IRC log for #maemo-ssu on 20121115

00:09.46*** join/#maemo-ssu andre__ (~andre@wikimedia/aklapper)
00:51.07*** join/#maemo-ssu Milhouse (~irc_milho@Maemo/community/contributor/Milhouse)
01:31.50*** join/#maemo-ssu peetah (~peetah@cha92-9-82-236-202-86.fbx.proxad.net)
02:20.21*** join/#maemo-ssu _rd (~rd@p57B49084.dip0.t-ipconnect.de)
03:29.44*** join/#maemo-ssu amiconn_ (amiconn@rockbox/developer/amiconn)
03:30.57*** join/#maemo-ssu Milhouse (~irc_milho@Maemo/community/contributor/Milhouse)
04:03.41*** join/#maemo-ssu DocScrutinizer05 (~HaleBopp@openmoko/engineers/joerg)
06:25.52*** join/#maemo-ssu _rd (~rd@p57B49084.dip0.t-ipconnect.de)
06:52.42*** join/#maemo-ssu luf (luf@nat/ibm/x-eeonivbxbcxtjtwl)
07:38.02*** join/#maemo-ssu andre__ (~andre@wikimedia/aklapper)
08:05.37*** join/#maemo-ssu kolp (~quassel@212.255.28.120)
08:52.00luffreemangordon: I think it's good time for testing new bluez-4.99 ;)
09:25.32kerioluf: i keep having minor issues when connecting my BT headphones to the n900
09:27.01keriobut they work fine once connected
09:27.12kerioand it only happens once in a while, so it's an unfindable bug :(
09:27.22lufkerio: no report no issues ...
09:27.33kerioi know :(
09:27.45lufWhat's exactly the problem?
09:28.25kerioi get the "connected" message and the icon in the status area, but after a tiny bit it disappears, the bluetooth icon still blue and the UI still reporting it's "connected"
09:28.32keriousually disconnecting it and reconnecting it fixes it right away
09:29.04lufHmmm ok.
09:29.21lufWhat are you tryiing to connect? HF or computer?
09:29.29*** join/#maemo-ssu iDont (~iDont@145.93.49.62)
09:30.47kerioluf: A2DP
09:30.55lufSo audio ...
09:31.03kerioyep
09:31.11keriohttp://www.lg.com/us/cell-phone-accessories/lg-HBS700-tone <- these ones
09:31.20kerio(they're *really* neat)
09:32.39lufI think it doesn't matter which kind. There is probably some ugly bug in audio. I'll try to find it but to be honest I'm short on time this week.
09:36.49Skrykerio: those do look neat, how's the audio (music) quality?
09:36.56kerioSkry: pretty good
09:37.30keriocompression artifacts tend to be exacerbated by the decode-encode-decode though, so you *really* want to start with some good quality audio
09:37.50keriosay, V0 VBR MP3s, or some high-bitrate CBR
09:38.03keriothe range is kinda poor, but i think that's also a problem of the n900
09:38.21kerioit's perfectly fine if you keep your phone in a jacket's pocket, a bit less so if you keep it in your pants' pockets
09:40.10kerioit's got some heavy bass though, so you might need to get used to it in the beginning
09:41.29*** join/#maemo-ssu BCMM (~BCMM@unaffiliated/bcmm)
09:42.20SkryI have mostly flacs so that's no prob, and stuff I mostly listen is not that bass heavy. I was thinking about spending ~80eur to wired headphones so might as well give those a try.
09:43.50keriomind you, they still sound worse than a good pair of wired headphones
09:43.56keriothey're just a lot more convenient, imo
09:45.12Skrywell, they must beat those headphones that came with n900 so should be improvement anyways
09:45.21kerioi said *good* :)
09:46.34lufkerio: I'm just curious had you change all pulseaudio rates in N900 to 44100 from 48000?
09:46.57kerioluf: i changed the default one iirc
09:47.00keriocould it be the reason?
09:47.23kerioi changed default-sample-rate from 48000 to 44100
09:47.25lufNo I thinking about totally different thing ;)
09:47.39kerioluf: how many resample methods are there, btw?
09:48.06lufVery few :(
09:48.46lufJust speex and two unusable as I remember.
10:03.33kerioluf: how do i check what's going on with pulseaudio at a specific time?
10:03.47keriofor instance, how do i check what's the samplerate of a certain stream?
10:08.58lufkerio: pactl list
10:09.05kerioluf: anyway, how do we configure audio sink?
10:09.44lufkerio: :) Look into the history ...
10:12.34kerioanyway, it feels like your bluez is a bit less reliable :(
10:12.58kerioit makes no sense though, it's the same codebase
10:13.45keriomaybe it's just that i notice these annoyances more, though?
10:16.24kerioluf: why is your pulseaudio +0m6?
10:18.01lufkerio: I like feelings :)
10:18.13lufIt's easy to find the problem with it.
10:18.32lufkerio: pulseaudio because of some bt patch + new module?
10:19.08kerioluf: so?
10:19.15kerio+0m6 would be harmattan
10:19.31*** join/#maemo-ssu Pali (~pali@unaffiliated/pali)
10:19.32lufNo.
10:20.20lufIt's just version of the package not the maemo version as I understand from packages (changelogs).
10:20.28keriowhy do you think +0m5 is called m5?
10:21.04kerioluf: anyway, there *should* be no resampling in pulseaudio
10:21.17keriosink input from mafw is s16le 2ch 44100hz
10:21.26kerioand the sink output to my headphones is s16le 2ch 44100hz
10:24.50lufAnd what do you mean with less reliable - worse sound? disconnecting? lagging?
10:25.02keriodisconnecting sometimes
10:25.13keriobut as i said, maybe it used to happen before too and i just didn't focus on it
10:25.24lufIt's the same as with stock ...
10:26.21kerioapophenia is a bitch
10:29.34lufhttp://talk.maemo.org/showthread.php?t=51894
10:29.41luffor example ...
10:30.12lufBut I can't say if it's more/less often with stock/4.99
10:30.28kerioyeah, yeah
10:30.34keriothat's what i was guessing
10:34.13kerioluf: what would really be awesome would be an ICD plugin to do PANU in a simple way :3
10:37.37luf:D Sure.
10:37.41*** join/#maemo-ssu vi__ (~PSleggs@proxy.ei-info.com)
10:37.55lufIt should be nice project for you, isn't it?
10:38.36kerioluf: that would involve me doing things
10:39.29lufAs I saw you have a lot of time at uni ;)
10:41.15lufkerio: bluez - disconnecting. I need some debug logs and maybe also hcidump if you're able to collect it ....
10:42.07kerioluf: i checked the syslog (it wasn't doing debug output though) and it only said "bluetoothd: audio device disconnected" or something similar
10:42.54lufkerio: Are you sure you had bluetoothd in debug (-d '*') ?
10:43.08kerioas i said, i didn't have it in debug mode :)
10:43.50lufIt's useless without debug.
11:51.05kerioluf: the issue is a lot simpler than i thought, i think
11:51.50kerioit's a range issue
11:52.38kerioluf: keeping the connection open requires less quality than opening a new one, or reopening it after powersaving
11:52.51keriois there a way to pump up the tx power?
11:52.54kerio:3
12:08.11*** join/#maemo-ssu M4rtinK (~M4rtinK@ip-89-102-207-166.net.upcbroadband.cz)
12:18.46*** join/#maemo-ssu kolp (~quassel@212.255.28.120)
12:42.00lufkerio: I don't know. I'm not such good in hw ...
12:43.11lufkerio: Do you have some debug? Or from what you judged that it's txt power? BTW increasing tx power leads to higher battery usage.
12:43.42kerioluf: it connects fine when i'm holding the n900 vaguely close to my headphones, it doesn't connect when the n900 is in my pocket :)
12:45.55lufkerio: And are you sure the problem is on N900 side?
12:46.27kerioi think it's a mix of bad range on the n900 *and* the headphones
12:46.57lufThat's for sure true :D You didn't miss any option ...
12:47.13kerioto be fair the range on the n900's bt is fairly awful
12:49.49lufI have no problem in car ... even in the back part (for baggage).
12:50.05lufWhat's awful for you?
12:55.05lufkerio: can you next time you'll meet the bug with disconnected x icon and status connected try to run hcitool to see what BT is thinking? I need to know if it's a problem of bluez or just AIS or whatever ...
12:57.59keriok
12:59.07lufkerio: have you started working on icd plugin for bluez PANU client? :)
12:59.29kerionope
13:07.36*** join/#maemo-ssu dafox (~dafox@ip51cc571d.speed.planet.nl)
13:31.25DocScrutinizer05kerio: most likely BT-TXlevel is fixed
13:33.26DocScrutinizer05kerio: my N900 connects to sony headset even when meters away, and range seems like going thru 2 walls from living room to kitchen
13:33.41DocScrutinizer05(~ 8m)
13:33.44keriomaybe it's these headphones then
13:34.26kerioDocScrutinizer05: also, A2DP is a lot more heavy than HSP/HFP
13:34.39DocScrutinizer05if you don't have N900 alu case, it probably is the headset, yes
13:35.03DocScrutinizer05and I'm talking about listening to music, so A2DP
13:35.10kerioi have a plastic case, it shouldn't affect radio
13:35.29DocScrutinizer05depends on the plastic
13:44.27lufDocScrutinizer05: do you mean some alu plactic?
13:49.24*** join/#maemo-ssu Pali (~pali@unaffiliated/pali)
14:21.24DocScrutinizer05luf: nope
14:22.01DocScrutinizer05WLAN is known to not work nicely when some quite common case-plastic is too close to antenna, it gets detuned
14:22.11DocScrutinizer05BT has same freq
14:22.20DocScrutinizer05even shares same antenna
14:22.25DocScrutinizer05on some devices
14:23.32DocScrutinizer05normal windowglass is known to give GPS a -6 to -18dB
14:23.57DocScrutinizer05Pastic isn't any different in that regard, depending on the particular material and thickness
14:24.02DocScrutinizer05plastic*
14:25.16DocScrutinizer05the funny example to figure: PIR motion detectors don't work thru window glass at all
14:25.26DocScrutinizer05quite counter-intuitive
14:26.19DocScrutinizer05something being transparent or non-conductive doesn't mean it's negligible for RF
14:28.07kerioDocScrutinizer05: i remember my range sucking even before i had the otterbox, though
14:29.40DocScrutinizer05kerio: a lot of 'egoistic' WLAN AP in the neighborhood may spoil BT as well
14:30.04DocScrutinizer05mind you, they share same frequency
14:30.45DocScrutinizer05even your own AP might suffice to give BT a hard living
14:30.55kerioDocScrutinizer05: hm, i should try when i'm at my countryside house
14:31.06DocScrutinizer05good plan
14:32.51DocScrutinizer05there've been several reports that streaming media from internet via WLAN to N900 to BT headset doesn't work at all, since WLAN and BT have to 'co-exist' andthat can't work when both have to do a 100% job
14:41.16jon_yisn't wifi and bt spread spectrum?
14:42.56jon_yas long as it isn't too crowded, it should work regardless of frequency overlap
14:52.31keriostarts counting the visible wifi APs
14:53.21kerio20 plus hidden
14:53.27kerioso yeah
14:55.59DocScrutinizer05jon_y: BT is freq hopping afaik
14:56.13DocScrutinizer05WLAN isn't exactly spread spectrum
14:57.01DocScrutinizer05but if a sufficient number of WLAN AP mess up each channel available for BT to hop on, BT can't work properly
14:58.17DocScrutinizer05WLAN rather is TDMA, leaving open timeslots between pkgs sent
14:59.33DocScrutinizer05if several WLAN AP use same freq (which is chan +/- 2, due to sidebands), then it becomes more and more unlikely that BT finds such an open timeslot in that range of BT channels
15:01.55DocScrutinizer05in my book WLAN is already zombie state in real life, due to absolute overload / crowding of RF band
15:03.02DocScrutinizer05and every luser and his dog ramp up both bandwidth used (by N etc), as well as TX power to shout down the neighbor
15:04.27DocScrutinizer05since none of those lusers gets the catch that a proper cable is a million times better than crappy WLAN for connecting their PC to the router 30cm away
15:05.28DocScrutinizer05waits for frist DSL-routermodem to come with WLAN only, no more ethernet ports at all
15:08.56DocScrutinizer05thne they bitch about their crappy DSL connections while they don't realize it's actually the WLAN that doesn't go beyond 1 or 2 MBit, due to them themselves and all their neighbors basically 'abusing' WLAN for a cable replacement for distances of 2m between stationary devices
15:09.32DocScrutinizer05and that
15:09.47DocScrutinizer05's where BT has to stand a living
15:12.48*** join/#maemo-ssu arcean (~arcean@aacv61.neoplus.adsl.tpnet.pl)
15:14.13keriopoor bt :(
15:27.10merlin1991DocScrutinizer05: did you try the new testing already?
15:35.53merlin1991who else in here tried the pre release -testing?
15:36.21*** join/#maemo-ssu jonwil (~jonwil@27-33-137-199.static.tpgi.com.au)
16:14.08*** join/#maemo-ssu Timo (~timo@unaffiliated/tiempjuuh)
16:23.02vi__DocScrutinizer05: I have a solution.
16:23.23vi__I have a 5dB gain enternal aerial to my AP.
16:23.28vi__It works.
16:23.50vi__So I going to replace it with 3 9dB antennas.
16:24.04vi__There are no free channels near my home.
16:24.42vi__If you cannot beat them...DESTROY THEM WITH POWER.
16:45.18keriovi__: still won't work
16:45.50kerioyou can't magically put bigger antennas and expect your wifi network to work better
16:46.24kerioit's still going to be crowded as fuck
16:57.44vi__Yes, but if everyone else gets really shitty service, they might fuck off my channel.
17:39.10*** join/#maemo-ssu MrPingu (~MrPingute@85.133.27.6)
18:15.53*** join/#maemo-ssu MrPingu (~MrPingute@86.92.226.97)
18:43.20*** join/#maemo-ssu NIN101 (~NIN@p5DD28C9F.dip0.t-ipconnect.de)
18:47.13*** join/#maemo-ssu arcean (~arcean@aacv61.neoplus.adsl.tpnet.pl)
19:02.20*** join/#maemo-ssu arcean (~arcean@aacv61.neoplus.adsl.tpnet.pl)
19:20.56tadzikis it advised to optify package which consists of a single 148K binary + manpages?
19:22.33Paliyes
19:22.51tadzikokay
19:23.22Palibtw, autobuillder doing optification if debian/optify file has auto
19:24.06tadzikthat's the first time I hear about autobuilder
19:24.11tadzikis that something garage-related?
19:27.57Paliit is builder for extrad-devel
19:28.02Palibut now not working
19:28.18Pali(also maemo wiki not working)
19:33.32*** join/#maemo-ssu M4rtinK (~M4rtinK@46.227.13.46)
19:35.23*** join/#maemo-ssu arcean (~arcean@aacv61.neoplus.adsl.tpnet.pl)
19:43.15Palimerlin1991, freemangordon, can you test new battery applet?
19:45.48Paliand next question: What about creating new fiasco rootfs image with pr1.3.1 + cssu-stable and without some preinstalled apps (cheery, ...)
19:59.40tadzikwhat package is responsible for the behaviour of a longpress in, for example, file manager?
20:03.42kerioPali: it's going to be tough to do
20:04.01kerioPali: but go right ahead :3
20:05.57kerioPali: the ubifs image is going to be a bitch to work with
20:06.17Palikerio, I can generate ubifs image, no problem
20:06.24Paliand I can also unpack and edit it
20:06.37kerioPali: cool
20:06.52kerioi doubt the scripts will run nicely in a chroot jail
20:06.57keriobut you could do the things they do manually
20:07.13kerioi mean things like HAM
20:07.16Paliand I have patches for block2mtd.ko driver which creating mtd device from block
20:07.31Paliwhich allows me to mount ubifs image on computer without nandsim
20:07.31kerioyou'd have to run ICD at the very least for HAM to work
20:36.31*** join/#maemo-ssu toxaris (~toxaris@s83-180-246-172.cust.tele2.se)
20:55.33*** join/#maemo-ssu Jade (~jade@modemcable021.180-203-24.mc.videotron.ca)
20:55.33*** join/#maemo-ssu Jade (~jade@unaffiliated/jade)
21:07.37DocScrutinizer05merlin1991: sorry, not yet
21:09.07DocScrutinizer05though I could install it on a fresh device, I'd not be able to find the time to test it at all, since I'm a tad ill (flu)
21:09.58*** join/#maemo-ssu Jade (~jade@Jade.broker.freenet6.net)
21:10.10*** join/#maemo-ssu Jade (~jade@unaffiliated/jade)
21:10.13tadzikwho can I ask questions related to hildon codebase?
21:15.20wmaroneprobably best to just ask
21:16.02tadzikthat got no response :)
21:16.07tadzikwhat package is responsible for the behaviour of a longpress in, for example, file manager?
21:16.14tadzikI believe that's just the context menu
21:16.52merlin1991tadzik: arcean knows best, but he's currently not here
21:17.22tadzikI'll look around for him, thanks
21:23.04*** join/#maemo-ssu M4rtinK (~M4rtinK@46.227.13.46)
21:38.49*** join/#maemo-ssu M4rtinK (~M4rtinK@46.227.13.46)
21:45.04*** join/#maemo-ssu MrPingu (~MrPingute@86.92.226.97)
21:55.08*** join/#maemo-ssu M4rtinK (~M4rtinK@46.227.13.46)
21:57.40*** join/#maemo-ssu arcean (~arcean@aacv61.neoplus.adsl.tpnet.pl)
22:12.06*** join/#maemo-ssu M4rtinK (~M4rtinK@46.227.13.46)
23:07.52*** join/#maemo-ssu nox- (noident@freebsd/developer/nox)
23:12.29*** join/#maemo-ssu M4rtinK (~M4rtinK@ip-89-102-207-166.net.upcbroadband.cz)
23:26.54*** join/#maemo-ssu FIQ (~fiq@unaffiliated/fiq)
23:30.35FIQok so I have a strange problem with CSSU installation, HAM first load something, then seems to do nothing. Then after a while, it simply said "Process failed" (might not be exact wording due to localized msg) and there's no icon
23:31.44FIQI had a very similar issue last time I installed cssu (pretty much first thing I did that time, so no devel unless backupmenu required it, which I've forgot), but then there was an icon
23:33.04DocScrutinizer05tadzik: I think you're asking for mime-types
23:33.54FIQWhile CSSU seemed to install fine then, any install afterwards of anything reissued the "Please accept these TOU" of cssu-install
23:33.59DocScrutinizer05tadzik: http://wiki.maemo.org/Mime_Types#Using_hildon-mime-open
23:34.54FIQso basically both of my cssu installations has failed strangely, and both that time and now, it's pretty much clean rootfs
23:37.52DocScrutinizer05FIQ: that for sure needs investigation. Please wait til the knowledgeable people respond
23:38.35DocScrutinizer05particularly merlin1991 (and mohammadAG) might want to have a look.
23:38.40FIQI have used apt to install a few packages from extras+extras-devel (the following: bash xchat vgb gpspmgui coreutils-gnu findutils-gnu mlocate unzip) and some from HAM (backupmenu, openssh server+client) and I've changed default to /bin/bash. That's basically everything from factory state to now (might have forgot something, but nothing major)
23:39.03FIQs/default/default shell/
23:39.22DocScrutinizer05oh lol, changed default shell
23:39.43merlin1991FIQ: the cssu enabler mechanism runs a script that was never tested under bash
23:39.52FIQohhh
23:40.04FIQso making /bin/sh default should work?
23:40.13FIQor has things broken already?
23:40.25merlin1991yeah, though I have nfc in what state you are now thanks to the broken install
23:40.38FIQhm
23:40.49FIQmerlin1991: that can explain *this* fail
23:41.01FIQbut not the previous one, which gave strange problems too
23:41.18FIQand then everything I did was installing backupmenu and making a backup
23:41.29FIQ(from factory state, that is)
23:42.06DocScrutinizer05well, it seems nobody can reproduce this, so far
23:42.16FIQok
23:42.24FIQI'll try to switch shell and try again
23:42.36DocScrutinizer05if you actually can reproduce failure to install CSSU on a factory image ülus BM, we got a new problem
23:42.45merlin1991FIQ: did you get as far as runing the enabler script from the app menu?
23:42.57FIQmerlin1991: no
23:43.05FIQi got to step 5 on wiki
23:43.39FIQi.e. adding repo (I think), wait to HAM to do something
23:43.42merlin1991FIQ: ok time to get some info
23:43.47DocScrutinizer05indly wonders if CSSU repo is broken now too
23:44.08merlin1991please pastebin /etc/apt/sources.list.d/hildon-application-manager.list
23:44.35merlin1991also the output of dpkg -l community-ssu-enabler
23:45.28FIQnot found
23:45.36merlin1991the package or the file?
23:46.00DocScrutinizer05or pastebin?
23:46.02FIQhttp://pastebin.com/6caU7zZx is the sources list
23:46.11FIQmerlin1991: the package
23:46.38DocScrutinizer05community-ssu-enabler not installed
23:46.40merlin1991FIQ: ok then you basically only added the repo but did nothing else
23:47.06DocScrutinizer05for some reason HAM blew chunks on installing enabler
23:47.31FIQDocScrutinizer05: no, No packages found matching community-ssu-enabler.
23:47.32merlin1991well change your shell back to /bin/sh and try again
23:47.40FIQmerlin1991: ok
23:47.52merlin1991so far nothing happened to you system cssu wise :)
23:47.58FIQok :å
23:48.56FIQs/å/p/
23:49.25DocScrutinizer05>>HAM first load something, then seems to do nothing. Then after a while, it simply said "Process failed" (might not be exact wording due to localized msg) and there's no icon<< conclusive
23:49.57DocScrutinizer05might even be connection to repo
23:50.06DocScrutinizer05HAM log...
23:50.45DocScrutinizer05or "out of memory"?
23:50.51DocScrutinizer05sth...
23:50.56FIQehm
23:51.03FIQi'm very skeptical to OOM
23:51.14FIQconnection could be the issue
23:51.31FIQalso how are you supposed to know when HAM is done?
23:51.35DocScrutinizer05there are weird checks for available free storage in original HAM
23:51.47FIQit just sits there, but wiki says that's supposed to happen
23:52.46DocScrutinizer05yeah, that might actually be a bit unfortunate with HAM
23:53.26DocScrutinizer05basically as soon as you can do a new operation in HAM it's supposed to be finished so much that you can close it
23:53.27merlin1991FIQ: yeah you can run top in a terminal
23:53.38DocScrutinizer05afaik
23:53.40merlin1991as long as there is an apt-worker process eating your cpu ham is working :)
23:53.57DocScrutinizer05apt-worker keeps working even when you close HAM
23:54.02DocScrutinizer05so no problem there
23:54.21FIQgood to know
23:54.32DocScrutinizer05(unless our new brilliant HAM kills apt-worker)
23:55.15FIQalso what happens if you attempt to start the process again?
23:55.33FIQwhen it's started, that is
23:56.31DocScrutinizer05since it just installs cssu-enabler, there shouldn't be much harm
23:56.36FIQbecause IIRC I did that when I tried to install cssu the first time because it stalled
23:56.42DocScrutinizer05usual disclaimer: aiui
23:57.30FIQas I didn't realize it was normal behaviour for HAM to do apparently nothing but in fact do something
23:58.37DocScrutinizer05merlin1991: could we add a notifier to cssu-enabler post-install?
23:59.58FIQhow long should the process take?

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