Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2011-03-11

00:09:19 Quit bertrik (Quit: :tiuQ)
00:09:24 Quit n1s (Quit: Ex-Chat)
00:29:05 Quit Zagor_ (Quit: Clint excited)
00:31:38***Saving seen data "./dancer.seen"
00:41:02 Quit sirrozha ()
00:42:09 Quit froggyman (Quit: Ex-Chat)
00:42:14 Join milk [0] (
00:45:18 Quit sideral (Quit: Leaving.)
00:49:18*pixelma wonders what the volume keys now do on Android
00:52:25JdGordonapparently change the volume
00:54:14pixelmathe discussion before was about the volume range
00:54:26JdGordonZagor: thats not really fixing fade-out :) why the heck isnt fade done in the playback engine or dsp? it is really out of place in wps.c
00:56:06 Join FoH [0] (
00:56:25 Quit Jerom (Quit: Leaving.)
00:59:53 Quit kadoban (Ping timeout: 255 seconds)
01:02:26 Quit Elfish (Ping timeout: 248 seconds)
01:03:16 Quit ranmachan (Read error: Operation timed out)
01:03:23 Join ranmachan [0] (
01:05:57jhMikeSJdGordon: that's a deep rabbit hole :)
01:06:07 Join kadoban [0] (
01:07:17 Join Elfish [0] (amba@2a01:4f8:100:90a1:abc:abc:abc:abc)
01:17:50 Join froggyman [0] (~seth@
01:17:51 Quit froggyman (Changing host)
01:17:51 Join froggyman [0] (~seth@unaffiliated/froggyman)
01:23:29scorcheB4gder: ping
01:27:47 Quit kevku (Ping timeout: 260 seconds)
01:31:37 Quit dfkt (Quit: -= SysReset 2.53=- Sic gorgiamus allos subjectatos nunc.)
02:02:29pixelmascorche: it looks like the akismet caught a "this topic has been moved to" post of mine
02:03:02pixelmamy impression is that it doesn't like links in posts in general and is a bit overcautious
02:04:43 Quit robin0800 (Quit: Leaving)
02:05:11pixelmadamnit, I hope the poster gets an email notification - or doesn't look for it in the old place :\
02:05:21pixelmauntil sorted
02:07:10 Quit timccc (Read error: Connection reset by peer)
02:08:48JdGordonyeah, akismet is either setup slightly wrongly or just sucks on SMF...
02:08:57jhMikeSwhy can't I find that really old pre-mpegplayer plugin patch - I seem to be mentally challenged over and over about locating it :\
02:09:28 Join timccc [0] (~timccc@
02:10:06JdGordonscorche: I see you're editing the /GSoCOrgApp2011 page... changing the title is probably a good idea :)
02:10:54scorchethere isnt really a "set up" for akismet on SMF
02:11:05scorcheit is just over-cautious
02:11:39scorchethough it should tell the poster that they were flagged and will be sorted out when an admin clears their post - i need to work on that
02:12:04scorchei can always disable akismet, though it has been catching plenty of spam
02:12:12JdGordonwell, ideally it shouldnt be sending reply notifications out until it gets through akismet
02:12:20pixelmait would be cool if it could "learn" about trustworthy members or at least user groups
02:12:28JdGordonand seeing the flagged reply in the reply window is also funny
02:13:00scorchepixelma: well, each time i hit "not spam" it *should* learn something
02:13:06saratogahow does it decide what to flag?
02:13:15pixelmaor allowed links - everything or so
02:13:16scorcheJdGordon: yeah - that is likely part of its crappy implementaton in SMF
02:13:27saratogalike why does it flag all of my replies to one thread, but not replies in other threads
02:13:40scorchepixelma: i cant really specify such things since akismet is a "cloud" thing
02:13:49scorchesaratoga: because it doesnt like you
02:13:56pixelmathe flagging of redirection posts surprised me now
02:15:11pixelmathanks for taking care of that so quickly
02:17:29 Join Judas_PhD [0] (
02:31:05 Quit mshathlonxp (Quit: Leaving)
02:31:39***Saving seen data "./dancer.seen"
02:34:07 Quit simonrvn (Read error: Operation timed out)
02:34:11 Join robin0800 [0] (
02:38:05 Quit robin0800 (Client Quit)
02:40:04 Join robin0800 [0] (
02:40:34 Join simonrvn [0] (simon@2001:470:8c85:11fe::c0a8:195)
02:41:58 Quit cjcopi (Read error: Operation timed out)
02:50:33 Quit MethoS- (Remote host closed the connection)
02:54:01 Join mystica555_ [0] (
02:54:43 Quit komputes (Remote host closed the connection)
03:00:06 Join cjcopi [0] (
03:02:40 Join GodEater_ [0] (
03:02:40 Quit GodEater_ (Changing host)
03:02:40 Join GodEater_ [0] (~bibble@rockbox/staff/GodEater)
03:04:57 Quit GodEater (Ping timeout: 240 seconds)
03:05:09scorcheas a reminder, if you are planning to be involved in GSoC (as a mentor capacity - or just want some say in it) - join #rockbox-gsoc
03:05:32 Quit simonrvn (Read error: Operation timed out)
03:06:53 Quit krabador (Quit: Sto andando via)
03:10:08 Join simonrvn [0] (simon@2001:470:8c85:11fe::c0a8:195)
03:13:57 Quit froggyman (Ping timeout: 240 seconds)
03:20:18 Join audio-i [0] (
03:23:57 Quit cjcopi (Ping timeout: 240 seconds)
03:26:47 Join froggyman [0] (~seth@
03:26:47 Quit froggyman (Changing host)
03:26:47 Join froggyman [0] (~seth@unaffiliated/froggyman)
03:27:11audio-ihi! if I was to use %?if to display a different viewport when current track time %pc is below 60 minutes, what is the time format I should use for the operand?
03:28:05JdGordonsomething like %?if(%pc, <, 3600)<>
03:28:36JdGordontime is in seconds im pretty sure
03:28:36audio-iI think I tried that and it didn't work... trying again
03:28:50JdGordoni cant check the code now though
03:29:38JdGordonactually... its prob a string so your shit out of luck
03:30:16audio-iyeah it seems so
03:30:59audio-ionly > O seems to return something usable
03:31:11audio-iwhat does it mean it's a string
03:32:06audio-ijust characters? not a number I guess therefore no > < etc would work
03:32:30 Join pjm0616 [0] (
03:32:31 Quit kadoban (Quit: bye)
03:32:49 Join kadoban [0] (
03:32:56JdGordonit just displays text
03:33:08JdGordonfile a bug that %pc isnt useable with %if
03:33:37audio-iok will do ;)
03:33:55audio-inot excatly requesting it :) but in this like other cases
03:34:14audio-iI find a "character length" tag would be very useful
03:35:32audio-ilike %cl on "rockbox" would return 7
03:35:48audio-ii think %cl is taken though
03:36:03JdGordonyeah, but how would that be useful?
03:36:20 Quit milk (Read error: Connection reset by peer)
03:36:36audio-iIn this case, if %pc is greater than 5, I would use a different viewport
03:37:11audio-ilaso useful in "controlled scrolling" routines
03:37:51audio-ior I could even use a bigger font if the number of characters allowes me to
03:38:16audio-iactually the ipod classic has that kind of effect
03:38:34audio-iwhen the title is long, it uses smaller fonts
03:39:02 Quit froggyman (Ping timeout: 260 seconds)
03:39:21audio-iavoids scrolling
03:45:05 Quit DerPapst (Quit: Leaving.)
03:50:22 Join cjcopi [0] (
03:50:39 Quit robin0800 (Quit: Leaving)
03:55:47 Quit balintx (Remote host closed the connection)
03:56:08 Join balintx [0] (
03:58:22audio-idamn! I thought I'd found another way of making it but it didn't work
03:59:29audio-iJdGordon: I was trying to use %pS with 3600 but it seems there is something like a 320 limit?
03:59:43JdGordoni doubt that
04:00:22audio-iwell, i set it to 3600 but it goes false at 5 min 20 sec
04:00:42JdGordonthe track is how long?
04:00:55audio-i2 hours something
04:01:18JdGordonhmm.... this could be 32bit overflowing maybe...
04:01:43JdGordon5:20 is how many milliseconds?
04:02:42JdGordonunsigned long time = token->value.i * (HZ/TIMEOUT_UNIT); <- /me cant do maths
04:04:13audio-iseems like you are right... and I'm certainly shit out of luck :/
04:04:35JdGordonmaking %pc do what you want should be relativly easy
04:04:58JdGordonis that track length? or remaining time?
04:05:11audio-icurrent track time
04:05:32audio-ielapsed time
04:07:00 Quit JdGordon (Quit: leaving)
04:09:18 Join JdGordon| [0] (
04:09:19 Quit JdGordon| (Changing host)
04:09:19 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
04:10:12 Quit z35 (Quit: Leaving)
04:10:48audio-ithanks! will try it later ;) somebody asked something of retroTape on the forums that could be resolved with this
04:13:17 Quit amiconn (Disconnected by services)
04:13:17 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:13:18 Join Llorean1 [0] (
04:13:20 Quit pixelma (Disconnected by services)
04:13:22 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:13:24 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:13:37 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:15:36 Quit Judas_PhD (Quit: This is a quitting message)
04:16:25 Quit Llorean (Ping timeout: 246 seconds)
04:18:35 Quit audio-i (Quit: CGI:IRC (EOF))
04:21:52 Join froggyman [0] (~seth@unaffiliated/froggyman)
04:22:09 Join Judas_PhD [0] (
04:23:28 Join notaguest [0] (
04:24:10 Quit notaguest (Client Quit)
04:28:49 Quit Dreamxtreme (Ping timeout: 255 seconds)
04:31:41***Saving seen data "./dancer.seen"
04:33:24 Join tails___ [0] (
04:34:42 Join kugel_ [0] (~kugel@rockbox/developer/kugel)
04:35:57 Quit tails__ (Ping timeout: 250 seconds)
04:37:49 Quit kugel (Ping timeout: 255 seconds)
04:39:30 Quit mystica555_ (Read error: No route to host)
04:41:55 Quit TheSeven (Ping timeout: 276 seconds)
04:46:05 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:47:58 Join mystica555_ [0] (
04:54:14 Join Barahir_ [0] (
04:57:38 Quit Barahir (Ping timeout: 276 seconds)
05:03:10 Quit ps-auxw (Ping timeout: 260 seconds)
05:05:26 Join Rob2223 [0] (
05:08:56 Quit Rob2222 (Ping timeout: 252 seconds)
05:12:10 Nick Llorean1 is now known as Llorean (
05:12:16 Quit Llorean (Changing host)
05:12:16 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
05:13:36 Join Dreamxtreme [0] (~Dre@
05:14:09 Join ps-auxw [0] (
05:14:46 Quit ps-auxw (Client Quit)
05:26:00 Quit simonrvn (Read error: Operation timed out)
05:30:21 Join simonrvn [0] (simon@2001:470:8c85:11fe::c0a8:195)
05:37:30 Join ps-auxw [0] (
05:41:42 Quit Judas_PhD (Quit: This is a quitting message)
05:50:00 Quit simonrvn (Read error: Operation timed out)
05:50:17 Join JdGord [0] (
05:50:38 Quit JdGord (Changing host)
05:50:38 Join JdGord [0] (~3a601b9e@rockbox/developer/JdGordon)
05:53:49 Quit JdGord (Client Quit)
05:58:49 Join JdGord [0] (
05:58:58 Quit JdGord (Client Quit)
06:00:08 Join JdGord [0] (
06:00:16 Quit JdGord (Client Quit)
06:00:19 Join simonrvn [0] (simon@2001:470:8c85:11fe::c0a8:195)
06:16:00 Quit kadoban (Read error: Connection reset by peer)
06:17:48 Quit simonrvn (Quit: see ya - n'multes)
06:21:03 Join simonrvn [0] (simon@2001:470:8c85:11fe::c0a8:195)
06:31:44***Saving seen data "./dancer.seen"
06:33:09 Join Judas_PhD [0] (
06:33:50 Quit Judas_PhD (Remote host closed the connection)
06:35:43 Quit simonrvn (Quit: see ya - n'multes)
06:40:05 Join Judas_PhD [0] (
06:40:17 Join simonrvn [0] (simon@2001:470:8c85:11fe::c0a8:195)
06:41:28 Join leavittx_ [0] (~lev@
06:50:55 Quit liar (Read error: No route to host)
06:53:45 Join Horscht [0] (~Horscht@xbmc/user/horscht)
06:56:21 Quit Horschti (Ping timeout: 246 seconds)
06:57:05 Join liar [0] (
07:00:30 Quit panni_ (Quit: ( :: NoNameScript 3.81 :: ))
07:03:35 Quit T44 (Read error: Connection reset by peer)
07:03:36 Join Topy [0] (
07:14:40 Quit liar (Ping timeout: 276 seconds)
07:34:41 Join MagusG [0] (
07:35:24 Quit bluefoxx_ (Quit: Can we, should we, will we?)
07:39:05 Join bluefoxx [0] (
07:44:05 Quit bluefoxx (Quit: Can we, should we, will we?)
07:44:43 Join bluefoxx [0] (
07:46:13 Join bthomson [0] (
07:50:37 Join kevku [0] (~kevku@2001:7d0:0:f9af:feed:feed:feed:feed)
08:04:30S_a_i_n_tJdGordon|: Is there a way to split lines in a .wps/.sbs/.fms/etc so you don't end up with some lines that are ~200 columns long?
08:05:15S_a_i_n_tI reduced my ASCII header to 80 char, kinda seems pointless to do so with truly massive lines in the skin code.
08:07:45S_a_i_n_tI was just wondering if there ws some magic tag for newline that's smart enough to stick the next line on the end of the previous line when it gets parsed or something, so you could have nice 80 char wide themes so everything looks nice and neat.
08:12:14JdGordon|# should do it
08:19:32 Quit simonrvn (Read error: Operation timed out)
08:21:19 Quit bluebrother (Read error: Operation timed out)
08:21:34 Join simonrvn [0] (simon@2001:470:8c85:11fe::c0a8:195)
08:21:52S_a_i_n_tWha? To split a line of actual code? so:
08:23:37 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
08:24:18JdGordon|S_a_i_n_t: yeah
08:24:31 Join mudd1 [0] (
08:25:05JdGordon|well... there has been plenty of debate over weather # should eat the \n or not...
08:25:10JdGordon|iirc svn has it eating it
08:26:55 Quit sasquatch (Ping timeout: 250 seconds)
08:27:56 Join LinusN [0] (~linus@rockbox/developer/LinusN)
08:31:46***Saving seen data "./dancer.seen"
08:35:21 Join ender` [0] (
08:37:04 Part LinusN
08:37:33 Join LinusN [0] (~linus@rockbox/developer/LinusN)
08:39:01 Part LinusN
08:42:11 Join LinusN [0] (~linus@rockbox/developer/LinusN)
08:45:41 Quit simonrvn (Read error: Operation timed out)
08:49:36 Join simonrvn [0] (simon@2001:470:8c85:11fe::c0a8:195)
08:51:12 Join sasquatch [0] (~username@
09:04:10 Join _Zagor [0] (~bjst@rockbox/developer/Zagor)
09:04:22 Nick _Zagor is now known as Zagor (~bjst@rockbox/developer/Zagor)
09:05:25 Quit sasquatch (Ping timeout: 252 seconds)
09:06:47 Join sasquatch [0] (
09:26:16 Join sideral [0] (~sideral@rockbox/developer/sideral)
09:27:55 Join einhirn [0] (
09:29:45 Quit leavittx_ (Ping timeout: 240 seconds)
09:29:47 Join esperegu [0] (~quassel@
09:32:06 Join pamaury [0] (
09:32:06 Quit pamaury (Changing host)
09:32:06 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:33:30 Quit Elfish (Ping timeout: 248 seconds)
09:35:35 Quit ranmachan (Ping timeout: 260 seconds)
09:35:49 Quit literal (Ping timeout: 276 seconds)
09:35:54 Join ranmachan [0] (
09:37:16 Join literal [0] (
09:44:08 Join Elfish [0] (~amba@2a01:4f8:100:90a1:abc:abc:abc:abc)
09:46:11 Quit mudd1 (Ping timeout: 250 seconds)
09:47:39 Quit mystica555_ (Ping timeout: 264 seconds)
09:48:32 Join n1s [0] (
09:48:33 Quit n1s (Changing host)
09:48:33 Join n1s [0] (~n1s@rockbox/developer/n1s)
09:52:34 Join mystica555_ [0] (
10:01:55 Quit sasquatch (Quit: WeeChat 0.3.2)
10:02:07 Join sasquatch [0] (
10:02:21 Quit S_a_i_n_t (Ping timeout: 276 seconds)
10:03:24 Join [Saint] [0] (S_a_i_n_t@
10:15:14 Quit Slasheri_ (Ping timeout: 250 seconds)
10:29:36 Quit jordan` (Quit: Coyote finally caught me)
10:29:56 Join jordan` [0] (
10:31:49***Saving seen data "./dancer.seen"
10:35:30 Quit JdGordon| (Ping timeout: 276 seconds)
10:35:41 Nick kugel_ is now known as kugel (~kugel@rockbox/developer/kugel)
10:36:01 Join JdGordon| [0] (
10:36:01 Quit JdGordon| (Changing host)
10:36:01 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
10:42:25 Join robin0800 [0] (
10:43:39 Quit JackWinter (Remote host closed the connection)
10:46:32 Join Slasheri [0] (
10:46:33 Quit Slasheri (Changing host)
10:46:33 Join Slasheri [0] (miipekk@rockbox/developer/Slasheri)
10:47:34pixelmaZagor: what will the volume keys now do on Android (regarding the volume steps and range)?
10:49:00Zagorit adjusts the rockbox volume
10:49:54Zagorbefore my commit, the volume buttons adjusted the *ringtone* volume so it's not like we're losing anything
10:50:34Zagorthough we should add proper android media volume adjustment asap
10:50:36kugelugh, apparently you need to add something to debian/changelog to bump the debian version of a package
10:50:55kugelffs, sorry
10:51:27pixelmaso I set the global volume (which affects the maximum available) outside Rockbox and then the percentage (or whatever it is) in Rockbox with the difference to now being that you can use the volume keys too instead of touchscreen things or the menu?
10:51:54kugelZagor: no, they didn't
10:51:55 Join JackWinter [0] (
10:52:18kugelthey changed the volume of the music if music was playing. that's just how it works on android
10:53:21Zagoroh, my mistake. though I do find it quite weird that "rockbox" adjusts ringtone volume just because I happened to be in pause
10:53:23pixelmabut the global volume (and I seem to remember that it showed the "ringtone volume" splash too)
10:53:45ZagorI suppose that is due to our disabling the service in pause?
10:54:06kugelyes, the global "media volume". I now need to go out of rockbox to change the global volume with the volume keys?
10:54:57Zagorat the moment, yes. we have to fix proper volume handling.
10:55:06kugelthen I object this change
10:55:06ZagorI'll look into that
10:55:22Zagorwe don't want two volume controls, it's just silly
10:56:03kugelI don't want different behvaior just because rockbox is in the foreground
10:56:10[Saint]glargh....I liked it :/
10:56:31Zagorkugel: so you don't want the rockbox high resolution volume?
10:56:47kugelI do, but that's not what you have committed
10:56:50Zagorwe had a long discussion about this yesterday, and several people felt the 16-step android volume is not sufficient
10:57:00pixelmait's confusing indeed and I'm looking forward to a better solution, if android would at least offer more fine grained control at least for media (I can even see the reason to have quicker controls for call volume)
10:57:12kugelI was part of that discussion, yes
10:57:36pixelma[Saint]: "it"?
10:58:00kugelIMO you just made it worse, there's still 2 volumes, but now there's inconsistent behavior
10:58:31[Saint]pixelma: it == having independant RB/Android volume control.
10:58:32Zagorthat's one way to view it. I'll try to improve it asap.
10:59:23kugelcan we get back to the old behavior until you improved it?
10:59:34pixelma[Saint]: you still have, just not as "easily" accessible
11:00:29Zagorkugel: why? it's not like we are distributing builds. just cut it out of your build if you can't stand it for a few hours/days
11:00:39Zagorit's a work in progress
11:00:40 Quit kevku (Read error: Operation timed out)
11:00:54kugelI will
11:02:23kugelI still think your commit was premature. "commit early, commit often" is fine but not if the intermediate commits make things just worse
11:02:59Zagorthat's not objective. I made android behave more like other targets, in that the volume buttons adjust the rockbox volume.
11:03:11Zagorif that is better or worse is a matter of opinion
11:03:58 Quit robin0800 (Quit: Leaving)
11:04:46kugelvolume handling just isn't the same thing on RaaA
11:05:19ZagorI know. but it should be.
11:06:13kugelthat's debatable
11:06:51Zagoryes, we debated it yesterday. and the consensus was that a single high-resolution volume control (optional high-resolution, even) is the desired solution
11:07:49[Saint]I seem to recall it being about 50/50 personally.
11:07:56 Join mudd1 [0] (~cmertes@2001:638:504:20e0:221:70ff:fe83:655e)
11:07:59kugelalright, we'll see what you come up with. but if it means rockbox can't change the global volume anymore then I will probably object
11:08:23Zagorkugel: so would I
11:09:53*kugel doesn't know if the global volume steps are the same on all android devices
11:10:10 Quit mudd1 (Client Quit)
11:10:53Zagorfrom what I have read, android offers 16 steps on all models. but the levels of each step might be different. we'll have to see what is possible.
11:13:01pixelma16 here for media and 8 for ringtone so they already offered additional steps
11:13:18pixelmaeven if it's not many
11:13:45 Join casainho [0] (
11:13:49kugeli have 9 levels (1 of them is mute)
11:15:09Zagorkugel: for ringtone or media?
11:15:44 Quit Horscht (Quit: Verlassend)
11:15:47kugelringtone is 8+vibrate+mute, media is 8+mute
11:16:12Zagorwow only 8! which phone is that?
11:16:18kugelhtc legend
11:16:39pixelmamaybe asking for Android version is needed too
11:16:53Zagorpixelma: if it is linked to android version
11:16:54kugel2.3.3 here
11:17:01Zagorthen it's not
11:17:08kugelbut I'm not surprised if vendors handle it differently
11:17:36pixelmaZagor: yeah, I know. But I thought running custom ROMs *could* have an influence
11:17:48Zagorpixelma: right, it could
11:19:08kugelwell, it means you can't rely on the exact volume steps, no?
11:19:10Zagorthough I can't see why a custom rom would make volume resolution worse than stock
11:19:45Zagorkugel: unless we can someone find out
11:39:42linuxstbMy HTC Desire HD has 16 for media, 8 for ringer.
11:42:19Zagorit looks like we could "just" expand to also adjust the stream volume
11:49:48 Join robin0800 [0] (
11:50:50 Join mcuelMobile [0] (b274a083@gateway/web/freenode/ip.
11:52:30mcuelMobileZagor: take a look at FS #11914 patch8
11:53:08Zagorlookie there!
11:54:54 Join mcuelMobile2 [0] (b274a083@gateway/web/freenode/ip.
11:55:05mcuelMobile2stupid crashin webclient
11:55:12 Quit robin0800 (Ping timeout: 255 seconds)
11:55:14Zagorgreat, I'll cut and paste a bit to see if I can make what I want
11:56:38mcuelMobile2I guess the hardest problem is to fix the fade on pause
11:56:48 Quit mcuelMobile (Ping timeout: 245 seconds)
11:57:28ZagormcuelMobile2: no, that's easy. we do that using pcm volume only, not affecting the stream.
11:58:05Zagoror, "easy". :-) it needs to make sense in the code too.
11:59:05 Join mshathlonxp [0] (
11:59:40mcuelMobile2but how will the set_volume function which volume to set?
12:00:21Zagorthe idea is to use pcm attenuation as small steps between the bigger stream volume steps
12:00:30 Quit JackWinter (Remote host closed the connection)
12:01:15Zagorso the function sometimes sets only pcm, and sometimes both
12:01:56mcuelMobile2well gtg, will look at it later
12:02:06 Quit mcuelMobile2 (Quit: Page closed)
12:03:15 Join JackWinter [0] (
12:03:17mshathlonxpcan anyone tell me why in rockbox "keyboard" aren't signs of baltic languages - lack of demand? :)
12:03:38gevaertsYou know you can change the layout?
12:03:50mshathlonxpnot yet ;>
12:03:51n1smshathlonxp: you can have whatever you want on the keyboard
12:03:57gevaertsIt's not possible to make one keyboard layout that suits all languages
12:04:22mshathlonxpwell, looks like default suits all, except for russian, asian and baltic :D
12:05:05gevaertsRight. It doesn't suit most of the world
12:05:24mshathlonxpah, I forgot arabs as well
12:05:37n1sits easy to suite different languages that use mostly the same characters
12:06:07mshathlonxpgevaerts it does
12:06:18gevaertsIt does?
12:06:50mshathlonxpchinese is only 1/6 of all world and other unsupported languages even less ;D
12:07:49gevaertsAre you aware of the fact that Chinese is *not* the only Asian written language?
12:08:41 Join HaimN [0] (~HaimN@
12:25:12 Quit timccc (Read error: Connection reset by peer)
12:25:31 Join timccc [0] (~timccc@
12:26:58pixelmahmm, I just found two MP3s that don't show the title tag on my Ondio even though it is there (it uses the filename fallback I have coded in my WPS). The difference to other working tracks is that they also have those itunesgapless tags set which the MAS probably can't make use of anyway. I'll try if removing these helps
12:31:45pixelmait does
12:31:51***Saving seen data "./dancer.seen"
12:33:50n1spixelma: sounds like a bug still
12:34:14n1sdid this happen before the semi-recent changes to id3 code?
12:35:25 Quit pamaury (Remote host closed the connection)
12:39:54pixelmacan't tell for sure, I don't think I noticed before and I only recently started keeping those tags, I don't think I have many of such files
12:40:50pixelmaand I don't use all my targets regularly
12:43:46pixelmait seems though that the grouping tag support now suddenly started working which I think it didn't before on hwcodec. Some weirdness going on with the different tagging fields
12:48:58mshathlonxpgevaerts well, the others don't have that much speakers :)
13:02:32 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
13:07:48n1smshathlonxp: how about hindi?
13:09:51pixelmait's getting a bit offtopic-y I think ;)
13:15:29 Quit user890104 ()
13:28:16 Join MethoS- [0] (~clemens@
13:31:43mshathlonxpn1s I dunno how hindi is being written, however I know that second official language in india is english anyway :>
13:32:45n1smshathlonxp: please take this to -community if you want to discuss it further
13:35:22HaimNHi, i need some help, my Sansa Fuze not read my SD card, the original OS read it bun rockbox not, its happened It only happens on one card, another SD card worked fine for me
13:40:11HaimNi found a patch that correct it but I don't know how to work with patches (I know compile scorce code but I don't know how put patches)
13:41:17HaimNI'm sorry for my bad english
13:44:34 Join liar [0] (
13:47:22kugelJdGordon|: why do you edit someone else's comment instead of just adding one?
13:48:40kugel(FS #12005)
13:49:12 Quit mshathlonxp (Ping timeout: 276 seconds)
13:50:21 Join mshathlonxp [0] (
13:50:53 Join Horscht [0] (
13:50:53 Quit Horscht (Changing host)
13:50:53 Join Horscht [0] (~Horscht@xbmc/user/horscht)
13:53:47pixelmanot the first time he did that, it was one of mine years ago and the bug report lost quite some info through the editing in my case
13:57:27kugelI don't understand the reason to do that
13:57:51pixelmame neither
13:58:21kugelit even requires more clicks
13:59:00pixelmaand is misleading - FS #8998 was mine
14:03:53HaimNsideral: I tried to work with it, but I got an error
14:04:28 Quit antil33t (Read error: Connection reset by peer)
14:04:38 Join antil33t [0] (
14:04:43sideralwhat's the error message? (use for copy&paste)
14:06:42 Quit mshathlonxp (Read error: Connection timed out)
14:07:27 Join mshathlonxp [0] (
14:08:38sideral"hunk failed" that means (a part of) the patch doesn't apply to your source tree. possibly the source tree has evolved since the patch was cut.
14:09:15sideralwhich of the patches did you want to apply? you should only need sd-init-fix.patch
14:10:04HaimNsideral: I tried only sd-init-fix.patch
14:10:32sideraland which version of the source code do you use?
14:11:09HaimNrevision 29559
14:12:06sideralThe patch applies perfectly for me to this revision
14:12:38sideralmaybe a DOS CRLF issue?
14:13:23HaimNI guess I do not have installed correctly the patch
14:13:35HaimNi work in ubuntu
14:15:14sideraltry downloading the patch again. Or apply the change manually; the patch is rather trivial
14:16:57sideralwget -O sd-init-fix.patch ""
14:17:57 Join milk_ [0] (
14:18:02HaimNHow can I download it? When I click it it opens it in a browser, (what I did before was to copy the text file with the same name)
14:18:17HaimNOK, thanks!
14:18:22 Quit Robdgreat (Ping timeout: 246 seconds)
14:18:58sideralBrowsers typically have a "Save as..." function that saves the page source verbatim
14:20:07HaimNits save it as a HTML file
14:20:23 Join Robdgreat [0] (~rob@unaffiliated/robdgreat)
14:21:13sideralthen either your browser is broken, or you're using an evil web proxy
14:21:28HaimNWhere should I put the patch?
14:22:27sideralwhere you want. you only need it once, for the patch command
14:23:38HaimNsideral: This is probably my ISP, I have content filtering (like porn sites or violent)
14:25:06gevaertsThis patch isn't very violent
14:25:19sideraldid the wget command manage to download the patch in plain text?
14:25:47sideralgevaerts: sure it is! have you ever been "deselected"?
14:26:01Zagoryay, this works pretty good
14:26:17gevaertssideral: never after a 100ms nap :)
14:26:29 Quit einhirn (Read error: Connection reset by peer)
14:26:43sideralgevaerts: that sounds quite cruel too
14:29:37 Join Highlander [0] (
14:30:24HaimNsideral: yes, but i'm still get the error massege
14:31:53***Saving seen data "./dancer.seen"
14:32:02sideralHaimN: Try doing the change the patch wants to apply by hand (find the right line, insert "udelay(100000);")
14:36:16HaimNI should delete these rows?:
14:36:44HaimNor just add udelay(100000);
14:36:46 Join user890104 [0] (
14:38:51gevaertsJust add the udelay
14:40:46HaimNok, thanks! now I compile it
14:47:05sideralHaimN: if you want to read up on the patch file format, check this out:
14:47:48 Join TheLemonMan [0] (
14:50:04HaimNsideral: thanks
14:56:07HaimNits work! Thank you all, (Especially you sideral)
14:56:41sideralNo worries, and thanks for trying Rockbox!
14:58:29HaimNI'm using Rockbox for almost a year and I enjoyed every minute, programmer already Rockbox on 30 more players of my friends!
14:59:27HaimN*I installed Rockbox already on another 30 musicians of my friends!
14:59:45 Join dfkt [0] (dfkt@unaffiliated/dfkt)
15:00:38sideralOK, then thanks for spreading the gospel! :)
15:02:35 Quit Highlander (Quit: Quitte)
15:07:23 Quit krazykit (Ping timeout: 246 seconds)
15:07:35 Join kevku [0] (~kevku@2001:7d0:0:f9af:feed:feed:feed:feed)
15:09:39 Join krazykit [0] (
15:10:27 Join sirrozha [0] (~sirrozha@
15:11:25 Quit sirrozha (Remote host closed the connection)
15:13:10 Quit TheLemonMan (Quit: Destructor called)
15:14:03 Join giovannizilli [0] (
15:14:14 Quit antil33t (Read error: Connection reset by peer)
15:14:22 Join antil33t [0] (
15:15:30giovannizilliI'm trying to use the convttf tool to convert anti alised fonts, but got some mitakes
15:16:17giovannizillicd /rockbox/tools; make; and get an error
15:16:18giovannizillimkzenboot.c:29: fatal error: zlib.h: File o directory non esistente compilation terminated. make: *** [mkzenboot.o] Errore 1
15:17:28gevaertsTry make convttf
15:17:56giovannizilliconvttf.c:21: fatal error: ft2build.h: File o directory non esistente compilation terminated. make: *** [convttf] Errore 1
15:18:23gevaertsInstall the freetype development package
15:18:58 Join benedikt93 [0] (
15:19:02 Quit benedikt93 (Changing host)
15:19:02 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
15:19:59 Quit simonrvn (Quit: see ya - n'multes)
15:20:26 Join simonrvn [0] (simon@2001:470:8c85:11fe::c0a8:195)
15:22:05giovannizilliok, done. Thank you very much.
15:22:20 Join robin0800 [0] (
15:24:16robin0800just uploaded a new MaximumArt Theme for the Sansa C200 could someone please delete the old one? many thanks
15:24:27 Quit krazykit (Ping timeout: 252 seconds)
15:26:26 Join krazykit [0] (
15:28:02 Quit giovannizilli (Quit: CGI:IRC)
15:28:43gevaertsrobin0800: done
15:30:16 Join facelessloser [0] (
15:30:54facelesslosercould someone help me out, i need to install an older version of rockbox
15:34:44robin0800gevaerts, thanks again .....managed to use %?if(%sC, <=, 35)<> as suggested last night by JdGordon
15:38:02 Quit Judas_PhD (Quit: This is a quitting message)
15:38:41 Join panni_ [0] (
15:39:15 Quit panni_ (Read error: Connection reset by peer)
15:39:31 Join panni_ [0] (
15:41:49 Quit facelessloser (Quit: CGI:IRC (EOF))
15:53:10 Join mcuelenaere [0] (~mcuelenae@rockbox/developer/mcuelenaere)
15:53:51mcuelenaerehas anyone run RB Android on a (recent version of) an emulator succesfully lately?
15:58:29Zagorgood news: the code works. getting the algorithm right is a bit of a challenge though.
15:59:28Zagori.e. how much should I lower the pcm volume when I bump the stream volume a step?
15:59:41 Part LinusN
15:59:56dionoeamcuelenaere: the audio output doesn't always work (I get it working like 10% of the time)
16:00:02dionoeais that the problem you were having ?
16:00:49mcuelenaereno, some crash in native code
16:00:50dionoeakugel had a pcm driver rewrite patch which I was supposed to try... maybe it fixes it
16:01:07mcuelenaerebut I was trying to pinpoint whether that was due to local changes or a stricter NDK
16:01:40dionoeaah hum ... never had that (but then I didn't try it in the last 2 weeks on an emulator)
16:02:02mcuelenaeredionoea: I did have trouble with the audio driver in the past though (emulator and my old G1)
16:02:08mcuelenaereso a rewrite sounds nice :)
16:03:43ZagorI have nice algorithms for "too little" and "too much" but not "just right" :)
16:05:31mcuelenaeredionoea: thanks, will look at it later
16:05:37kugelZagor: how much steps does the stream have via code?
16:05:50 Quit HaimN (Ping timeout: 255 seconds)
16:06:45kugelmcuelenaere: audio in the emulator just stopped working for me at some point (during early development)
16:06:58kugelI suspected it was a qemu issue since there's no problem on my phone
16:07:18 Quit [Saint] (Ping timeout: 240 seconds)
16:07:24kugelbut when it worked it was so utterly slow that it wasn't useful anyway
16:07:32Zagorkugel: stream is 0 to getStreamMaxVolume(). which is 15 on my phone and 8 on yours.
16:07:41mcuelenaereZagor: how about aligning it to the nearest 990/getStreamMaxVolume()?
16:07:41Zagorpcm is, as you know, 0 to 1.
16:07:51mcuelenaerekugel: did you look at logcat?
16:07:57mcuelenaere(at that point in time)
16:08:12kugelcan't remember
16:08:13Zagormcuelenaere: how do you mean?
16:09:18mcuelenaereI remember having some thread issues in the audio driver (e.g. invoking methods loaded from Java thread 1 on Java thread 2)
16:09:37mcuelenaereZagor: that was RE your how much should I lower the PCM volume when I bump the stream volume a step
16:10:22Zagoryes. but I don't understand your suggestion. getStreamMaxVolume() is a constant.
16:11:01mcuelenaereRockbox PCM volume is -990..0, so going to the nearest 990/16=61 (getStreamMaxVolume on some devices) should fix that, shouldn't it?
16:11:50kugelwe want more than 16 steps
16:11:52Zagorno, the problem is I have to control two different volume controls: the music stream volume and the pcm volume. the stream volume is 0-15 (on my phone) and the pcm volume is 0.0 to 1.0.
16:12:07 Quit Barahir_ (Read error: Operation timed out)
16:12:13mcuelenaereso if you're on -500, and you bump up the stream volume you'd go to -~488
16:12:33mcuelenaerehmm ok, different problem
16:13:07Zagorwhen I go from -930 to -920 I go from stream volume 1 to 2. then I also lower the pcm volume to make it a smooth step.
16:13:22 Join [Saint] [0] (S_a_i_n_t@
16:13:44Zagorthe challenge is getting the audible difference the same when going from -930 to -920 as when going -920 to -910.
16:14:23mcuelenaerecan't you leave the PCM volume at 1 and only touch the stream volume? (I'm guessing I'm not familiar with the full problem)
16:14:47Zagormcuelenaere: I can. then we get the standard android low-resolution volume.
16:15:13Zagorthe goal of the excercise is to add high resolution volume control by also using the pcm volume.
16:15:38Zagorbecause the difference between 1 and 2 in stream volume is really big
16:15:38mcuelenaereah yes, I somehow thought Rockbox did some additional volume mixing
16:16:22 Join Barahir [0] (
16:16:23Zagoron android we leave that to the AudioTrack class
16:19:12mcuelenaeredo you know the relation between PCM volume & stream volume, e.g. does PCM=1.0, stream=1 equal to PCM=0.5, stream=2?
16:19:17 Quit sideral (Ping timeout: 252 seconds)
16:19:18 Quit antil33t (Read error: Connection reset by peer)
16:19:30 Join antil33t [0] (
16:19:54Zagormcuelenaere: my testing indiciates it is unfortunately not that simple, no
16:20:12 Quit mystica555_ (Read error: Connection reset by peer)
16:20:17mcuelenaereand I fear that it won't be the same across all devices..
16:20:30Zagorprobably not
16:22:41 Quit tails___ (Ping timeout: 250 seconds)
16:22:57Zagorone workaround is to only add high resolution volume to stream volume 1. that's where it is most needed, and fully deterministic.
16:24:17mcuelenaerebut then the user can't change the volume the same way when he/she is out of the Rockbox activity
16:25:59Zagorrockbox can listen to volume changes and remove the pcm attenuation when the user bumps volume from outside rockbox
16:26:20Zagorbut no, pcm volume can only be controlled from rockbox
16:26:53 Join u42p [0] (
16:27:14Zagorwell actually if rockbox listens to the volume changes, it can still do pcm volume adjustments
16:27:15mcuelenaerekugel: ever got "JNI WARNING: ref 0x... should not be visible to native code"?
16:27:44kugelnever seen that one
16:27:46Zagorobviously, the pcm attenuation only applies to rockbox but that will always be the case
16:28:38 Quit panni_ (Read error: Connection reset by peer)
16:28:39mcuelenaereanyone willing to test an Android-related patch? (just to eliminate my environment as a problem)
16:29:47 Join LinusN [0] (~linus@rockbox/developer/LinusN)
16:29:51ZagorI can do it later tonight. I have to go soon.
16:31:36 Part LinusN
16:31:55***Saving seen data "./dancer.seen"
16:32:44 Join tails_ [0] (
16:32:54mcuelenaereright, seems like the emulator has somehow gone stricter (or it hasn't been tested for some time)
16:33:06mshathlonxpn1s btw how can I change keyboard layout?
16:33:40*mcuelenaere really wants native backtraces on Android
16:36:04 Join Judas_PhD [0] (
16:41:20pixelmamshathlonxp: read on, follow links...
16:45:28CIA-2New commit by mcuelenaere (r29560): Android port: generate initial config when none is present with dynamically ...
16:45:38CIA-2New commit by mcuelenaere (r29561): Android: update project to SDKv11
16:45:42CIA-2New commit by mcuelenaere (r29562): Android: future-proof the RunForegroundManager code to Honeycomb
16:45:48CIA-2New commit by mcuelenaere (r29563): RaaA: move Android apps-code to separate dir under apps/hosted
16:45:54CIA-2New commit by mcuelenaere (r29564): Android: remove rockbox.apk when doing 'make clean'
16:46:01CIA-2New commit by mcuelenaere (r29565): Android: replace "launching Rockbox activity"-behaviour with ...
16:46:06CIA-2New commit by mcuelenaere (r29566): Android: get rid of warnings (no functional changes)
16:46:48Zagorwhammo! :)
16:47:03dionoeamcuelenaere: are you runing honeycomb in the emulator ?
16:47:18mcuelenaerenope, my poor laptop can't handle it :)
16:47:58mcuelenaerebut this code was tested on Android 2.2.1 (and based on reference code from the SDK)
16:48:13dionoeathe crashes you get are on a froyo emulator then ?
16:48:41mcuelenaerecurrently, a 1.6 one
16:49:05CIA-2r29560 build result: All green
16:49:08dionoeaah. I was runing 2.2 last time i tired
16:49:13linuxstbmcuelenaere: Shouldn't you use appextra instead of this?;r2=29563;pathrev=29563
16:49:13mcuelenaerebut I read somewhere that the emulator has some checking enabled not in release builds
16:49:29linuxstbmcuelenaere: (I removed that code recently...)
16:49:30*mcuelenaere greps appextra
16:49:49pixelmafor what it's worth - I also get the occasional "audio stops working" on my phone. Considering that I haven't listened for a long time and just do a bit of testing, it happens quite often
16:50:26linuxstbmcuelenaere: It's used in apps.make just about that code you added. It's defined in tools/configure for each target
16:50:28mcuelenaerelinuxstb: this was based on some "old" patches from FS #11914
16:50:32scorcheB4gder: Zagor: ping
16:50:41kugelmcuelenaere: what changed with honeycomb w.r.t. foreground running?
16:50:45linuxstbmcuelenaere: Then surely they conflicted?
16:50:50mcuelenaerekugel: it removed setForeground
16:51:00Zagorscorche: yes?
16:51:11mcuelenaerelinuxstb: it did, but I didn't see anything appextra related
16:51:16mcuelenaerelinuxstb: will change it now
16:51:27kugelmcuelenaere: that's removed long ago, isn't it?
16:51:32scorcheZagor: I am getting ready to submit the GSoC org app - have you had a look and can give it your final blessing?
16:51:43kugelmcuelenaere: IIRC 2.2 already removed it, and we handled that already
16:51:44*Zagor takes a look
16:52:04kugelah, wait I understand. you mean the function is removed entirely (instead of being a no-op)?
16:52:06mcuelenaerekugel: perhaps it deprecated it, but in Honeycomb the actual call is removed
16:52:39CIA-2r29565 build result: All green
16:52:57kugelmcuelenaere: why didn't you just change the other api. your change seems larger than needed
16:54:05linuxstbmcuelenaere: A question unreleated to your commits today - why the #ifdef APPLICATION around target/hosted/android/app/button-application.c in firmware/SOURCES ?
16:54:55mcuelenaerekugel: hmm I didn't really look at it much, due to setForeground not being available anymore I just looked at the Android reference and found this code and went with it (after fixing it)
16:55:37mcuelenaerelinuxstb: hmm did I add that?
16:55:48linuxstbsvn blame says you did ;)
16:56:25mcuelenaeregit blame says kugel ;)
16:56:47linuxstbIt was this commit, no?
16:57:13scorchealso, if you havent yet and have interest, look over and
16:57:34scorche(also, join #rockbox-gsoc
16:57:53mcuelenaerelinuxstb: hmm I have this;revision=27310
16:58:25*kugel sees no #ifdef APPLICATION in either commit
16:58:38mcuelenaereah sorry, wrong grep
16:59:05*kugel wonders if people look at the revisions they post here :p
16:59:09mcuelenaerelinuxstb: I guess I based myself on the SDL port?
16:59:14mcuelenaerekugel: I did :)
16:59:33linuxstbkugel: You can't see it in the link I posted?
16:59:38mcuelenaerekugel: and
17:00:04 Part u42p ("Leaving")
17:00:23kugellinuxstb: I can't see it
17:01:07pixelmait's very clear from linuxstb's link
17:01:07kugelah you talked about SOURCES
17:01:10mcuelenaerelinuxstb: so do I add hosted/android to APPEXTRA now or ..? (doesn't seem very clean)
17:02:47 Quit Farthen (Ping timeout: 240 seconds)
17:03:06linuxstbmcuelenaere: That's what the APPEXTRA mechanism is for - extra, target-specific subdirs in apps/. IMO it's cleaner than putting target-specific #ifs in apps/Makefile
17:03:11linuxstbI mean apps/apps.make
17:03:46linuxstbmcuelenaere: But going back to firmware/SOURCES, it does just look like a copy/paste from the SDL code (which IIUC is used for both the sim and SDL apps), so it's needed there, but not for Android.
17:03:59 Join Farthen [0] (
17:04:03mcuelenaerelinuxstb: so the Makefile for Android should containt APPEXTRA=recorder:gui:radio:hosted:hosted/android?
17:04:18linuxstbmcuelenaere: No need for ":hosted:" any more I think.
17:04:19mcuelenaerelinuxstb: yea, probably
17:04:32linuxstb(unless you left something there)
17:04:41mcuelenaerecurrently, no
17:05:31scorchelast call before i submit the application...
17:05:53mcuelenaerelinuxstb: you're sure hosted isn't needed anywhere else?
17:06:16mcuelenaereoh it's only in apps/ ok
17:07:12 Part Zagor
17:07:59mcuelenaerekugel: RE foreground, you want me to revert to the old way and do it that way?
17:08:43CIA-2New commit by mcuelenaere (r29567): Android: use APPEXTRA instead of makefile hack
17:11:50mshathlonxpor maybe ? :)
17:12:47CIA-2r29567 build result: All green
17:12:56kugelmcuelenaere: I don't feel strong, but I guess I made the two separate interfaces for exactly this case
17:14:49scorche"Organization Application submitted."
17:16:52linuxstbscorche: When are the successful ones announced?
17:17:35*linuxstb wanted a date, not a link ;)
17:18:36linuxstbNext Friday,18th March at 19;00 UTC
17:18:42scorchethe date is in the link! ;)
17:22:29 Quit Judas_PhD (Quit: This is a quitting message)
17:25:38 Quit mshathlonxp (Ping timeout: 276 seconds)
17:26:30 Join mshathlonxp [0] (
17:28:05 Quit timccc (Ping timeout: 250 seconds)
17:28:13 Join Judas_PhD [0] (
17:28:55 Quit Judas_PhD (Client Quit)
17:30:18 Quit [Saint] (Quit: I'm only going to Heaven if it feels like Hell, I'm only going to Heaven if it tastes like caramel...)
17:30:47 Quit tails_ (Ping timeout: 252 seconds)
17:31:43 Join tails_ [0] (
17:36:22 Join [Saint] [0] (S_a_i_n_t@
17:40:39 Quit [Saint] (Ping timeout: 250 seconds)
17:46:31 Join [Saint] [0] (S_a_i_n_t@
17:47:55 Join sideral [0] (~sideral@
17:47:55 Quit sideral (Changing host)
17:47:55 Join sideral [0] (~sideral@rockbox/developer/sideral)
17:49:37 Join bmbl [0] (
17:49:37 Quit bmbl (Changing host)
17:49:37 Join bmbl [0] (~bmbl@unaffiliated/bmbl)
17:50:56 Join tails__ [0] (
17:52:08 Quit jhMikeS (Ping timeout: 240 seconds)
17:53:44*gevaerts looks for people who know a bit about playlists and bookmarks to finish those gsoc ideas. JdGordon, pondlife, Llorean: you were most active on the mailing list about this
17:54:29gevaertskugel: can you write a bit about the Android native widgets thing?
17:55:18kugelshouldn't people that actually want that do it?
17:55:32 Quit tails_ (Ping timeout: 276 seconds)
17:55:51kugelI don't need/want native widgets so I haven't even thought about it any further. but others keep popping that up
17:56:42kugeliirc linuxstb is a strong advocate of it
17:57:21gevaertsCan you write up something else then? ;)
17:57:31*gevaerts is always hopeful!
17:59:39 Join stoffel [0] (
18:04:36 Join toffe82_ [0] (
18:05:29 Quit mshathlonxp (Quit: Leaving)
18:08:06 Quit sideral (Quit: Leaving.)
18:09:26 Join sideral [0] (~sideral@rockbox/developer/sideral)
18:09:52 Join komputes [0] (~komputes@ubuntu/member/komputes)
18:20:33 Quit sideral (Ping timeout: 246 seconds)
18:25:51 Quit literal (Quit: leaving)
18:25:57 Join literal [0] (
18:26:31 Quit n1s (Quit: Lämnar)
18:29:46 Quit Xerion (Quit: )
18:31:58***Saving seen data "./dancer.seen"
18:36:29 Join Xerion [0] (
18:43:00 Quit sasquatch (Quit: WeeChat 0.3.2)
18:43:24 Join sasquatch [0] (
18:50:48 Join LambdaCalculus37 [0] (~3f74f70d@rockbox/staff/LambdaCalculus37)
18:52:35 Quit pamaury (Read error: Operation timed out)
18:55:57*B4gder added himself to the gsoc page
18:57:45rasherNot that it matters much, but shouldn't the third tier here:
18:57:55rasher be the other mentors?
19:00:26 Join {phoenix} [0] (
19:03:16 Join kronflux [0] (
19:06:54 Join leavittx [0] (~lev@
19:11:35 Join bertrik [0] (
19:11:35 Quit bertrik (Changing host)
19:11:35 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
19:27:26 Join mshathlonxp [0] (
19:29:44 Quit LambdaCalculus37 (Quit: CGI:IRC 0.5.9 (2006/06/06))
19:32:26 Quit user890104 ()
19:34:35CIA-2New commit by mcuelenaere (r29568): Android: use NewGlobalRef for references that are used globally
19:34:40CIA-2New commit by mcuelenaere (r29569): Android: Don't share the JNI environment across threads, but obtain it the ...
19:36:23mcuelenaere(r29568 fixes the emulator)
19:38:40CIA-2r29568 build result: All green
19:40:22kugelmcuelenaere: huh?
19:40:36kugelthe environment was not shared acress threads. that wouldn't work anyway
19:40:55kugelI think your commit is not needed
19:41:19 Join Stummi [0] (~Stummi@rockbox/developer/Stummi)
19:41:26mcuelenaerekugel: I had some warnings that there was an access from the wrong thread to an jobject/jclass/whatever a long time ago
19:41:33mcuelenaerethis should fix that
19:41:40kugel"a long time ago"?
19:42:06CIA-2r29569 build result: All green
19:42:09mcuelenaereyes, can't remember when but it was using the emulator and when the CPU was very busy
19:42:16mcuelenaere+ playing audio
19:42:19kugelthe other commit is also completely unneeded
19:42:24mcuelenaereno it's not
19:42:30kugelyou don't need global refs
19:42:31mcuelenaerethat fixes the emulator
19:43:00kugelI read that yes
19:43:25kugelI explicitely avoided global refs so that the VM can do it's cleanup when main() returns
19:43:26mcuelenaerekugel: more precisely, making RockboxFramebuffer_instance a global ref fixed it
19:43:57kugelah right, that one was probably buggy
19:44:00kugelbut the others not
19:44:49kugelobjects created in native code are GC'd when the entry point function returns. you add unneeded complexity
19:45:21kugelmcuelenaere: I wonder why you put up rather trivial patches on FS but this ones not
19:45:40mcuelenaeretrue, I had forgotten that main() doesn't return
19:45:55mcuelenaerekugel: I've made these two just today
19:46:04mcuelenaerethe others were from the FS item
19:46:07kugelIMO both are unneeded
19:46:35 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
19:46:46mcuelenaereperhaps so, but the problem with RockboxFramebuffer_instance remains
19:46:51kugelI have actually work that lets main return (to shut down the service), in which case normal references would be much more convinient
19:48:00kugelright, the RockboxFramebuffer_instance is different because it's not created by native code since recently so it may need other handling (albeit its reference shouldn't change anyway)
19:49:06 Quit mshathlonxp (Read error: Connection timed out)
19:49:51 Join mshathlonxp [0] (
19:53:50 Quit MethoS- (Read error: Connection reset by peer)
19:54:15 Quit kronflux (Quit: Leaving)
19:55:55 Quit B4gder (Read error: Operation timed out)
19:56:59kugelmcuelenaere: I think I'd rather see those reverted and fix the issues in another way
19:57:16kugel(until now I wasn't even aware there are issues)
19:57:16mcuelenaerekugel: I'm working on partly reverting r29568
19:57:33mcuelenaereI don't think r2959 is wrong
19:57:55kugelcan you use the env pointer from the Surface callbacks instead of the global one?
19:58:08 Quit kkit|sh (Read error: Operation timed out)
19:58:27kugelI don't say r29569 is wrong, it's just unneeded
19:59:47mcuelenaerewell, even apart from the functional change, it does clean up the files a bit, there's no need for the external env_ptr anymore
19:59:47 Join B4gder [0] (~daniel@rockbox/developer/bagder)
20:01:55mcuelenaerekugel: doesn't it already use the env pointer from the Surface callbacks?
20:02:00 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
20:02:13mcuelenaereand I don't see how that's connected to the global reference problem
20:02:26kugelnot for lcd_update/_rect
20:05:00kugelI thought you need to use the env of the calling thread, but perhaps you need the one of the thread that created the object
20:05:20kugelor the same which you used for GetMethodID
20:05:49mcuelenaereyou need the env associated with the current thread
20:06:51 Quit pamaury (Ping timeout: 252 seconds)
20:07:17mcuelenaerekugel: is the revert, and doesn't revert the fix
20:07:56 Join Barahir_ [0] (
20:08:04kugelmcuelenaere: so so everything should've been alright before your commits
20:08:28mcuelenaereno, when extended JNI checks are enabled it triggers that JNI warning I mentioned earlier
20:08:31mcuelenaere(which is on emulators)
20:08:48 Join user890104 [0] (
20:09:59 Quit Barahir (Read error: Operation timed out)
20:10:03mcuelenaere(this is all about r29568)
20:10:04kugelbut can you tell which access was wrong?
20:10:40mcuelenaerewhen making RockboxFramebuffer_instance a global ref, it is fixed so probably in lcd_update(_rect)
20:10:57kugelbut they use the env of the current (calling) thread
20:11:32kugelI suspect it wants the same env used for GetMethodID (or the warning is just wrong)
20:13:03mcuelenaerekugel: this is what I get when I don't make it global:
20:13:36mcuelenaeremethod ID's can be shared across environments I think
20:13:50mcuelenaereit's only jobject and it derivates that can be free'd
20:14:19mcuelenaere"Note: method and field IDs are just 32-bit identifiers, not object references, and should not be passed to NewGlobalRef."
20:14:39mcuelenaereoops, mixing up separate things
20:14:58kugelthe activity creates the framebuffer, from the UI thread. does that help any?
20:16:02mcuelenaereI don't think the warning is related to referencing it from the wrong thread, it's probably because we shouldn't dereference a local reference
20:17:23mcuelenaerethe RB fb instance is passed to surfaceCreated as a local reference, thus when that method returns it could potentially be freed
20:17:35kugeldoesn't local/global reference just mean something about object lifetime?
20:18:10kugelbefore the surface is freed SurfaceDestroyed will be called
20:18:40mcuelenaeresure, but that doesn't have anything to do with the rules wrt local/global reference
20:19:26kugelthe reference is valid until it's garbage collected
20:19:38kugelglobal reference just menas "don't garbage collec this"
20:19:45mcuelenaereno, it's valid until the function returns
20:20:17mcuelenaere"Every object that JNI returns is a "local reference". This means that it's valid for the duration of the current native method in the current thread."
20:20:18 Quit casainho (Quit: ChatZilla 0.9.86 [Firefox 3.6.15/20110303171539])
20:20:29mcuelenaere"Even if the object itself continues to live on after the native method returns, the reference is not valid."
20:20:42mcuelenaere"(Dalvik VM will warn you about most reference mis-uses when extended JNI checks are enabled.)"
20:20:51kugelah, interesting. I missed that sentence
20:20:52mcuelenaerewhich I think is enabled on emulators
20:21:15kugelso yes, we need a global reference for the fb instance
20:23:03CIA-2New commit by mcuelenaere (r29570): Partly revert "Android: use NewGlobalRef for references that are used globally" ...
20:23:32 Join enthdegree [0] (
20:23:46mcuelenaereugh, forgot one part
20:25:11CIA-2New commit by mcuelenaere (r29571): Android: no need to keep RockboxPCM_class around
20:25:18kugelmcuelenaere: still not entirely happy with lcd-android.c
20:26:05kugelI think you can unconditionally create the global reference in SurfaceCreated and unconditionally delete it in SurfaceDestryoed (without need for lcd_deinit()), can't you?
20:26:20CIA-2r29570 build result: All green
20:26:54mcuelenaerehmm yes, probably, that could make lcd_deinit go away
20:29:28CIA-2r29571 build result: All green
20:31:59***Saving seen data "./dancer.seen"
20:35:40 Quit stoffel (Remote host closed the connection)
20:37:32 Quit kevku (Quit: KVIrc 4.0.2 Insomnia
20:37:41 Join stoffel [0] (
20:38:33CIA-2New commit by mcuelenaere (r29572): Android: delete the global reference in surfaceDestroyed ...
20:41:47CIA-2r29572 build result: All green
20:42:08 Quit mshathlonxp (Quit: reb00t)
20:45:23 Join mtstickney [0] (
20:45:35 Join msh_ [0] (
20:45:41 Nick msh_ is now known as mshathlonxp (
20:48:34kugelmcuelenaere: I think you also want to create the global ref uncontitionally, otherwise you possibly delete a non-existent ref?
20:49:43 Quit mc2739 (Ping timeout: 248 seconds)
20:50:05mcuelenaerekugel: when's that? the only time RockboxFramebuffer_instance gets set is by a NewGlobalRef() call (and when it's released, it's set to NULL)
20:50:48kugelah ok
20:51:06kugelI still find conditional "alloc" and unconditional "free" very confusing
20:51:32 Quit simonrvn (Quit: see ya - n'multes)
20:51:56 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
20:52:05mcuelenaerewell, I'm not sure what would happen when we globally reference something more than once, but doing it conditionnaly seems cleaner
20:52:27 Join simonrvn [0] (simon@2001:470:8c85:11fe::c0a8:195)
20:52:38mcuelenaereit's basically: if(cur != new) cur = new;
20:54:28 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
20:59:30mtstickneyis there a way to bypass an initial filesystem check in rockbox?
21:00:28 Join gbl08ma [0] (
21:00:30 Quit {phoenix} (Remote host closed the connection)
21:00:32linuxstbmtstickney: What do you mean? What initial filesystem check?
21:00:37Zagormcuelenaere, kugel: what's the easiest way to get a debug printout/dialog/whatever in android?
21:00:50mcuelenaereZagor: you know of logcat?
21:01:01mcuelenaereenable logf, and sprinkle your code with logf codes
21:01:09mcuelenaereah, do you use the emulator or an actual device?
21:01:14Zagoractual device
21:01:17mtstickneylinuxstb: I'm assuming that's what the "scanning disk" message refers to
21:01:36mcuelenaereZagor: enable adb on the device and just run "adb logcat" on your pc
21:01:50mtstickneylinuxstb: the trouble is I've got a corrupt fs, but I get a panic during the scan, so I can't repair the fs with another computer
21:01:51mcuelenaereand don't forget to connect your device to your pc :)
21:02:33mcuelenaereenabling adb can be done in settings->apps->development->USB-debug (loosely translated from dutch)
21:02:43 Quit enthdegree (Ping timeout: 276 seconds)
21:04:05kugelmcuelenaere, Zagor: DEBUGF/debugf works as well (in debug builds)
21:04:06mcuelenaerehmm, pausing and pressing play again makes the audio stop on my SGS
21:04:19Zagorand what do I call from java?
21:04:25linuxstbmtstickney: OK, that's the Rockbox "directory cache" (dircache) feature. Which device are you running Rockbox on?
21:04:25 Quit leavittx (Ping timeout: 252 seconds)
21:04:29mcuelenaerekugel: is debugf available all the time? (without adding LOGF_ENABLE)
21:04:36mcuelenaereZagor: Log.d("Rockbox", "....");
21:04:45mcuelenaereor Log.w if you want warnings etc
21:04:45kugelLOG (if it's there, but that's just a wrapper for Log.d("Rockbox", ...)
21:05:00kugelmcuelenaere: in debug builds
21:05:11mtstickneylinuxstb: it's a sansa e200v2
21:05:19kugelbut yea, without adding LOGF_ENABLE
21:07:34mcuelenaeredoes Rockbox already handle .nomedia?
21:07:52kugelno, should we?
21:08:11mcuelenaereI think that's the Android equivalent to .nodb (or whatever that is on Rockbox)
21:08:27saratogayou can disable dircache, but its probably easier to just fix your file system
21:08:32kugelyes, but for the android's database
21:08:46kugeland it includes pictures
21:09:03mtstickneysaratoga: I can't do either, because the initial dircache scan crashes rockbox (so it doesn't function as a drive when plugged in via usb)
21:09:03kugelI actually have .nomedia in my rockbox folder so that the photo viewers don't pick up the album art
21:09:33saratogamtstickney: you'd fix the file system from USB mode, not rockbox, we don't support USB on your player yet
21:09:40mcuelenaerewell some apps seem to create some .nomedia files, so why not also ignore those folders upon scanning?
21:09:45gevaertssaratoga: yes we do
21:10:01gevaertsmtstickney: boot to the original firmware, and fix it from there
21:10:15saratogado we?
21:10:55saratogaah yes, i'm mixing up my AS chips again
21:12:12saratogafor some reason i read e200v2 as as3525v2
21:12:13mtstickneygevaerts: good call, I forgot about the OF's usb mode
21:12:38linuxstbBut that sounds like a bug to me - Rockbox should be able to enter USB mode even if the disk is corrupted.
21:13:16gevaertslinuxstb: I'd say we need bootloader USB for this sort of thing
21:14:05saratogaor failing that at least go to USB mode before doing DB and disk caching if you boot up with it inserted
21:14:43linuxstbgevaerts: Yes, I guess that makes more sense on most targets.
21:14:54gevaertssaratoga: I'm not against that, but *I*'m not touching that code :)
21:15:25saratogayeah i'm not going near that one either
21:15:32saratogaUSB bootloader is probably a lot easier
21:15:36gevaertsyes :)
21:15:57kugelmcuelenaere: this way the code makes more sense to me (as a reader)
21:15:59gevaertsI guess we'd hide bootloader USB behind a button though
21:16:17saratogaon most targets don't we always use it if theres a USB cable connected?
21:16:21saratogamy gigabeat F does I think
21:16:29gevaertsYes, and that's *annoying*
21:16:32saratogaunless thats the OF USB mode with a similar font
21:16:52gevaertsIs your gigabeat flashed?
21:16:57saratogano stock
21:17:03gevaertsThen it's the OF
21:17:15gevaertsActually, I'm not sure
21:17:22saratogait looks like sysfont IIRC
21:17:53gevaertsBootloader USB tends to break car adapter mode
21:17:56saratogaideally it would boot back into rockbox after USB disconect so you didn't realize the bootloader handled the transfer
21:18:13saratogaon AMS we can actually tell the difference between a charger and a USB cable
21:18:30gevaertsOnly if you wait
21:18:59Zagoryes, got it!
21:20:27Zagorit actually is linear, my code was just wrong
21:21:17gevaertssaratoga: the detection is done by waiting for USB bus activity. I guess the bootloader could wait a second or so (I don't know what the best time is) if it sees power
21:21:49saratogathat seems sensible
21:21:59saratogaif USB host isn't detected, just keep booting, rather then crash
21:22:09Topic"Rockbox 3.8 is released | Please read before speaking: | Please direct offtopic/social chat to #rockbox-community | This channel is logged at | An upgrade caused issues with the wiki. We are on the case | Please finish GSoC ideas from -dev and put them on" by ChanServ (ChanServ@services.)
21:22:34gevaertsI'd still add a button. Some hosts may be slow, and we don't want to wait for a minute
21:22:40 Join pamaury [0] (
21:22:40 Quit pamaury (Changing host)
21:22:40 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
21:23:20gevaertsThat only works for software stack targets of course
21:26:28mcuelenaerekugel: it is it guaranteed that for every surfaceCreated call, there's an equivalent surfaceDestroyed call?
21:26:39mcuelenaereif so, then it would simplify it yes
21:26:47mcuelenaerethen, commit away :)
21:26:59kugelwith various SurfaceChanged calls inbetween
21:27:27mcuelenaerebut we don't handle that in native code, do we?
21:27:57mcuelenaereoh, it isn't even handled in java code I see
21:28:34mcuelenaereperhaps that should emulator surfacedestroyed+surfacecreated?
21:28:41mcuelenaereif the size changes
21:29:00mcuelenaerebut I don't see what would trigger that at runtime
21:33:20kugelI couldn't parse that
21:40:23CIA-2New commit by zagor (r29573): Use the pcm volume to add finer steps between every android stream volume step.
21:42:03 Quit esperegu (Remote host closed the connection)
21:42:20Zagorkugel: when you have a minute, I would be interested to see how/if that works with your 9-step music stream
21:43:43kugelZagor: I'll have a look
21:43:51CIA-2r29573 build result: All green
21:44:36Zagornow we need to add volume event listening so we adapt to changes done outside rockbox
21:45:09kugeldidn't mcuelenaere patch have that?
21:46:57 Quit factor (Read error: Connection reset by peer)
21:49:14bluebrotherZagor: your volume button changes for Android have an ugly side effect: the volume steps of the buttons are different depending on if Rockbox UI is showing or not
21:49:35Zagorside effect? that is the whole point!
21:50:10bluebrotherseriously? I find it strange to have the buttons behave differently depending on if Rockbox is showing or not (or even the screen is locked)
21:50:11Zagorinside rockbox, you have higher volume resolution than android has
21:50:51Zagorwell, the only "solution" to that is to disable the high-resolution volume control
21:51:21bluebrothersure, but this makes the button behaviour inconsistend.
21:51:40Zagoryes. I suggest we make it optional.
21:52:18bluebrothersounds like an option :)
21:52:56AlexPI don't like the idea of an option, can we have an option to show the option?
21:53:22Zagorconfigurable menus
21:53:26B4gderwe can have the app guess if you want to see the option!
21:53:50AlexPCan I enable curmudgeonly old bugger mode? :)
21:54:07AlexPConfigurable menus done well would be nice
21:54:15 Quit stoffel (Remote host closed the connection)
21:54:19*gevaerts wants someone to implement an option to increase interest in gsoc
21:54:35AlexPgevaerts: To harass people you really have to say something :)
21:54:55gevaertsAlexP: I said things, and I changed the topic!
21:55:00AlexPAnd if there is something that you feel I can contribute to, then do prod :)
21:55:11AlexPI didn't notice either :)
21:55:20kugelI actually think in-app vs out-of-app is understandable by the user, but perhaps screen-on vs screen-off not
21:55:30gevaertsAnd that shouldn't even have been necessary, there have been several calls on the mailing list
21:55:35kugelbut it's probably less confusing than 2 entirely separate levels
21:55:40AlexPgevaerts: indeed there have
21:56:27Zagorkugel: is screen-off the same as out-of-app in this case?
21:56:38kugelI expect so, I haven't tried yet
21:56:41*AlexP is looking forward to a bit of an android bashing-things-out session at devcon
21:56:58kugelbut IIRC screen-off is handled by the lockscreen app (if that exists)
21:57:00bluebrotherit at least behaves so :)
21:57:02B4gderjune 3rd is looking more and more as the date now
21:57:13gevaertsAlexP: can we do the bashing without Android?
21:57:22Zagorkugel: yes, tested now. biiig volume steps :-)
21:57:43AlexPB4gder: I'd say it is the 3rd even
21:58:00AlexPIt'd take a late run of 3 or 4 people to even think about changing it
21:58:11B4gderyes, unlikely to happen
21:58:26AlexPWhich is a bugger - holiday weekend (everywhere except UK) = more expensive eurostar
21:58:41AlexPBut we shall cope!
21:58:47B4gderI'm sure the fund can sponsor
21:58:51AlexPheh :)
21:59:00gevaertsCan the fund buy a train? :)
21:59:17bluebrotherpersonally I would prefer the 17th. But I guess doesn't make much of a difference given the numbers :)
21:59:31bluebrothergevaerts: Rockboxers on a train?
21:59:51AlexPgevaerts: We could mount a hostile take over of Deutsche Bahn - they arwe bidding for a eurostar service
21:59:56kugelI would really prefer 3rd june since (as I mentioned) my internship starts mid- or end-june
22:00:07*gevaerts considers the idea of building a Tower on a train
22:00:10AlexPbluebrother: can you make the 3rd?
22:00:19kugelI don't want to take a free friday early on
22:00:26bluebrotherAlexP: yep, it's just a bit inconvenient for me
22:00:31AlexPah, OK
22:00:33bluebrotherbut I'll manage that :)
22:00:50AlexPcool :)
22:01:16bluebrotherI just need to figure how I want to handle extending my stay. If I'm around I really have to do that :)
22:04:43 Join DJ_PUNK2011 [0] (
22:04:56 Join factor [0] (~factor@
22:06:05DJ_PUNK2011Does anybody know the band name of this music or maybe song name
22:06:20 Join Mir_ppc [0] (
22:06:39bertrikI haven't planned ahead to june, I'll probably be able to make nearly any date
22:06:49Mir_ppci just installed roxkbox on a Clip (v1) 2gb
22:06:50bluebrotherDJ_PUNK2011: the "This video is not available in your country" band?
22:06:53Mir_ppcits crashed
22:07:06Mir_ppcthis is the error message
22:07:06Mir_ppcMir_ppcPrefetchabort at 60000016
22:07:07Mir_ppcMir_ppcfsr 0x8
22:07:07DBUGEnqueued KICK Mir_ppc
22:07:07Mir_ppcMir_ppcdomain 0, fault 8)
22:07:12DJ_PUNK2011Hey u guys
22:07:14DJ_PUNK2011how are you
22:07:28DJ_PUNK2011I have a question 2 you maybe you know this link
22:07:30AlexPDJ_PUNK2011: Please stay on topic
22:07:31DJ_PUNK2011I found a music
22:07:32bluebrotherDJ_PUNK2011: this is a technical channel.
22:07:54DJ_PUNK2011blue brother just need to write a question which might know you
22:08:08AlexPDJ_PUNK2011: If it is not to do with Rockbox, then don't
22:08:12DJ_PUNK2011does anybody from you know the singer or a band of this music? Please find youtube link below:
22:08:18AlexPDJ_PUNK2011: Stop it
22:08:49 Quit ej0rge (Read error: Operation timed out)
22:08:52DJ_PUNK2011ok I stopped
22:09:03DJ_PUNK2011just wrote a question
22:09:22AlexPThis channel is for support and development of Rockbox only
22:09:27AlexPNot random quetions
22:09:34DJ_PUNK2011a I c
22:09:38DJ_PUNK2011Sorry for that Alex
22:09:50DJ_PUNK2011I thought this is just normal chat conversations
22:09:53Mode"#rockbox +o B4gder" by ChanServ (ChanServ@services.)
22:11:57Zagormcuelenaere: which API is internal in your fs#11914 patch8?
22:12:43 Quit DJ_PUNK2011 (Quit: DJ_PUNK2011)
22:14:24 Join kevku [0] (~kevku@2001:7d0:0:f9af:feed:feed:feed:feed)
22:15:18Mode"#rockbox -o B4gder" by ChanServ (ChanServ@services.)
22:18:16mcuelenaereit isn't documented, but you can find more about it in AOSP
22:19:05 Part mtstickney
22:21:23Zagorok. do you want to rework and commit that patch, or do you want me to?
22:21:41mcuelenaereZagor: RE r29573, the RockboxService variable was made final because that's a requirement when accessing it from an anonymous class, so strictly speaking it isn't needed in your case
22:22:50Zagorok. my java fu is very limited.
22:23:49mcuelenaereyeah, I'm not sure why it's needed either, Eclipse just suggested that :)
22:24:22 Join ej0rge [0] (
22:27:44kugelZagor: android's volume popup is gone :(
22:31:55Zagoryes, disabled since we use our own volume display.
22:32:01***Saving seen data "./dancer.seen"
22:32:12kugelcan we enable it?
22:32:48Zagorof course we can, but it doesn't show the actual volume. only the low-res android steps.
22:32:55Zagorso it is rather misleading
22:35:10kugelI think I would still like it
22:35:41ZagorI think it only makes sense if you disable high resolution volume
22:36:35Zagorotherwise the popup will show the same "value" several times, indicating volume adjustment is broken
22:36:45kugelI realize that
22:45:41kugelbut I don't think it'd be perceived as borken, but just as "different"
22:46:07ZagorI disagree. the user presses the button repeatedly, sees no effect and hears no effect. ergo, broken.
22:46:20kugelhe hears the effect?
22:47:14Zagorno. the steps are so small it is very easy to miss one or two. especially if the user expects the normal big steps.
22:48:23kugelwe could have larger steps anyway
22:48:28kugel100 are a bit too much
22:48:39saratogahow many does android have
22:49:02Zagorsaratoga: different for different models. nexus one has 16 steps
22:49:04kugelmcuelenaere: I'm still not a fan of r29569
22:49:20saratogawow not very many
22:49:38Zagorno. kugels has as few as 9.
22:49:46Zagorhence the need for this fix
22:49:59saratogawhy does it vary so widely by device?
22:50:06ZagorI don't know
22:50:09AlexPmine has 8 I think
22:50:15kugelZagor: anyway it seems to work on my phone
22:50:22 Join bimbel [0] (~bmbl@unaffiliated/bmbl)
22:50:22Zagorkugel: great!
22:50:23kugelsteps are smooth as far as i can tell
22:50:36 Quit bmbl (Ping timeout: 241 seconds)
22:50:50AlexPOh, no mine has 16
22:50:58AlexP8 is the ringer volume :)
22:51:08Zagorright, it is different for different stream types
22:51:21saratogahuh mine has 15 as well
22:51:50kugelwait, it seems the volume goes lower for a fraction of second when pressing the key
22:51:59mcuelenaerekugel: it could be reverted, but then we need to ensure that we don't yield when a native function is called with a different JNI environment (which isn't guaranteed in the PCM code IIRC)
22:53:13kugelPCM uses it's own env in the callback
22:53:13saratogaFWIW i don't see the graphic not updating ever step as that big a deal, i'm sure i've seen various devices that do that, and of course a lot of rockbox skins that use a volume icon
22:53:13kugelwhat do you mean by yield?
22:53:13saratogaalthough making sure the steps were 1.5 to 2dB would be nice
22:53:13saratogaso you can hear them easily
22:53:13kugela rockbox yield? all rockbox threads have the same env because it's within 1 pthread
22:54:08Zagorc200 has 80 steps
22:54:18kugelsaratoga: on cabbie it was never a problem that the graphic doesn't change with every step
22:54:36mcuelenaerekugel: yes, a rockbox yield
22:54:59kugelI don't understand why it is a problem that the android display doesn't change, while cabbie also doesn't update
22:55:08kugelcabbie volume display has even less states I think
22:55:26Zagorkugel: because users expect the volume widget to change when the press the button. it always does.
22:55:58AlexPWe really have to start to think a bit differently for touchscreen vs standard devices IMO
22:56:02Zagorwith rockbox-specific things, nobody has any expectations and we can do more whatever we like.
22:56:26kugelmcuelenaere: a rockbox yield is safe
22:56:28Zagorwith host OS things, we cannot be too confusing
22:57:04CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
22:57:04*mcuelenaere tries to remember
22:57:13mcuelenaerekugel: does the PCM java code run in a separate (p)thread?
22:57:24kugelonly the callback
22:58:01mcuelenaereIIRC the problem I had that "long time ago", was that in the callback there was a yield() somehow, which triggered an lcd_update() in a different rockbox thread
22:58:23mcuelenaereand as the JNI environment in the LCD subsystem belongs to a different thread, it crashed
22:58:28mcuelenaereor it gave a warning at least
22:58:40kugelI thought I ensured the env handling is done safely, but I didn't add comments regarding that (my mistake). and if there's something I overlooked please point me at it
22:59:06kugelin that long time the lcd subsystem was rewritten several times :)
22:59:36kugelwe know update from within the rockbox threads, early we posted to the UI thread
23:00:03Zagorheh, ak4537 has 128 steps
23:00:29kugelI'd be happy with just 50 steps, I mostly only use the top 20 steps anyway
23:00:46Zagoryeah 50 is probably enough
23:01:00Zagorconfigurable! ;-)
23:01:03 Quit Stummi (Quit: Bye!)
23:01:24kugeland we should either make it a real db scale or change the volume display in rockbox to be percentage based
23:01:33kugel-99 to 0 is missleading too
23:01:53Zagorwe don't have db information on android
23:02:10 Join Buschel [0] (
23:02:20ZagorI don't think it's a big problem though. it is attenuation, the unit is not terribly important
23:03:16BuschelJdGordon|: did you see FS #11997 ? It seems this is connected to some WPS stuff when displaying album art
23:03:39mcuelenaerekugel: at 00.57.30 and following
23:04:13kugelthat can't be :)
23:04:14mcuelenaerelong time ago ~= 5 monhts ;)
23:04:28kugelI think I talked about this with someone else recently
23:04:57kugelit should be possible to sleep() from the callback. that would even deadlock on target
23:05:13kugelbecause it's most often called from a DMA isr
23:05:41kugelit should *not*
23:05:50mcuelenaereI know :)
23:06:00 Quit FoH (Quit: process management issues)
23:06:01Zagorsurely sleep has an interrupt check?
23:06:29mcuelenaereI couldn't figure out any other way to explain it at that time
23:06:45mcuelenaereproblem was that it's hard to trigger
23:07:17kugelit's hard to prove without backtrace but it really would cause weird things to happen on a native taret
23:08:58mcuelenaereI'm currently trying to get native backtracing working :)
23:09:13mcuelenaeres/trying/looking into/
23:09:34kugelZagor: no it hasnt
23:10:04kugelthere isn't even an easy way to tell if you're in interrupt context
23:11:12mcuelenaeresetting a variable to true when entering the interrupt handler? ;)
23:11:22Zagorthe kernel knows
23:11:46kugelZagor: I'm pretty sure it doesn't
23:11:54mcuelenaereZagor: rockbox kernel?
23:12:05kugelthe new semaphores have a semi-check. it's only an indicator though
23:12:11Zagorkugel: uh, yes it does
23:12:31Zagormcuelenaere: linux. I thought you were discussing android?
23:12:46mcuelenaereZagor: yes, but this is Rockbox internals?
23:13:03mcuelenaereI mean, Rockbox interrupts aren't *real* interrupts on hosted targets
23:13:05kugelwe talk about the rockbox kernel and our sleep()
23:13:08 Quit gbl08ma (Ping timeout: 246 seconds)
23:16:06 Quit mshathlonxp (Quit: Leaving)
23:16:14kugelmcuelenaere: you didn't by chance debug what threadid=10 back then did you?
23:16:17 Join mshathlonxp [0] (
23:16:32mcuelenaerekugel: I'm afraid the IRC logs are all we can rely on
23:16:57mcuelenaerekugel: what do you mean with what threadid=10 is? I think it just means the thread with ID=10 :)
23:17:11kugelyea, but what thread :)
23:17:16mcuelenaereah no, sorry
23:17:24 Join gbl08ma [0] (
23:18:03gbl08ma[this IRC client is driving me nuts]
23:18:39kugelI think our env handling is pretty safe and we shouldn't need to add additional bloat to it
23:20:54mcuelenaereperhaps I could eliminate some getJavaEnvironment() calls, by passing the env around
23:22:12*kugel maintains that getJavaEnvironment() isn't needed at all
23:22:14 Quit mshathlonxp (Quit: Leaving)
23:22:20kugelhave you actually been able to repro it sense then?
23:23:25mcuelenaerelet me try that now
23:23:32kugelZagor: do you also hear what I described?
23:24:03 Join timccc [0] (~timccc@
23:24:07Zagorkugel: a dip in volume when pressing the button? I can't hear anything like that
23:24:22kugelit's only for a very short moment
23:24:46Zagorevery step, or only some?
23:24:50 Join mshathlonxp [0] (
23:25:09kugelit goes the other direction as I press actually
23:25:23Zagoris it more audible if you hold the button to repeat it?
23:25:55*kugel starts thinking this could be a CM issue
23:26:33kugelperhaps there are spurious volume button presses. that
23:27:19kugelthat'd explain my 8 steps and the fact that there's always also a volume bump when I use long-vol press to switch tracks
23:28:30kugelZagor: it's very audible with short presses, not so much with long presses
23:28:54kugelif I press vol down the volume is bumped very shortly (and vice versa with vol up)
23:29:56 Quit tails__ (Ping timeout: 246 seconds)
23:30:05 Join kronflux [0] (
23:30:08kugelit doesn't seem to happen when I choose the volume in the list with the trackball
23:30:09kronflux #harmonymc
23:33:21 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
23:33:58 Part toffe82_
23:35:23mcuelenaerekugel: btw, threadid=10 was the "audio thread": '"audio thread" prio=5 tid=10 NATIVE' (
23:35:31kugelZagor: btw, thanks for that fade on stop commit :)
23:35:48Zagoryou're welcome! :)
23:36:03kugelmcuelenaere: oh, I thought tid=7 was that one
23:36:28mcuelenaerekugel: I'm not sure, but I think that log is pretty right :)
23:36:58kugelthere's only one audio thread, and that calls pcmSamplesToByteArray
23:37:51mcuelenaerewow :) I just played an MP3 in the emulator and it just crashed fully :)
23:38:09kugelthat's my last experience with it also from a few month back
23:38:19mcuelenaerehaven't had that one before
23:38:37kugelI generally stopped using the emulator because of that
23:41:45gbl08mawhat emulator? I have been playing mp3 to take screenshots of my theme ports on the sim of nano1g and it plays without problems (don't know about sound quality, but that's not the goal of the sim)
23:42:01mcuelenaerecrap, that's reproducable..
23:42:28Zagorthe volume works nicely on my old 1.6 acer liquid too
23:43:14AlexPgbl08ma: android emulator, not Rockbox simulator
23:43:35gbl08mait's been the whole day developing for android, hasn't it?
23:43:44*mcuelenaere hopes this is due to his Makefile changes
23:44:00mcuelenaerestill, QEMU shouldn't crash..
23:48:16CIA-2New commit by kugel (r29574): Create GlobalRef in surfaceCreated, in accordance to r29572.
23:52:24CIA-2r29574 build result: All green
23:53:37 Quit ej0rge (Ping timeout: 260 seconds)
23:54:17mcuelenaereperhaps this is also one of those "extended" JNI checks which are disabled on real targets?
23:56:13kugelperhaps. can you enable them on real?
23:56:32mcuelenaereyes and yes
23:56:40mcuelenaereadb shell setprop dalvik.vm.checkjni true
23:56:42mcuelenaererequires root
23:56:46*mcuelenaere is trying right now

Previous day | Next day