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 2007-07-10

00:03:18 Join My_Sic [0] (
00:05:06*amiconn wonders what's so different between the menu and the browser
00:05:24amiconnThe menu is significantly slower than the browser on ipod video unboosted
00:05:26*ender thoughtit'seasiertoalignwithtabs
00:05:34amiconnBut it's the same list widget....
00:05:43 Join midkay [0] (n=midkay@rockbox/developer/midkay)
00:06:13 Quit ender (" If you're going to do something tonight that you'll be sorry for tomorrow morning, sleep late.")
00:09:46 Join Llorean [0] (
00:10:13 Quit jhMikeS (Nick collision from services.)
00:10:19 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
00:13:22amiconndan_a: around?
00:14:46 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
00:18:09 Quit jgarvey ("Leaving")
00:19:05 Join webguest60 [0] (i=4157837d@gateway/web/cgi-irc/
00:19:09 Quit bluebrother ("leaving")
00:20:39 Quit webguest60 (Client Quit)
00:25:49lazkasansa still not working and it's definitly the "divide-by-zero" commit that's causing it.
00:26:45Lloreanlazka: The "divide by zero" commit tells it to stop ignoring errors. It's not causing it, it's calling attention to it.
00:28:16lazkaLlorean: ok :)
00:30:18obolazka: what address is it reporting the error at?
00:30:34 Part maffe
00:30:52 Join maffe [0] (
00:32:09lazkaobo: screen gets white.. then black... wheel light is on -> freeze
00:32:16lazkano output
00:32:30amiconndiv0 error in the lcd driver perhaps
00:32:31lazkasimulator works fine
00:32:37pixelmaLlorean: unless there is a bug in the related code... ;)
00:32:53Lloreanamiconn: I'm getting a little bit of green garbage at the top of my screen instead of booting.
00:33:31LloreanDivide by Zero at 400051F8
00:33:55amiconnNow check the .map what's there
00:34:12LloreanDoing that
00:35:23 Quit Wiwie ("Miranda IM! Smaller, Faster, Easier.")
00:37:03Lloreanamiconn: Just says playback.o
00:37:11 Quit seclein2 (Connection timed out)
00:37:15linuxstbDoes it say .ibss to the left?
00:37:29linuxstbThen that's data...
00:37:43linuxstb(my ipod build is the same)
00:38:47linuxstbExactly... ;)
00:38:55amiconnThat would mean execution is run away
00:39:08*amiconn would expect this address to be icode
00:39:16LloreanOr the div by 0 reporting is reporting the wrong address.
00:40:29linuxstbIt seemed to report the correct address for the first problem though (in fat_mount).
00:40:40LloreanThat div by 0 is in the bootloader
00:40:48LloreanI'd just updated it too.
00:40:57linuxstbOK, that makes sense.
00:41:13linuxstbThe whole of the bootloader runs from IRAM.
00:41:56linuxstbWhat does the bootloader map say?
00:44:00LloreanCould fat_mount be giving the same problem, before lcd_init, in Rockbox as well and that's what's giving people freezes?
00:44:13 Quit amiconn (
00:44:13 Quit saratoga (
00:44:13 Quit HellDragon (
00:44:13 Quit entheh (
00:44:13 Quit toffe82 (
00:44:13 Quit lazka (
00:44:13 Quit bospaadje (
00:44:13 Quit mbr (
00:44:13 Quit BHSPitLappy (
00:44:13 Quit merbanan (
00:44:13 Quit ptw419 (
00:44:13 Quit XavierGr (
00:44:13 Quit Caliban (
00:44:13 Quit gromit` (
00:44:13 Quit Mouser_X (
00:44:13 Quit DataGhost (
00:44:14 Quit crwl (
00:44:14 Quit sbeh (
00:44:14 Quit sneakums (
00:44:14 Quit krp (
00:44:14 Quit Hadaka (
00:44:14 Quit hannesd (
00:44:27linuxstbLlorean: OK, that's reassuring. You could try adding "if (!fat_bpb->bpb_bytspersec) return -2;" at line 315. before the division using that variable.
00:44:31NJoinhannesd [0] (
00:44:36linuxstb(line 315 of firmware/drivers/fat.c)
00:45:00LloreanLet me just recover my Sansa first, gotta track down some firmware. It can't load the OF right now, ;)
00:45:25linuxstbDo you have a SD card?
00:45:47LloreanI do.
00:46:06linuxstbHave you tried booting without it?
00:46:14LloreanDidn't even realize it was in
00:46:27 Quit Xerion (Read error: 104 (Connection reset by peer))
00:47:10LloreanStill dividing by zero
00:48:01linuxstbIs the FAT partition is the first partition on the Sansa?
00:48:12LloreanI believe so
00:48:52 Quit jhMikeS (
00:48:52 Quit Arathis (
00:48:52 Quit lee-qid (
00:48:52 Quit Farp (
00:48:52 Quit andrew__2 (
00:48:52 Quit h4ngedm4n (
00:48:52 Quit mtoups (
00:48:52 Quit crashmatrix (
00:48:52 Quit pabs (
00:49:15linuxstbThen I can't understand why it's getting invalid values and trying to divide by zero. IIUC, the partition probing would try that partition first.
00:49:34 Join RaRe` [0] (
00:50:43 Join Xerion_ [0] (
00:51:04 Nick Xerion_ is now known as Xerion (
00:51:30 Quit Jon-Kha (
00:51:30 Quit Slasheri (
00:51:30 Quit w0rd54 (
00:51:30 Quit luckz (
00:51:36 Quit mirak (Remote closed the connection)
00:51:37NJoinamiconn [0] (
00:51:48Lloreanlinuxstb: With that added to fat.c, the bootloader works fine again
00:52:44linuxstbGood. Tried a normal build?
00:52:52LloreanAbout to
00:53:08NJoinsaratoga [0] (i=9803c6dd@gateway/web/cgi-irc/
00:53:08NJointoffe82 [0] (
00:53:08NJoinlazka [0] (
00:53:08NJoinbospaadje [0] (
00:53:08NJoinmbr [0] (
00:53:08NJoinBHSPitLappy [0] (
00:53:08NJoinmerbanan [0] (n=banan@
00:53:08NJoinptw419 [0] (
00:53:08NJoinXavierGr [0] (
00:53:08NJoinCaliban [0] (
00:53:08NJoingromit` [0] (
00:53:08NJoinMouser_X [0] (
00:53:08NJoinDataGhost [0] (n=dataghos@
00:53:08NJoinsneakums [0] (
00:53:08NJoinsbeh [0] (
00:53:08NJoincrwl [0] (
00:53:42NJoinHellDragon [0] (i=JD@unaffiliated/helldragon)
00:53:51DerPapstGood night everyone :)
00:53:55 Quit DerPapst ("So Long And Thanks For All The Fish!")
00:54:04NJoinkrp [0] (
00:54:04NJoinHadaka [0] (
00:54:33Lloreanlinuxstb: Seems to fix the regular build too.
00:54:47NJoinJon-Kha [0] (
00:54:47NJoinw0rd54 [0] (
00:54:47NJoinluckz [0] (
00:54:47NJoinSlasheri [0] (i=miipekk@rockbox/developer/Slasheri)
00:54:56linuxstbLlorean: OK, I'll commit that.
00:55:02 Quit hannesd (
00:55:02 Quit qwx (
00:55:02 Quit preglow (
00:55:02 Quit markun (
00:55:02 Quit amiconn (
00:55:02 Quit Xerion (
00:55:02 Quit maffe (
00:55:02 Quit Robin0800 (
00:55:02 Quit Stribbs (
00:55:02 Quit gtkspert (
00:55:02 Quit lostnihilist (
00:55:02 Quit z35 (
00:55:02 Quit tedrock (
00:55:02 Quit hax (
00:55:02 Quit crashd_ (
00:55:02 Quit [Ray] (
00:55:02 Quit J (
00:55:02 Quit Overand (
00:55:02 Quit idnar (
00:55:02 Quit enyc (
00:55:02 Quit [mbm] (
00:55:02 Quit lostlogic (
00:55:11 Quit funky (Remote closed the connection)
00:55:41 Join funky [0] (n=repulse@unaffiliated/funky)
00:56:13NJoinjhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
00:56:13NJoinArathis [0] (
00:56:13NJoinlee-qid [0] (
00:56:13NJoinpabs [0] (
00:56:13NJoincrashmatrix [0] (
00:56:13NJoinandrew__2 [0] (
00:56:13NJoinFarp [0] (i=Farp@
00:56:13NJoinmtoups [0] (n=mtoups@HENSON.ISR.CS.CMU.EDU)
00:56:13NJoinh4ngedm4n [0] (
00:56:55 Join thegeek_ [0] (
00:56:55 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
00:56:55NJoinhannesd [0] (
00:56:55NJoinmaffe [0] (
00:56:55NJoinRobin0800 [0] (
00:56:55NJoinStribbs [0] (i=dcf0795f@gateway/web/cgi-irc/
00:56:55NJoinmarkun [0] (n=markun@rockbox/developer/markun)
00:56:55NJoingtkspert [0] (
00:56:55NJoinlostnihilist [0] (
00:56:55NJoinqwx [0] (
00:56:55NJoinz35 [0] (
00:56:55NJointedrock [0] (
00:56:55NJoinhax [0] (n=hax@httpcraft/hax)
00:56:55NJoincrashd_ [0] (
00:56:55NJoinpreglow [0] (n=thomj@rockbox/developer/preglow)
00:56:55NJoinidnar [0] (i=mithrand@unaffiliated/idnar)
00:56:55NJoinJ [0] (
00:56:55NJoinlostlogic [0] (n=lostlogi@rockbox/developer/lostlogic)
00:56:55NJoin[Ray] [0] (
00:56:55NJoinOverand [0] (i=overand@pdpc/supporter/active/Overand)
00:56:55NJoinenyc [0] (
00:56:55NJoin[mbm] [0] (i=mbm@openwrt/developer/mbm)
00:59:51 Part dan_a
01:01:41 Join funky_ [0] (
01:01:53***Saving seen data "./dancer.seen"
01:01:54 Quit bdgraue (Remote closed the connection)
01:03:18 Quit funky (Success)
01:05:20 Quit stripwax (Read error: 110 (Connection timed out))
01:07:03 Quit spiorf_ (Remote closed the connection)
01:07:50 Quit BHSPitLappy (Read error: 104 (Connection reset by peer))
01:08:39 Quit RaRe (Read error: 110 (Connection timed out))
01:09:56 Join relaxed [0] (n=relaxed@unaffiliated/relaxed)
01:11:33 Quit thegeek (Read error: 110 (Connection timed out))
01:12:51 Quit petur (Read error: 104 (Connection reset by peer))
01:13:04 Quit FOAD_ (Read error: 104 (Connection reset by peer))
01:13:15 Join FOAD [0] (
01:13:23 Join petur [0] (n=petur@rockbox/developer/petur)
01:15:15 Quit Robin0800 (" HydraIRC -> <- Po-ta-to, boil em, mash em, stick em in a stew.")
01:24:47 Quit obo ("bye")
01:28:30 Join Megaf [0] (
01:30:19 Join mike__ [0] (
01:32:07 Quit lee-qid ("aufwiederbyebientotsayonara")
01:34:22 Part Llorean
01:34:44MegafRockBox is Linux?
01:35:42mike__but its similar in many ways
01:35:42mike__they are both open source
01:36:23mike__they are both operating systems, although rockbox is only for mp3 players, where linux is mostly aimed towards full computers (although some projects bring it to other devices)
01:36:50 Part toffe82
01:37:03Megafmy iPod Nano runs iPod Linux
01:37:09 Part hcs
01:37:21mike__rockbox is similar
01:37:22Megafi will intall rock box now
01:37:27 Quit petur ("Zzzzz")
01:37:35Megaftri-boot :P
01:37:56mike__google "ipod rockbox installer"
01:38:02mike__it will take you to a forum with a nice installer to install rockbox on the ipod
01:38:18Megafrockbox installer for Linux?
01:38:23scorchemike__: dont recommend that here
01:38:23mike__i think this will not tri-boot with ipl
01:39:04scorcherockbox will tri boot
01:40:28 Quit chrisjs169 (Read error: 110 (Connection timed out))
01:40:48mike__just replace it
01:40:49mike__so nevermind.. thats not what you want
01:42:36Megafi lost my iPod!!!
01:42:41Megafis not here
01:42:46Megafwhere is my ipod?
01:42:53scorchewherever you left it last?
01:43:05Megafi dont know
01:43:12Megafmy grandfather house
01:43:17scorchewell, it sint like we can help you find it...
01:43:33Megafmy ipod is there =/
01:44:09 Part Megaf ("Ex-Chat")
01:46:05mike__i wish the IRC could help find lost goods
01:46:25scorchethat would bloat the protocol
01:52:53 Join Llorean [0] (
01:53:05 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
02:03:48 Quit mike__ ("ChatZilla [Firefox]")
02:04:50 Quit perplexity (Read error: 104 (Connection reset by peer))
02:05:18 Join perplexity [0] (
02:05:27 Join iamben [0] (
02:05:44iambenhow do i reset my sansa if i get a divide by zero error?
02:05:59 Join TiMiD[FD] [0] (n=TiMiD[
02:06:07LloreanIt can be shut down by simply holding the power button for 15 seconds
02:06:35iambenah i just wasnt patient enough, thanks =)
02:06:36 Quit pixie (Remote closed the connection)
02:06:40 Join pixie [0] (
02:15:16 Quit saratoga ("CGI:IRC")
02:19:08 Quit Guile` (Read error: 104 (Connection reset by peer))
02:19:28 Join Guile` [0] (n=Guile@
02:21:32 Quit Thundercloud (Remote closed the connection)
02:24:02 Quit Arathis ("Bye, bye")
02:24:11*jhMikeS didn't expect divX0 to be showing up so soon, hehe. important bugs to fix though.
02:24:34 Join saratoga [0] (i=9803c6dd@gateway/web/cgi-irc/
02:25:20 Quit perplexity (Excess Flood)
02:25:48 Join perplexity [0] (
02:28:12linuxstbjhMikeS: Did it work OK on your Sansa?
02:28:52jhMikeSI no longer have a lockup when going to the file browser from the WPS after extracting the card and trying to read from ti
02:30:31jhMikeSperhaps the problem with freezing with the SD enabled will go away now. I got word that some units seem to indeed have bit 7 stuck low until a card is inserted the first time.
02:34:16linuxstbI was talking about the divide by zero commit...
02:35:33jhMikeSin the FAT driver? I think it's relevant to what I was just saying and may have fixed that problem.
02:37:04linuxstbNo, I meant did Rockbox work on your Sansa after you enabled the divide by zero check earlier today?
02:37:39jhMikeSI never ran it on sansa. I checked it out on gigabeat only to make sure the trap works then has to run out.
02:39:03linuxstbAre SD cards typically formatted as "superfloppies", or do they have a partition table?
02:40:02jhMikeSI don't know. I reformatted mine as FAT32. They were originally FAT16.
02:41:12 Quit sneakums (Read error: 110 (Connection timed out))
02:41:17jhMikeSwhat the linux command to see if there's a table?
02:41:36linuxstbIt's just that the probing for partitions caused my Elio to crash (when I was working on that port). It's hard disk doesn't have a partition table.
02:42:14linuxstbfdisk -l /dev/sdX will give garbage (I think) if there's no partition table.
02:43:18 Join mike__ [0] (
02:43:32mike__is there a way to change the controls for my ipod 5.5 gen 30gb?
02:43:54mike__i want to make the center button select, rather than the fast foreward button
02:44:20pixelmamike__: are you using a custom build?
02:44:43pixelmain an official build center = select
02:45:08pixelma(right works too in most cases)
02:45:29mike__like in the main menu, center should choose my option?
02:45:36mike__ill have to look into another build then..
02:45:40mike__thanks for your help!
02:45:58linuxstbOut of curiosity, which build are you using?
02:46:06pixelmaI've read that some custom build include a patch that changes that behaviour
02:46:14mike__let me check (its the most popular ipod one in the forum)
02:46:16 Quit Soul-Slayer (Read error: 104 (Connection reset by peer))
02:46:21 Quit maffe (Remote closed the connection)
02:46:31scorcheuse the official build...
02:46:32mike__evilG (
02:46:42 Quit perplexity (Read error: 104 (Connection reset by peer))
02:46:51mike__im using these for improved theme support
02:46:53linuxstbSo what does center do in the main menu?
02:46:53mike__but I will try now
02:47:08mike__nothing... if I hold it it brings up a settings for the option im on
02:48:00linuxstbYou should probably report it in the thread for that build - it doesn't sound like intended behaviour to me (or at least, not sensible behaviour...)
02:48:33mike__its not that bad
02:48:36mike__just a little annoying
02:48:38mike__ill do that
02:49:15 Join adrift [0] (
02:49:45 Quit hcs ("Leaving.")
02:50:02 Join perplexity [0] (
02:51:55mike__after reading in the forum, i see discussion of this (a patch disables some of its functionality)
02:52:33 Quit mike__ ("ChatZilla [Firefox]")
02:55:02 Join secleinteer [0] (
02:56:30 Quit relaxed (Remote closed the connection)
02:56:59jhMikeSlinuxstb: i see windows showing partitions and I don't think rockbox has any partition stuff disabled when reading the card.
03:01:56***Saving seen data "./dancer.seen"
03:04:56 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
03:05:17 Quit jhMikeS (Nick collision from services.)
03:05:23 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
03:06:23 Quit RaRe` (Read error: 110 (Connection timed out))
03:06:57 Join RaRe` [0] (
03:07:38linuxstbjhMikeS: OK, it sounds like a different issue then.
03:10:25 Join __tumu [0] (
03:14:06saratogaoh god theres a static VLC decoder in ffmpeg
03:14:15saratogawhy couldn't i notice this sooner
03:15:28linuxstbWhat's the advantage of that?
03:15:50 Part pixelma
03:16:09saratogai spent 3 days writing one
03:16:17saratogaso that we wouldn't need to use malloc
03:16:30saratogawell not really 3 days, most of it was spent cursing at ffmpeg rather then working
03:17:40 Quit tumu (Read error: 104 (Connection reset by peer))
03:17:48linuxstbAre you going to use it?
03:19:10saratogai'm looking at it now
03:19:15 Join seclein_ [0] (
03:19:25saratogathey've tagged it as "broken/missdesigned"
03:19:31saratogawhich isn't encouraging
03:19:39saratogabut mine is pretty broken right now anyway
03:20:51 Quit secleinteer (Read error: 104 (Connection reset by peer))
03:20:59saratogalooks like they just wrap realloc around an array, so its not really any better then my solution
03:21:10saratogabut would have been infinately easier to implement had i noticed it
03:21:31 Quit seclein_ (Remote closed the connection)
03:22:02saratogaon the upside, it seems they're annoyed at how horribly stupid their VLC code is to label it depreciated, so maybe we'll get a better one eventually
03:23:08linuxstbBTW, I've been working a little on the asf parsing - the problem with packets being split into multiple payloads. It seems to be worse than I first thought - a single packet can contain payloads belonging to different streams (in the case of video files). I still don't know if an entire superframe will be in a packet (mixed in with video payloads).
03:24:21saratogais this a concern for pure audio files or just mixed audio/video files?
03:24:59linuxstbSo far I've only seen it in mixed audio/video and in pure audio files encoded by ffmpeg.
03:26:13linuxstbThe mixed audio/video file was encoded with wmp11
03:26:21linuxstbOr rather, wmsdk v11
03:29:43saratogai'd be ok with not dealing with that case if its particularly difficult
03:30:18saratogaalso, I'd really like to know where Jones got this stupid VLC function from, the static one was added to ffmpeg almost 3 years ago . . .
03:31:09linuxstbWouldn't he have just used whatever the WMA decoder used?
03:31:53linuxstbi.e. without thinking to look for alternatives?
03:35:25 Join secleinteer [0] (
03:35:28 Join Nation-of-Baby-G [0] (
03:35:38 Nick Nation-of-Baby-G is now known as Fusion-Pompadour (
03:36:17saratogalinuxstb: this is the one the wma decoder used, at least as of 2004
03:36:28saratogaevidently he didn't take all the code directly from ffmpeg
03:36:45 Join Nico_P [0] (
03:36:48saratogathe wmadecoder itself seems to be from last summer, but not all of the other parts
03:36:58Fusion-PompadourStupid question:
03:38:02Fusion-PompadourI d/l'd the daily build this morning; and for about 20 mn the iPod displayed a screen saying I needed to re synch w/iTunes
03:38:11linuxstbsaratoga: What are your plans? Are you trying to sync with current ffmpeg?
03:38:27Fusion-Pompadourafter about 20 mn in the carger it works fine. What's up w/that
03:38:34saratogalinuxstb: i'm merging the changes into one of my PC based test programs to see what difference they really make
03:39:02saratogai don't really like their static version, so I see no reason to use it, but if they've improved anything else it'd be worth having
03:40:48LloreanFusion-Pompadour: Rockbox has no such screen, so there's not much that we can answer regarding that.
03:40:57LloreanNico_P: Hello.
03:41:09linuxstbsaratoga: What's the status of the < 32kbps files? I assume they still don't work?
03:41:11Fusion-PompadourI may be off topic
03:41:35 Part Fusion-Pompadour ("Konversation terminated!")
03:43:41Nico_PLlorean: got my email ?
03:46:34JdGordonhey Nico_P
03:46:39Nico_Phi :)
03:46:42iambendang, this rockbox aint bad
03:47:00JdGordonI had a very quick look at playback.c yesterday, got any ideas where to start merging our code in?
03:47:35iambenthe devs & theme makers must have really good vision, these fonts are tiny
03:47:47iambennice for LOTS of info on the screen though
03:48:13Nico_PJdGordon: places like audio_load_track, audio_fill_file_buffer, audio_rebuffer
03:48:24Nico_Phopefully we can get rid of most of this
03:48:51saratogalinuxstb: I'm actually working on them now
03:48:52Nico_Pthe place to start looking is probably audio_thread()
03:49:00saratogai found the VLC thing while debugging them
03:49:08JdGordonright :)
03:49:28*JdGordon goned
03:53:58 Join RoC_MasterMind [0] (
03:54:55 Join aliask [0] (
03:56:49 Join chrisjs169 [0] (
03:57:16scorcheiamben: you can change the font...
03:59:05 Join devon [0] (
04:06:33 Quit devon ("using sirc version 2.211+KSIRC/1.3.12")
04:40:15 Join sneakums [0] (
04:42:37 Quit XavierGr ("One firmware to rule them all!")
04:58:19 Quit perplexity (Read error: 113 (No route to host))
04:58:44 Quit mbr (Read error: 110 (Connection timed out))
04:58:47iambensome of the dirs on my sansa like music & playlists seem to be hidden in the file browser, is this normal?
05:01:59***Saving seen data "./dancer.seen"
05:02:44saratogarockbox follows the hidden attribute, and also hides folders that start with a dot
05:02:57saratogaeither change the folder status or enable the show hidden files option in rockbox
05:04:55iambenhidden, is this a vfat attribute? and how do i change it from linux?
05:05:49 Quit Nico_P (Remote closed the connection)
05:12:12 Quit RoC_MasterMind ("Leaving")
05:12:18iambenthanks for the info... dunno how to actually toggle "hidden", but recreating the dir worked
05:12:33 Join RoC_MasterMind [0] (
05:14:09 Join cmos_ [0] (
05:14:29 Nick cmos_ is now known as bagawk (
05:15:35 Quit ptw419 ()
05:16:23 Join BigMac [0] (
05:18:47 Join festol [0] (
05:19:26linuxstbiamben: I think the only way to do it in Linux is to use mtools. But as you found out, just recreating the dir will create it without the hidden bit.
05:21:46 Join perplexity [0] (
05:24:17 Quit billytwowilly (Remote closed the connection)
05:24:33 Join billytwowilly [0] (
05:32:52saratogalinuxstb: do you have a test driver somewhere for your parser ?
05:33:02saratogasomething that compiles independently of rockbox?
05:33:26saratogamine is now really out of date with SVN
05:33:47 Quit DogBoy ("Leaving")
05:34:39linuxstbYes, I do have one, but it's currently broken - I'm working on getting it working with packets with multiple payloads. I should hopefully clean it up and get it working again tomorrow.
05:36:01saratogadoes it have the latest wmadeci included?
05:37:48saratogalinuxstb: I found an optimized VLC routine in ffmpeg that we're not using for some reason, but I'm having a hell of a time getting it to work
05:38:47 Quit festol (Connection timed out)
05:42:43 Quit BigMac ("Ex-Chat")
05:47:24saratogalinuxstb: if you have a test driver thats capable of decoding wma files on the disk and that uses something close to the current svn decoder, could you post it?
05:47:34saratogait will save me the trouble of updating mine
05:47:44 Quit saratoga ("CGI:IRC")
05:56:19 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
06:10:33 Quit linuxstb (Read error: 110 (Connection timed out))
06:12:36 Join kkurbjun [0] (
06:17:45 Quit Nibbier (Read error: 110 (Connection timed out))
06:18:39 Join Nibbier [0] (
06:24:47 Quit linuxstb_ (Read error: 104 (Connection reset by peer))
06:24:52 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
06:25:47 Nick qwx is now known as qwm (
06:26:01 Quit lazka (Remote closed the connection)
06:26:24 Part qwm
06:34:14 Quit Llorean ("Leaving.")
06:37:22 Join Llorean [0] (
06:48:19 Join ptw419 [0] (
06:54:22 Quit Stribbs ("CGI:IRC (Ping timeout)")
06:57:25 Join BHSPitLappy [0] (
06:59:35alienbiker99does rockbox use the podcast tag?
07:00:33RoC_MasterMindWhat does that mean
07:01:13 Join rudefyet [0] (
07:01:18RoC_MasterMindBeryl can see it's there.
07:01:23RoC_MasterMindIf it looks I guess
07:01:30 Quit RoC_MasterMind ("Leaving")
07:02:03***Saving seen data "./dancer.seen"
07:02:04rudefyetberyl on rockbox?!
07:03:03rudefyetjust loaded rockbox on my new 80gig iPod, it's been awhile, has mpegplayer gotten anywhere?
07:11:20 Join scorche` [0] (n=scorche@rockbox/administrator/scorche)
07:11:34 Quit scorche (Read error: 104 (Connection reset by peer))
07:12:16alienbiker99if only i could have coverflow for mediamonkey
07:16:51 Join davina [0] (
07:18:06 Quit kkurbjun (Read error: 110 (Connection timed out))
07:19:43 Quit davina (Remote closed the connection)
07:21:47 Join davina [0] (
07:27:37 Quit lymeca ("...and don't forget: 09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0 !")
07:32:14 Quit kubiix ("Miranda IM! Smaller, Faster, Easier.")
07:38:42 Nick scorche` is now known as scorche (n=scorche@rockbox/administrator/scorche)
07:42:17 Quit billytwowilly (Remote closed the connection)
07:42:36 Join billytwowilly [0] (
07:48:28amiconnjhMikeS: It's really odd. I just checked what coldfire does. Back when we started using the divu/divs instructions, we introduced a div0 exception. However, I've never seen div0 problems popping up un cf...
07:49:32amiconnNote though that not all targets have a div0 exception. SH doesn't, and gcc doesn't offer a way to hook one like on arm
07:51:54jhMikeSamiconn: I did add the string for it to coldfire awhile ago. It does pop up if you force one.
07:53:02jhMikeSbtw, audio clicks are back on e200 since the lower CPU frequency makes a cache flush take too long (during lcd updates).
07:54:40jhMikeSthough it does appear the battery is lasting a good bit longer
07:56:06amiconnI couldn't have found that myself. Wonder why saratoga didn't notice it
07:56:36amiconnI wonder why it's necessary to flush cache on sansa during normal operation though
07:56:50jhMikeSIt wouldn't be a problem with 1) COP running pcm 2) a nice a of forcing a refill just before flushing the display
07:57:37jhMikeSbecause the data written to the driver frambuffer won't be current
07:58:43jhMikeSit also rather sluggish in general at 30MHz
08:04:56 Join GodEater_ [0] (
08:05:47amiconnThat nasty lcd hookup on sansa :(
08:06:21amiconnI wonder whether the PP cache controller allows marking some areas as uncacheable
08:06:47jhMikeSI think that would bog it down horribly
08:07:14amiconnI would just mark the lcd framebuffer as uncacheable
08:07:53jhMikeSright, but I'm worried that would bog down writes to the driver framebuffer.
08:08:13amiconnShouldn't be slower than writing to cache+then flush
08:08:24jhMikeSI'm rather fond of the memory mapped graphics ram myself on that and gigabeat
08:09:25jhMikeSmmmm...I don't know. I think it burst transfers that stuff.
08:16:04amiconnPossibly. But flushing the whole cachae all the time doesn't sound efficient
08:18:26amiconnI need to investigate 0x70000034 further. My current guess is that it sets the ram timing (but why 2x?)
08:19:05jhMikeSI wonder about flushing specific ranges and invalidating ones as well
08:19:29amiconnHowever, the ipod OF sets these timing values to (roughly) clock_in_MHz/10, while the sansa OF sets it (roughly) to clock_in_MHz/5
08:19:50amiconn..and the H10 OF to clock_in_MHz/4
08:20:18jhMikeSis there any use of the MRC instructions or anything?
08:20:57 Quit rudefyet ("Leaving")
08:21:00amiconnThe PP doesn't seem to have CP15
08:21:40amiconnI found no mrc or mcr in the disassemblies
08:32:47jhMikeSf000f040, 60006044 seem to have association to f000f044 (where the flush is initiated)
08:33:27 Join RaRe [0] (
08:36:06 Join spiorf [0] (
08:38:40jhMikeSI've been listening for hours now and I still show 72% (3.95V) :)
08:45:48 Join kaaloo [0] (
08:46:01 Part kaaloo
08:47:17jhMikeSI guess f000f040 is written in ipod_init_cache but the emu dump shows it used consistently.
08:47:53 Join ender` [0] (
08:52:38 Join Rob222241 [0] (
08:52:49 Quit RaRe` (Read error: 110 (Connection timed out))
08:56:18 Quit Dcider (Read error: 110 (Connection timed out))
08:56:19 Join Wiwie [0] (
08:56:56*jhMikeS can't seem to set the boost counter on the screen :\
09:02:06***Saving seen data "./dancer.seen"
09:02:42 Join BHSPitMonkey [0] (
09:03:16 Join petur [0] (n=petur@rockbox/developer/petur)
09:04:02 Quit Wiwie ("Miranda IM! Smaller, Faster, Easier.")
09:04:50amiconnHmm. Given the reports about instability on some H10s, I refined my guess for 0x70000034
09:05:00amiconnIt might be the ram waitstates
09:05:25amiconnWill do some experiments to find out. Need to disable the cache for that...
09:07:18amiconnIf this is true, it would also explain why ipl doesn't manage stable operation above 75MHz
09:09:26jhMikeS0x70000034 is set to 0x00000707 in the emu by of but it's set to 0x00001010 in ata_init. ??
09:10:47 Quit Rob2222 (Read error: 110 (Connection timed out))
09:13:08amiconnipod OF sets it proportional to cpu clock, roughly cpucclock/10
09:13:25amiconn...i.e. 0x00000808 at 80MHz, 0x00000303 at 30MHz etc
09:13:51amiconnThe sansa loader sets it to 0x00001010 (running at 80MHz)
09:14:05amiconn...and H10 to 0x00001515 (also at 80MHz)
09:17:01jhMikeSthen we've got more wait stats than we need atm
09:17:54 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
09:17:54amiconnRockbox uses the ipod formula. If my theory is correct, we use too few waitstates on sansa and h10
09:19:01petursounds logical...
09:19:11jhMikeSyou said sansa sets to 0x00001010 running at 80MHz but we're only running 75, right?
09:19:37jhMikeSwe already set to 0x1010
09:21:46amiconnAt least not in the clock setup
09:22:45jhMikeSif that's what it is, it should definitely get out of ata_init on e200
09:24:53 Quit BHSPitLappy (Read error: 104 (Connection reset by peer))
09:26:34jhMikeSamiconn: would you be able to check the pcm-pp.c rework if I send a patch? if it plays and stops without incident then it checks out.
09:26:52 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
09:27:44 Join Splosh123 [0] (
09:28:08Splosh123I need some help? :)
09:29:01jhMikeSas for pp5002, I can't check it. I just need to determine if writing to the FIFO actually is enough to ack the request.
09:30:11Splosh123I had rockbox on my ipod Video 30GB. And then i deleted everything on my ipod (.rockbox folder, hidden ipod folder and all music) now the ipod is not being reconised by windows at all? :S
09:30:12amiconnI can't test atm, no time redoes some stuff on 5002 besided for 502x but no cable for the 3g :(
09:35:59jhMikeSfrankly I should just switch all the 502x over to 16-bit samples and get rid of the << 16 for them
09:38:51JdGordondoes the gigbeat have recording?
09:39:46jhMikeSit appears it could but in a limited way since the IIS SDIN is exposed to the world according to the block diagram
09:44:39JdGordoncool, when was the timer added to the current builds page?
09:45:54jhMikeStime on gigabeat?
09:46:19jhMikeSoh, never mind :P
09:47:15 Nick funky_ is now known as funky (
09:48:05jhMikeSit seems indeed that writing to the FIFO is an FIQ ack for 502x and nothing else is needed
09:48:49JdGordonthats a bit odd...
09:48:58JdGordon+88bytes on non rec targets!
09:50:04 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
09:53:09amiconnjhMikeS: Why do you think I could test on pp5002?
09:53:18peturJdGordon: isn't English lang build-in?
09:53:40JdGordonbah, yeah forgot about that
09:53:45jhMikeSamiconn: just wondering if you could since I wasn't sure. pp5020 tests would also be good though.
09:53:49*JdGordon goes back to not worrying about it
09:53:55 Join obo [0] (
09:54:02pixelmathat reminds me of the langV2 patch...
09:54:04 Join Crash91 [0] (i=c4dbb954@gateway/web/cgi-irc/
09:54:29Crash91hi guys
09:54:42JdGordonpixelma: do you know if the groundowrk can be done for that patch so we can slowly go through and fix the strings instead of doing it in 1 massive patch?
09:55:36Crash91im just here to say - > Great work! you guys were like dead for a week then all these improvements :D *rubs hands with glee*
09:56:21markunCrash91: there is always stuff going on even if there are no commits
09:56:28Jthis might be a silly request, but on could somebody please say what the colours indicate?
09:56:37pixelmaJdGordon: I think the groundwork is pretty much done - but the english.lang shouldn't be altered too often (especially because of voice). You could help syncing it again ;)
09:56:45scorcheJ: in which chart?
09:56:51Zagorjhulst: green is good, red is bad :-)
09:56:52Jeither, both!
09:57:24JZagor: good how? Bad why?
09:57:43Zagorjhulst: top char is about compiler errors and warnings. green = no errors/warnings.
09:57:53Jit would be useful to add the information to the page IMO
09:58:09petur2 more (good) commits and we're all green :)
09:58:14ZagorJ: bottom chart is about binary size. green means size went down (good), red means size went up (bad)
09:58:23pixelmaJdGordon: hmm... weird way of mine to put these things together in one sentence
09:59:06Jand the numbers which are neither colour?
09:59:35Zagorjhulst: they are so small they don't count :-)
09:59:46Crash91stayed neutral i think
09:59:54Crash91deleted and added same amount?
09:59:56Zagormeh, xchat expands nicks wrong
09:59:59jhMikeSgcc's treatment of r12 also seems somewhat randon hence the data aborts after recording. I cache the data in r10 and r11 FIQ banked and just use banked r12 as scratch
10:00:02 Quit billytwowilly (Remote closed the connection)
10:00:18 Join billytwowilly [0] (
10:00:31Jcould that information be added to the page? It's information which is good to know...
10:00:35amiconnjhMikeS: I hope to be able to test tonight (CET). But it might be that I won't have the mini g1 available anymore by then
10:01:13ZagorCrash91: no, all numbers are actual changes. but iirc up to 16 bytes can change simply due to difference in struct padding etc so no real fault of the patch.
10:01:27Zagordon't quote me on that last bit though :)
10:01:46jhMikeStime to give it back? :)
10:03:29Crash91> im downloading something so im quit laggy
10:03:32 Join safetydan [0] (n=dan@rockbox/developer/safetydan)
10:05:02Crash91finished now :)
10:06:01Crash91one last question
10:06:36Crash91does anyone know how to use media coder to convert all the files in a directory that has loads of subdirectories?
10:07:16Crash91in my case, /Music/<Artist>/<Album>/<mp3 file(s)>
10:07:50ZagorCrash91: do you use linux?
10:08:08JdGordonare we still short on mini 1g's in the dev group?
10:08:33ZagorCrash91: oh media coder only runs on windows. I don't know then...
10:08:48Crash91nope, today windows for a change :)
10:11:12Crash91i have been using linux for a week
10:13:48 Quit Crash91 ("CGI:IRC (EOF)")
10:14:51 Join XavierGr [0] (
10:18:45 Join Crash91 [0] (n=Crash91@
10:19:21Crash91is mpegplayer good enough for playing full length movies on the sansa?
10:21:22scorcheas long as you can watch it in one sitting without having to resume
10:23:13 Join pondlife [0] (
10:23:25Crash91hehe i have a trick for that
10:23:29Crash91slpit files
10:24:02Crash91how big will an mpeg file be compared to a 700mb divx avi?
10:24:17 Join Xerion [0] (
10:24:28markundepends on the bitrate you use
10:24:38aliaskAnd the resolution.
10:24:45markunno, only the bitrate
10:24:47aliaskWell, indirectly.
10:25:02aliaskYou can use a lower bitrate with a smaller resolution...
10:25:17Crash91well what should i use? for 208x176 video
10:25:36 Join Wiwie [0] (
10:25:46Crash91and whats mpeg4 ASP and mpeg 4 AVC
10:26:03 Join freqmod [0] (
10:26:41 Quit Wiwie (Client Quit)
10:27:02Crash91here are my settings in the iPod extension of Media coder (yes! it has sansa resolution)
10:27:51TiMiD[FD]Crash91: you should check google, there are tons of articles explaining that
10:28:01TiMiD[FD]wikipedia also
10:28:16Crash91video: Mpeg4 ASP, audio mp3, 208x176, expand to fit, 20fps, vid bitrate 400 and audio quality : "lower"
10:28:37Crash91im just checking the mpegplayer wiki..
10:30:17TiMiD[FD]do they really mention mpeg4 ASP/AVX ?
10:30:29TiMiD[FD]the site seems dead
10:30:53Crash91no, but they do have an extension for mediacoder which i saw recently
10:31:00Crash91it should do everything for me :)
10:31:28TiMiD[FD]ah right
10:31:54TiMiD[FD]never used that soft
10:32:11 Quit perplexity (Read error: 110 (Connection timed out))
10:33:45Crash91sansa media coverter made 693MB of mov files in 4 parts
10:33:53Crash91im expecting better from this :)
10:36:53 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
10:36:59 Join PaulJam [0] (
10:37:16Crash91i think now that scaling has been enabled on sansa, it will work much better
10:37:17 Join miepchen^schlaf [0] (
10:39:58PaulJamCrash91: why? how should cpu scaling improve the performance.
10:40:48Crash91rockbox>5hours (if youre lucky) OF> 15 hours
10:40:58PaulJamah, i thought you were talking about the mpegplayer performance
10:41:10Crash91yeah and i think that might help
10:41:14Crash91mpegplayer as well
10:41:57PaulJammpegplayer will propably always boost on a fullscreen video on sansa
10:42:33Crash91this is very irritating, mediacoder keeps encoding my Xvid to an mp3!
10:42:45TiMiD[FD]you could use mencoder
10:43:04TiMiD[FD]you can control very finely the parameters
10:44:20Crash91mencoder is bundeled with media coder, just tell me the params
10:45:46Crash91got it before you :p
10:46:17 Join AceNik [0] (n=AceNik@
10:46:21TiMiD[FD]there is also avidemux
10:46:37TiMiD[FD]actually I used that one to encode my videos if I recall
10:47:02AceNikguys i opened up my setero pin of my headphones to repair, the thin is there are 2golden wires, n 1 red n 1 green , how to solder them on 3 terminals
10:48:49 Quit AceNik (Client Quit)
10:52:05 Quit Splosh123 (Read error: 110 (Connection timed out))
10:54:01Crash91mencoder isnt working
10:54:13Crash91it keeps saying system cannot find spec. file
10:54:27 Join desowin [0] (
10:58:33aliaskI'm playing around with making a tree for the gigabeat S, but it looks like the stuff in the target tree isn't being included. I've edited the stuff in the configure script, but is there something else?
10:59:04markunaliask: I think ptw419 has done the same, so maybe you should work together
10:59:13 Join Entasis [0] (
10:59:43 Join Wiwie [0] (
10:59:48markunaliask: yes, you need to add files to firmware/SOURCES too
10:59:51aliaskHe hasn't finished, and he asked if I could take a look.
11:00:29aliaskAh cool. Thanks for the pointer.
11:01:41scorcheoh wow... Llorean: since when has the forums had "/me"?
11:01:59Lloreanscorche: Day one, I believe. They're sneaky like that.
11:02:08***Saving seen data "./dancer.seen"
11:02:25scorchehrm...well, that is my "learn something new" for today...
11:02:40LloreanThough it looks like it didn't work. worked in the preview
11:03:29LloreanAnd it's not disabled, I just checked.
11:03:41LloreanWrap it in [me] and [/me] for the same effect
11:03:56scorchehrm....odd how it would work in the preview, but not in the post..
11:04:33scorchethat didnt even work int he preview
11:05:11*scorche nods in agreement
11:05:31*Crash91 never knew that ogg was so much better and smaller than mp3!
11:05:40 Join perplexity [0] (
11:05:46scorchewell, that is a bit of a generalization...
11:05:55 Join webguest46 [0] (i=bd91d39d@gateway/web/cgi-irc/
11:05:59Crash91yeah i know
11:06:12Lloreanscorche: Odd, I don't know. I swear I've seen it used in the past, and can't think of what I've changed that might disable it.
11:06:12Crash91but for me, encoding my files to ogg is just fine
11:06:24 Join w1ll14m [0] (i=546851d0@gateway/web/cgi-irc/
11:07:01w1ll14mi have a question, what is the recommended gcc version to compile rockbox arm ?
11:07:04Crash91how do you op yourself if youre op in that channel and you have it registered?
11:07:22dionoea /msg chanserv op #channel nick
11:07:22 Quit webguest46 (Client Quit)
11:07:23scorche4.0.3 , rather
11:07:38Crash91but doesnt do it for you?
11:07:47scorcheit does
11:08:04w1ll14mcrash thanx :) still got 3.4.5 so i'm getting a new version ;)
11:08:15w1ll14mscorche also thanx!
11:08:28Crash91lol i missed a dot, sorry
11:08:29scorchew1ll14m: well, a new version wont need it to be 4.0.3 specifically
11:08:48Crash91what OS are you using>
11:08:54scorcheCrash91: you dont need the nick portion if you are opping yourself
11:09:03w1ll14mi know :) i forgot which version it was..... but i allready know that i would need a specific gcc
11:09:32Crash91w1ll14m: are you using linux?
11:09:46w1ll14mgentoo linux to be exact
11:10:32Crash91ok hold on for a sec
11:11:11Crash91ok, have you co-svnned?
11:11:34w1ll14m:) of course i've been compiling rockbox since december 2006 :)
11:11:49w1ll14mi just needed to know the exact gcc version ;)
11:12:16Crash91now goto
11:12:19w1ll14mbut thanx for the help anyway :)
11:12:40Crash91goto /tools/configure
11:12:46Crash91then ./
11:12:50Crash91and follow
11:12:58Crash91it will install the compiler for you
11:13:01w1ll14mi could use that one too :)
11:13:04Crash91the correct version
11:13:24w1ll14mat the moment i'm trying crossdev with gcc 4.0.3 arm-elf
11:13:25Crash91from mail archive
11:13:25Crash91When run, it asks you what target architecture you want a compiler setup for.
11:13:25Crash91Then, it downloads, unpacks, builds and installs binutils and the gcc
11:13:25DBUGEnqueued KICK Crash91
11:13:25Crash91cross-compiler for you.
11:13:39w1ll14mif that won't work ll use
11:13:49Crash91good luck
11:14:01w1ll14mthanx! :) i think i will suceed ;)
11:14:32scorcheit also will patch gcc with a few patches we have for it...there is actually one patch we have for arm gcc, but it will only affect you if you delve into thumb code, so you dont *really* need it
11:14:59w1ll14mahh ok :)
11:15:26scorchethe other patch is to sh gcc, so likely wont affect you
11:15:33Crash91omg! i need 30more mb of free space.. to get another movie on :p
11:15:51JdGordonpixelma: apart from being out of sync, whats holding the langv2 patch from commit?
11:17:53 Quit obo ("bye")
11:18:04safetydanJdGordon, it's one of those things that you only want to do once. And not all the strings have been properly categorised.
11:18:37 Quit Wiwie ("Miranda IM! Smaller, Faster, Easier.")
11:18:43freqmodCould somebody tell me what the second column of a dissasembly means (the one which i have underlined): "0:_4553_ 042523" I can't find a good enough manual
11:18:54freqmodafter 0:
11:19:19JdGordonsafetydan: do you know how many strings need cataegorising still?
11:20:39 Join thegeek [0] (
11:21:45JdGordonthere is a bug in features.txt btw
11:22:11JdGordon#if defined(CONFIG_TUNER) is wrong... it needs to be #if CONFIG_TUNER because its defined to 0 on targets without
11:23:05JdGordon... and just thinking, shouldnt features.txt be in apps/lang to make is easier to find?
11:27:07 Quit PaulJam (".")
11:27:19JdGordonsafetydan: pixelma: have the order of strings in english.lang been changed? and have they been worked on in order? so the last one with features is the last one which has been done?
11:27:31JdGordon.. I mean, everytihng above that one is done?
11:28:44CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
11:28:44*Crash91 loves OGG
11:28:59 Part Crash91
11:35:11 Quit Xerion (Read error: 104 (Connection reset by peer))
11:35:34safetydanJdGordon, the strings have been sorted yes.
11:35:52safetydanThe big thing stopping its commit is that it will break all voice files and then we'll need to have per-target voice files.
11:36:18pixelmaJdGordon: I think it was complete against the last svn it was synced to
11:36:56JdGordonthe english.lang.rej is longer than english.lang by 400 lines!
11:37:36aliaskI remember giving up trying to sync that patch at that point.
11:37:46pixelmawhen I synced it once, I gave up working with the .rej
11:38:06 Quit thegeek_ (Read error: 110 (Connection timed out))
11:38:57JdGordonhmm, well if the lang was finshed at that point its not really that hard to sync it up again, and then fix the strings added since then
11:40:27pixelma(the english.lang.rej that is - took the others though)
11:42:31 Join kaaloo [0] (
11:42:46 Part kaaloo
11:42:55w1ll14mdamn my build is from 07-05-2007, i need to update :)
11:43:06 Quit desowin (Read error: 113 (No route to host))
11:44:46 Quit Llorean (Read error: 104 (Connection reset by peer))
11:45:06 Join desowin [0] (
11:45:51 Join webguest37 [0] (i=c023110f@gateway/web/cgi-irc/
11:45:59pixelmaZagor: just wanted to point you to the last one who registered at twiki (AntiqX)
11:46:09pondlifeamiconn: Has anyone monitored the CPU frequency on PortalPlayer under the OF? Not sure how you could do this, but I wonder if it uses a sliding scale.
11:47:50Zagorpixelma: thanks. killed.
11:48:08*scorche cleans up the remains
11:48:26peturthere are more twiki users with illegal names, iirc there's even on that has write access (unless it was his real name)
11:50:16scorchesome of those are touchy...
11:52:02Zagoryeah we've been sloppy at times
11:53:37peturbah... can't find him anymore :/
11:53:54pixelmayeah... I remember someone called himself ThiasHias (and got write access). His email address was something with "matze" which is only a german nickname for Matthias... I believe he was deleted later (not sure)
11:56:36midkayMatthias != Thias.. so he should have been deleted if he wasn't :)
11:57:25markunor renamed
11:58:18midkaywould he know that he was renamed? he'd try to login and wouldn't be able to as Thias...
12:00:01 Join Xerion [0] (
12:00:06pixelmafound it odd especially with the "Hias" part ;)
12:01:11 Join maffe [0] (
12:02:05 Join lee-qid [0] (
12:03:44 Quit Hadaka (Read error: 60 (Operation timed out))
12:03:58 Quit TiMiD[FD] ("leaving")
12:18:20 Quit billytwowilly (Remote closed the connection)
12:18:35 Quit perplexity (Read error: 113 (No route to host))
12:18:36 Join billytwowilly [0] (
12:20:43 Quit XavierGr (Nick collision from services.)
12:20:46 Join XavierGr [0] (
12:20:52 Quit webguest37 ("CGI:IRC")
12:22:20 Part maffe
12:23:05JdGordon:'( I hate trying to figure out stupid annoying bugs
12:23:21JdGordonwhy the hell is the root menu passing 0x1 to do_menu() :'(
12:24:27w1ll14mhehe :) i've got my cross-platform compiler up and working again :)
12:25:07w1ll14mi see we have a wma codec?
12:29:54 Join Bawitdaba [0] (
12:30:56safetydanHuzzah! I finally can get assembler interweaved with the C source with objdump
12:35:01 Quit XavierGr (Nick collision from services.)
12:35:04 Join XavierGr [0] (
12:37:48 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
12:37:53linuxstbsafetydan: Please share...
12:38:14safetydanlinuxstb, compile with -g then you can use objdump −−source <object file>
12:39:00linuxstbpondlife: Regarding CPU frequency on PP, I remember that some of the older ipods (up to the 4G greyscale and minis IIRC) had a hack available where if you created a text file with a specific name in the iPod_Control folder, it would display the current CPU frequency in the status bar. But I can't remember the details.
12:39:46GodEater_how people find little easter eggs like that never ceases to amaze me
12:39:54linuxstbI tried it once on my ipod Color, and it didn't work. It's also possible that it was removed in newer firmware versions for the older ipods.
12:40:18linuxstbProbably by searching for strings in the firmware. Or someone from Apple leaked it.
12:40:39pondlifeIt was just a random thought I had that might save battery life on the PP targets... no biggie, but I thought I should share it.
12:40:51linuxstbThe ipodwizard people probably spent a long time staring at hex dumps of the ipod firmware.
12:41:43GodEater_very helpful if you can't sleep I find
12:42:04linuxstbAh, here are the details -
12:42:06pondlifeamiconn was saying he had rather got into the OF style...
12:43:06linuxstbGodEater: BTW, is there any arrangement for DevConPub?
12:43:14*linuxstb has forgotten what day was arranged...
12:44:17linuxstbAnd do you still need my 20GB hard disk?
12:44:37pondlifelinuxstb: It's tomorrow
12:44:42linuxstbThought so.
12:44:47pondlifeI definitely can't make it :(
12:44:57GodEater_meeting at the hotel
12:45:12GodEater_I've been a bit ill over the weekend so didn't manage to explore properly
12:45:18linuxstbIs there anyone else apart from us and Mr & Mrs Febs?
12:45:29GodEater_there were a couple of other people posted in the thread
12:45:35GodEater_which I updated this morning
12:45:45GodEater_not sure if they're actually going to come though
12:45:56GodEater_and yes I'd still like the 20G disk please :)
12:53:42 Join Llorean [0] (
13:00:18 Join RaRe` [0] (
13:02:09***Saving seen data "./dancer.seen"
13:04:41 Join yohannrjm [0] (
13:04:52 Join LinusN [0] (i=linus@gateway/web/cgi-irc/
13:05:01amiconnpondlife: Most OFs do a different thing than rockbox wrt clock frequencies
13:06:40pondlifeI know, I was just considering that this might be a factor towards the battery life. We have a step-boost, perhaps it's more of a sliding scale with other firmwares?
13:06:47amiconnOne has a huge table with codecs and their performance requirements (base clock) and additional requirements (dsp stuff), and calculates the necessary clock from that
13:07:12amiconnThe rockbox way is more flexible with less different frequencies
13:08:08pondlifeNo problem, just mentioning it in case it were useful.
13:08:32amiconnIt adapts the effective clock to the *actual* requirements (think e.g. vbr files) and hence is more effectife as long as the power drain varies proportional to cpu clock
13:08:34pondlifeI don't have a PP target, but would love to see Rockbox beating the OF for battery life on all targets
13:09:34amiconnThe problem on PP is that (1) our most common codec (libmad) is considerably less efficient than the coldfire version,
13:09:44 Quit yohannrjm ()
13:10:26amiconn(2) we don't know much about the chip hence there are most probably some modules running which are normally powered down by the of
13:10:41amiconn(3) We don't know how to power down the hdd on ipods
13:11:23amiconnhi LinusN :)
13:11:32 Quit adrift ("ChatZilla [Firefox]")
13:13:14amiconnLinusN: What do you think is the better approach for PP clock scaling: making the clock setup routine calculate registers at runtime (more flexible, can set all possible clock frequencies e.g. in 1MHz steps) or using precalculated values like on coldfire?
13:13:32amiconnI want to clean up this stuff after some more research
13:13:39LinusNi like precalculated
13:14:05*LinusN gets some coffee
13:14:16amiconnIs there a special reason?
13:16:08pondlifeAs long as the on-the-fly calculation isn't too slow, why not go with that? Smaller binsize, more flexible.... I suppose the boost mechanism could then go up a step at a time (depending on the boost count) and that might save CPU?
13:16:14jhMikeSsome day we might need more gradual frequency scaling ala low latency audio. it's one of those major sticky points about it.
13:16:16 Quit RaRe (Read error: 110 (Connection timed out))
13:16:40LinusNwell, it'
13:17:12LinusNs a philosophical thing really, why calculate it in runtime when you always end up with the same result
13:17:22LinusN(as it works now)
13:17:44amiconnEasier to adapt if you want different frequencies
13:17:56pondlifeDepends on how big you want the precalculated table to grow too..
13:18:13LinusNpondlife: i don't think 3 entries would kill us
13:18:14amiconnpondlife: Well, 3 entries atm
13:18:32dionoeawell some decoding could be done halfway between the low and high freqs used in rockbox. So it might help lower battery consumption right? (or does 1sec @low + 1sec @high drain the same as 2sec @mid ?)
13:19:28pondlifeI wouldn't expect it to be linear, based on limited desktop CPU experience. Isn't higher frequency normally an exponential power hog?
13:19:38pondlife(or thereabouts)
13:19:43LinusNi don't think we should go for more frequencies until we have a very good reason to
13:19:48Lloreandionoea: There's a small cost in time during the frequency change, but its pretty tiny.
13:19:59*jhMikeS would like to know power frequency vs. power and current curves
13:20:03LinusNKISS as always
13:20:28pondlifeI could argue that having a calculation is simpler than a lookup table.
13:20:42pondlife(Depends on the calculation though!)
13:20:52amiconnpondlife: As long as voltage is constant, cmos chips have a linear frequency vs. power dependency
13:20:53LinusNwell, if the calculation is there, and it works...
13:21:18 Join spiorf_ [0] (
13:22:26amiconnSome notebook chips can lower their voltage at lower clock, and then it becomes nonlinear
13:23:15amiconnLinusN: Full calculation is not there atm
13:23:55amiconnThe OF calculates everything, and uses MHz units (PP5020+) or even kHz (PP5002)
13:24:44 Join DerPapst [0] (
13:24:50pondlifeUltimately, whichever strategy can give the best battery life (whilst being 100% stable) is the one... we need to experiment, right?
13:24:51DerPapsthi :)
13:25:47DerPapsti like the idea at calculating the frequency at runtime...
13:26:33amiconnI don't
13:26:35w1ll14mwell actualy to be honest i'm still using my ipod at 24/60/90 MHz it runs better than normal frequency (especialy when using dsp)
13:26:40LloreanIt might be useful for game/graphical-plugins that require a bit more oomph than normal, but don't need full boost?
13:27:13pondlifeI was thinking it could result in less frequency switching.
13:27:13 Quit w1ll14m ("CGI:IRC (EOF)")
13:27:31jhMikeSThe supply voltage is more important to power consumtion that f
13:27:41amiconnpondlife: I hope to be able to implement switch-to-idle-on-sleep, that might help too
13:28:16Lloreanpondlife: VBR files could result in either a heck of a lot of switching, or if you have it 'hold peaks' as it were, time wasted at peaks.
13:28:20jhMikeSonly thing is, nothing can sleep during playback since codec yield(), not sleep()
13:30:39amiconnIf there's really nothing to do atm, codecs should sleep(0)
13:31:01jhMikeSoh, just forget does sleeping when waiting for a buffer available
13:34:11markunLlorean: oops, was I not supposed to move the "Best custom build for nano? Senab, EvilG, PJulius or what?" thread?
13:34:27 Join Thundercloud [0] (
13:36:13Lloreanmarkun: The unsupported builds forum is just supposed to have one build per thread. It's not really a discussion forum. But he'd already asked those questions anyway, and completely ignored the answer.
13:36:19LloreanErr, one thread per build
13:36:52LloreanThere's not really a good place for general questions about unsupported builds, because really outside of "Which one's the best" which nobody can answer anyway, theres nothing you can ask generally anyway.
13:38:22amiconnIs there a H10 firmware dump (unencrypted) available somewhere?
13:39:03Lloreanamiconn: There's an encrypted one at Bagder's site, I think. mi4code works on it, I believe.
13:39:13 Join Nico_P [0] (
13:40:16linuxstbamiconn: You need mi4code.c from here - - and the firmware from here - or
13:41:06linuxstbWhen you unencrypt the mi4 file, you'll need to strip the 512 byte (IIRC) header.
13:42:55jhMikeSI did PP5022.mi4 and remember a header strip thingy somewhere (option?)
13:43:48 Join krazykit_ [0] (
13:43:56 Quit spiorf (Read error: 110 (Connection timed out))
13:44:03jhMikeSof course it might be one of those days where everything I say is just plain wrong
13:47:57*pondlife has those most days
13:48:35*linuxstb has learnt not to talk during those days
13:49:32*jhMikeS just talks and hopes something is right by pure, random chance
13:52:38 Quit safetydan ("Ex-Chat")
13:54:35 Join crop [0] (i=c27f0812@gateway/web/cgi-irc/
13:55:35cropHas putting some data/code into iram sped up the wma codec? In the comments it reads "should"
13:55:57 Quit krazykit (Read error: 110 (Connection timed out))
13:57:25aliaskcrop: By the sounds of it they most likely will speed up the decoding process, but it hasn't been benchmarked.
13:59:01 Join rogelio [0] (n=rogelio@
14:02:47linuxstbsaratoga doesn't have a Coldfire target, so needs others to benchmark for him.
14:03:44croplinuxstb: how can it be benchmarked? Is there a plugin or something?
14:05:24 Join webguest63 [0] (i=c023110f@gateway/web/cgi-irc/
14:06:46 Quit lee-qid ("aufwiederbyebientotsayonara")
14:08:05linuxstbcrop: Yes, the test_codec plugin (not compiled by default - add it to apps/plugins/SOURCES)
14:08:42cropcrop: why was it excluded from SOURCES btw?
14:09:10linuxstbBecause it's not aimed at users. Same as the other test_* plugins
14:09:43linuxstbAlthough maybe (once the bugs are ironed out), it could be useful to include if users want to compare codec performance.
14:10:47 Quit Siltaar (Remote closed the connection)
14:11:45 Quit DataGhost (Nick collision from services.)
14:11:48 Join [Sur`DataGhost] [0] (
14:13:40pondlifeAnyone think that should be rejected? It sounds a familar request...
14:13:56pondlifeA configurable quick menu...
14:14:38peturtell 'm to use cfg files for quick changes ;)
14:14:56*GodEater_ agrees with petur
14:14:56pondlifeI'll let the experts do that...!
14:15:40croplinuxstb: and how can I use the plugin once I've built it? This won't happen the next two weeks −− my PC is bricked
14:19:02GodEater_crop: you're doing a good job of logging into IRC without a working PC ;)
14:19:54cropI'm using a friend's one. No RB dev env here.
14:21:17*pondlife misread that as "dev envy" :)
14:21:49*LinusN envies those who have time for rockbox development
14:22:25pondlifetime for lunch though
14:28:59croppondlife: how did you manage to find the "follow playlist" bug?
14:29:10cropit was rather tricky IMHO
14:29:41pondlifeI couldn't get the sim to show it, so just put lots of logfs in.
14:30:25cropSo it didn't occur on the sim?
14:30:53pondlifeNo. But the logfs let me know which routine was first "generating" the wrong filename.... then I spotted that playlist_peek(0) was being used elsewhere to get the current track.
14:31:08pondlifeSo I simplified, and it seems to work ok.
14:31:29pondlifeIs it ok for you, or have I borked something else?
14:31:41 Join Soul-Slayer [0] (n=Administ@
14:31:46cropHm... But miika must have had a reason for putting such an index there
14:31:55cropIt's ok
14:31:57pondlifeIt's been there a very long time
14:32:16pondlifeMaybe scrobbling? Is that working still?
14:32:19 Join w1ll14m [0] (i=546851d0@gateway/web/cgi-irc/
14:32:29pondlifeOr database playcount?
14:32:41 Join Crash91 [0] (n=crash@
14:32:54cropI don't use the DB
14:32:55Crash91any wps experts around?
14:32:56Soul-SlayerHey, I made a battery bench of my gigabeat playing a 192kbps album on repeat using default settings... If I upload the results to paste bin could someone tell me if they seem right? I'm not sure how to read it properly :$
14:33:05pondlifecrop: Me neither
14:33:34pondlifeMaybe some kind soul who uses scrobbling and/or the database runtime stats will read this and test it for me.
14:33:43pondlifeAnyway, gotta go and eat
14:33:48Crash91test what
14:33:58croppondlife: me too. Have to go. Bye
14:34:00 Part crop
14:34:32Soul-Slayer Does this seem right for a healthy gigabeat battery playing a 192kbps album?
14:34:32 Quit billytwowilly (Remote closed the connection)
14:34:48 Join billytwowilly [0] (
14:35:02JdGordonany opinions on FS #7417?
14:35:12markunSoul-Slayer: looks normal
14:35:24Soul-SlayerPhew... Thanks
14:35:34Soul-SlayerWhich means I still need to work out why it drained so fast yesterday
14:35:35markunalthough 16 hours is not the highest value we have seen
14:35:49Soul-SlayerWell.. I only managed like 4 from it yesterday which was making me wonder
14:36:06markunthe hdd and the backlight use a lot of power
14:36:39pondlifeJdGordon: I say reject it (although I'm technically at lunch)
14:36:41Crash91JdGordon: I would like that
14:37:00markunSoul-Slayer: so if you browse a lot without dircache you drain the battery fast for example
14:37:02JdGordonpondlife: well.. i was more asking about the patch i attatched
14:37:08pondlifeOops, sorry
14:37:22*JdGordon would like to replace the existing quickscreen with that
14:37:25pondlifeI was going to say "Bloat!" and "Support Nightmare!"
14:37:35pondlifeWill look later though
14:38:04Soul-Slayermarkun: There's not really a way of getting around using the backlight though if I want to see what on earth I'm doing :P
14:38:26markunSoul-Slayer: not really no, but you can lower the brightness and have a short timeout
14:38:44Soul-SlayerAlso, for 'antiskip time', I was trying to work out which would require the least disk spinups, which would it be? 5s?
14:39:14markunI don't think that option has much to do with spinup
14:40:35Soul-SlayerRight, well I'll lower my backlight brightness and time it's on for and hope for the best, thanks a lot :)
14:41:00markunand do you have dircache on?
14:41:45Soul-Slayer90% sure I did, I had an H10 before I got the gigabeat and if I didn't have dircache on on that it took forever to navigate anywhere, so I think I'd have turned it on out of habit, but I'll make sure I do this time
14:42:40markunSoul-Slayer: here is some info on power consumption;filename=current-07-06-26.txt
14:43:32Soul-SlayerSo the backlight practically halves battery life? Youch
14:43:56 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
14:44:25Soul-SlayerAlthough, that means if the backlight was on 100% of the time yes?
14:44:49Crash91any wps experts awake!?!
14:46:00Soul-SlayerI doubt the button lights drain significant power, but they don't really make a difference to me so I'll leave them off
14:46:04GodEater_Crash91: as I believe I've said before - not many of the "WPS Experts" hang out here
14:46:46Crash91CodEater_:Yes, you have, but it was a "not many"
14:46:50 Quit webguest63 ("CGI:IRC")
14:47:01GodEater_Crash91: okay - change that to "none of"
14:47:02Crash91GodEater_:I mean GodEater_ :p
14:47:09Crash91not CodEater
14:47:19 Join axlgreasetires [0] (i=a233e050@gateway/web/cgi-irc/
14:47:44axlgreasetiresWhat's up noo0bs
14:47:56 Join saratoga [0] (i=9803f670@gateway/web/cgi-irc/
14:48:05Crash91i need someone who is willing to check this out on a sansa sim build, pref. windows
14:48:15saratogaamiconn: I've been listening to my sansa for 3 days now without hearing any skipping
14:48:37axlgreasetiresCrash whats wrong with the build?
14:48:48Crash91i have an evil feeling that my build is messed up
14:48:53Crash91even after reverting svn
14:49:04saratogathough i've been listening to mostly high bit rate mp3, so maybe i'm boosted frequently enough that theres no issue?
14:49:09w1ll14mcrash: i would do a clean svn
14:49:35w1ll14minstead of reverting....... i screwed up some times even when a revert was a succes
14:49:39 Quit linuxstb (Nick collision from services.)
14:49:41 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
14:49:42 Quit saratoga (Client Quit)
14:49:56Crash91i need somewone to test that wps
14:50:09Crash91*for SANSA*
14:50:52 Quit axlgreasetires (Client Quit)
14:50:53 Quit spiorf_ (Remote closed the connection)
14:51:20Soul-SlayerI'm not sure if this will help, but I don't have wiki write permissions, so if anyone wants to put my results ( here (, was running a 192kbps album at default settings, 2007-07-09, SVN of that day
14:51:43 Join lee-qid [0] (
14:53:10 Join Arathis [0] (
14:53:17markunSoul-Slayer: I can give your write permission
14:53:21 Join PaulJam [0] (
14:53:41Crash91someone! anyone! test this on a sim build of that
14:53:47Crash91the Sansa*
14:55:22PaulJamCrash91: you could download an uisim fron rashers page and test it yourself:">
14:55:45Soul-Slayermarkun: Thank you, my account name is JohnathonMihalop
14:55:47 Join saratoga [0] (i=980398fe@gateway/web/cgi-irc/
14:56:04Crash911. Im running linux, 2. i think my sim is screwed (even after original svn revert) 3. TEST IT!
14:56:05 Join _Veseliq_ [0] (
14:56:38Crash91saratoga, i believe you have a sansa?
14:56:57saratogaSansa and Gigabeat
14:57:08markunSoul-Slayer: done
14:57:14Crash91 < if you have a sim of the sansa handy, pls test this
14:57:29LloreanCrash91: Please don't repeat messages over short periods.
14:57:31Soul-SlayerThanks markun
14:58:16Soul-SlayerNow then... Editinga a table... This could prove interesting
14:58:22Crash91Llorean: ok, but seeing as Saratoga is helpful and he signed in after my last post (1mins 59 exactly)
14:58:33pixelmaCrash91: I see an error in line 10 -> there's a single questionmarks after the %ar (seems a leftovers)
14:58:46markunSoul-Slayer: you can just attach the file
14:59:00saratogaCrash91: what do you want me to do?
14:59:17markunI noticed that some results got overwritten because people want to use the same filename
14:59:36Crash91please test that as a wps file, its pure text
14:59:57 Quit _Veseliq_ (Client Quit)
15:00:05Soul-SlayerI can see what I'm doing with the table now, I'll do both though
15:00:06Crash91pixelma: there are multiple %ar tags, you may not be able to see the symbols
15:00:11saratogaCrash91: how do I do that?
15:00:22pixelmaCrash91: sorry for singular/plural mixup but I saw one first but there's a second - that would make the wps parser fail
15:00:35Crash91saratoga: save it as a wps file and use it on sim/sansa
15:00:43pixelmayou have "%ar?" two times...
15:00:46Crash91saratoga: tell me any errors
15:00:59Crash91pixelma: do you mean this symbol? ᛣ
15:01:12saratogaCrash91: how are WPS files used on the device?
15:01:38Crash91paste it as "text.wps" in .rockbox/wps/
15:01:56Crash91then use file browser and click select when the file is highlighted
15:02:00pixelmaI don't mean symbols (and I don't see your because of codepage mismatch in my IRC) - I mean the wps tags
15:02:10***Saving seen data "./dancer.seen"
15:03:15Crash91pixelma: ill repeat for sake of clarity > " ᛣ " and the %ar tags have that same symbol after them, try copy-pasting the stuff in quotes in notepad etc.
15:03:45pixelmano... I don't mean that
15:03:45saratogaCrash91: it fills my screen with random symbols
15:03:53Crash91i forgot to mention
15:04:01 Quit SirFunk (Connection timed out)
15:04:09amiconnsaratoga: There is probably an issue with lcd updates happening when the cpu is unboosted and music is playing (cache flush taking too long). Ask jhMikeS for details
15:04:11PaulJamCrash91: if you have a simulator bilt under linux then you could try to run the uisim with −−checkwps. this could give you a hint where it fails.
15:04:13Crash91use the font "6x10"
15:04:37 Join krazykit [0] (
15:04:47Crash91PaulJam: its not failing, i seem to be having loads of alignment probs recently
15:05:10pixelmalike in this part (start of line 10): %?mp<Stop|?? -%pr _%ar?_ %?ig... (underscored what I mean) - that's wrong syntax
15:05:47Soul-SlayerAll done
15:06:18saratogaCrash91: is there some reason you can't do this in the sim?
15:06:19linuxstbsaratoga: I've now got the "multiple payloads in a single packet" working. The ffmpeg encoded test file now plays without giving any decoding errors, but the output is just static. The wmv file (containing mixed audio/video payloads within the same packet) is playing perfectly.
15:06:26Crash91pixelma: confirmed, the "?" you see are special symbols, gedit under linux can see them
15:07:01saratogalinuxstb: is this the v1 test file?
15:07:16saratogai may not have comitted all the fixes on my end
15:07:23saratogasince i couldn't test them
15:07:29linuxstbAnd I've given up maintaining my standalone test programs - I'm just using the simulator and DEBUGF().
15:07:35 Join _Veseliq_ [0] (
15:07:36Crash91saratoga: did you try to use the "6x10" font?
15:07:40Crash91with the wps
15:07:49pixelmaCrash91: wonder why I see the other (double vertical lines) in the pastebin then
15:08:01saratogaCrash91: i wasn't sure how to change fonts so i gave up
15:08:15Crash91they must be included in the spec. charset you are using
15:08:16saratogaregarding the parser, do you have a test driver I could use with it and the decoder?
15:09:42saratogamy current one is ffmpeg based and now gettting rather far out of sync with recent changes
15:10:07saratogaamiconn: should i revert the frequency scaling commit then?
15:10:29saratogajhMikeS: maybe i should have asked you about that
15:10:34DerPapstCrash91: i don't get it.. why can't you checkout a clean version of the trunc and test your wps then? is SDL broken?
15:11:18Crash91DerPapst: I did svn revert -R . and then made a fresh sim build
15:11:46Crash91DerPapst: everything worked fine, but now wps is acting wierd
15:11:53jhMikeSsaratoga: 'bout what now?
15:12:01DerPapstCrash91: ah ok
15:12:02saratogathe frequency scaling on the sansa
15:12:14linuxstbsaratoga: That's what I just said - I don't have working standalone test programs any more, I find it easier to just use the sim.
15:12:14 Quit XavierGr (Nick collision from services.)
15:12:17 Join XavierGr [0] (
15:12:43saratogalinuxstb: sorry, missed that comment
15:13:07w1ll14mwhere should i place the .voice file? was it in voices or langs ?
15:13:09jhMikeScould be a good thing to do that since it borks recording really badly too. I have some ideas about fixing the problem with crackling
15:13:34w1ll14mdamn so that's why it didn't worked :)
15:13:44saratogajhMikeS: ok, i'll revert scaling
15:13:51Crash91if you need english voice, then there would have to be english.lng and english.voice
15:13:51linuxstbsaratoga: I'll commit my changes to the parser soon. It seems to be working well, I just need to clean it a little and test on some targets.
15:14:10 Join krazykit1 [0] (
15:14:13jhMikeSJust removing the HAVE_ADJUSTABLE_CPU_FREQ from the header?
15:14:17saratogalinuxstb: ok good
15:14:32saratogathats all i did to enable it
15:14:37PaulJamCrash91: what exactly is the problem with the wps? i see that tere is a problem with ling titles and on line 6 you shoud remove the spaces at the end of the line
15:14:55jhMikeSother than that it works fine and the batt life seems a good bit better
15:15:36jhMikeSwell, the sluggish UI too
15:15:49 Quit krazykit_ (Read error: 110 (Connection timed out))
15:16:02saratogaso what is the long term solution to the problem likely to be?
15:16:15Crash91PaulJam: ling titles? and i will try to remove the spaces
15:16:42PaulJamlong titles
15:17:22PaulJam(i have too much stuff lying in front of my keyboard)
15:17:33Crash91PaulJam: Long Titles would be a problem, yes, but i cannot help that
15:18:28Crash91%s scrolls the whole line and i do not want to use bmps as the whole point of this is to be pure text
15:18:50Crash91i wish screen rotation (i.e. 90 and 270 degrees) was enabled
15:19:20 Join richcoosa19 [0] (n=richie@
15:19:41PaulJami just wanted to mention it, because i missed the part where you said what doen't work with the wps.
15:19:53Crash91Well, its working, but my sim keeps saying codec failure
15:19:55LloreanCrash91: It's not a case of simply enabling something.
15:20:27Crash91Llorean: i know, whole lots of code rephrase that to "implemented"
15:20:54Crash91PaulJam: thanks a lot, i guess the stuff at the end of the line was to blame
15:21:21jhMikeSsaratoga: something to fill the FIFO just before flushing any caches. of couse i'd like to try out having the framebuffer in non-cached memory just to see how it works...when someone finds out how
15:22:40jhMikeSI'm thinking a forced interrupt right before the flush or maybe messing with the FIFO thresholds just to cause one...not sure yet what will work nicely.
15:23:03PaulJamCrash91: another thing in line 10: behind the "No Tag->" you have on playback one and if paused 2 spaces.
15:23:39 Quit GodEater_ (Read error: 110 (Connection timed out))
15:23:46Crash91PaulJam:If you read carefullt=y, you will see two %t0.5
15:24:03Crash91this means if paused, || will blink every half a second
15:24:31PaulJamCrash91: line 10 not 11
15:24:31 Quit w1ll14m ("CGI:IRC (EOF)")
15:24:32 Join bdgraue [0] (
15:25:43 Quit krazykit (Read error: 110 (Connection timed out))
15:25:45Crash91PaulJam it is correct
15:26:06PaulJamCrash91: this causes the genre string to move when pausing/unpausing
15:26:28Crash91if stopped, display stop, if paused/playing, display genre, if fwding/rwding display position in song
15:27:00Crash91PaulJam: aah, true, thanks
15:27:23amiconnjhMikeS: Does that mean even the fiq is delayed during cache flush?
15:28:17Crash91PaulJam, fixed
15:29:10jhMikeSamiconn: yes, it doesn't occur and the FIFO runs out of samples
15:30:04jhMikeSoddly enough the OF flushes the cache on every single DMA start when communicating with the SD card so some solultion must be possible
15:30:16Crash91PaulJam: final version looks like:
15:30:49 Quit DerPapst (Read error: 110 (Connection timed out))
15:31:21jhMikeSit always checks CACHE_CTL just before doing it and I see 0x00004007 there
15:37:08 Quit richcoosa19 (Remote closed the connection)
15:37:23 Join richcoosa19 [0] (n=richie@
15:37:23 Quit XavierGr (Nick collision from services.)
15:37:26 Join XavierGr [0] (
15:37:38 Nick richcoosa19 is now known as richcoosa18 (n=richie@
15:38:07linuxstbsaratoga: I've committed my parser changes.
15:38:47 Quit krazykit1 (Read error: 110 (Connection timed out))
15:43:26 Join SirFunk [0] (
15:43:50 Join low_light [0] (i=c730180b@gateway/web/cgi-irc/
15:44:04 Join richcoosa19_ [0] (n=richie@
15:46:21 Quit XavierGr (Nick collision from services.)
15:46:24 Join XavierGr [0] (
15:48:46 Part LinusN
15:49:00 Join chrisjs169_ [0] (
15:49:14JdGordondoes anyone else get heaps of "warning: pointer targets in passing argument 2 of ‘gui_syncsplash’ differ in signedness" type warnings doing a target build?
15:50:26 Join perplexity [0] (
15:50:34linuxstbsaratoga mentioned something similar. Works fine for me though...
15:52:11linuxstbAre you using the recommended gcc?
15:52:13Crash91why is the progbar on the recorder different from that of others?
15:52:26JdGordonyeah, just recompiled using
15:52:38linuxstbCrash91: How is it different?
15:52:59pixelmado you mean on the Player?
15:53:02Crash91instead of filling up the bar,
15:53:12Crash91it is like a scrolling box
15:53:22Crash91yes player and recorder
15:54:02JdGordonis there a way to disable a specific warning in gcc?
15:54:29Crash91Archos Player/Studio: This will display a 1 character "cup" that empties as the time progresses. < i do not see how that is a "cup"
15:55:00pixelmathe Player has a charcell (no bitmap) display - ah and these screenshots are old. On the Recorder (other bitmap Archos) it's the same now too - look at the last entries of the site you linked to
15:55:05Zagorcup sounds better than square :-)
15:55:50Crash91ok :)
15:56:19 Join bluebrother [0] (i=IbaDdDzE@rockbox/staff/bluebrother)
15:56:31 Quit bdgraue (Remote closed the connection)
15:58:00 Join w1ll14m [0] (i=546851d0@gateway/web/cgi-irc/
15:58:23pixelmaCrash91: and there is no screenshot of the "cup"
15:59:44LloreanIsn't the cup basically what he was doing with his volume indicator, but in a single cell rather than three?
16:00:15Crash91who do you mean by he? me?
16:00:35Crash91mm, i kinda get it
16:01:00Crash91but the ||s were to separate them as it was a conditional
16:01:12JdGordoncan I bug everyone about FS #7417 again? (and hopefully, someone with a working sh compiler can tell me the bin size hit please?)
16:02:06 Quit richcoosa18 (Read error: 110 (Connection timed out))
16:03:11LloreanI thought the consensus leaned toward "customizable menus are bad"
16:03:28 Quit chrisjs169 (Read error: 110 (Connection timed out))
16:04:11JdGordonyeah, but I think this menu is more usefull than the current quickscreen
16:04:32JdGordonits not really customizable menus anyway.. its just a more flexible one
16:04:52LloreanTwo of the three buttons in the quickscreen are functions that many MP3 players have assigned directly to a button anyway.
16:05:23Crash91 And it will include the items that the user chooses to have < sounds like customisable
16:05:42 Join bdgraue [0] (
16:05:46Crash91and what do you mean by 5 quickscreen items, how is that possible?
16:06:14 Quit XavierGr (Nick collision from services.)
16:06:14amiconnJdGordon: The quickscreen is just not that useful atm because actions broke it, and it's still not back to normal
16:06:17 Join XavierGr [0] (
16:06:46*JdGordon knows that... it also never works the way it used to on any targets except the recorder anyway
16:07:37 Quit pondlife ("disconnected has pondlife")
16:07:38 Quit low_light ("CGI:IRC")
16:07:39LloreanI do think that with the Database having a separate main menu option, and the existence of the playlist catalog, that maybe the third option in the quick menu needs rethought.
16:08:35JdGordoncan someone with a working sh please try a compile and let me know the bin cost?
16:09:17JdGordonI mean, if its less than 2 or 3 hundred bytes then id be inclined to go for it because imho its a nice addition which wont negativly affect anyone..
16:09:33*JdGordon suspects it will be a bit more than that though
16:09:35pixelmathe Ondio doesn't have a quickscreen at all and I very seldom miss on.
16:09:42pixelma*one even
16:09:53Crash91It would be nice if the rec button on the sansa activates the recording immediately
16:10:13Crash91unless youre in a plugin, the rec button isnt really used for anything
16:10:19JdGordonit would, but we cant agree on the correct button mapings for it
16:10:53LloreanI'm pretty sure nearly everyone thought a long press of it should go to the recording screen.
16:11:08*JdGordon thinks thats counterintuitive
16:11:17JdGordon... but id rather talk about the menu patch :p
16:11:50Crash91in the menu and wps at least, it servers no purpose, and the user could choose whether it asks/starts recroding immediately/does nothing
16:12:38LloreanMaybe make "tap" enter the recording screen, and "hold" enter the recording screen *and* start recording?
16:12:41pixelmaah dionea... want to take a look at new tiles for the jewels puzzle?
16:12:51amiconnI'd certainly not want the recording button to start recording directly
16:13:07LloreanI wouldn't want a short press to ever immediately start recording.
16:13:07pixelmasorry, dionoea :)
16:13:18amiconnThis could start an unwanted recording too easily
16:13:24JdGordonLlorean: I would be happy with that
16:13:34LloreanJdGordon: The short/long suggestion?
16:13:36JdGordon/usr/local/sh-elf/lib/gcc/sh-elf/4.0.3/../../../../sh-elf/bin/ld: region IRAM is full (/home/jonno/rockbox/recv2/apps/rockbox.elf section .ibss)
16:13:37dionoeapixelma: it's ok, I highlight a few versions of my nick :)
16:13:40dionoeawhere are they ?
16:13:45Crash91amiconn: you could set it to ask,
16:14:12aliaskThen it's just annoying.
16:14:22aliaskLong press makes the most sense IMO.
16:14:24Crash91or maybe the abov mentioned working with a long press
16:14:26*pixelma remembers that being the case (start recording accidentally) from the OndioFM radio screen until that was changfed to a "double-click"
16:14:42 Join Naked [0] (
16:15:00 Nick Naked is now known as Hadaka (
16:15:04JdGordondouble click would be fine also, as long as a short click entered the rec screen
16:15:12Crash91mm, double click is good
16:16:17Crash91how about default>long press, user can customise whether to ask/record/nothing and also choose if it is long/short/dbl click which activates it
16:16:49JdGordonuser configurable buttons is bad...
16:16:55JdGordonmenus on the other hand....
16:18:05Crash91JdGordon:Its not configurable buttons, but whatever, ill settle on long press bcos if you can have accidental clicks, youll probably have accidental double clicks as well
16:18:55amiconnAccidental double clicks are a lot less likely than an accidental long press
16:19:11dionoeadouble long press ?
16:19:15 Quit Zagor ("Client exiting")
16:19:20amiconne.g. squeezing the button when carrying it in a case
16:19:33Lloreanamiconn: As long as a single click doesn't go to the screen at all, then yeah. If it's double click as in "first click goes to the screen, second click starts recording" then it's quite easy to happen.
16:19:37JdGordonI'd like to have it record while rec is being held down so it can be used as a quick dictophone thing
16:19:49 Join low_light [0] (i=c730180b@gateway/web/cgi-irc/
16:20:06 Join XavierGr_ [0] (
16:20:08 Quit XavierGr (Nick collision from services.)
16:20:23amiconnJdGordon: What the hell are you doing?? (wrt iram problem)
16:20:33JdGordonnothing... its a fucked gcc
16:20:38JdGordonit happens with svn also
16:21:30Crash91amiconn:ok whatever, you win :p
16:21:30LloreanWhat file needs comparing to see bin size increase?
16:21:47amiconnJdGordon: Forgot to apply rockbox #1?
16:22:06JdGordondid it with
16:22:11JdGordonLlorean: yeah, i tinhk so
16:22:27Crash91amiconn: but what if rockbox is getting a bit laggy, then you press it a few times before rockbox catches up
16:22:37Crash91and it starts recording
16:22:40LloreanCrash91: If Rockbox is getting laggy, something else is wrong.
16:23:11Crash91Llorean, sometimes it is laggy if youre playing music and have just exited from a plugin
16:24:22Crash91nvm lets not argue about this until someone actually makes a patch
16:25:31Crash91i guess its something like ACTION_TREE_?RECORD?, BUTTON_RECORD|BUTTON_REPEAT, BUTTON_RECORD
16:26:02Crash91thats a very wild guess as i do not know C
16:26:10JdGordonthats C ?
16:26:37aliaskPretty much. The patch itself is fairly trivial, so 90% of the "work" will be in discussing the best option.
16:27:04LloreanJdGordon: Comparing two ajbrec.ajz for Recorder, shows 280 bytes.
16:27:04 Quit Entasis (Read error: 104 (Connection reset by peer))
16:27:15JdGordonok thanks
16:27:26JdGordonis that OK?
16:28:07JdGordon+280bytes... half of them is the added strings :p
16:28:19LloreanI'm still against it.
16:28:31LloreanAnyone who needs custom menus can just make a folder and throw a couple .cfg files in it.
16:28:39LloreanAll with 0 bin size increase.
16:28:58JdGordonyour gonna make a seperate cfg for every possible settings?
16:29:06aliaskSpeaking of space in the added strings, I think I missed the status of the langv2 cleanup in the conversation earlier. Is it anywhere near completed?
16:29:49LloreanJdGordon: How many "quick" options are going to have a wide range of settings? The whole point of a quick menu are frequently changed options, and most of those only have a couple choices.
16:29:55LloreanOr at least, a couple choices that the user uses.
16:30:19Crash91Its a good idea, but not that essential
16:30:29Crash91i could do withoutit
16:30:35JdGordonI dunno... i'd probably put volume, shuffle, brightness, backlight in it
16:30:40amiconnJdGordon: Comparing ajbrec.ajz sizes does *not* compare true binary size (for recorder v1)
16:30:46LloreanJdGordon: Volume? Seriously?
16:31:02*JdGordon is a bit odd...
16:31:20amiconnAlways compare rockbox.bin
16:31:28*Llorean has a lovely gigabeat that lets him adjust volume in the menus.
16:31:31*JdGordon actually just realised it would be 1 extra keypress from the menus :p
16:31:42Crash91ok, brightness> i use VERY rarely, backlight settings, not really
16:31:56JdGordonCrash91: who's side are you one!!??
16:31:57LloreanDo you use all the brightness levels, or just two or three?
16:32:03Crash91i mostly use browse themes and plugins, which are the main menu
16:32:09 Join n1s [0] (
16:32:14amiconnHmm, and I should have addressed that to Llorean
16:32:28JdGordonamiconn: force of habbit there... ? :D
16:32:28Lloreanamiconn: I'll remember that next time. Thanks. :)
16:32:31Crash91deafault (5/6?) and 1 are the only ones ive ever used
16:33:19Crash91JdGordon: I have already said, good idea, not really essential, and im showing you jsut how useless it is
16:33:21JdGordonout of 180+ settings, I think its still bad to limit the quick access ones to 3 hardcoded ones
16:33:42dionoeabtw, about plugin browsing, anyone against applying the patch to sort plugins in different subdirectories if they're games, demo plugins, viewers, etc... ?
16:33:52JdGordonis it finished?
16:34:04Crash91dionoea, nope
16:34:08LloreanJdGordon: As I said, two of those settings, many MP3 players have directly on buttons.
16:34:25amiconnI only use one single quickscreen option more often than once a month, and a second one (in the second quickscreen only available on recorder) maybe twice per year...
16:34:29LloreanThey're basically essential playback options.
16:34:47LloreanMost CD players have a "Shuffle/Repeat" button as well.
16:34:55amiconnThe first is 'file view', and the second is screen flip
16:34:55LloreanAt least, every one I've owned, I cant say "most"
16:35:09Crash91i usually get the quickscreen by mistake when im trying to press power
16:35:28 Quit low_light ("CGI:IRC")
16:35:44Crash91JdGordon: Which target do you intend it for?? It says only sansa
16:35:52JdGordonall of them
16:36:12amiconnAnything else that's in the quickscreen is changed so infrequently by me that I don't even remember it's in the quickscreen(s)
16:36:20 Quit bdgraue (Remote closed the connection)
16:36:22 Join XavierGr [0] (
16:36:33JdGordonLlorean: amiconn: i;m not suggesting (just yet) that the QS be removed and this added...
16:36:46LloreanJdGordon: A normal menu there would be decidedly less quick.
16:37:01JdGordonits an added menu in the main menu...
16:37:19LloreanThen I really don't see the difference between that and a folder of .cfg files.
16:37:20Crash91i think select button in WPS is more useful than this
16:37:25JdGordon<menu > <down> <select> to your 16 favorite settings
16:37:31amiconnI never change repeat mode except for testing, and I also never used shuffle mode since I got accustomed to 'insert shuffled'
16:38:03amiconnWhen I used to use shuffle mode, I always forgot to disable it again...
16:38:12JdGordonamiconn: I am exactly the same, i never use the QS becuase imho the optinos are stupid... this is why i want to put this in
16:38:46*Llorean uses both the Shuffle and Repeat options on the quickscreen regularly
16:39:05amiconnJdGordon: So.. what does this offer that can't be done the normal way? And how does it deal with the 2 now-quickscreen buttons on recorders?
16:39:07LloreanAny option is stupid if you don't use it.
16:39:37*JdGordon wonders if the patch was actually looked at.. or just the description....
16:39:46JdGordonthis adds a new menu item in the root menu
16:39:51JdGordondoesnt touch the QS at all
16:40:15bluebrothertalking about some "favourites" like menu?
16:40:24Lloreanbluebrother: Yes, a menu of favorite menus, basically
16:40:34JdGordonno, a menu of favorite settings
16:40:39bluebrotherah, that feature request ...
16:40:41LloreanJdGordon: Every setting is a menu
16:40:46JdGordonno its not
16:40:47LloreanUnless you're talking about settings actually preset to something
16:40:51JdGordon.... not in the code....
16:40:52LloreanLike Volume:-37
16:40:55amiconnDoesn't sound like a good idea to me...
16:41:12bluebrotherwell, I'm not sure if I really like a favourites menu
16:41:19LloreanBut "Volume" is a menu with options ranging from max volume to min volume
16:41:37LloreanOr at least a list.
16:41:43amiconnJdGordon: Menu of favourite settings: Settings->Manage setting->Browse .cfg file
16:42:09bluebrotherwell, the volume setting is reachable throuth the sound settings menu
16:42:17JdGordonI could have done the patch with about the same amount of complexity to allow it to really be a favorites menu (i.e a menu of any menu in the settings menu), but I DIDNT.... this is for settings only
16:42:19bluebrotherso it is a setting
16:43:05amiconnAnd unlike a "favourites" menu, "Browse .cfg file" is more flexible and powerful
16:43:15amiconnAnd it already exists
16:43:18JdGordonand more annoying
16:43:28bluebrotherwell, that entry could get renamed "Configurations" or something similar
16:43:35*Crash91 agrees with JdGordon on this one
16:43:37bluebrotherwhat's annoying about that?
16:43:44LloreanThe fact that you have to set it up first.
16:43:56LloreanBut if we're talking "favorites" anyway, you're setting up something either way.
16:44:00JdGordonand have 1 cfg for each possible setting/value you want to use
16:44:01bluebrotheryou have to do that for any favourites-like menu
16:44:05Crash91the fact that manage cfgs dumps everything
16:44:19LloreanCrash91: JDGordon's method requires use of a text editor anyway
16:44:26bluebrotherwell, you can always create a customized cfg that only changes what you want to change
16:44:29 Join Nudger [0] (
16:45:01Crash91how about an "add to custom.cfg in root" in the context menu of a setting
16:45:32bluebrotherwhy not create a plugin that can enable or disable lines from a cfg?
16:45:34Crash91it would automatically add that to the file "custom.cfg" in the root, then it would be easier
16:45:42JdGordonbah, the point isnt that you can easily change a setting _back_ to the value you like, its to be able to change a setting to any of the legal values...
16:45:49Crash91bluebrother: ever hear of text_editor?
16:45:56Lloreanbluebrother: A simple .cfg editor? Basically just give it the ability to add or remove a # from a line?
16:46:02 Quit XavierGr_ (Connection timed out)
16:46:06NudgerI was wondering if anyone uses Rockbox in conjunction with ml_ipod for Winamp.
16:46:09bluebrotherLlorean: yup, something like that
16:46:17NudgerBasically, I stream all my music from my iPod straight through Winamp. Can I still do that with Rockbox?
16:46:18LloreanCrash91: A slimmed down interface even on that. Line highlight, select inserts a # or removes one if present, no other features.
16:46:22bluebrotherCrash91: I know text_editor. But it's too difficult to use for that case IMO
16:46:32LloreanNudger: Rockbox doesn't affect its ability to be used as a mass storage device
16:46:36Crash91true on both counts
16:46:38NudgerLlorean Excellent :)
16:47:01NudgerOne other question - if I do go ahead and install this, it's the equivalent of formatting the iPod, right? I'll have to back up all the music first?
16:47:02bluebrotherjust present option and value as list, and clicking it enables or disables it by commenting in the file
16:47:09LloreanNudger: Nope.
16:47:11aliaskNudger: Nope
16:47:16Nudgerhurrah! :)
16:47:23Crash91Nudger: Nope ?
16:47:27JdGordonmodifying text_editor could be done easily.. but thats not the point of the damn patch
16:47:30 Join webguest80 [0] (i=cdae0804@gateway/web/cgi-irc/
16:47:36 Join GodEater_ [0] (
16:47:38bluebrotherNudger: you might be interested in the manual and the FAQ
16:48:00 Join ackbahr [0] (
16:48:05LloreanJdGordon: Our point is, there are lost of alternatives that can help avoid binsize increase.
16:48:07GodEater_linuxstb: btw - unless Bagder's parcel is waiting for me at home tonight - we shall be T-Shirtless come tomorrow :(
16:48:26 Quit Nudger (Client Quit)
16:48:48JdGordonLlorean: right, but really, we have much bloatier patches in that add nothing...
16:48:50LloreanJdGordon: The ultimate being to recreate the menu system with .cfg files, and offer it as a download from the extras page. Users can delete the options they don't want, and rearrange the folders. You still get a fully customizable set of settings, and no binsize increase still. It's OCD, but it'd work.
16:49:15LloreanIt'd be functionally equivalent.
16:49:17JdGordon... and they say our UI is bad currently :D
16:49:33LloreanJdGordon: It'd look just like our UI, in "Supported" mode.
16:49:38bluebrotherwell, do we really care if "they" call our ui bad? It's quite good.
16:49:48 Quit w1ll14m ("CGI:IRC (Ping timeout)")
16:50:08LloreanMy point is, there's a functionally equivalent way to do what you're suggesting without any binsize increase at all. So why is your way better?
16:50:26JdGordonbecause 99.9% of devs/users are not OCD
16:50:40LloreanBut the menu system needs to be created exactly once.
16:50:48LloreanThen users just copy what they want out of the folder structure.
16:50:50Crash91Llorean:We could put it as a plugin in extras (the CFG editor)
16:51:06LloreanCrash91: Plugins don't need to be extras, they don't follow the same rules.
16:51:10LloreanOr rather, restrictions
16:51:15JdGordonexactly ones, plus every time the settings change... i.e new ones or removed ones
16:51:23 Join fed [0] (i=4a7b3e0c@gateway/web/cgi-irc/
16:51:31LloreanJdGordon: And couldn't a script generate them?
16:51:45webguest80My H10 20gb has been spordadically shutting down during play, usuallly on the first track. This started about 5 days ago upon updating the firmware. Is there a known issue for this? I want to keep the latest version, since wma is getting more stable.
16:51:56LloreanAbsolutely, since a script can create .cfg files.
16:51:56fedI would like to ask about the new changes to recoding.c Can anyone help me?
16:52:22bluebrotherwhy have customizable menus anyway?
16:52:37Arathiswebguest80: it's a known thing, but as far as I know there is no solution yet.
16:52:48Lloreanbluebrother: Because he's unhappy with the options offered in the quickmenu.
16:52:50*JdGordon sends his best of luck to Llorean to generate a script to build cfg menus from settings_list.c
16:52:55Arathisat least amiconn didn't come up with one when I was around ;)
16:52:56LloreanArathis: Is there a flyspray task containing information on it?
16:53:08LloreanJdGordon: *I* don't want customizable menus
16:53:24LloreanYou do, but you want to do it in a way that's an infinite percent more bloated for binsize than necessary.
16:53:26bluebrotherI don't think customizable menus are good.
16:53:34 Quit Soul-Slayer (Read error: 110 (Connection timed out))
16:53:53ArathisLlorean: don't know. but as said amiconn knows about it and gatherd some information about it from me yesterday
16:53:57bluebrotherwho wants to support users messing up their menus? Especially without getting paid?
16:54:07webguest80Thanks, which version can I go back to, should I decide I'm hurting bad enough? Many Thanks. On another note, how do custom view get used in the database?
16:54:19Lloreanbluebrother: It'd just be one user-menu, the option wouldn't be there to mess up or modify any of the static menus.
16:54:21Crash91why have it anyway, its not like you need to change a setting before the world ends/your car crashes/your plane crashes/you get shot
16:54:38Lloreanbluebrother: But it's still 100% redundant to .cfg files.
16:54:49fedI guess I'll try again later...
16:54:51Crash91i would rather waste a whole 5 seconds of my life going to settings and getting whatever i want
16:54:52 Part fed
16:55:00*JdGordon bangs head on wall... this isnt customizable menus... its an additional one which te user has to setup manully so they cant break the inbuilt system.. and it add <300bytes
16:55:04bluebrotherwell, if it's just settings then I see absolutely no reason for it.
16:55:14LloreanJdGordon: It's been pointed out my measurement was wrong.
16:55:44JdGordonunless i get another number, im using that one!
16:55:44*Crash91 bangs his head. The fact of life is, not everyone is running out of time like Jack Bauer
16:55:45LloreanJdGordon: And it's also 300 bytes (or however much) that can be avoided 100%. You can recreate any menu but the graphical ones (EQ, colors) by way of .cfg files, as I said.
16:55:53 Join lazka [0] (
16:56:02LloreanThey even look the same.
16:56:05LloreanMinus the icons.
16:56:36bluebrotheralso, if we add a customizable settings menu, why not make everything customizable?
16:56:43bluebrotherusers will be asking that.
16:56:46LloreanI still want to know what advantage coding it offers over .cfg files.
16:56:52bluebrotherand I am really against customizable menus
16:57:00Llorean.cfg files let you have an unlimited selection of customizable menus with unlimited entries in each.
16:57:08LloreanHis gives you one menu with 16 entries *and* increases binary size
16:57:29aliaskBut with the added advantage that it's accessible from anywhere.
16:57:33*Crash91 wil be right back to rant/argue about this in 10 mins
16:57:50Lloreanaliask: Anywhere?
16:57:56Lloreanaliask: It's an entry in the main menu.
16:58:00aliaskWell, anywhere where you can access the menu
16:58:09LloreanAnd the root of your player is one press away from the menu.
16:58:39JdGordonas apposed to having to go through 14 menu levels to get to some settings.
16:58:50bluebrotherwell, you could move the "browse configurations" entry to the main menu and have the same effect.
16:59:03Lloreanbluebrother: Especially if it supported nested directories.
16:59:40amiconnJdGordon: If your favourites menu is about changing single settings, I can't see any advantage over using the stock configuration menu itself...
16:59:59*JdGordon gives up
17:00:01bluebrotheralso, how often do you really change settings? Users crying about such a thing are just playing around
17:00:07Arathiswebguest80: I think the commit was "3 Jul 00:42by Jens Arnold". not really sure though. regarding the custom view I don't have a clue
17:00:14amiconnThe big advantage of .cfg files is that they can change multiple settings at once, e.g. for switching between usage cases
17:00:36Lloreanbluebrother: And any setting you change frequently is usually between 2 values, On/Off, or long/short backlight delays, etc.
17:00:51LloreanAn "In-car" vs "at-home" vs "on-wall-power" setting
17:01:00bluebrotheryes, or you change between setting sets. Like amiconn said.
17:01:06 Part Crash91
17:01:10amiconnI have 3 .cfg files to switch between headphone, hifi and in-car use, adjusting volume/treble/bass/channel config, font, and wps
17:01:16amiconnOh, and voice settings
17:01:20bluebrotherwe should rather clean up the menus a bit.
17:01:31aliaskbluebrother: Definitely.
17:01:38 Quit billytwowilly (Remote closed the connection)
17:01:50bluebrotherIMO "general settings" could move up a level
17:02:12***Saving seen data "./dancer.seen"
17:02:16aliaskThere are some settings in LCD settings I noticed which have nothing to do with the LCD.
17:03:04 Quit Nico_P (Remote closed the connection)
17:03:12 Join Nico_P [0] (
17:03:17 Join RaRe [0] (
17:03:18bluebrotheron the h100, there are "LCD settings", and "Remote LCD" is a subitem. It should be next to "LCD settings" IMO
17:03:27 Join billytwowilly [0] (
17:03:27 Quit Thundercloud (Remote closed the connection)
17:03:35 Join Thundercloud [0] (
17:03:50Lloreanbluebrother: I agree
17:03:55aliaskLine selector, bg/fg colour, backdrop stuff belongs in display, rather than LCD settings.
17:04:13 Join Shaid [0] (
17:04:56 Join chrisjs169__ [0] (
17:05:18webguest80Thanks, I'll dig around. On another note, I was noodling with the mpeg plugin and was confused about the parameters used to make test files. I am a video engineer and I was trying to build a stack of test files, but going from the website guidelines, I couldn't make out fixed settings that were milestone targets. For instance 196X128 1.5Mbps (MPEG 1) /29.97fps Simple profilew/192kps stereo plus assorted flags
17:05:34 Quit Xerion (" ")
17:06:11Lloreanwebguest80: 1.5Mbps sounds really high for that small of a screen. And particular reason for preferring MPEG 1 over 2?
17:06:24bluebrother"clear backdrop" is in "LCD settings". Should move one up IMO
17:07:09amiconnbluebrother: It really belongs to the LCD. Targets with (non-mono) remote LCD can have 2 backdrops...
17:07:57bluebrotherthat's a point, but it still doesn't look fitting there to me.
17:09:30 Join ackbahr_ [0] (
17:09:41 Quit PaulJam (".")
17:13:05 Quit _Veseliq_ (Remote closed the connection)
17:13:17 Join Shaid` [0] (
17:13:48webguest80MPEGI was spec'd out for computer viewing/lowfi VCD's from the start. This came from the days when DVD's required hardware decoding. The codec likes making video on frame rez up to 355x288, after that MPEG II taes over, because of the extra math overhead needed to accurately render NTSC/PAL/SECAM grade video for commercial use. 1.5MBps is the absolute top for MPEG 1 and below the commercially accepable bottom
17:14:34 Quit ackbahr_ ("ChatZilla [Firefox]")
17:16:11 Quit chrisjs169_ (Read error: 110 (Connection timed out))
17:19:57 Quit webguest80 ("CGI:IRC")
17:19:59 Quit RaRe` (Success)
17:20:03 Join webguest80 [0] (i=cdae0804@gateway/web/cgi-irc/
17:20:10 Quit desowin (Read error: 113 (No route to host))
17:20:16*linuxstb doesn't like the trend of features being added to Rockbox that can't be accessed via the UI - what commercial firmware requires you to edit text files to use a DAP?
17:20:36Nico_PI know I'm late but I have to say I quite like the idea of a "shortcuts menu"
17:20:59Nico_Phaven't looked at JdGordon's patch, but I like the idea
17:22:09 Quit webguest80 (Client Quit)
17:22:21Nico_Plinuxstb: what are you referring to ?
17:22:24Nico_Picons ?
17:22:32Lloreanlinuxstb: I'd have to agree, though I can't think of features beyond Colors and Icons that hit that requirement?
17:23:02JdGordonlinuxstb: nothing in rockbox requires you to use a text editor.... its all added extras for people who want to tinker
17:23:20pixelmaswitching horizontal scrolling (the one with holding left/right) on or off...
17:23:45JdGordonis that setting still in? I thought it was removed ages ago?
17:24:01Nico_PLlorean: I don't quite get what you mean when you say it's possible to have a kind of custom menu with .cfg files
17:24:07 Join maffe [0] (
17:24:30LloreanNico_P: It's somewhat of a crazy idea. The point was that one can reproduce the rockbox menus though.
17:24:42Nico_Phow ?
17:24:42linuxstbJdGordon: That's what I mean. IMO, we should either implement things properly (meaning fully available via the UI) or not at all. Otherwise it just feels like a hack.
17:25:02LloreanNico_P: Create a folder called Volume. In it, create files 00.cfg, -01.cfg, -02.cfg, each with one line in them setting the volume to the same level as their filename.
17:25:04Nico_Plinuxstb: I agree
17:25:05JdGordonif bin size wasnt a consideration then I would agree
17:25:32Nico_PLlorean: wow indeed that's a bit crazy :)
17:25:33linuxstbLlorean: I guess there is just colours and icons (and JdGordon's proposed "customisable" menu) at the moment - but it seems to be a trend.
17:25:41 Quit ackbahr (Read error: 110 (Connection timed out))
17:26:16Lloreanlinuxstb: I think for themeing options, I don't feel they necessarily need a UI interface. Like Icons, etc, it makes sense to only have 'offline' editing as it were. Colors is very borderline though, for me.
17:26:16Nico_Plinuxstb: the colors can be edited with the texy editor, although last time I tried it wasn't a big success
17:26:52amiconnlinuxstb: It's not a trend. Colours and icons are part of themability. And then you would have to mention .wps as well
17:26:55LloreanNico_P: It's *very* crazy. But it'd reproduce a nearly visibly identical menu structure that is functionally identical to that of Rockbox. And no binsize increase. A viable alternative to the other proposition. :-P
17:27:02pixelmaJdGordon: re. the horizontal scroll is enabled by default. If you want to switch it off you can only do so via editing the .cfg - usually one isn't affected at all and won't notice.
17:27:11Lloreanamiconn: I think colors could have a menu, since it's restricted to known file extensions anyway.
17:27:15Nico_PLlorean: I still prefer the "coding way"
17:27:27LloreanNico_P: I don't see an advantage to it, at all.
17:27:51amiconnAnd the comparison with commercial firmwares doesn't count - which other commercial firmware allows you to customise the UI to a similar degree?
17:28:09Nico_PLlorean: allow the user to have his favorite options at hand without needing to remember where they are
17:28:09Lloreanamiconn: I think the fear is more related to JDGordon's patch than existing options.
17:28:39LloreanNico_P: yes, but if the menu system were recreated as .cfg files by a script (as a downloadable extra) they could simply copy the appropriate folders from that .zip, into a folder on their device.
17:28:50Nico_PLlorean: ...without telling him to write a bunch of cfg files
17:28:57LloreanNico_P: He doesn't have to write them
17:29:11LloreanHe just copies over a few folders, one per menu entry he wants.
17:29:18amiconnLlorean: How would this approach handle i18n? And voice?
17:29:35JdGordonlinuxstb: if thats your only concern ill make a gui .rock in the morning
17:29:40*amiconn prefers 'Lautstrke' over 'Volume'
17:29:42Nico_Pyeah maybe but it would be much more user friendly if there just was a list of the possible items and one just selects the wanted ones
17:29:53Lloreanamiconn: It wouldn't.
17:30:25Lloreanamiconn: My point is mainly that it's fully possible to accomplish customizable menus without a code increase.
17:30:32Nico_PLlorean: this interface could even just be a plugin that writes a text file
17:30:38 Join kaaloo [0] (
17:30:39 Quit petur ("*plop*")
17:30:53bluebrotherwhy is there a need to reorder the settings anyway?
17:30:58 Join Psilonaut [0] (
17:31:12Nico_PLlorean: point well taken, but I think the coding way offers better usability
17:31:17 Part kaaloo
17:31:55LloreanNico_P: Once the menus are created (and applicable voice files) the usability would be identical.
17:32:06Nico_Pbluebrother: it's not reordering the settings, it's having a selection of the favorite ones available immedately... exactly like the quickscreen
17:32:08LloreanIE, once a user has 'configured' the menus to their liking.
17:32:20Nico_PLlorean: what if you want to change the configuration ?
17:32:26JdGordonexcecpt voiceing wouldnt work
17:32:33Nico_Pthen to me it becomes a nightmare
17:32:40 Quit Shaid (Read error: 110 (Connection timed out))
17:32:55 Join Shaid [0] (
17:33:16LloreanJdGordon: .talk clips wouldn't work? Why not?
17:33:20Nico_Pgregj was right when he said rockbox needed to be more userfriendly and cfg are a solution but not a userfriendly one
17:33:38JdGordonLlorean: oh, you want .talk clisp for every .cfg also?
17:33:45LloreanNico_P: I'm against adding a configurable menu at all. My point is, that those few users who "need" one, can create one if necessary.
17:33:46*amiconn thinks the general consenus should stay: no user configurable menus, no user configurable buttons
17:33:53JdGordonyoud run out of disk space very quickly doing it this way :p
17:34:09LloreanJdGordon: Not really. .talk clips are quite small, and numbered options don't need them.
17:34:10bluebrotherNico_P: well, be it reordering all settings or just having some favourites, my question is still the same: why is there a need for that?
17:34:29bluebrotherimo, users looking for something like that are just playing around with settings. Nothing Rockbox is intended for ...
17:34:40JdGordonbecause having to dig through menus is a PITA... and the patch is effectivly free
17:35:29linuxstbJdGordon: I have to say I'm against the idea of a custom menu, at least until after the settings menus have been reorganised. I would hope they could be organised in such a way that the need for a custom menu disappears.
17:35:41pixelmait's free? And (I have to admit there aren't many) but if you hook it to the quickscreen, you exclude the Ondio users
17:36:04bluebrotherthe quickscreen should stay the way it is.
17:36:12JdGordonpixelma: its an extra item in the root menu... noting to do with QS at all....
17:36:20Nico_Pbluebrother: the quickscreen is good, but not everyone likes the options that ar in it... thus the need to be able to customise it
17:36:22*JdGordon is fairly sure he said that at least 3 times now
17:36:44Nico_P... or customise something similar :)
17:36:52JdGordonlinuxstb: do you honestly see that happening any time soon though?
17:37:04bluebrotherNico_P: I still don't see any need. You just can't satify everyone, and I don't think customizable menus are good.
17:37:27bluebrotherJdGordon: why not start reordering the settings right now? ;-)
17:37:38JdGordonbluebrother: so its better to annoy everyone than satisfy 80% of users?
17:37:56bluebrotherannoy anyone? By which means?
17:38:03Nico_Pbluebrother: I don't see why they are bad if done right
17:38:04JdGordon... because ive redone the code for the menus once already...
17:38:23aliaskAnd it was a huge step in the right direction, might I add.
17:38:44 Quit Shaid` (Read error: 110 (Connection timed out))
17:38:45bluebrotherwell, as far as I understood reordering the settings shouldn't be hard.
17:38:52bluebrotheror did I got something wrong there?
17:39:09amiconnThe quickscreen is only of very little use imo
17:39:15JdGordonbluebrother: your attitute is that if the users dont like the avaiable optins then they better just shut up because there is no point making life easier for those that we could...
17:39:40JdGordonand no, reordering will be a pice of cake because of the recode... just a bit time consuming
17:39:55Nico_Pamiconn: I tend to agree, and that's why I think it should be made customisable or even replaced with something better
17:40:15amiconnI can't think of anything that could replace it though
17:40:16bluebrotherwell, my point is, how many users don't like the available options? 1%? 10%? 100%? We know only about those screaming
17:40:23bluebrotherand I bet those aren't much at all.
17:40:49Nico_Pamiconn: a "shortcuts menu" with most freqently used settings
17:40:49amiconnYet another menu that just reiterates existing options (be it in fixed or user selectable order) just doesn't make sense to me
17:40:59 Join Shaid` [0] (
17:41:06Nico_Pamiconn: that's what the quickscreen is
17:41:19amiconnNico_P: I have *no* most frequently used options I could put there...
17:41:22JdGordonso dont use it then.... noone has actually said a _real_ reason against it...
17:41:36LloreanJdGordon: Being 100% pointless isn't a reason against it?
17:41:46*JdGordon doesnt think its pointless at all
17:41:52amiconnit is
17:41:59bluebrothercustomizable menus are a support nightmare. Be it some favorites menu, the whole menus or the quick screen
17:42:01Nico_PLlorean: if it's not pointless to everybody, no
17:42:06LloreanIgnoring the crazy solution, normal .cfg files can cover almost every usage case.
17:42:11JdGordonso is keeping support for the archos player...
17:42:16JdGordonbut thats another argument....
17:42:37amiconnIf you have to go to the menu anyway, you can as well use the existing option. I can't really see the difference
17:42:41LloreanIf you want favorite settings, go with favorite configurations rather than some cluttered menu.
17:43:33chrisjs169__What's the chance of the Sansa Express being Rockbox'd?
17:43:33amiconn.cfg files, otoh, *are* a very useful measure to speed things up if you want to set multiple options at once
17:43:42JdGordonI just cant see how a cluttered file tree is nicer than a menu...
17:44:02LloreanNico_P: 100% pointless in that "it entirely replicates existing functionality without allowing the user to do anything new"
17:44:04JdGordonand I agree there..... but thats not what the patc is about...
17:44:04linuxstbJdGordon: I think it's the wrong solution to the problem - the problem is that settings are hard to access, so we can try to solve that by reorganising the structure of the menus. I also agree that .cfg files are a nice existing solution to quickly changing settings, and they should be encouraged more - possibly by writing a plugin to help manage them.
17:44:13JdGordonits about changing one setting at a time
17:44:19LloreanJdGordon: I can't see how it's *worse* than one. I think they're equal, but the filetree doesn't bloat bin size.
17:44:32*amiconn doesn't think the options are hard to access
17:44:46LloreanWhy not table this discussion entirely until the menus are optimized?
17:44:49amiconn*Some* of them need a little reordering, that's all
17:44:56*bluebrother agrees with amiconn
17:45:02aliaskThe only settings I could possibly think about changing often enough to warrant them being duplicated in another menu would be shuffle and repeat, but we already have them in the QS.
17:46:05amiconnWhat would be useful would be a context menu (yes!) for settings, allowing to send their value (and *only* this one) to a selectable .cfg file, in order to create such multi-settings .cfg files on target
17:46:19Lloreanamiconn: I agree
17:46:35*JdGordon gasps
17:46:37LloreanVery much like the add to favorites playlist option.
17:46:43amiconnSo you could "click together" a .cfg file
17:47:08linuxstbThe original feature request (FS #7417) was because the user thought that settings were hard to access and wanted a shallower hierarchy. My point is that we can do that for everyone, without the need for customised menus.
17:47:15 Join Wiwie [0] (
17:47:40amiconnBut the problem how to remove a setting from a .cfg needs to be solved as well
17:47:44*JdGordon passes a big cup of "open mindness" to the crowd
17:47:52JdGordonthe patch is about chaning a single setting
17:48:00JdGordonnot multple which obviously are better done in a cfg
17:48:10LloreanJdGordon: Out of curiosity, are we allowed to call you close-minded for disagreeing with us too?
17:48:19JdGordonof course
17:48:31amiconnYes, sand for single settings I can't see any action require apart from reordering settings a bit
17:48:33 Join crop [0] (i=c27f0812@gateway/web/cgi-irc/
17:49:00cropHow do I put an image to a wiki page?
17:49:08bluebrotherwell, like "add this value to cfg" there could be a "delete this value from cfg"
17:49:24JdGordonamiconn: you of course reaslise that the only actual reason against the patch is bin size, and a patch to do your suggestion would be much larger.....
17:49:44LloreanJdGordon: His patch would add *new* functionality.
17:50:00bluebrotherJdGordon: that would do it the "right" way :P
17:50:01LloreanRight now the only way to create a single line .cfg file requires text_editor, something that may not be present.
17:50:01linuxstbJdGordon: I agree that fast access to a particular setting is different to .cfg files, I'm just saying I would prefer fast access to _all_ settings, not just some I've put in a custom menu.
17:50:05JdGordonno it wouldnt... it would add exactly the same about of new features
17:50:11LloreanJdGordon: No.
17:50:21LloreanJdGordon: Tell me, without any plugins, how do you create a one-line cfg file in Rockbox?
17:50:32LloreanJdGordon: Then tell me, without any plugins, how do you change one setting? Which can be done without code changes.
17:50:33JdGordoni dont
17:51:12JdGordon... that was of course to the first one
17:51:26LloreanHis suggestion adds something new you can do. Yours adds a different way of accessing existing functions.
17:51:56LloreanOtherwise any new menu is a new function, independent of what it contains.
17:52:34LloreanI think the first thing that needs to be done, as linuxstb keeps suggesting, is work on the existing menu layout which I think everyone agrees is at some degree of sub-par
17:55:38 Quit merbzt ("Leaving.")
17:56:22*JdGordon hasnt conceded yet.... but going to bed
17:56:30 Quit JdGordon ("Konversation terminated!")
17:58:20 Quit lee-qid ("aufwiederbyebientotsayonara")
17:59:02 Quit Shaid (Read error: 110 (Connection timed out))
17:59:03 Nick Shaid` is now known as Shaid (
18:00:46amiconnLlorean: In fact you can create single line .cfg files without plugins... just it's incredibly cumbersome
18:02:43amiconn(1) Save your current config (2) Reset all settings (3) Change the one setting you want in the .cfg (4) Shutdown and reboot (5) rename config.cfg to your new .cfg's name (6) reload the saved settings
18:02:56 Join webguest21 [0] (i=186e1ca3@gateway/web/cgi-irc/
18:02:57Lloreanamiconn: Ah, yes, I'd realized that as soon as you said it could be done
18:03:21amiconnBut this way you can't put the default into a .cfg
18:03:40webguest21Hello people. Is it normal for the Gigabeat to lose battery power even if its not being used?
18:03:48markunwebguest21: a little
18:03:57 Quit aliask ("ChatZilla [Firefox]")
18:04:06markunwebguest21: it uses 0.71mA while 'off'
18:04:19markun(on a 830mAh battery)
18:04:57webguest21probably from the resume playback I guess
18:05:27markun(so should last almost 49 days)
18:06:35webguest21should I have certain settings OFF to minimize the loss of battery power
18:06:48markunno, it's always 0.71
18:07:21markunbut if you don't use your gigabeat for a long time it's better to switch the little battery switch to OFF
18:07:53*amiconn wonders what the gigabeat uses that much power for while being off
18:08:05amiconnThe rtc shouldn't draw more than a few A
18:08:15markunamiconn: indeed, quite strange
18:08:24webguest21I am surprised at the battery bench results because with normal use I am getting about 8 or 9 hours
18:08:34markunperhaps we can make it lower, but it's the same value if you sleep from the OF
18:09:11markunwebguest21: these benchmarks are very artificial
18:09:16Lloreanwebguest21: And how much do you get under the original firmware?
18:09:25markunif you use the backlight and hdd a bit it will already drop a lot
18:10:06linuxstbmarkun: Presumably you can turn the battery switch off to preserve power?
18:10:18markunlinuxstb: yes
18:10:26markunbut the clock will be reset of course
18:10:36linuxstbMine always is anyway...
18:10:47*linuxstb has a watch ;)
18:10:51webguest21don't know. did not keep it long enough. pretty much switched to Rockbox right away
18:11:50webguest21the original firmware was too crappy
18:12:13markunwebguest21: I agree :)
18:13:11 Quit XavierGr (Nick collision from services.)
18:13:14 Join XavierGr [0] (
18:13:28 Quit Nico_P (Remote closed the connection)
18:14:26webguest21well how much are you guys getting with normal use in regards to battery hrs.
18:15:02markunI've never played until the battery was completely drained
18:15:23markunI guess I usually recharge after 6-8 hours of listening
18:16:19webguest21Is the WMA codec 100% yet or does there need more cleaning up?
18:16:56 Join Domonoky_ [0] (
18:16:58linuxstbIn terms of functionality, it still can't decode all WMA files (e.g. low bitrates, wma v1), and there's no seeking.
18:17:38linuxstbAnd then there is hopefully still a lot more optimistation potential - I doubt that very many codecs could be called fully-optimised in Rockbox.
18:18:07 Join iPodPhoto [0] (i=46311c65@gateway/web/cgi-irc/
18:18:10linuxstb(more optimisation means longer battery life)
18:19:06iPodPhotoHello. Is there someone here with twiki access
18:19:24linuxstbLots of people - do you want write rights?
18:19:51webguest21oh ok, great accomplishment even without the seeking
18:20:19iPodPhotoYes i would. I have been trying to get wrtie rights for almost 2 months now. But no one is ever on the IRC when i ask.
18:20:33linuxstbWhat's your wiki name?
18:20:52iPodPhotoMark Sikora
18:21:01webguest21so WMA at this point could use more battery then because of its infancy?
18:21:32 Quit XavierGr (Nick collision from services.)
18:21:35 Join XavierGr [0] (
18:22:10linuxstbiPodPhoto: OK, you're added.
18:22:23 Join My_Sic [0] (
18:22:43iPodPhotook. thank you kindly
18:23:06linuxstbwebguest21: It depends on the device you're using, and what other codecs you're comparing to.
18:23:26 Quit My_Sic (Client Quit)
18:25:58 Quit crop ("CGI:IRC")
18:25:59 Quit iPodPhoto ("CGI:IRC (EOF)")
18:26:05 Quit webguest21 ("CGI:IRC (Ping timeout)")
18:31:37 Join BigMac [0] (
18:31:51 Join obo [0] (
18:31:59BigMacHey any *nix users around?
18:32:32Domonoky_BigMac: only rockbox users here..OS doesnt matter :-)
18:33:41*bluebrother uses linux ... so no match on *nix
18:34:25BigMacbluebrother: I was using *nix to refer to any flavor of linux, do you know how to export something to your path?
18:34:42bluebrotherhehe, know what you meant ;-)
18:34:43BigMacIsn't it export PATH="path"
18:35:04bluebrotheryou usually do "export PATH=$PATH:/the/path/you/want/to/add
18:35:17bluebrotherwell, on bash at least.
18:35:25BigMacalright lemme give it a go
18:35:46BigMacBut wait, is that the temporary way?
18:35:55BigMacAs in I would have to do that every reboot?
18:35:56bluebrotherit exports in your current session
18:36:13bluebrotheryou want to add something permanently?
18:36:39Domonoky_BigMac: then add this command to your profile script..
18:36:48bluebrotherwell, that's in ~/.bash_profile on my box
18:37:00BigMacOk, I will look for it
18:37:21bluebrotherbut you could also add it to /etc/bashrc to have it globally for all users
18:38:03 Join Megaf [0] (
18:38:08Megafhello all
18:38:12Megafremember me?
18:38:18BigMacbluebrother: bash.bashrc right?
18:38:25Megafim just instaled rock box on my iPod Nano
18:38:28Megafvery nice
18:38:36bluebrotherI don't have a bash.bashrc no my box
18:39:24BigMacWhat distro?
18:39:42linuxstbBigMac: IIRC, it should be added to $HOME/.bashrc
18:40:01linuxstb.bash_profile is for login shells
18:40:03Megafcan rockbox play videos?
18:40:18BigMaclinuxstb: alright, I will go look for it
18:40:25linuxstbMegaf: Yes - see the manual and/or the PluginMpegplayer wiki page.
18:40:40 Join petur [0] (n=petur@rockbox/developer/petur)
18:40:59 Part pixelma
18:41:17 Join relaxed [0] (n=relaxed@unaffiliated/relaxed)
18:41:22bluebrotherwell, my box exports PATH from within ~/.bash_profile
18:41:38 Quit XavierGr (Nick collision from services.)
18:41:41 Join XavierGr [0] (
18:41:49bluebrotherbut I never completely understood (or tried to understand) the exact differences between those various bash configuration files
18:42:26BigMacWell I probably should restart x right?
18:42:29linuxstbOn Debian at least, the default .bash_profile runs .bashrc.
18:42:38linuxstbNo, you should just need to open a new terminal.
18:42:53linuxstbThat's the advantage of using .bashrc instead of .bash_profile
18:43:11LloreanI've added to .bashrc as well.
18:43:25LloreanBut then I think my login shell is dash, so I don't believe .bash_profile gets executed.
18:43:37*Domonoky_ thinks .bash_profile is valid for the whole session, from login to logout, .bashrc is only valid as long as the shell exists..
18:43:47bluebrotherBigMac: no, you only need to restart the shell so it rereads its configuration
18:44:10bluebrotheryou can just run bash from the command line ;-)
18:45:20BigMacNah, I am one of those crazy characters who needs a gui
18:46:20Megaflinuxstb: ok, thanks
18:46:33*linuxstb wonders what the ASF_FLAG_SEEKABLE flag actually means in an asf file - most files seem to have it set to 0...
18:49:22dionoeathe SEEKABLE and BROADCAST flags aren't used anywhere in the code in VLC (they're just defined)
18:50:00linuxstbI've just noticed that mplayer can seek in asf files with that flag unset, and even adding an index (with asfbin) to a file doesn't set the flag.
18:50:06linuxstbSo it seems completely useless...
18:51:23 Join AceNik [0] (n=AceNik@
18:52:20 Join kaaloo [0] (
18:52:25AceNikguys thr is a new problem in my H10 suddenly while playing music for sometime it freezes
18:52:31 Join dan_a [0] (n=dan_a@
18:52:43 Part kaaloo
18:52:56AceNikno respond to any button, no lcd nothin
18:52:58 Quit petur ("switch")
18:53:07 Join petur [0] (n=petur@rockbox/developer/petur)
18:53:29 Quit Wiwie ("Miranda IM! Smaller, Faster, Easier.")
18:55:28amiconnlinuxstb: Regarding codecs which could be called really optimised in rockbox, I can only think of 2: libmad (on coldfire), and flac (all swcodec)
18:55:55 Quit XavierGr (Nick collision from services.)
18:55:58 Join XavierGr [0] (
18:56:14amiconnAnd of course the uncompressed ones
18:56:37 Quit SirFunk (Remote closed the connection)
18:57:36amiconnHmm, seems there are a bit more...
18:57:52*amiconn only really uses mp2/mp3 and flac so far
18:57:55 Join Davide-NYC [0] (
19:00:25Davide-NYCrecording menu patch... skimmed the IRC logs and still don't get how it works. How does it work?
19:02:05 Quit relaxed ("")
19:02:14***Saving seen data "./dancer.seen"
19:04:18amiconndan_a: ping...
19:04:44dan_aamiconn: pong
19:04:50amiconnah :)
19:05:16dan_aHow did the clock investigations go?
19:05:43GodEatermy replacement ipod battery arrived. My god is it small!
19:05:50amiconndan_a: COuld you perhaps run some tests on your G3 in order to confirm some details?
19:06:02AceNikguys any solution to H10 suddenly closing down
19:06:54dan_aamiconn: Of course - except that bootloaders I compile don't work for some reason
19:06:55amiconnIt looks like the PP5002 clock setting works similar to PP502x, only that the settings are scattered across the various registers in a different way (and there are fewer settings)
19:07:34amiconndan_a: You don't need to compile bootloaders (that is, if the current bootloader doesn't touch the clock setting registers)
19:08:02amiconnI have a patch; I could either provide this or a ready-to-run test build
19:08:13 Quit BigMac (Read error: 104 (Connection reset by peer))
19:08:19dan_aA patch will be fine
19:08:28amiconnThe patch adds a couple of readouts to the debug menu
19:08:38amiconnThere is a clock estimation that needs to be calibrated
19:09:31amiconnThe first test build would be only with this patch applied, and then checking what the estimation displays at 24, 30 and 75 MHz, and adjust a factor that it matches approximately
19:09:44Davide-NYCI now understand the recording patch. Is there a way to set input dependant directories in a cfg file?
19:09:59 Quit AceNik ("bye guys have fun , enjoy !!!!!")
19:10:20 Quit billytwowilly (Remote closed the connection)
19:10:27dan_aamiconn: That all sounds simple enough
19:10:35 Join billytwowilly [0] (
19:10:41amiconnThe next step involves disabling both cpu scaling (in the config-* file) and the complete content of ipod_set_cpu_speed() (so rockbox doesn't touch the registers) and then read a number of values from the debug menu
19:11:20amiconnThe goal is to find out what the apple loader sets, and what the cpu speed resulting from these settings is (hence the calibration)">
19:15:03dan_aOK, I'll let you know when I'm done or need some help
19:24:15 Quit ptw419 ()
19:29:28rogeliohi, what software does rockbox use to do automated bulds? it looks great and I want to do something similar
19:29:52linuxstbCustom scripts written in a combination of perl and bash afaik.
19:29:56GodEaterrogelio: a combination of bash scripts and perl
19:30:03GodEaterlinuxstb: hehe - jinx
19:30:33 Quit richcoosa19_ (Remote closed the connection)
19:31:08rogelioOh, ok, I thought you were using something like buildbot. Are they accesible somewhere?
19:31:29GodEaterrogelio: I think you'd need to ask Bager or Zagor for them
19:31:40GodEaterneither of whom appear to be here
19:32:16 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
19:32:24GodEaterthey're not fool proof either - we sometimes get stuck builds
19:32:50rogelioGodEater_: ok, thanks, I'll look for them
19:33:06*petur thought they were copied in svn
19:33:15GodEaterpetur: really ? where ?
19:33:55 Quit HellDragon (Nick collision from services.)
19:34:41 Join jgarvey [0] (
19:34:57peturGodEater: ?
19:37:18rogelioI've found some info here: :-)
19:37:40GodEaterrogelio: that's how to become a client of the master build server though
19:37:55GodEaterwhich doesn't contain the scripts which triggers the autobuild
19:39:39 Join Wiwie [0] (
19:43:12 Join HellDragon [0] (i=JD@unaffiliated/helldragon)
19:43:26 Join low_light [0] (i=c730190b@gateway/web/cgi-irc/
19:47:33dan_aamiconn: The Apple loader sets the clock to 48MHz. Which registers do we need to check?
19:48:52amiconnWhat's the adjusted count for PP5002 in perfcheck() ?
19:49:09amiconnI want to know all values below the gpio values
19:51:23 Join ptw419 [0] (
19:51:52dan_aThe adjusted value from perfcheck is 48MHz. CLOCK_ENABLE is 0x2, CLOCK_SOURCE 0xe4, CLOCK_DIV 0xE000, PLL_DIV 0x18, PLL_MULT 0x4b, TIMING1_CTL 0xa0ff, TIMING2_CTL 0x80ff
19:53:55amiconnI mean what is needed to make the estimation work (instead of the 10000 with the /* TODO... */ behind)?
19:55:31dan_aIt was out by a factor of 3.2, so I guess 32000
19:55:37andrew__2does anyone know if I2C interrupts work on the sansa?
19:55:58andrew__2I am thinking of implementing the headphone detection in the AS3514
19:56:06 Nick andrew__2 is now known as andrewg867 (
19:56:41amiconndan_a: Wow... that's slow compared to PP502x...
19:57:37dan_aYes, even taking account of the broken cache
19:57:53amiconnThe pll is prepared for 75MHz... did you make ipod_set_cpu_speed() an empty function for this test?
19:58:00 Join lee-qid [0] (
19:58:10dan_aI did
19:58:22amiconnHmm, interesting...
19:59:45amiconnAnd it runs at 48MHz?
20:00:14dan_aAccording to perfcheck(), yes
20:00:23*amiconn would want a G3 for experimentation... this is becoming interesting
20:01:17LloreanWasn't someone in the forums giving one away? Who ended up with that one?
20:01:20LloreanOr was it on the list?
20:01:30amiconnjhMikeS I think
20:01:37amiconnBut he has no ipod cable...
20:01:55LloreanAh, problematic.
20:02:23*amiconn wonders whether G3's are available from ebay(.de), and at what price
20:02:30*amiconn should check
20:02:39*amiconn should also check for H10s
20:03:10dan_aBroken ones are frequently on ebay .uk
20:03:19 Quit Wiwie ("Miranda IM! Smaller, Faster, Easier.")
20:04:28amiconndan_a: Some further tests you could do: reenable ipod_set_cpu_speed(), and change the value written to 0xcf005010 from 0xe000 to 0xe011, and in another step to 0xe022, and check the est. clock
20:04:44Davide-NYCI gave jhMikeS the 3G, I presumed someone would pony up for the cable. (grumble)
20:05:06amiconnBefore doing that, you could place the 32000 in debug_menu.c line 432 and check whether the estimation is correct
20:05:16*amiconn has an ipod cable...
20:05:26Davide-NYCproblem solved
20:05:28amiconnAssuming the cable is the same for *all* ipods
20:05:48Davide-NYCno, the 3G requires the Firewire cable
20:05:50 Quit chrisjs169__ (Read error: 110 (Connection timed out))
20:05:52amiconnNah, I mean I could test on a G3 without its own cable
20:05:55Davide-NYCno idea why
20:06:01amiconnAh, but the same plug?
20:06:35amiconnI also have a fw cable (found out fw detection from that on mini G1, G2 and video (power only))
20:06:50amiconnDidn't implement it yet though :\
20:07:24Davide-NYCQQ: can I implement input dependant recording directories using CFG files, and if so where would I find docs?
20:07:24amiconnipods started firewire only. At some point, apple added usb support. The newest models only support usb
20:08:20amiconndan_a: The two least hex digits in 0xcf005010 are pll post-dividers iiuc
20:08:27Davide-NYCgotta run. I'll check the logs for an answer. (please)
20:08:30LloreanDavide-NYC: Do you mean "change recording directories by loading .cfg files"?
20:08:31 Quit Davide-NYC ("ChatZilla [Firefox]")
20:08:43amiconnUnlike on PP5020+, they seem to work without an offset of 1, but 0 acts like 1
20:09:05amiconnSo, 0xe011 should give 75MHz like 0xe000, but 0xe022 should give 37.4MHz
20:09:10amiconn*37.5 of course
20:09:24amiconnOtherwise I have to correct my theory
20:09:35saratogaI've been meaning to ask why we still support the 3G givent that the port is basically nonfunctional and no one seems interested in working on it
20:09:51 Join DerPapst [0] (
20:10:09dan_asaratoga: Define nonfunctional - it plays MP3s
20:10:10 Join Siltaar [0] (
20:10:12amiconnsaratoga: dan_a has a G3 and it's already a mostly working port (plays music), so we shouldn't drop it imo
20:10:19low_lightamiconn: I've noticed a couple of things olympus m:robe bootloader (another pp5020 device)...
20:10:30amiconnWe never dropped a target that reached the stage of playback yet
20:10:34DerPapstat least 2 are interested in that port ;)
20:10:37saratogadan_a: when I tried it a while back on my 3G, it played for a couple minutes at most before crashing
20:10:42low_lightwrites to 0x60006020 are always followed by outl(0x4800000F, 0x60007000)
20:10:50saratogathough maybe thats been fixed
20:10:51*amiconn would even like to get ipod G2 and G1 running
20:11:13DerPapstamiconn: you can copy data to the 3G iPod via firewire and usb. but you can charge it only via firewire
20:11:35amiconnlow_light: If I knew what that actually does... :/ I found other values written to 0x60007000 (and 0x6007004):
20:12:01amiconn0x4800001f, 0x420000ff, 0x4800000f, 0x42000000, 0x4100000a
20:12:05low_lightand the value written to 0x70000034 for freq change is (R0 | (R0 << 8)) | (inl(0x70000034) & 0xFFFFF0F0)
20:12:08 Join Wiwie [0] (
20:12:08dan_aamiconn: I have a G1 or G2 here (not sure which) - I don't know if it works at all, but you're welcome to it if you want
20:12:17amiconn0x80000000 and 0x0 are known - sleep/wakeup core
20:12:45DerPapstdan_a: the wheel of the 2G is touch sensible
20:12:51amiconnlow_light: I am almost sure what 0x70000034 does, and we already handle that
20:13:14DerPapstdan_a: nvm ;)
20:13:31dan_aDerPapst: Then it's a 1G
20:13:50amiconndan_a: I'm not interested in non-working units. YOu never know what's required to make it work again when it doesn't react at all, and you don't have an identical unit to check against
20:13:52DerPapstah ok.. i thought you were talking about the minis.
20:14:05DerPapstand i wanted to correct myself already ^^
20:14:49amiconnThings like broken hdd are often obvious, but anything else - hmm.....
20:14:54dan_aamiconn: I'll buy a firewire cable and get it charged to see if it will boot, then
20:15:32*DerPapst has rockbox on his 3G :)
20:15:36andrewg867does anyone know if charging the sansa from USB works in rockbox?
20:16:21*DerPapst hunts something to eat.
20:16:21amiconnlow_light: Oh, btw, I found some sequences like you described in the apple OF as well. But the main clock change routine doesn't do that
20:16:25DerPapstsee you
20:16:26 Quit DerPapst ("So Long And Thanks For All The Fish!")
20:17:15amiconnThe 0x420000xx values seem to be some kind of preparation for deep sleep (32kHz clock), but I don't have an idea what the lower 8 bits do...
20:18:35amiconnIf that chinese site is correct, the G1 runs PP5000, and the G2 runs PP5001. But somehow I don't believe that...
20:19:03dan_aI understood that G1-G3 used PP5002
20:19:19amiconnYes, that's what the ipl site says
20:20:19low_lightwell, the bootloader doesn't actually change cpu freq as far as I can tell, but the routine appears to be there
20:22:07 Quit atsea- (Remote closed the connection)
20:28:25 Quit Megaf ("Ex-Chat")
20:29:25 Join atsea- [0] (i=ariel@gateway/tor/x-d69c8b7f9a28bfff)
20:31:39 Join Xerion [0] (
20:33:16dan_aamiconn: Writing 0xe011 to 0xcf005010 gives 66MHz. 0xe022 gives 33MHz
20:33:44amiconnAnd 0xe000 gives 75?
20:35:13dan_aI'll check...
20:35:13 Join entheh [0] (
20:35:38amiconnSeems I have to re-check the readout routine...
20:36:37 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
20:37:25dan_aamiconn: 0xe000 gives 132MHz
20:37:42amiconnLooks like your calibration isn't correct then
20:38:04amiconnBut the divider works different from what I though. It works like on PP5020+
20:38:24 Join mirak [0] (
20:40:13*amiconn calculates that 18182 would be the correct value
20:40:50amiconnThat would also make a lot more sense in comparison to PP5020+
20:41:11amiconndan_a: Is the 48MHz estimation based on the same correction factor?
20:45:02 Join chrisjs169_ [0] (
20:47:51amiconndan_a: I think the loader sets 24MHz
20:49:26 Quit miepchen^schlaf ("Verlassend")
20:49:38 Quit atsea- (Read error: 104 (Connection reset by peer))
20:49:42 Join miepchen^schlaf [0] (
20:56:10 Join SirFunk [0] (
20:56:10 Quit jgarvey ("Leaving")
20:56:26 Join jgarvey [0] (
20:56:46 Join chrisjs169__ [0] (
21:01:46 Join richcoosa19_ [0] (n=richie@
21:02:17***Saving seen data "./dancer.seen"
21:02:23gregjwhat's up folks
21:04:51 Quit chrisjs169_ (Read error: 110 (Connection timed out))
21:08:37 Quit low_light ("CGI:IRC (Ping timeout)")
21:10:08 Join linuxstb_ [0] (i=5343d4aa@gateway/web/cgi-irc/
21:15:55linuxstb_gregj: Hi. Your wmv file (I think it was your's - solo_as_i_am.wmv) now plays in Rockbox.
21:16:47*amiconn needs to find an iram function in the g3 firmware :(
21:17:54 Join bdgraue [0] (
21:18:20gregjlinuxstb_: cool, I will certainly try it - thanks :)
21:19:29gregjisn't it funny, if you put in google
21:19:45gregjfirst few answers will be direct opposite
21:20:04gregjI guess chanining that to 'ask her out' would be better ;)
21:20:33amiconnlinuxstb: Do you agree that uses too many different bitrates for the lossy codecs?
21:22:03linuxstb_amiconn: Yes, I already mentioned that to the people doing those tests. I think we agree that something like a "typical" and "maximum" is all that's really needed.
21:22:10linuxstb_(we = you and I)
21:22:57dan_aamiconn: With your calibration, when the CPU frequency menu says it is running at 30MHz, perfcheck shows 17MHz. When the CPU frequency menu says 78MHz, perfcheck says 44.21
21:23:33amiconnhmm, but how does that match your 132MHz reading?
21:24:37amiconnohh, let me check something...
21:24:39 Join Bagder [0] (n=daniel@rockbox/developer/bagder)
21:25:37amiconnOoops, seems I cut the clock on PP5002 in half when doing my PP502x changes
21:25:52amiconnOnly if cpu scaling is enabled, that is
21:26:08dan_aOoops! That will explain why the results were weird!
21:26:50linuxstb_dan_a: How stable is Rockbox on your 3G? A few people have come to IRC over the last few months reporting frequent crashes (whilst you weren't around).
21:27:00 Join low_light [0] (i=c730190a@gateway/web/cgi-irc/
21:27:31 Quit billytwowilly (Remote closed the connection)
21:27:36amiconndan_a: could you change line 104 outl(8, 0xcf005018); to
21:27:40amiconnoutl(4, 0xcf005018);
21:27:43amiconnand re-check?
21:27:46dan_alinuxstb_: I've not been using it recently, but it had started crashing - it seemed to be a CPU scaling problem
21:27:48 Join billytwowilly [0] (
21:28:04linuxstb_Ah so CPU scaling was enabled?
21:28:17amiconnI changed the pll divider from 8 to 4 on PP502x, and this is the matching change for PP5002
21:29:12dan_alinuxstb_: IIRC, that was my first commit
21:29:17 Join atsea- [0] (i=ariel@gateway/tor/x-f6f9442b8b91dd20)
21:29:30amiconnThat should make it possible to match the estimations with scaling enabled vs. disabled
21:29:56amiconnlinuxstb: I guess its the timing setup issue we also had on PP502x
21:30:33 Quit HellDragon (Client Quit)
21:30:36amiconnI do not fully understand the clock setup yet, as there are many special cases. Sometimes the code also seems to repeat stuff
21:31:03*amiconn should do the test with disabled cache first
21:32:32 Join HellDragon [0] (n=JD@unaffiliated/helldragon)
21:39:05 Join Buschel [0] (
21:47:10Buschelhi there, does anybody know how much CPU (in MHz) is consumed on an iPOD when rockbox is idle? from my measurements i can see that rockbox needs decoding-core + 12 MHz while playback. i am interested in possible reduction of the residual 12 MHz
21:47:22dan_aamiconn: With updated calibration, 0xe000 gives 66MHz
21:48:23 Quit petur ("reboot")
21:50:23dan_a0xe011 gives 33MHz
21:50:41amiconnYea, so the divider divides by (n+1)
21:50:42 Join petur [0] (n=petur@rockbox/developer/petur)
21:50:56amiconnBut the estimation is still not 100% correct
21:51:07 Join Robin0800 [0] (
21:51:30amiconnI assume you're running with the changed pll divider, scaling enabled, and cpu set to 78MHz?
21:52:01amiconn..and calibration value == 18182?
21:52:31 Quit richcoosa19_ (Remote closed the connection)
21:52:43linuxstb_Buschel: What are you measuring? I 'm not sure I understand your question - when Rockbox is idle, by definition it's consuming no CPU.
21:52:46amiconnBtw, the divider is evaluated the wrong way in the OF...
21:54:06dan_aI'm using calibration = 16000, which gives 24MHz and 30MHz at those frequencies, but 66MHz when the CPU Frequencies page says 78MHz
21:55:03amiconnNow that is interesting...
21:55:23Buschellinux: ok, i'll try explain better :) lets say the decoder-core for some format needs 30 MHz for realtime decoding and the CPU needs 42 MHz for playback. then there is 12 MHz left which is consumed by all other activities (which include dsp and i2c to DAC)
21:55:30dan_a(with the divider at 8 and the calibration at 32000 I got the correct values consistently)
21:55:47amiconnYeah, but that is *not* correct
21:56:04amiconnSo there seems to be a 66MHz limiter in the chip ?!?
21:56:18Buschellinux: 12 MHz is quite a lot when comparing it to all the stuff the decoder does with 40 MHz
21:56:42amiconnThe PLL base clock is 24MHz and the multipliers are 4, 5 and 13 now, giving 24, 30 and 78MHz
21:56:47dan_aNo, I understand that it's not correct - so either there is a limiter or the divider isn't simply a divider, but controls other things too
21:58:27linuxstb_Buschel: How are you measuring those frequencies?
21:59:48amiconndan_a: That sounds unlikely
22:00:14enycIpod mini 2nd generation... seems to have real battery life problems using rockbox just for audio playing.....
22:00:31enyc(playing mostly 128kbps mp3 and 128mpbs aac)
22:00:38enycwithout useng scopes and suchlike
22:00:43enyc... is this normal?
22:00:58 Quit Bagder ("*plopp*")
22:01:08bagawkenyc: yea, my video only gets about 9-10 hours
22:01:29enycbagawk: hrrm ipod mini 2g ... 1 hour life apparently ;-)
22:01:30bagawkenyc: it's the big reason I still use the apple software
22:01:36amiconnBattery life with rockbox on ipods is less that original firmware
22:01:43enycbagawk: used to do more before esing rockbox
22:01:49amiconnHowever, mini g2 is one of the better targets
22:01:55enycamiconn: ok.. and what sort of 'ratio' is normal?
22:02:04 Join hannesd_ [0] (
22:02:05amiconnYou should get 6..8 hours on that
22:02:16enycmaybe this has a buggy version on there? waht about hard drive idle not working?
22:02:18*amiconn has a mini g2 himself
22:02:30enycamiconn: maybe the batt is dying?
22:02:33Buschellinux: 78 MHz CPU / 2,6x realtime (test_codec) = 40 MHz, 25% boost-ratio for playback = 0,75*30 MHz + 0,25*78 MHz = 42 MHz
22:02:45linuxstb_enyc: How long do you get with the Apple firmware?
22:02:46amiconnDoing a runtime test (playing ~200kbps mp3, but no backlight) recently yielded 8:40
22:02:47enycamiconn: should vorbis work well on the mini 2g?
22:02:57enyclinuxstb_: i'd need to ask.. its not actually mine ;-)
22:03:28enycamiconn: are the backlight/hdd timeouts configurable?
22:04:43*enyc waits patiently ;-)
22:05:04linuxstb_Yes, they are, but the defaults are very short (4s for backlight and 5s for disk iirc)
22:05:26enyclinuxstb_: right... so it SHOULD be spinning down the disk when just playing files... ok..
22:05:52amiconnDuring playback it always spins down fast
22:05:56linuxstb_Yes. the disk spins down, but Rockbox doesn't fully power them off on the ipods (we don't know how).
22:06:09enyclinuxstb_: i understand
22:06:45enychave to do a playback test etc. on both firmwares
22:06:59enycmay well be worth installing a newer rockbox... maybe there si a bug in that version etc.
22:07:18linuxstb_Yes, that's the only way to compare. Choose a long MP3 album, and play it on repeat in both firmwares until the battery dies.
22:07:33linuxstb_(long means larger than the available RAM - 32MB)
22:07:59Buschellinux: ahem, 78 / 2.6 = 30 of course :/
22:08:09enycalso... if the battery is dying (not unusual on ipod units) then this sort of low bottery life is usual ;-)
22:08:23 Join Bagder [0] (n=daniel@rockbox/developer/bagder)
22:08:37enycill try to get that tested ;-)
22:08:42 Quit perplexity (Read error: 113 (No route to host))
22:08:46 Quit XavierGr (Nick collision from services.)
22:08:49 Join XavierGr [0] (
22:09:39linuxstb_Buschel: I'm not sure how accurate the boost ratio is - e.g. boost happens when the disk is spinning, regardless of CPU load (to keep disk spinup time to a minimum).
22:10:12linuxstb_Your measurements also don't take account of how much time the CPU is sleeping.
22:10:19Buschellinux: boost ratio was measured after several minutes of playback for several files/codecs
22:10:27*amiconn wonders what Buschel is trying to achieve
22:11:44Buschelamiconn: i try to understand where the cpu-time is consumed -> for optimization
22:12:32amiconnThere is such a load of factors that it's probably not possible to find out during normal playback
22:12:47Buschelamiconn: lower cpu-time may decrease power consumption. further optimization of e.g. mpc-decoder will not have that large impact, if lots of cpu-time spent on other tasks
22:12:48amiconnI think you need some kind of test_dsp plugin for that
22:13:15Buschelamiconn: good point
22:13:35amiconnThere's cpu time consumed for all kinds of things, even lcd updates in the audio debug screen take a measurable amount
22:14:07Buschelamiconn: you do not know how much by accident?
22:14:15amiconnWhen playing the same track on H1x0 and H300 and watching boost ratio in audio debug, the boost ratio on H300 is higher
22:14:41amiconn...because of its colour screen, even when neutralizing other factors (like the adc read)
22:14:51amiconnBoth adc and lcd timing cannot be optimised further
22:15:08 Join Soul-Slayer [0] (n=Administ@
22:15:13amiconnAnd on ipod, we don't even know how to change the lcd interface timing...
22:15:14 Join BigMac [0] (
22:15:53Buschelbut do we know how much cpu-time such tasks (lcd-update, dsp, dac) consume?
22:16:08Buschelsomething like profiling results?
22:16:39 Quit hannesd (Read error: 110 (Connection timed out))
22:16:40linuxstb_There's a test_fps plugin which measures lcd update speed.
22:17:28Buschelis this contained in the standard package or do i need to enable it before compiling?
22:17:40 Nick hannesd_ is now known as hannesd (
22:17:49linuxstb_You need to enable it - similar to test_codec
22:19:31Soul-SlayerDoes anyone elses gigabeat sometimes refuse to turn on unless they turn the battery off and on again, or am I alone there?
22:21:26linuxstb_I sometimes get the Toshiba "recharge battery" splash. But it always works at the second attempt.
22:21:35Soul-SlayerYes I get that also
22:21:50Soul-SlayerAre you using the empty system files or do you still have the OF ones?
22:22:03linuxstb_I have the OF ones.
22:22:13Soul-SlayerRight, then I have no idea why mine does this
22:22:26*linuxstb_ is waiting impatiently for kkurbjun's flashing work to progress
22:23:08saratogadoes rockbox boost when buttons are pressed?
22:23:25XavierGrsaratoga: I don't think so
22:23:41linuxstb_saratoga: No, but I think it could be nice, at least on some targets
22:23:48saratogathat might be a nice feature for the PP targets, at least until smp works better
22:24:00saratogaif I make a patch for PP, would it be accepted?
22:24:05amiconndan_a: So the calibration value is 16000?
22:24:31 Part Domonoky_
22:24:57linuxstb_saratoga: I think it's definitely a worthwhile experiment. I don't think it's needed for all PP targets though - just ones with large colour LCDs (e.g. Color/Photo, Sansa and Video)
22:25:26linuxstb_saratoga; It's probably also not needed in all screens - e.g. boosting in the WPS isn't needed
22:26:12amiconnsaratoga: Boosting/unboosting in quick succcession is a bad idea and must be avoided
22:26:39amiconnEach transition needs to wait for the pll to relock, that's why
22:27:20linuxstb_I had in mind a boost with something like a 5-10 second timeout.
22:27:36amiconnWhat should that be good for, btw?
22:27:46Buschelbtw, is it known that pc-sim does not compile? i have to comment wma.
22:27:58amiconnYou need to reconfigure
22:28:06linuxstb_Make the UI more responsive. It's noticably more reactive when the CPU is boosted on my Color.
22:28:10amiconnThe sim compiles fine, see build table
22:28:35amiconnlinuxstb: There seems to be a problem in the menus that doesn't exist in the browser
22:28:58Buschelhmm, maybe i should really get a clean trunk and re-do my patches...
22:29:14amiconnThis has nothing to do with boost
22:29:37amiconnClearly noticeable on the video: menus are sluggish at 30MHz, browser is okay
22:30:01 Join kubiix [0] (
22:30:12 Quit Buschel ()
22:31:05 Part rogelio ("Kopete 0.12.4 :")
22:31:06*amiconn thinks that this problem needs to be hunted down & fixed
22:31:47dan_aamiconn: I think 16000 is correct
22:32:09amiconnOkay. SHould I commit the fixed pll divider + added debug stuff?
22:32:18amiconnThe debug stuff is there for PP50x as well
22:33:08 Quit low_light ("CGI:IRC (Ping timeout)")
22:33:19*amiconn wonders where that limiter hides :(
22:33:43 Quit BigMac (Read error: 110 (Connection timed out))
22:33:51amiconndan_a: If there's really a clock limiter, this explains the especially poor performance on G3
22:34:02 Join kubiixaka [0] (
22:34:19amiconn66MHz is only 10% more than what the iFP CPU has, plus the PP5002 has a broken cache...
22:34:36amiconn(otherwise the calibration factor would be the same as on PP502x)
22:34:58dan_aThe 5002 is rated up to 90MHz though, IIRC
22:35:31 Join Rincewind [0] (
22:36:42 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
22:36:50 Quit amiconn (Nick collision from services.)
22:36:51 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
22:37:09 Quit jhMikeS (Read error: 104 (Connection reset by peer))
22:37:20amiconndan_a: yes it is
22:37:53amiconnMaybe it's a safety measure that you have to unlock >66MHz.
22:38:18amiconnSomehow I'm not convinced though...
22:45:31 Quit jhulst (Read error: 104 (Connection reset by peer))
22:46:38 Quit Guile` (Read error: 104 (Connection reset by peer))
22:46:59 Join Guile` [0] (n=Guile@
22:50:27 Quit kubiix (Read error: 110 (Connection timed out))
22:50:51amiconndan_a: If you disable cpu scaling and then use the clock estimation, does it yield 66MHz or 75MHz?
22:51:58dan_aamiconn: 66. I then dropped the clock speed to 66MHz (outl(66, 0xcf00501c) in ipod_set_cpu_speed) and the estimation was still 66MHz
22:52:28 Join jhulst [0] (
22:52:31amiconnVery interesting... there seems to be a limit somehow...
22:53:40amiconnAnd the non-scaling setup uses a different pll divider (24) compared to the scaling setup (4), so it's independent from that one
22:56:06dan_aI drop the clock speed to 24Mhz and the estimation is 24MHz. There is definitely some kind of limit
22:56:14 Quit jhulst (Remote closed the connection)
22:56:22amiconnYes, and I guess if you set 53 you get 53
22:57:54linuxstbGodEater or GodEater_ : Around?
22:59:18dan_a53MHz gives 53MHz, as expected.
22:59:24amiconnYou could try something: Change line 113 (or line 138, depending on whether you have clock scaling enabled):
22:59:51amiconnfrom outl(0xa8, 0xcf00500c); to outl(0xac, 0xcf00500c);
23:00:08amiconnThis will use the other post-divider setting
23:00:18amiconnMaybe one is limited and the other is not
23:00:27dan_aTrying it now
23:00:58amiconnIiuc, this register defines the mode for the 4 selectable clock sources
23:01:16amiconnBit 0..1 define source 1, bit 2..3 define source 2 etc
23:01:49amiconnWe're using source 2, so this change sets the mode from 2 to 3
23:02:06 Join DerPapst [0] (
23:02:21***Saving seen data "./dancer.seen"
23:02:31amiconnMode 0 is 32kHz (but that probably requires tons of other registers to be set up properly), mode 1 is 24MHz, mode 2 and mode 3 use the PLL but with different post-dividers
23:02:35DerPapstGood evening :)
23:03:42*dan_a is an idiot
23:04:05dan_aI should have moved the multiplier back to 75 before doing that test!
23:04:15amiconnhehe, true
23:05:12dan_aStill 66MHz
23:05:25amiconnok, so the modes act the same
23:06:00amiconnnasty.... :\
23:07:11dan_aI wonder if IPL know anything
23:07:27dan_a(about this issue, I mean!)
23:11:27 Quit SirFunk (Read error: 113 (No route to host))
23:12:09 Quit Wiwie ("Miranda IM! Smaller, Faster, Easier.")
23:12:41 Join stripwax [0] (
23:14:03 Quit freqmod (Read error: 104 (Connection reset by peer))
23:14:23amiconndan_a: The ipl code doesn't tell anything else than what is in rockbox, and the wiki doesn't even know about the clock setup registers
23:14:45amiconnActually the clock setup code seems to be copied verbatim from ipl, including a mistake
23:15:07amiconnCheck the calculation in the comment for the #if block ...
23:15:26amiconn#if 0 block I mean
23:15:49dan_a64 is almost 66...
23:16:25amiconn line 147ff
23:21:07dan_aIt explains why raising the clock speed never had the effect I'd hoped for
23:23:33linuxstb_dan_a: I don't know if you've seen tomal's libmad optimisations (FS #6705), but it makes mp3 decoding on the 3G around 154% realtime (192kbps file) compared to 118% without the patch.
23:24:01dan_alinuxstb_: I haven't, but I'll take a look now
23:24:03amiconnSomeone really needs to check output accuracy...
23:24:27 Quit Soul-Slayer (Read error: 104 (Connection reset by peer))
23:26:43saratogalinuxstb: to boost for a couple seconds, I'd have to register a timer to turn off the boost ?
23:26:55saratogaor is there some ready made function to do it
23:27:57dan_aamiconn: How do we do that? Especially given that we know that the output will be different from the existing decoder?
23:28:11linuxstb_saratoga: Do you experience the same as amiconn - faster UI in the file browser than the menu?
23:28:49 Join PaulJam [0] (
23:28:55saratogaon the sansa with or without boosting?
23:29:30linuxstb_Yes, I guess without is the better test.
23:29:30DerPapstmenus should be laggy and the filebrowser not
23:29:33saratogabah sorry, i mean scaling
23:29:44DerPapsteven without that ;)
23:30:03linuxstb_DerPapst: So you notice it?
23:30:34DerPapstactually haven't tried yet.. i just repeat what i've read a few hours ago ;)
23:30:38DerPapstwill test now.
23:30:49saratogawith scaling disabled, everything is fast
23:31:38saratogaat least playing with it now, i couldn't notice a difference between the database, file browser or settings menus while playing a lame aps mp3
23:32:09*linuxstb_ switches computers
23:32:11 Quit linuxstb_ ("CGI:IRC")
23:33:09saratogayou know, a very easy way to handle boosting on button touches would be to just have the back light handle boosting
23:33:21saratogasince presumably people who have the light on are interacting with the player
23:33:23amiconndan_a: Some time ago I've read about some test method on the libmad website
23:33:32saratogaat least for targets like the sansa that are useless without backlight on
23:33:42amiconnThat would be short sighted
23:33:59amiconnIt would mean the cpu would be always boosted when the backlight is always on
23:34:23amiconnAnd there are situations where you want the backlight always on without operating the ui
23:34:48saratogadoesn't the back light use many times more power then boosting?
23:34:49*PaulJam keeps his backlight almost always on
23:34:55amiconnBoosting more isn't solving the problems
23:35:15amiconnIt only shortens battery runtime
23:35:33saratogano but its a trival to implement solution until SMP support is working better
23:35:39saratogathen we can simply revert it
23:35:40davinawith headphones on and your music machine in your pocket, who needs a backlight
23:35:42linuxstbsaratoga: I think that if boost-on-button-press is implemented, it's only needed in the list widget, not everywhere.
23:35:49DerPapstfilebrower is as reactive as the plugin menu... but browsing themes is slower.
23:35:52 Quit saratoga ("CGI:IRC")
23:35:57 Join saratoga [0] (i=980398fe@gateway/web/cgi-irc/
23:36:07 Nick saratoga is now known as saratoga2 (i=980398fe@gateway/web/cgi-irc/
23:36:07linuxstbThe plugin menu _is_ the file browser...
23:36:16saratoga2linuxstb: good point
23:36:20DerPapstthat expalin it..
23:36:25DerPapsti didn't know that.
23:36:39amiconnThe PP5002 datasheet states 170mW for mp3 decoding. That's in the same order of magnitude as the backlight
23:37:13amiconndavina: The music machine isn't only used in the pocket
23:37:23 Quit billytwowilly (Remote closed the connection)
23:37:39amiconnI use it in my car, and my car.cfg sets the backlight to always on (even on mono/greyscale targets)
23:37:40 Join billytwowilly [0] (
23:37:56linuxstbBut yes, on my Color, the menus do seem slower than the file browser, although both are slower than I would like, and are nice and fast when the CPU is boosted.
23:38:19amiconnSo something must be wrong with the menus
23:38:38linuxstbThe difference is even more noticable when boosted.
23:39:00davina amiconn thats true, in the car mine plugs into its ipod dock and charges from the battery so i can mount it on the dashboard with the light on, and also at home in the stereo doc it charges and also i can have the screen on. but i dont have screen lit up on battery alone except to change tracks
23:39:22amiconnI'm not sure, but I *think* I remember having seen the menu code updating the lcd more than once per step in the sim
23:39:48davinaanyway its night night from me, sweet dreams all
23:39:52 Quit davina ("xchat on Ubuntu 7.04")
23:40:06 Join Hilikus [0] (n=aaa@penhune-lab-172-107.Concordia.CA)
23:40:53saratoga2amiconn: at least for the sansa, toni measured 40ma x 1.5(?) volts == 60mw for mp3 playback w/ scaling
23:41:07Hilikushey guys, i want to record a live show usingmy dap, i was wondering if an uncompressed recording would give me longer battery life because theres no encoding or less since theres more disk spin?
23:41:18saratoga2oh that would be 40ma at battery voltage i guess
23:41:35saratoga2so probably what you said
23:42:07amiconnBut of course the PP5022/5024 draws less than the PP5002, with the PP5020 _probably_ inbetween
23:42:33linuxstbHilikus: I would guess less due to the increased disk activity, but it's just a guess. I don't think anyone has done recording battery benchmarks with different encoders.
23:42:47saratoga2i think its just color targets (IE sansa, nano and video) where this would be a big issue
23:43:07Hilikusso is there any recommendation on which format to use? what do you use for a long recording?
23:43:08 Quit Hadaka (Read error: 60 (Operation timed out))
23:43:11saratoga2maybe photo too
23:43:28linuxstbsaratoga2: The Nano's LCD is small, so I would guess it's fine. The Color/Photo is slow (it's what I have)
23:44:28linuxstbAnyone know if the list widget uses partial lcd updates? I know there was a patch for it, but don't recall it being committed.
23:44:59amiconnIt does. But perhaps the menu code does an additonal update, perhaps even fullscreen?
23:45:05 Join iamben_ [0] (
23:45:21 Quit iamben (Read error: 104 (Connection reset by peer))
23:45:23amiconn-> sim, and add DEBUGF() messages to lcd_update() and lcd_update_rect()
23:45:35Hilikusany idea how big would a 2hour wave file mono at 44k would be??
23:46:09DerPapstunder 4GB ;)
23:46:25petura fair bit below 2GB
23:46:35peturoh mono.. even less
23:46:43linuxstbIIRC, a CD is 650MB for 74 minutes of Stereo.
23:46:54Hilikusi thought it wouldnt fit in my 20gb free space
23:46:55peturwhy wiould you want to record in mono?
23:47:04Hilikusbuilt in mic
23:47:16peturdon't use that...
23:47:25Hilikusi cant use a real mic
23:47:41peturwhat device?
23:47:49Hilikuspetur you this its too bad?
23:48:10peturinternal mic picks up HDD spinup noises
23:48:15DerPapstHilikus: a bigger problem would be the 4GB limit of fat32 and wave :P
23:48:25peturdon't know how bad it is on x5
23:48:32Hilikusso ive heard, but its either that or not record it at all
23:48:39HilikusDerPapst mmm good point
23:48:44peturDerPapst: we split at 2G
23:48:50DerPapstah ok.
23:48:55DerPapstdidn't know that
23:49:09*DerPapst has never recorded with rockbox
23:49:10Hilikusim thinking on using lossless encoding
23:49:13peturyou can record untill disk full or batt empty ;)
23:49:36Hilikuscause i will probably sound edit the file after the show
23:49:38peturHilikus: if you go for internal mic, I would record to mp3
23:50:02lazkais there a way to exclude all plugins from compiling?
23:50:10peturyou'll have less spinups and internal mic quality is bad anyway
23:50:16Hilikuslazka change the SOURCES file
23:50:19linuxstblazka: "make bin" (that also excludes codecs)
23:50:22DerPapstlazka: removem from SOURCES
23:50:35saratoga2wow didn't know about make bin
23:50:38DerPapst^ what linuxstb said :P
23:50:41saratoga2would be handy
23:51:05 Join rotator [0] (n=e@rockbox/developer/rotator)
23:51:08Hilikuspetur, thanks for your oppinion, thats what i was looking for
23:51:08saratoga2is there one that excludes plugins but includes codecs?
23:51:12linuxstb"make help"
23:51:19peturHilikus: try to get an external mic ;)
23:51:22linuxstbsaratoga2: No.
23:51:42Hilikusits not exactly "legal" to record the show eventhough im paying a lot of money to see it
23:52:01Hilikusbut i understand, some people start selling it
23:52:16Hilikusi just want to have the memory of the show
23:52:17*petur taped bjork on sunday and was lucky to get through security :)
23:52:32Hilikusand you had an external mic?
23:52:41peturand mic amp :)
23:52:42DerPapstjust claim that it can't record :P
23:52:55 Part maffe
23:53:09 Join maffe [0] (
23:53:39Hilikuspetur so you think mp3 is fine even if i want to edit/retouch the file afterwards?
23:53:40linuxstbHmm, moving in the menu causes three almost full-screen lcd_update_rects (the status bar is excluded).
23:53:49Hilikusjust uncompress it and edit it?
23:54:00 Quit ompaul ("Leaving")
23:54:27peturHilikus: there's always quality loss... if it is a loud show, go wav, if it is quiet, go mp3 to avoid spinup noises
23:54:38linuxstbAnd moving in the file browser just causes a single lcd_update_rect (same area)
23:54:45Hilikusi think its going to be loud! its tool
23:54:56DerPapstHilikus: good audio editing programs can work with encoded formats as good as with wave
23:55:02peturHilikus: then go wav...
23:55:19peturbut I don't know if the internal mic won't overload though...
23:55:21Hilikuswhat does the volume have to do with the format petur?
23:55:36pixelmathe problem with the internal mic in the X5 is that you can only have two gain settings... using an external one through line in will give 30 (or something like that)
23:55:48peturthe mic will record disk spinups
23:56:07linuxstbDerPapst: Not really - mp3 limits the manipulations you can do without reencoding.
23:56:22Hilikusi see
23:56:25peturif it is loud, I would go for the lowest gain settings and pray
23:56:40Hilikusim going to run a test with the opening band
23:56:51*petur always does...
23:56:53Hilikusjust to make sure its not clipping all the time
23:57:01DerPapstlinuxstb: didn't have any difference with audidtion and cubase afair
23:57:18linuxstbDerPapst: Are you sure they didn't just transparently re-encode?
23:57:27DerPapstthey probably did
23:57:37Hilikusyah thats what i was thinking too
23:57:39*DerPapst hides
23:58:17 Join jhulst [0] (
23:58:18linuxstbSeems the menus perform two lcd updates when you press a button, and another one when you release it...
23:58:34Hilikuspetur, last question, if i used mp3, what bitrate would you recommend>
23:58:47linuxstbWhereas the file browser is just performing one.
23:58:49peturthe highest?

Previous day | Next day