**** BEGIN LOGGING AT Mon Jan 11 02:59:58 2016 Jan 11 03:19:24 ((vodafone says...)) I witnessed a 1 week complete breakdown of the 2G functionality of the cell servicing the area where I live. O2 never admitted any problems Jan 11 03:23:31 anyway a complete flashing can't hurt, particularly after using fapman and doing so since years without flashing. Jan 11 04:55:05 I'm pretty sure I did apt-get upgrade with the testing repo in there Jan 11 04:55:08 lol Jan 11 06:59:35 fishbulb: check fmms.log and see what error is reported when its trying to fetch the mms Jan 11 07:02:42 and as DocScrutinizer05 said, theres about a million different reasons it can randomly fail, most of them on the provider side Jan 11 07:03:04 frals: nslookup cannot resolve the host needed to fetch the content Jan 11 07:03:30 ah, right Jan 11 07:04:10 try manually connecting to their MMS apn and resolve the host, otherwise the operator is most likely having issues with their service Jan 11 07:04:37 frals: "manually connecting" - how? Jan 11 07:05:14 easiest is probably modifying the "internet apn" through internet conenction settings to use the same settings as the MMS APN and just connect thru the normal ui Jan 11 07:05:26 but its been a couple of years since i last used a n900 :) Jan 11 07:06:03 hmm, yeah Jan 11 07:06:16 fishbulb: ^^^ Jan 11 07:13:22 just scrolled waaaay back, make sure the proxy is set as well Jan 11 07:13:49 he said the proxy is set-up Jan 11 07:14:05 been asked several times Jan 11 07:14:09 aight Jan 11 07:14:48 hi frals :-) Jan 11 07:15:00 thanks for helping out Jan 11 07:15:07 shame the log pastebin expired, but easiest way to see if its fMMS or operator is by connecting through the standard interface and see if it works there Jan 11 07:15:40 the magic scripts i hacked together to bring up 2 connections at the same time arent as good as they should be.. but they mostly work ;) Jan 11 07:15:44 no worries :) Jan 11 07:15:51 frals: is sending MMS also involving sending of a SMS, or is that SMS only on receiver side? Jan 11 07:16:19 sms only on receiver, http on sender Jan 11 07:16:25 ack, ta Jan 11 07:17:13 the "problem" is that sending is usually just posting to the mmsc, which some oeprators have accessible through all their APNs, while getting the mms *might* be on a different host thats only accessible through their mms apn Jan 11 07:17:25 so when fMMS says "MMS sent", this means it actually could upload the content via HTTP Jan 11 07:17:31 yeah Jan 11 07:18:05 and that the server on the other side didnt say http 4xx as reply iirc Jan 11 07:18:06 then it seems the APN is correct Jan 11 07:18:20 the TX APN yes Jan 11 07:18:20 if he is able to send Jan 11 07:18:28 it is only one APN Jan 11 07:19:02 mmsc might not be accessible if hes not connected to the "right" APN, which is used for retrieving Jan 11 07:19:28 I suspected wrong APN all the time actually Jan 11 07:19:32 frals: could we have different APNs for sending and receiving? Jan 11 07:19:49 freemangordon: see frals' post 3 up Jan 11 07:20:22 ah, you mean he is connected to the "internet" APN and is able to send, but not receive? that makes sense Jan 11 07:20:28 or are you asling if fMMS supports this per default Jan 11 07:20:50 yes, if this is supported Jan 11 07:21:03 not supported in fmms Jan 11 07:21:09 frals: is there a way to ask modem which is the APN currently used? Jan 11 07:21:24 uh icd had some stuff for that iirc Jan 11 07:21:34 AT cmd? Jan 11 07:21:44 gets bit tricky since fmms connects to 2 APNs in most cases... ) Jan 11 07:21:52 sorry, shouldn't pester you when google would do Jan 11 07:22:27 basically fmms has a setting to automatically connect you to the mms apn in the background without disrupting the internet connection.. might be that vodafone stopped supporting multiple contexts at the same time Jan 11 07:22:27 so the BB5 actually supports 2 concurrent APN? Jan 11 07:22:40 yeah, not sure what the upper limit was but 2 was fine Jan 11 07:22:49 4 iirc Jan 11 07:22:53 frals: he said he uses "havoc" mode Jan 11 07:22:54 upper limit Jan 11 07:23:06 ok, havoc tries to connect automagically in the background Jan 11 07:23:19 i hope i logged something if it failed to do that... Jan 11 07:23:26 hehehe Jan 11 07:23:42 but could be that he was connected to internet apn, and then the mms apn connect failed, and then that could explain what he was seeing Jan 11 07:23:55 changing connection mode to the safe one could be worth trying Jan 11 07:24:15 I suggest using pnatd to ask modem which contexts/APN (there was even another better name, no?) are active Jan 11 07:24:27 anyway have to run to work, please ping me when he shows up / if i can help with anything else :) Jan 11 07:24:50 * freemangordon has to run too Jan 11 07:25:08 another idea: aiui fMMS creates a second network device for the second context/APN. So we could use wireshark to see what's going on there Jan 11 07:25:33 frals: freemangordon: have a nice day both of you guys! Jan 11 07:27:09 (wireshark) actually a simple ifconfig before and after sending or trying to receive a MMS should already do, to see if any and how much traffic wnet via each of the two APN Jan 11 07:28:54 iirc (long ago that I last looked at MMS from some distance) you even need to tell provider that your device can handle and wants to use MMS, so the provider enables some services (possibly the special APN) for you Jan 11 07:30:56 unless fishbulb gets rootsh and network tools (dig/host, ifconfig, etc pp) installed, and learns how to connect to phone via ssh session from PC, we won't get anywhere debugging this Jan 11 07:32:27 first pretty simple test: spoil the standard APN for data connectivity, so we can be sure fMMS doesn't use "the internet" to send MMS Jan 11 07:33:40 anyway I have to run as well. Need to sort my 2015 taxes Jan 11 07:33:48 bbl Jan 11 07:35:31 heck! when you google for 3GPP AT commands, 2nd hit is "radekp" :-) Jan 11 07:35:57 http://radekp.github.io/qtmoko/api/atcommands.html Jan 11 07:36:48 HAH! "PDP context" was the name Jan 11 07:41:09 ok, BB5 (aka N900) supports THREE PDP contexts: http://paste.opensuse.org/30269195 Jan 11 07:41:38 frals: does fmms come with apn db or it fetches it from somewhere? Jan 11 07:42:19 http://ukblog.im-mobility.com/public/support/Nokia/Nokia_CS-10_AT_Command_Set.pdf search: "AT+CGDCONT" Jan 11 07:42:40 or jump to page 180 Jan 11 07:42:45 o/ l8rs Jan 11 07:45:47 MEH!!! Jan 11 07:45:49 AT+CGPADDR=1 Jan 11 07:45:51 ERROR Jan 11 07:48:12 KotCzarny: the device ships with mms apn db (since pr1.1 or 1.2) :) Jan 11 07:49:13 duh, what? Jan 11 07:49:31 sorry i'm missing context Jan 11 07:49:42 * DocScrutinizer05 opens chanlog Jan 11 07:54:21 frals, that's what i thought, because n900 didnt support mms, and settings were magically filled :) Jan 11 08:02:47 mms apn db? really now? Jan 11 08:03:07 heck that thing must be out of date now Jan 11 08:24:58 DocScrutinizer05: yeah, if fishbulb is comfy in the terminal it should be pretty easy to figure out where things go wrong Jan 11 08:25:22 yeah, probably a tad out of date by now.. :) Jan 11 08:26:00 otoh from what i heard back while working on fmms operators rarely changed those settings Jan 11 08:26:38 someone could always write a small script to parse the OTA provisioning messages the operator sends out, but i didnt want to even look at that after reading the mms "standard" :D Jan 11 08:26:44 but it can happen Jan 11 08:28:52 yeah Jan 11 08:29:19 biggest issue is usually new mvnos Jan 11 08:42:10 hey did I mention I can send MMS to myself Jan 11 08:44:28 I think you did Jan 11 08:45:24 yappari seems to work Jan 11 08:45:30 and be available on most phones Jan 11 09:01:11 fishbulb: can you post the logfile to pastebin so i can have a look? Jan 11 09:01:21 include sending to yourself + failing to retrieve one if possible Jan 11 09:02:42 hi I'm here for a few minutes Jan 11 09:03:15 I'll post the one log file I have and make another one too Jan 11 09:06:10 fishbulb: yappari works indeed, but do not create any group or whatsapp will ban you Jan 11 09:06:17 that's something i'd really like to fix, but i don't know how Jan 11 09:06:57 ceene: did you try dumping traffic and compare it to what does an android phone to create groups? Jan 11 09:07:01 ok. I just need one way of sending pictures to normal people's phones Jan 11 09:07:35 bencoh: it's easier than that, we have decompiled code of official whatsapp Jan 11 09:07:46 and i can't for the life of me get which is the difference between theirs and mine Jan 11 09:07:54 ceene: oh and, maybe you should disable the "create group" function meanwhile Jan 11 09:08:04 that's true Jan 11 09:08:13 i placed a big red warning explaining that Jan 11 09:08:20 but still a couple people got banned Jan 11 09:08:21 ceene, then packet dump Jan 11 09:08:25 could be easier Jan 11 09:08:25 ceene: this is why dumping traffic might be a good idea :) Jan 11 09:08:34 traffic is encrypted Jan 11 09:08:35 if you cant work it out from the decompiled source Jan 11 09:08:38 and all that Jan 11 09:08:53 xmpp protocol is first binary encoded Jan 11 09:08:57 and then it goes through ssl Jan 11 09:09:03 so it's quite a headache Jan 11 09:09:11 TPROXY + ssltunnel (for instance) Jan 11 09:09:38 but yeah, it's a pain Jan 11 09:09:59 http://pastebin.com/KVTqCMFV Jan 11 09:10:37 this is me trying to send one and trying to open one Jan 11 09:10:46 I think I try to open one first Jan 11 09:10:53 and then I send one maybe? Jan 11 09:13:45 hmm Jan 11 09:17:00 im thinking it doesnt work because of the failure to resolve the mmsc host to ip Jan 11 09:17:15 and the fmms_magic script doesnt early out there, it pretends everything is fine even though it goes bonkers Jan 11 09:17:43 can you try changing connection mode to the safe one Jan 11 09:18:29 "polite" or "rude" Jan 11 09:18:52 try "rude" first, it'll bring down the current connection though Jan 11 09:20:33 or edit your normal 3G internet connection to use the MMS settings and see if you can receive mms while using that Jan 11 09:27:34 frals: what those rude/polite modes are for? Jan 11 09:32:32 * Polite (Default behavior <0.6): Only switch to MMS connection if no connection is active, otherwise do nothing (pretty much) Jan 11 09:32:35 * Rude: Force a disconnect of the active connection regardless of what it is, connect to MMS APN and then reconnect to the old one. Jan 11 09:32:38 * Havoc: Try to bring up a connection to the MMS APN in parallell with whatever is active, this is what most other phones do. Note that to get this functionality on the N900 there are a few ugly hacks active which might break stuff, but it *should* work Jan 11 09:32:42 http://talk.maemo.org/showthread.php?t=39737 ;) Jan 11 09:33:34 fishbulb: so, change 3g internet connection settings to look like the mms one, set connection mode in fmms to polite, manually ocnnect to your 3g connection and then try to fetch the mms Jan 11 09:34:19 frals: thx Jan 11 09:34:19 oh some info Jan 11 09:34:29 I don't really have the internet going all the time Jan 11 09:35:59 frals: may i propose names: 'lazy connect', 'force mms apn' and 'parallel' ? Jan 11 09:36:23 it's never going actually. Jan 11 09:36:35 there's no autodisconnect stuff Jan 11 09:36:39 whoa, bowie is dead? Jan 11 09:36:44 KotCzarny: needed one word to fit in the UI ;) Jan 11 09:37:34 :) Jan 11 09:55:40 hey Jan 11 09:55:50 sorry I'm not gonna be much use right now Jan 11 09:57:49 it's like 9pm and there's stuff happening Jan 11 11:36:20 hi Jan 11 11:41:58 hi Jan 11 11:42:05 jonwil: i don't know if you read me last night Jan 11 11:42:10 but i've hacked something onto qt4 Jan 11 11:42:18 and i've got it to connect to tlsv1.2 servers :) Jan 11 11:42:21 yeah I saw you did something with QT4 and newer openssl Jan 11 11:42:23 thats a good start Jan 11 11:42:24 yeah! :) Jan 11 11:42:47 something nice about qt Jan 11 11:43:02 is that it doesn't link against openssl, but uses dlopen Jan 11 11:43:11 and searches for the most up to date version Jan 11 11:43:18 nice one Jan 11 11:43:34 so the patch i made last night is only needed to support new protocols Jan 11 11:43:42 but at least, simply installing openssl1 will make qt use it Jan 11 12:05:12 there are quite a few changes on qt's network/ssl module between 4.7 and 5 Jan 11 12:05:25 so i din't know how feasible it is to backport all of that Jan 11 12:14:45 we probably dont need everything just to use the right TLSv*_method (or whatever the enum/functions is) anyway Jan 11 12:15:07 if that's the case, probably something very similar to what i've already done is enough Jan 11 12:15:19 I'd just patch 4.7 with the few lines need to enable tlsv>1.0 Jan 11 12:15:56 which should be the default protocol, if any? Jan 11 12:17:16 most probably yeah Jan 11 12:17:18 \/43 Jan 11 12:17:19 woops Jan 11 12:18:37 i can see you're pretty excited about this :P Jan 11 12:19:30 i'm thinking Jan 11 12:19:41 that a priority based system could work Jan 11 12:19:56 try first tlsv1_2, if can't connect, try again tlsv1_1 Jan 11 12:19:57 etc Jan 11 12:20:08 KISS, we're talking about ssl/tls ;p Jan 11 12:20:15 yeah, but Jan 11 12:20:18 if i change the enums Jan 11 12:20:25 we'd need to recompile every qt app out there Jan 11 12:20:30 and i don't want that Jan 11 12:20:38 i want all qt applications to use the most secure protocol by default Jan 11 12:21:01 even if they don't know nothing about tlsv1.2, which they don't because qt4.7 didn't support it Jan 11 12:21:02 current code goes through the protocols list in enum order? Jan 11 12:21:11 nope Jan 11 12:21:13 current one Jan 11 12:21:20 uses either user chosen protocol Jan 11 12:21:32 or a default one, which is case QSsl::AnyProtocol: Jan 11 12:21:32 default: Jan 11 12:21:32 ctx = q_SSL_CTX_new(client ? q_SSLv23_client_method() : q_SSLv23_server_method()); Jan 11 12:21:35 break; Jan 11 12:21:45 then you dont need to insert values in-between two enums, just add it at the end (?) Jan 11 12:22:10 but in that case i'll need to change the default to one of the new protocols Jan 11 12:22:18 otherwise, they wouldn't be usuable by anyone Jan 11 12:22:28 i can add the tlsv1.2 code, but no application will know to chose it Jan 11 12:22:42 and sslv3 is completely deprecated, so I guess we shouldn't even allow it Jan 11 12:22:49 just change the default: line Jan 11 12:22:58 and set what? Jan 11 12:23:01 tlsv1.2? Jan 11 12:23:01 sslv2 is deprecated as well Jan 11 12:23:06 lol Jan 11 12:23:08 ok Jan 11 12:23:17 so should i just set tlsv1.2 as the default Jan 11 12:23:23 and let applications choose something older if they want Jan 11 12:23:30 I think there is a generic tlsv1 call in openssl1.x Jan 11 12:23:42 there may be, let me grep for it Jan 11 12:24:45 TLS_method(), TLS_server_method(), TLS_client_method() Jan 11 12:24:46 A TLS/SSL connection established with these methods may understand the SSLv3, TLSv1, TLSv1.1 and TLSv1.2 protocols. Jan 11 12:25:03 but it alows sslv3 Jan 11 12:25:19 I think you can also pass it some flags to disable any protocol Jan 11 12:25:31 (same goes for ciphers) Jan 11 12:25:49 I dunno if we want to completely drop sslv3 from default Jan 11 12:26:03 maybe not Jan 11 12:26:09 what does qt5 do for default? Jan 11 12:26:40 i think... Jan 11 12:27:08 // SSLv2 and SSLv3 will be disabled by SSL options Jan 11 12:27:14 i don't know where it does that Jan 11 12:27:21 but that comment is under SecureProtocols Jan 11 12:27:33 so i think it accepts everything and then disables those Jan 11 12:28:11 In regards to what we enable/disable by default we should follow the practice of whatever others are doing (i.e. what QT5 does or what Firefox does or what Chrome does) Jan 11 12:28:27 qt5 doesn't disable it completely Jan 11 12:28:33 but seems to exclude it from the list of secure protocols Jan 11 12:28:34 Any site out there that actually needs SSLv3 and will fail with TLS is probably not a site worth caring about Jan 11 12:29:07 else if (protocol == QSsl::SecureProtocols) Jan 11 12:29:08 options = SSL_OP_ALL|SSL_OP_NO_SSLv2|SSL_OP_NO_SSLv3; Jan 11 12:29:15 but you are free to choose sslv3 if you want Jan 11 12:29:46 if i can invoke tls_method() and then disable sslv3, the i could set that as the default Jan 11 12:30:01 but let people choose sslv3 if they explicitely want to Jan 11 12:30:57 Both browsers here on my system disable SSLv3 completly and only enable TLS (1.0, 1.1 and 1.2) Jan 11 12:31:54 uhm Jan 11 12:32:03 that documentation i pasted before about TLS_method() seems to be wrong Jan 11 12:32:10 or non matching my openssl-1 version Jan 11 12:32:22 ceene: be careful, I'm sure TLS_client_method() is in openssl-0.9.x Jan 11 12:32:28 as there's no match for TLS_client_method Jan 11 12:32:33 yep, it's not there on 1 Jan 11 12:32:33 right ;) Jan 11 12:32:46 so... what's the alternative on openssl-1? Jan 11 12:32:51 internet explorer lets you enable SSLv2 and SSLv3 if you need to but they are off on my system Jan 11 12:33:05 and my other browser doesn't even have options for SSL Jan 11 12:33:06 ceene: I dont remember :) Jan 11 12:33:16 ok, don't worry Jan 11 12:33:47 Its great that we are updating OpenSSL and QT and stuff but we still need to figure out how to update nss in microb Jan 11 12:33:57 and I am still getting nowhere with updating the root CA store Jan 11 12:36:19 it seems tls_method will be back on openssl-1.1 Jan 11 12:36:55 or maybe not Jan 11 12:37:00 i don't understand openssl changelogs Jan 11 12:37:14 I think nobody understands openssl as a whole Jan 11 12:37:39 lol Jan 11 12:37:44 yeah, i think that's it Jan 11 12:37:47 it'll be on openssl-1.1 Jan 11 12:37:50 which is yet to be released Jan 11 12:39:09 ceene: SSLv23_client_method in openssl-1.0.x allows tlsv1 Jan 11 12:39:26 and you can later restrict to disable sslv2 (or v3 for that matters) using SSL_set_options Jan 11 12:39:41 but only v1.0 Jan 11 12:39:51 and in case you're wondering, yes, this API is plain shit. Jan 11 12:40:05 yep Jan 11 12:40:09 the names don't make any sense Jan 11 12:40:18 nothing makes any sense in fact Jan 11 12:40:31 i just want a secure_client() and a secure_server() Jan 11 12:40:35 :)) Jan 11 12:41:11 that's why it's called open-ssl instead of o-secure-sl Jan 11 12:41:19 because it's open, but being secure is something yet to proof Jan 11 12:42:02 anyway, unless you want to rewrite the qt/ssl glue to use libressl, we'll have to deal with it Jan 11 12:42:21 i don't think so Jan 11 12:42:41 ? Jan 11 12:43:23 that i'll rewrite that Jan 11 12:43:27 i'll just deal with it :P Jan 11 12:46:15 :) Jan 11 13:00:02 i gotta go now Jan 11 13:00:08 but i'm doing that we've talked about Jan 11 13:00:20 i hope this week i can post somewhere a couple of .deb so they can be tested Jan 11 13:00:29 cya! Jan 11 13:00:35 please rather send a couple of git links :) Jan 11 13:00:39 but thx :) Jan 11 14:42:34 hmm, something for Pali, a pity he's not around: Jan 11 14:42:54 >>Device units do exist on the file system, solely for ordering purposes. They are generated only for devices tagged “systemd” for udev to interpret in its rulesets. They have no options, and have their primary use to implement a so-called “device-based execution”, which is a conditional execution of units based on udev device availability – this is, in fact, done with the cooperation of udev itself which understands rules like Jan 11 14:42:55 SYSTEMD_WANTS=. << http://blog.darknedgy.net/technology/2015/10/11/0/ Jan 11 14:50:57 OOOUUUUUUUUuuuuuUUUU! >>Automount units then are a further variation, as they drive autofs, though do explicitly queue jobs. In practice, they’re used to extend regular mount units by having mount points be attached lazily, i.e. deferring only until they are accessed.<< Jan 11 15:00:33 I've read ~40% now (up to >>The transaction manager<<) and I feel a strong nausea Jan 11 15:01:14 not because of the article but because of what it describes Jan 11 20:34:49 Wizzup_: you have gentoo booting on n900 with mainline kernel, ain't? Jan 11 21:34:51 Pali: poweroff -f on device does nothing, is that expected? Jan 11 21:36:19 freemangordon: no idea... look into /etc/init.d/mini* scripts Jan 11 21:36:35 there are scripts which do shutdown/reboot Jan 12 02:03:18 The weirdest: when I opened lasagne (cl-launcher is neat, by the way), and kept it in light-on mode for a long time, the entire OS froze to the extent that camera-lights would not switch off (not even when camera lens is closed), keyboard backlight would not switch on (entirely untypical), GUI would not move (quite common), and power button cannot shutdown (haven't seen this one in a long time) Jan 12 02:04:04 Oh, and side-slider would not switch off the display to blackness (also entirely untypical). Had to take out the battery, in the end, to shut down. Jan 12 02:04:39 By the way, camera light switched itself off before I took out the battery, probably a built-in defence against burning out **** ENDING LOGGING AT Tue Jan 12 02:59:59 2016