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 2010-07-05

00:08:21 Quit saratoga_ (Quit: Page closed)
00:14:23 Quit ender` (Quit: I will not include a self-destruct mechanism unless absolutely necessary. If it is necessary, it will not be a large red button labelled "Danger: Do Not Push". The big red button marked "Do Not Push" will instead trigger a spray of bullets on anyone stu)
00:22:44bertrikfunman, I'm tempted to commit my fix for the alarm poweroff problem too. For me, the problem went from completely reproducible to not reproducible (tried 5 times or so).
00:23:28bertrikmaybe tomorrow, after the battery bench ... :)
00:25:29funmancan you send me the patch?
00:25:45funmanmy clipv2 just acted strange: the screen lit up itself and now it's not responsive
00:25:47 Join Luca_S [0] (
00:26:16bertrikI thought you already tried it:
00:26:19Luca_Sfunman: my girlfriend borrowed my fuzev2, sorry, I can't help with benching for a couple of days
00:26:48 Quit einhirn (Read error: Connection reset by peer)
00:27:12Luca_Sgood night rockboxers :)
00:27:16 Quit Luca_S (Client Quit)
00:27:21funmanbertrik: well yes if it doesn't work then it's not a fix?
00:27:38bertrik"works for me" :/
00:28:58 Nick detaos_ is now known as detaos (
00:29:11funmando you actually use this feature ?
00:29:32bertrikno, not until I started testing it
00:30:09 Quit Jerom (Read error: Connection reset by peer)
00:31:14funmanThe lock-up happens more often when you want to enable it and go to bad (Murphy's law)
00:31:21funmanjust go to bed*
00:32:39funmanbertrik: can you reproduce the bug without the patch ?
00:32:45bertrikhm, so maybe this bug is more likely at a particular time of day (or night) :)
00:33:24bertrikyes, for me it always happens the second time, when it has woken up from the first alarm
00:33:38CIA-6New commit by pixelma (r27285): Close the \ActionStdCancel correctly so that the following space won't be eaten.
00:34:42 Quit pamaury (Ping timeout: 265 seconds)
00:35:30 Join pamaury [0] (
00:35:30 Quit pamaury (Changing host)
00:35:30 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
00:35:34CIA-6r27285 build result: All green
00:36:20 Quit S_a_i_n_t (Read error: Connection reset by peer)
00:39:21 Join krabador [0] (
00:39:34 Join S_a_i_n_t [0] (S_a_i_n_t@
00:39:39 Quit pamaury (Client Quit)
00:45:33***Saving seen data "./dancer.seen"
00:50:55 Quit bertrik (Quit: :tiuQ)
00:52:55 Quit liar (Quit: ...)
01:02:53 Quit BHSPitMini (Ping timeout: 260 seconds)
01:07:42 Quit funman (Quit: leaving)
01:11:49 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
01:36:48 Quit Dhraakellian (Ping timeout: 240 seconds)
01:43:31 Join Dhraakellian [0] (
02:11:20 Join HommeBarbu [0] (
02:11:36HommeBarbu:join #mhvlug
02:12:11 Quit HommeBarbu (Client Quit)
02:12:55 Join HommeBarbu [0] (
02:13:08 Part HommeBarbu
02:21:02 Quit S_a_i_n_t (Ping timeout: 265 seconds)
02:21:22 Join S_a_i_n_t [0] (S_a_i_n_t@
02:24:54 Quit xavieran (Remote host closed the connection)
02:26:34 Join xavieran [0] (
02:27:56 Join halmi [0] (
02:29:00 Quit halmi (Read error: Connection reset by peer)
02:42:47 Quit BlakeJohnson86 (Ping timeout: 245 seconds)
02:44:21 Join BHSPitMini [0] (
02:45:36***Saving seen data "./dancer.seen"
02:47:11 Quit simonrvn (Read error: No route to host)
03:00:01 Join [Saint] [0] (S_a_i_n_t@
03:00:32 Quit balintx (Read error: Connection reset by peer)
03:00:47 Quit S_a_i_n_t (Ping timeout: 260 seconds)
03:02:42 Join balintx [0] (
03:08:23 Quit GeekShadow (Quit: The cake is a lie !)
03:14:18 Join wincent [0] (
03:14:37 Quit wincent (Changing host)
03:14:37 Join wincent [0] (~Miranda@rockbox/developer/wincent)
03:15:46 Quit rado1 (Quit: Leaving.)
03:22:24 Quit BHSPitMini (Ping timeout: 276 seconds)
03:26:43CIA-6New commit by wincent (r27286): pdbox: Added different sampling rates based on hardware capabilities. ...
03:28:38CIA-6r27286 build result: All green
03:31:28 Join simonrvn [0] (
03:31:58CIA-6New commit by wincent (r27287): pdbox: Added Sansa e200 to supported devices.
03:33:38CIA-6r27287 build result: All green
03:34:03 Quit [Saint] (Ping timeout: 245 seconds)
03:34:36 Quit JdGordon (Quit: Leaving.)
03:35:06CIA-6New commit by wincent (r27288): pdbox: Optimizations in the audio output code.
03:36:56CIA-6r27288 build result: All green
03:41:00 Quit togetic (Ping timeout: 240 seconds)
03:55:30 Join togetic [0] (~togetic@unaffiliated/ibuffy)
04:03:20 Quit leavittx_ (Ping timeout: 240 seconds)
04:03:20 Quit leavittx (Ping timeout: 240 seconds)
04:03:43 Join leavittx [0] (~leavittx@
04:03:45 Join leavittx_ [0] (~leavittx@
04:04:58 Quit sneak (Remote host closed the connection)
04:08:48 Join sneak [0] (debian-tor@gateway/tor-sasl/sneak)
04:11:34 Quit MethoS- (Remote host closed the connection)
04:19:25 Quit amiconn (Disconnected by services)
04:19:27 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:19:47 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:19:47 Quit pixelma (Disconnected by services)
04:19:49 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:20:04 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:26:24 Join S_a_i_n_t [0] (S_a_i_n_t@
04:31:25 Quit binaryhermit (Read error: Connection reset by peer)
04:31:48 Join binaryhermit [0] (
04:35:57 Join DerPapst [0] (
04:38:14 Quit DerPapst1 (Ping timeout: 265 seconds)
04:38:20 Quit TheSeven (Ping timeout: 260 seconds)
04:42:04 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:45:40***Saving seen data "./dancer.seen"
04:50:16 Quit panni__ (Quit: ( :: NoNameScript 3.81 :: ))
04:53:42 Join BlakeJohnson86 [0] (
04:55:16 Join JdGord [0] (~jd@
05:17:01CIA-6New commit by saratoga (r27289): Move the test and benchmark plugins to the top of the viewers list, since if someone compiles a build with them, they're probably going to use them ...
05:17:18 Quit JdGord (Quit: Bye)
05:17:38*S_a_i_n_t wonders why ther aren't alphabetical.
05:17:45S_a_i_n_t(plugins, viewers, etc.)
05:18:01S_a_i_n_tI have a patch I made to do that floating around somewhere...
05:18:46CIA-6r27289 build result: All green
05:29:25saratogafinally got test_codec to decode a 20MB flac on my Clip+
05:30:57 Quit Horscht (Quit: Verlassend)
05:43:58 Quit krabador (Remote host closed the connection)
05:57:08 Quit anewuser (Quit: Sixth edition of 25 samples + 48 hours + detroit party (july 31) (maybe streamed online!))
06:04:26 Quit simonrvn (Read error: No route to host)
06:07:26 Quit S_a_i_n_t (Read error: Connection reset by peer)
06:07:28 Join jennifur [0] (
06:15:41 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
06:16:58 Join simonrvn [0] (
06:24:11 Nick fleebailey33 is now known as flee_freenode (~fleebaile@unaffiliated/fleebailey33)
06:45:41***Saving seen data "./dancer.seen"
06:57:03 Nick flee_freenode is now known as fleebailey33 (~fleebaile@unaffiliated/fleebailey33)
06:57:11 Quit kramer3d (Quit: Leaving)
07:01:38 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
07:05:10 Quit froggyman (Ping timeout: 265 seconds)
07:23:34 Quit binaryhermit (Ping timeout: 245 seconds)
07:25:33saratogaanyone feel like running test codec on the fuzev2?
07:25:40jennifurwhat codec?
07:25:46saratogaall of them!
07:25:50jennifursure, why not
07:26:37saratogayou'll need FS #11458 and the test_files folder off of
07:27:29jennifurdo i need to do svn update first?
07:27:48saratogayeah you should run this on an up to date otherwise unmodified SVN checkout
07:32:10 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
07:36:09saratogai just increased PCLK to 40MHz on my clipv2 and the speed up is huge
07:36:36saratogaassuming I did this right, WMA decodes in under 30MHz
07:42:34jennifurwell, just waiting for the test files to download, then i can transfer them over
07:43:06saratogaonce you have them, compile a build with test_codec.c added to apps/plugins/SOURCES
07:43:32saratogaand then "open with" one of the test_files and choose test codec, then do the whole directory when it asks
07:46:03 Join stoffel [0] (
07:47:46pixelmasaratoga: you know of the test plugins advanced build option? ;)
07:48:49pixelmano need to edit SOURCES, although with the advanced build option you'll get the whole bunch of test plugins
07:50:23jennifursaratoga, i'm getting an err when it tries to compile, pastebin of the err is here:
07:51:25jennifuroh wait never mind
08:00:33 Quit scorche (Ping timeout: 276 seconds)
08:00:55 Join scorche [0] (~scorche@rockbox/administrator/scorche)
08:07:29 Join kugel [0] (kugel@rockbox/developer/kugel)
08:07:40jennifursaratoga, which one? there is more than one option here.
08:09:43saratogatest whole directory or something like that
08:09:58jennifurtrying the speed test one first
08:09:58saratoga"speed test folder"
08:09:59kugelsaratoga: how did you get test_codec to work with the 20mb file?
08:11:42saratogakugel: lots of rebuffering
08:11:51kugelis there a patch?
08:11:59saratogakugel: yeah its on the tracker
08:12:06saratogaFS #11458
08:13:02saratogafunman: (for the logs) PCLK has a huge impact on vorbis and WMA performance, but on other formats the difference is smaller
08:13:26saratogavorbis at 24MHz is just 62% the speed of vorbis at 40MHz
08:14:44kugelnice (the patch)
08:15:17kugelsaratoga: PCLK affects the RAM speed IIUC
08:16:04saratogakugel: IRAM mainly is whats being tested here
08:16:24saratogavorbis seems almost entirely limited by memory speed at 250MHz
08:17:13saratogabut MP3 sees only a very minor difference, i guess since its filterbank is so multiplier heavy
08:18:35kugelsaratoga: sorry, I mean RAM&IRAM.
08:19:05kugelon amsv2 the codec buffer is entirely in IRAM
08:20:45saratogai need to do a test with no boosting at all
08:23:31kugelsaratoga: the tests where made with the PCLK change but still with boosted @240MHz?
08:26:45saratogakugel: yeah, i just now started a test with pclk = fclk = 40mhz
08:27:11saratogai bet vorbis is a lot faster per MHz
08:28:00jennifurthe test is still running by the way saratoga
08:29:10saratogapastebin the results when you get them, should be the same as mine I think
08:29:50jennifurabout how long did it take for you?
08:30:07saratogamaybe half an hour
08:36:03saratogaoh thats right, we disabled boosting on amsv2 . . .
08:36:43 Quit kugel (Quit: exit(0);)
08:37:35 Join chid [0] (
08:45:08 Join einhirn [0] (
08:45:46***Saving seen data "./dancer.seen"
08:46:14 Join Keaft [0] (43a4cbd3@gateway/web/freenode/ip.
08:46:30KeaftAnyone alive?
08:46:36jennifuri am
08:46:40 Join ender` [0] (
08:46:49jennifurthough, i could be a bot
08:46:51Keaftoh good I thought that the world had ended for a moment there
08:47:09Keaftare you a dev?
08:47:22Keaftor someone that has answers to questions?
08:47:33jennifuri might have an answer to your question
08:47:38jennifurnever know until you ask
08:47:42Keaftvery true
08:48:42 Join mitk [0] (~mitk@
08:48:49jennifurso, what is your question Keaft?
08:48:58Keaftok so I'd imagine there has been a lot asked about the ipod nano 5g so instead of saying WHEN I figured I'd ask what would be the best mailing list to get on?
08:49:29saratoga20.1MHz for vorbis 128k when PCLK = FCLK!
08:49:45Keaftif that made any sense... shoot it's way late for me lol
08:49:49saratogaverses 45MHz when PCLK=24 and FCLK=240
08:50:36saratogatheres no obvious list to get on, although i guess you could watch the rockbox or linux4nano dev lists and see if anyone ever mentions working on it
08:50:48 Join S_a_i_n_t [0] (S_a_i_n_t@
08:50:53Keaftsounds good.
08:51:12KeaftI could probably sign up and then filter in gmail for "Ipod" or something like that
08:52:56saratogathese results really call into question our assumptions about optimization, at least on AMS
08:53:45saratogaif performance can vary so dramatically and unpredictably with clock speed, then we should probably not test things with the clock much higher then its likely to be
08:55:25 Quit Keaft ()
09:00:54S_a_i_n_tAh, crap. He left.
09:01:26S_a_i_n_tAFAIK noone's even got further past considering how hard the Nano5G may or may not be.
09:09:01 Quit stoffel (Remote host closed the connection)
09:10:47 Join halmi [0] (
09:18:55 Join flydutch [0] (
09:31:28 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
09:37:03 Quit DerPapst (Quit: Leaving.)
09:45:14 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
09:53:52 Join Rob2222 [0] (
09:57:02S_a_i_n_tHmmmm...%aL and %aR don't seem to work correctly.
09:57:19 Quit Rob2223 (Ping timeout: 245 seconds)
10:06:53S_a_i_n_tYeah, I'm trying to figure out exactly what the problem is though.
10:07:19S_a_i_n_tit's one of those "I know something is wrong, but what the fuck is it?" things.
10:08:26S_a_i_n_tI couldn;t find any documentation on %Sr do you know if it is %Sr<LTR|RTL> or %Sr<RTL|LTR>?
10:12:53JdGordonumm.. the "yes" branch is rtl
10:12:57 Join dfkt [0] (dfkt@unaffiliated/dfkt)
10:13:06JdGordonso <rtl|ltr> I think?
10:13:18*JdGordon can never remember if it is true|false or false|true
10:13:33S_a_i_n_tKnowing that will certainly help me put my finger on this.
10:13:47S_a_i_n_ttrue|false btw
10:17:37 Join tagtype [0] (
10:17:53tagtypedoes rockbox read ID3v2.4 and v2.3, or only v2.3 ?
10:18:46S_a_i_n_tGood question, I believe the manual was the answer to that.
10:18:59S_a_i_n_tI can't answer off the top of my head sorry.
10:19:59S_a_i_n_tJdGordon: It seems like just normal old %al actually right aligns text when the text is RTL?!?
10:20:05pixelmait reads ID3v2.4
10:20:12S_a_i_n_tperhaps it has something to do with the %ax tag.
10:20:27JdGordonax might swap al/ar
10:20:48S_a_i_n_tIt seems it does.
10:20:51JdGordonhmm, no it sholdnt
10:21:14JdGordonax before the viewport?
10:21:16S_a_i_n_t%aL left aligns the text when it is RTL in my WPS
10:21:46S_a_i_n_tI'll try taking out %ax and see if %aL works then.
10:22:55 Join bimbel [0] (~Miranda@unaffiliated/bmbl)
10:23:03 Quit kramer3d (Quit: Leaving)
10:23:04 Quit tagtype (Quit: CGI:IRC (EOF))
10:24:11S_a_i_n_tJdGordon: Yeah, just tested it then, %ax is flipping the alignment tags also.
10:24:27JdGordonnot according to the code
10:24:29JdGordonpastebin me
10:24:57 Quit bmbl (Ping timeout: 240 seconds)
10:25:04 Nick bimbel is now known as bmbl (~Miranda@unaffiliated/bmbl)
10:27:22S_a_i_n_tJust remove %ax and the alignment works as intended.
10:28:02JdGordonyeah because it is telling the viewport to automagically set the alignment correctly
10:28:16JdGordonthat is a wierd sideeffect
10:28:28S_a_i_n_tand, if I use %ax, and %al instead of %aL then it right aligns instead of left.
10:28:48JdGordonmaybe %aL/R are redundant
10:29:07S_a_i_n_tThis "feature-bug" kinda makes them so, yes.
10:29:35S_a_i_n_tWith this oddity, then %Sr is plenty.
10:30:09JdGordonyou cant put ax inside sr
10:31:25S_a_i_n_tIt wasn;t inside it.
10:32:13S_a_i_n_tThe section of code I pasted didn;t actually use %Sr at all.
10:32:36S_a_i_n_tPerhaps I misunderstood.
10:33:27S_a_i_n_tWhat I meant was, instead of %ax, if you don'T WANT TO ACTUALLY MIRROR THE VIEWPORT THEN 5sR WILL WORK SUFFICIENTLY.
10:33:35S_a_i_n_targh! stupid caps.
10:36:30 Join DerPapst [0] (
10:39:22 Join swilde [0] (
10:40:47S_a_i_n_tJdGordon: Here's an interesting one...can the %Fl line include/be included in a conditional statement? ie. ?Sr<%Fl(fontAdetails)|%Fl(fontBdetails)>
10:42:21 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
10:42:25 Quit gevaerts (Ping timeout: 265 seconds)
10:45:50***Saving seen data "./dancer.seen"
11:02:13 Join giovanni_ [0] (~giovanni@
11:03:54giovanni_pixelma: I've seen that you've committed my patch, be aware that I've also modified the table that explains the touch screen 3x3 grid in the manual.
11:05:39pixelmayeah, I saw that and actually wanted to have a look where the difference is but forgot :/
11:06:56JdGordonS_a_i_n_t: of course it can
11:07:09pixelmagiovanni_: could you describe how it looks differently now in a few words?
11:07:10JdGordonit wont do what you want.. but you can put it there :p
11:07:27S_a_i_n_tHardy har har...
11:08:11S_a_i_n_tI need a way of switching to Unifont for the RTL languages.
11:08:28S_a_i_n_tPreferably one that isn't making every text viewport conditional.
11:08:48pixelmaconditional viewports should work, I guess
11:08:59giovanni_yes, now the table is a little bit larger and the words are centered.
11:09:05S_a_i_n_tyeah, that was what I was trying to avoid.
11:10:16pixelmagiovanni_: that's ok I think, just not mentioned in the commit message (the commit was messed up anyways :/ )
11:10:49giovanni_no problem for me, I only hope I've made it the right way
11:11:54 Join Jaykay [0] (
11:13:47giovanni_pixelma: I can paste the diff part on pastebin so you can see
11:14:27pixelmayou don't need to, I still have the full patch around
11:15:31 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
11:16:38 Quit jennifur (Remote host closed the connection)
11:16:45JdGordonS_a_i_n_t: stick it on the wishlist... I might be able to come up with a system to make conditional font loading work
11:17:01JdGordonor at least load both but font '8' can be either of them
11:18:00mitkHi. Is here an dev who can commit FS #11436 - Polish charset for 08-Nedore font?
11:18:53S_a_i_n_tHmmm, for now I basically have to make a copy of each text viewport with Unifont loaded and use %Sr to switch it.
11:19:10S_a_i_n_tthat's going to make for a *massive* conditional switch lol ;)
11:19:35bertrikmitk, sure, I can do that
11:20:00giovanni_I need some hint to change the key assignment for pacbox
11:20:41pixelmamitk: can you remind me in the evening - and would you add them to 09-Nedore too?
11:20:52JdGordonS_a_i_n_t: why not just always use unifont?
11:20:58pixelmaoh, bertrik was quicker
11:21:04S_a_i_n_tIt's hideous.
11:21:15mitkbertrik: Great. Thanks.
11:21:37 Quit balintx (Remote host closed the connection)
11:21:48*S_a_i_n_t hates fixed width fonts, especially ugly ones.
11:21:52 Join balintx [0] (
11:22:10S_a_i_n_tThe only reason to use it is becasue it has hebrew/arabic chars
11:23:17CIA-6New commit by bertrik (r27290): Polish charset for 08-Nedore font, FS #11436 by Tomasz Kowalczyk
11:23:24mitkpixelma: Yes, I can do also 09, but following "one change in one FS" rule I should do it in another patch I think
11:24:07mitkbertrik is so fast...
11:24:25bertrikhm, yeah, I just closed the task
11:24:28 Join robin0800 [0] (
11:25:01CIA-6r27290 build result: All green
11:27:16 Quit andash (Read error: Operation timed out)
11:28:15CIA-6New commit by bertrik (r27291): Serbian translation update, FS #11456 by Ivan Pesic
11:29:52CIA-6r27291 build result: All green
11:43:19 Join GeekShad0w [0] (
11:45:38 Quit GeekShadow (Ping timeout: 264 seconds)
11:46:48pixelmamitk: you are right about one step at a time, I was just asking :)
12:01:56 Join watto [0] (~watto@
12:05:23 Quit JdGordon (Read error: Connection reset by peer)
12:15:34 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
12:28:08 Quit togetic (*.net *.split)
12:30:53bertrikcool, the problem with a keypress getting ignored after exiting a plugin is reproducible in the sim
12:31:51bertrikat least, the one I see with the dict plugin
12:34:36 Join togetic [0] (~togetic@unaffiliated/ibuffy)
12:36:18 Join andash [0] (
12:36:19 Quit andash (Changing host)
12:36:19 Join andash [0] (~andash@unaffiliated/andash)
12:43:33 Join liar [0] (
12:45:53***Saving seen data "./dancer.seen"
12:46:26 Join halmi_ [0] (
12:49:14 Quit halmi (Ping timeout: 264 seconds)
12:58:19 Join halmi [0] (
13:00:29 Quit halmi_ (Ping timeout: 240 seconds)
13:11:05 Join halmi_ [0] (
13:11:49 Quit GeekShad0w (Ping timeout: 240 seconds)
13:13:29 Quit halmi (Ping timeout: 248 seconds)
13:14:18 Join pamaury [0] (
13:14:18 Quit pamaury (Changing host)
13:14:18 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
13:23:18mitkpixelma, bertrik: FS #11460 - Polish charset for 09-Nedore font. Can you?
13:23:38 Quit pamaury (Ping timeout: 265 seconds)
13:24:00bertrikyeah, in 20 minutes or so
13:24:30 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
13:26:12pixelmamitk: currently not. But thanks anyways :)
13:27:53bertrikgevaerts, I think a 1 minute difference in battery bench runtime is probably not significant
13:28:04gevaerts*probably* ;)
13:29:30bertrikdfkt found a 7-minute improvement from using PLLB on his clip+, I found a 1-minute deterioration, so I'd say the effect is negligible
13:40:58 Quit GeekShadow (Read error: Connection reset by peer)
13:53:37 Join BdN3504 [0] (
13:54:05CIA-6New commit by bertrik (r27292): Polish charset for 09-Nedore font, FS #11460 by Tomasz Kowalczyk
13:55:56 Join BdN3504-2 [0] (
13:55:56 Quit BdN3504 (Client Quit)
13:56:02CIA-6r27292 build result: All green
13:56:55 Join teru [0] (
13:57:25mitkbertrik: thank you
13:59:22BdN3504-2sandisk sansa clip+ r27220:How do enable "hold" on this device?
13:59:51BdN3504-2spellcheck: "how do i..."
13:59:55S_a_i_n_tI believe it has a software hold
14:00:02S_a_i_n_t(a button combination)
14:00:35BdN3504-2i searched the pdf version of the manual for hold, but that didn't turn up anything useful, that's why i ask here
14:00:45mitkhome+select combo
14:01:18BdN3504-2ah thanks!
14:02:12 Quit BdN3504-2 (Quit: CGI:IRC (EOF))
14:07:40 Quit YPSY (Ping timeout: 260 seconds)
14:14:22 Quit halmi_ (Quit: halmi_)
14:14:34CIA-6New commit by uchida (r27293): text viewer: fixed StatusBar randomly appears (FS #11455).
14:16:10CIA-6r27293 build result: All green
14:18:26 Join MethoS- [0] (~clemens@
14:22:26 Join Ypsy [0] (
14:22:35 Nick Ypsy is now known as YPSY (
14:26:24 Join Rob2223 [0] (
14:29:36 Quit Rob2222 (Ping timeout: 260 seconds)
14:34:15 Join panni_ [0] (
14:34:48 Join Rob2222 [0] (
14:38:32 Quit Rob2223 (Ping timeout: 276 seconds)
14:45:57***Saving seen data "./dancer.seen"
14:46:16 Quit bieber (Ping timeout: 240 seconds)
14:46:29 Join bieber [0] (
14:51:05 Quit bieber (Ping timeout: 248 seconds)
14:51:22 Join bieber [0] (
14:52:16 Join kugel [0] (~kugel@rockbox/developer/kugel)
14:53:44 Quit Kitar|st (Ping timeout: 252 seconds)
15:00:14 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
15:01:17 Quit antil33t (Read error: Connection reset by peer)
15:01:23 Join antil33t [0] (
15:02:22 Join komputes [0] (~komputes@ubuntu/member/komputes)
15:17:04 Join kugel_ [0] (
15:17:27 Quit kugel (Disconnected by services)
15:17:30 Nick kugel_ is now known as kugel (
15:17:35 Quit kugel (Changing host)
15:17:35 Join kugel [0] (~kugel@rockbox/developer/kugel)
15:23:39 Quit teru (Quit: Quit)
15:25:27 Quit storm` ()
15:26:54 Join krabador [0] (
15:28:00 Quit krabador (Remote host closed the connection)
15:28:16 Quit komputes (Ping timeout: 240 seconds)
15:40:14 Join komputes [0] (~komputes@ubuntu/member/komputes)
15:42:33 Join Xerion [0] (
15:46:23 Join Kitar|st [0] (
15:46:48 Quit mitk (Quit: Leaving)
15:46:55 Join Kitr88 [0] (
15:47:05 Quit Kitr88 (Client Quit)
16:16:41 Quit einhirn (Read error: Connection reset by peer)
16:20:32 Quit kugel (Ping timeout: 265 seconds)
16:27:24 Join halmi [0] (
16:29:11 Quit bluebrother (Disconnected by services)
16:29:14 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
16:42:32 Join merbanan [0] (
16:45:59***Saving seen data "./dancer.seen"
16:49:57 Join einhirn [0] (
16:52:32 Join kugel [0] (~kugel@rockbox/developer/kugel)
17:22:13 Join kugel_ [0] (
17:22:35 Quit kugel (Disconnected by services)
17:22:39 Nick kugel_ is now known as kugel (
17:22:43 Quit kugel (Changing host)
17:22:43 Join kugel [0] (~kugel@rockbox/developer/kugel)
17:26:32 Join marines [0] (~quassel@
17:32:58 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
17:33:42 Join mt [0] (~mtee@rockbox/developer/mt)
17:36:59 Quit marines (Remote host closed the connection)
17:44:09 Join AzureWurk [0] (
17:45:21 Join Moikkis [0] (
17:49:01 Quit MethoS- (Read error: Connection reset by peer)
17:51:01 Join MethoS- [0] (~clemens@
17:51:13 Join kugel_ [0] (
17:51:35 Quit kugel (Disconnected by services)
17:51:39 Nick kugel_ is now known as kugel (
17:51:43 Quit kugel (Changing host)
17:51:43 Join kugel [0] (~kugel@rockbox/developer/kugel)
18:06:21 Join YDS10-YDS11SL [0] (
18:10:46 Quit komputes (Quit: I haven't slept for ten days, because that would be too long.)
18:15:07 Join [Saint] [0] (S_a_i_n_t@
18:17:30 Quit S_a_i_n_t (Ping timeout: 248 seconds)
18:22:15 Join Buschel [0] (
18:29:14 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
18:32:38 Quit kramer3d (Client Quit)
18:33:22 Part YDS10-YDS11SL
18:36:51 Join anewuser [0] (anewuser@unaffiliated/anewuser)
18:38:43 Quit Xerion (Read error: Connection reset by peer)
18:39:04CIA-6New commit by torne (r27294): Provide the option to automatically update existing bookmark files on stop, without creating ones that don't already exist. ...
18:41:32CIA-6r27294 build result: All green
18:45:37 Quit Jaykay (Quit: ChatZilla 0.9.86 [Firefox 3.6.6/20100625231939])
18:45:51 Quit swilde (Quit: ERC Version 5.3 (IRC client for Emacs))
18:45:57 Join Jaykay [0] (
18:46:00***Saving seen data "./dancer.seen"
18:48:43 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
18:52:03 Join Xerion [0] (
19:01:49 Quit pamaury (Remote host closed the connection)
19:08:16 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
19:11:07anewuserlately, I've been having hiccups on sound
19:11:21anewuserlike it stops for half a second every 10-20 seconds
19:11:35anewuserthink it started after doing a recording
19:12:59 Join Jerom [0] (~jerome@
19:14:02 Join komputes [0] (~komputes@ubuntu/member/komputes)
19:15:30 Quit leavittx (Ping timeout: 240 seconds)
19:18:49 Quit DerPapst (Quit: Leaving.)
19:19:30 Quit leavittx_ (Ping timeout: 240 seconds)
19:22:26 Join leavittx_ [0] (~leavittx@
19:25:43 Join stripwax [0] (
19:26:18bertrikthe problem with the first button not working after exiting the dict plugin seems to be a case of a missing release event, it is queued by the button code, but cleared by button_clear_queue in plugin.c
19:26:56Tornethe code that handles changing contexts? or solmething else
19:27:23Tornethe action code tries to ignore releases after context changes, but i don think it does it very well
19:27:58bertrikI don't think I fully understand the problem yet
19:28:00 Join leavittx [0] (~leavittx@
19:28:16Tornewell, that might be worth looking at :)
19:28:34Torneit doesnt handle button combinations rightp i know
19:28:40bertrikTorne, how and where does the action code try to ignore releases after context changes?
19:29:09Tornei forget where exactly, can't look right now
19:29:18Torneit's in action.c though
19:31:58bertrikok, I think I see it :(
19:32:44Torneit may not be your problem
19:32:54kugelbertrik: I think the problem also happens with some splashes in the core
19:32:55Tornebut i am fairly sure it's not bug-free
19:32:59kugelif it's the same anyway
19:34:13 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
19:37:19bertrikOk, this could be it: button release event is eaten by button_clear_queue() when exiting the plugin. Then in action.c all keys are ignored until a release is seen.
19:37:45bertrikSo action.c only starts working after pressing and releasing one more time
19:38:29Tornethat sounds plausible.
19:38:58Torneation.c also stops ignoring after *one* release even if the last thing pressed was a combination
19:39:13Tornewhich gives you spurious, instead of missed, events
19:41:08 Quit wincent (Ping timeout: 265 seconds)
19:43:19 Quit flydutch (Quit: /* empty */)
19:43:37 Join stoffel [0] (
19:44:12 Join Lear [0] (chatzilla@rockbox/developer/lear)
19:51:45bertrikyup, this seems to be it, the button is eaten by the if-statement in action.c line 207
19:52:54 Join wincent [0] (
19:53:17bertrikTorne, basically exactly the code you suggested
19:53:50Tornewell while you're there can you make it work right for released combinations too?
19:54:14bertrikI was hoping to find a simple-to-fix typo or something :|
19:55:08bertrikMaybe we should not eat buttons when changing context, but maybe rather defer the context change a bit until we see a fully "idle" key situation
19:59:35 Quit BlakeJohnson86 (Quit: Leaving.)
19:59:47 Quit BSOD90 (Ping timeout: 240 seconds)
19:59:50 Join BSOD90 [0] (
20:00:35 Join BlakeJohnson86 [0] (
20:01:14Buschelbaaahhh, this faad code is soo ugly...
20:03:11 Quit simonrvn (Read error: No route to host)
20:03:27Tornebertrik: the context is just whatever someone passes o get_action though
20:03:42Torneyou get a new one, you can't refuse it
20:07:38 Join simonrvn [0] (
20:09:20bertrikI don't see any obvious solution to this problem
20:11:51bertrikThe hack to eat buttons until release was probably introduced to fix another problem
20:12:47 Quit r0b- (Ping timeout: 240 seconds)
20:14:58 Join jennifur [0] (
20:19:07 Join r0b- [0] (
20:24:11 Quit r0b- (Ping timeout: 276 seconds)
20:24:21 Join r0b- [0] (
20:24:39 Join CGL [0] (~CGL@
20:28:49 Quit stripwax (Quit:
20:32:17 Part Moikkis
20:39:03Buschelstripwax: did you try to replace faad's sbr_dct stuff with your new one?
20:39:26Buschelstripwax: or is it not possible?
20:44:34 Join evilnick_i4 [0] (
20:46:02***Saving seen data "./dancer.seen"
20:49:46 Join funman [0] (~fun@rockbox/developer/funman)
20:53:09 Quit evilnick_i4 (Changing host)
20:53:09 Join evilnick_i4 [0] (~evilnick@rockbox/staff/evilnick)
20:55:17 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
21:02:42CIA-6New commit by bieber (r27295): Theme Editor: Fixed bug that caused WPS skins without SBS parents not to render their backdrops correctly. Added support for RTL-dependant text ...
21:04:38 Part AzureWurk
21:04:47CIA-6r27295 build result: All green
21:10:31 Join evilnick_ [0] (~evilnick@2001:0:4137:9e76:388f:be6:b5a6:f655)
21:11:05 Join halmi_ [0] (
21:11:07 Quit evilnick (Quit: Leaving)
21:12:54CIA-6New commit by bieber (r27296): Theme Editor: Enabled conditional display of ID3/file info with checkbox on the device panel
21:13:52 Quit halmi (Ping timeout: 264 seconds)
21:14:16 Join froggyman [0] (~me@unaffiliated/froggyman)
21:14:55CIA-6r27296 build result: All green
21:16:27saratogaBuschel. funman, kugel: I did a test with the clipv2 forced to 40MHz core and iram clock:
21:16:31saratogaresults are very interesting
21:17:32 Join funman_ [0] (~fun@2001:0:53aa:64c:14db:17a4:b23e:7c46)
21:17:59 Quit funman (Ping timeout: 260 seconds)
21:18:00 Nick funman_ is now known as Guest11508 (~fun@2001:0:53aa:64c:14db:17a4:b23e:7c46)
21:18:19 Nick Guest11508 is now known as funman (~fun@2001:0:53aa:64c:14db:17a4:b23e:7c46)
21:18:26 Quit funman (Changing host)
21:18:26 Join funman [0] (~fun@rockbox/developer/funman)
21:19:00bertrikvery efficient?
21:19:47saratoga7MHz for flac decode
21:20:02saratogabut also very different then the results for boosted
21:20:34funmanwhat do you mean with boosting disabled?
21:21:03funmanyou put the CPU to 40MHz ?
21:21:14saratogaliterally commented out the boost commands
21:21:15funman(sorry i can't read, it's written in the wiki)
21:21:49bieberAny opinions on how I should handle the built-in fonts with the theme editor?
21:22:09funmanwe should benchmark 40MHz pclk versus 24MHz pclk
21:22:21funman(for battery)
21:22:41bieberi.e., should I copy them over into the editor's resources directory, or just have references to their location in place in the source tree?
21:24:21saratogafunman: updated it
21:24:26 Quit halmi_ (Quit: halmi_)
21:24:38saratogafunman: I think 24MHz probably uses less power, since according to those tests almost nothing should boost
21:24:56saratogabut boosting with PCLK higher is probably more efficient if you're going to do it
21:25:10saratogabut not if it means your unboosted clock has to be higher too
21:25:55saratogaso best is probably 24MHz pclk=fclk unboosted, and 64MHz PCLK, 200+ FCLK
21:26:10saratogaassuming you can chage pclk on boost and want to max fclk
21:26:33funmandynamic PCLK implies dynamic clock switching for some peripherals
21:26:47saratogaboosting to fclk = ~150 MHz might be more efficient though, since it seems like the faster clock really doesn't help most codecs as much since the pclk stays too low
21:27:23domonokybieber: do you mean the built-in systemfont, or the fonts which are delivered with the normal zip ?
21:27:31 Quit GeekShadow (Ping timeout: 240 seconds)
21:27:41bieberdomonoky: Both
21:27:42 Join Frampis [0] (
21:27:50saratogaAPE and Wavpack seem to be the only formats with really good FCLK scaling
21:27:50bieberIs the built-in system font not loaded from a FNT file?
21:28:01Frampiswhere does the rockbox installation utility save the backup when upgrading on windows?
21:28:10funmansaratoga: first thing is to be sure we can boost/unboost correctly. I think my last patch on fs#11297 has too much changes because it also changes PLLA to 248MHz
21:28:44Frampisit removed my theme and icons so I need to find the backup
21:29:07domonokyFrampis: it should create a .backup folder on your mp3player
21:29:11CIA-6New commit by Buschel (r27297): Submit FS #11450. Change compile option for libfaad to -O2 for ARM target and -O1 for other CPUs. This speeds up decoding by 1-5%.
21:29:24Frampisdomonoky, thanks
21:30:21pixelmabieber: IIRC the built-in font is only the ascii part of the 08-Schumacher-Clean one, so just the needed parts to show errors etc. to not bloat bin size (and the name built-in is there for a reason ;) )
21:30:48pixelmanot sure what happens at compile time though
21:31:04CIA-6r27297 build result: All green
21:31:09Buschelbtw, I have some optimizations for aac-he. on PP5022 decoding is sped up from 130 MHz to ~116 MHz.
21:31:39bieberGot it
21:32:05 Join mirak_ [0] (
21:32:49Frampisare 3.5 themes incompatible with the latest version?
21:33:09domonokyFrampis: yes, the syntacs was heavily changed..
21:33:18FrampisI have selected the theme I used previously for the while playing screen but it looks completely different
21:33:35domonokybut the themesite wa upgraded, and there is a converter somewhere.
21:34:03 Quit merbanan (Quit: Leaving)
21:34:12domonokyso if you got the theme from the themesite, just redownload it. If it was self created.. try to find the upgrader tool :-)
21:34:14Frampisif anyone knows where to find said converter, that would make my day
21:34:38Frampisthe theme I used has disappeared from the site
21:34:49Frampisit was jbox for the ipod video
21:35:03saratogafunman: FWIW I think boosting to 200MHz and keeping voltage low (i think 200mhz is the same for amsv1 v2 voltage?) makes most sense now
21:35:27domonokythe converter is in svn under: utils\skinupdater but you have to compile it, dont know if there is a readymade binary avaiable somewhere.
21:37:12 Quit chid (Ping timeout: 260 seconds)
21:38:07Frampisare themes that have been deleted from the site archived somewhere?
21:38:55funmansaratoga: i'll test fs#11458 on clipv1
21:38:57Frampisoh, the Rockbox Utility finds a theme called jbox but it's not listed on the site, weird
21:40:33Frampisit doesn't work, though
21:40:44Frampisso apparently outdated themes are stil lbeing hosted
21:41:23mc2739Frampis: if you are a windows user, there is a link to a windows version of the convertor here:
21:41:34Frampismc2739, thank you very much
21:42:04 Nick evilnick_ is now known as evilnick (~evilnick@2001:0:4137:9e76:388f:be6:b5a6:f655)
21:45:49Frampisthe patched skin doesn't work any better, though
21:45:49Frampisthe only file for jbox is jBox.wps, there aren't files with the extensions sbs or fms
21:47:06 Quit evilnick (Ping timeout: 248 seconds)
21:47:11 Join evilnick_i4_ [0] (~evilnick@
21:47:17 Quit evilnick_i4 (Ping timeout: 265 seconds)
21:47:18 Nick evilnick_i4_ is now known as evilnick_i4 (~evilnick@
21:49:12funmanfields in WAV files are all big endian, except the filesize which is little endian? (at offset 4)
21:49:58funmanah no, some other fields are LE
21:51:01 Join evilnick [0] (
21:52:47 Part watto
21:54:32Frampisthe resulting file is different from the original but it works no better, damn
21:57:22 Quit evilnick (Ping timeout: 245 seconds)
21:58:01 Join evilnick [0] (~evilnick@2002:18bc:9495::18bc:9495)
21:58:09domonokyFrampis: themes on the themesite can be hidden because they fail all checkwps checks, or because they manually hidden, or they can get deleted manuall.
21:59:13FrampisRockbox Utility shows this particular theme but it is not displayed on the site
21:59:29domonokyFrampis: i took a shot look from the admin interface, and jBox is hidden because it fails checkwps, ie it is broken.
21:59:52bieberAnyone care to discuss the FNT format?
22:00:51Frampisdomonoky, can a skin be "broken" but still function?
22:01:12FrampisjBox functioned back when 3.5 was current
22:01:36Frampisor is it only "broken" only for the latest version?
22:01:46domonokyFrampis: but rockbox changed, so jbox broke.. :-)
22:02:18Frampisso there are things wrong with the skin that even the converter can't fix?
22:02:35 Quit evilnick (Ping timeout: 248 seconds)
22:03:23domonokywhen did you see jbox in rbutil ? before or after you installed a current build ?
22:03:52Frampisif I open it up right now and select extras -> install themes, it shows up
22:03:53Frampisin the list
22:04:12domonokyrbutil uses the old-site (which host non-converted themes) if it thinks it has a 3.6 or older build installed.
22:06:44FrampisI think I'll download a new skin and make new graphics similar to jBox with paint or something :P
22:07:14domonokyperhaps it would be easier to just try to fix jBox ?
22:07:17mc2739Frampis: here is a fixed jbox.wps file:
22:07:43 Quit simonrvn (Read error: No route to host)
22:08:10Frampismc2739, thanks, I'll try it out
22:09:13Frampisit doesn't work, though
22:09:32Frampisstill falls back to the default wps
22:09:51mc2739what version of Rockbox are you using>
22:10:37FrampisI just upgraded and that's why I'm trying to fix the skin
22:10:50mc2739ahh, I fixed it for current builds.
22:11:25FrampisI also tried a skin called FSM that is listed on the skin site but it also does not work, is it not a requirement for wps's to be compatible with 3.6 to be listed?
22:11:54 Join simonrvn [0] (
22:12:07mc2739you have to download themes for 3.6 here:
22:12:28mc2739oops, wrong link:
22:13:46Frampisi.e. this site is outdated?
22:14:11funmanwho thought it was a good idea to add the name of a function as a comment on the ending brace?
22:14:27funmanit's confusing as it removes focus from other comments
22:14:56LearI see that from time to time...
22:15:00FrampisjBox downloaded from your link does not function
22:15:30FrampisI'm starting to wonder if my installation of Rockbox is somehow broken here and not the skins
22:15:35domonokyFrampis: no, you link is not- outdated, its newer then 3.6 :-)
22:15:51 Quit ranma (Remote host closed the connection)
22:16:16domonokymc2739: would be the correct link for a 3.6 build..
22:16:24 Quit Jaykay (Quit: ChatZilla 0.9.86 [Firefox 3.6.6/20100625231939])
22:17:08Frampisbut isn't 3.6 the current stable release? says 3.6
22:17:41domonoky3.6 is the "current stable release" but not the "current build" which i though you had installed :-)
22:18:45Frampisso does 3.6 still use the old skin syntax? the only reason I'm looking for a working jBox is the fact that it broke when I upgraded from 3.5 to 3.6
22:19:21mc2739Frampis: try this one: (updated wps file for 3.6)
22:19:51FrampisI triend it when you linked it before and it doesn't work either
22:19:57mc2739Frampis: there were some changes made between 3.5 and 3.6
22:20:17mc2739it is updated - try again
22:21:45Frampisnow it loads, thanks
22:22:01 Quit simonrvn (Ping timeout: 276 seconds)
22:22:04Frampisbut some of the graphics are misplaced, its seems I have to change some of the coordinates
22:22:39 Quit anewuser (Quit: Sixth edition of 25 samples + 48 hours + detroit party (july 31) (maybe streamed online!))
22:23:52saratogafunman: (sorry was out) looking at the results more, the only actual format that needs FCLK > 100MHz is APE c4000, all others don't even need that
22:24:12saratogai wonder if boosting so high makes any sense for just one APE profile
22:25:04saratogait'd probably be more efficient to boost lower where the CPU uses less power and operates more efficiently due to pclk
22:25:40saratogathough i guess this is really just one more argument for a 3 step boost system
22:26:01 Quit Jerom (Quit: Leaving.)
22:26:51funmanboosting still doesn't work though
22:26:54 Join Jerom [0] (~heidi@
22:27:37saratogadetails ....
22:28:01 Join evilnick [0] (~evilnick@2002:18bc:9495::18bc:9495)
22:28:33 Quit balintx (Read error: Connection reset by peer)
22:28:46 Join balintx [0] (
22:28:54saratogaignoring that pesky fact, how feasible would it be to have a function in the codec/plugin API that requests boost_more
22:29:09saratogawhich would basically increase the boost clock
22:29:24saratogasay from 100MHz to 250MHz (or similar amounts depending on PCLK)
22:30:29pixelmabieber: what is there to discuss about fnt?
22:30:45mc2739bertrik: I think the hack to eat buttons is for quickscreen, if I remember correctly, I #if 0'd that section and the quickscreen was the only thing that had a problem. See FS #10097 for additional information that I found.
22:31:07 Join simonrvn [0] (
22:31:08 Join ranma [0] (
22:31:17bieberI was just wondering if anyone could give me a little more info about the format than we have in the wiki, before I go digging into the font loading code for it
22:32:18 Quit bmbl (Quit: Bye!)
22:33:24 Quit kugel (Ping timeout: 240 seconds)
22:35:18saratogafunman: any idea why test_codec output gets corrupted sometimes on the clip if you let the back light go out and then come back on?
22:35:46saratogai assume its a driver issue since no other player but the clips does it
22:36:34funmanyou mean some lines are missing?
22:36:43 Quit evilnick (Ping timeout: 248 seconds)
22:37:33saratogafunman: no its even weirder, the text doesn't update, so you get what was on the screen before the backlight went out, even though its been updated since then
22:38:02 Join evilnick_ [0] (~evilnick@2001:0:4137:9e74:44b:86c:b5a6:f655)
22:38:06CIA-6New commit by bieber (r27298): Theme Editor: Added a small PHP script to auto-generate a qrc entry for a directory with aliases for each file
22:38:50funmandunno, looks correct so far (been decoding ape_c5000 for a dozen of minutes)
22:39:50saratogabasically i think what happens is that LCD updates are ignored if the screen is off, so if the test finishes while its off, the results don't print and you just see what was on the screen before it went off
22:39:58CIA-6r27298 build result: All green
22:40:13saratogaso if I test a single long file and dont keep hitting buttons to keep the screen on, the output is never printed
22:41:33saratoga2.4MHz per EQ band when the clip is unboosted
22:42:00 Nick evilnick_ is now known as evilnick (~evilnick@2001:0:4137:9e74:44b:86c:b5a6:f655)
22:42:08funmanscreen should be updated at the next log_text()
22:42:10CIA-6New commit by bieber (r27299): Theme Editor: Fixed redundant directory separator bug in addresourcedir script
22:42:21 Join Drise [0] (
22:42:55saratogaon my player at least log_text seems to be ignored if the screen is off
22:43:15DriseJust curious, does anyone know how far along the Sansa Fuze V2 USB support is?
22:43:15funmanthere's something weird: lame_320.mp3 : 5533.33% realtime
22:43:33funmansaratoga: yep, lcd_update() is a no-op if the backlight is off
22:43:59CIA-6r27299 build result: All green
22:44:14funmanshould be the same for fuze
22:44:19 Quit Drise (Client Quit)
22:44:46bieberMeh, apparently all the Rockbox fonts amount to 15MB of data
22:44:48saratogafunman: so I should just make the plugin turon the screen before writing to it?
22:44:57saratogaturn on
22:45:14bieberSo if I include them in the Theme Editor, I get a huge executable, but if I don't all the themes that rely on fonts bundled with RB won't work :/
22:45:41funmansaratoga: usually log_text() is called regularly, so you would see weird content only a bit between turning the backlight on and the next log_text() ?
22:45:56saratogafunman: hmm looks like I broke something :)
22:46:04***Saving seen data "./dancer.seen"
22:46:07bluebroth3rbieber: how about downloading the fonts pack when needed? Or asking the user to do so?
22:46:16saratogain the test_codec program it draws a bunch of stuff to the screen without turning on the backlight
22:46:27saratogawhich works fine on the e200v1 but not on the clip FWIW
22:46:32funmanaache and mp3 results are broken
22:46:37bieberbluebroth3r: Is there a dedicated download location just for fnt files?
22:47:01bluebroth3rbieber: well, there's the font pack (on the extras section for the downloads)
22:47:06saratogathe small buffer may expose additional bugs
22:47:16saratogabut weird that lame broke but not a lossless file with far more rebuffering
22:47:30funmanhm weird aache plays much too fast
22:47:39bluebroth3rthough there are two problems: you need to download it and you need to extract it
22:47:59saratogafunman: blame buschel :)
22:48:15*Buschel hides away
22:48:46bluebroth3rbieber: IMO it's acceptable to have the user download the fonts pack and place it in some specific location for a first implementation.
22:48:47funmanyeah i think mp3 stops at the first rebuffer
22:49:07funmanBuschel: it looks like it plays 'realtime' = as fast as i can
22:49:35saratogarealtime or 2x too fast?
22:49:35bluebroth3rembedding 15MB fonts doesn't sound like a good idea, as it would also require updating the theme editor every time a font gets updates. Doesn't happen too often, but well ...
22:49:55funmanmore than 2x too fast
22:50:08funmanhow can i check samplerate in rockbox?
22:50:11saratoga2x would have made sense, no idea then
22:50:26saratogathe standard test file is a 44.1khz HE file
22:50:59bluebroth3rbieber: I guess it shouldn't be a problem to have the fonts separately available for download in a specific location. Needs to get done on the download server of course.
22:51:02saratogabut HE encodes everything at half sample rate and then upsamples during decoding, so to the AAC-LC stuff in rockbox it looks like a 22.05khz file (but rockbox playback only get its after its converted to 44.1)
22:51:26bieberbluebroth3r: What is RBUtil using for ZIP files? I just found out that Qt has QCompress()/QUncompress(), which use zlib
22:51:56funmansaratoga: perhaps it would make sense to just force the backlight on in test_codec?
22:52:09saratogaalready added it to my build :)
22:52:17saratogajust wondered if it was something more generally wrong
22:52:36bluebroth3rbieber: see rbutilqt/zip
22:53:40saratogaadvance_buffer_loc looks wrong to me
22:53:43saratogaand only mp3 uses it
22:53:55saratogaand only targets with < 8MB of RAM :)
22:54:01 Join evilnick_ [0] (~evilnick@2001:0:4137:9e76:2c6e:1d42:b5a6:f655)
22:54:51 Quit evilnick (Ping timeout: 248 seconds)
22:55:08bieberbluebroth3r: Should we try to move the zip code into a common directory, or just copy/paste it?
22:56:11amiconnbertrik: Regading leaving plugins - why is button_clear_queue() called?
22:57:01amiconnThis function should only be used if absolutely necessary - it's nasty to get your events eaten unpredictively
22:57:51 Quit Lear (Quit: ChatZilla 0.9.86 [Firefox 3.6.6/20100625231939])
22:58:09CIA-6New commit by bieber (r27300): Theme Editor: Added the built-in font to the project resources
22:58:15amiconnBuschel: Do you think aac could be made dual core on PP? (thinking of aac-he...)
22:58:35 Quit evilnick_ (Ping timeout: 248 seconds)
22:59:17saratogaamiconn: yes it can be
22:59:23saratogabut first it needs to be better optimized
22:59:32saratogaat the moment the HE parts take more then 80MHz on PP
23:00:00CIA-6r27300 build result: All green
23:00:03saratogathis could quite easily be fixed, the ffmpeg people have been nice enough to rewrite the 3 slowest functions using the best known algorithms
23:00:09saratogaand they're not very complicated
23:00:17BuschelPP5022: 130 MHz (with current svn and 64kaache.m4a)
23:01:46 Quit komputes (Remote host closed the connection)
23:01:46saratogai think with effort real time decoding on one core shoul dbe possible
23:02:04 Quit n17ikh (Ping timeout: 260 seconds)
23:02:17Buschelsaratoga: yes
23:02:41*amiconn wonders how fast aac-he is on coldfire
23:02:59 Quit Xerion (Quit: )
23:03:02Buschelmaybe we should also think about splitting aac-he to be able to allocate more data to iram
23:03:17saratogabasically right now the problem is that AAC uses really dumb subband filterbanks
23:03:27Buschelamiconn: want to try a aac-he patch? on coldifre? :o)
23:03:31saratogawhich are then badly optimized by gcc
23:03:33amiconnIRAM is not always a good idea. Beware of the non-single-cycle parts on PP5020
23:03:37 Quit r0b- (Read error: Connection reset by peer)
23:03:41 Quit [Saint] (Ping timeout: 276 seconds)
23:03:56Buschelamiconn: this can be handled for different CPUs of course
23:04:11saratogaFWIW i bet we could share the same IRAM buffers between the HE and LC parts, they're never run at the same time (on single core anyway)
23:04:30amiconnIronically this means that with proper optimisation, codecs are often faster on PP5002 in spite of its broken cache
23:04:55amiconn(faster than PP5020, obviously. Never faster than PP5022)
23:05:49amiconnsaratoga: They're separate code paths?
23:06:03amiconnI thought -he is some kind of extension of -lc
23:07:07funmanAAC doesn't use range coding?
23:07:21saratogaamiconn: AAC+ is an AAC-LC file with an added part
23:07:32saratogaso basically you completely decode a normal LC file, then feed it into a second decoder
23:07:47amiconnHmm, but then you do need the -lc code path
23:08:07saratogayeah, but they're run sequentially, so buffers could be shared between them without loss of data
23:08:31amiconnHmm, and you don't need the old buffer to decode the next part?
23:08:38saratogaand fortunately it looks like they can share at least some of the same look up tables (e.g. the ffmpeg use the same trig tables for both parts)
23:09:04saratogayou'll need to save at least 512 samples IIRC to decode the next block, but thats not much memory
23:09:11amiconnDifferent q - isn't sbr similar to qmf, or am I mixng up things?
23:09:21saratogaits implemented with a QMF typically
23:09:25 Quit ender` (Quit: You know you created god in your image when he hates the same people you do.)
23:09:42amiconnWell, qmf should be very well emac'ifiable on coldfire
23:10:02saratogaessentially what the SBR encoder does is 2 subband decomposition and then create an AAC-LC of the lower frequency subband
23:10:27saratogathe decoder then reverses the process, decoding the LC file, and then doing an iQMF to get the time domain samples
23:10:44saratogaexcept since the upper subband isn't transmitted some tricks are needed to guess what it looked like
23:10:58 Join evilnick [0] (
23:11:05bluebroth3rbieber: sorry, had a phone call
23:11:22saratogai need to dig into the ffmpeg code more, they actually use an MDCT in place of the QMF in one part, which would be a very nice trick since we have a very fast MDCT in rockbox already
23:11:37saratogathe rest could probably share the ATRAC3 QMF which is also very fast
23:11:54bluebroth3rfor now I would simply use the source files via relative paths. I was considering to merge rbutil/ and utils/ anyway, and we could do that in that run
23:11:56 Join DerPapst [0] (
23:12:15saratogasince AAC-HE is so similar to ATRAC (except ATRAC uses 3 or 4 subbands depending on the flavor while AAC-HE only uses 2)
23:12:48saratogai don't think AAC-HE optimization is a hard problem, its just a boring one since who uses AAC HE :)
23:13:16*amiconn doesn't use aac at all
23:13:45bieberbluebroth3r: Okay
23:13:51saratogayeah its not very useful if you can use mp3, wma or vorbis
23:13:56bieberI think I'm going to be spending a little while figuring out the font format anyways, so we've got some time ;)
23:14:18saratogajust layers and layers of crap designed to hit as many mpeg-la patents as possible
23:14:31bluebroth3rgood. I was planning some Rockbox work for tonight but RL kept me away :(
23:15:04 Quit evilnick_i4 (Quit: Colloquy for iPhone -
23:15:24 Join evilnick_i4 [0] (
23:15:33saratogaFWIW I'm pretty impressed with WMA Pro, in some ways its a remarkably simple format, yet it seems to do as well as AAC
23:15:48bluebroth3rReal Life.
23:15:57Buschelsaratoga/amiconn: in the meanwhile (until anyone of us or somebody else will start to rework libfaad) ->
23:15:58bertrikamiconn, button_clear_queue is called unconditionally after just after the plugin returns, it looks like it's been this way since svn r4907
23:16:27bieberOh, lol
23:16:31bluebroth3rhmm, simply putting all below rbutil/ would create a utils/tools folder. Sounds a bit weird :o
23:16:47saratogabuschel: you sure I can't convince you to port the ffmpeg replacement for sbr_qmf_analysis_32 ?
23:16:51saratogai've got it right here :)
23:16:52bluebroth3rthough I guess that shouldn't be much of a problem :)
23:17:15saratogahuh got MP3 working in test codec but AAC-HE still breaks
23:17:28Buschelsaratoga: my wife will kill me ;o)
23:17:52funmanencoder plugin progresses: now loading an encoder from the plugin fails ;)
23:17:58Buschelsaratoga: hwo does this aac-he breaking look like?
23:18:37saratogaBuschel: i think its my fault trying to fix test codec
23:18:54Buschelgood :) (for me)
23:18:58saratogaoh no
23:19:01saratogano its your fault :)
23:19:14saratogajust tried playing it outside of the test_codec and its all screwe dup too
23:19:18saratogahaha thought I broke it
23:19:33Buschelwhich revision do you use?
23:19:38saratogacurrent SVN
23:19:42saratogatime scale is off
23:19:50saratogait advanced 10 second in the first 2 seconds of playback
23:19:58saratogaand the audio is way pitch shifted
23:20:07saratogadid you touch aac.c lately?
23:21:14 Quit togetic (Ping timeout: 276 seconds)
23:21:28BuschelI have touched a lot of aac code in the last days... nevertheless the 64kaache file works fine on my iPod and my sim
23:21:42saratogai tested on a clipv1
23:22:07saratogadid you use the standard aache test track?
23:22:30saratogai wonder whats different about the clip
23:22:38saratogaoh wait
23:22:47Buschelthere might be other #defines?
23:22:54saratogathis is probably my fault
23:24:13saratoga"Disable AAC-SBR, AAC-SSR, and AAC-PS on Sansa clip. Clean up warnings introduced by this. Enable compiling the AAC codec on the Clip. Due to lack of memory for seek tables, files over 10 minutes long or so will probably fail."
23:24:17Buschelwhat is MEMORYSIZE for a clip?
23:24:28saratogayou'd think I'd remember committing that :)
23:25:03funmanBuschel: 2
23:25:05saratogaso basically its just decoding the LC part of the file, but not setting the sample rate correctly
23:25:28funmansaratoga: tables don't fit in iram?
23:25:48saratogafunman: it fits but then theres hardly any left over for the seek table, so files have to be very short or they fail
23:26:29saratogaso i disabled the HE stuff until someone got around to fixing the MP4 parser to not need tones of memory for seek tables
23:26:52saratogathere needs to be a check for MEMORYSIZE < 2 that corrects for the reduced sample rate of the LC file
23:27:19 Quit simonrvn (Quit: yay normal quit, trying some h/w trick...)
23:27:21saratogaor I guess SBR_DEC would be a better check
23:27:46BuschelSBR_DEC is not set if MEM <= 2
23:28:15saratogaBuschel: did adding that check for SBR_DEC in aac.c the other day cause this?
23:29:53Buschelsaratoga: talking of r27231? no, not I know of.
23:29:54 Join Russel-Athletic [0] (
23:30:34Russel-Athletici made a custom theme but with the latest svn release it does not work, what is the best way to find out what is going wrong?
23:30:42Russel-Athletic(theme will not load)
23:30:50saratogaBuschel: if forceUpsampling is set it should probably still allocate that buffer even if SBR is disabled though?
23:31:31bieberRussel-Athletic: Is your theme using the old format for your WPS files? i.e. do your tags look like %tagname|option1|option2|option3?
23:31:40 Quit MethoS- (Read error: Connection reset by peer)
23:31:51Buschelsaratoga: this sequence was copied from libfaad/specrec.c
23:32:24Buschelsaratoga: can you just try to go back to pre-r27225 for aac.c and libfaad?
23:32:34saratogatrying now
23:32:35Russel-Athleticbieber: probably, at least i have a line like %?m<stopped|playing|paused|ffwd|rew>
23:32:46 Quit funman (Quit: leaving)
23:33:22bieberRussel-Athletic: That's the correct form for conditionals, but tags by themselves should now look like %tag(option, option)
23:35:10Russel-Athletici am sorry, i am not sure, here is the theme
23:35:57Russel-Athletici removed the # at the start
23:36:03bieberYeah, that's in the old format
23:36:29bieberThere's a converter in the SVN tree somewhere to convert it to the new format, but I don't know exactly where it is. I'm sure someone around here does, though
23:37:06bieberFound it: it's in /utils/skinupdater
23:37:15Russel-Athleticthanks i will try it
23:38:17Buschelsaratoga: I can reproduce the failure in the pcsim if I use "#if 0" instead of "#if MEMORYSIEZ > 2" in libfaad/common.h
23:38:42saratogaBuschel: did you configure a sim with < 2MB of RAM?
23:39:21Buschelsaratoga: just a local change in common.h
23:39:45saratogabut you needed it because your sim had > 2MB of RAM right?
23:40:04saratogaok sounds right
23:40:29saratogacompiling an older build now, just taking a while since i didn't have a checkout ready
23:40:46 Join togetic [0] (~togetic@unaffiliated/ibuffy)
23:42:30 Join MethoS- [0] (~clemens@
23:42:38saratogaBuschel: no still broken in r27200
23:43:07saratogamaybe its been broken all year and no one noticed
23:43:39 Join simonrvn [0] (
23:44:10Buschelsaratoga: you were 1 min faster. tested with r27224 -> same error
23:44:31saratogacan we use forceUpSampling to make it output the correct number of samples?
23:45:23BuschelforceUpSampling is hrdwired to the SBR_DEC define
23:45:47 Join r0b- [0] (
23:47:47bieberIn a FNT file, shouldn't noffset and nwidth be the same? Does each character not need an offset and a width specified?
23:47:56 Quit Russel-Athletic (Quit: leaving)
23:48:10saratogasample rate is off by about 4x according to my watch which doesn't really make sense
23:49:46 Join fml [0] (
23:49:59 Join evilnick_ [0] (~evilnick@rockbox/staff/evilnick)
23:50:20 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
23:50:52fmlTorne: Hello! Will you also provide the description of the new feature for the manual? What bookmarks are updated? Just in the current dir? Or all (where)?
23:50:53 Quit evilnick_ (Client Quit)
23:53:37 Quit evilnick_i4 (Ping timeout: 245 seconds)
23:53:43TorneOh, I forgot about the manual..
23:53:58Torneit saves a bookmark if the file it would save it to already exists
23:54:16amiconnbieber: The offset table is optional. If the font is fixed width *and* the defined glyphs have no gap, the offsets can be calculated
23:55:17amiconnThe font format is derived from microwindows (called nano-x nowadays)
23:56:29bieberSo let me see if I'm more or less understanding the format. After the header, there's just a dump of the image data for all the fonts, it's padded to a 16 bit boundary, then there's an array of offsets into the image data for each character, followed by width data if necessary?
23:57:40saratogaBuschel: just correcting the sample rate in rockbox doesn't help
23:58:28amiconnBoth the width table and the offset table are optional
23:58:47Buschelsaratoga: the failure definately comes in when disabling SBR_DEC
23:58:54amiconnThe width table is only present for proportional fonts, for fixed width it is unnecessary and hence missing

Previous day | Next day