IRC log for #maemo-ssu on 20140211

01:34.32*** join/#maemo-ssu sixwheeledbeast (~paul@host-78-144-100-179.as13285.net)
01:47.40*** join/#maemo-ssu MohammadAG (~MohammadA@Maemo/community/contributor/MohammadAG)
01:53.53*** join/#maemo-ssu Milhouse (~Milhouse@Maemo/community/contributor/Milhouse)
02:28.46MaxdamantusJust wondering .. what's the difference between zImage-2.6.28.10-cssu1 and zImage-2.6.28.10-cssu1.fiasco?
02:32.27MaxdamantusThere seems to be some metadata at the beginning of the .fiasco, but flasher doesn't unpack it with -u
02:33.26MaxdamantusOh, nvm .. was using the flasher wrong.
02:37.12Maxdamantusalso answers the question I asked in #maemo the other day of what "FIASCO" means .. I think the people there thought I was asking about etymology, but I just wanted to know that it referred to essentially an archive format (and not a software distribution or something).
03:01.34*** join/#maemo-ssu g3kk3r (torrkull@lehtori.cc.tut.fi)
03:04.48*** join/#maemo-ssu amizraa (~amizraa@gateway/tor-sasl/amizraa)
03:08.15*** join/#maemo-ssu silviof (~silviof@unaffiliated/silviof)
03:27.00*** join/#maemo-ssu sixwheeledbeast (~paul@host-92-28-73-86.as13285.net)
03:29.00*** join/#maemo-ssu amiconn_ (quassel@rockbox/developer/amiconn)
04:23.39*** join/#maemo-ssu kostola (~quassel@antani.ispirata.biz)
04:43.22*** join/#maemo-ssu unclouded (~neil@2001:4428:200:80fc:7e05:7ff:fe0d:d421)
06:44.50*** join/#maemo-ssu sunny_s (~sunny_s@business-092-079-020-027.static.arcor-ip.net)
06:50.10*** join/#maemo-ssu Maxdamantus (~Maxdamant@2001:470:f078::dead:beef:cafe)
07:18.47*** join/#maemo-ssu thedead1440 (~thedead14@unaffiliated/thedead1440)
07:26.11*** join/#maemo-ssu Maxdamantus (~Maxdamant@2001:470:f078::dead:beef:cafe)
07:49.55DocScrutinizer05(i386 packages of cssu) would that change anything after all? I mean, does xephyr have orientation support? and how? would MHD shortcuts really work on xephyr? Anything else difference you could notice UX wise?
07:51.03DocScrutinizer05xephyr stuff is meant to test *apps* during development, no?
07:51.28DocScrutinizer05we got no CSSU-only apps (well, not really)
07:52.17DocScrutinizer05and that's fine the way it is, and expected to stay this way
07:54.37DocScrutinizer05finally tests rollback of wlancond, after err 2(?) years of IRC not reconnecting after connection roaming
07:54.56silviofDocScrutinizer05: Its a matter of test of software. I use my maemo-based mobile on daily basis. If I develop some components than I will test it. But without a second phone I have only the xephyr-way. But the xephyr way is only functional on X86.
07:55.21DocScrutinizer05so?
07:55.27silviofso!
07:55.31DocScrutinizer05how's that related to cssu?
07:56.33DocScrutinizer05you can't test cssu "components" on x86
07:56.41DocScrutinizer05generally
07:56.54DocScrutinizer05and any app isn't supposed to need cssu to run
07:57.44silviofno you cant, because if you change something in hildon-desktop you need to build a package and to install with dpkg. Thats is not able because some depencies are not provided for the x86.
07:57.54DocScrutinizer05so, depending on what "components" you want to develop, you're either fine with x86 emu as is, or can't use it at all
07:59.52DocScrutinizer05and not because there's no cssu x86 available, but because x86 isn't the right thing to use for some components to test
08:03.28DocScrutinizer05IroN900:~# ll /usr/sbin/wlancond*
08:03.30DocScrutinizer05-rwxr-xr-x 1 root root 59764 2012-08-22 19:53 /usr/sbin/wlancond
08:03.31DocScrutinizer05-rwxr-xr-x 1 root root 57692 2011-12-24 17:38 /usr/sbin/wlancond.bak
08:03.33DocScrutinizer05MHM!
08:05.42DocScrutinizer05IroN900:~# mv /usr/sbin/wlancond /usr/sbin/wlancond.new
08:05.43DocScrutinizer05IroN900:~# cp /usr/sbin/wlancond.bak /usr/sbin/wlancond
08:05.45DocScrutinizer05IroN900:~# uptime
08:05.46DocScrutinizer05<PROTECTED>
08:05.48DocScrutinizer05IroN900:~# dsmetool -b
08:07.42DocScrutinizer05let's see if that will fix this annoying obscure incapability of xchat to detect a change of own IP -> connection stall -> reconnect_to_server
08:19.03*** join/#maemo-ssu MohammadAG_ (~MohammadA@Maemo/community/contributor/MohammadAG)
08:21.25*** join/#maemo-ssu sixwheeledbeast (~paul@host-2-97-92-123.as13285.net)
08:23.53DocScrutinizer05merlin1991: http://mg.pov.lt/maemo-ssu-irclog/%23maemo-ssu.2011-12-26.log.html#t2011-12-26T22:46:07
08:24.27DocScrutinizer05merlin1991: what's been the purpose of patched wlancond? I can't find "wlancond" string in cssu-t changelog
08:38.52*** join/#maemo-ssu freemangordon_ (~ivo@printec.bg)
09:20.21merlin1991the package name is osso-wlan, it contains 1 fix for the injeciton driver that has no side effects and a fix for eapd
09:21.37merlin1991commits: https://gitorious.org/community-ssu/osso-wlan/commit/d33e64330937dd1812bed257a4525513ebaf8fe1 and https://gitorious.org/community-ssu/osso-wlan/commit/22c5e804febd00195dc48dae652b120a3e289369
09:22.45*** join/#maemo-ssu jon_y (~enforcer@2001:e68:4053:2e8f:e01f:8e6d:cdca:d0e5)
09:24.56merlin1991DocScrutinizer05: ^^
09:27.59DocScrutinizer05thanks
09:28.17DocScrutinizer05what's eapd?
09:29.23merlin1991its part of the wpa auth stuff
09:29.33DocScrutinizer05aah
09:30.22DocScrutinizer05I'm finally fed up with xchat taking 60min to reconnect to ZNC when I leave WLAN range and roam to GPRS
09:30.51DocScrutinizer05checking all possible culprits now, but I can't find anything that makes sense
09:33.08merlin1991there's always the chance that xchat is the culprit ;)
09:45.38DocScrutinizer05yeah, but I can't recall it changed since it worked, which it evidently did
09:45.59DocScrutinizer05only change I can think of: update to ZNC1.0
09:47.32DocScrutinizer05might ZNC ping-keep-alive not be aware of IP changes?
09:48.03DocScrutinizer05which would mean it answers pings from arbitrary source IP
09:48.28DocScrutinizer05no matter if a client authenticated there or not
09:50.14DocScrutinizer05and even if that was the case, then where from are the 3600 seconds from IP change on N900 till xchat re-registering with ZNC
09:51.40DocScrutinizer05or maybe xchat doesn't accept "net_ping_timeout = 90" and defaults to 3600?
09:52.52*** join/#maemo-ssu freemangordon_ (~ivo@printec.bg)
09:58.59DocScrutinizer05merlin1991: do you have the xchat code at hand? could you check if xchat rejects "net_ping_timeout = 90" and defaults to 3600?
09:59.36DocScrutinizer05it seems I might have changed the net_ping_timeout from 240 to 90 eventually
10:10.22*** join/#maemo-ssu arcean (~arcean@aaep128.neoplus.adsl.tpnet.pl)
10:14.11DocScrutinizer51let's see what happens now when roaming to GPRS
10:15.01DocScrutinizer05ok
10:15.21DocScrutinizer05starts a timer in mind
10:17.08*** join/#maemo-ssu jon_y (~enforcer@2001:e68:4053:15f9:e01f:8e6d:cdca:d0e5)
10:17.36*** join/#maemo-ssu sunny_s (~sunny_s@business-092-079-020-027.static.arcor-ip.net)
10:22.23DocScrutinizer05ok, the 301 seconds I've sent net_ping_timeout to have expired. No reconnect of xchat >:-(
10:23.15DocScrutinizer05changing wlancond didn't help
10:23.38DocScrutinizer05changing all suspicious settings in ZNC didn't help
10:44.19DocScrutinizer05hmm, relogin after 25 min
10:45.08*** join/#maemo-ssu luf (~luf@80.188.29.62)
11:02.16*** join/#maemo-ssu joshgillies (~josh@ppp108-14.static.internode.on.net)
11:10.25DocScrutinizer51hmmm
11:10.40DocScrutinizer51*something* changed
11:11.02DocScrutinizer51it now does relogin after 25min
11:11.30DocScrutinizer05with "net_ping_timeout = 301"
11:11.57DocScrutinizer05strace of xchat looks strange
11:12.52DocScrutinizer05anyway no packets sent between ZNC and xchat for the whole duration of disconnect   => xchat is buggy, or fremantle system is buggy
11:28.41*** join/#maemo-ssu lizardo (lizardo@nat/indt/x-jorcdsrzdnfsdqru)
11:38.34*** join/#maemo-ssu NishanthMenon (~nmenon@nat/ti/x-myiadimugvhtbhba)
11:54.13*** join/#maemo-ssu sixwheeledbeast (~paul@host-2-98-100-159.as13285.net)
11:57.48*** join/#maemo-ssu javispedro (~javier@Maemo/community/contributor/javispedro)
12:15.35*** join/#maemo-ssu dhbiker (~dhbiker@95.87.145.172)
12:53.21*** join/#maemo-ssu hbib (~user@tmo-098-6.customers.d1-online.com)
13:08.27*** join/#maemo-ssu hbib (~user@tmo-108-92.customers.d1-online.com)
13:19.31*** join/#maemo-ssu hbib1 (~user@tmo-098-162.customers.d1-online.com)
13:22.31*** join/#maemo-ssu M4rtinK (~M4rtinK@2001:718:801:22e:a11:96ff:fe16:10e4)
14:16.27FIQDocScrutinizer05: for me, xchat is usually having a hard time to figure out that it disconnected if that's what you're referring to
14:16.34FIQnot just on N900, but on desktop as well
14:17.37DocScrutinizer05well, th<t been different for me, a maybe 1 or 2 years ago. eventually it changed to xchat needing exactly 60min after IP change to reconnect
14:18.00DocScrutinizer05and now it seems my tweaks changed that 25min
14:19.22DocScrutinizer05the only things i changed on N900 were: "net_ping_timeout = 90" -> "net_ping_timeout = 301" (SIC!), and replacing CSSU wlancond by a two years old orginal version
14:19.59DocScrutinizer05I can't see any logic in either of both changing time to reconnect form 60 to 25min
14:22.36DocScrutinizer05I nevertheless wonder if maybe ZNC is somehow remote-controlling what xchat does, regarding timeouts and whatnot
14:28.13DocScrutinizer05if that wasn't 25 but 18min 8possible, since I only tested 2 times and not very precisely) then this would suggest timeout is  5400min / net_ping_timeout
14:28.42DocScrutinizer05s/timeout/time till reconnect/
14:32.42DocScrutinizer05I wonder if IRC protocol does something similar like SIP, for keepalive timeout negotiation. I.E. both the server (here ZNC) and the client (xchat) negotiate the timeout based on some prefrences set on either side. In the end this would mean xchat actually gets "remote controlled" by ZNC like I suspected above
14:34.11*** join/#maemo-ssu NIN101 (~core@n900.quitesimple.org)
14:34.26DocScrutinizer05particularly: when ZNC tells xchat "hey, I gonna send a keepalive ping to you every 60 min" then it's of course clear why xchat won't notice the IP change resp "disconnect" any sooner
14:36.53DocScrutinizer05maybe xchat asks for 90 seconds interval, and ZNC answers 60 but some bug multiplies that by 60 to convert from seconds to minutes (or simply places wrong unit in there)
14:38.51DocScrutinizer05anyway, you can easily adjust xchat net_ping_timeout:  /set net_ping_timeout <value in seconds>
14:39.13DocScrutinizer05just in case you want to experiment a bit with this
14:40.51DocScrutinizer05note that the new value doesn't get saved to .xchat2/xchat.conf until you gracefully exit xchat (or find the "save" menu entry)
14:42.20DocScrutinizer05now if only I'd find a more easy method than changing the IP of N900, to test this
15:05.08*** join/#maemo-ssu Vazay (1fd2b0a2@gateway/web/freenode/ip.31.210.176.162)
15:15.12*** join/#maemo-ssu FIQ (~FIQ@fiq.se)
15:35.47*** join/#maemo-ssu arcean_ (~arcean@aacx252.neoplus.adsl.tpnet.pl)
15:57.31*** join/#maemo-ssu anYc (anyc@home.hadiko.uni-karlsruhe.de)
16:04.07*** join/#maemo-ssu Guest53546 (~FIQ@unaffiliated/fiq)
16:06.43FIQDocScrutinizer05: do you know how the irc protocol looks like?
16:07.20FIQthe only thing the server sends continuosly is a ping timeout check
16:08.11FIQthe server doesn't tell the client to do anything regarding connectivity (besides these pingpongs)
16:09.02FIQthe server *does* "remote control" irc clients, but only by issuing IRC commands to it, i.e. to set chanmodes/usermodes, join/part channels, change your nick and host, etc
16:09.45FIQand since there's a bajillion different irc servers and there's not really a strict standard, the client has to obey to be able to function properly everywhere
16:09.50*** join/#maemo-ssu joshgillies (~josh@ppp108-14.static.internode.on.net)
16:10.04DocScrutinizer05FIQ: does >>I wonder if IRC protocol does something similar like SIP<< sound like I know IRC protocol?
16:10.11FIQno it doesn't
16:10.16FIQthat's why I told a bit :D
16:10.29FIQi.e. the fact that the server doesn't tell the client anything regarding timeouts
16:11.49DocScrutinizer05aha
16:16.14DocScrutinizer05tbh i'm not particularly interested in IRC protocol. What i'd like to know are details about how xchat handles net_ping_timeout
16:17.46DocScrutinizer05I'd even think that keep-alive-pings are only loosely related to IRC protocol, if at all
16:18.20keriothe IRC protocol doesn't deal with it i think
16:18.34kerioi mean, there's no specified timeout
16:18.44DocScrutinizer05at least xchat has it under net_*, not under irc_*
16:24.50*** join/#maemo-ssu hbib (~wurmel@pD9E3DDF7.dip0.t-ipconnect.de)
16:42.42*** join/#maemo-ssu hbib (~wurmel@pD9E3DDF7.dip0.t-ipconnect.de)
16:55.06*** join/#maemo-ssu dos1 (~dos@81.219.200.186)
16:55.09*** join/#maemo-ssu dos1 (~dos@unaffiliated/dos1)
16:59.53ShadowJKThe irc server independently decides how often to send pings towards clien (if any)t, the client independently decides how often to send pings towards server, if any.
17:00.39ShadowJKxchat in maemo was patched to remove most timers, so all pinging is disabled by default.
17:01.34DocScrutinizer05DAMN! who decided THAT to be a good idea?
17:02.20ShadowJKBecause it ate humongous amounts of power before that change
17:02.31DocScrutinizer05I can't see any reasonable rationale behind that
17:02.48ShadowJKThere were about half a dozen 1s timers, iirc
17:03.03DocScrutinizer05it still spins in a poll() = timeout loop with ~2s loop time
17:03.35ShadowJKnet_ping_timeout is probably not even meant to send periodical pings
17:03.56DocScrutinizer05I don't care if it sends pings
17:04.29DocScrutinizer05it however shall notice missing inbound pings from server, or otherwise detect that server lost connection
17:04.49DocScrutinizer05this worked a year or max 2 years ago, nof it's fubar
17:04.57DocScrutinizer05now*
17:05.46DocScrutinizer05whoever decided to implement that patch err botch for sure didn't do a clever thing
17:05.53ShadowJKxchat doesn't know how often the server intends to send pings, or if it will send pings at all.
17:06.13ShadowJKThat it ever worked on your network was probably a coincidence or accident :P
17:06.34DocScrutinizer05well, that's why I thought it sends pings itself, and period should be the net_ping_timeout
17:06.58DocScrutinizer05no, definitely no coincidence
17:07.12DocScrutinizer05I carefully tuned it to work like it should
17:07.20ShadowJKAmusingly, googling turns up people on hexchat with same "issue"
17:07.42*** join/#maemo-ssu FredrIQ (~FIQ@fiq.se)
17:08.14*** join/#maemo-ssu okias (~okias@berger.cust.centro-net.cz)
17:08.32DocScrutinizer05and xchat never "ate battery like mad" om my device, see bq27200 logs on wiki power
17:09.00ShadowJKWe didn't have bq27200 access whenever xchat was ported
17:09.16DocScrutinizer05irrelevant
17:09.43DocScrutinizer05I had a working xchat and I metered it with bq27200.sh
17:09.49ShadowJKxchat hasn't been modified since 2009 or so anyways :P
17:10.18DocScrutinizer05then your whole story about removed timers is moot
17:10.51DocScrutinizer05since I told that my xchat worked a 2 years ago
17:11.26ShadowJKYeah that stuff was all done in Maemo4 and Maemo5 2009 until powertop showed saner number of wakeups
17:11.43ShadowJKI dont think xchat has been touched since
17:13.32*** join/#maemo-ssu FredrIQ (~FIQ@unaffiliated/fiq)
17:19.12ShadowJKxchat on PC behaves the same, I just tried set net_ping_timeout to 10, and nothing is sent
17:19.26*** join/#maemo-ssu FIQ (~FIQ@prgmr/customer/fiq)
17:21.07*** join/#maemo-ssu BCMM (~BCMM@unaffiliated/bcmm)
17:22.41ShadowJKugh, actually, it causes reoccurring disconnects, since no ping is sent, no pings can be received,  it is now is on a 10s reconnect cycle
17:22.45ShadowJKlol
17:23.17*** join/#maemo-ssu LauRoman (~LauRoman@5-14-92-17.residential.rdsnet.ro)
17:23.22DocScrutinizer05most weird stuff: the only things i changed on N900 were: "net_ping_timeout = 90" -> "net_ping_timeout = 301" (SIC!), and replacing CSSU wlancond by a two years old orginal version. which made xchat re-register after ~25min instead of 60min
17:25.21DocScrutinizer05I should try /set net_ping_timeout 600 and see if it reconnects after 5 minutes then
17:25.22ShadowJKTurning on gui_lagometer makes it send stuff every 10s or so on PC
17:25.32ShadowJKI use /timer -repeat 0 600 allserv ping
17:25.35ShadowJKon N900, to send a ping every 10 minutes
17:25.44DocScrutinizer05oooooh! lag-o-meter!!
17:25.51DocScrutinizer05I think I disabled that
17:26.02DocScrutinizer05DANG!
17:26.39DocScrutinizer05(/timer) niiiice!
17:27.27DocScrutinizer05but... does it notice when ping fails? and does it reconnect then?
17:28.33DocScrutinizer05lemme find where's that goddamn lagometer option
17:28.35ShadowJKIt would take 5 minutes for tcp to timeout, or if the other end is actually alive, it would probably send a RST and it'd fail immediately
17:28.42ShadowJK/set gui_lagometer
17:29.04DocScrutinizer05didn't use that, I'm sure
17:30.07DocScrutinizer05erheehee#@!!$%  line 54: gui_lagometer = 0
17:30.30DocScrutinizer05but I guess I used settings
17:33.19DocScrutinizer51ok
17:34.36DocScrutinizer05so, I just 2roamed" wlan->gprs
17:45.51DocScrutinizer511836 recdonnected
17:46.36DocScrutinizer05so it took 3 minutes when lagometer = 1
17:46.53DocScrutinizer05miracle solved
17:47.19DocScrutinizer05problem spotted - now how to eliminate it?
17:49.23DocScrutinizer05seems net_ping_timeout just sets a threshold for lagometer, eh?
17:49.41DocScrutinizer05no lagometer, no threshold, no timeout, no reconnect
17:51.07*** join/#maemo-ssu amizraa (~amizraa@gateway/tor-sasl/amizraa)
17:51.09DocScrutinizer05what we actually would want is a config parameter for ping interval of lagometer, while lagometer=1|0 SHALL only be a GUI thing
17:51.44DocScrutinizer05ShadowJK: thanks a megaton!
17:55.21*** join/#maemo-ssu sixwheeledbeast (~paul@cl-1547.lon-02.gb.sixxs.net)
17:58.43ShadowJKI wonder if net_ping_timeout also works on pings sent with /timer
17:59.20DocScrutinizer05that's what I asked
17:59.36DocScrutinizer05[2014-02-11 18:27:27] <DocScrutinizer05> but... does it notice when ping fails? and does it reconnect then?
19:25.21*** join/#maemo-ssu NishanthMenon (~nmenon@nat/ti/x-zjinypnxicvqshxy)
19:30.43*** join/#maemo-ssu Guest21837 (~FIQ@fiq.se)
19:46.43*** join/#maemo-ssu FredrIQ (~FIQ@fiq.se)
19:50.13*** join/#maemo-ssu FIQ- (~FIQ@fiq.se)
21:12.59*** join/#maemo-ssu honestly (~luke@unaffiliated/luketheduke)
21:20.14*** join/#maemo-ssu okias (~okias@berger.cust.centro-net.cz)
22:17.53*** join/#maemo-ssu M4rtinK (~M4rtinK@ip-78-102-146-9.net.upcbroadband.cz)
22:28.48*** join/#maemo-ssu FIQ- (~FIQ@prgmr/customer/fiq)
23:09.40*** join/#maemo-ssu Gagliano (959ad3aa@gateway/web/freenode/ip.149.154.211.170)
23:09.42Gaglianohttp://ip.gagliano.fr Free call over the world, free service finance by ads. No registration required, please help me to grow !!! Thank you all of you !
23:09.44Gaglianohttp://ip.gagliano.fr/ = Appel gratuit dans le monde, financé par les pubs, merci de cliquez sur les pubs si vous êtes satisfait du service. Merciiiiii de m'aider à améliorer mon service egalement.
23:09.50GaglianoMerci thank you ! ;-)
23:19.36*** join/#maemo-ssu xes (~xes@unaffiliated/xes)
23:25.56*** part/#maemo-ssu Gagliano (959ad3aa@gateway/web/freenode/ip.149.154.211.170)
23:31.58Maxdamantus05:16:14 < DocScrutinizer05> tbh i'm not particularly interested in IRC protocol. What i'd like to know are details about how xchat handles net_ping_timeout
23:32.09MaxdamantusDunno if it's been answered, but XChat sends pings itself to the server.
23:32.20Maxdamantusiirc, fairly frequently, when the lag-meter is enabled.
23:32.29MaxdamantusI think it avoids sending them when it's disabled.
23:33.29MaxdamantusI think you can open a raw log tab in XChat.
23:33.35Maxdamantusmaybe with the /rawlog command .. /me checks.
23:34.08MaxdamantusWindow → Raw log
23:34.51Maxdamantuseh, looks like that was already said.
23:35.02Maxdamantus(the ping thing)
23:36.06*** join/#maemo-ssu nox- (noident@freebsd/developer/nox)

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.