there is no parser at all Jul 28 03:32:11 Speaking of shells, I want to try and change my default shell into zsh (offtopic, not on maemo of course). Jul 28 03:32:29 if ( $?somevar) echo $somevar <- this fails when somevar was not defined Jul 28 03:32:43 even if you are already checking it Jul 28 03:32:51 it must be on 2 different lines Jul 28 03:33:16 lol, that's strict C program. Every variable literally has to be defined, whether it be const, char, double, float, etc. Jul 28 03:33:27 s/am/aming/ Jul 28 03:33:28 psycho_oreos meant: lol, that's strict C programing. Every variable literally has to be defined, whether it be const, char, double, float, etc. Jul 28 03:33:38 *facepalms self* programming* Jul 28 03:33:47 psycho_oreos: it's worst, $?var is supposed to check if it was set beforehand Jul 28 03:34:05 oh right Jul 28 03:34:06 jon_y, haha wut. That's messed up. Jul 28 03:34:12 all the vars are strings Jul 28 03:34:17 no types Jul 28 03:34:30 I don't think it supports functions Jul 28 03:35:08 "some string \"escape\"" <- this fails Jul 28 03:35:14 unmatched " Jul 28 03:35:39 some funny things I can't even understand Jul 28 03:35:52 Ah, but still C programming requires it to be defined regardless, before that variable could be used. If lets say you were to do a cruel comparison between BASH and maybe say CSH, BASH allows variables to sort of be defined "on-the-fly" so to speak. Jul 28 03:36:17 http://www.grymoire.com/Unix/CshTop10.txt Jul 28 03:36:43 if(1) != if (1 ) != if ( 1) Jul 28 03:37:41 oh yes, syntax errors like exiting with "unknown action" Jul 28 03:38:04 good luck debugging a thousand line script from a 2 word error message Jul 28 03:42:39 if you're changing 1000 lines at once, you're doing something wrong -.- Jul 28 03:43:16 it was due to if ( !$?var ) Jul 28 03:43:22 can't put ! next to $ Jul 28 03:43:36 it must be if ( ! $?var ) Jul 28 03:43:45 useless debugging message Jul 28 03:59:13 hmm, nat64d fails if it's gprs1 :/ Jul 28 04:00:42 I never could get nat64d to work Jul 28 04:01:18 I understand it is mapping a range of ipv4 address to a /96 space Jul 28 04:01:30 but couldn't get the config right Jul 28 04:01:47 it has a config? :P Jul 28 04:01:50 I think I had to recompile it Jul 28 04:01:53 a few times Jul 28 04:02:14 yes it has, how else do you set where to map? :) Jul 28 04:02:36 IIRC it was hardcoded in the source Jul 28 04:02:39 hence my recompiles Jul 28 04:02:51 huh, must be a different nat64d Jul 28 04:02:57 or older Jul 28 06:41:17 * brolin_empey wonders if he can successfully use a USB to serial converter or a USB VGA controller with the USB Host capability of Maemo 5. Jul 28 06:43:40 With Kermit, I could have a frog in addition to a Python in my pants pocket. ;-) Jul 28 06:44:27 Would Python try to eat Kermit? :P Jul 28 06:46:18 yes brolin_empey .. i did use usb <> serial to configure a cisco router with miniom on maemo :) Jul 28 06:47:12 *minicom Jul 28 06:49:51 VGA output on a handheld computer would be useful for testing displays before buying them: then I would have one fewer reason of already few reasons to need a notebook computer which is significantly less functional than an original IBM Personal Computer machine type 5150 from 1981. Jul 28 06:52:47 * brolin_empey still needs to create his comparison of the IBM 5150 desktop computer from 1981 to most models of notebook computers from over 3 decades later. Jul 28 06:57:48 A comparison table of the Apple II desktop computer from 1977 to all models of mobile computers running Apple iOS from at least 3 decades later would be (highly) amusing too. Jul 28 07:00:05 Apple Inc. is comedy gold to me. Jul 28 07:00:54 ps/2 model 55 Jul 28 07:31:50 http://toby-lou.tumblr.com/post/56530005774/skype-drawing-with-my-brother-c-tlr-2013 ♡ Jul 28 07:56:34 psycho_oreos: why not use pali's battery applet? Jul 28 09:31:30 Sicelo: I am now successfully using minicom on Maemo 5 as a serial terminal emulator with ftdi_sio but I forgot that I have to use Ctrl+M instead of just pressing Enter in minicom. Jul 28 09:33:50 Sicelo: Next project: USB VGA controller on Maemo 5. Jul 28 09:37:03 regarding minicom Jul 28 09:37:05 use GNU screen Jul 28 09:37:26 screen /dev/ttyUSB0 115200 Jul 28 09:37:27 (etc) Jul 28 09:39:50 brolin_empey: \o/ Jul 28 10:11:51 kerio: Darn, bmerep has some detection for charging amount versus realtme use and it gives you a yellow banner every 3 minutes. Jul 28 10:12:15 "The PC you're charging from doesn't allow for enough power" Jul 28 10:12:39 That's nice but I'm on a wall charger that just isn't powerful enough I guess Jul 28 10:12:48 I wish I could disable that message. Jul 28 10:16:00 Sicelo: Actually, using a composite video to VGA converter seems (much?) easier. Jul 28 10:22:52 i would guess so brolin_empey. i have no experience at all with usb-vga controllers. never had one Jul 28 10:29:45 Sicelo: Judgeing by my quick search on eBay, VGA to composite video converters are small and very low cost but composite to VGA converters are larger and cost much more, which is unsurprising considering converting RGB to composite is much simpler than the reverse. Jul 28 10:36:26 Sicelo: http://talk.maemo.org/showthread.php?t=12051 This post predates the Nokia N900. Jul 28 10:39:00 brolin_empey: n900 composite -> vga is quite possible, I bought a device that can do it Jul 28 10:39:13 (I got it for my SNES, but yeah) Jul 28 10:39:15 n900 works too Jul 28 10:46:08 Wizzup: Which model of converter do you have? Jul 28 10:47:53 I don't have it here ... :( I can tell you in a few hours Jul 28 10:48:24 Incidentally, my sleep schedule is b0rk3d Yet Again™. :/ Jul 28 10:49:25 Wizzup: OK, I will still be here when you have it. ;-) Jul 28 10:50:06 Wizzup: Do you know the approximate physical dimensions of the converter? Jul 28 10:50:38 yes. larger than the n900 for sure Jul 28 10:50:41 also needs external power Jul 28 10:57:06 If I want a handheld computer with a proper GNU/Linux distribution using the X Window System, external USB Host capability, and an integrated hardware keyboard, do I even have any alternative to the Nokia N900? Jul 28 10:57:17 nope Jul 28 10:57:28 well Jul 28 10:57:39 you can usually hold laptops with your hands Jul 28 10:58:14 does that count? Jul 28 10:59:16 brolin_empey: pandora? also, I wouldn't count on n900's X and hildon to work well with an external usb video card Jul 28 10:59:36 oooh right Jul 28 10:59:43 the openpandora also works Jul 28 11:00:00 expensive though Jul 28 11:00:48 it's pretty much a n900 Jul 28 11:03:19 kerio: No, by “handheld computer” I meant a mobile computer I can easily fit in my pants pocket and which has integrated converged telecommunication capability. I simply refuse to call such a computer a telephone even if I occassionally use it as a mobile telephone. Jul 28 11:05:28 meh Jul 28 11:05:36 buy pants that are large enough Jul 28 11:09:39 A Nokia 1011 from 1992 is literally a mobile telephone. A Nokia N900 from 2009 is a mobile computer which can be used as a telephone but is still very useful even without cellular connectivity. Jul 28 11:13:57 kerio: I thought the (Open)Pandora lacks integrated converged telecommunication capability? Jul 28 11:14:41 pandora lacks design Jul 28 11:14:48 fugly thing Jul 28 11:15:16 pandora also had no gsm radio Jul 28 11:15:19 has* Jul 28 11:23:00 oh I'm sorry I didn't know how pretty a device is mattered to brolin_empey Jul 28 11:24:14 i dont think it matters to him, he didnt say it Jul 28 11:24:26 friend of mine has one of those pandora failures Jul 28 11:25:32 oh, they did upgrade the hw Jul 28 11:25:36 friend still has the 600mhz orso Jul 28 11:25:56 several friends of mine have openpandoras and they are all very happy with it Jul 28 11:26:46 so are my friends, which is a good thing Jul 28 11:37:42 Beauty is in the eye of the beholder. Jul 28 11:38:57 when you have an openpandora, it sure isnt in your hands Jul 28 11:39:04 so the eye of the beholder looks like a good place to search Jul 28 11:45:49 lol Jul 28 11:45:55 the thing is Jul 28 11:46:16 it's marketed as some sort of handheld console Jul 28 11:46:18 for emulators Jul 28 11:46:29 yeah it shows itself running gta3 Jul 28 11:46:30 but good emulators require lots of cpu to be accurate Jul 28 11:46:31 (lol) Jul 28 11:46:49 well, the snes emus etc probably work well enough on the 'new' 1ghz pandora Jul 28 11:48:08 it has a horrible keyboard tho Jul 28 11:48:12 buttons feel very bad Jul 28 11:48:20 bsnes/higan requires 3ghz to run smoothly Jul 28 11:48:30 most people dont care for 100% accurate emulation Jul 28 11:48:38 because it doesnt really matter Jul 28 11:48:54 except when playing speedy gonzales Jul 28 11:48:54 good enough is fine for most folks Jul 28 11:49:13 or star fox (!) Jul 28 11:49:36 lylat wars! Jul 28 11:50:21 no, that's starfox 64 Jul 28 11:50:29 oh yeah Jul 28 11:50:40 and that reminds me, i should probably finish it on my n64 Jul 28 11:50:42 star fox was the superfx game that everybody was so hyped about Jul 28 11:50:45 but that actually was really shitty Jul 28 11:50:52 i 100%ed the 3ds version Jul 28 11:50:52 with horrible graphics Jul 28 11:50:56 yeah Jul 28 11:51:13 it used an early version of the superfx chip Jul 28 11:51:16 it had some problems Jul 28 11:51:19 i think i still have the cart Jul 28 11:51:21 somewhere Jul 28 11:51:31 starfox 2 was supposed to be better Jul 28 11:51:34 oh yeah now it all comes back Jul 28 11:51:37 talking frog Jul 28 11:51:38 and the beta is actually half finished Jul 28 11:51:54 but it was scrapped because of the n64 Jul 28 11:51:56 i think the only reason it was awesome was because hi 3d graphics on snes Jul 28 11:52:04 yeah Jul 28 11:52:19 i was firlmy PC-nerd anyway Jul 28 11:52:23 console games always were lol simple Jul 28 11:52:35 starfox 64 is unbelievably good Jul 28 11:53:00 i spent 160 dutch guilders on Conker's Bad Fur Day Jul 28 11:53:07 then i quit buying Nintendo games Jul 28 11:53:13 why? Jul 28 11:53:14 games for nintendo consoles* Jul 28 11:53:21 bad fur day was amazing Jul 28 11:53:32 because it was an overly expensive game that was overly expensive because at the last moment nintendo decided to not publish it Jul 28 11:53:52 the game itself was pretty fun, especially in multiplayer Jul 28 11:54:03 kerio: seen the playthrough of Conker's by some of the original devs? Jul 28 11:54:05 on youtube Jul 28 11:54:07 no Jul 28 11:54:11 wauw, you should Jul 28 11:54:28 http://www.youtube.com/playlist?list=PLFP9r6vJL2qCCERMBP_ZWUEV0vaSL4rHf Jul 28 11:54:29 there you go Jul 28 11:54:33 Real Life™ is a real game. Jul 28 11:54:34 NSFW language, but its so awesome Jul 28 11:55:30 also, they are very bad at their own game Jul 28 11:58:32 http://choualbox.com/Img/20130220160021W.jpg :P Jul 28 11:58:37 kerio: also a small story in there about the N64 expansion pack Jul 28 11:58:37 :) Jul 28 11:58:43 in relation to DK64 bug Jul 28 12:00:23 anyway, MAME also takes ton of cpu Jul 28 12:00:40 *tons Jul 28 12:00:52 I dont think I've ever seen a European care about MAME Jul 28 12:00:57 which makes it kinda useless on the n900, sadly :( Jul 28 12:09:37 pandora? HAH! Raster and me approached them about a custom build with GSM radio, in whatever form of cooperation. They were not interested, and we came to think they were not interested to ever produce real hw Jul 28 12:10:27 seeing how openmoko is dead and pandora is not (yet), maybe not a bad decision Jul 28 12:13:05 (([2013-07-28 13:09:40] [...] A Nokia N900 from 2009 is a mobile computer which can be used as a telephone ...)) Which is pretty much what Nokia told us in their advertisments around 2008/2009, no? Jul 28 12:13:23 and.... good (UGT) morning! Jul 28 12:14:02 good (UGT) morning Doc Jul 28 12:14:14 good afternoon Jul 28 12:14:21 Lava_Croft: you're mixing up things. Raster alias carsten haitzler never said he's openmoko Jul 28 12:14:30 thats not what i said either Jul 28 12:14:46 so what sense makes your statement? Jul 28 12:14:47 i just imagine that adding gsm radios adds a host of work and design they might not want to undertake, given the experience of openmoko Jul 28 12:14:58 and it might not really make the pandora a better device Jul 28 12:15:08 aaah, well. No it doesn't, actually Jul 28 12:15:31 i imagine that the pandora guys have rather limited funds etc Jul 28 12:15:39 if they had good funds, they would have hired a proper designer Jul 28 12:15:44 * Lava_Croft chuckles a bit lame Jul 28 12:15:47 adding a module is utterly lightweight, from an engineering perspective Jul 28 12:16:17 it will add work on hardware, software, maybe certification of some sort Jul 28 12:16:57 and we just asked if they'd be interested when we use their desin to improve on it and add some phone functionality. No expense of any kind on their side, first instance Jul 28 12:16:58 and maybe they looked at Nokia and saw the trouble that even Nokia had in producing a device with proper hardware quality Jul 28 12:17:10 (maemo devices) Jul 28 12:17:22 just an opportunity to earn royalties Jul 28 12:20:43 oh, by the way Jul 28 12:20:52 Darkplaces no run on N900 Jul 28 12:20:57 at least not satisfactory Jul 28 12:26:01 hi Jul 28 12:27:14 my gps doesn't work, what's correct location server? Jul 28 12:27:37 supl.nokia.com or supl.google.com Jul 28 12:27:49 I think that nokias one doesn't work anymore Jul 28 12:31:55 ok, thx, I try googles server Jul 28 12:34:33 how I can see gps info? like speed? Jul 28 12:35:00 i see just map Jul 28 12:36:39 http://repository.maemo.org/extras-devel/pool/fremantle/free/g/gpsjinni/ Jul 28 12:36:46 try location test app Jul 28 12:37:20 it's basically a must to deal with N900 GPS Jul 28 12:37:53 app manager is loading.. takes ages Jul 28 12:38:17 cant there be done improvement so loading doesnt take ages? :p Jul 28 12:38:22 like cssu Jul 28 12:39:14 HAM sources are free, go ahead! Jul 28 12:40:12 hehe Jul 28 12:41:34 oh FSCK, not this Konversation hickup again!! Jul 28 12:42:07 sorry OT Jul 28 12:42:37 installing gpsjinni Jul 28 12:43:32 none of the supl servers are working on N900 now fwiw Jul 28 12:46:55 gpsjinni is good, thx Lava_Croft Jul 28 12:54:46 just in case i've missed some key page on Google - it's not possible to wifi tether in AP (infrastructure) mode on the N900, is it? Jul 28 12:57:08 afaik there's no way to do hostAP on the n900's wifi Jul 28 14:35:00 ~weather EDDN Jul 28 14:35:02 Nuernberg, Germany; (EDDN) 49-30N 011-03E 318M; last updated: 2013.07.28 1420 UTC; Dew Point: 57 F (14 C); Pressure (altimeter): 29.85 in. Hg (1011 hPa); Relative Humidity: 30%; Temperature: 93 F (34 C); Visibility: greater than 7 mile(s); Wind: from the SE (130 degrees) at 2 MPH (2 KT) Jul 28 14:35:32 +3 downtown penalty Jul 28 15:06:46 DocScrutinizer05: 35.5 here, a pretty horrid couple of days Jul 28 15:08:31 n900 have thermal security sensors? Jul 28 15:12:02 drathir: it certainly has at least one thermometer Jul 28 15:12:36 but shutdown device when overheat? Jul 28 15:14:27 Temperature : 35 C Jul 28 15:14:37 in the air Jul 28 15:17:12 after night and chaging device was little hot maybe was to little covered but now temp back to normal state, but i think about if temp will to high, the phone auto shutdown? Jul 28 15:19:05 btw yea thunderstorm should come 300km distance temp is Temperature : 26 C Jul 28 15:31:07 who used rescueOS? where do I get initrd.img ? Jul 28 15:31:28 http://206.253.166.96/N900/rescueOS/documentation.txt for flasher-3.5 -k 2.6.37 -n initrd.img -l -b"rootdelay root=/dev/ram0" Jul 28 15:32:43 initrd.img is rescueOS-1.1.img Jul 28 15:33:36 ah, I had just the same idea :D Jul 28 15:51:27 "Error claiming USB interface: Device or resource busy" :-( Jul 28 15:52:30 I have aur/flasher 3.5_2.5.2.2-2 Jul 28 15:54:10 you must be root Jul 28 15:54:17 yes Jul 28 15:56:39 and i must hold "u" on N900 while switch on? Jul 28 15:57:40 hold u while inserting the USB cable Jul 28 16:00:03 then this happens: "USB device found found at bus 002, device address 045. Error claiming USB interface: Device or resource busy" Jul 28 16:09:53 drathir: yes, afaik dsme has thermal management and will shutdown on overtemperature Jul 28 16:11:34 NeutrinoPower: blacklist phonet! Jul 28 16:12:51 general advice at Jul 28 16:12:54 ~flashing Jul 28 16:12:55 it has been said that maemo-flashing is http://wiki.maemo.org/Updating_the_tablet_firmware Jul 28 16:12:58 applies Jul 28 16:13:44 while you're not programming your flash stroage on N900, the procedure is identical on PC side Jul 28 16:14:05 so are the caveats Jul 28 16:22:37 YES! it works Jul 28 16:44:18 heh, Cyanogen is having a Google+ discussion about dropping the ability to have root on Android, and so many people are like "sure, why would i ever need it, i'm not an adblocker or a criminal" (<- parody)... wait i said "heh"? i meant "meh". i guess this is why i just bought an N900 even though i had no reason to ;( Jul 28 16:46:35 Why are they entertaining the idea? Jul 28 16:47:09 To protet the dumb user? Jul 28 16:47:19 *protect, from themselves Jul 28 16:49:44 Raimu: well, it's not very clear from the G+ thread, and i haven't read other sources yet - but my understanding is that Android itself (version 4.3) is in the process of integrating SELinux and losing the concept that as long as you're root, you can do anything, so Cyanogen is wondering whether CM should stick to that concept, or break it. specifically, he wants to know "what people use root for", so he can provide alternate ways to do it without root ( Jul 28 16:49:46 additional CM-specific capabilities?) instead Jul 28 16:54:51 Oh, OK. I'm not very knowledgeable on where Android is heading. Jul 28 16:56:02 Is the SElinux change something like an attempt at clamping down on user abilities to do more than Google sanctions? Jul 28 17:11:58 Raimu: well i wouldn't know, i suspect there might be a bit of genuine security concern (especially in the wake of a couple big exploits on Android), and a bit of "clamping down". i don't know what exactly they're planning and doing, i should read up. but, it's certainly true that the current Android permission model is very weak... it basically gives you a list of things an app can do (and it's very easy for apps to list "everything" and assume the user won't Jul 28 17:12:01 care), plus some degree of isolation among apps Jul 28 17:12:08 that's not a stellar security system Jul 28 17:12:35 Maemo doesn't even have that, afaik, but then Maemo's app are mostly open source things Jul 28 17:13:04 security or liberty, which do you want. Jul 28 17:13:25 LjL: Heh, yeah, Chrome has the same list. Jul 28 17:14:11 And some apps just make it some point of honour to keep using more and more of them even though they're off the scope of what the apps should do. Jul 28 17:14:27 ecc2g: i don't buy that they're incompatible Jul 28 17:14:35 Raimu: yeah. Jul 28 17:14:56 or inconvenience. that's the third option. Jul 28 17:15:22 maybe i'm the only one, but i didn't mind how Symbian actually asked me every time an app wanted to use data, unless i gave a "sticky" go-ahead Jul 28 17:16:19 I don't actually trust the symbian model. Jul 28 17:17:11 ecc2g: i don't like the Symbian model in general, those "signed capabilities" are something i can *definitely* live without, i was just talking specifically about the concept of *asking* you about doing things Jul 28 17:17:45 Android is like, "here's what i can and possibly will do, if you're not ok with that, don't install". that seems way too black and white. fortunately with CyanogenMod, i can also *deny* some capabilities, but then apps start crashing Jul 28 17:19:36 It's kind of depressing that 5 of the top 6 calculators on Play have all the permissions they need to wake up at 3AM and start scanning your network and reporting back what they find to china. Jul 28 17:19:52 heh :( Jul 28 17:20:03 ouch. Jul 28 17:20:12 I mean - sure - they probably just 'legitimately' use that permission to show you ads. Jul 28 17:20:22 But... Jul 28 17:20:45 Unfortunately it sounds like you have to get rid of Linux to have these fine grain controls over apps Jul 28 17:20:47 SpeedEvil: and yeah, another factoid with Android is how applications can mostly freely schedule themselves to be started up at any time, so when you add that to wildcard-capabilities... Jul 28 17:20:58 ecc2g: nah Jul 28 17:21:02 either you VM these things or no Linux Jul 28 17:21:08 ecc2g: well, Android VMs them Jul 28 17:21:11 because binaries can always do ioctls directly to the kernel Jul 28 17:21:25 ecc2g: but Android just stops short of actually *letting you control* them Jul 28 17:21:30 ecc2g: hm, what about SELinux then Jul 28 17:21:39 (which, apparently, Android is going to start using) Jul 28 17:23:51 ecc2g: You don't quite have to get rid of linux. You have to have a deeply well thought out permissions system with more than normal linuxy permissions. Jul 28 17:24:06 ecc2g: And someone actually vetting these permissions at least to some degree. Jul 28 17:24:24 'No, you're a calculator, you don't say anything about why you might need to use the mic - denied'. Jul 28 17:24:30 either way if an app starts ioctling to the kernel somehow a hook needs to go to the GUI to allow it or not... Jul 28 17:24:40 SpeedEvil: yeah, i think that can easily enough be the user though, perhaps contrary to current trends Jul 28 17:24:41 this interface is very klunky - and is a convenience issue Jul 28 17:24:43 Or it's just flat denied. Jul 28 17:25:03 Aegis had in principle the right technology to a large degree - but the wrong policy. Jul 28 17:25:13 ecc2g: ah yeah, i remember reading how adding a "feedback" thing in Linux was difficult (i read that about firewalls, but i guess it goes for most things) Jul 28 17:25:31 If it was possible to add other community 'stores' with the same permissions as nokia, few people would have had much issue with aegis. Jul 28 17:25:51 ^legitimately Jul 28 17:25:55 the kernel will happily report these things but basically have to go to a completely new api. Jul 28 17:26:12 SpeedEvil: i don't know/remember what Aegis is, where they the ones who handled Symbian capabilities? Jul 28 17:26:50 Aegis was a wrapper around many normal system attributes. It did fancy parsing of sourcecode at compile time to generate a permission list for the binary. Jul 28 17:27:04 http://i.imgur.com/HuGr8On.jpg #offtopic #someonewaitedforthis Jul 28 17:27:14 Those permissions were the only ones the binary had - if it tried to do other things it got an error. Jul 28 17:28:05 then there's the case where you allow it, and then change your mind later... but the damage is already done... Jul 28 17:28:13 It's a sane and sensible approach - but only if the user has some means to bypass it. Jul 28 17:29:03 SpeedEvil: uhm, i don't understand - i fail to see the use of that. it lets you generate a "manifest" (in Android terms) automatically based on source, and then you can be sure the binary won't do things the source didn't mention, but how's that different from just having someone trusted compile it? Jul 28 17:29:05 And yes, ideally shut off individual capabilities. Jul 28 17:29:52 LjL: No - even if the binary does unexpected calls - they don't work. Jul 28 17:30:08 SpeedEvil: yeah Jul 28 17:30:52 this problem isn't new...it still boggles me why is this such a big problem now, it has been this way ever since binaries were distributed... Jul 28 17:30:54 So in principle - in the event of stack smashing or other bugs in the binary - it can't be exploited Jul 28 17:30:58 SpeedEvil: but that's the same in Android, if your binary tries to do something it didn't declare a permission for in the Manifest, then the call fails. only difference compared to what you described is, the developer has to list the permissions manually in the Manifest, they're not "guessed" from the source code, but that doesn't seem a big deal to me unless i'm missing something Jul 28 17:31:08 SpeedEvil: ah, so that was the goal, i see Jul 28 17:31:22 It's a problem now as people install lots more apps typically, with lots less review, and no payment. Jul 28 17:31:23 also if you have fine control on ever call, it is a performance hit Jul 28 17:31:29 s/ever/every/ Jul 28 17:31:31 ecc2g meant: also if you have fine control on every call, it is a performance hit Jul 28 17:31:37 No payment up front. Jul 28 17:31:51 Whereas the app authors expect payment from ads. Jul 28 17:32:18 This tends to lead to privacy violations, amongst other things, and apps that are hard to tell from malicious ones. Jul 28 17:32:46 i've been trying to use as many open-source apps as possible on Android lately, and only get them from F-Droid Jul 28 17:33:30 Then there is security/convenience tradeoffs. Jul 28 17:33:40 For example 'hotspot shield'. Jul 28 17:34:08 Yes - you make yourself secure to the local network even if it's unencrypted - for the cost of giving a third party all of your traffic Jul 28 17:36:49 that doesn't sound like a wholly wonderful idea :P Jul 28 17:56:40 uhm, stuff about SELinux as it's being deployed on Android can be read at http://selinuxproject.org/page/SEAndroid and https://source.android.com/devices/tech/security/se-linux.html but i honestly understand little of the mumbo jumbo. seems to me like they're doing the usual "Android 'itself' is very permissive, then hey, if manufacturers want to use the tools we provide to lock your system down..." stunt, but i am a pessimist Jul 28 19:14:17 Hi, i have very dark video during skype video call Jul 28 19:14:36 Any possible idea, how can i fix it ? Jul 28 19:15:54 SpeedEvil: ssh tunnel to a trusted server? Jul 28 19:35:10 XATRIX: not much you can do.. althoughfrom my reading, it appear that the front cam is actually better on skype than with XMPP, for example. you just have to use your N900 where lighting is good. Jul 28 19:35:35 s/ghfr/gh fr/ Jul 28 19:35:36 Sicelo meant: XATRIX: not much you can do.. although from my reading, it appear that the front cam is actually better on skype than with XMPP, for example. you just have to use your N900 where lighting is good. Jul 28 19:40:41 :) Jul 28 19:41:16 so there's no way to fix the exposition ? Jul 28 19:41:41 skype is closed source first of all :P Jul 28 19:42:01 yea Jul 28 19:42:11 but i meant, the kernel module Jul 28 19:42:18 you can play with the 'raw' camera using mplayer or gstreamer tweaks, but it's still terrible. Jul 28 19:44:36 ok, will try to Jul 28 19:44:39 thank! Jul 28 21:27:42 Sc0rpius: ping Jul 28 23:01:16 FYI: Jul 28 23:01:19 [2013-07-29 00:45:18] we want to remove version "1.1.99.6+0m5+withxcb" Jul 28 23:01:21 [2013-07-29 00:45:34] but what we have is "2:1.1.99.6+0m5+withxcb" Jul 28 23:03:51 http://maemo.org/packages/package_instance/view/fremantle_extras-devel_free_armel/libx11-data/1.1.99.6+0m5+withxcb/ Jul 28 23:09:22 http://maemo.org/packages/view/libx11-data/ Jul 28 23:09:37 kudos to freemangordon Jul 28 23:10:24 with a bit of luck this also explains the "removing old package version" spamming Jul 28 23:12:14 anyway we cleaned out the queue (db table) that had some 800k+ of bogus records Jul 28 23:13:49 however we failed to fix stuff like http://maemo.org/packages/view/modrana/ Jul 28 23:14:24 repo-management enabled again Jul 28 23:20:49 freemangordon: Man! Jul 28 23:21:59 still: http://wstaw.org/m/2013/07/29/plasma-desktopVa3743.png Jul 28 23:22:12 killing *my* PC Jul 28 23:22:22 don't ask what it does to maemo.org Jul 28 23:22:31 it's STILL loading Jul 28 23:22:38 89% Jul 28 23:24:36 LOL, seems same shite happened already 2010-02-09 12:12 Jul 28 23:24:49 (it finished loading meanwhile) Jul 28 23:25:59 then starting again 2012-12-28 11:46 every ~5min till today Jul 28 23:27:39 funny enough it first imported diskusage 0.3.20090322-1 by system, then immediately started trying to remove old package diskusage 0.3.20090322-1 Jul 28 23:28:28 according to /var/log/repository-queue-proc.log we seem to have zillions of such packages still Jul 28 23:29:20 really zillions Jul 28 23:29:53 or wait, is this always the same 14 packages? Jul 28 23:32:01 I leave it up to you to find out: http://privatepaste.com/1a5869910c Jul 28 23:32:21 These freezes are getting worse and worse Jul 28 23:32:33 Nevermind me Jul 28 23:46:17 sixwheeledbeast: I try to delete bogus spinning packages as of pastebin above, you might try to pick up the rest I missed (if you really are such brave) Jul 28 23:47:59 pid 0.2.20090322-1 <-- lolwut Jul 28 23:48:01 sixwheeledbeast: so far I only "cleaned" http://maemo.org/packages/view/diskusage/ Jul 28 23:49:19 psycho_oreos: can you feel my pain? Jul 28 23:49:36 I feel with poor maemo.org servers Jul 28 23:51:02 figure what happens when a spider searches the whole package interface Jul 28 23:51:46 DocScrutinizer05, seems like its really messy that log :/ yeah Jul 28 23:52:03 you can bet on every of the packages in http://privatepaste.com/1a5869910c having a webpage like pid or diskusage Jul 28 23:54:26 http://privatepaste.com/d7647180f8 Jul 28 23:55:00 I dunno if every of those packages has an "epoch" Jul 28 23:55:55 but every of that log lines indicates another "removing old package instance" event logged on that package's webpage Jul 28 23:56:30 so I stopped repo-management again until we fixed that shit for good Jul 28 23:57:52 (while freemangordon went to sleep ;-P ) Jul 29 00:03:27 That last link to pastebin contained what freemangordon was on about I suppose. Jul 29 00:40:57 yep Jul 29 00:41:52 we inserted a few "printf()" into that friggin php script, and those revealed the epoch (2:) issue Jul 29 00:51:19 Ahh nice Jul 29 00:56:07 now I wonder if that's the reason for all those packages that are listed as "File not found, not removing anything." Jul 29 01:41:02 DocScrutinizer05: thanks that little calm me, that overheat modules egsist and react when phone need that... Jul 29 01:43:35 anyway a lot of android users are affected broken signing package system, and that threat is more important than root acces and broke smthing by user i think... Jul 29 01:47:38 LjL: symbian also signing jad/jar files if not signed you cant even mark auto accept data acces... Jul 29 01:48:58 oter things is that most of ppl even dont read permissions what app have... Jul 29 01:52:25 True... when we trust an app, we dont have to read all permissions Jul 29 01:55:10 aanyway block root console is stupid idea because installer if im not wrong have root permissions anyway, and if bad app comming and accepted also destroy the whole system... Jul 29 02:14:47 Docscrutinzer51: I was looking at your pastebin with the errors about removing packages. I think part of the problem may be the fremantle/fremantle-1.3 issues. check the pid package for instance, the version that's trying to be cleaned is only in fremantle-1.3 Jul 29 02:15:16 nice find Jul 29 02:15:56 lemme pastebin the full php script for you Jul 29 02:17:56 doesn't explain everything though, as I see those "old version cleaned" messages in the cc65 package page and this doesn't even exist in fremantle-1.3 at all... Jul 29 02:18:41 I'd like to see the script, sure Jul 29 02:21:53 http://privatepaste.com/2f11e8829b Jul 29 02:22:34 watch out, your eyes might start bleeding Jul 29 02:24:36 thx, taking a look... if i see anything jump out at me I'll let you know Jul 29 02:26:59 you can run this manually? Jul 29 02:29:01 I wonder if you ran it against all the failing packages from your pastebin with php package-remove.php fremantle-1.3 extras-devel packagename version it might sort out a lot of those "file not found" messages Jul 29 02:33:27 no, we didn't do anything like that yet Jul 29 02:35:17 so, aiui, you suggest that instead of Jul 29 02:35:31 /usr/bin/php5 /usr/local/bin/packages/package-remove.php fremantle extras-devel pid 0.2.20090322-1 Jul 29 02:35:49 I should run on a rootshell sth like Jul 29 02:35:58 /usr/bin/php5 /usr/local/bin/packages/package-remove.php fremantle-1.3 extras-devel pid 0.2.20090322-1 Jul 29 02:36:02 right? Jul 29 02:37:22 shawnjefferson: ^^^ Jul 29 02:37:52 maybe worth a try with one package... Jul 29 02:38:13 yeah, that's one package, no? Jul 29 02:38:26 that file only exists in fremantle-1.3 Jul 29 02:38:32 yes Jul 29 02:38:46 ok Jul 29 02:39:10 root@stage:~# /usr/bin/php5 /usr/local/bin/packages/package-remove.php fremantle-1.3 extras-devel pid 0.2.20090322-1 Jul 29 02:39:11 No repository found. (fremantle-1.3 extras-devel) Jul 29 02:40:02 didn't you say you want to send your sshkey again? Jul 29 02:40:47 hmm, crap. :( Jul 29 02:41:03 aiui fremantle-1.3 is crap Jul 29 02:41:15 yes, I can email it to you again Jul 29 02:41:54 unixname been sjeffers? Jul 29 02:42:07 there is a mismatch in the repos with fremantle and fremantle-1.3 that is definitely causing some problems. Jul 29 02:42:27 yes. the 1.3 is bogus aiui Jul 29 02:42:45 on my ssh key? I think it was shawnjefferson actually, but I could create a new user Jul 29 02:43:23 I always claim incorrectly that it been invented to be a symlink to fremantle, but now I recall somebody told me that applies for extras only, NOT for extras-devel Jul 29 02:43:39 nah, your username Jul 29 02:44:04 for your account Jul 29 02:45:04 yes, my username on your side is sjeffers or should be Jul 29 02:45:22 which VM? Jul 29 02:47:11 I think it's worse than that though... looking at the Packages file of each for pid shows more packages in fremantle1.3 but the filename is pool/fremantle/free/p/pid/pid_0.2.20090322-1 which seems wrong to me Jul 29 02:47:45 I guess the one that's got these repo scripts and autobuilder stuff Jul 29 02:50:11 let's start with www Jul 29 02:51:21 damn, your key has linebreaks, prolly from your mail agent Jul 29 02:51:34 oops, yep my fault Jul 29 02:54:02 it looks like the php takes the 2: "epoch" into account... at first glance anyway Jul 29 02:56:11 that message you got about "no repository found" is from a failed db lookup for fremantle-1.3. guess it doesnt exist in midgard? Jul 29 02:56:11 I wonder if that script has any elements that introduced a incompatibility when updating PGP Jul 29 02:56:31 probably not Jul 29 02:56:44 don't ask me about midgard Jul 29 02:57:32 rather ask me about gardening, or cooking, I have more clue about those than about midgard Jul 29 02:57:33 it seems like at some point the midgard package db got out of sync with the repo, probably around this fremantle/fremantle-1.3 setup Jul 29 02:57:48 (me has zilch idea about gardening) Jul 29 02:58:56 no surprise, honestly. midgard had a corrupted repligard table for several months it seems Jul 29 02:59:15 thus nuttin worked in midgard **** ENDING LOGGING AT Mon Jul 29 02:59:58 2013