IRC log of #maemo-ssu for Wednesday, 2012-06-06

merlin1991the string is going to be in the binary :D00:00
merlin1991though you can leave it out of git ;)00:00
freemangordonforget about .deb, I am sleepy00:00
freemangordon:D00:00
freemangordonyeah00:00
merlin1991you can simply call the build like this "dpkg-buildpackage -rfakeroot -us -uc -Ifbkey.h"00:00
freemangordonyep00:01
merlin1991also adding -I.git speeds up all our compiles / uploads :D00:01
freemangordonand won't add it to git00:01
freemangordonok, I need 2 repos then00:01
merlin1991hm?00:01
freemangordonfeedservice2-dev00:01
freemangordonand feedservice-plugin-fb-common00:02
merlin1991ah yeah true :D00:02
merlin1991why does it have a -dev package?00:02
merlin1991o.O?00:02
freemangordonthat is a -dev package for another closed-source library00:02
merlin1991ah00:02
merlin1991what does that one do? :D00:02
freemangordonsame as libconnui-dev for cbs-widget00:03
freemangordonuses libcurl to connect to FB00:03
freemangordonadn sqllite and some other things00:03
*** lizardo has quit IRC00:05
*** NIN101 has quit IRC00:06
merlin1991any particular names/ descriptions?00:07
freemangordonfeedservice-plugin-fb-common00:07
freemangordonand feedservice2-dev00:07
merlin1991https://gitorious.org/community-ssu/feedservice2-dev00:08
freemangordonhttp://maemo.org/packages/view/feedservice-plugin-fb-common/00:08
merlin1991https://gitorious.org/community-ssu/feedservice-plugin-fb-common00:08
freemangordonfor the description00:08
merlin1991makes me rofl :D00:09
freemangordonmerlin1991: you kan check on your n900 for the descriptions00:10
freemangordon*can00:10
freemangordonas those on maemo.org are older versions00:10
merlin1991I'm about to headdesk due to sleep deprivation, I don't think so :D00:11
freemangordonyeah, same here00:11
freemangordonlets continue tomorrow00:11
*** Estel_ has quit IRC00:43
*** Estel_ has joined #maemo-ssu00:43
MohammadAGwhy not include the key00:47
*** Estel_ has quit IRC00:47
*** Estel^ has joined #maemo-ssu00:47
MohammadAGit's not a matter to worry about00:47
freemangordonMohammadAG: osso-gnome-vfs2 is missing int the last *-pr00:47
Palifreemangordon, I do not have fb (and I do not know how is working...) but what is that key?00:50
Paliand what is problem with fb?00:50
freemangordonhmm my bad, it is in -pr, but not in the changelog00:50
freemangordonPali: FB widget does not work00:50
freemangordonFB pphoto uploader too00:50
Palibecause fb changed api?00:51
freemangordonno, there is some bug in curl00:51
Palifreemangordon, bug in curl? bug which depends on current date/time?00:52
freemangordonPali: for some reason curl tries to use IPv6 adress when connecting to FB00:52
PaliI know that FB added AAAA record last week (or earlier)00:53
freemangordonyep00:53
freemangordonand that triggers the bug00:53
Paliah, stupid bug00:53
Palibut then it is not only problem with fb?00:54
freemangordonPali, so far that is the only reported bug00:54
Paliand solution is to update curl to new version?00:55
freemangordonyep00:55
Paliok00:55
Palibut what is that key for fb?00:55
freemangordonin the meanwhile I REed a part of the FB infra responsible for login00:56
freemangordonwhen an application tries to use FB API it must provide 2 keys00:56
freemangordonapplication ID and a secret key00:56
freemangordongoogle for FB API key00:57
*** Estel^ has quit IRC00:57
freemangordonPali, have to take some sleep00:57
freemangordonbb for now :)00:57
Paliand you cannot use key from that nokia binary?00:58
Paliok, bye00:58
freemangordonI can, the problem is if I publish it on gitorious00:58
freemangordonbye00:58
DocScrutinizer05I don't see how that's a problem, esp when you obfuscate it a little01:03
merlin1991it only is a problem if fb dev tos says something against it01:04
DocScrutinizer05even ship it as binary01:04
DocScrutinizer05you "publish" that key anyway, in each binary package some user downloads01:05
merlin1991You must not give your secret key to another party, unless that party is an agent acting on your behalf as an operator of your application. You are responsible for all activities that occur under your account identifiers.01:05
DocScrutinizer05errr, jaggedijagg01:06
merlin1991in other words, app yes, source no01:07
DocScrutinizer05shipping with the app either *is* or *is not* "giving key to other party"01:07
merlin1991and then there is We can take enforcement action against you and any or all of your applications if we determine in our sole judgment that you or your application violates Facebook Platform Terms and Policies. Enforcement action is both automated and manual, and can include disabling your application, restricting you and your application's access to Platform functionality, terminating our agreements with you, or any other action as we01:07
merlin1991 in our sole discretion deem appropriate.01:07
merlin1991in other words we *might* ruin fb for all n900 users of the stock plugin if we publish the nokia keys01:08
DocScrutinizer05so what? do as I suggested: extract the key from original nokia binary you ship partially to gitorious, by a nifty combination of dd, od, and sed, create a temporary file with some hex values which you can #include01:11
DocScrutinizer05though I'd do that with my own key, sure01:11
DocScrutinizer05not really the Nokia one01:11
DocScrutinizer05you could do that on device01:11
merlin1991and why do all the nifty stuff?01:12
DocScrutinizer05in pre-install01:12
merlin1991why not simply "not" put the key on in the source package and gitorious and have it just in the binary?01:12
merlin1991could be argued to be "configuration" and not part of the "software"01:12
Palibut this violate GPL01:13
Palifirst choose other license01:13
merlin1991freemangordon: wrote it, he chooses the license :)01:13
DocScrutinizer05look, we're just patching original nokia lib ;-D every single byte of it except the key01:14
merlin1991rofll01:14
Pali:D01:14
merlin1991well if anything freemangordon can slap any license on it, there is no (c) nokia on his code01:14
merlin1991and since it is very maemo specific I wouldn't worry about people stealing the source, so I guess anything bsd/mit like would be perfect01:15
merlin1991though as I wrote earlier it's up to freemangordon01:15
PaliI do not see any differences between distributing binary with key and source code with key in hexdump01:16
Paliboth are same01:17
DocScrutinizer05honestly, the lib has same name, runs on same device, for same main() with same API01:17
DocScrutinizer05I only half was joking01:17
DocScrutinizer05we're indeed patching a lib, not stealing a key01:18
merlin1991yeah :D01:18
merlin1991btw I think we would be GPL compliant if we load the key from a file at runtime and not distribute that file :)01:18
merlin1991(in source)01:18
DocScrutinizer05about the non-disclosure, this can get handled by actually using whatever key is in original lib as found during install of our patch01:18
merlin1991tricky01:19
merlin1991we replace, thus apt removes the old lib before we place the new lib01:19
DocScrutinizer05ouch01:19
merlin1991before *any* inst hook of our package gets called01:19
DocScrutinizer05even before pre-install?01:19
DocScrutinizer05dang01:19
Paliwhy to create some stupid hack because of license?01:20
DocScrutinizer05just for the fun01:20
Pali:D01:20
merlin1991Pali: I'd simply not license it gpl and be happy :D01:20
PaliI think key should be included in source code in some hexdump way01:21
PaliIt is really same as key included in elf binary01:21
DocScrutinizer51indeed01:22
merlin1991yeah, but the fb tos explecitely forbid to publish the key outside of an application01:23
Palisource code is application part01:23
DocScrutinizer51but then you can't download the original Nokia lib anywhere01:23
DocScrutinizer51without 'auth'01:23
merlin1991Pali: the wording is: You must not give your secret key to another party, unless that party is an agent acting on your behalf as an operator of your application. You are responsible for all activities that occur under your account identifiers.01:23
*** BCMM has quit IRC01:24
PaliSecurity through obscurity?01:24
merlin1991apparently :/01:25
Palithere always exists way how to obtain key from app01:25
merlin1991though I loved the security through insanity article on tdwft :D01:25
Palifrom any app01:25
DocScrutinizer51merlin1991: thisx doesn't say which encoding your app has01:25
merlin1991yeah but operator != guy reading your code01:25
Paliso you cannot write application in bash/python?01:25
DocScrutinizer51exactly01:26
Palithen operator != guy which hexdump/objdump your code01:26
DocScrutinizer51that's not what the TOS says01:26
DocScrutinizer51incluse the key in hex, 'decode' it rot1301:27
DocScrutinizer51enough to "not share"01:27
Paliit is also encoded in nokia binary?01:28
merlin1991DocScrutinizer51: but my rot26 is so much more secure ;)01:28
DocScrutinizer51who can read c source and understand it does rot13 can as well hexdump the binary01:28
PaliI think char key[] = {  0x00, 0x01, ... }; is enought01:28
DocScrutinizer51probably01:29
Palino encoding but binary format01:29
Palisame as in elf binary01:29
DocScrutinizer51yup01:29
DocScrutinizer51we should ask quim01:30
merlin1991dunno how quim can help us here01:30
DocScrutinizer51he's the one to worry about Nokia's key, not we01:30
DocScrutinizer51or lemme put it that way, when Nokia allows us to use the key, we are an 'agent(?)'01:31
Paliso why then ask? :D We do not worry about Nokia :) same as Nokia does not worry about us :)01:31
merlin1991I only wory about $random people on the stock plugin suddenly not being able to use it anymore01:32
merlin1991(because of our actions)01:32
Palifreemangordon wrote that stock plugin does not working due to AAAA record01:32
Paliproblem with curl01:33
merlin1991yeah, but a different libcurl fixes that again01:33
DocScrutinizer51merlin1991: when Nokia allows us resp declares cssu replacement official, there's no reason anybody feels pissed01:33
merlin1991which even is in extras-devel (even though it shouldn't be there)01:33
DocScrutinizer51the point is we patch a lib but the app and platform stays the same01:35
DocScrutinizer51the key is for the app01:35
merlin1991I'm only jumping those hoops here to prevent the unlikely occasion of fb revoking the official nokia key01:35
DocScrutinizer51they CANT when Nokia says thats all OK01:36
DocScrutinizer51well they can, but why should they?01:36
merlin1991also I'm not worried about the .deb at all, I'm thinking about the tar.gz and gitorious01:37
DocScrutinizer51in the end for fb nuttin changed01:37
merlin1991except that the key now is easily avaiable to anybody interested01:37
DocScrutinizer51I see this different01:37
merlin1991btw Pali technically we can't make the lib gpl anyway01:38
DocScrutinizer51not any easier than with any non-encrypted binary01:38
merlin1991it's links against the closed feedservice2 lib from nokia01:38
*** Atarii has quit IRC01:38
Paliand LGPL?01:38
merlin1991and as per gpl we would have to provide source for that aswell01:38
merlin1991Pali:  does not help us01:39
DocScrutinizer51the closed feed is a point01:39
DocScrutinizer51the hex in source, not01:39
merlin1991"our" product inclueds parts that we can't provide source to01:39
PaliLGPL allow linking with closed libary01:39
merlin1991yeah but only if the lib is lgpl01:39
merlin1991no the other way around01:39
merlin1991or rather it allows you to link gpled library, but not to link against non gpl stuff and make yours lpgl01:40
merlin1991the fun of closed components :/01:40
DocScrutinizer51merlin1991: take some sleep ;)01:40
merlin1991DocScrutinizer51: 3 contradicting statements? not tired enough ;)01:41
Palihttps://www.gnu.org/licenses/gpl-faq.html#FSWithNFLibs01:45
Palihttps://www.gnu.org/licenses/gpl-faq.html#GPLPluginsInNF01:46
DocScrutinizer51I'd extract the key from original lib. If we got no pre-install hook to do that, so heck this will be the first 'app' (except cssu?) that needs two packages to install completely - a chain-install01:46
Palimerlin1991, you can write program under GPL license and write exception to link closed source libary xyz into your program01:46
*** nox- has joined #maemo-ssu01:48
DocScrutinizer51and my TV just asks "is FB still in or alreafy out?"01:50
DocScrutinizer51or wait, can't 'our' lib download the original one?01:54
DocScrutinizer51during install01:54
merlin1991please no01:54
DocScrutinizer51please no? :-P01:55
merlin1991we'd have to put proper headers in order to get the nokia ssu server to behave and then work with the fun packaging of an debian package (tar.gz inside ar) and still don't mess aynthing up01:55
DocScrutinizer51eeeew01:55
merlin1991especially the first part is not fun01:55
merlin1991unpacking the deb actually goes quite good (ofc only if we have ar avaiable on device, I dunno that)01:56
DocScrutinizer51can't we just use apt-get?01:56
merlin1991since we're in install dpkg is locked thus apt-get will error out before running01:56
DocScrutinizer51hahaha rrrright01:56
merlin1991but we ofc could patch apt-get to allow download with locked dpkg database01:57
merlin1991but that is just tooooooooooooo fucking much for a single api key01:57
DocScrutinizer51yep01:57
DocScrutinizer51btw I wonder why cssu and this 'app' don't use alarmd or a simple subshell with sleep to start that 2nd phase of install02:00
* DocScrutinizer51 also ponders there's obviously some hook for doing backup prior to PR upgrade02:05
DocScrutinizer51might be feasible to exploit that to get a copy of original lib02:06
Paliso this is reason why we should include cron into maemo and cssu :D02:07
merlin1991nah alarmd would do fine02:08
merlin1991but would still be a stinkin dirty hack02:08
DocScrutinizer51or we simply install under alternative name, and do a rename in postinstall02:08
merlin1991thou shall never touch the content of other packages with yours02:08
PaliI'm suggesting to include key into source code in hex02:08
DocScrutinizer51haha02:08
merlin1991part of the debian bible ;)02:09
*** Sc0rpius has joined #maemo-ssu02:09
DocScrutinizer51well, we're not allowed to redistribute Nokia (C), neither complete binaries nor excerpts02:11
DocScrutinizer51are we?02:11
merlin1991that key?02:11
merlin1991we got it listening on the traffic ofc02:11
DocScrutinizer51I think that's irrelevant02:15
merlin1991yep, jokes usually are02:15
DocScrutinizer51as irrelevant as encoding of our app, machinecode or c test02:15
DocScrutinizer51honestly, make the key a user config option. ship a tool to extract key from original lib and mv original to backup plus mv patched into place02:31
*** arcean has quit IRC02:46
*** M4rtinK has quit IRC02:49
*** M4rtinK has joined #maemo-ssu03:06
*** M4rtinK has quit IRC03:11
*** Estel_ has joined #maemo-ssu03:26
*** Estel^ has joined #maemo-ssu03:51
*** Estel_ has quit IRC03:51
*** Pali has quit IRC04:21
*** trbs has quit IRC04:30
*** amiconn_ has joined #maemo-ssu05:17
*** amiconn has quit IRC05:17
*** amiconn_ is now known as amiconn05:17
*** Estel^ has quit IRC05:19
*** Estel_ has joined #maemo-ssu05:20
*** int_ua has quit IRC05:53
*** amiconn_ has joined #maemo-ssu06:03
*** amiconn has quit IRC06:03
*** amiconn_ is now known as amiconn06:04
*** M4rtinK has joined #maemo-ssu06:04
*** Estel_ has quit IRC06:10
*** Estel_ has joined #maemo-ssu06:10
*** M4rtinK has quit IRC06:17
*** Estel_ has quit IRC06:18
*** Estel_ has joined #maemo-ssu06:18
*** M4rtinK has joined #maemo-ssu06:18
*** Estel_ has quit IRC06:43
*** Estel_ has joined #maemo-ssu06:44
*** cbeau has left #maemo-ssu06:45
*** Estel_ has quit IRC06:46
*** Estel_ has joined #maemo-ssu06:48
*** nox- has quit IRC07:07
*** Estel_ has quit IRC07:36
*** Estel_ has joined #maemo-ssu07:38
freemangordonDocScrutinizer51, Pali, merlin1991: the situation is in a no way different to cbs-widget: FOSS library linked to closed source08:38
freemangordonand I think putting the keys (and only the keys) in a different header file, which won't be on gitorious and won't get included in tar.gz solves our problem pretty easy08:40
*** M4rtinK has quit IRC08:57
freemangordonmerlin1991: here http://gcc.gnu.org/releases.html is the ling for gcc 4.6.2 source code, just follow "GCC 4.6.2" link and choose the best mirror for you09:08
freemangordon*link09:08
*** amiconn has quit IRC09:37
*** amiconn has joined #maemo-ssu09:40
*** M4rtinK has joined #maemo-ssu10:01
Estel_freemangordon, may I buggy You with interesting request?10:22
freemangordonsure10:23
Estel_do you remember idea about using Rpi as hdmi out for N900... During conversation with SpeedEvil and Hurrian, the latter come up with extra interesting idea10:24
Estel_http://mg.pov.lt/maemo-irclog/%23maemo.2012-06-06.log.html#t2012-06-06T05:06:4010:24
Estel_from this line to end of related discussion10:24
Estel_IMO, it's our best chance to achieve it without - practically - any overhead on N900's side (CPU, RAM), and ability mto transfer everything, using Rpi as processing machine10:25
freemangordonEstel_, cloning is done on a HW level10:25
Estel_and, as my limited knowledge goes implementing it *shouldn't* be very hard, given fact, that TV-Out kernel module is open and we need only small change + implement it to userland as additional option10:25
Estel_he?10:25
freemangordondriver os FOSS (DSS2 driver)10:26
Estel_SpeedEvil suggested, that hardware element read framebuffer10:26
Estel_yea10:26
Estel_and it compress etc10:26
Estel_read on, I proposed other way later10:26
Estel_SpeedEvil was in soubt, if it's doable using USb speeds, but it was confirmed later10:26
freemangordonand it supports VFI (virtual framebuffer), not sure is it enabled in .config10:27
Estel_basically, something that would allow Rpi to read our framebufer - in raw mode - and process it on Pi10:27
Estel_hm, itneresting10:27
Estel_it would be 800x480*3*FPS bits - for 10 fps, it's only ~11Mb/s, would run flawlessly even over wifi10:27
Estel_on good connection wifi, it should work even for 20 and 30 FPS10:27
Estel_for USB, it would be possible to use higher resolutions10:28
Estel_Pi would need to, basically, display captured framebuffer, without or with additional processing10:29
Estel_this way, everything would work, even video, and - except dropping to framebuffer like in Tv-out - everything would be processed using Pi resources10:29
Estel_I knwo You're (both) overprojected, but it would be KILLER feature - better than anythinbg made for any mobile device without hdmi out, including apples and paid software10:30
Estel_of course You know my limited (barely existent) skills re coding, but as far as I can tell, the bets thing is that it *shouldn't* require cosmic ammounts of effort10:30
* Estel_ teases10:30
Estel_~tease10:31
Estel_~striptease10:31
infobotHoogah Hoogah wah wah *takes of the box* *dances around showing of the cpu and memory* Ah yeah you likey my little HD no?10:31
freemangordonEstel_, scratch that, what is supported is RFBI, not VFI10:36
*** dafox has joined #maemo-ssu10:36
freemangordonEstel_, I am not sure I get the idea. You want to use VENC (that is TV out) module to do what exactly?10:45
Estel_to cannibalsie way it's reading screen data *before* processing via hardware parts...10:46
Estel_and to change tv-out applet to generl purpose external view applet (or better name) that have tv-out as one of, lets say 3 functions. or two10:46
Estel_then, on Raspberry Pi side, capture screen data (in form it is presented before processed by N900 hardware tv-out) and display it via HDMI10:47
freemangordonEstel_, there is no dedicated TV out memory, fremebuffer is read line by line and scaled in real-time10:47
*** M4rtinK has quit IRC10:47
Estel_= HDMI output10:47
Estel_OK10:47
Estel_so couldn't it by read line by line over USB networking by Pi?10:47
Estel_is there a way to allow Pi to access N900's framebuffer?10:47
Estel_so scratch tv-out module10:47
Estel_if it's not needed, then, even better10:47
Estel_we don't need scalling10:48
Estel_only exporting data to framebuffer, in native resolutions, for beginning10:48
freemangordonto access as in? it is /dev/fbN (n=0,1,2)10:48
Estel_any sensible way to read it from another linuxbox?10:48
Estel_Ok, do we have any implementations in linux world, where something read framebuffer and display it?10:48
freemangordonya, do a netcat :P10:48
Estel_as per my calculations, bits per second are10:49
Estel_800*480*3*FPS10:49
freemangordonwrong10:49
Estel_?10:49
freemangordonwe are not using rgb,10:49
Estel_ops10:49
freemangordonFB is 16 bpp10:49
Estel_don't tell me that it's 16 bits :p10:49
Estel_eh10:49
freemangordon565 or something10:50
Estel_and we have this pout into framebuffer, without any drawback for performance, when using tv-out?10:50
Estel_I don't talk about scaling10:50
Estel_oh god10:50
freemangordonplease rephrase10:50
Estel_any way to process it in sensible way before sending?10:50
freemangordonprocess as in?10:50
Estel_I jsut wonder how Tv-out is processing so much raw data, even if we drop scaling to Pal/NTSC10:51
freemangordoncomress? scale?10:51
Estel_IDK, change to RGB, whatever10:51
freemangordonEstel_, you annot drop the scaler10:51
freemangordon*cannot10:51
freemangordonit is a part of VENC module10:51
Estel_understood.10:51
freemangordonEstel_, look at DSS subsystem in TRM and you will get the picture10:51
Estel_so after reading from framebuffer, it's either PAL/NTSc or nothing?10:51
Estel_i mean tv-out of course10:52
freemangordonit is 56510:52
freemangordonwhich is passed to VENC module in chunks of (iirc) 2048 bytes10:52
Estel_it doesso 800*480*565*FPs, which make it unsuitable for sending10:52
freemangordonEstel_, that is the FB content, not the result of downscaling10:53
Estel_now, I wonder if we can process it a little - in any way - to make it "smaller" in terms of byte,s without using too much resources on N900 side, and, actually, leave considerable ammount for other functions10:53
Estel_yea10:53
Estel_but we *don't* want to downscale for HDMI out10:53
*** dafox has quit IRC10:53
Estel_for that we have Tv-out10:53
freemangordonMJPEG?10:53
freemangordonor some other fast algo10:53
Estel_sounds ncie. Basically, main limit is *realistic* thoroghput of USB networking10:54
Estel_...and mjpeg would allow us to do 20 fps without killing N900?10:54
freemangordonBTW I still wonder what is wrong with X forwarding10:54
Estel_freemangordon, remember thread in TMO? AFAIK, You were able to forward EasyDebian to external machine...10:55
Estel_but not Maemo as per se10:55
freemangordonnever tried hard10:55
Estel_lack of xephyr or something10:55
freemangordonit was just a POC10:55
Estel_You know, if You would try hard, and it would succeed, it would be same *killer* feature with even less hassle ;)10:55
Estel_for sure less than RE fb plugin :P10:55
* Estel_ teases again10:55
freemangordonwhy you would want maemo through HDMI?10:55
freemangordonit won't fit10:56
freemangordonmost of the UI is optimized (read hardcoded) fro 800x48010:56
freemangordon*for10:56
Estel_because I would like to sue higher resolutions for 3rd party FOSS programs (i.e. not hildon-desktop itself)10:56
Estel_because Tv output is downscaled and small fonts are hard to read10:56
freemangordonexample? ED?10:57
freemangordonED works like charm with X forwarding10:57
Estel_because I love to work on trips with N900, but hate fact that quality of everything I see on big screen is worse than holding small screen close to eyes :P10:57
Estel_no no,m lets leave debian alone10:57
freemangordonbut what application then, gimme an example10:58
Estel_basically, any FOSS program in repos that could run it higher ress but is hardcoded now to 800x480 could be updated to work using other resolutions10:58
Estel_leafpad :P10:58
Estel_no, seriously10:58
Estel_I'm afraid to tell "games", bnecause I will hear "then do it Yourself and don't bug me"10:58
Estel_but, other than games:10:59
freemangordonEstel_, have in mind that even if there is a way to forward FB to Pi, what you would yse as an inout device?10:59
freemangordonn900 itself?10:59
freemangordon*input11:00
* freemangordon is afk11:00
Estel_sorry, got dced11:00
Estel_freemangordon, other than some educational tools, videos and images Document viewers?11:01
Estel_Ed doesn't have DSP acceleration for video, at least not the easy way11:01
Estel_BTW, AFAIK, in your testing Ed was exported to external screen via LXDE, yep?11:01
Estel_some programs are veryt resource hungry, and run better when started from within Maemo, without LXDE and all middle-mans (GIMP)11:02
Estel_(Chromium, even old Libre Office)11:02
Estel_s/old/good old/11:02
infobotEstel_ meant: (Chromium, even good old Libre Office)11:02
Estel_but generally, You're right11:02
Estel_most programs we need in better resolution are in ED11:02
Estel_if we can't run hildon-desktop and so goes on in higher resolutions, it may be quite unnecessary, for sure11:03
Estel_honestly, i'm a little pissed off by PAL/NTSC downscaling11:03
Estel_it's just quite annoying, that on big big BIG screen, You have lower resolution/quality of image than on native 800x480 screen,, but maybe it's only psychological/purist point of view11:04
Estel_but, another *important* thing - other systems, like Ubuntu 12.04. should it work just like with Easy Debian?11:04
Estel_Your's and other's experiences show, that running programs from such antively booted systems is much faster than chroot (why, BTW?)11:04
Estel_I think that they should be exportable via xephyr etc, but You know better, probably11:05
Estel_(i.e. confirm/deny? tried that? interested to try?)11:05
freemangordonEstel_, forget about streaming media from n900 to Pi, OMAP has one GFX and 2 video overlays which are scaled/mixed in HW to produce the image you have on LCD11:11
Estel_ah11:12
Estel_ so, I msut summarize, that You're right - most like as for Maemo, we need only Ed and such things via HDMI-out11:12
Estel_downscaling to PAL (little better) or NTSC (little worse) is irritating, but we can live with it11:12
Estel_and question about other systems booted independently of Maemo?11:12
Estel_I suspect it should be no problem to send them?11:13
freemangordonand that is done with export DISPLAY=a.d.c.d:0.0 startlxde11:13
freemangordonEstel_, NFC, but it should be the same11:13
Estel_I think You agree that it's very interesting thing for, lets say, Lubuntu 12.0411:14
Estel_especially, considering how fast things like Chromium or LibreOffice works there11:14
freemangordonno, that was for ED chroot11:14
Estel_BTw, any clue why it works much slowe rin chroot under Maemo? chroot should be like native speed?11:14
freemangordonslower? no, it is the same here, but my ED is on uSD11:15
freemangordonand not in a loopback device11:15
Estel_same here, no loopback11:15
Estel_but in eMMC. i use swap; on uSD...11:15
freemangordonme too11:16
Estel_hey, I'm sure i remember You saying, on ubuntu 12.04 thread, that things work faster than in ED11:16
freemangordonhmm, can't remember saying such thing11:16
Estel_hm, shouldn't we avoiud I/O conflicts with swap, just like with Maemo on eMMC? ED during operation, also generates quite ammount of I/O, doesn't it conflict with swap I/O?11:16
freemangordonseems like no11:16
Estel_ok, will search and ask later, maybe ti was just me missunderstanding it11:17
Estel_sorry for bugging You then, and thanks for clarification11:17
Estel_You're right, it seems, that we have everything we need possible already11:17
*** dafox has joined #maemo-ssu11:28
*** Estel_ is now known as maemobot11:41
*** maemobot is now known as Estel_11:45
freemangordonMohammadAG: hiding FB secret key is enough, ain't? API key is useless without it, correct?11:45
*** Estel_ is now known as safsdgdfhg11:45
*** safsdgdfhg is now known as Estel_11:45
*** arcean has joined #maemo-ssu11:52
MohammadAGNo11:55
MohammadAGFB secret key is used for more advanced changes11:55
MohammadAGOr calls11:55
MohammadAGAnyway, why hide it when it's clear in the public?11:55
*** Pali has joined #maemo-ssu12:22
freemangordonMohammadAG: so I should remove API key from the sources too?12:24
freemangordonit does not make sense, sekret key is used fro signature generation12:24
freemangordon*secret12:24
MohammadAGDude? Just keep it12:25
MohammadAGJust keep both keys12:25
MohammadAGThe same way everyone keeps them12:25
freemangordonMohammadAG: where it is clear? in binary?12:25
MohammadAGYes12:26
MohammadAGYou can't do anything harmful with them12:26
freemangordonMohammadAG: what one can do with API key only?12:26
MohammadAGNothing12:27
freemangordonok, that was my question :)12:27
MohammadAGSo just keep the API key and the secret key in the source12:27
DocScrutinizer51those keys are just so FB stays in control in case somebody comes up with a rogue app12:27
freemangordonso, now we have API key(only) on gitorious, secret key is not there12:27
MohammadAGExactly, they can just delete/blacklist the key12:27
MohammadAGWHY?!12:27
freemangordonjust in case12:28
freemangordonto keep merlin1991 happy :D12:28
DocScrutinizer51weird rationale12:28
freemangordon?12:28
freemangordonone needs both keys to be able to send a request to FB servers12:29
MohammadAGIn case what12:29
MohammadAGVery weird12:29
MohammadAGNokia made no effort to hide12:29
MohammadAGNeither should they make any effort12:29
MohammadAGA request to do what12:29
MohammadAGJust put both keys on gitorious12:29
MohammadAGThey won't cause a nuclear strike12:29
MohammadAGIf you're not putting it on gitorous I am12:30
freemangordonok, application id (API key) is visible from your FB profile12:31
MohammadAGAgain any argument you make won't make sense12:31
MohammadAGBoth keys can be obtained by stringing the lib12:31
MohammadAGBoth keys aren't tied in any way to a certain device's imei12:32
freemangordonMohammadAG: but it is not that easy to have the lib12:32
MohammadAGBoth keys are in plaintext12:32
DocScrutinizer51that's what Pali ans I say whole morning12:32
freemangordonand the concern is that FB could blacklist n900 if we put the keys on gitorious12:32
MohammadAGFB don't give a fuck12:32
DocScrutinizer51the ONLY issue / concern being it's part of Nokia (C) IP12:33
MohammadAGThat's false, I can bet on it12:33
freemangordonTBH I am ok either ways12:33
DocScrutinizer51unless we use our own key, which kinda defeats purpose of 100% compatibility12:33
freemangordonMohammadAG: just make and agreement with merlin1991 and i will put secret key on gitorious too12:34
freemangordonok?12:34
Palifreemangordon, my idea was to include key in hex form like: char key[] = { 0x11, 0x12, ... };12:34
DocScrutinizer51again. please ask Quim!12:35
freemangordonPali, it is in such form12:35
Paliit is same as hexfump of elf binary12:35
freemangordonhttps://gitorious.org/community-ssu/feedservice-plugin-fb-common/blobs/master/include/facebook/common.h#line2912:35
DocScrutinizer51If Quim gives OK, we *are* OK and basically Nokia's agent12:35
DocScrutinizer51in that regard12:36
freemangordonhmm, or not :)12:36
MohammadAGiOS keys are also plaintext12:43
freemangordonMohammadAG: where?12:44
freemangordonyou have the source code?12:44
MohammadAGNo, strings12:46
MohammadAGNo one hides these12:46
MohammadAGYou can get the N9 keys in the sane way12:47
freemangordonMohammadAG: lets end that discussion, I told you, make an agreement with merlin1991 and I will but the second key on gitorious too, ok?12:54
freemangordon*put12:55
freemangordonMohammadAG: on the other hand - were you able to find X-Fade?13:02
freemangordonmerlin1991 told me there is some problem with -testing repo13:02
*** xmlich02 has joined #maemo-ssu13:03
MohammadAGmerlin1991's pestering him afaik13:03
*** int_ua has joined #maemo-ssu13:14
*** xnt14 has quit IRC13:17
*** MohammadAG has quit IRC13:19
*** MohammadAG has joined #maemo-ssu13:20
*** xnt14 has joined #maemo-ssu13:21
*** lizardo has joined #maemo-ssu13:30
*** MohammadAG has quit IRC13:32
*** MohammadAG has joined #maemo-ssu13:32
Estel_freemangordon, MohammadAG, if some contact with X-Fade is needed, You can always ask me14:16
Estel_i'm in almsot constant contact with him (no joke :D )14:16
Estel_...so I can easily forward Your problems. Generally, anyone from council is good to ping regarding infra14:17
*** arcean has quit IRC15:05
PaliEstel_, promoting package kernel-power-settings to extras: http://maemo.org/packages/view/kernel-power-settings/15:05
Estel_yea, told X-Fade about that one15:06
Estel_he is working on it15:06
Estel_You remind me old problem, or something new failed?15:06
*** DocScrutinizer06 has joined #maemo-ssu15:06
*** DocScrutinizer has quit IRC15:06
*** DocScrutinizer has joined #maemo-ssu15:06
Estel_Pali^^^15:08
Paliold problem15:08
*** DocScrutinizer05 has quit IRC15:10
*** Estel_ has quit IRC15:19
*** maemobot has joined #maemo-ssu15:20
*** lizardo_ has joined #maemo-ssu15:22
*** arcean has joined #maemo-ssu15:24
*** lizardo_ has quit IRC15:55
*** M4rtinK has joined #maemo-ssu17:38
*** maemobot is now known as Estel_18:40
*** M4rtinK has quit IRC18:53
*** int_ua has quit IRC19:01
*** NIN101 has joined #maemo-ssu19:18
*** mirandir has joined #maemo-ssu19:23
DocScrutinizer51hmm19:42
*** dafox has quit IRC19:42
DocScrutinizer51that might explain a lot :-P19:42
DocScrutinizer51alas if it's really a bot, it will blow each turing test outa the water19:43
DocScrutinizer51by simply redefining the test parameters radically19:44
*** Atarii has joined #maemo-ssu20:14
*** Atarii has joined #maemo-ssu20:14
*** krayon has joined #maemo-ssu20:18
*** krayon has quit IRC20:18
*** krayon has joined #maemo-ssu20:18
*** arcean_ has joined #maemo-ssu20:27
*** arcean has quit IRC20:30
DocScrutinizer51turing redefined: if tester headbangs before 15min are over, obviously bot wins20:31
DocScrutinizer51other forms of giving up count as well: going insane, shoot yourself...20:32
DocScrutinizer51shoot the bot/non-bot...20:35
DocScrutinizer51well, in thois latter case it's arguable if the bot won20:35
* merlin1991 headdesks20:36
merlin1991which sane person timestamps deltas with the beginning of the delta?20:36
freemangordonmerlin1991, hi20:37
freemangordonread the log?20:37
merlin1991just did, sry I actually wanted to pester x-fade today but didn#t manage20:37
freemangordonok, as I was under the impression that you told meyesterday it is mag who is pestering X-Fade20:38
DocScrutinizer51merlin1991: lol20:38
merlin1991nah I wanted todo that myself20:38
freemangordonthe second think - I uploaded FB crap on gitorious with only one of the two keys :)20:38
DocScrutinizer51sure thing, keep the fun for yourself ;-zp20:38
freemangordona kind of compromise :D20:39
freemangordonbut mag wants them both. please, do some agreement with him and tell me what to do20:39
freemangordonmerlin1991 ^^^20:39
merlin1991jsut load the 2nd one20:39
freemangordonupload the second one on gitorious?20:40
freemangordonok20:40
merlin1991I'm the minority in here saying no so I guess I'm overruled20:40
DocScrutinizer51could any of you guys please ask quim?!20:40
freemangordonDocScrutinizer51: I can bet qgil has NFC about FB API and keys20:40
DocScrutinizer51I'd rather trust in his decision on it20:40
freemangordonBut qhat kind of decisiion one can make if he is not informed on the matter at all?20:41
DocScrutinizer51he for sure has an idea of API access keys in general20:42
freemangordonDocScrutinizer51, I have move than idea for security, keys and such (that is a part of my job). But it does not mean I know a shit about FB20:42
*** int_ua has joined #maemo-ssu20:43
DocScrutinizer51it's not FB which is the problem here, it's a key handed out to NOKIA by FB20:43
freemangordonIt is MohammadAG that is the most knowledgeable, so maybe we should trust him20:43
freemangordonand that key flls under the TC of FB, not Nokia20:44
freemangordon*falls20:44
freemangordonas it is FB who is the key issuer and SP20:44
DocScrutinizer51I'd trust him if we hadn't that debate and 'minority report'20:44
freemangordonwe have the debate because we lack knowledge20:45
freemangordonand it is the same for qgil20:45
DocScrutinizer51and it is NOKIA wo owns the key20:45
freemangordonno, it is FB who owns the key20:45
DocScrutinizer51not at all20:45
*** mase76 has joined #maemo-ssu20:45
freemangordonwant a bet?20:45
DocScrutinizer51evidence: it's in a nokia blob20:45
freemangordonbut it is authourized by FB, not Nokia20:46
DocScrutinizer51that's nothing to do with ownership20:46
freemangordonDocScrutinizer51, it has all to to with the ownership, it is the same if you say that you are owner of the public key in an RSA pair20:47
freemangordonthe owner is the issuer20:47
DocScrutinizer51itMs NOKIA who's responsible to obey the EULA for that ke20:47
freemangordonnot the user20:47
DocScrutinizer51key20:47
DocScrutinizer51rabulism20:48
freemangordonbut it is the user who will be declined access to FB with that key (if FB decides) not Nokia20:48
freemangordonand that is the poit20:48
DocScrutinizer51ask quim, instead of speculating20:48
freemangordon*point20:48
DocScrutinizer51handwaving20:49
freemangordon[20:38] <freemangordon> DocScrutinizer51: I can bet qgil has NFC about FB API and keys20:49
DocScrutinizer51so what20:49
DocScrutinizer51search somebody to hol.d the bet or what?20:49
DocScrutinizer51I'm not interested20:50
DocScrutinizer51I know quim some time20:50
DocScrutinizer51and it's irrelevant what he got a clue of or not20:50
DocScrutinizer51asking him is the only PC thing20:50
freemangordonwell, I would like to know what merlin1991 and mag will decide20:51
DocScrutinizer51they already did20:51
freemangordonqgil is not a CSSU maintainer last time I chacked20:51
freemangordon*checked20:51
freemangordonso, I will upload the second key20:51
freemangordongtg20:52
freemangordonbye for now20:52
mase76hi! does anybody know, if openmediaplayer reads the gain tagged with mp3gain?20:53
*** mase76 has quit IRC21:05
merlin1991bah21:06
merlin1991was just about to answer21:06
merlin1991(close to om uses mafw thus NO!)21:06
Sc0rpiusman21:11
Sc0rpiusdo you know if there's a way to make the middle top button (the one that unlocks the screen) behave EXACTLY like the one at the side ?21:11
Sc0rpiusbecause my side button (the one that slides) is so worn that it's hard for me to slide it anymore21:12
Sc0rpiusand the button that unlocks (but then swipe) is totally useless!!! I would love to make it unlock/lock the screen without sliding shit21:12
Sc0rpiusI guess there's a way since there are people that have made custom screensavers or something.21:13
DocScrutinizer06in former times my advice would've been "delegate it to the council, they'll feature it out" - nowadays it seems in sovjet russia out of features councils your delegation21:18
DocScrutinizer06anyway fact is we're redistributing non-foss stuff owned by Nokia, even if we change that key. This means we ought talk to them prior to doing so21:19
DocScrutinizer06change the notation of that key (to ocatl for example)21:20
DocScrutinizer06octal21:20
DocScrutinizer06yet another point nobody mentioned yet: only Nokia (and FB, but they won't) can tell if that key FB granted usage to Nokia actually runs under same TOS like the ones somebody (merlin?) quoted here21:22
DocScrutinizer06their "contract" might differ vastly from what you find on FB's webpage for end users and other ants21:23
DocScrutinizer06Sc0rpius: straight brute force approach: redefine GPIO numbers in kernel21:26
DocScrutinizer06but that might fail, as... I mean it's the *power* button, not some arbitrary switch like camdoor or kbd-slider21:28
DocScrutinizer06it's probably as nasty as redefining ctrl-alt-del for ISA21:28
Sc0rpiushmm21:28
Sc0rpiuswell I would want to unlock/lock the screen with single push21:29
Sc0rpiusand a long push to be the power button just like it is now but without showing menu or annoying swipe screen21:29
DocScrutinizer06how about fixing the slider ?21:29
Sc0rpiusI would need another button, I've slided too much and the little bar is flat so the whole button is flat21:30
DocScrutinizer06a long push to power button does (TZZZDUM) power off21:30
Sc0rpiusyeah that's fine.21:30
Sc0rpiusbut I would like to redefine the short push21:30
Sc0rpiusto lock/unlock without annoyances21:30
DocScrutinizer06the short push also does power off, though with an option for software to intercept21:31
DocScrutinizer06it's a quite special kind of button21:31
Sc0rpiuswell I saw an application in tmo these days21:32
Sc0rpiusI forgot the name21:32
DocScrutinizer06I'd get a spare part for the slider button21:33
Sc0rpiusQtLockScreen21:33
Sc0rpiusI guess the guy that wrote it somehow intercepted the button21:33
Sc0rpiusand the whole thing actually21:33
Sc0rpiushe made a video21:33
Sc0rpiushttp://www.youtube.com/watch?v=hUK7OvJZGdo21:33
Sc0rpiusI should test his application first since it doesn't have a slide (but it has a button)21:34
Sc0rpiusand then read his source code21:34
DocScrutinizer06no, it's probably low level maemo (even kernel) that intercepts it21:34
DocScrutinizer06the higher level stuff prolly will use dbus to learn about power button pressed21:34
Sc0rpiuswell but he definitely made it21:35
DocScrutinizer06well sure, you can replace what's there and already doing it by some own stuff21:35
DocScrutinizer06but I guess it's easy to make it behave exactly like the lockslider21:36
DocScrutinizer06it's NOT easy21:36
DocScrutinizer06honestly, why don't you fix the friggin plastic lever21:36
Sc0rpiusand where I can find a new one21:36
DocScrutinizer06or get a spare21:36
Sc0rpiusthe N900 has been discontinued for so long21:37
Sc0rpiusit's very hard to find replacements for it21:37
DocScrutinizer06I could dremel one for you, from solid gold ;-D21:37
Sc0rpiushehehehe21:37
DocScrutinizer06or just glue a layer of solid gold handle on top of the worn plastic21:38
* DocScrutinizer06 wonders what users do to lockslider, as Sc0rpiusisn't the only one with exactly this problem21:39
Sc0rpiusthat's the ONLY hardware button I use21:43
Sc0rpiuseverytime I check my phone I slide it21:44
Sc0rpiusthat's like 129035468912346589127364981273465 times a day21:44
*** xnt14 has quit IRC21:46
*** MohammadAG has quit IRC21:46
*** xnt14 has joined #maemo-ssu21:48
*** MohammadAG has joined #maemo-ssu21:48
DocScrutinizer06sure thing, nevertheless this plastic lever has to be a defective part as well21:49
DocScrutinizer06otherwise I can't see how anybody except Krueger would ruin it by operating it21:49
DocScrutinizer06except of course when dust and fine sand blocks the slide22:03
DocScrutinizer06anyway, looking at it I think you even can drill a 1mm hole into it ans screw a knob on top22:05
DocScrutinizer06and*22:05
*** MohammadAG has quit IRC22:05
*** MohammadAG has joined #maemo-ssu22:05
DocScrutinizer06or glue a 0.8mm steel bolt into the 1mm hole, with a nice round end that just protrudes 1.5mm over the flat surface of the bump sliding in the hole22:06
*** trbs has joined #maemo-ssu22:07
DocScrutinizer06s/in the hole/in the case apperture/22:07
infobotDocScrutinizer06 meant: or glue a 0.8mm steel bolt into the 1mm hole, with a nice round end that just protrudes 1.5mm over the flat surface of the bump sliding in the case apperture22:07
MohammadAG<freemangordon> It is MohammadAG that is the most knowledgeable, so maybe we should trust him22:10
MohammadAGhmm?22:10
*** int_ua has quit IRC22:24
*** int_ua has joined #maemo-ssu22:39
*** int_ua has quit IRC22:45
*** int_ua has joined #maemo-ssu22:55

Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!