IRC log for #maemo-ssu on 20140417

00:00.20DocScrutinizer05when every app is only another "server" connecting to the already running "client" browser process that renders and displays the GUI, then no surprise it's fast at startup
00:00.40DocScrutinizer05I say *when*, I have no clue how that stuff actually works internally
00:01.07okiasDocScrutinizer05: qt5 should be same as qt4, so classic app
00:01.24DocScrutinizer05ooh, so it's not QML ?
00:01.35okiasDocScrutinizer05: of course UI is abstracted little bit, but that seems file (different screen size, rotation)
00:01.40okiasit's QML :D
00:01.51okiass/file/fine/
00:01.55DocScrutinizer05aiui QML is basically HTML
00:01.59DocScrutinizer05more or less
00:02.00merlin1991the qt5 qml renderer is a totally new pipeline compared to the qt4 one
00:02.05merlin1991and also nope
00:02.17merlin1991qml != html on so many levels
00:02.31okiaswe run browser anyway.. there is not much use these days without browser :(
00:02.41DocScrutinizer05haha
00:02.44merlin1991the only thing they have in common is javascript
00:02.51okiasso, it can have html/xml parser already loaded in memory
00:02.56DocScrutinizer05OUUUUUUUUUuuuUUUCH
00:03.23okias;)
00:03.24merlin1991DocScrutinizer05: meaning javascript the language, not javascript the dom api
00:03.50DocScrutinizer05so it's a renderer anyway
00:04.05okiasbtw. did you guys tried measure, how much take gtk vs qt and pygtk vs pyqt on n900?
00:04.16DocScrutinizer05even with a matching whalecrap scripting language
00:04.22okiasI guess it running both at same time (as there is mix of qt/gtk apps)
00:04.25DocScrutinizer05aka "a browser"
00:04.26merlin1991well I never touched pyqt only pyside, and that is slooooooow
00:04.50okiaswell, that may be because python :D
00:04.58merlin1991nope
00:05.07merlin1991the thing that takes forever is the startup
00:05.24merlin1991which essentially is the linker resolving ALL the Qt symbols when you do import pyside ...
00:05.55merlin1991err not linker but loader, but you get the picture
00:07.11okiasget it
00:07.32okiaswell, I think write anything for low-memory mobile device in python is crazy
00:07.33DocScrutinizer05gotta love OO. Not only takes it ages to instantiate all objects at startup. even better: it takes almost as long (or even longer) on shutdown to run every ~destructor. vlosing a huge ram footprint Konqueror can take ages, while a kill -9 does same job in milliseconds
00:07.57bencoh:))
00:08.01okiasKonqueror, there is anyone who use that these days?
00:08.09DocScrutinizer05yes
00:08.50okiaswith kwebkit kpart?
00:09.18DocScrutinizer05with lots of goodies other shit never heard about
00:09.30bencohlike ?
00:09.42DocScrutinizer05like user defined services
00:10.16bencohwhat is it ?
00:10.34DocScrutinizer05yeah, I knew he's say that ;-P
00:11.35DocScrutinizer05mailto: is a service that's not exactly user defined
00:12.07DocScrutinizer05gg: for "http://google.com/q=$1" is user defined
00:12.47DocScrutinizer05prolly beats QML hands down X-P
00:14.56DocScrutinizer05then there are all those nice plugins for embedded display, like kate which even allows editing *and storing* files in provser
00:15.05DocScrutinizer05dang, *browser*
00:15.37DocScrutinizer05of course with syntax highlight and all
00:16.19DocScrutinizer05absolutely natural handling of e.g sftp://
00:17.10DocScrutinizer05I got my N900 filesystem in Konqueror as a bookmark
00:18.21DocScrutinizer05name: N900   location: sftp://root@iron900/root
00:21.57DocScrutinizer05one click and I'm renaming and deleting files on N900, and copying files and dirs to and from it to *any* other location, even my web server box at cloud-7.de
00:22.32DocScrutinizer05since, you guessed it, sftp://maemo.cloud-7.de is just another bookmark in konqueror
00:24.06DocScrutinizer05I created http://maemo.cloud-7.de/maemo5/session-log_enable-catalogs_README.txt via konqueror context menu "new -> textfile", then opened it in konqueror via "open with: kate"
00:24.30DocScrutinizer05kate, on ^S, saves the file directly to the webserver box
00:25.57DocScrutinizer05I admit a lot of that magic happens via kio_* daemons
00:27.27okiaswell, dolphin can do that out-of-box too :) it's all kio
00:27.31okiasand that's why I love KDE
00:27.33DocScrutinizer05I anyway can't see any other tool that would allow me to browse a remote sftp:// filesystem, and open files there in local editor, and save them to remote after editing
00:27.45DocScrutinizer05dolphin is retarded
00:27.53okiasI'd like see some kind of working KDE on N900, but that dream already passed :(
00:29.58DocScrutinizer05dolphin also can't render html files I click on
00:31.00DocScrutinizer05I guess dolphin *could* open FF to do that
00:32.09bencohtalking of whic ... any tried (the defunct) konqeror/embedded on the n900 ?
00:32.19bencohanyone*
00:32.31DocScrutinizer05never heard about
00:33.05DocScrutinizer05I'd guess on 800*480 - minus systray - it would be rather useless anyway
00:33.48DocScrutinizer05not that you couldn't tweak it to do some semi-useful work with it on such a screen
00:34.17bencohit used to have "kiosque" mode
00:34.27DocScrutinizer05the nice thing about KDE being that you can configure virtually *everything*
00:34.30bencohactually I tried it a long time ago on some lowres/MIPS device
00:35.11DocScrutinizer05time for killing nepomuk again... ;-P
00:35.45bencoh:]
00:36.32DocScrutinizer05http://privatepaste.com/4b9c2d2c4b
00:38.01DocScrutinizer05finally has to get around to setting up a cronjob for this
00:40.44DocScrutinizer05not that other crap works that flawless, regarding RAM footprint: http://wstaw.org/m/2014/04/17/plasma-desktopS17816.png
00:42.17dos1I just disabled nepomuk, as I don't use it anyway
00:42.55DocScrutinizer05breaks kmail, no more search, no addresses
00:43.20DocScrutinizer05#kde tells me "wait for kde4.13! Nepomuk is dead!"
00:43.33DocScrutinizer05\o/
00:43.37dos1oh; I didn't care as I don't use kmail as well ;)
00:43.52dos1yeah, they're replacing it with something simplier afaik
00:44.04DocScrutinizer05then OTOH knotify4 113MB RAM, *really* now? WHAT FOR?
00:44.49dos1baloo
00:44.51okiasDocScrutinizer05: 5.1
00:45.30DocScrutinizer05hmm?
00:45.53DocScrutinizer05a joke?
00:46.04okiasno, my Gentoo
00:46.09DocScrutinizer05oooh
00:46.18okiasI could probably run my whole full-featured KDE on N900 :D
00:46.29dos1knotify4 - 18.5 MB here
00:46.31okias4.12.6 now, updating to 4.13.0
00:46.40okias*updating right now
00:46.45DocScrutinizer05dos1: still too much
00:47.49DocScrutinizer05I think I will go blame PukeAudio
00:47.52dos17.6 MB of it shared
00:48.05dos1no pulseaudio here as well :)
00:48.11okiasDocScrutinizer05: Pulseaudio here
00:48.16DocScrutinizer05makes a lot of sense
00:48.25okiasbut still 5.1M ;)
00:48.49DocScrutinizer05yeah, when I start up session, it's like 1.5MB
00:48.54DocScrutinizer05or less
00:49.03DocScrutinizer05it "memleaks"
00:49.30okiasDocScrutinizer05: yes, that is big problem of knotify :(
00:49.32dos1yeah, probably
00:50.07dos1indeed, my session is pretty young atm - 18 days
00:50.31DocScrutinizer05watch CPU-time, I bet it's like 30MB per minute of CPU time
00:51.14DocScrutinizer05same with konqueror when running JS on a website
00:51.42dos129 min
00:51.52DocScrutinizer05after a few days you need to close the window or your RAM is getting a rare resource
00:52.10DocScrutinizer05WUT, yout knotify ate 29min of CPU?
00:54.52dos1s/min/sec/ - seems it's time to go to bed, I was pretty sure that dot was a colon :)
00:55.41DocScrutinizer05:-)
00:56.00DocScrutinizer05frowns at "Phonon backend: VLC"
00:56.20DocScrutinizer05really WTF? Phonon, my ass
00:56.37*** join/#maemo-ssu Milhouse (~Milhouse@Maemo/community/contributor/Milhouse)
00:56.37dos1gstreamer here
00:56.55dos1but I have one very annoying bug with phonon I couldn't tackle yet
00:56.57DocScrutinizer05I should try gstreamer
00:57.31DocScrutinizer05phonon is maximum useless crap
00:58.07dos1I have speakers connected to monitor, and providing sound to it via hdmi
00:58.40DocScrutinizer05a audio api wrapper around a soundserver for a stremer for a soundserver for a proper audio stack for the soundcard hardware
00:58.51dos1as I didn't want pulseaudio I have set up alsa device with dmix for hdmi
00:59.05DocScrutinizer05sounds reasonable
00:59.33DocScrutinizer05since PA is
00:59.37DocScrutinizer05~poettering
00:59.37infobot'sth is poettering' means it acts invasive, possessive, destructive, and generally in an egocentric exacerbating negative way. ``this cancer is extremely poettering''
00:59.47dos1and everything is fine, music plays, I can switch devices in phonon settings between hdmi and internal sound card and it works
01:00.18dos1but when I play music via phonon app, and some notification plays, the music switches to internal
01:01.08DocScrutinizer05friggin PA is absolutely useless as soon as audio config gets a tad more complex than "line out via standard codec"
01:02.09DocScrutinizer05in a *decent* system you'd define the correct ALSA device *inside* your mp3-player config, and profit
01:02.27dos1actually, I tried PA few times and never had problems with configuring it as I want
01:02.39DocScrutinizer05huh?
01:02.52DocScrutinizer05please, educate me, oh master!
01:03.10dos1but I *always* had some problems with implementation bugs, and I always rageuninstalled it after some time :P
01:04.08DocScrutinizer05I wouldn't even know how to configure PA so twinke ringtone goes to speakers audiocard while calls audio goes to headset
01:04.20DocScrutinizer05not even to think about phonon
01:04.43dos1for instance, using audacity with pulseaudio was always completely unreliable on my machine, no matter what I tried
01:05.02DocScrutinizer05actually I think I don't know of a single app that has PA config settings
01:05.39dos1well, in PA you're rather supposed to use mixer and its global settings for things like that
01:05.55dos1you can configure per-application streams to go to specific device
01:05.57DocScrutinizer05and honestly, it completely eludes me what for you need PA anyway
01:06.51DocScrutinizer05I can't see how I would configure mixer / PA for an app that's not recently playing audio
01:07.33DocScrutinizer05that aside, even when it would be possible, it feels idiotic
01:07.59okiasbtw. did you seen this? http://www.etalabs.net/compare_libcs.html
01:08.09okiasthis maybe useful for n900
01:08.40*** join/#maemo-ssu RST38h (marat@wsip-184-180-40-182.ri.ri.cox.net)
01:08.44dos1I always liked it's network-transparent approach, I was trying few times to get seemless audio switch between n900 and laptop, or seamless streaming from raspberry pi
01:09.51dos1but even though it was marketed as a great feature of pulseaudio when it was being introduced everywhere, it's obviously forgotten and second-tier and also, I never managed to get it right thanks to bugs ;P
01:10.08dos1some of them were already reported for a looong time
01:10.29DocScrutinizer05and ALSA is most likely better on that too
01:10.49okiasoh, I found on forum that Skry tried musl on n900
01:10.55dos1but tbh it's been a while since I tried pulseaudio the last time
01:11.04dos1so maaaaaybeeeee it's a bit better now
01:11.06dos1maybe :P
01:11.12DocScrutinizer05the *only* thing that ALSA admittedly can't do OOTB is seamless switching from one device to another
01:11.36DocScrutinizer05but you can really ask if that's a flaw or rather a feature
01:12.53DocScrutinizer05after all in virtually every case it's desirable that the *app* gets to know about such change in audio sink and the associated properties
01:13.51DocScrutinizer05and in real live there's no such thing like "seamless switching of audio between two sinks"
01:14.19DocScrutinizer05you *always* get some discontinuity
01:14.47DocScrutinizer05unless you do proper crossfade between two alternative cables to same amp
01:15.23DocScrutinizer05which I for the life of mine can't come up with any RL usecase for
01:16.53DocScrutinizer05even PA must fill buffers of the new audiocard to use as sink
01:18.19DocScrutinizer05in ALSA I suggested a way to handle such "seamless handover" between two defined and configured audio sinks. You simply split the stream and feed it to both sinks, and you mute one of them
01:19.38DocScrutinizer05for "handover" you simply unmute the "new" muted one, and mute the "old" one
01:20.57DocScrutinizer05per-app volume controls in mixer: no problem with ALSA. Just with ALSA they are persistent, so you can adjust them *before* your audio starts playback
01:22.10*** join/#maemo-ssu drathir87 (~kamiljk8@s51.linuxpl.com)
01:22.15DocScrutinizer05mixing concurrent audio? dmix did it before Poettering even started to wonder if it works, and it actually worked better at that time than PA does nowadays
01:23.28DocScrutinizer05latencies? TZZZ, every ALSA plugin has absolutely defined latency, easy to calculate, you don't even need any code to do this at runtime
01:24.38DocScrutinizer05but Poettering went "DAMN! how do I query latency of the ALSA stack? No function call to do that! I need to invent something where you actually can query the latency, since otherwise how is shit supposed to work??"
01:25.41DocScrutinizer05result: sometimes I had latencies of 46 HOURS in PA - no kidding!
01:27.27dos1I can see some value in PA-style handling of devices and handovers - with PA it is (should be) easy to just connect random usb headphones and instantly switch everything to it
01:27.46dos1with pure alsa, especially when you didn't fine-tune your configs, it might be not so easy
01:29.59dos1but it's not enough to make me want to struggle with pulseaudio and its issues one after another :P
01:30.30DocScrutinizer05instantly switch everything to it, and enjoy earshot when next notification comes in, thanks to no way to configure volume beforehand
01:30.44DocScrutinizer05;-P
01:31.11DocScrutinizer05that's a moot argument
01:31.41dos1any time I gave pulseaudio a chance I disabled that brain damaged windows-style volume mangling feature
01:32.06DocScrutinizer05you need proper set up for quite a number of things (like kernel drivers) anyway, or otherwise your USB headset will do nothing at all
01:32.39dos1I'm usually running all the apps at maximum volume and just adjusting the main controls
01:32.40DocScrutinizer05sorry, please don't use alien idioms
01:32.43DocScrutinizer05;)
01:33.22DocScrutinizer05what's a "windows-style volume mangling feature"?
01:33.50dos1I hate that "feature" where when I have music at say, 30%, and main volume at 50%, and then when I increase the volume of music to 80% main volume increases too
01:34.01dos1that's default pulseaudio behaviour
01:34.10DocScrutinizer05OMG
01:34.51dos1by default, each app-volume is relative to 100% of the device max
01:35.16dos1I always disable that, so each app-volume is relative to the volume capped by main volume control
01:36.00dos1so no earshot notifications for me, as I usually keep all app-volumes at 100% ;)
01:37.21dos1and well, usb headsets seem to be handled well by udev etc. these days
01:37.42DocScrutinizer05but you still dunno if your USB hs is maybe 10 times louder than your current speakers
01:37.52dos1it just worked
01:38.26dos1yep, but I never put headphones on my head before music starts playing
01:38.57dos1no matter if I do handover or just start to play ;)
01:39.09DocScrutinizer05and claiming udev handles USB hs great is just moot invalid argument, since you can as well say "ok, when somebody would just take a bit of care then ALSA config was equally working OOTB for USB hs"
01:39.43dos1of course
01:40.18dos1the thing is, for some reason noone took a bit of care for having that out of box in popular distros
01:40.33dos1before Poettering, that is
01:40.43dos1and now we have to deal with his crap :P
01:41.09dos1(that is, uninstalling and reconfiguring audio after default distro installation ;))
01:41.49dos1ok, I'm out, good night o/
01:42.20DocScrutinizer05and honestly, writing a PCM plugin for ALSA that would protect the app from a device-vanished error coming up from down the ALSA stack, and rather simply opens same stack/device again, so another plugin further down could open another alternative audiocard when you unplug or unload the current one, that's no rocket science and whould've made PA completely redundant dispensible crap from very beginning
01:42.50DocScrutinizer05n8 dos1
01:45.26DocScrutinizer05but IIRC ~poettering 's main concern that lead to invention of PA been the "unpredictable latency" of ALSA - a complete aberration
01:46.52DocScrutinizer05plus the allegedly not working or even missing mixing, dunno if he never heard of dmix or simply was tricked into thinking ALSA didn't know to mix because some distros had dmix not configured in default stack
01:47.18DocScrutinizer05iirc dmix per default came as late as ALSA0.9
01:47.27DocScrutinizer05but it existed way longer
02:01.55*** join/#maemo-ssu silviof1 (~silviof@unaffiliated/silviof)
02:09.11*** join/#maemo-ssu wmarone_ (~wmarone@c-24-10-24-149.hsd1.ca.comcast.net)
02:25.40*** join/#maemo-ssu amiconn_ (quassel@rockbox/developer/amiconn)
04:34.42*** join/#maemo-ssu LauRoman (~LauRoman@5-14-87-52.residential.rdsnet.ro)
04:41.07*** join/#maemo-ssu amizraa (~amizraa@gateway/tor-sasl/amizraa)
04:44.00*** join/#maemo-ssu radekp (~radek@82.113.39.213)
05:39.33*** join/#maemo-ssu _rd (~rd@p5088F333.dip0.t-ipconnect.de)
05:43.10*** join/#maemo-ssu amiconn (amiconn@rockbox/developer/amiconn)
06:53.23*** join/#maemo-ssu luf (~luf@80.188.29.62)
07:10.30*** join/#maemo-ssu sunny_s (~sunny_s@business-092-079-020-027.static.arcor-ip.net)
07:10.51ApicA wonderful nice splendid gorgeous Thursday Morning (UGT) on this fine Boomtime!
07:56.52*** join/#maemo-ssu delphi (~ns-team@devbin/founder/trx)
08:27.44*** join/#maemo-ssu NIN101 (~core@n900.quitesimple.org)
09:03.29*** join/#maemo-ssu kolp (~quassel@55d4663d.access.ecotel.net)
09:23.02*** join/#maemo-ssu MrPingu (~MrPingute@ip503c83e7.speed.planet.nl)
09:37.54merlin1991Apic: :)
09:38.03merlin1991also morning
09:38.21ApicBpO]]
09:38.26ApicB-) even
09:38.29Apic2u2
09:57.17*** join/#maemo-ssu LauRoman (~LauRoman@5-14-87-52.residential.rdsnet.ro)
10:07.03*** join/#maemo-ssu arcean (~arcean@aacv219.neoplus.adsl.tpnet.pl)
11:03.54*** join/#maemo-ssu lizardo (lizardo@nat/indt/x-cibdjzqllifobjrj)
11:14.28*** join/#maemo-ssu okias (~okias@berger.cust.centro-net.cz)
12:26.29*** join/#maemo-ssu _rd (~rd@p5088F333.dip0.t-ipconnect.de)
12:50.41*** join/#maemo-ssu NishanthMenon (~nmenon@nat/ti/x-qysmkeyrlnmayinc)
13:45.55*** join/#maemo-ssu sixwheeledbeast (~paul@cl-1547.lon-02.gb.sixxs.net)
13:46.32*** join/#maemo-ssu arcean_ (~arcean@aaep119.neoplus.adsl.tpnet.pl)
13:52.31*** join/#maemo-ssu _rd (~rd@p5088F333.dip0.t-ipconnect.de)
14:09.41*** join/#maemo-ssu BCMM (~BCMM@unaffiliated/bcmm)
15:08.49*** join/#maemo-ssu scoobertron (~tom@66.172.11.27)
16:00.44*** join/#maemo-ssu javispedro (~javier@69.Red-80-32-146.staticIP.rima-tde.net)
16:00.44*** join/#maemo-ssu javispedro (~javier@Maemo/community/contributor/javispedro)
17:00.03*** join/#maemo-ssu sixwheeledbeast (~paul@host-92-15-167-6.as43234.net)
17:04.48*** join/#maemo-ssu sunny_s (~sunny_s@business-092-079-020-027.static.arcor-ip.net)
17:09.03*** join/#maemo-ssu sixwheeledbeast1 (~paul@host-92-29-204-252.as13285.net)
17:32.58*** join/#maemo-ssu M4rtinK (~M4rtinK@ip-78-102-146-131.net.upcbroadband.cz)
17:35.22*** join/#maemo-ssu okias (~okias@berger.cust.centro-net.cz)
17:58.48*** join/#maemo-ssu sixwheeledbeast (~paul@host-80-41-221-137.as13285.net)
18:05.16*** join/#maemo-ssu _rd (~rd@p5088F333.dip0.t-ipconnect.de)
18:22.01*** join/#maemo-ssu sixwheeledbeast1 (~paul@2a01:348:6:860a:94de:190b:dd1:4b93)
18:36.22*** join/#maemo-ssu nox- (noident@freebsd/developer/nox)
18:53.37*** join/#maemo-ssu _rd (~rd@p5088F333.dip0.t-ipconnect.de)
19:56.54*** join/#maemo-ssu xes (~xes@unaffiliated/xes)
20:39.22*** join/#maemo-ssu trx (~ns-team@212.200.199.163)
20:39.22*** join/#maemo-ssu trx (~ns-team@devbin/founder/trx)
20:40.04*** join/#maemo-ssu dos1 (~dos@unaffiliated/dos1)
20:58.58*** join/#maemo-ssu LauRoman (~LauRoman@79.116.44.28)
23:06.58*** join/#maemo-ssu LauRoman (~LauRoman@5-14-85-30.residential.rdsnet.ro)
23:28.14*** join/#maemo-ssu amizraa (~amizraa@gateway/tor-sasl/amizraa)

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