**** BEGIN LOGGING AT Thu Apr 28 02:59:58 2016 Apr 28 03:07:33 wait! you say the "outer" structure ahd a *pointer* to the inner extended structure? then is _should_ be safe indeed Apr 28 03:08:41 I don't understand how the actual data structure type of any of both structures would have any impact on that Apr 28 03:09:52 the library hands you a pointer to struct_outer, which has a pointer aka memory address to struct_inner. When you extend struct_inner by appending data to the end, nothing bad should happen Apr 28 03:15:26 I thought you were referring to struct_outer actually embedding struct_inner verbatim. so addr(struct_outer) <= addr(struct_inner) < addr(end(struct_inner)) <= addr(end(struct_outer)) Apr 28 03:16:54 end(struct){return addr(struct) + len(struct)} Apr 28 03:17:40 actually Apr 28 03:17:55 end(struct){return (pointer) addr(struct) + len(struct)} Apr 28 03:26:28 luke-jr: mind to elaborate on the problem you ran into? Apr 28 03:27:50 DocScrutinizer05: in glibc or my lib's ABI? Apr 28 03:28:03 the example you gave above Apr 28 03:28:35 https://github.com/bitcoin/libblkmaker/blob/master/blktemplate.h Apr 28 03:28:39 I added to struct blktxn_t at the end Apr 28 03:29:49 and where did that break? Apr 28 03:30:32 when software using the library accessed blktemplate_t->txns[1] Apr 28 03:32:36 I see https://github.com/bitcoin/libblkmaker/blob/master/blktemplate.h#L96 Apr 28 03:32:46 I don't see where the problem happened Apr 28 03:33:48 txns[1] used to be txns+32, now it is txns+36 Apr 28 03:34:03 tbh I don't find txns[] in blktemplate_t Apr 28 03:35:13 ooha, for arry the compiler assumes it knows the array element size Apr 28 03:36:09 that's compiletime, indeed Apr 28 03:36:39 nasty Apr 28 03:37:52 so the problem is not the ABI change per se, neither the embedding superstructure, it's a pointer to a datatype that changes while the compiler generated code that assumes it's fixed Apr 28 03:39:03 IOW when you get a pointer to an array, your compiler has to rely on size of each array element didn't change since compile time Apr 28 03:40:08 that's implicit knowledge about datatype properies that actually shouldn't get entangled to ABI. Alas it obviously is Apr 28 03:40:14 *aiui* Apr 28 03:41:29 you probably should update http://www.tldp.org/HOWTO/Program-Library-HOWTO/shared-libraries.html :-) Apr 28 03:41:43 nice finding Apr 28 03:43:41 actually the problem is your data type of the array element changed Apr 28 03:46:05 while it's OK to chenge the data type of a single structure by simply appending data to the end of the structure, it's breaking your ABI when you use that changed dataype in any other data structure Apr 28 03:46:41 which you did in the array definition Apr 28 03:48:15 so yes, the "embedding outer structure", even hile using a pointer, used a pointer to a different datatype, and that broke the ABI Apr 28 03:48:23 while* Apr 28 03:49:02 nice catch nevertheless :-) Apr 28 03:54:10 tbh arrays in C always been quite messed up a implementation concept Apr 28 03:56:18 all the aray pointer arithmetics are prone to confuse the one using it Apr 28 03:56:51 way too much implied knowledge that alas isn't ubiquitous Apr 28 03:56:58 in compiler Apr 28 03:57:37 see the above obscure data type change of pointer to array Apr 28 04:01:26 did the ABI datatype definitions cover and catch this ("undefined or missing parameter" on binary startup), or did you simply run into errors from wrong index into array during runtime? Apr 28 04:02:32 the latter would be *very* unfortunate Apr 28 04:53:21 >> a recent Gentoo glibc update broke fork() << So, while generally installing new versions of libs should not damage anything, glibc was shown to be a risk once due to developers-testers making a mistake, as in, >> implicit knowledge about datatype properties that actually shouldn't get entangled to ABI. Alas it obviously is << Apr 28 04:55:24 And, if there are libraries of entirely different versions (like 2 and 3), they are not backwards-compatible, and they do not conflict, they simply co-exist. If the libraries are of same version, the new one will be "replacing" the old one, and it should be backwards-compatible, unless there is a mistake somewhere... Apr 28 07:22:47 sure Apr 28 07:27:18 what's sure? Apr 28 07:28:55 backlog ^ if there are libraries of entirely different versions (like 2 and 3), they are not backwards-compatible, and they do not conflict, they simply co-exist. If the libraries are of same version, the new one will be "replacing" the old one, and it should be backwards-compatible, unless there is a mistake somewhere... Apr 28 07:32:52 and even then you *could* make them coexist Apr 28 07:33:02 though it gets nasty then Apr 28 07:33:29 you need to resort to LD_PRELOAD or similar stuff Apr 28 07:34:01 to tag either the apps that need the new version or the apps that need the old version Apr 28 07:34:26 freemangordon: ༼ つ ◕_◕ ༽つ Give GCC Apr 28 07:34:47 o.O .oO(???) Apr 28 07:35:11 sure Apr 28 07:35:22 what's sure? Apr 28 07:35:28 ;-D Apr 28 07:35:41 :D Apr 28 07:37:24 anyway that array issue is a tricky one Apr 28 07:37:50 I wonder if ABI symbols catch it Apr 28 07:38:33 I guess they should Apr 28 07:40:31 related: http://ispras.linuxbase.org/index.php/ABI_compliance_checker Apr 28 07:41:57 http://abi-laboratory.pro/tracker/timeline/glibc/ Apr 28 07:51:24 krrhrrhrr http://blog.famillecollet.com/public/reports/mysql__-3.0.9-3.1.0.html#Type_Problems_High Apr 28 07:52:05 >>When I start maintaining this package (version 2.x), the library was even not versionned. So I have to work with developers to include this feature. Now I need to convince them to respect the pratices.<< Apr 28 07:53:32 so much about proving my claim above regarding >>it's amazing how little lib devels seem to understand about how libs actually are supposed to work<< Apr 28 07:54:53 why do you think docker is popular Apr 28 07:55:05 instead of actually figuring out what's needed and how's needed by your shitty webapp Apr 28 07:55:13 you can just ship the whole distro you've been developing on Apr 28 07:55:19 because some fools think it's a great idea to run each app in its own VM basically Apr 28 07:55:27 no, that's not it Apr 28 07:55:30 jails do that Apr 28 07:55:32 jails are fucking awesome Apr 28 07:55:51 sure Apr 28 07:56:03 docker is roughly jails on linux Apr 28 07:56:04 they should link sttaically instead Apr 28 07:56:09 statically Apr 28 07:56:11 but it's used by awful devs Apr 28 07:56:42 see what king of fools thinks about all that: Apr 28 07:56:50 ~systemd cabal Apr 28 07:56:51 systemd cabal: a bunch of people (Lennart Poettering, Kay Sievers, H. Hoyer, D. Mack, T. Gundersen, D. Herrmann) who want to turn linux into their wet dream perverted version of windows-me-too: http://0pointer.net/blog/revisiting-how-we-put-together-linux-systems.html -- Rumor has it that 2016 systemd will have replaced kernel. See ~nosystemd" Apr 28 07:57:11 i guess the approach is different Apr 28 07:57:25 with jails it's the sysadmin who's installing the app that decides how to contain it Apr 28 07:57:32 perverted dream with cabal? Apr 28 07:57:42 or windows? Apr 28 07:57:45 so you start from a thing that already works and has good documentation and well thought dependencies Apr 28 07:57:55 and you install that in a separate container for ease of mind, pretty much Apr 28 07:58:37 nowadays you have bullshit forum software whose only supported install method is to run a launcher on a docker *host* Apr 28 07:58:43 not even just a container Apr 28 07:59:24 honestly, docker... they can eat it Apr 28 08:00:06 if I wanted such stuff, I'd go android or windows-phone already Apr 28 08:00:52 *_* Apr 28 08:01:01 honestly, install freebsd Apr 28 08:01:16 it's incredibly refreshing to actually use a system that's well designed Apr 28 08:01:52 i'm tempted to Apr 28 08:02:26 alas, it kinda sucks as a desktop system Apr 28 08:02:36 pc-bsd is probably better for that Apr 28 08:03:00 should we expect to see maemo freebsd based? Apr 28 08:03:14 nah Apr 28 08:03:22 ofc not Apr 28 08:03:29 maemo is held together with the rust of backwards compatibility Apr 28 08:03:44 remove that, and it's a knockoff phone os with no community Apr 28 08:11:05 indeed Apr 28 08:15:06 actually maemo has _no_ inherent security features except the extras-testing scheme, compared to all other phone OS. It's only thanks to early testing and now "thanks to" no more new apps because of lack of economical perspective, that we don't see same amount of malware and privacy breach in maemo as we see e.g. in android Apr 28 08:16:43 for HARM Nokia tried to tackle that aspect with their store and aegis, which basically checked all apps for any malicious aspects before offering them to the user masses Apr 28 08:17:20 this is somewhat similar to what apple does, aiui Apr 28 08:18:43 android afaik has this "this app wants access to: your contacts, your bank account, your sex life..." and you have to nod off each to make the app run Apr 28 08:19:28 not to mention that any parvency of security is instantly gone when the typical M.O. is to google " apk" on google and click on the first link Apr 28 08:24:40 yeah Apr 28 08:26:17 android resp the user still would need to allow the app to access whatever the app wants to access, but it is plain to see that joe average-androiduser is nodding off everything as long as the app is for free Apr 28 08:27:26 "send SMS? duh why not?" - meh actually they don't even read that far, only see "requester? where's the OK button?" Apr 28 08:31:12 ~wtf m.o. Apr 28 08:31:14 usage: wtf . Apr 28 08:31:25 err yes? Apr 28 08:31:31 ~wtf mo Apr 28 08:31:32 Gee... I don't know what mo means... Apr 28 08:36:23 more Apr 28 08:36:28 oh Apr 28 08:36:39 modus operandi Apr 28 08:36:50 Modus operandi is a Latin phrase, approximately translated as "method of operation".[1] The term is used to describe someone's habits of working, particularly in the context of business or criminal investigations. In English, it is often shortened to M.O. Apr 28 08:37:03 mo money <- not modus operandi Apr 28 08:37:09 can we order pyra yet, dangit Apr 28 08:37:45 nope Apr 28 08:56:53 Anyone know if the last 2.3.7 nitdroid ended up ever getting calls working? Or did it all stop without being fixed? Apr 28 09:00:12 nitdroid is dead I'd think Apr 28 09:01:31 Yeah, all my googling says they went dead at around 4.0.4? Just wasn't sure if they ever got voice working in 2.3.7 (gingerbread) before they started working on the next one Apr 28 09:02:03 Seems odd to me to spend so much time porting a phone OS, and not focus on getting the actual phone function to be the first thing to get working :P Apr 28 09:02:27 is remixos still a thing Apr 28 09:05:22 Seems pretty active at xda Apr 28 09:06:29 The way my N900 is going I feel like I'm just going to end up having it around as an mp3 player ): Generally I would be fine using it, but it really doesn't like handling websites these days Apr 28 09:09:11 azkay: get a vps with vnc / rdp Apr 28 09:09:15 and run browser there ;) Apr 28 09:09:53 I've got a vps, but it's in another country and I only have 1.5gb/mo on my phone :P Apr 28 09:10:46 Using reddits mobile site, after waiting literally 5 minutes to load the page (if it doesn't stop half way), I can only open maybe 3 links before the browser freezes up lol Apr 28 09:14:11 use /usr/bin/links then Apr 28 09:14:18 should render a lot faster :D Apr 28 09:19:41 lynx? Apr 28 09:27:23 moin :) Apr 28 09:34:08 I'll have to jump into the water head first and see what happens :P Apr 28 09:54:06 midori on easy debian's chroot works pretty good Apr 28 09:54:10 reddit and the like work fine Apr 28 09:54:36 not mobile, though Apr 28 09:54:43 mobile loads a heck of images at the same time Apr 28 09:55:33 he he Apr 28 09:55:59 maybe we should have some system wide proxy that would resize images on the fly? Apr 28 09:56:55 it's very easy to implement, and i have both pieces of code done already (proxy and tiny/fast imageresizer) Apr 28 09:57:15 I meant the i.reddit.com, sorry not m.reddit :) Apr 28 09:57:50 Yeah I had set up a proxy on my vps that was supposed to compress images, grayscale them etc but for some reason I couldn't get it working through the n900 Apr 28 09:57:50 ah Apr 28 09:57:53 didn't know that thing Apr 28 09:58:02 No matter what I did the proxy settings were ignored Apr 28 09:58:03 that should work pretty fine on midori, for sure Apr 28 09:58:13 anyway imgur > reddit Apr 28 09:58:18 KotCzarny: ain't that what opera does? Apr 28 09:58:23 I'll have to check it out' Apr 28 09:58:33 I think opera stopped the servers or something recently Apr 28 09:58:34 yes, but it does it outside your control Apr 28 09:58:52 imgur is useless if I don't look at pictures on reddit to begin with :P Apr 28 09:59:04 pah :P Apr 28 09:59:22 http://imgur.com/r/AbandonedPorn :P Apr 28 10:00:00 http://imgur.com/r/cableporn too Apr 28 10:00:34 and of course basis of the whole internet: http://imgur.com/r/StartledCats Apr 28 10:01:09 :D Apr 28 10:01:36 https://www.meneame.net/backend/media?type=post&id=1722162&version=0&ts=1385224111&image.jpeg Apr 28 10:01:39 this is mine :) Apr 28 10:01:42 well, she's now bigger Apr 28 10:02:08 doesnt look startled at all Apr 28 10:02:14 nope Apr 28 10:02:23 https://www.meneame.net/backend/media?type=post&id=1768394&version=0&ts=1391426230&image.jpeg Apr 28 10:03:20 http://imgur.com/r/Eyebleach Apr 28 10:04:39 you can just try any word after /r/ because it just matches user added tags for pics Apr 28 10:05:42 lol Apr 28 10:06:34 and yes, imgur is the best source for free pr0n (static and animated ;) Apr 28 10:07:19 i've got work to do, damn it! Apr 28 10:07:32 :> Apr 28 10:09:41 But can it update my phone? :P Apr 28 12:18:58 hi people, I need flasher for N9 and all links are dead, is there any hope for me Apr 28 12:34:52 tnx found it Apr 28 13:23:26 how much I would right now my yappari... Apr 28 13:23:42 oh my... this sorrow won't fade soon Apr 28 13:24:30 please someone tell me that I can install and use yappari Apr 28 13:24:52 Someone is addicted to social media :) Apr 28 13:25:41 infobot, tell me that you have my yappari Apr 28 13:25:56 it won't invalid target says Apr 28 13:26:03 what's that?! Apr 28 13:26:07 why!? Apr 28 13:26:40 even you infobot Apr 28 13:30:24 i think ive found a fail-safe solution to running whatsapp on linux, download virtualbox, create android-x86 vm, and install whatsapp app :D Apr 28 13:31:07 the obvious solution Apr 28 13:41:16 zzap-x: not viable on maemo Apr 28 13:41:31 aye true Apr 28 14:45:04 ~sb Apr 28 14:45:04 methinks scratchbox is a cross-compiling system that uses binfmt_misc, rpc calls, and an nfs mount to make a cross-build appear to be 100% native, and is found at http://www.scratchbox.org/, hosted by maemo now. Also at http://maemo.merlin1991.at/files/SB Apr 28 14:54:54 ~emmc Apr 28 14:54:55 well, emmc is is http://nds2.fds-fire.nokia.com/fdp/interface/FiRe/2010/5/--FID--A0A22YHFSICNA/--LID--FiRe1275051276916/AE98ED9D_RX-51_2009SE_10.2010.13-2.CENTRAL-EUROPE_PR_EMMC_MR0_ARM.bin or see ~emmc2 Apr 28 14:55:01 tnx dear Apr 28 14:55:15 you probably want ~flashing Apr 28 14:55:21 ~flashing Apr 28 14:55:22 hmm... maemo-flashing is http://wiki.maemo.org/Updating_the_tablet_firmware, or - on linux PC - download&extract http://maemo.cloud-7.de/maemo5/patches_n_tools/maemo-my-private-workdir.tgz, cd into it, do sudo ./flash-it-all.sh; or see ~flashing-cmdline Apr 28 14:55:49 KotCzarny: oh no I have those :) btw, can you show me the hashes if you have the files? Apr 28 14:55:58 I have them downloaded long time ago, not sure if the same Apr 28 14:56:14 grab the script from the last link Apr 28 14:56:18 it has hashes inside Apr 28 14:56:21 KotCzarny: tnx! Apr 28 15:04:59 gentoo users, how to get 32bit libusb? Apr 28 15:05:13 ~flasher Apr 28 15:05:14 [flasher] at http://www.jedge.com/n810/flasher/maemo_flasher-3.5_2.5.2.2.tar.gz (also .exe!), or http://www.chakra-project.org/ccr/packages.php?ID=5027 or generally http://www.google.com/search?q=maemo_flasher-3.5_2.5.2.2.tar.gz. HARMattan(N9): https://aur.archlinux.org/packages/fl/flasher-harmattan/flasher-harmattan.tar.gz, or -- list of filenames/md5sums: http://pastebin.com/sYKdNJSH, or http://galif.eu/nokia/ Apr 28 15:05:22 there was also the open source one hm Apr 28 15:05:34 ~0x0ffff Apr 28 15:05:36 no? Apr 28 15:05:42 just grab it from slackware or something Apr 28 15:06:27 ~0xFFFF Apr 28 15:06:27 i heard 0xffff is https://github.com/pali/0xFFFF, or http://talk.maemo.org/showthread.php?t=87996 Apr 28 15:06:43 KotCzarny: ah yea, forgot that my OS is pure64, crap :( Apr 28 15:07:05 just run some livecd Apr 28 15:07:09 Pali: right on time, lol Apr 28 15:07:10 problem solved Apr 28 15:07:17 there's usually a libusb32 or sth like that Apr 28 15:08:51 http://paste.opensuse.org/50854736 Apr 28 15:09:55 antranigv: what? Apr 28 15:10:59 Pali: just was gonna try 0xFFFF, flasher is not working cz of missing libs :)) Apr 28 15:11:23 antranigv: both 0xFFFF and flasher-3.5 needs libusb 0.1 Apr 28 15:11:29 version 0.1 Apr 28 15:11:37 not 1.0; not compat!!! Apr 28 15:12:25 http://paste.opensuse.org/50854736 Apr 28 15:13:03 Pali: AFAIK in gentoo pre-1.0 is named libusb-compat Apr 28 15:13:26 I'll give it a try Apr 28 15:13:30 I had included the lib stuff into that one if there was a generic way to do so: Apr 28 15:13:37 ~slashing-cmdline Apr 28 15:13:37 there is "compatibility" library which has libusb 0.1 API, but just wraper around libusb 1.0 Apr 28 15:13:44 ~flashing-cmdline Apr 28 15:13:44 somebody said flashing-cmdline was http://mg.pov.lt/maemo-irclog/%23maemo.2016-04-09.log.html#t2016-04-09T01:18:12 Apr 28 15:13:47 that version of libusb will not work correctly Apr 28 15:13:59 there are bugs in libusb 1.0 which upstream ignore Apr 28 15:14:13 and so... 0xFFFF does not work with libusb 1.0 correctly Apr 28 15:14:32 to have 0xFFFF working you need to use *original* libusb 0.1 Apr 28 15:14:43 aha~ Apr 28 15:15:18 now I got it :) time to try Apr 28 15:17:11 hmm ctcp-version doesn't give away the distro Apr 28 15:18:05 maybe we could list the according instructions for the major distros somewhere? Apr 28 15:18:24 it worked! it was the libusb-compat package :) Apr 28 15:18:43 DocScrutinizer05: agree, gonna do now for gentoo/funtoo Apr 28 15:19:02 the question is "where?" Apr 28 15:19:33 iirc I locked http://wiki.maemo.org/Updating_the_tablet_firmware Apr 28 15:21:04 should be unlocked now Apr 28 15:22:23 please edit http://wiki.maemo.org/index.php?title=Updating_the_firmware&action=edit§ion=7 Apr 28 15:27:02 well, seems it already has all it needs Apr 28 15:27:10 locking it again Apr 28 15:28:02 antranigv: or is your gentoo approach different to the apt based distro approach? Apr 28 15:30:47 DocScrutinizer05: not that much, just the package names are different, etc. Apr 28 15:31:20 then please add it at bootom after the apt distro instructions Apr 28 15:31:26 bottom Apr 28 15:32:15 before "some notes: 2:" Apr 28 15:39:42 tell me when you're done so I can lock the page again Apr 28 15:49:16 okey, just broke smthn :) Apr 28 15:49:50 Raw data transfer EP found at EP2. Apr 28 15:49:50 SU_GET_UPDATE_STATUS_REQ terminated with error code 1. Apr 28 15:50:52 you should listen what our guru Pali toldya Apr 28 15:51:20 [2016-04-28 Thu 17:11:29] version 0.1 Apr 28 15:51:21 [2016-04-28 Thu 17:11:37] not 1.0; not compat!!! Apr 28 15:52:25 but it worked with the rootfs :/ Apr 28 15:52:35 well Apr 28 15:53:36 btw that's the reason why the wikipage was locked Apr 28 15:53:49 it worked! Apr 28 15:53:52 what reason? :/ Apr 28 15:54:19 the reason that this page only allows 100% correct and verified info Apr 28 15:55:39 and when Pali is using a THREE exclamation marks - sth he usually never does Apr 28 15:55:42 ... Apr 28 15:58:44 it's locked again Apr 28 16:09:21 wtf, nothing is working :/ can it be hardware problem? Apr 28 16:09:44 to be more specific Apr 28 16:09:45 your description means nothing Apr 28 16:09:50 for example Apr 28 16:09:53 yes, please be more specific Apr 28 16:09:54 opening a terminal Apr 28 16:10:03 and then it closes, saying internal error, etc :) Apr 28 16:10:11 what did you flash? Apr 28 16:10:14 only emmc? Apr 28 16:10:47 emmc and ./flasher-3.5 -F RX-51_2009SE_10.2010.13-2.VANILLA_PR_EMMC_MR0_ARM.bin -f AND ./flasher-3.5 -F RX-51_2009SE_21.2011.38-1_PR_COMBINED_MR0_ARM.bin -f Apr 28 16:10:57 vanilla and combined Apr 28 16:11:06 and both completed without error? Apr 28 16:11:14 yup Apr 28 16:11:25 check dmesg for badblocks? Apr 28 16:11:47 though it would be tricky without terminal Apr 28 16:11:50 does ssh work? Apr 28 16:12:08 first I need to install ssh, but if no program is running, how can I install ssh :)) Apr 28 16:12:26 btw Apr 28 16:12:35 first I did emmc, then the rootfs Apr 28 16:12:42 I did it right, right? :/ Apr 28 16:13:31 that flash_it_all script does full flash, did you try it? Apr 28 16:14:32 its kind of flash and forget Apr 28 16:16:08 greate, so it also downloads, I should've read the script from the begining Apr 28 16:16:15 yup Apr 28 16:16:19 told you Apr 28 16:26:28 KotCzarny: same problem :/ Apr 28 16:26:39 sorry, flashing RX-51_2009SE_10.2010.13-2.VANILLA_PR_EMMC_MR0_ARM.bin (content of eMMC, incl /home/user/) failed with errorcode 0 Apr 28 16:26:59 then its not flashed? Apr 28 16:27:21 nop, not that file, only the first file Apr 28 16:29:19 use decent cable and a primary USB jack of your PC Apr 28 16:29:43 it;s a lappie ^^ ok, gonna try another cable :) Apr 28 16:49:59 lol Apr 28 16:50:01 usbfs: process 11361 (flasher-3.5) did not claim interface 2 before use Apr 28 17:06:39 you should use the correct lib# Apr 28 17:22:04 i got yowsup working, seems to send and recieve ok, im too scared to create a group incase i get banned Apr 28 17:44:40 zZap-X: do eeet Apr 28 17:44:51 #yolo Apr 28 17:45:11 #yowao Apr 28 17:48:48 yowsupppp Apr 28 19:05:24 INFO:yowsup.common.http.warequest:{"status":"fail","reason":"no_routes","retry_after":3600,"sms_wait":3600,"voice_wait":3600} Apr 28 19:05:27 grrrr Apr 28 19:25:20 what's up with whatsapp? Apr 28 19:26:56 zamn900: dunno i cant register Apr 28 19:27:09 who are you? Apr 28 19:27:18 lol me Apr 28 19:27:26 sure Apr 28 19:31:55 I'm Mr Zuckerberg, hello. Apr 28 19:34:01 I am sure someone did tell you to say so... that's pointless Apr 28 19:34:27 :-/ Apr 28 19:35:19 even aimless Apr 28 19:35:23 * zZap-X contacts Mr Zuckerberg for update Apr 28 19:35:42 not accurated at all Apr 28 19:36:47 and you know why? because I am Zuckerberg Apr 28 19:37:12 :O Apr 28 19:37:48 suckerzwerg Apr 28 19:38:50 lol Apr 28 19:39:38 right... Apr 28 19:40:04 I would appreciate to say that face to face Apr 28 19:40:16 not by facebook Apr 28 19:40:26 farcebook Apr 28 19:40:44 facecrook Apr 28 19:41:05 * DocScrutinizer05 feels silly Apr 28 19:43:17 i didn't advance anything today on yappari :( Apr 28 19:45:11 ceene, then tell us what did you do swapping your hero's duty? Apr 28 19:45:53 i tried Apr 28 19:46:05 but i'm facing strange things Apr 28 19:46:16 like? Apr 28 19:46:35 batman calls you to help him? Apr 28 19:47:00 real life job Apr 28 19:47:02 flash was having a relation too fast with your girlfriend? Apr 28 19:47:06 and segfaults Apr 28 19:47:09 oh ok... Apr 28 19:47:17 those are the biggest obstacles right now Apr 28 19:47:47 nothing bad as penguin or the jocker Apr 28 19:47:57 or worse? Apr 28 19:48:14 nope Apr 28 19:49:07 even for all the money in the world would't happen you to manage that... Apr 28 19:51:23 do you know I recall a film in which there was a hacker and a big boss of crimanls who wanted the hacker to crack a password in a few minutes with a gun on the head and a whore sucking his cock too... both handicaps while doing some hack... aiming to kill him at the end... but unsuccessfully... Apr 28 19:51:41 lol Apr 28 19:52:14 I am serious... I can't forget that particoular scene Apr 28 19:53:47 http://www.effecthacking.com/2015/05/top-10-best-hacking-scenes-from-movies.html Apr 28 19:55:13 oh my thank suckerberg Apr 28 19:57:32 btw I'd suggest much more to see mr.robot the serie Apr 28 19:58:14 veryvery blowing of mind stuff Apr 28 20:00:30 already seen Mr Robots, excellent looking forward to next season Apr 28 20:00:49 ofc it's addictive Apr 28 20:01:02 you gotcha man... Apr 28 20:01:17 and did you see the kingsmen? Apr 28 20:01:39 not hacker stuff but very cool film Apr 28 20:01:40 nope Apr 28 20:01:46 * zZap-X will make note Apr 28 20:01:49 awesome Apr 28 20:02:40 quote: "huh huh... you got your shhhhhhhit!" Apr 28 20:02:50 lol Apr 28 20:16:53 is there an easy way to have both wlan0 and gprs0 connected at the same time? what scripts control the ifdown of wlan0 when gprs0 is up and vice versa? Apr 28 20:18:33 I use "advanced interface switcher" Apr 28 20:18:41 on condition that wlan0 is being connected to AP in STA mode, that is Apr 28 20:21:27 sometimes my sim card slips away from it's plug little case... and that's what caused me sometime to have no gsm connection at all... I solved that by adding to my sim some adesive stripe... and now is thicker... now plugs well... Apr 28 20:22:16 afaik, advanced interface switcher does different thing, it lets you choose what is disabled and enabled before any connection could be made. the problem is that attempt to connect to gprs while wlan0 is connected results in wlan being disconnected and the other way around. Apr 28 20:23:28 what you want ult is very possible. connect gprs0 using maemo's connection system, and use your command line tool of choice for wlan0 Apr 28 20:24:59 any other than wpa_supplicant you can tell me about? Apr 28 20:30:25 i use wpa_supplicant with a simple shell script made by Sicelo ;) Apr 28 20:31:48 great, wpa_gui can't found the interface...the list is just empty. where could i have a look at this script? Apr 28 20:32:46 my phone is too private... I can't share how I do it... even if I understood one word of yours Apr 28 20:35:16 i think i edited wpa_supplicant.conf manually for my wlan Apr 28 20:37:44 then i have script "modprobe w112xx, stop wlancond, ifconfig wlan0 up, wpa_supplicant -i wlan0 -c /home/user/wpa_supplicant.conf -B, ifconfig wlan0 ip Apr 28 20:39:14 wpa_supplicant.conf is just my wlan name and password Apr 28 20:40:19 the problem is that for each network you may encounter you've to add an entry in conf file Apr 28 20:41:12 true, I don't know any other way and this works for me becuse i only use wlan at home Apr 28 20:41:22 is it hardcoded in wlancond to treat wlan and gprs as mutually exclusive? Apr 28 20:41:46 i don't know Apr 28 20:41:59 but would guess, yes it is... Apr 28 20:44:08 you can't manipulate wlan0 while wlancond owns it :) Apr 28 20:44:59 i hoped it's up for some patchable script that checks what's coming up up and deconfigures the conflicting ifaces Apr 28 20:47:42 and there is no way to bring up gprs through any other facility than wlancond? Apr 28 20:50:13 there might .. but not easy Apr 28 20:50:21 i gave you the easiest way :) Apr 28 21:03:47 when i start wpa_supplicant -i wlan0 -c /etc/wpa_supplicant.conf which contains only ctrl_iface it goes CTRL-EVENT-DISCONNECTED bssid=00:00:00:00:00:00 reason=0 like crazy Apr 28 21:06:10 ult: what's the use of having two parallel internet interfaces when you don't set up according routes? Apr 28 21:06:59 +1 :) Apr 28 21:07:18 you didn't explain what you need 2 iface up Apr 28 21:07:20 *why Apr 28 21:07:51 custom rt_table could be put in place, ip rules configured to route certain traffic via gprs while other stuff through ap Apr 28 21:08:19 ((is it hardcoded in wlancond to treat wlan and gprs as mutually exclusive?)) that's "hardcoded" into the way internet connectivity works Apr 28 21:09:14 I think what you want to look into is dnsmasq Apr 28 21:09:43 no way, multiple default routes via multiple ifaces are possible. what's the connection to dnsmasq? Apr 28 21:10:09 look into it and you'll see Apr 28 21:10:36 icd2, dnsmasq, wlancond Apr 28 21:10:47 "per interface forwarders" is what i've seen, am i missing smth? Apr 28 21:14:36 http://paste.opensuse.org/17098343 Apr 28 21:16:17 >>The dnsmasq DHCP server supports static address assignments and multiple networks.<< Apr 28 21:22:26 elaborate a bit. there's no problem in having it manage several interfaces at once. my goal is to pinpoint the script which brings down one iface when other comes up. IF it's done by script and not some logic inside something which is closed source. Apr 28 21:25:16 https://talk.maemo.org/showthread.php?t=96108 Apr 28 21:26:08 ICD = Internet Connection Deamon Apr 28 21:27:56 >>To be honest I recommend avoiding ICD like the plague since most of it seems to come from symbian...<< Apr 28 21:28:54 http://maemo.org/api_refs/5.0/beta/icd2/group__dbus__api.html Apr 28 21:29:08 DocScrutinizer05: do we have easy way to control gprs0 on N900? Apr 28 21:29:22 dunno Apr 28 21:29:44 tbh I never looked into it Apr 28 21:30:38 http://maemo.org/api_refs/5.0/beta/icd2/group__srv__provider__api.html Apr 28 21:32:55 http://maemo.org/api_refs/5.0/beta/icd2/group__network__module__api.html Apr 28 21:35:30 http://wiki.maemo.org/N900_dbus Apr 28 21:47:51 ALOT of info, thx. any idea how to discover what exactly brings device down? strace wlancond - megs of garbage. gdb breakpoint on some syscall? any other way to install hook on event? Apr 28 21:50:00 /etc/init.d/network /etc/dnsmasq.conf Apr 28 21:50:45 oops sorry Apr 28 21:51:20 /etc/init.d/networking Apr 28 21:54:06 seems /sbin/ifup is basically what you want to look into Apr 28 21:54:22 though it's entangled with ICD(") Apr 28 21:54:27 2 even Apr 28 21:55:32 obviously a connection gets enabled and marshalled to sustem only when DHCP succeeded Apr 28 21:55:40 system* Apr 28 21:55:54 tgus dnsmasq is also entangled Apr 28 21:56:29 I even think dnsmasq is the last element in that chain, that finally sets up the route and gives the OK to use the IF Apr 28 21:57:09 someone must send AT commands to the modem in the case of gprs0. who is that? i guess ICD? Apr 28 21:58:41 libicd2 libisi dunno Apr 28 21:59:56 I'm not even sure it tears down the PDP_context Apr 28 22:00:31 dnsmasq serves as dns forwarder and cached reply relay, plus dhcp server. but in both cases with gprs and wlan in STA mode we're on the client side. what dnsmasq has to do with connection completion in this scenario, i'd rather assume that udcpd is involved, no? Apr 28 22:00:37 ask frals, he might know details Apr 28 22:01:11 a ps will enlighten you Apr 28 22:01:25 udhcpc is what our dhcp client. dnsmasq is our dns server Apr 28 22:02:20 root 31334 0.0 0.2 1628 576 ? S Apr28 0:00 /sbin/udhcpc -i wlan0 -s /etc/udhcpc/libicd_network_ipv4.script -H IroN900 -f -R 15 Apr 28 22:02:44 dnsmasq is not used as dhcp server on maemo's default config. of course the ability is there and you can use it for yourself Apr 28 22:03:16 IroN900:~# cat /etc/udhcpc/libicd_network_ipv4.script Apr 28 22:03:17 #!/bin/sh Apr 28 22:03:19 # Place your script to /etc/maemo-dhcp.d directory so that it gets run. Apr 28 22:03:20 run-parts -a "$1" /etc/maemo-dhcp.d Apr 28 22:04:19 less /etc/maemo-dhcp.d/50_ipv4_network_setup Apr 28 22:05:38 /usr/bin/dbus-send \ --system \ --dest=com.nokia.icd \ /com/nokia/icd/autoconf \ com.nokia.icd.autoconf.autoconf_changed \ string:$interface string:DHCP string:CONNECTED Apr 28 22:08:02 i wonder if that dbus cmd is what i'm missing in my wpa_supplicant string .. i have dummy network, and all applications 'see' the connection. pidgin does not however. it still works, but it keeps showing "waiting for connection" Apr 28 22:08:42 system informs maemo aps about connectivity status Apr 28 22:08:46 will try it tomorrow at work (which is the only place i need wpa_supplicant .. stupid windows-based radius network) Apr 28 22:09:06 ~usb-networking Apr 28 22:09:07 [usb-networking] http://wiki.maemo.org/USB_networking, or http://wiki.maemo.org/N900_USB_networking, or http://wiki.maemo.org/User:Joerg_rw/usb-networking Apr 28 22:09:55 i have all that :) Apr 28 22:10:10 and works perfectly for all applications, except pidgin Apr 28 22:13:06 hmmm Apr 28 22:13:33 does pidgin need to authenticate anew? Apr 28 22:14:19 I mean, for all my IRC clients they need to log in again after internet connection got changed Apr 28 22:14:38 not only on maemo Apr 28 22:17:05 ult: sorry I was a tad grumpy, I'm hungry and tired. Actually this whole internet connection stuff on maemo is a maze and a mess Apr 28 22:17:26 I got no real clue either Apr 28 22:19:16 what I can tell you is a maybe interesting observation that trying to switch from GPRS to WLAN doesn't break the connection until WLAN actually provided a DHCP lease Apr 28 22:19:50 if it doesn't, the connection try will get aborted and user gets a requester "failed! try again?" Apr 28 22:20:06 during all that the GPRS connection happily works on Apr 28 22:25:51 yes .. i like that particular feature. saves you getting disconnected into a bad network :) Apr 28 22:26:14 indeed, the NITs are meant for always-on connectivity Apr 28 22:26:20 unrelated: half a day of IroN900 idling, as seen by router http://wstaw.org/m/2016/04/29/plasma-desktopxb2219.png Apr 28 22:27:41 nice router Apr 28 22:27:46 indeed Apr 28 22:27:50 ubiquiti Apr 28 22:28:14 great Apr 28 22:28:17 tried to enclose/etc/maemo-dhcp.d/50_ipv4_network_setup deconfig) section in { if [ "$interface" != "gprs0" ]; then} but to no avial. but with added {echo "$(date) $interface" >> /tmp/2} before the condition check it looks like wlan gets deconfigured even when you're changing connection from gprs to wlan...weird Apr 28 22:29:31 well, tear it down just in case user messed with it meanwhile ;-) Apr 28 22:30:45 you quite easily could just use wifieye despite not being connected to an AP and rather using GPRS Apr 28 22:31:15 in that case _obviously_ WLAN is active, but somewhat incorrectly configured Apr 28 22:32:30 and maybe that's just what you want to do: disallow "automatic switching to WLAN when available" in settings, then absolutely normally bring up WLAN and connect and add a route Apr 28 22:33:27 I guess it's a nightmare to teach ICD2 that trick to keep 2 concurrent connections active Apr 28 22:34:18 so best and simlest thing is to not tell "maemo" aka ICD2 / wlancond etc pp about any AP it's allowed to connect to Apr 28 22:34:48 then connect "manually" like on any other linux system Apr 28 22:35:15 see wifieye for a prolly helpful example how to bring up and tear down WLAN Apr 28 22:36:00 though wifieye only does scans, connecting to an AP can't be significantly more complicated Apr 28 22:38:58 "auto switch to wifi" doesn't provide a way to have both gprs and wlan simultaneously up. connect manually - do you mean via wpa_supplicant? Apr 28 22:45:50 http://processors.wiki.ti.com/index.php/WL127x_WLAN_API_Information#mac802.11_Basics Apr 28 22:46:18 >>"auto switch to wifi" doesn't provide a way to have both gprs and wlan simultaneously up<< is *exactly* what I said Apr 28 22:47:52 "auto switch to wifi disabled" doesn't seem to do the trick either Apr 28 22:51:10 but wpa_supplicant succeded in connection to AP while gprs0 was ( and remained) active, provided it had needed network descr block in conf. all that is left is to automate the process of doing it with 1 command Apr 28 22:58:58 * DocScrutinizer05 waves Apr 28 23:00:02 glad to hear my suggestion worked for you **** ENDING LOGGING AT Fri Apr 29 02:59:58 2016