IRC log of #maemo for Sunday, 2016-02-07

*** fishbulb has joined #maemo00:13
fishbulbhi, I'm getting messages not delivered with yappari00:14
fishbulband I think I might have missed a call00:14
fishbulbthe 3g modem is on all the time00:14
fishbulbis there a way I can test that somehow? some diagnostic tests or whatever?00:14
Sicelo009Ntest what?00:15
fishbulbI have no idea. the modem?00:15
fishbulbthat modem00:16
Sicelo009Nopen other websites, etc.00:16
fishbulbyeah that all works00:17
fishbulbmaybe yappari isn't working I dunno.00:17
fishbulbthough it does.. pretty much. it's only skipping messages from one chick on an iphone00:17
Sicelo009Nknown problem00:18
*** shamus has joined #maemo00:36
fishbulbit is?00:45
fishbulbthese are pretty important messages to me00:45
fishbulbthat's just an ongoing problem?00:46
*** heroux has quit IRC00:47
Sicelo009Nyou may have to buy an iphone or android phone then00:50
Sicelo009Nit's a pretty 'big' problem ... we donm5 currently have full textsecure on N900 ... that's the cause of your problem00:51
fishbulbwhat is textsecure00:53
*** heroux has joined #maemo00:54
fishbulbI have no money for a friggen android00:54
ceeneencryption protocol of whatsapp00:54
fishbulbI have a tablet...00:54
fishbulbyeah but these aren't encrypted messages00:54
ceenethey are00:54
ceenethere are android phones for like 60 bucks00:55
ceenesecond hand shitty ones should cost a fraction of that00:55
fishbulbwhat are?00:56
fishbulbyou know they are encrypted for sure?00:56
fishbulbdude I have this phone partly because I'm broke00:57
ceeneyes, they are encrypted00:57
fishbulbhow come half of them work then00:57
fishbulbit's just one contact that doesn't00:57
*** jonwil has joined #maemo01:02
ceenethings that happen when a protocol is kept secret and undocumented01:04
ceenesometimes thing work, sometimes don't01:04
*** RedM has quit IRC01:05
*** RedW has joined #maemo01:05
*** Muumipeikko has joined #maemo01:06
*** arossdotme-planb has joined #maemo01:19
fishbulbhow'd you get this far then?01:19
fishbulbit's not like I'm complaining either this is totally awesome, but then one person freaks out if their messages don't go through, because it's like... one of the reasons I can still keep this phone01:20
*** arossdotme has quit IRC01:21
*** disco_stu_droid has joined #maemo01:28
*** disco_stu has quit IRC01:28
*** disco_stu_droid is now known as disco_stu01:29
*** heroux has quit IRC01:33
*** heroux has joined #maemo01:34
*** sunshavi has joined #maemo02:01
fishbulbI like the improvements too02:10
fishbulbthe images go to the image directory with the other photos now02:10
bencohceene: have you looked at the other opensource implems?02:13
bencoh(yowsup for example)02:15
bencoh(and whatsapp-purple)02:16
bencohah, looks like whatsapp-purple has the same issue02:21
*** ruskie has quit IRC02:22
*** DrCode has quit IRC02:22
*** ruskie has joined #maemo02:24
*** akossh has quit IRC02:25
*** RedM has joined #maemo02:35
*** DrCode has joined #maemo02:35
*** RedW has quit IRC02:35
*** florian has quit IRC02:46
*** lopx has joined #maemo02:53
lopxlong time02:53
lopxanyone on?02:54
*** fishbulb has quit IRC02:58
lopxi know someone is always on :)03:04
*** hashcore has quit IRC03:11
*** lopx has quit IRC03:22
*** sunshavi has quit IRC03:27
*** xes_ is now known as xes03:42
*** Muumipeikko has quit IRC03:48
*** arossdotme has joined #maemo04:08
*** arossdotme-planb has quit IRC04:10
*** arossdotme-planb has joined #maemo04:13
*** arossdotme has quit IRC04:15
*** arossdotme has joined #maemo04:17
*** arossdotme-planb has quit IRC04:20
*** arossdotme-planb has joined #maemo04:23
*** arossdotme has quit IRC04:25
*** eMHa has joined #maemo04:33
*** arossdotme-planb has quit IRC04:36
*** eMHa__ has quit IRC04:38
*** vakkov has quit IRC04:38
*** heroux has quit IRC04:42
*** Defiant has quit IRC04:45
*** Defiant has joined #maemo04:48
*** heroux has joined #maemo04:52
*** Hurrian has quit IRC04:58
*** LauRoman has joined #maemo05:04
*** hashcore has joined #maemo05:08
*** Hurrian has joined #maemo05:12
*** arossdotme-planb has joined #maemo05:12
*** heroux has quit IRC05:13
*** Pali has quit IRC05:18
*** heroux has joined #maemo05:27
*** Kabouik_ has joined #maemo05:29
*** LauRoman|Phone has quit IRC05:30
*** Kabouik has quit IRC05:31
*** lxp has joined #maemo06:01
*** lxp1 has quit IRC06:03
*** jonwil has quit IRC06:05
*** heroux has quit IRC06:10
*** heroux has joined #maemo06:10
*** DocScrutinizer05 has quit IRC06:21
*** DocScrutinizer05 has joined #maemo06:21
*** protem has quit IRC06:56
*** heroux has quit IRC06:58
*** pagurus has quit IRC07:11
*** heroux has joined #maemo07:11
*** auenf has quit IRC07:17
*** heroux has quit IRC07:33
*** heroux has joined #maemo07:33
*** auenf has joined #maemo07:43
*** arossdotme has joined #maemo07:47
*** sparetire has quit IRC07:47
*** arossdotme-planb has quit IRC07:49
*** arossdotme-planb has joined #maemo07:53
*** arossdotme has quit IRC07:57
*** Vajb has quit IRC07:59
*** Vajb has joined #maemo08:06
*** heroux has quit IRC08:07
*** Hurrian has quit IRC08:26
*** Sicelo009N has quit IRC08:30
*** Sicelo009N has joined #maemo08:32
*** robink_ has quit IRC08:38
*** heroux has joined #maemo08:41
*** robink_ has joined #maemo08:41
*** Hurrian has joined #maemo08:48
*** auenf has quit IRC08:55
*** Sicelo009N has quit IRC08:55
*** auenf has joined #maemo08:59
*** Sicelo009N has joined #maemo09:00
*** heroux has quit IRC09:17
*** heroux has joined #maemo09:24
*** Vajb has quit IRC09:25
*** auenf has quit IRC09:36
*** auenf has joined #maemo09:37
*** Sicelo009N has quit IRC09:38
*** heroux has quit IRC09:52
*** auenf has quit IRC09:52
*** heroux has joined #maemo09:53
*** futpib has joined #maemo09:55
*** auenf has joined #maemo09:57
*** heroux has quit IRC10:17
*** heroux has joined #maemo10:17
ceenebencoh: yep, and yowsup implements textsecure, which is really the only way to go now10:20
*** Vajb has joined #maemo10:22
*** eijk_ has joined #maemo10:50
*** arossdotme has joined #maemo11:08
*** arossdotme-planb has quit IRC11:09
*** eijk_ has quit IRC11:15
*** heroux has quit IRC11:18
*** hashcore has quit IRC11:24
*** hashcore has joined #maemo11:25
*** heroux has joined #maemo11:31
*** xelo has joined #maemo11:40
*** Vajb has quit IRC12:08
*** Pali has joined #maemo12:08
*** heroux has quit IRC12:13
*** eMHa_ has joined #maemo12:14
*** eMHa has quit IRC12:14
*** heroux has joined #maemo12:15
*** xelo has quit IRC12:22
*** arossdotme-planb has joined #maemo12:24
*** xelo has joined #maemo12:24
*** arossdotme has quit IRC12:27
*** jonwil has joined #maemo12:28
*** arossdotme has joined #maemo12:29
*** lobito1 has joined #maemo12:31
*** arossdotme-planb has quit IRC12:32
*** lobito has quit IRC12:34
ceeneso i have libaxolotl packaged for maemo12:41
ceenestill have to check which are the .h files needed for the -dev package, but that's okay12:41
ceenei can start now more or less to implement textsecure on yappari12:41
*** arossdotme-planb has joined #maemo12:44
*** arossdotme has quit IRC12:45
*** xelo has quit IRC12:51
*** xelo has joined #maemo12:52
ceenemost of the work is really from coderus12:57
ceenei just backported it to qt4 and made the debian/ dir12:57
*** EgS has quit IRC13:05
*** EgS has joined #maemo13:06
ceenelinking error :(13:25
ceene/usr/lib/ undefined reference to `CiphertextMessage::CURRENT_VERSION'13:25
ceenethat's simply false13:25
ceenei'm just about to replace that with a #define13:31
ceenei can't see why it doesn't found that symbol13:31
ceenein fact, there are other symbols defined in the very same place in the very same way that don't trigger that error13:35
KotCzarnydifferent data type?13:36
KotCzarnythough it looks like a define13:36
ceeneit's the same13:36
ceeneclass CiphertextMessage13:36
ceenepublic: static const int UNSUPPORTED_VERSION         = 1; static const int CURRENT_VERSION             = 3;13:36
ceeneand there are a few more like that13:36
ceenei mean, i could replace that by a define13:36
ceenebut... i'd like to know what's happenning13:36
KotCzarnytry comparing to the working code?13:37
ceeneUNSUPPORTED_VERSION doesn't trigger that error13:37
ceenenor the other consts13:37
ceenethere's not much code i can test13:37
KotCzarnytry defining them like: public: static const int UNSUPPORTED_VERSION=1, CURRENT_VERSION=3;13:38
ceeneok, it won't hurt13:38
ceenebut doesn't make much sense, does it?13:39
KotCzarnywhen in doubt, ask the wiser, if wiser is not available try black magic13:39
ceenethe same13:39
KotCzarnythen its in the headers, but not in the .so13:40
KotCzarnycheck .so which symbols are defined13:40
ceenedpkg-shlibdeps: warning: symbol _ZN17CiphertextMessage15CURRENT_VERSIONE used by debian/libaxolotl/usr/lib/ found in none of the libraries.13:40
ceenethat's after compiling the .so13:41
DocScrutinizer05public: ?13:41
ceenebut i can't see why13:41
KotCzarnycurrent_versionE ?13:41
ceenei'll check the warnings, there are a few13:41
ceenena, it's just noise of the output13:41
ceenei've checked the source and there's not versionE anywhere13:41
DocScrutinizer05I don't think a "public:" is for the whole line of code13:42
ceeneit should affect everything that is below the public:13:42
KotCzarnytill private or something else changes types13:42
ceenemy paste before was fixed by irssi13:42
ceenebut it's public13:42
DocScrutinizer05err hmm, that's not how I thought the syntax of c and c++ would work13:42
DocScrutinizer05though I admit I have no clue13:43
KotCzarnyceene: nm|grep _VERSION13:43
DocScrutinizer05anyway if "public:"  was a switch that changes context for everything after, there must be a sort of "endpublic:"13:44
ceene[sbox-FREMANTLE_ARMEL: /tmp/N900/libaxolotl] > nm|grep _VERSION U _ZN17CiphertextMessage15CURRENT_VERSIONE13:44
ceeneDocScrutinizer05: it'd be private: or whatever13:44
ceenebut in its abscense it should just define verything as public13:44
KotCzarnycheck the lib compilation output13:45
KotCzarnyor maybe its in some other sublib13:45
KotCzarnyis it the only .so in that tree?13:45
ceeneyep, there's no other .so13:46
ceenei'll be checking the warnings13:46
ceenethere must be something somewher13:46
ceenei gotta go now, i'll tell you if/when i find something13:46
ceenethanks, cya!13:47
*** arossdotme has joined #maemo13:55
DocScrutinizer05I'd do a stupid thing and simply add a "public: " in front of the " static const int CURRENT_VERSION             = 3;"13:56
jonwilBy default things in C++ are private13:58
jonwilWait no, that's not strictly true, it depends on whether its a struct or a class13:58
*** arossdotme-planb has quit IRC13:59
KotCzarnyhis problem is undefined symbol in LIBRARY14:00
KotCzarnynot in his code14:00
DocScrutinizer05it suggests that a access_specifier affects exactly ONE member14:05
DocScrutinizer05like I would have thought the syntax of a proper language is defined14:06
DocScrutinizer05wait, I should read to bottom14:07
DocScrutinizer05heck no, ceene is right: >>...and two member functions with public access: the functions set_values and area,<<14:08
DocScrutinizer05c++ is a mess14:10
DocScrutinizer05compare >> int a, b; <<  to  >> public: t x; t y; << and the different syntax14:12
DocScrutinizer05in first the semicolon is delimiter for the type 'int' specifier. in second there is no delimiter for 'public:'14:14
* jonwil is no closer to figuring out what libmaemosec and this strange absolute_pathname function is really doing (not just "it gets the absolute pathname" but what the actual bits of code are doing)14:15
jonwilIf I can figure this out I can commit a new set of root certificates no problems :)14:16
* DocScrutinizer05 adds a >>goto public: << at random places14:16
DocScrutinizer05jonwil: which line of code is unclear?14:17
DocScrutinizer05and why is it a problem to start with? doesn't it do just what it's supposed to do?14:18
jonwilThe issue is that /etc is a symlink in Scratchbox14:18
jonwiland that function is returning /blah/blah/blah/etc instead of /etc14:18
jonwilmeaning the output of cmcli is not what we need it to be for the device14:18
DocScrutinizer05well, that's exactly what it's supposed to do, no?14:19
jonwilok, I think I understand it a bit more now14:19
jonwilIt gets information about the passed in path (the lstat call)14:20
jonwilthen it says "is this a symbolic link"14:20
DocScrutinizer05actually iirc it does a chdir into the provided path and then gets the cwdir (which is supposed to be the plain raw pathname without symlinks) and returns that14:21
jonwilno, the code definatly does more than that14:22
DocScrutinizer05sure it does more, it also checks if the path is absolute or relative, and in latter case it prepends the path to current dir in front14:23
DocScrutinizer05it's quite weird code, but maybe there's a reason for that14:24
DocScrutinizer05there might be attack vectors otherwise14:24
DocScrutinizer05or problems with process specific name spaces, or whatever14:25
DocScrutinizer05alas there's little comment in source on why all that is needed14:26
DocScrutinizer05that's why I asked what's the source of trouble, after all I think it worked so far, no?14:27
jonwilThe issue is that /etc in scratchbox is a symlink.14:28
DocScrutinizer05or do we run into a sleeping bug in that code, in a specific environment that hasn't been tested so far?14:28
KotCzarnyfor full path it should use realpath()14:29
KotCzarnynot some cd&pwd14:29
DocScrutinizer05well, maybe the code is exactly for handling stuff in an unexpected manner in situations like a scratchbox environment14:29
KotCzarnyjonwil, you can try changing symlink into mount --bind, if its only for the building package means14:29
jonwilBecause /etc is a symlink when cmcli puts path names for certificates into files under /etc/secure/s it stores the target of the symlink14:29
KotCzarnyor just change the code in cmcli?14:30
jonwilnot the actual link14:30
DocScrutinizer05so e.g. a chrooted process would still use the system wide cert store? or whatever14:30
*** heroux has quit IRC14:31
*** heroux has joined #maemo14:31
ceenei don't really see why would anything other than tar, backup utilities, etc, would mind if some path is symbolic or not14:31
ceene(hi again!)14:31
DocScrutinizer05it seems the problem isn't the code but the semantics and rationale behind it14:32
DocScrutinizer05ceene: same here14:32
*** arossdotme has quit IRC14:33
DocScrutinizer05ceene: there are other things as well, like bindmounts and chroot14:33
jonwilBasically the issue is that when you use cmcli to install a certificate, it puts a path to that certificate in a file under /etc/secure/s. On Scratchbox that path is not /etc but /blah/blah/etc or whatever when we need it to be /etc since that's where things live on the device14:33
ceenei can't see why cmcli cares about the nature of the path14:34
DocScrutinizer05might actually be a bug that processes symlinks in a way originally meant for e.g. bindmounts or chroots14:34
DocScrutinizer05unless we find out about the rationale behind that code, we prolly never will know how it's supposed to work and what's wrong14:37
DocScrutinizer05it might be everything from noob's cargo cult coding to a highly sophisticated (though maybe flawed) security feature to handle special attack vectors or environments14:39
DocScrutinizer05knowing the origins of this code might help a lot. e.g. if it originally been meant for android or docker or whatever14:40
DocScrutinizer05can't you do a 'git blame' or whatever, to see who and when and why introduced that code snippet?14:42
jonwilThe project is clearly 100% Nokia14:42
jonwilmaemo-security-certman was 100% the work of Nokia14:43
DocScrutinizer05so odds are it's actually already meant to handle scratchbox environment in a special manner14:43
DocScrutinizer05possibly never tested though14:44
DocScrutinizer05".. err humm, let's use the host's certificate store when we run in scratchbox..." ?14:45
*** Vajb has joined #maemo14:45
jonwilIts got nothing to do with where it gets the certificate store from and no it wont use the host certificate store on Scratchbox as such14:45
DocScrutinizer05<spews a few dozen lines of code>  "... looks good..."14:46
jonwilThe symlink in question has nothing to do with the certificate store and is a generic symlink for all of /etc14:46
KotCzarnyjonwil, maybe you should provide certs as source and run cmcli during install/reinstall?14:46
jonwilso for whatever reason the people behind Scratchbox decided that /etc in a scratchbox environment would be a symlink elsewhere14:46
DocScrutinizer05please don't focus on the symlink, it's evidently the pathological case that hasn't been thought of14:46
jonwilThe root problem is that the files under /etc/secure/s contain absolute paths to the certificate files (which makes a lot of sense to me) but under Scratchbox, its storing the /blah/blah/etc symlink target rather than just /etc (we need it to store /etc)14:47
DocScrutinizer05that code is meant to handle something else14:47
DocScrutinizer05how about you simply comment out that code?14:48
DocScrutinizer05on target platform it doesn't do anything useful anyway aiui, and in scratchbox it doesn't work14:49
DocScrutinizer05hey, when you have the author, you could simply send a mail and ask for a hint what's the deal with absolute_path14:50
jonwilyeah I was about to do that14:56
DocScrutinizer05however see the very next function in code, which is using absolute_path along with a few other tricks to get a process context14:59
DocScrutinizer05also for processes handled by maemo-launcher15:00
DocScrutinizer05in this context stuff like optification and the involved bindmounts come to mind15:01
jonwilmail sent :)'15:01
DocScrutinizer05I hope it's no æ mail addr, or you rather had to use aunt google to find a more recent addr to contact the author15:02
*** DocScrutinizer05 has quit IRC15:03
*** DocScrutinizer05 has joined #maemo15:03
DocScrutinizer05while searching for æ15:04
ceenei still can't see why this const is then undefined15:04
* DocScrutinizer05 neither15:05
ceenei'm just about to replace it with a #define15:05
ceeneit's just a number15:05
ceeneand not used for anything too useful, in fact15:05
DocScrutinizer05a define is compile time, a public: is exported to the ABI, no?15:05
ceeneyes, but given that it's just a const defined in a .h file that can be simply included15:06
ceeneit doesn't matter at all15:06
KotCzarnyceene, its undefined in the library object itself15:06
ceeneKotCzarny: but why?15:06
DocScrutinizer05not if that thing is a .so or whatever15:06
ceenei'm the one compilnig the library15:06
KotCzarnysome compile error?15:06
ceenei've checked warnings and nothing related to it15:07
DocScrutinizer05optimization kicked it out?15:07
ceenecould be, i guess15:07
DocScrutinizer05use it in code, for some nonsense15:07
jonwilThe email address I found when I made contact with the author of maemo-security-certman came from github15:08
ceenei think it can be that same question15:08
ceenec++ is fucking weird15:09
KotCzarnymildly put15:09
DocScrutinizer05[Sun 07 February 2016] [13:10:54] <DocScrutinizer05> c++ is a mess15:10
DocScrutinizer05jonwil: sounds good15:10
ceenei believe this is the answer15:10
ceenewhich of course is nuts15:10
ceenebut it's probably going to work, i guess15:10
ceeneso much for code reusability15:11
ceenei have to type the exact same thing twice15:11
ceeneso i can make twice the mistakes i'd do otherwise15:11
DocScrutinizer05hail c++15:12
DocScrutinizer05use inline assembler ;-)15:12
DocScrutinizer05at very least you know *exactly* what's going to happen then15:12
ceenethere's someone right now, with an unwritable name, who is proud of himself for inventing c++15:13
ceenehe sleeps well at nights15:13
ceeneprekeywhispermessage.o: In function `.LANCHOR0':15:15
ceeneprekeywhispermessage.cpp:(.rodata+0x40): multiple definition of `CiphertextMessage::UNSUPPORTED_VERSION'15:15
ceenekeyexchangemessage.o:(.rodata+0x0): first defined here15:15
ceenei'm done with it15:15
ceenethat's it15:15
ceene#defines it's going to be15:15
*** Vajb has quit IRC15:15
DocScrutinizer05Bjarne Stroustrup ?15:16
ceenethat one!15:16
ceenenow it compiles, and yappari links against it without any problem15:19
* DocScrutinizer05 puts B. Stroustrup next to L. Poettering on his secrete list15:19
ceenethat's a list that secretes people you hate?15:20
DocScrutinizer05>>links against it without any problem<< doesn't tell much yet, you might run into an assert at yappari runtime, like "version of doesn't fit"15:20
DocScrutinizer05((list)) up to you to figure what that list is for ;-)15:21
ceeneDocScrutinizer05: i'm the one who is going to write yappari code to match my libaxolotl version15:22
ceeneso that won't happen15:22
DocScrutinizer05:-) good15:22
ceenemaybe my libaxolotl will differ from mainstream15:22
ceenebut i don't care much15:22
DocScrutinizer05I never heard of that strange critter ;-)15:22
ceenemoreso if this stupid thing is the only one that's blocking my development15:22
ceeneyeah, it's a funny name15:22
KotCzarnycompile statically? :>15:22
ceeneit's the thing that implements textsecure ciphering15:22
DocScrutinizer05axolotl, wasn't that an ugly amphibian?15:23
ceenenah, i already have .debs for libaxolotl and libcurve2551915:23
ceeneyeah, it's a salamander15:23
ceeneor something like that15:23
ceenei don't think it's ugly, just funny looking :)15:24
ceenelooks like a fish with legs15:24
DocScrutinizer05>>As of 2010, wild axolotls were near extinction<< ;-P15:25
ceenepoor thing15:25
DocScrutinizer05wonder what that implies for libaxolotl15:25
ceeneas long as wa uses it... it'll have a long live15:26
ceenenot that i'm against crypto, but i find that whenever i have to use some crypto library they are all a mess15:26
DocScrutinizer05so wa is the zoo for axolotls15:26
ceenei guess it's because crypto is a hard subject in itself, so implementation are also hard problems15:26
ceenebut man, all this initialization, keys, prekeys, shared secrets...15:27
DocScrutinizer05yeah, it's mindboggling15:27
DocScrutinizer05 are the very simple looking basics15:31
DocScrutinizer05 already gets nasty15:32
*** jonwil has quit IRC15:32
ceenei've read agatha christie's And Then There Were None, which had less people than that and much more gratifying to read and guess whodunnit15:33
bencohthat one was really quite enjoyable, actually :)15:36
DocScrutinizer05 is where we get to the more interesting things15:36
bencoh(and then there was none)15:36
DocScrutinizer05PFS isn't really simple to understand or to implement15:44
DocScrutinizer05axolotl is even listed in as one of the implementations15:45
*** disco_stu has quit IRC15:51
*** disco_stu_droid has joined #maemo15:51
*** disco_stu_droid is now known as disco_stu15:52
*** arossdotme has joined #maemo15:53
*** msava has joined #maemo16:02
*** pagurus has joined #maemo16:16
*** disco_stu_droid has joined #maemo16:17
*** disco_stu has quit IRC16:17
*** Vajb has joined #maemo16:18
*** disco_stu_droid is now known as disco_stu16:18
freemangordonPali: any idea how to prevent screen going blank after trying to reboot or shutdown?16:18
Palino idea...16:19
*** trumee has quit IRC16:25
*** trumee has joined #maemo16:28
DocScrutinizer05freemangordon: hmm? that sounds a tad fuzzy16:32
DocScrutinizer05am I missing some context?16:32
freemangordonDocScrutinizer05: when you reboot N900, display goes black16:33
DocScrutinizer05well, eventually the video buffer gets initialized, yes16:33
freemangordonI want to prevent that, to be able to capture console aoutput16:33
DocScrutinizer05ooh you meant after console already went active16:34
freemangordonwell, I guess commenting out acx565akm_disable(dssdev); in remove() will do the jib :)16:34
DocScrutinizer05aaah, during shutdown16:35
freemangordonhmm, no :(16:35
freemangordonyes, during shutdown16:35
DocScrutinizer05yes, what you said then. sounds reasonable16:36
enycDocScrutinizer05: is there a new draft hackerbus spec to look at?16:36
DocScrutinizer05however you also need to prevent system shutdown then16:36
freemangordonbut doesn't work16:36
DocScrutinizer05enyc: wpwrak hasn't provided a public version yet16:36
enycokies =)16:37
DocScrutinizer05maybe prod him about it in neo chan16:38
enycooh oops16:38
freemangordonDocScrutinizer05: what should be done so I to have working serial output from N900?16:38
DocScrutinizer05freemangordon: a good question. the classical approach would be to use serial (or other low complexity) interface to export all printout in realtime16:39
freemangordonwell, there is serial, I just don't have the way to interface it :)16:40
DocScrutinizer05frewuently UART console via a special debug adapter is used for that16:40
freemangordonis there such a way, some standard device to connect to omap serial outs16:40
freemangordon"debug adapter"?16:41
bencohthe debug port below battery?16:41
bencohmore pads than port, but ...16:41
DocScrutinizer05for N900 yes16:41
DocScrutinizer05it's 1V8 TTL UART16:42
freemangordonTTL 1.6V?16:42
freemangordonops, 1.816:42
freemangordonno way TTL :)16:42
bencohdoes it go straight to omap?16:42
DocScrutinizer05well "TTL"16:42
bencohsounds like instant fry the moment you do something wrong16:43
freemangordonyou mean 0V for 0 and 1.8V of 1?16:43
freemangordonand what about the connector? can I use some standard plug?16:44
freemangordonyep, read those16:44
freemangordondoesn;t help much :)16:44
DocScrutinizer05tbh I'd suggest to solder some thin wires to the testbads16:45
freemangordonwon;t do that on my "new" device16:45
KotCzarnydo it on the old then?16:45
freemangordon"old" one is my daily device :)16:45
DocScrutinizer05there are generic "phoenix" adaper "pliers"16:46
bencoh/dev/ttyS0 might be wrong on modern kernels16:46
bencoh(it's probably something like ttyO0)16:47
Defiantbencoh: srsly? I thought they abandon ttyO* again16:47
freemangordonDocScrutinizer05: what I am supposed to search on ebay for?16:47
DocScrutinizer05lemme try16:48
bencohDefiant: ah? maybe then16:48
bencoh£55.99 :/16:51
DocScrutinizer05bencoh: those testpads are not designed for convenient low cost connectors to get attached16:52
bencohI wonder how it works16:52
freemangordonme too16:53
DocScrutinizer05freemangordon: an alternative would be to hack the kernel to use the UART implementation in musbcore instead. the USB allows doing simple UART TX/RX via the D+/- lines16:53
freemangordonI understand nothing from those pictures16:53
bencohme neither16:54
bencohespecially the one with the stripped board16:54
DocScrutinizer05it's a simple clip (acrylic) that presses some pogopins to the DUT16:54
bencohhow does it attach to it16:54
freemangordonDocScrutinizer05: the problem is that the bug I chase needs usb gadget driver loaded16:55
DocScrutinizer05the pogopins sit on a small adapter PCB 'plugged' into the tip of the clip16:55
freemangordonok, but how to power that assembly?16:55
DocScrutinizer05I guess the crocodile clips are meant to contact to the battery blades16:56
freemangordontoo complicated for my test16:56
DocScrutinizer05well, it gets more convenient and "simple" if you shell out way more bucks16:57
DocScrutinizer05the Nokia test jig16:57
*** xelo has quit IRC16:58
*** xelo has joined #maemo16:58
DocScrutinizer05a possibly more appealing approch is to hack something like this:
DocScrutinizer05 ... 009.jpg17:00
freemangordonwhat is this?17:02
*** shentey has joined #maemo17:02
freemangordonI mean - someone made it or you haev it?17:02
Sicelosre made one too .. and gave step-by-step instructions for someone else to make for themselves17:03
DocScrutinizer05I bought it on fleabay/$seller long ago, for a 5 bucks. Alas I can't find it in the interwebs anymore, only stuff like this:
freemangordonDocScrutinizer05: oh, so you have a way to catch serial output?17:04
* freemangordon gives DocScrutinizer05 zImage and modules and waits for the logs :)17:04
DocScrutinizer05no, that thing connects to BB5 F-Bus, not to the UART pads17:06
*** heroux has quit IRC17:07
DocScrutinizer05you however can hack that thing moreately easily, you just need to 'move' the pogopins17:07
*** heroux has joined #maemo17:08
freemangordonbut I don;t have the serial converter chip as well17:08
freemangordonand I feel too old for hacking such things :)17:08
freemangordonso I'd rather buy something prebuild, or ask somebody on the channel  to capture the logs for me17:09
Siceloask sre then ;)17:10
Siceloor he's busy nowadays?17:10
freemangordonI don;t see him around17:10
bencohI suspect he might want to keep his adapter anyway :)17:10
Sicelolol, yes .. meant he could get him to test17:11
bencohsince he works on n900+kernel from time to time17:11
Sicelofreemangordon: tried xmpp? he usually responds instantly17:11
freemangordonSicelo: never used xmpp, besides fbchat and google chat :)17:11
Sicelookay. email then? :p17:13
Siceloor i could let him know you need him? but then how does that help17:13
DocScrutinizer05hacking a standard Nokia FBUS cable should be simple enough. Nokia F-Bus is 3V3 but you should easily be able to modify the 3V3 voltage to 1V8, there must be level shifters anyway17:13
freemangordonwell, you can ask him to join #maemo :)17:13
freemangordonSicelo: but yeah, maybe I should mail him17:14
DocScrutinizer05Nokia data cable it's dubbed17:14
bencohnokia data cable, the thing that existed back when 3310 was mainstream?17:15
freemangordonor simply wait for Pali's best friend amongst the kernel maintainers to return from holiday :D17:15
DocScrutinizer05or finally share your UPS address ;-)17:15
freemangordonDocScrutinizer05: hmm? you have my address17:16
DocScrutinizer05I do?17:16
Sicelonever seen sre here17:16
DocScrutinizer05then I probably could find it in my data desert if I already knew it ;-)17:17
freemangordonDocScrutinizer05: I gave it to you to send me BB-XM iirc17:17
DocScrutinizer05and did I?17:18
freemangordonbecause we've solved the problem so there was no need, or somesuch17:18
DocScrutinizer05yeah I remember now17:18
DocScrutinizer05I connected the BB-xM to Europa instead ;-)17:19
freemangordonDocScrutinizer05: 8.04.2015 22:1817:21
DocScrutinizer05will see what I can do17:21
DocScrutinizer05seems you're not eager to do any soldering17:21
freemangordonI miss the tools needed17:21
DocScrutinizer05ok, I think I got the tools17:21
freemangordonI don't want to risk breaking a device17:22
DocScrutinizer05meh, it takes breaking eggs to make an omelet17:22
freemangordonI have only one old(> 20 years) 10MHz russian scope and 40W iron17:23
DocScrutinizer05I see17:23
DocScrutinizer05I think my scope and iron are up to the task17:23
freemangordonsee, if there is noone to do it, I will, but knowing there are guys better suited for the job, I prefer to not ;)17:23
DocScrutinizer05stay tuned17:23
* DocScrutinizer05 idly wonders how to *send* data from Rigol scope to RS232 ;-)17:24
freemangordonPali: i am thinking to enable 550MHz before sending the patch, unless you remember why it is disabled17:25
Palifreemangordon: no idea why 550 is disabled17:27
freemangordonok, will enable it17:28
PaliDocScrutinizer05: any idea?17:29
Paliwhy could be 550MHz freq obsolate and disabled?17:29
DocScrutinizer05ooh, maybe instable?17:29
Palifor n900?17:29
*** sparetire has joined #maemo17:30
freemangordonnaah, it is not that one17:30
freemangordon550 is table17:30
DocScrutinizer05there are quite a few theoretically possible clock rates that simply don't work and thus are deprecated17:30
Palican you check nokia kernel and kernel-power?17:30
freemangordonit is enabled in stock kernel17:30
DocScrutinizer05well, then I dunno. however I don't see it listed in e.g. zzztop17:31
Palion my N900 with kernel-power 550 is enabled17:31
freemangordonPali: is it used, according to stats?17:31
*** kerio has quit IRC17:31
DocScrutinizer05 seems zzztop only lists *used* clock factors17:32
freemangordonDocScrutinizer05: zzztop ignores freq that were not used during the measurement17:33
Palikernel-config show that 550 is supported and enabled17:33
*** kerio has joined #maemo17:33
*** trumee has quit IRC17:34
DocScrutinizer05*cough*  cifsdnotifyd17:35
*** krnlyng has quit IRC17:36
Palihere in log is some info about 550Mhz freq17:38
freemangordonPali: BTW operating-points-v2 allows for so-called "turbo" freqs, which is an euphemism for OC imo :)17:39
*** trumee has joined #maemo17:39
freemangordonPali: I see nothing useful there, besides someones preferences17:40
Paliok, enable it17:41
Paliand what for 250MHz?17:41
freemangordon250 or 125?17:41
Pali250 is changed to 249.6MHz17:41
Paliand 125 disabled17:41
Paliboth are correct??17:41
freemangordon/sys/devices/system/cpu/cpufreq/policy0/stats # cat *17:42
freemangordon249600 1801517:42
freemangordon500000 4117:42
freemangordon550000 017:42
freemangordon600000 485917:42
freemangordonPali: ^^^17:43
Palihm... 550 is 0?17:43
freemangordonthis is qemu17:43
Palirigth, so 550 is disabled (so 0)17:44
freemangordonbut even on the device, 550 is rarely used17:44
freemangordon125 is disabled, so it is missing at all :)17:44
freemangordonPali: see
freemangordonthis is KP17:44
freemangordonwith a charger attached17:44
freemangordon(so 250 is not hit)17:45
*** pagurus has quit IRC17:46
*** heroux has quit IRC17:46
freemangordonPali: this is what I am going to send
*** heroux has joined #maemo17:47
DocScrutinizer05   I'm also lazy. freemangordon would you be content with this USB jack at the end of the cable from your new N900 devel device to your PC??17:47
Palistill... from where we have that 249.6MHz value??17:48
freemangordonDocScrutinizer05: I think it will be even more convenient17:48
freemangordonI guess this is ftdi RS232<->USB?17:48
Palialso see sre email: and his dmesg output:
bencohiirc I used 550 for some time on my device17:49
bencohbefore switching back to 60017:50
*** krnlyng has joined #maemo17:50
*** florian has joined #maemo17:50
*** arossdotme-planb has joined #maemo17:50
DocScrutinizer05FTDI TTL-232RG-VREG1V8-WE17:51
freemangordonPali: from some commit, can't find it now, gimme some time17:53
DocScrutinizer05freemangordon: >>This cables require USB drivers, which are used to make the FT232R in the cable appear as a virtual COM port.<<  I hope your linux has such a driver17:54
*** arossdotme has quit IRC17:54
freemangordonwell, it is ubuntu 14.04, I guess it has :)17:54
freemangordonat least there is no problem with FTDi adapters17:54
DocScrutinizer05ok, so I'll send you such a cable and a device where it's connected to17:55
freemangordonPali:   p, li { white-space: pre-wrap; }  df5c1e60a5213755d31631b7c12bf33bc1a6bb7017:55
Palifreemangordon: I know that commit, but do not know *why* is needed now17:56
Paliit is only in linux-N900 tree17:56
freemangordonPali: so, you think we should use 250 instead of 249.6?17:56
PaliI do not know, I'm just asking :-)17:57
DocScrutinizer05sounds sane, LOL17:57
Palianyway, see sre dmesg:
Paliand search for "Cannot find matching frequency"17:57
freemangordonPali: hmm, seems there were some bug which seem to be fixed17:58
*** arossdotme has joined #maemo17:58
DocScrutinizer05more puzzling is >> [   27.884918] clock: dpll1_ck: cannot round to rate 250000000  <<17:58
Palithere is 124.8MHz 249.6Mhz and 549.6Mhz17:59
freemangordonyes, because it cannot find a combination of mul/div17:59
*** arossdotme-planb has quit IRC17:59
freemangordonPali: ok, I'll try 250 instead of 249.618:00
DocScrutinizer05it tries to round but says it can't18:00
*** pagurus has joined #maemo18:01
DocScrutinizer05makes a damn lot of sense to round 249600000 to 25000000018:01
*** eMHa_ has quit IRC18:02
*** eMHa_ has joined #maemo18:02
DocScrutinizer05maybe there's a error threshold that's too small to round that value to next 'even' clock freq?18:02
*** shentey has quit IRC18:05
*** arossdotme-planb has joined #maemo18:05
freemangordonseems we can drop that commit18:06
freemangordonthis is N900, not qemu18:06
freemangordonPali: what about enabling 125MHz?18:07
*** florian has quit IRC18:08
Palitry... we will see :-)18:08
freemangordonPali: if we enable 125, there is no need to send a patch18:08
*** florian has joined #maemo18:08
Paliis Skry (Joni Lapilainen) somewhere around?18:09
PaliI would like to know motivation for that patch18:09
*** arossdotme has quit IRC18:09
freemangordonPali: i remember him hanging on #armlinux and #linux-omap18:11
*** pagurus has quit IRC18:11
freemangordon~seen skry18:12
infobotfreemangordon: i haven't seen 'skry'18:12
*** pagurus has joined #maemo18:12
Pali"Adjust MPU OPP values for RX51 -- The OPPs defined in the OPP table don't have exact matches with available MPU frequencies because N900 uses a different system oscillator than other 3430 boards. This is a hack to fix the 250MHz frequency, and also disable the obsolete 125MHz and 550MHz ones."18:13
Palithis is written in commit message18:13
*** arossdotme-planb has quit IRC18:14
freemangordonPali: yes, but it seems this is no longer valid18:15
DocScrutinizer05err, sounds like it couldn't get obsolete18:15
freemangordonPali: see
infobotDocScrutinizer05 meant: err, sounds like it couldn't get invalid18:15
DocScrutinizer05freemangordon: I assume you're quoting output of a stock kernel on N900? then please check the oscillator frequency used in the kernel sources, vs the real one in N90018:17
freemangordonDocScrutinizer05: this pastebin is from 4.5-rc118:17
*** pagurus has quit IRC18:18
freemangordonDocScrutinizer05: what to check then?18:18
*** pagurus has joined #maemo18:18
*** pagurus has left #maemo18:18
*** pagurus has quit IRC18:19
DocScrutinizer05when kernel simply assumes a "standard" oscillator freq, then of course the clock freq calculations will be just fine, however the data they base on is wrong, possibly introducing other nasty trouble with timing18:19
DocScrutinizer05N900 uses a 19.2MHz osc on HFCLKIN18:21
*** arossdotme has joined #maemo18:22
Palifreemangordon: so patch is not needed...18:22
Palior you can specify turbo-mode (disable by default) to have all data :-)18:22
Palianyway DT patch for C-states is needed18:24
freemangordonDocScrutinizer05: It is rather that patch is wrong, stock kernel uses those freqs as well18:24
*** RST38h has quit IRC18:24
freemangordonPali: for C-states?18:24
Paliexit latency18:24
DocScrutinizer05hmm ok then, strange18:24
freemangordonPali: but that seems too complicated for me :)18:25
*** RST38h has joined #maemo18:25
*** eMHa_ has quit IRC18:25
freemangordonPali: also, we don;t use DT cpu-udle driver :)18:25
freemangordonPali: in omap2plus_defconfig, there is CONFIG_CPU_IDLE=y, but no CONFIG_ARM_CPUIDLE18:31
freemangordonthe same for rx51_defconfig18:31
* freemangordon wonders how that works at all18:31
freemangordonand if18:31
DocScrutinizer05freemangordon: s/stock/mainline/18:34
DocScrutinizer05Nokia kernel uses an allegedly unusual oscillator of aiui 19.2MHz. I wonder what oscillator the mainline 4.5-rc1 uses18:35
DocScrutinizer05my fault to use the term "stock kernel"18:36
DocScrutinizer05this is from mainline kernel?18:40
DocScrutinizer05cool shit18:40
freemangordonwanna more?18:40
DocScrutinizer05looks like 1920018:40
DocScrutinizer05nah, thanks18:40
*** arossdotme has quit IRC18:41
*** arossdotme has joined #maemo18:45
*** krnlyng has quit IRC18:45
*** krnlyng has joined #maemo18:46
*** florian has quit IRC18:49
*** Vajb has quit IRC19:01
*** arossdotme-planb has joined #maemo19:06
*** arossdotme has quit IRC19:09
*** pagurus has joined #maemo19:12
*** pagurus has quit IRC19:14
*** pagurus has joined #maemo19:14
*** krnlyng has quit IRC19:17
*** Sicelo009N has joined #maemo19:18
*** Sicelo009N has quit IRC19:18
*** Sicelo009N has joined #maemo19:18
Sicelo009Nceene: ping19:19
*** eMHa_ has joined #maemo19:23
*** arossdotme-planb has quit IRC19:24
*** eMHa_ has quit IRC19:28
*** eMHa_ has joined #maemo19:28
*** heroux has quit IRC19:28
ceeneSicelo009N: pong19:30
Sicelo009Nany idea why my sister can't send me a message on whatsapp (Lumia 520)? she gets a yellow triangle with exclamation mark.19:32
Sicelo009Nthat's encryption maybe?19:32
ceenedo you see something in your log?19:32
ceenethough i've never seen that icon on any android19:33
Sicelo009Nhmm, i will check further. let me fire up sqlite19:33
ceeneyou don't need sqlite for this19:33
ceenejust old yappari.log19:33
ceenethe thing tha happens now is that someone sends you an encrypted message19:34
ceeneyappari then tells them "sorry, only clear text"19:34
ceeneand the other party sometimes resends the same message without encryption19:34
ceenebut some other times, they won't resend the message19:34
ceenealthough the next message they send is now clear19:34
ceenethis back and forth is visible on yappari.log19:35
Sicelo009Nokay. looking. problem is it's been some days19:36
Sicelo009Nonly now my other sister tells me of this problem19:36
Sicelo009N2nd sister using Android19:36
ceeneyou can grep for that19:38
ceeneand check if it matches what she tells you, by looking at the context around, maybe19:38
ceenewhen yappari receives a ciphered message it doesn't send receipt or read acknowledges19:39
ceeneso at least i'm sure she won't see green nor blue ticks19:39
Sicelo009Nyes, she does say there aren't ticks19:40
Sicelo009Nas far as you know, can a user 'control' their side not to send me encrypted messages?19:40
ceenenot that i know19:40
ceenebut i believe that at least after the first non clear message19:41
ceenethe following ones should arrive you19:41
ceenesince she can see that you haven't received it, she can send it again19:41
Sicelo009Nomg, i see lots of these encrypted things in log, lol19:41
ceeneit's tiresome, because her client will forget from time to time that you don't want encrypted messages19:41
ceeneso it will happen frequently :/19:41
*** arossdotme has joined #maemo19:41
Sicelo009Nthanks for efforts thus far :)19:42
ceenei'm afraid they won't be enough19:43
ceenei'm trying to implement textsecure19:43
*** krnlyng has joined #maemo19:43
ceenebut it's quite a hellish thing to do, even more so when i don't really know c++19:43
Sicelo009Ni can imagine19:43
*** msava has quit IRC19:44
*** heroux has joined #maemo19:47
*** msava has joined #maemo19:47
*** arossdotme-planb has joined #maemo19:53
Sicelo009Nbtw, noticed Tizen has Whatsapp :)19:53
*** xelo has quit IRC19:53
Sicelo009Nofficial, i will assume19:53
bencohceene: let's just rewrite everything in C/gtk :*19:55
Sicelo009Nceene: do you think/know the supported S40 phones are able to read/send encrypted messages?19:55
ceenecan i download that?19:55
ceeneSicelo009N: i'm looking right now at s40 version and yes, they implement axolotl19:56
*** arossdotme has quit IRC19:56
ceenebencoh: i'd have to learn gtk then...19:56
Sicelo009Ndownload what? Tizen's Whatsapp? No idea. I saw it on tmo thread regarding update of Z1 or some such...19:56
*** xelo has joined #maemo19:56
ceeneyes, i'd like to see it19:57
bencohceene: :]19:57
bencohceene: looks like you're as fond of GUI development as I am :]19:57
ceenebencoh: i've never coded any gui for anything19:58
ceenei'm just not able to19:58
ceenewell, i did something on visual basic 4.0 :)19:58
ceenei was a hax0r back then, when i was 1419:58
*** Vajb has joined #maemo19:59
ceeneif i had to write yappari's ui myself, we wouldn't have any at all19:59
*** arossdotme has joined #maemo19:59
bencohceene: sounds like me :D19:59
Sicelo009NTizen -->
ceenei can't get a link to download it20:01
*** arossdotme-planb has quit IRC20:01
Sicelo009Ni doubt there is. likely bundled. :(20:02
ceenei'm not sure if it's native client or if it's android's20:06
ceenethey have one thing called ACL that allows you to run android apps20:06
Sicelo009Ncould be20:06
ceeneACLTM for Linux20:07
ceenethat'll allow us to run whatever20:08
ceeneso we have to go and pay these people a heck of money, probably20:08
*** arossdotme has quit IRC20:09
ceeneandroid's whatsapp is a java written gui + libwhatsapp.so20:09
ceenei guess that all the interesting bits are there20:09
ceenesomeone could probably use libwhatsapp.so20:10
ceeneand forget about all the protocol intrincacies20:10
ceenerunning, of course, closed source code20:10
Sicelo009Nwould that .so be useable somehow on N900?20:11
ceenei think so20:12
DocScrutinizer05or someone could just build a gateway to that whatsapp universe, for more sane protocols like IRC or XMPP20:12
ceenealthough i haven't studied it much20:12
ceeneDocScrutinizer05: the thing is that whatsapp is indeed xmpp, pretty much20:12
ceenereplacing all the xml strings with binary codes, just a dictionary translation20:12
DocScrutinizer05well, a gateway would be highly useful20:13
ceenei find however all xml-based protocol to be insufferable20:13
ceenei think there exists one20:13
ceenebased on libpurple20:13
*** msava has quit IRC20:13
ceenethat attachs to a bitlbee instance20:13
ceeneso you can bridge irc<->whatsapp20:13
DocScrutinizer05I thought of a webservice20:13
ceeneif even thought of running an android emulator20:14
ceeneand reading whatsapp conversations database20:14
ceenewhich is just sqlite20:14
DocScrutinizer05such a webservice could even provide connection to plain SMS20:14
DocScrutinizer05outboubd would cost money of course20:15
*** msava has joined #maemo20:16
ceenesomething like that, using real wa as the backend would be useful basically forever20:18
ceenewithout much burden20:18
ceenei could even charge people for that :P20:18
ceenei've managed to get a tizen image20:19
ceeneit's basically lots of partition images zipped into one file20:19
ceeneand i've mounted a couple of them20:19
ceenethat thing runs systemd20:19
ceeneyes, i could also become an evil master overlord20:20
*** arossdotme has joined #maemo20:21
ceenebin/TizenInstallerStub: ELF 32-bit LSB shared object, ARM, EABI5 version 1 (SYSV), dynamically linked, interpreter /lib/, BuildID[sha1]=cc65549772dd68d7523bd8e0c0f2eeba5c7e4410, not stripped20:22
ceenethis thing seems to be the one that installs whatsapp on tizen20:22
DocScrutinizer05>>well, that's where I got stuck<< ;-P20:25
*** vectis3 has quit IRC20:26
ceenedon't know20:27
ceenereally, maybe the easiest thing would be that in fact20:27
ceenerunning a bunch of android emulators20:27
ceeneor whatever20:27
ceeneor android for x86, don't know if that exists and if whatsapp runs on that20:28
bencohceene: there is a purple-like bitlbee20:29
*** heroux has quit IRC20:29
bencoherr, strike that. there is a bitlbee-like purple-based gateway20:30
bencoh(there's also a bitlbee purple plugin, but ...)20:30
ceenethe best wa implementation right now is yowsup20:30
ceeneit implemented textsecure even before whatsapp20:30
ceenei don't know, this whatsapp business bores me a little more each day20:32
*** heroux has joined #maemo20:32
ceenei understand why scorpius got fed up with it20:32
bencohceene: do you use it?20:32
ceeneyes, but not very much, to be honest20:33
Sicelo009Nwhat makes yowsup hard to port for N900?20:33
ceenei don't speak with anyone directly, but i'm on three groups20:33
ceeneand my gf is in all of them too20:33
*** xelo has quit IRC20:34
ceeneSicelo009N: nothing at all, it's written on python, so i think it would just work TM20:34
ceenebut there's no gui for it20:34
ceeneit'll be slow and power hungry20:34
ceenebut it'll work20:34
ceenemaybe i can use that server side20:34
bencohI'm not sure you want to go that way20:34
ceenenot at all20:35
ceenepython2.7 doesn't know how to do a non busy wait20:35
ceeneeven sleep(30) uses power20:35
Sicelo009Nthen bencoh could port that python to c for us :)20:35
ceenei was in conversations with the offlineimap folks and the conclusion was that20:35
ceenedunno, i'm starting to think seriously about yowsup server side20:36
tmceene: yes - we can!20:36
bencoh(talking about offlineimap, you should try mbsync)20:36
ceenetm: lol !20:36
ceenebencoh: in the end i built an exchange server on a linux server of mine20:36
ceeneso my server gets mail, passes it through procmail so i don't get spam on the phone20:37
ceenewhich in the end is more battery, less data comsumption20:37
bencohSicelo009N: no way I'd read python code and parse/pack some binary-encoded-xml in C :p20:37
ceeneand exchange on n900 supports server push20:37
ceeneso i get mail instantly20:37
bencohwhy don't you use modest+dovecot?20:37
*** heroux has quit IRC20:38
bencoh(or any other imap server on your linux box)20:38
DocScrutinizer05  is actually tempting to like whatsapp, but then I heard it got raped by suckerberg20:38
ceenebecause modest by itself doesn't support push email20:38
ceeneIDLE command on IMAP servers20:38
bencohceene: it has no support for IDLE?20:38
ceenebut exchange does20:38
bencohI wonder how hard it would be to add it to tinymail20:39
ceeneso less battery and faster email20:39
ceenei don't like modest either20:39
ceenebut that's another issue20:39
ceenegotta go now, guys20:39
ceenesee ya!20:39
Sicelo009Nbencoh: i've asked many times ... probably lack of interest, or something like that? :)20:39
Sicelo009Nmyself too dumb for those kinds of things20:40
bencohSicelo009N: I don't use it (whatsapp), I don't like reading (or writing/debugging, for that matters) python code, and I dont fancy xml parsing ;)20:41
Sicelo009Ni meant tinymail ;)20:42
Sicelo009Nwhatsapp ... i can quit it anytime. just going to have to persuade my sisters to use something else (easier said than done though)20:43
bencoh(well I dont use it either, but that doesn't as bad as the other one ^^)20:43
*** arossdotme-planb has joined #maemo20:44
*** geaaru has joined #maemo20:45
Sicelo009Ni use it :)20:46
Sicelo009Nalmost every little N900 feature i use20:46
*** heroux has joined #maemo20:46
*** arossdotme has quit IRC20:46
Sicelo009Nstock applications mostly20:48
Sicelo009Nqalendar is the one main replacment i'm using. omp i use, but do use stock one as well. in addition to modest i use mutt for my gmail. everything else .. stock20:52
*** xelo has joined #maemo21:08
*** Sicelo009N has quit IRC21:09
*** kerio has quit IRC21:23
*** kerio has joined #maemo21:24
*** M4rtinK has joined #maemo21:30
*** Vajb has quit IRC21:32
*** arossdotme has joined #maemo21:37
*** Vajb has joined #maemo21:38
*** arossdotme-planb has quit IRC21:40
*** hashcore has quit IRC21:42
Vajbok value 0x28 turned surprisingly when i hit over ten cycles.22:03
Vajbnow it is 0x0b which i believe means 1122:04
Vajbin /sys/class/power_supply/bq27200-0/registers22:05
*** hashcore has joined #maemo22:07
*** luf_ has joined #maemo22:08
*** arossdotme-planb has joined #maemo22:09
Siceloof course, hehe22:11
*** arossdotme has quit IRC22:11
Vajbwell maybe for u, but for me it was surprise :p22:11
Sicelo0x means Hexadecimal22:12
Vajbso 0x0j would be 20?22:12
*** vectis3 has joined #maemo22:12
Sicelono. no such :)22:12
Sicelostops at F22:12
SiceloF = 1522:13
Vajbnot 16?22:13
Maxdamantus16 is 0x1022:13
Vajbah right22:13
Sicelomax beat me to it :)22:13
Vajbforgot 10 from my count22:13
Vajbalso that comic doesn't make sense to me :p22:15
Maxdamantus10_2 = 222:16
Maxdamantus10_8 = 822:16
Maxdamantus10_16 = 1622:16
Maxdamantusgenerally, 10_x = x22:16
*** florian has joined #maemo22:16
Maxdamantus(where `a_b` means the sequence `a` interpreted in base `b` (where b is described in base 10))22:17
Siceloi think _ in Maxdamantus' convention means base22:17
Vajbbut let me be clear 0x10 is 16, so is 0x19 25?22:17
Sicelopatent that Maxdamantus! before someone steals your idea :p22:17
MaxdamantusThat's already fairly typical notation.22:18
Maxdamantususing a subscript to denote the base.22:18
Siceloyes, that one i knew, but not _ :)22:18
Maxdamantus_ is also typically used to denote a subscript.22:18
Sicelookay. first time i see it myself22:18
MaxdamantusIt might come from TeX, though in TeX you'd have to write `10_{16}` for the last one, otherwise the `6` won't come out in subscript form.22:20
Sicelookay. never learned TeX/LaTeX .. still wish to22:21
MaxdamantusIt's a horrible system. You shouldn't wish to learn it.22:22
MaxdamantusAlthough I'm not aware of any practical alternatives.22:22
bencohhaha, usually you dont really want to but have to, because it's the only "sane" solution22:22
Vajbokies. Now i know what value to wait to know when to recalibrate battery. That is 0x1e :)22:23
bencohMaxdamantus: it's turing-complete! :]22:23
Vajbequals 30 cycles =)22:23
bencohVajb: why dont you use i2cget?22:23
MaxdamantusI tried writing a parser combinator library in it once.22:23
bencohplaying with the bq27x registers is fun, but...22:23
Maxdamantusbut it's just too confusing a system.22:24
Vajbwell i have bnf and i like it. It just lags that one feature i wish to have22:24
Sicelobencoh: i think he has replacement bme22:24
DocScrutinizer05>> numbers take the form [base#]n, where the optional base is a decimal number between 2 and 64 representing the arithmetic base, and n is a number in that base.  If base# is omitted, then base 10 is used<<  I suggest ********#077 however, to avoid the base base problem. At least for Seti and for gold plates on voyager22:24
bencohMaxdamantus: I know a small non-profit that used to have its subscriber information system written in... tex :]22:24
MaxdamantusThere aren't really any clear semantics. It's just this massive stack of macro libraries that determine how your program reduces.22:24
Vajbbencoh ^22:24
SiceloVajb: 0x1e in register 0x28? why?22:25
bencohSicelo: you can still use it :)22:25
Siceloyes. not recommended though :)22:25
Siceloand you've to add some other flag, was it -y or some such.. forgot22:26
bencohsomeone (pali iirc) went through the different i2cget calls in the bq27*.sh script and it was safe22:26
bencohyeah, -f or something22:26
VajbSicelo: well to know to let my device turn off by low power to get it calibrated again22:26
DocScrutinizer05Vajb: bq27200 will _tell_ you when CI=122:26
Maxdamantuscc bq27200.c -o bq2720022:26
Maxdamantusthen you can ./bq2720022:26
luf_DocScrutinizer05: How is it going with neo900?22:27
DocScrutinizer05we're stumbling from one obstacle into next. otherwise fine22:27
Vajbhmm DocScrutinizer05 maybe i've missed that value change, but so far i have noticed the need for calibration when battery measurements go back to stock values and bnf stops working22:28
luf_That's pitty.22:28
DocScrutinizer05bencoh: well, actually I wrote the script and I checked the i2cgets are safe ;-)22:28
bencohah :)22:29
bencoh(are we talking about the same script?)22:29
Maxdamantusi2cget doesn't usually seem to work while you have the bq27200 module loaded.22:29
Maxdamantus(on those registers)22:29
*** LauRoman has quit IRC22:29
Siceloreplacement bme won't allow the 'regular' i2cget read22:29
Siceloyes, Maxdamantus +`22:29
DocScrutinizer05maybe you talk about ShadowJK's version which is derived from mine22:29
bencohDocScrutinizer05: indeed22:29
MaxdamantusYou can just use my C program, which is derived from ShadowJK's script but doesn't take a second to run.22:30
Siceloeven yours doc. won't run without modification.22:30
Sicelojust tried it now :)22:30
SiceloError: could not sett address to 0x55: Device or resource busy22:30
DocScrutinizer05yes, that's because of bq27200.ko occupies the chip22:31
Vajbyes bq refuses to work with bme-replacement22:31
Sicelo22:26 < bencoh> yeah, -f or something22:31
Sicelothat's what's needed for that one22:31
Vajband after bme-replacement goes to "not calibrated" state bnf stops working too22:31
DocScrutinizer05and the dang module exports gibberish to /sys/*/registers22:31
Vajbbut if i unload some module (maybe rx51_battery) bq starts to work22:32
Sicelowhile everyone's here .. my battery is now never really 100% full. too old i know. or might something else be 'bad' software-wise?22:33
DocScrutinizer05I felt like "fsck it" when realizing that it needs a whole parser to make sense from /sys/*/registers22:33
*** LauRoman|Phone has joined #maemo22:33
MaxdamantusLike that.22:33
Vajbbut today i learnt somethign valuable of hexadecimals. Maybe from now on i give my phone number on hex when asked :)22:34
Siceloyou might be contacted by the NSA22:34
Vajbwho knows, maybe they offer me a job :p22:34
bencohI kinda remember someone having its phone number in hex form on his CV22:35
bencohpretty efficient against spam or unwanted job offers, while still reaching the relevant people22:35
SiceloVajb: or they remind you of your unpaid taxes  haha22:35
Vajbbut now im off to sleep, nite22:36
*** arossdotme has joined #maemo22:56
*** arossdotme-planb has quit IRC22:59
*** arossdotme-planb has joined #maemo23:05
*** qwazix has quit IRC23:06
*** qwazix has joined #maemo23:07
*** xelo has quit IRC23:07
*** arossdotme has quit IRC23:09
*** kerio has quit IRC23:10
*** cpt_nemo has quit IRC23:10
*** parazyd has quit IRC23:10
*** parazyd has joined #maemo23:10
*** kerio has joined #maemo23:13
*** eijk_ has joined #maemo23:13
*** arossdotme-planb has quit IRC23:14
*** futpib has quit IRC23:15
*** vectis3 has quit IRC23:20
*** xes has quit IRC23:32
*** arossdotme-planb has joined #maemo23:35
*** vectis3 has joined #maemo23:35
*** arossdotme-planb has quit IRC23:46

Generated by 2.15.1 by Marius Gedminas - find it at!