**** BEGIN LOGGING AT Sun Feb 07 02:59:58 2016 Feb 07 04:05:39 http://talk.maemo.org/showthread.php?p=1498036#post1498036 Feb 07 08:20:02 bencoh: yep, and yowsup implements textsecure, which is really the only way to go now Feb 07 10:40:55 right Feb 07 10:41:01 so i have libaxolotl packaged for maemo Feb 07 10:41:19 still have to check which are the .h files needed for the -dev package, but that's okay Feb 07 10:41:30 i can start now more or less to implement textsecure on yappari Feb 07 10:56:20 wow Feb 07 10:57:12 most of the work is really from coderus Feb 07 10:57:22 i just backported it to qt4 and made the debian/ dir Feb 07 11:25:24 oh Feb 07 11:25:27 linking error :( Feb 07 11:25:33 /usr/lib/libaxolotl.so: undefined reference to `CiphertextMessage::CURRENT_VERSION' Feb 07 11:25:37 that's simply false Feb 07 11:31:38 i'm just about to replace that with a #define Feb 07 11:31:48 i can't see why it doesn't found that symbol Feb 07 11:35:51 in fact, there are other symbols defined in the very same place in the very same way that don't trigger that error Feb 07 11:36:04 different data type? Feb 07 11:36:18 though it looks like a define Feb 07 11:36:19 nope Feb 07 11:36:21 it's the same Feb 07 11:36:25 class CiphertextMessage Feb 07 11:36:25 { Feb 07 11:36:25 public: static const int UNSUPPORTED_VERSION = 1; static const int CURRENT_VERSION = 3; Feb 07 11:36:29 and there are a few more like that Feb 07 11:36:41 i mean, i could replace that by a define Feb 07 11:36:49 but... i'd like to know what's happenning Feb 07 11:37:02 try comparing to the working code? Feb 07 11:37:29 UNSUPPORTED_VERSION doesn't trigger that error Feb 07 11:37:44 nor the other consts Feb 07 11:37:53 there's not much code i can test Feb 07 11:38:20 try defining them like: public: static const int UNSUPPORTED_VERSION=1, CURRENT_VERSION=3; Feb 07 11:38:40 ok, it won't hurt Feb 07 11:39:11 but doesn't make much sense, does it? Feb 07 11:39:34 when in doubt, ask the wiser, if wiser is not available try black magic Feb 07 11:39:50 the same Feb 07 11:40:03 then its in the headers, but not in the .so Feb 07 11:40:16 check .so which symbols are defined Feb 07 11:40:54 dpkg-shlibdeps: warning: symbol _ZN17CiphertextMessage15CURRENT_VERSIONE used by debian/libaxolotl/usr/lib/libaxolotl.so.1.0.0 found in none of the libraries. Feb 07 11:41:00 that's after compiling the .so Feb 07 11:41:08 public: ? Feb 07 11:41:13 but i can't see why Feb 07 11:41:36 current_versionE ? Feb 07 11:41:38 i'll check the warnings, there are a few Feb 07 11:41:44 na, it's just noise of the output Feb 07 11:41:54 i've checked the source and there's not versionE anywhere Feb 07 11:42:00 I don't think a "public:" is for the whole line of code Feb 07 11:42:16 it should affect everything that is below the public: Feb 07 11:42:20 yup Feb 07 11:42:35 till private or something else changes types Feb 07 11:42:39 my paste before was fixed by irssi Feb 07 11:42:43 but it's public Feb 07 11:42:47 err hmm, that's not how I thought the syntax of c and c++ would work Feb 07 11:43:05 though I admit I have no clue Feb 07 11:43:08 ceene: nm libaxolotl.so.1.0.0|grep _VERSION Feb 07 11:44:19 anyway if "public:" was a switch that changes context for everything after, there must be a sort of "endpublic:" Feb 07 11:44:24 [sbox-FREMANTLE_ARMEL: /tmp/N900/libaxolotl] > nm libaxolotl.so.1.0.0|grep _VERSION U _ZN17CiphertextMessage15CURRENT_VERSIONE Feb 07 11:44:32 DocScrutinizer05: it'd be private: or whatever Feb 07 11:44:39 undefined Feb 07 11:44:42 but in its abscense it should just define verything as public Feb 07 11:45:01 check the lib compilation output Feb 07 11:45:30 or maybe its in some other sublib Feb 07 11:45:42 is it the only .so in that tree? Feb 07 11:46:39 yep, there's no other .so Feb 07 11:46:43 i'll be checking the warnings Feb 07 11:46:47 there must be something somewher Feb 07 11:46:59 i gotta go now, i'll tell you if/when i find something Feb 07 11:47:01 thanks, cya! Feb 07 11:47:22 bb Feb 07 11:56:03 I'd do a stupid thing and simply add a "public: " in front of the " static const int CURRENT_VERSION = 3;" Feb 07 11:58:16 By default things in C++ are private Feb 07 11:58:42 Wait no, that's not strictly true, it depends on whether its a struct or a class Feb 07 12:00:11 his problem is undefined symbol in LIBRARY Feb 07 12:00:13 not in his code Feb 07 12:05:14 http://www.cplusplus.com/doc/tutorial/classes/ Feb 07 12:05:50 it suggests that a access_specifier affects exactly ONE member Feb 07 12:06:24 like I would have thought the syntax of a proper language is defined Feb 07 12:07:19 wait, I should read to bottom Feb 07 12:08:44 heck no, ceene is right: >>...and two member functions with public access: the functions set_values and area,<< Feb 07 12:10:54 c++ is a mess Feb 07 12:12:55 compare >> int a, b; << to >> public: t x; t y; << and the different syntax Feb 07 12:14:50 in first the semicolon is delimiter for the type 'int' specifier. in second there is no delimiter for 'public:' Feb 07 12:15:39 * 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) Feb 07 12:16:09 If I can figure this out I can commit a new set of root certificates no problems :) Feb 07 12:16:16 * DocScrutinizer05 adds a >>goto public: << at random places Feb 07 12:17:10 jonwil: which line of code is unclear? Feb 07 12:18:07 and why is it a problem to start with? doesn't it do just what it's supposed to do? Feb 07 12:18:27 The issue is that /etc is a symlink in Scratchbox Feb 07 12:18:42 and that function is returning /blah/blah/blah/etc instead of /etc Feb 07 12:18:58 meaning the output of cmcli is not what we need it to be for the device Feb 07 12:19:00 well, that's exactly what it's supposed to do, no? Feb 07 12:19:56 ok, I think I understand it a bit more now Feb 07 12:20:15 It gets information about the passed in path (the lstat call) Feb 07 12:20:22 then it says "is this a symbolic link" Feb 07 12:21:48 actually 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 that Feb 07 12:22:56 no, the code definatly does more than that Feb 07 12:23:41 sure 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 front Feb 07 12:24:06 it's quite weird code, but maybe there's a reason for that Feb 07 12:24:35 there might be attack vectors otherwise Feb 07 12:25:36 or problems with process specific name spaces, or whatever Feb 07 12:26:39 alas there's little comment in source on why all that is needed Feb 07 12:27:36 that's why I asked what's the source of trouble, after all I think it worked so far, no? Feb 07 12:28:17 The issue is that /etc in scratchbox is a symlink. Feb 07 12:28:27 or do we run into a sleeping bug in that code, in a specific environment that hasn't been tested so far? Feb 07 12:29:03 for full path it should use realpath() Feb 07 12:29:11 not some cd&pwd Feb 07 12:29:25 well, maybe the code is exactly for handling stuff in an unexpected manner in situations like a scratchbox environment Feb 07 12:29:40 jonwil, you can try changing symlink into mount --bind, if its only for the building package means Feb 07 12:29:59 Because /etc is a symlink when cmcli puts path names for certificates into files under /etc/secure/s it stores the target of the symlink Feb 07 12:30:01 or just change the code in cmcli? Feb 07 12:30:03 not the actual link Feb 07 12:30:09 so e.g. a chrooted process would still use the system wide cert store? or whatever Feb 07 12:31:51 i don't really see why would anything other than tar, backup utilities, etc, would mind if some path is symbolic or not Feb 07 12:31:56 (hi again!) Feb 07 12:32:00 it seems the problem isn't the code but the semantics and rationale behind it Feb 07 12:32:53 ceene: same here Feb 07 12:33:21 ceene: there are other things as well, like bindmounts and chroot Feb 07 12:33:30 Basically 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 device Feb 07 12:34:21 i can't see why cmcli cares about the nature of the path Feb 07 12:34:26 might actually be a bug that processes symlinks in a way originally meant for e.g. bindmounts or chroots Feb 07 12:37:17 unless we find out about the rationale behind that code, we prolly never will know how it's supposed to work and what's wrong Feb 07 12:39:09 it might be everything from noob's cargo cult coding to a highly sophisticated (though maybe flawed) security feature to handle special attack vectors or environments Feb 07 12:40:13 knowing the origins of this code might help a lot. e.g. if it originally been meant for android or docker or whatever Feb 07 12:42:39 can't you do a 'git blame' or whatever, to see who and when and why introduced that code snippet? Feb 07 12:42:52 The project is clearly 100% Nokia Feb 07 12:42:58 ouch Feb 07 12:43:20 maemo-security-certman was 100% the work of Nokia Feb 07 12:43:55 so odds are it's actually already meant to handle scratchbox environment in a special manner Feb 07 12:44:21 possibly never tested though Feb 07 12:45:01 ".. err humm, let's use the host's certificate store when we run in scratchbox..." ? Feb 07 12:45:48 Its got nothing to do with where it gets the certificate store from and no it wont use the host certificate store on Scratchbox as such Feb 07 12:46:09 "... looks good..." Feb 07 12:46:11 The symlink in question has nothing to do with the certificate store and is a generic symlink for all of /etc Feb 07 12:46:39 jonwil, maybe you should provide certs as source and run cmcli during install/reinstall? Feb 07 12:46:45 so for whatever reason the people behind Scratchbox decided that /etc in a scratchbox environment would be a symlink elsewhere Feb 07 12:46:48 please don't focus on the symlink, it's evidently the pathological case that hasn't been thought of Feb 07 12:47:13 The 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) Feb 07 12:47:20 that code is meant to handle something else Feb 07 12:48:26 how about you simply comment out that code? Feb 07 12:49:19 on target platform it doesn't do anything useful anyway aiui, and in scratchbox it doesn't work Feb 07 12:50:21 hey, when you have the author, you could simply send a mail and ask for a hint what's the deal with absolute_path Feb 07 12:56:23 yeah I was about to do that Feb 07 12:59:44 however see the very next function in code, which is using absolute_path along with a few other tricks to get a process context Feb 07 13:00:21 also for processes handled by maemo-launcher Feb 07 13:01:04 in this context stuff like optification and the involved bindmounts come to mind Feb 07 13:01:46 mail sent :)' Feb 07 13:01:51 :-D Feb 07 13:02:45 I hope it's no ænokia.com mail addr, or you rather had to use aunt google to find a more recent addr to contact the author Feb 07 13:03:39 dang! Feb 07 13:03:53 ctrl-q Feb 07 13:03:57 ;-P Feb 07 13:04:30 while searching for æ Feb 07 13:04:40 i still can't see why this const is then undefined Feb 07 13:05:09 * DocScrutinizer05 neither Feb 07 13:05:21 i'm just about to replace it with a #define Feb 07 13:05:28 it's just a number Feb 07 13:05:37 and not used for anything too useful, in fact Feb 07 13:05:55 a define is compile time, a public: is exported to the ABI, no? Feb 07 13:06:15 yes, but given that it's just a const defined in a .h file that can be simply included Feb 07 13:06:19 it doesn't matter at all Feb 07 13:06:21 ceene, its undefined in the library object itself Feb 07 13:06:27 KotCzarny: but why? Feb 07 13:06:33 not if that thing is a .so or whatever Feb 07 13:06:36 i'm the one compilnig the library Feb 07 13:06:56 some compile error? Feb 07 13:06:58 nope Feb 07 13:07:03 i've checked warnings and nothing related to it Feb 07 13:07:25 optimization kicked it out? Feb 07 13:07:37 could be, i guess Feb 07 13:07:50 use it in code, for some nonsense Feb 07 13:08:20 http://stackoverflow.com/questions/5508182/static-const-int-causes-linking-error-undefined-reference Feb 07 13:08:21 The email address I found when I made contact with the author of maemo-security-certman came from github Feb 07 13:08:28 i think it can be that same question Feb 07 13:08:42 o.O Feb 07 13:09:10 c++ is fucking weird Feb 07 13:09:20 mildly put Feb 07 13:10:07 [Sun 07 February 2016] [13:10:54] c++ is a mess Feb 07 13:10:26 http://stackoverflow.com/questions/5391973/undefined-reference-to-static-const-int/5392134#5392134 Feb 07 13:10:28 jonwil: sounds good Feb 07 13:10:30 i believe this is the answer Feb 07 13:10:44 which of course is nuts Feb 07 13:10:50 but it's probably going to work, i guess Feb 07 13:11:13 yes Feb 07 13:11:24 so much for code reusability Feb 07 13:11:30 i have to type the exact same thing twice Feb 07 13:11:39 so i can make twice the mistakes i'd do otherwise Feb 07 13:11:52 hehe Feb 07 13:12:01 hail c++ Feb 07 13:12:13 use inline assembler ;-) Feb 07 13:12:42 at very least you know *exactly* what's going to happen then Feb 07 13:13:13 j/k Feb 07 13:13:39 there's someone right now, with an unwritable name, who is proud of himself for inventing c++ Feb 07 13:13:45 he sleeps well at nights Feb 07 13:15:06 prekeywhispermessage.o: In function `.LANCHOR0': Feb 07 13:15:06 prekeywhispermessage.cpp:(.rodata+0x40): multiple definition of `CiphertextMessage::UNSUPPORTED_VERSION' Feb 07 13:15:09 keyexchangemessage.o:(.rodata+0x0): first defined here Feb 07 13:15:10 i'm done with it Feb 07 13:15:13 that's it Feb 07 13:15:15 #defines it's going to be Feb 07 13:16:46 Bjarne Stroustrup ? Feb 07 13:16:53 that one! Feb 07 13:19:21 there Feb 07 13:19:21 done Feb 07 13:19:29 now it compiles, and yappari links against it without any problem Feb 07 13:19:36 * DocScrutinizer05 puts B. Stroustrup next to L. Poettering on his secrete list Feb 07 13:20:10 that's a list that secretes people you hate? Feb 07 13:20:45 >>links against it without any problem<< doesn't tell much yet, you might run into an assert at yappari runtime, like "version of blabla.so doesn't fit" Feb 07 13:21:24 ((list)) up to you to figure what that list is for ;-) Feb 07 13:22:06 DocScrutinizer05: i'm the one who is going to write yappari code to match my libaxolotl version Feb 07 13:22:09 so that won't happen Feb 07 13:22:16 :-) good Feb 07 13:22:21 maybe my libaxolotl will differ from mainstream Feb 07 13:22:29 but i don't care much Feb 07 13:22:36 I never heard of that strange critter ;-) Feb 07 13:22:46 moreso if this stupid thing is the only one that's blocking my development Feb 07 13:22:51 yeah, it's a funny name Feb 07 13:22:55 compile statically? :> Feb 07 13:22:57 it's the thing that implements textsecure ciphering Feb 07 13:23:13 axolotl, wasn't that an ugly amphibian? Feb 07 13:23:27 nah, i already have .debs for libaxolotl and libcurve25519 Feb 07 13:23:40 yeah, it's a salamander Feb 07 13:23:53 https://en.wikipedia.org/wiki/Axolotl Feb 07 13:23:57 or something like that Feb 07 13:24:06 i don't think it's ugly, just funny looking :) Feb 07 13:24:16 looks like a fish with legs Feb 07 13:25:09 >>As of 2010, wild axolotls were near extinction<< ;-P Feb 07 13:25:29 poor thing Feb 07 13:25:42 wonder what that implies for libaxolotl Feb 07 13:26:03 as long as wa uses it... it'll have a long live Feb 07 13:26:12 hehe Feb 07 13:26:31 not that i'm against crypto, but i find that whenever i have to use some crypto library they are all a mess Feb 07 13:26:34 so wa is the zoo for axolotls Feb 07 13:26:52 i guess it's because crypto is a hard subject in itself, so implementation are also hard problems Feb 07 13:27:03 but man, all this initialization, keys, prekeys, shared secrets... Feb 07 13:27:23 yeah, it's mindboggling Feb 07 13:31:18 https://en.wikipedia.org/wiki/Alice_and_Bob are the very simple looking basics Feb 07 13:32:03 https://en.wikipedia.org/wiki/Alice_and_Bob#Cast_of_characters already gets nasty Feb 07 13:33:57 i've read agatha christie's And Then There Were None, which had less people than that and much more gratifying to read and guess whodunnit Feb 07 13:36:21 that one was really quite enjoyable, actually :) Feb 07 13:36:24 https://en.wikipedia.org/wiki/Perfect_forward_secrecy is where we get to the more interesting things Feb 07 13:36:50 (and then there was none) Feb 07 13:36:56 were* Feb 07 13:44:56 PFS isn't really simple to understand or to implement Feb 07 13:45:17 axolotl is even listed in https://en.wikipedia.org/wiki/Perfect_forward_secrecy as one of the implementations Feb 07 14:19:00 Pali: any idea how to prevent screen going blank after trying to reboot or shutdown? Feb 07 14:19:25 no idea... Feb 07 14:32:14 freemangordon: hmm? that sounds a tad fuzzy Feb 07 14:32:35 am I missing some context? Feb 07 14:33:15 DocScrutinizer05: when you reboot N900, display goes black Feb 07 14:33:44 well, eventually the video buffer gets initialized, yes Feb 07 14:33:46 I want to prevent that, to be able to capture console aoutput Feb 07 14:34:26 ooh you meant after console already went active Feb 07 14:34:45 well, I guess commenting out acx565akm_disable(dssdev); in remove() will do the jib :) Feb 07 14:35:22 aaah, during shutdown Feb 07 14:35:27 hmm, no :( Feb 07 14:35:35 yes, during shutdown Feb 07 14:36:00 yes, what you said then. sounds reasonable Feb 07 14:36:16 DocScrutinizer05: is there a new draft hackerbus spec to look at? Feb 07 14:36:20 however you also need to prevent system shutdown then Feb 07 14:36:23 but doesn't work Feb 07 14:36:46 enyc: wpwrak hasn't provided a public version yet Feb 07 14:37:49 okies =) Feb 07 14:38:09 maybe prod him about it in neo chan Feb 07 14:38:39 ooh oops Feb 07 14:38:44 DocScrutinizer05: what should be done so I to have working serial output from N900? Feb 07 14:39:53 freemangordon: a good question. the classical approach would be to use serial (or other low complexity) interface to export all printout in realtime Feb 07 14:40:30 well, there is serial, I just don't have the way to interface it :) Feb 07 14:40:43 frewuently UART console via a special debug adapter is used for that Feb 07 14:40:50 is there such a way, some standard device to connect to omap serial outs Feb 07 14:41:04 "debug adapter"? Feb 07 14:41:36 the debug port below battery? Feb 07 14:41:39 "port" Feb 07 14:41:47 more pads than port, but ... Feb 07 14:41:48 for N900 yes Feb 07 14:42:13 it's 1V8 TTL UART Feb 07 14:42:23 TTL 1.6V? Feb 07 14:42:29 ops, 1.8 Feb 07 14:42:43 no way TTL :) Feb 07 14:42:44 does it go straight to omap? Feb 07 14:42:47 yes Feb 07 14:42:51 :/ Feb 07 14:42:52 well "TTL" Feb 07 14:43:07 sounds like instant fry the moment you do something wrong Feb 07 14:43:13 yes Feb 07 14:43:21 you mean 0V for 0 and 1.8V of 1? Feb 07 14:43:26 yes Feb 07 14:44:33 and what about the connector? can I use some standard plug? Feb 07 14:44:40 http://wiki.maemo.org/N900_Hardware_Hacking#Debug_ports Feb 07 14:44:47 yep, read those Feb 07 14:44:55 doesn;t help much :) Feb 07 14:45:06 tbh I'd suggest to solder some thin wires to the testbads Feb 07 14:45:27 won;t do that on my "new" device Feb 07 14:45:45 do it on the old then? Feb 07 14:45:58 "old" one is my daily device :) Feb 07 14:46:04 there are generic "phoenix" adaper "pliers" Feb 07 14:46:23 /dev/ttyS0 might be wrong on modern kernels Feb 07 14:47:11 (it's probably something like ttyO0) Feb 07 14:47:26 bencoh: srsly? I thought they abandon ttyO* again Feb 07 14:47:48 DocScrutinizer05: what I am supposed to search on ebay for? Feb 07 14:48:00 lemme try Feb 07 14:48:50 Defiant: ah? maybe then Feb 07 14:50:15 http://www.fonefunshop.co.uk/Unlocking/servicecableuniversal1.htm Feb 07 14:51:15 £55.99 :/ Feb 07 14:52:06 bencoh: those testpads are not designed for convenient low cost connectors to get attached Feb 07 14:52:23 I wonder how it works Feb 07 14:53:36 me too Feb 07 14:53:45 freemangordon: 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+/- lines Feb 07 14:53:48 I understand nothing from those pictures Feb 07 14:54:08 me neither Feb 07 14:54:25 especially the one with the stripped board Feb 07 14:54:37 it's a simple clip (acrylic) that presses some pogopins to the DUT Feb 07 14:54:40 how does it attach to it Feb 07 14:55:05 ah Feb 07 14:55:17 DocScrutinizer05: the problem is that the bug I chase needs usb gadget driver loaded Feb 07 14:55:21 the pogopins sit on a small adapter PCB 'plugged' into the tip of the clip Feb 07 14:55:42 ok, but how to power that assembly? Feb 07 14:56:16 I guess the crocodile clips are meant to contact to the battery blades Feb 07 14:56:28 hmm Feb 07 14:56:37 too complicated for my test Feb 07 14:56:42 *taste Feb 07 14:57:15 well, it gets more convenient and "simple" if you shell out way more bucks Feb 07 14:57:38 example? Feb 07 14:57:48 the Nokia test jig Feb 07 14:59:58 a possibly more appealing approch is to hack something like this: http://maemo.cloud-7.de/share-service/20130719_002.jpg Feb 07 15:00:44 http://maemo.cloud-7.de/share-service/20130719_003.jpg ... 009.jpg Feb 07 15:01:29 http://maemo.cloud-7.de/share-service/20130719_007.jpg Feb 07 15:02:07 :nod: Feb 07 15:02:11 what is this? Feb 07 15:02:40 I mean - someone made it or you haev it? Feb 07 15:03:21 sre made one too .. and gave step-by-step instructions for someone else to make for themselves Feb 07 15:03:43 I 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: http://www.phoneunlock.com/Nokia-Service-FBus-Cable-Rj45-for-JAF-MX-Box-Saras-Box-x-5-cables-SKU176972.html Feb 07 15:04:26 DocScrutinizer05: oh, so you have a way to catch serial output? Feb 07 15:04:56 * freemangordon gives DocScrutinizer05 zImage and modules and waits for the logs :) Feb 07 15:04:59 http://elektranox.org/n900/serial-adapter.html Feb 07 15:05:41 https://www.sstparts.com/product.php?aId=5072 Feb 07 15:06:10 no, that thing connects to BB5 F-Bus, not to the UART pads Feb 07 15:06:34 :( Feb 07 15:07:22 http://www.fonefunshop.co.uk/Unlocking/service_cables/box/atf-lightning-nokia-service-cables.htm Feb 07 15:07:48 you however can hack that thing moreately easily, you just need to 'move' the pogopins Feb 07 15:08:20 but I don;t have the serial converter chip as well Feb 07 15:08:37 and I feel too old for hacking such things :) Feb 07 15:09:35 so I'd rather buy something prebuild, or ask somebody on the channel to capture the logs for me Feb 07 15:10:07 ask sre then ;) Feb 07 15:10:13 or he's busy nowadays? Feb 07 15:10:35 I don;t see him around Feb 07 15:10:50 I suspect he might want to keep his adapter anyway :) Feb 07 15:11:04 lol, yes .. meant he could get him to test Feb 07 15:11:05 since he works on n900+kernel from time to time Feb 07 15:11:09 ah Feb 07 15:11:16 freemangordon: tried xmpp? he usually responds instantly Feb 07 15:11:54 Sicelo: never used xmpp, besides fbchat and google chat :) Feb 07 15:13:04 okay. email then? :p Feb 07 15:13:14 or i could let him know you need him? but then how does that help Feb 07 15:13:47 hacking 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 anyway Feb 07 15:13:50 well, you can ask him to join #maemo :) Feb 07 15:14:18 Sicelo: but yeah, maybe I should mail him Feb 07 15:14:21 Nokia data cable it's dubbed Feb 07 15:15:07 nokia data cable, the thing that existed back when 3310 was mainstream? Feb 07 15:15:14 yes Feb 07 15:15:16 or simply wait for Pali's best friend amongst the kernel maintainers to return from holiday :D Feb 07 15:15:54 or finally share your UPS address ;-) Feb 07 15:16:14 DocScrutinizer05: hmm? you have my address Feb 07 15:16:20 I do? Feb 07 15:16:24 yes Feb 07 15:16:46 never seen sre here Feb 07 15:17:19 then I probably could find it in my data desert if I already knew it ;-) Feb 07 15:17:44 DocScrutinizer05: I gave it to you to send me BB-XM iirc Feb 07 15:17:55 duh! Feb 07 15:18:01 and did I? Feb 07 15:18:04 no Feb 07 15:18:08 why? Feb 07 15:18:30 because we've solved the problem so there was no need, or somesuch Feb 07 15:18:37 aah Feb 07 15:18:59 yeah I remember now Feb 07 15:19:50 I connected the BB-xM to Europa instead ;-) Feb 07 15:20:18 :nod: Feb 07 15:20:20 :) Feb 07 15:21:04 DocScrutinizer05: 8.04.2015 22:18 Feb 07 15:21:10 ta Feb 07 15:21:16 will see what I can do Feb 07 15:21:32 seems you're not eager to do any soldering Feb 07 15:21:36 no Feb 07 15:21:42 I miss the tools needed Feb 07 15:21:56 ok, I think I got the tools Feb 07 15:22:08 I don't want to risk breaking a device Feb 07 15:22:33 meh, it takes breaking eggs to make an omelet Feb 07 15:23:05 I have only one old(> 20 years) 10MHz russian scope and 40W iron Feb 07 15:23:17 I see Feb 07 15:23:35 I think my scope and iron are up to the task Feb 07 15:23:43 see, if there is noone to do it, I will, but knowing there are guys better suited for the job, I prefer to not ;) Feb 07 15:23:55 np Feb 07 15:23:57 stay tuned Feb 07 15:24:00 ok Feb 07 15:24:41 * DocScrutinizer05 idly wonders how to *send* data from Rigol scope to RS232 ;-) Feb 07 15:25:26 Pali: i am thinking to enable 550MHz before sending the patch, unless you remember why it is disabled Feb 07 15:25:37 https://www.youtube.com/watch?v=wEmdHCXhrNk Feb 07 15:27:52 freemangordon: no idea why 550 is disabled Feb 07 15:28:44 ok, will enable it Feb 07 15:29:20 DocScrutinizer05: any idea? Feb 07 15:29:31 hmm? Feb 07 15:29:51 why could be 550MHz freq obsolate and disabled? Feb 07 15:29:52 ooh, maybe instable? Feb 07 15:29:53 for n900? Feb 07 15:30:19 naah, it is not that one Feb 07 15:30:25 550 is table Feb 07 15:30:26