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 2008-05-13

00:00:32 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
00:01:13 Join voltagex_ [0] (
00:01:40voltagex_well, the segfault i was experiencing on osx 10.5 is fixed by rbutil svn apparently
00:05:11 Quit waldo (Remote closed the connection)
00:07:55 Quit jgarvey ("Leaving")
00:08:14 Quit voltagex (Read error: 110 (Connection timed out))
00:08:55 Quit bluebrother ("Verlassend")
00:12:47 Quit mf0102 ("Ex-Chat")
00:14:13evilnickDoes the gigabeast with fresh bootloader mount under windows or just linux?
00:14:42voltagex_evilnick: do you mean with rockbox installed?
00:15:04evilnickwell... with the bootloader installed but before I put a build onto it
00:15:21 Nick JdGordon|zzz is now known as JdGordon (i=jonno@rockbox/developer/JdGordon)
00:15:25evilnick(I'm not comfortable enough in linux to mount and fix the partition)
00:15:40voltagex_evilnick: should boot into Bootloader USB Mode
00:16:55evilnickvoltagex_: I have done, but in Ubuntu it won't let me mount (through a right click in Computer)
00:17:16voltagex_evilnick: what about in Windows?
00:17:24voltagex_evilnick: what's the problem with the MP3 player anyway?
00:18:05evilnickThere's no problem with it, just I would prefer to do the rest of the install in a more familiar environment
00:18:17 Join Rob2223 [0] (
00:18:19 Quit dabujo_ (Read error: 104 (Connection reset by peer))
00:22:39 Quit matsl ("Riece/3.1.2 XEmacs/21.5-b28 (fuki, linux)")
00:23:16evilnickGot it working! Fantastic!
00:27:59 Quit RBPFC1 ()
00:28:18saratogapixelma: the wma page is up to date
00:28:35zesweet, my CF card in this <$4 CF-IDE adaptor is working fine in my spare rio karma bits instead of an HD
00:28:58saratogawe only support WMA std, none of the special flavors (lossless, Pro, Voice, etc)
00:30:28 Quit brent113_ (Read error: 110 (Connection timed out))
00:30:31pixelmasaratoga: aha, will keep it in mind for next time. Maybe you can mention the latter somewhere on that page?
00:34:53saratogaok updated
00:35:40 Join lilith634 [0] (
00:36:01saratogaalso i'm hesitant to say so on the wiki page just yet, but I think we're pretty much 100% WMA compatabile, at least for files above 16kbps
00:36:10pixelmagreat, thanks :)
00:36:13 Quit Rob2222 (Read error: 110 (Connection timed out))
00:37:31 Quit OlivierBorowski (Remote closed the connection)
00:39:02*Nico_P reads about a nice looking 160 GB DAP that isn't an iPod
00:40:07Nico_P"Aigo Patriot MV5920"
00:40:26 Quit lilith634 ("")
00:42:02 Quit ender` (" The propeller is just a big fan in front of the plane used to keep the pilot cool. When it stops, you can actually watch th")
00:44:12zeNico_P: interesting looking unit/specs
00:46:19Nico_P"MP5" doesn't mean much to me though
00:46:57zefirst thing i found about it says "MP5 Player (a PMP in China that supports RealMedia)"
00:47:20zeVideo: MPEG-1 (DAT, MPG), MPEG-2 (VOB), MPEG-4 (AVI, Divx3-5.x, Xvid), ASF, WMV9, RM / RMVB
00:47:24zeAudio: MP3, WMA, WAV, AC3, OGG
00:47:35jhMikeSanother thing replete with NDA hardware I'm sure :)
00:49:22zedamn i accidently grabbed an image of the partition instead of the device when i backed this card up
00:51:00 Quit Nico_P (Remote closed the connection)
00:51:04 Quit gevaerts ("nearly morning")
00:53:45 Join Shaid [0] (
00:57:39 Quit ompaul (Client Quit)
01:04:53voltagex_jhMikeS: once rockbox runs on something does an NDA really matter?
01:13:21 Nick gregzx is now known as GregZzzZZX (
01:13:51 Quit GregZzzZZX ("ChatZilla 0.9.82 [Firefox]")
01:14:18 Quit mackes ("Mackes is out")
01:27:14jhMikeSvoltagex_: as far as docs, sure it does. even a driver won't tell you everything about the reason it's written that way.
01:28:29 Part toffe82
01:32:16voltagex_jhMikeS: ah ok, I'm not skilled enough for low level programming like that anyway :(
01:32:49voltagex_Anyone here from Canberra? I'm looking for someone to replace the hard drive and battery in my H300
01:33:02voltagex_or rather, one of my H300s
01:38:39 Quit Falafel (" HydraIRC -> <- Organize your IRC")
01:40:15***Saving seen data "./dancer.seen"
01:40:54zebtw if anybody ever has anything to try re firmware flashing or anything on the rio karma, i can now use flash cards with my spare board and would be willing to throw anything at it hehe
01:43:14 Quit evilnick ("ChatZilla 0.9.82 [Firefox 3.0b5/2008041515]")
01:45:06 Join csc` [0] (n=csc@archlinux/user/csc)
01:48:42 Quit SirFunk (Remote closed the connection)
01:57:52 Join brent113_ [0] (n=brent113@
02:00:21 Quit einhirn_ (Read error: 104 (Connection reset by peer))
02:04:17 Quit dan_a ()
02:14:51tkoodadoes rockbox on an 4/5th gen ipod support being controlled by the docking port?
02:17:09soapnot at this time, though there is a patch-in-progress in the tracker.
02:25:06 Join toffe82 [0] (
02:27:09 Join fdinel [0] (
02:36:24*jhMikeS wonders why the fancy term "digital exponential filter" was used instead "moving average" in the powermgmt code :D
02:53:02 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
03:00:55 Part pixelma
03:03:52 Quit XavierGr ()
03:05:32 Quit linuxstb (Read error: 110 (Connection timed out))
03:07:53 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
03:16:41jhMikeSactually it reduces to a leaky integrator
03:29:07 Join SirFunk [0] (
03:40:19***Saving seen data "./dancer.seen"
03:47:11 Nick fxb is now known as fxb__ (
03:47:40 Quit homielowe ()
03:51:45 Quit DerDome (Nick collision from services.)
03:51:46 Join DerDom1 [0] (
03:51:56 Nick DerDom1 is now known as DerDome (
03:56:43 Join low_light [0] (i=4cbde1a4@gateway/web/ajax/
03:59:05 Join nuonguy [0] (
04:04:46JdGordonanyone know what needs to be fixed to get a port showing the correct remaning batt life?
04:07:05jhMikeSsomehow know how much current drain there is and how many mAh remain in the cell
04:08:27JdGordoncan the numbers be fiddled with untill we get a better approx?
04:08:44JdGordonatm the mini2g says 3 hours for full when its alot closer the 13
04:09:19low_lightJdGordon: you were looking for me the other night?
04:11:17*nuonguy is a first-time rockbox user
04:11:34nuonguywhere do I put the songs so that a database refresh will find them?
04:12:46low_lightrockbox should scan the whole disk
04:12:57low_lightbut I don't use the database
04:13:40nuonguyI have a sansa e280 with music on it that it plays natively
04:14:08nuonguyI tried this on the plane a coupla days ago, and I couldn't seem to figure out what I was doing
04:14:31jhMikeSJdGordon: It might be possible to calulate a discharge rate dV/dT around the current voltage and extrapolate that from the battery curve.
04:15:31JdGordonlow_light: I was but too late, got a mini instead of a mr100
04:18:25low_lightare you doing a CF mod?
04:22:05low_lightI plan to try it in the mr100, but need to get an adaptor
04:22:34JdGordondoes it use the same disk as the mini?
04:24:00JdGordonjhMikeS: can that be done from a batt bench log?
04:26:27jhMikeSJdGordon: just the curve part. the deltas need to be calculated to get a discharge rate.
04:30:25low_lightJdGordon: the mr100 disk has the flex cable build in. The adapter is actually for the mini, but I think it should work.
04:30:59JdGordonah, so you cant just hook a CF card in?
04:35:24 Join miepchen^schlaf_ [0] (
04:36:52 Join homielowe [0] (n=eric_j_l@
04:39:47 Quit amiconn (Nick collision from services.)
04:39:53 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
04:45:24 Join djuniah [0] (n=47e15003@gateway/web/cgi-irc/
04:47:25 Quit djuniah (Client Quit)
04:47:38 Join djuniah [0] (n=47e15003@gateway/web/cgi-irc/
04:49:26 Quit miepchen^schlaf (Connection timed out)
04:53:06 Quit djuniah ("CGI:IRC (Ping timeout)")
04:57:28 Quit Status (Read error: 113 (No route to host))
04:58:57 Join cool_walking_ [0] (
04:59:13 Quit Thundercloud (Remote closed the connection)
04:59:39cool_walking_ is completely blank. "Content-Length: 0" is in the headers.
04:59:58 Quit csc` (Read error: 110 (Connection timed out))
05:00:24JdGordonbecause you didnt add index.php
05:00:38cool_walking_same thing
05:00:47cool_walking_and /tracker/ has worked before
05:01:09 Quit kugel (Remote closed the connection)
05:03:01 Join Evilnick [0] (
05:05:33 Quit homielowe (Read error: 110 (Connection timed out))
05:06:22cool_walking_with "type" argument, it's fine
05:27:05 Quit Horscht ("I am root. If you see me laughing, you better have a backup")
05:40:20***Saving seen data "./dancer.seen"
05:40:53 Join ctaylorr [0] (
05:47:01Shaidno failsafe in the code for users not passing a value, and script erroring and aborting, I'd guess
05:53:52cool_walking_It used to show all task types (patches, bugs, and feature requests). Just checked today and got the blank page.
06:01:44 Quit ctaylorr (Read error: 110 (Connection timed out))
06:08:10 Join Booth1nator [0] (n=cb0a7954@gateway/web/cgi-irc/
06:11:16 Join homielowe [0] (
06:12:44Booth1natordoes anyone know how to load rockbox and linux on ipod when it says ipod not detected when i know it is plugged in>
06:13:10advcomp2019Booth1nator, what ipod
06:13:41Booth1natorclassic 80gb running 5.5G
06:14:19advcomp20196g or 5.5g?
06:14:45Booth1nator5.5G im pretty sure
06:15:09scorcheif it is a classic, it is a 6G
06:15:50Booth1natork. so the linux and rockbox are not compatible
06:16:09scorcheneither will work ont he classic
06:16:28Booth1natorthat sux alot
06:16:36 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
06:16:41BHSPitLappyIf it's got an aluminum exterior, you can't hack it (unless it's an old Mini)
06:17:11Booth1natorsomeone has to be able to hack it eventually
06:17:19Booth1natorcurse apple
06:17:29scorchewell, it ownt happen with no kneo trying
06:18:02Booth1natorso no1 is even bothering to try?
06:18:18scorcheno one is working on it
06:18:52Booth1natorthat sux heaps
06:22:17 Join Compaq_Owner [0] (
06:22:52 Join homielowe_ [0] (
06:23:26 Nick Compaq_Owner is now known as FyshY_ATF (
06:23:42FyshY_ATFI need help with my IPOD
06:24:05FyshY_ATFmy install didn
06:24:25FyshY_ATFdidn't work and now i dont know what to do
06:24:44advcomp2019FyshY_ATF, what ipod
06:25:14FyshY_ATFmy ipod ...
06:25:48Booth1natorwats ur ipod?
06:26:05LloreanBooth1nator: Please use real words, not things like "ur" as per the channel guidelines
06:26:06FyshY_ATF1st gen nano
06:27:20LloreanFyshY_ATF: What do you mean by "my install didn't work"
06:27:37FyshY_ATFit says Error ! ant load rockbox.ipod: hold menu+select ro reboot then select play for disk mode
06:28:08LloreanFyshY_ATF: You haven't properly installed a build then, just a bootloader. Go into disk mode, like it says, and follow the "Installing the Firmware" instructions in the manual.
06:28:36FyshY_ATFbut i cant get into disk mode to reset it
06:28:43LloreanYes you can.
06:29:03LloreanJust hold Menu+Select, and as soon as the screen clears, hold Select+Play
06:29:58FyshY_ATFit keeps going to the black screen
06:30:11LloreanI don't know what you mean by "the black screen"
06:30:17 Join homielowe__ [0] (
06:30:34FyshY_ATFwhen it says error that black screen
06:30:52LloreanThen you're not holding Select+Play soon enough
06:31:44 Quit homielowe (Read error: 104 (Connection reset by peer))
06:32:46FyshY_ATFstill the same ... as fast as my thumbs will allow
06:33:02LloreanAnd you're holding the center button, and the bottom button, and not releasing them?
06:33:08 Quit low_light (" ajax IRC Client")
06:33:23FyshY_ATFyup ... even befor the apple appears
06:34:05LloreanWell, the thing is, the bit of code that reacts to the Select+Play is permanent code put there by apple. It runs before any of our code does, and it's the same code that goes and loads our code if you don't hit Select+Play
06:35:13LloreanSo, keep trying. Try doing it a little later, try being more careful not to wiggle your fingers, but it's permanent in-ROM code that handles this, so either it works or something has physically gone wrong with your iPod, something unrelated to Rockbox at all.
06:37:51FyshY_ATFlike my buttons not working .. is that what you mean ?
06:38:15LloreanThe most likely cause really is you just not quite getting it right.
06:38:40FyshY_ATFokay ... im trying .. =(
06:38:42 Quit homielowe_ (Read error: 104 (Connection reset by peer))
06:38:49LloreanTry just turning on Hold instead of holding Select+Play, this will cause the rockbox code to load the original firmware.
06:40:05FyshY_ATFThat worked !
06:40:21FyshY_ATFnow what do it do to finish install ?
06:40:38LloreanFollow the "Installing the Firmware" instructions in the manual. I already told you this.
06:41:07FyshY_ATF*thank you
06:42:30 Quit homielowe__ ()
06:50:40 Quit Booth1nator ("CGI:IRC")
06:59:20JdGordonanyone know what voltage goes out the ipod remote pins?
07:00:49JdGordon3.3v apparently :(
07:04:22 Join linuxstb__ [0] (
07:05:12 Join homielowe [0] (
07:06:09JdGordonany suggestions where to get 5v from the ipod mainboard?
07:06:29*JdGordon miht have to take the easy way and cut the battery wires
07:13:36voltagex_Anyone here who can help me with a bricked H340? My second one no longer works
07:16:19 Quit linuxstb_ (Read error: 113 (No route to host))
07:16:58 Quit FyshY_ATF ("ChatZilla 0.9.81 [Firefox]")
07:18:51 Join austriancoder [0] (n=austrian@rockbox/developer/austriancoder)
07:31:29 Part toffe82
07:40:21***Saving seen data "./dancer.seen"
07:40:48ShaidJdGordon: How about where the battery wires join the motherboard?
07:52:06 Join benime [0] (
07:52:52benimedunno if this is the place to address it, but the the links to the pcb scans for the iriver h1x0/h3x0 on this page are broken
07:54:49 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:58:47austriancoderPlease review:
08:00:54Lloreanaustriancoder: You're really best off posting to the -dev mailing list, so you don't have to worry about people spotting comments you make while nearly everyone isn't around
08:03:58 Quit homielowe (Read error: 104 (Connection reset by peer))
08:03:58LinusNaustriancoder: shouldn't you have a BALANCE_CAP as well?
08:04:14 Join homielowe [0] (
08:04:23LinusNi can't see the connection between clipping and balance
08:05:43austriancoderLinuN: hmm.. you are right.
08:07:22 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
08:08:54 Quit BigBambi (Remote closed the connection)
08:09:12 Join voltagex [0] (
08:09:35 Join thechamp [0] (n=4b13ae09@gateway/web/cgi-irc/
08:09:43 Quit thechamp (Client Quit)
08:09:49 Join codingtigger [0] (
08:10:02 Join thechamp [0] (n=4b13ae09@gateway/web/cgi-irc/
08:10:03 Quit thechamp (Read error: 104 (Connection reset by peer))
08:10:05 Quit miepchen^schlaf_ ()
08:14:17 Join Rob2222 [0] (
08:20:45LinusNhehe, it's a little embarassing to have the W3C XHTML badge in the forum when the markup contains 165 errors... :-/
08:21:13LloreanI went through and fixed it, long ago. But the new menus contain most of 'em
08:21:20LloreanNew menu on the left, that is.
08:21:20 Nick linuxstb__ is now known as linuxstb (
08:22:25*Llorean just needs to sit down and step through them again sometime soon
08:22:34 Quit voltagex_ (Read error: 110 (Connection timed out))
08:22:50 Join Romanian [0] (
08:23:48 Join tvelocity [0] (
08:23:55 Quit BHSPitLappy (Remote closed the connection)
08:24:43 Join Mathiasdm [0] (
08:25:03 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
08:25:16JdGordonBUGGER! seems I lost my rec screen patch
08:25:26*JdGordon not having a good day :/
08:25:27voltagexIs it possible to source capacitors and resistors to fit the H340?
08:25:49RomanianIs there a way to scandisk in Ubuntu?
08:26:12voltagexRomanian: FAT32 or EXT2/3?
08:26:24voltagexRomanian: that's a question for #ubuntu
08:26:24Romanianvfat, which = FAT32 afaik
08:26:34RomanianRight, thanks.
08:26:37voltagexbut fsck.fat32 i think
08:27:07Romanianer, it's for my external HDD MP3 player
08:27:40voltagexRomanian: yeah. try fsck /mnt/mp3player
08:28:32RomanianMany thanks.
08:29:29Romanianit wasn't connected :\
08:29:54voltagexthat generally helps
08:30:11Romanianhmm, I'm not sure what the mount point is
08:30:16Romanianoh, IAUDIO
08:30:30RomanianNo, I lied.
08:30:46voltagexopen a console and type mount
08:30:47Romaniangah, it's not showing up
08:30:57voltagexin ubuntu it should be /media/IAUDIO
08:31:10Romanianaye, it should
08:31:12Romanianthere we go
08:31:31cool_walking_You shouldn't have it mounted when doing fsck.
08:31:42cool_walking_And fsck wants the device node, not the mount point
08:31:53RomanianSo it'd be /dev/*** ?
08:32:07RomanianHow do I know what the *** would be?
08:32:13 Quit DerDome (Read error: 110 (Connection timed out))
08:32:24cool_walking_"fdisk -l" as root
08:32:31GodEatercheck the output of dmesg when you plug it in too
08:32:32 Join ender` [0] (
08:32:52*austriancoder Build should have been done 6m 32s ago, at 08:26:06
08:32:54Romanianaye, sdb
08:33:10cool_walking_it should be a partition i.e. have a number at the end
08:33:10 Quit Rob2223 (Read error: 110 (Connection timed out))
08:33:26RomanianUse first FAT or second FAT?
08:33:30RomanianFAT32, LFN
08:33:47Romanian"FATs differ but appear to be intact"
08:33:54 Join spiorf [0] (
08:34:13 Quit saratoga ("CGI:IRC (Ping timeout)")
08:34:40Romanianit says "free cluster summary wrong", and I say "correct", but then leaves me with "Leaving file system unchanged."
08:35:01GodEaterI don't think fdisk will fix the file system if it's mounted
08:35:08Romanianit's not mounted
08:35:37GodEaterhave you tried fsck -a ?
08:35:43codingtiggerAny one had problems installing rockbox on a Sansa e250 from Vista?
08:36:22JdGordonUAC probably causing trouble?
08:36:24GodEatercodingtigger: yes I believe quite a few people have - it's normally the UAC stuff getting in the way
08:36:25cool_walking_codingtigger: Never tried, but the first thing I'd check when dealing with Vista is that you have admin rights.
08:36:48RomanianI just did "fsck -a /dev/sdb1" and it appeared to have worked, since "sudo fsck /dev/sdb1" gives no more info
08:36:57codingtiggercool_waling_: I have admin rights.
08:37:05codingtiggerGodEater: What's
08:37:14bughunter2google it :P
08:37:47codingtiggerk, got it. I'll fiddle with it for a bit...
08:37:50GodEaterUAC == Stupid Vista feature that causes more trouble than it solves (IMO)
08:38:13Shaidlike most vista features
08:38:28bughunter2GodEater: why? i thought it was just one popup?
08:38:32GodEaterlike vista period
08:38:35bughunter2if you admin rights that is :)
08:38:48bughunter2s/you/you have/
08:39:00GodEaterbughunter2: because people keep pestering us about it ?
08:39:40codingtiggersorry, didn't mean to bug you guys about UAC
08:39:55GodEatercodingtigger: we don't blame you, we blame M$ :)
08:40:18JdGordonwhats the point of flashing the remote LCD when recording starts?
08:40:28codingtiggerMaybe a note in the users manual would help too. :)
08:40:52JdGordonno, it doesnt flash.. it just says "remote lcd off" then "remote lcd on"!?
08:41:10GodEatercodingtigger: it would - but no-one's written one, and none of the regular manual contributors own vista to my knowledge
08:41:14RomanianOn a side note, do you think a Conroe-L (Celeron) would be enough for a home-built NAS?
08:41:24GodEaterit would be nice if someone who actually HAD vista could write such an item and submit it to us
08:41:26GodEaterhint hint
08:41:47GodEaterRomanian: what has that to do with Rockbox ?
08:41:49codingtiggerworking on it...
08:41:57LloreanRomanian: NAS has nothing to do with Rockbox...
08:41:59RomanianWhoa, what? fsck -a /dev/sdb1 seemed to have added a hell of a lot of files to my player.
08:42:08RomanianI know it doesn't. Hence the side-note. No response was required.
08:42:22scorchethere are no side-notes here..
08:42:24LloreanRomanian: This channel is on-topic, please keep it that way.
08:42:26RomanianI see.
08:42:26voltagexRomanian: PM or #ubuntu
08:42:43*cool_walking_ trying not to say "..only Zuul"
08:44:21 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
08:44:23cool_walking_Romanian: Or read the channel topic for somewhere else to go.
08:47:03codingtiggerSo, I have admin rights, I disabled the UAC grab-your-desktop thingy, and when I try to drag the .rockbox folder from the zip file to my device, Vista doesn't let me do it.
08:47:09LinusNLlorean: the markup is valid again
08:47:28LloreanLinusN: Thanks. :)
08:47:40codingtiggerIf I use 'Extract all" like the manual says, I can't selec my device as the destination
08:47:55 Join Bagderr [0] (n=daniel@gateway/web/cgi-irc/
08:48:06GodEatercodingtigger: can you not just type the drive letter in ?
08:48:20codingtiggerVista doesn't give me a drive letter
08:48:21 Nick Bagderr is now known as B4gder (n=daniel@gateway/web/cgi-irc/
08:48:31GodEaterthen you have it in MTP mode, and not MSC
08:48:43codingtiggerof course!
08:48:54codingtigger*operator error*
08:49:00GodEaterPEBKAC :)
08:49:08*JdGordon just found the rec trigger screen... wtf?
08:49:45 Quit leox_ (Read error: 104 (Connection reset by peer))
08:49:49 Join leox [0] (
08:49:50codingtiggerAnd it was clearly written in the manual too. I feel kinda stupid
08:50:18GodEaterif a day goes by and I've not done something that makes me feel stupid, I mark it with a smiley on the calendar
08:50:33JdGordongot many smileys? :D
08:50:37*amiconn really wonders what austriancoder is trying to achieve with his audio code shuffling :/
08:50:47GodEaterJdGordon: I've never needed any yet ;)
08:51:02amiconnBesides, if you really want to structure the audio code properly, the MAS code needs one more split.
08:51:37amiconnThe combination of MAS3507D + DAC3550 has quite different audio handling than the later MAS35xxF family
08:52:46voltagexreally, really need to source parts for my H340... :(
08:52:49voltagexwindow 4
08:52:55voltagexyou're not irssi
08:56:06codingtiggerOk, all good. Rockbox is up and running on my device. Thanks again!
08:56:54austriancoderamiconn: clean up sound.c and move all target specific code out of sound.c... later I want to use dsp in the sim
08:57:18 Quit BHSPitLappy (Remote closed the connection)
08:58:39JdGordondoes anyone here use the rec trigger screen?
08:59:14 Quit codingtigger ()
09:00:31LloreanJdGordon: I don't use recording at all, but is that the screen where you can set it up to start recording if there's a certain amount of noise?
09:00:50JdGordoni guess so, I also dont use it
09:02:13markunaustriancoder: I think it's a good idea. Makes it a lot easier to add a new sounds codec without having to changed the many #ifdefs in sound.c
09:02:14JdGordonit uses the sysfont, inline settings, way too much code
09:02:17*amiconn always wondered whether anyone apart from the original implementer uses it
09:02:34JdGordonwho put it in?
09:02:56LloreanI think it's popular among the REP crowd.
09:03:30JdGordonit adds *ALOT* of code... would use much less if it didnt do its own drawing
09:03:57 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
09:04:01amiconnIt does its own drawing because it show a peakmeter for adjusting the trigger levels
09:04:26JdGordonI see that... and understand why before vp's this was needed
09:05:06JdGordonreally all I want to know is do all the settings need to be visible on the screen at the same time? or can we out the settings into a list like normal ones with the peakmeter under it?
09:05:24 Quit ompaul (Client Quit)
09:05:55amiconnThe idea is that you adjust those settings and be able to immediately see the effect through the peakmeter
09:06:08JdGordonyes, thats fine... the pm will stay
09:06:31amiconnI *hate* those inline settings, but for this purpose I see no other way that would work equally well :/
09:06:44markunamiconn: why do you hate them so much?
09:07:37amiconnThey're confusing and ugly
09:07:37JdGordonwould it be OK to put a regular list above the peakmeters? (making sure that entering the setting will still show the peakmeters?) the list would then only show the setting name, not the values also
09:07:50markunamiconn: in what way confusing?
09:08:02 Quit spiorf (Remote closed the connection)
09:08:28markunI tried a inline settings patch once and rather liked it. You can check the value without the need to enter it
09:08:44JdGordonpixelma: hey, do you use the rec trigger screen?
09:08:56 Join petur [50] (n=petur@rockbox/developer/petur)
09:08:58JdGordonmarkun: seen 8968?
09:09:28amiconnThe values have different lenghts, and on a crowded screen (read: archos) they're looking strange in conjuction with the settings. And adjustment of those settings is inconsistent with the rest of rockbox
09:09:29JdGordonpetur: hey, sorry about last night, seems I've lost the patch
09:09:29JdGordondo you use the rec trigger screen?
09:09:54peturno, but is that a reason to dump it? ;)
09:10:16*JdGordon not looking to dump it
09:10:23JdGordonjust get some advice on vp-ing it
09:10:31peturI got a bit distracted so I didn't get that much further
09:10:41*JdGordon thinks that will give a hefty green delta
09:10:46amiconnJdGordon: I think that if the peakmeter stays visible (including the special indicators) while adjusting the setting, it would be ok
09:11:26JdGordonthats what I was hoping
09:11:28amiconnIf this works, I'd think the settings should use the standard settings lists as well ("just" viewported to leave room for the peakmeter)
09:11:30pixelmanot really no (except trying how that works at all). The point were it gets ugly/confusing is when the item names get cut off because they are too long (either small screen, and/or large font, and/or long setting name (maybe in a different language))
09:11:56pixelmathe first sentence was an answer to JdGordon ;)
09:12:00amiconnThe trigger screen uses the sysfont....
09:12:27amiconn...which is ugly for non-latin languages
09:12:36JdGordonamiconn: thats what I was going to do
09:12:48amiconnAh, no it doesn't
09:12:59peturJdGordon: I was a bit surprised I couldn't get the REC buttonpress through from the list (after adding it to the STD keymap), REC_REPEAT does get throuh. Then again, I know little of keymaps and viewports :/
09:13:34amiconnThe items outright overlap with the values in a few cases...
09:14:31amiconnUrgh, and on H10, adjustment is totally unintuitive...
09:14:54peturH10 keys and their maps suck big time
09:15:27amiconnTry the rectrigger screen...
09:16:12amiconnUp/Down adjust the value. FF/Rew move through the list but in the opposite direction from what you'd expect...
09:19:24 Quit shodanX (Remote closed the connection)
09:21:32 Join CaptainSquid [0] (
09:23:29JdGordonwho's smart enough to translate the threshold setting code into a min/max/step size for me ? (menus/recording_menu.c:406)
09:23:41 Part tkooda ("thanks")
09:25:14 Join DerDome [0] (
09:30:32*pondlife wishes that gui_synclist_init took a const void *data...
09:30:51pondlifeSadly, some code modifies this block (which is reasonable, I guess).
09:31:34pondlifeI could do a big const police raid if so.
09:31:58pondlifeBut I'm not planning on reworking the bookmark selector and debug menus right noe.
09:32:05JdGordonoh, +const.. righto
09:32:20JdGordonwhy is it bad if stuff modify that?
09:32:38JdGordonobviously thye only do it if they know what it is...
09:32:48pondlifeIt's not particularly, but there's a tree of const-iness that it would enable..
09:32:59pondlifeSo.. no worries.
09:33:25pondlifeDid you look at my peephole fix for the file deletion issue (over track transition) btw?
09:33:44 Nick fxb__ is now known as fxb (
09:34:19 Join tdtooke [0] (
09:34:44tdtookeAnybody awake out there?
09:34:56*pondlife yawns
09:35:31tdtookeDo you have any idea if the metadata tags on a mp3/aac/whatever file appear in a specific order or is it a random thing?
09:35:51pondlifeI don't think that's specified, for MP3 at least.
09:36:18pondlifeCertainly we shouldn't care too much (as long as they fit within our memory restrictions).
09:36:20tdtookeI had an idea for getting the sort tags patch working if it did
09:36:59linuxstbBut Rockbox will fill the metadata buffer on a "first come, first served" basis. So if you have long tags at the start, you may not see the last one.
09:37:16tdtookeMy idea was to set sortartist to whatever artist was, provided it came first and later in the parse tags function in metadata_common.c if it was found sortartist did exist then set it to the real value and change over tagnavi.config to use sortartist instead
09:38:04 Part Romanian
09:38:07tdtookeIf the ordering is arbitrary that won't work, but I'm thinking the answer is somewhere along those lines....maybe
09:38:43linuxstbI don't understand - why not just modify the tags after everything has been read?
09:39:17pondlife(Or tag your files properly...)
09:39:53tdtooketagging files properly is what I'm after, it's been decided that the sort tag is the right way to ignore a the so I'm wanting to get it off the ground
09:40:25***Saving seen data "./dancer.seen"
09:40:43tdtookeI haven't got enough of a grip on the code yet to modify tags after they've been read linux, :)
09:41:32pondlifeI thought we already supported the sort tags?
09:42:06tdtookewe have a patch which will build a database that supports it, but it doesn't do anything yet, I'm trying to come up to speed on it to figure out how to change that
09:42:46Lloreanpondlife: The idea is for it to sanely default to the "normal" tags if sort tags aren't present.
09:43:15linuxstbI would have thought that the sortartist should always be populated - pointing to "artist" if there was no sortartist in the file. Then tagcache needs to use sortartist when sorting by artist.
09:43:36pondlifeYes, so as long as the artist = "The Beatles" and sort artist = "Beatles, The", it should work. Or is this not yet implemented?
09:43:48tdtookenot, it's not yet implemented
09:43:57tdtookekinda redundant that
09:44:00Lloreanpondlife: The problem though is that for files where there's no "sort artist" because the "artist" is fine.
09:44:00pondlifeAh, ok. I thought it was
09:44:21LloreanThe way it's implemented you can explicitly use one or the other, but it won't default back to "artist" if "sort artist" isn't present (iiuc)
09:44:24tdtookeI was just basically asking about a bad idea I had on how to do this
09:44:36linuxstbif (sortartist == NULL) { sortartist = artist; }
09:44:44linuxstb(somewhere after the metadata has been read)
09:44:51pondlifeWe should be able to do that on the fly....
09:45:06 Quit austriancoder (Remote closed the connection)
09:45:08tdtookemakes sense, right now I'm in parse tags, I guess I need to read on a bit further
09:45:23pondlife...or retag your files so sortartist=artist where sortartist is blank...?
09:45:51tdtookeI'm not about to put up a patch that requires mass retagging to work right for everybody
09:46:10pondlifeI meant that the existing code should work if you retag.
09:46:18LloreanThe existing code would work if you retag.
09:46:37LloreanBut tagnavi should handle files with both sort artist and artist tags, or "normal" collections where the only tag is the "artist" tag equally well.
09:46:49pondlifeI agree
09:47:17LloreanBut database uses presorted lists, right?
09:47:38LloreanWouldn't it be something to do during the commit? For all songs with an artist but no sort artist, populate the sort artist list with their artist string instead, etc?
09:48:10tdtookethat's what I'm wanting to do
09:49:26tdtookenot good, I have 7 files open and am not thinking clearly....:)
09:49:53LloreanI thought database harvested all the metadata before starting the commit anyway. In which case I'd imagine the order of it within the file would be more or less irrelevant.
09:50:06*Llorean hasn't looked into it though, so he's kinda guessing.
09:51:18tdtookemy first idea was very bad, i've been up too long, i'm poking around in tagcache.c now
09:51:31tdtookei think initially i must've been trying to come up with the most complicated way possible to do this
09:51:56JdGordonpixelma: petur: any complaints about having the last few trig durations in seconds instead of minutes? (it goes 0s, 1s, 2s, ... 30s, 1min, 2min, 5min, 10min)
09:52:18*petur never used trig
09:52:28JdGordonmaybe we could add a flag to make time values normal
09:52:48 Join davina [0] (
09:58:14linuxstbtdtooke: I would simply add the line "if (!id3->grouping) { id3->grouping = id3->artist; }" to the end of the get_metadata() function in apps/metadata.c
09:58:36linuxstb(and probably the same thing to somewhere for hwcodec...)
09:58:43 Quit tdtooke ("sleep")
10:00:05 Quit Shaid (Read error: 110 (Connection timed out))
10:04:14 Quit cool_walking_ (Read error: 104 (Connection reset by peer))
10:04:22 Join cool_walking_ [0] (
10:06:30 Join cool_walking__ [0] (
10:24:34*JdGordon has no idea how to fix these trigger settings...
10:24:40 Quit cool_walking_ (Read error: 113 (No route to host))
10:24:55JdGordonthe global_Settings value is being used nastily
10:25:33JdGordon<0 is when its in dB mode, >=0 in the other mode
10:25:51JdGordondunno wtf to do here for the setting screen
10:26:37 Nick cool_walking__ is now known as cool_walking_ (
10:27:03 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
10:27:04pondlifeReal there a Linux command to do find/replace on multiple files? (Or rather, a Cygwin command...)
10:27:26gevaertssed ?
10:27:38JdGordonor awk?
10:27:51*JdGordon has no idea how to use either though
10:31:54B4gdersed, awk, perl
10:31:58cool_walking_sed -i 's/from/to/g' files_to_change
10:32:59gevaertsOr, if you need / in from or to, change the syntax-/ to something else
10:33:40*gevaerts challenges people to come up with an even less understandable explanation of this feature
10:38:01GodEateror escape the / with a \
10:38:21GodEateris it a competition to see how confused we can make pondlife?
10:38:29pondlifeIt's working
10:38:35pondlifeI mean the confusion
10:38:46GodEaterI assumed that's what you meant ;)
10:38:49pondlifeOK, how can I sed recursively?
10:39:20pondlifei.e. s/static struct plugin_api*/static const struct plugin_api*/g on all Rockbox
10:39:21*GodEater waits for someone to bring in either xargs, find, or preferably both
10:39:25gevaertssed -i 's/from/to/g' `find . -type f -other-find-args-that-you-want`
10:39:45*pondlife runs off to find a Windows app .
10:39:46gevaertsit's easy :)
10:39:49B4gdernah, that breaks on too many args
10:39:50GodEateronly half points to gevaerts
10:39:57*petur mentions he once wrote a windows commandline tool to find and execute
10:40:31GodEatergevaerts: I think your line there is likely to choke on the number of files in the rockbox tree too
10:40:32*gevaerts recommends as a sed tutorial :)
10:40:55 Join OlivierBorowski [0] (
10:40:56pondlifeSo no way to sed recursively?
10:41:09B4gderpondlife: find . -type f | xargs sed ....
10:41:23GodEaterfull marks to B4gder
10:41:28pondlifeObvious, isn't it...
10:41:36*GodEater awards B4gder the t-shirt
10:41:39gevaertsBut be careful about the .svn directories
10:41:58B4gderright, the .svn dirs...
10:42:18pondlifefind *.c ... ?
10:42:25B4gder'find . -name "*.[ch]" | grep -v '\.svn' | xargs sed ....
10:42:38B4gder(without the bonus ')
10:42:54*GodEater thinks that deserves some sort of hat too
10:43:24*gevaerts still prefers the find in a subshell way if there aren't too many results
10:43:26B4gderusing a *nix shell is all about using pipes
10:43:42scorchepondlife: you should use pv too! ;)
10:43:42GodEaterand the occasional fifo to really impress people
10:44:11pondlifeI'm happy with pipes, but I don't know any commands.
10:44:23GodEaterwell now you know four
10:44:31pondlifexargs was the missing link, I think
10:44:48peturfyi, the windows tool I wrote was
10:44:53pondlifeI'm not sure if it's worked - not much feedback is there?
10:44:57GodEaterwe'll expect a report written on what you've learned by lunchtime okay pondlife?
10:45:08*gevaerts has a standard shell exam question. Reverse the order of lines in a text file using only standard commands
10:45:09B4gderpondlife: svn diff should tell you ;-)
10:45:10*pondlife passes GodEater a blank sheet of A3
10:45:10GodEaterno feedback is a good thing
10:45:15scorchepv -cN find . -name "*.[ch]" | pv -cN grep -v '\.svn' | pv -cN xargs sed
10:45:23pondlifeI'm far away from SVN sadly
10:45:38pondlifeAs in, I've already done lots of stuff manually
10:45:46*pondlife gives up
10:45:47scorchepondlife: pv will show you progress :)
10:46:00pondlifeNot on Cygwin....
10:46:20*scorche wonders if pv is packaged for cygwin
10:46:28pondlifeok.. no
10:46:37pondlifenow how do I undo that sed stuff ? :)
10:46:48 Join austriancoder [0] (n=cgm6393@rockbox/developer/austriancoder)
10:46:57gevaertswith sed 's/to/from/' ?
10:47:07pondlifeJust joking
10:47:08amiconnThe plugin api struct is already defined const. It's just not declared const so as to avoid changing all plugins
10:47:57GodEatertrust amiconn to show up late with a good reason for not confusing pondlife in the first place ;)
10:48:11amiconnIn fact it's decalred const (would give a warning otherwise), but it's cast when calling the plugins
10:48:38pondlifeI've changed all plugins anyway for const void *parameter, so thought I might as well do the API too.
10:49:28pondlifeThis being part of a bigger tree-of-constiness, of course...
10:50:08amiconnWell, nothing needs changing in the core apart from plugin_.c line 722 (remove the cast) then
10:50:38pondlifeHence my attempt to sed the plugins into submission...
10:51:03*amiconn would just do that kind of thing manually
10:51:06gevaertsWait, was there a * in the 'from' part ?
10:51:37*pondlife is doing it manually, only 70 to go
10:51:47amiconnSearch and replace can break for many reasons here (varying whitespace and line breaks)
10:52:11gevaertsthen you would need more magic
10:52:56GodEaterreplace your spaces with \W (or \w, I can never remember which)
10:54:24gevaertsAnd of course escape the *s
10:55:18 Join ch4os_ [0] (n=ch4os@unaffiliated/ch4os/x-059673)
10:55:55 Join PaulJam [0] (
10:56:44 Quit cool_walking_ (Remote closed the connection)
10:57:52 Join dabujo [0] (
10:58:39 Join J-23 [0] (
11:01:13markunhi J-23
11:02:09 Quit linuxstb (Read error: 110 (Connection timed out))
11:02:39 Join crash91 [0] (n=crash@
11:04:08 Quit ch4os (Read error: 110 (Connection timed out))
11:04:51 Quit voltagex (Read error: 110 (Connection timed out))
11:05:47 Quit goffa_ (Remote closed the connection)
11:05:57 Join goffa [0] (n=goffa@
11:08:18 Part crash91 ("Ex-Chat")
11:08:30 Quit ch4os_ ("leaving")
11:09:07 Join ch4os_ [0] (n=ch4os@unaffiliated/ch4os/x-059673)
11:09:26 Nick ch4os_ is now known as ch4os (n=ch4os@unaffiliated/ch4os/x-059673)
11:11:24 Join mak3 [0] (
11:11:31mak3hi all
11:12:06mak3my ipod video is broken so i need a new player.
11:12:24scorchehow is it broken, may i ask?
11:13:32mak3i wanted now buy a ipod classic as replacement.
11:14:13GodEaterI wouldn't do that if you want to run Rockbox
11:14:15mak3looked at rockbox to show if its supported, doesnt seems to have a port for it.
11:14:18GodEaterunless you're planning on doing the port
11:14:34mak3yeah i saw that.
11:14:40mak3scorche: physical. ;)
11:15:15mak3well, i need to find another replacement, any recommends?
11:15:25J-23Sansa E200
11:15:44mak3should work good with rockbox.
11:15:45J-23or new iPod Video ;)
11:15:59GodEateryeah, I'd get another of those, of a Gigabeat F
11:16:08GodEaterthe Sansa won't stack up against the video in terms of storage
11:17:32mak3hm gigabeat F seems to be big.
11:17:36mak3should be not too big.
11:18:28GodEaterit's nearly identical to the video in terms of size
11:19:04mak3ok, then this is a nano? ;)
11:19:19GodEaterno, that's a mini
11:19:30mak3hmm ok mini
11:20:42gevaertsthe F is thicker than the video
11:20:55GodEaterhence "nearly" ;)
11:21:00GodEaterit's lighter though
11:24:52 Join Falafel [0] (
11:24:53mak3hm it seems that no online shop sells it in switzerland. :/
11:25:35mak3even google search no useful results.
11:26:03mak3is it that old? or rare?
11:30:32mak34looking for sansa e2000 now.
11:31:08mak3J-23: ipod classic seems to be a lot cheaper than ipod video... :/
11:33:07mak3i could repair the ipod video 60gb but nearly for the same amount as a new ipod classic 80gb
11:33:52scorchemak3: we dont talk about the classic here...this is a channel about rockbox
11:34:14scorche(we dont talk about it in that context at least)
11:34:17gevaertsmak3: none of the currently supported rockbox targets is built any more, so unless you're lucky you need to buy refurbished or second hand
11:34:37J-23iPod Classic isn't supported by Rockbox :/
11:34:56mak3yeah thats why i ask here for another replacement...
11:35:04gevaertsAnd when looking for an e200, make sure you get a v1. The port to v2 models isn't done yet (although it now seems to have started at least)
11:35:42gevaertsHave you looked at ?
11:35:46mak3gevaerts: k thx good to known.
11:35:55 Quit amiconn (Nick collision from services.)
11:36:02 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
11:38:08 Join Jayden [0] (
11:38:13 Join crash91 [0] (n=crash@
11:38:14JaydenHello, anyone here?
11:38:39 Quit crash91 (Remote closed the connection)
11:39:23*gevaerts isn't here
11:39:47*Jayden nods.
11:39:49JaydenOk then.
11:40:07 Join n1s [0] (
11:40:08gevaertsIf you have a question, just ask (as stated in the guidelines)
11:40:27***Saving seen data "./dancer.seen"
11:40:28JaydenAlright then, but I like to make sure someone's here first generally
11:41:15gevaertsThere are a lot of people around, but most of them aren't looking at their irc window all the time
11:41:23JaydenOk, here you go: about a week ago, my rockbox.ipod file spontaneously erased itself from my iPod, making it unbootable into rockbox as well as the apple firmware. Whenever I try to get it to load into the apple firmware, it stalls, my comp won't recognize either version, even in disk mode.
11:41:30*JdGordon knows what he is going to do about these bloody threshold settings and hopes it doesnt annoy too many people :/
11:41:33JaydenMy question do I fix this
11:42:30n1sJayden: is it recognised in emergency disk mode?
11:43:31JaydenWould that be different than regular disk mode?
11:43:32 Quit nuonguy ("This computer has gone to sleep")
11:44:03n1sthe one you get from rebooting with select+menu and then holding select+play
11:44:18JaydenOh yeah, that's the one it doesn't recognize it
11:44:24JaydenWell I mean my comp recognizes it
11:44:27JaydenBut it won't open it
11:44:34JaydenAnd I've tried 3 diff comps
11:44:40JaydenTo see if they could access it
11:44:48*gevaerts recommends checking the filesystem
11:44:52JaydenIt DOES detect that I plugged it in there and such
11:45:01n1sthat's kind of different... try fdisk and see if it can read the partition table
11:45:32n1s(if gevaerts' suggestion doesn't work out)
11:45:42gevaertsI would expect this to be either a corrupted filesystem or a broken disk
11:45:42JaydenLemme see here. Bbs.
11:45:43n1sand if all else fails, restore with itunes
11:45:50JaydenI don't know why it would be a broken disk
11:46:11n1sdisks are pretty good at breaking for no apparent reason :/
11:47:40gevaertsTry chkdsk (or whatever it's called on windows) or fsck first
11:51:08JaydenIn progress
11:51:59 Join thegeek_ [0] (
11:57:09Jayden Brb
11:57:26 Quit Jayden (Read error: 104 (Connection reset by peer))
11:58:25*pondlife crosses fingers and hopes for the entire front page
11:59:04 Nick fxb is now known as fxb__ (
12:00:33gevaertsimpressive :)
12:00:52*gevaerts assumes that pondlife also hopes for green
12:01:00pondlifeCould be an impressive high score too if I get a single red
12:01:06pondlifeI'll blame sed, of course...
12:01:10 Part J-23
12:02:05 Join J-23 [0] (
12:02:05pondlifeNow, if I could only get gui_sync_list to take a const, I could constify most of the playlist code (which is my aim).
12:02:47*n1s hands pondlife the const police hat, badge and baton :D
12:03:13pondlifeNah, it's not fun - you can keep them for now :)
12:03:41 Join Jayden [0] (
12:04:01pondlifeThe static police at least get green deltas to keep them interested.
12:06:27 Quit J-23 (Read error: 104 (Connection reset by peer))
12:06:38 Join J-23 [0] (
12:07:35*pondlife borrows the baton from n1s, and whacks the build server.
12:08:10gevaertsWith the amount of code you changed, I'm not surprised at the build taking much longer than usual
12:08:34pondlifeI thought it built from clean anyway.
12:08:45gevaertsDoesn't it use ccache ?
12:08:51pondlifeAh, true
12:09:07pondlifeCygwin has no ccache... :/
12:09:33peturI thought it had but was actually slower
12:09:53pondlifeWouldn't surprise me at all...
12:21:31 Quit goffa (Read error: 110 (Connection timed out))
12:21:38*gevaerts thinks that the build time now starts to look suspicious
12:21:48 Join SC [0] (n=3a6a56ee@gateway/web/cgi-irc/
12:23:45 Join ikari [0] (n=53be794d@gateway/web/cgi-irc/
12:24:53ikarihi, my wps "style" is not changing as i choose different .wps files and themse.. can you tell me why?
12:25:42gevaertsThe most likely cause is thatthe wps files contain errors
12:25:55 Quit kharo (Read error: 110 (Connection timed out))
12:26:07ikarithis happends with lots all wps files :(
12:26:22pixelmaor they use a feature you'd need a patch(ed build) for
12:26:59JdGordonOH CRAP! peakmeters dont play nice unless they are redrawn regularly
12:27:08ikarii'll search for patches. is there a need to restart rockbox after changing wps?
12:27:55ikarithank you very much guys
12:27:57pixelmaikari: no. What themes/WPSs do you try and on which player?
12:28:25ikariipod 5.5 - themes: crashBox, jClix and ubuntu (linux user :) )
12:30:03 Quit SC ("CGI:IRC")
12:31:46ikarit is strange since "ubuntu" page says: This WPS works with the standard build and does not require any patches.
12:32:25 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
12:33:22PaulJammight br that the ubuntu theme worked when the WPS pharser was less strict on errors
12:33:23 Quit ikari ("CGI:IRC (EOF)")
12:33:57GodEaterikari: unfortunately theme authors forget to put the date or revision of the build they tested them with
12:34:10GodEaterwhich means that "works with the standard build" doesnt' help you much
12:34:28GodEaterI'm pretty sure jClix needs an unofficial build to work with though
12:35:19gevaertsSomewhere in april wps loading was made more strict, so incorrect wpses that loaded before stopped working
12:36:01gevaertsubuntu.wps can't find stop.bmp
12:36:48 Join mib_39pm2v [0] (i=86ceaf40@gateway/web/ajax/
12:36:51 Part mib_39pm2v
12:37:12 Join BigBambi [0] (i=86ceaf40@gateway/web/ajax/
12:38:30 Quit desowin ("KVIrc 4.0.0 Insomnia")
12:39:19pondlifeBagder: You able to kick the build server?
12:41:13B4gderthese new problems are really annoying and weird
12:41:23B4gderthe build server just freezes occaionally
12:41:56pondlifeWhat? The whole machine locks?
12:42:06pondlifeOr the build script?
12:42:08gevaertsDid you use the playlist code from a few months ago in the build server ?
12:42:10B4gderthe build master process freezes
12:42:38pondlifeIt always seems to be my (rare) commits that do it ;)
12:42:56B4gderit seems to be a deadlock in the threads code or similar
12:44:41B4gderstrace shows a call to futex() and nothing else
12:48:17 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
12:48:22pondlifeCan you put a timeout on futex() ?
12:48:24 Quit mcuelenaere (Remote closed the connection)
12:48:30 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
12:48:51 Quit K4rP4D ("Leaving")
12:48:51B4gderthe futex is just what's used under the hood, they are "simple" locks in the perl code
12:49:00B4gderand no, they can't have timeouts
12:49:04 Join K4rP4D [0] (n=KrPD@unaffiliated/krpd)
12:49:05pondlifeSo, what's the lock for?
12:49:13B4gderlocking shared data
12:49:21B4gderthat each build server thread touches
12:50:01pondlifeSo one thread might have hung somehow with the futex upped (or whatever the terminology is)?
12:50:01 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
12:50:04B4gderI need to review the code and clean it up to see if I can figure out why this happens
12:50:13B4gderyes, something like that
12:51:04 Quit K4rP4D (Client Quit)
12:51:05amiconnJdGordon: Of course, that's why there even is a special preakmeter-redraw-while-waiting function
12:51:24 Join K4rP4D [0] (n=KrPD@unaffiliated/krpd)
12:51:35pondlifeHmm, this new build isn't looking too good either...
12:53:39pondlifeAll of the current builds are reporting r17492, so I guess that means it succeeded on the targets...
12:54:31JdGordonamiconn: arg, yeah, but i totaly forgot about that :p any ideas for it?
12:55:13B4gderand it hung again
12:55:26B4gdera little bit annoying to say the least
12:55:47B4gderProcess 8973 attached - interrupt to quit
12:55:47B4gderfutex(0xf22eebd8, FUTEX_WAIT, 9539, NULL <unfinished ...>
12:55:55 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
12:59:02 Join MTee [0] (n=MT@
12:59:28*MTee says hello !
13:00:46 Join goffa [0] (n=goffa@
13:00:51*JdGordon thinks about totally abusing the thread system to keep redrawing the peakmeter on this screen in a seperate thread way higher than the ui one...
13:01:17 Part J-23
13:02:56 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
13:07:36markunhi MTee
13:07:48 Quit chrisjs169 (Read error: 113 (No route to host))
13:07:56B4gderdamn this is serious
13:08:12pondlifeWhat did I do? :/
13:08:36B4gderthe same freeze happens over and over now
13:08:41B4gderit just can't complete
13:08:55pondlifeWhat's holding the futex? Another process perhaps?
13:09:07pondlifeWhat updates the website on completion?
13:09:10B4gderit's a multi-threaded program
13:09:23B4gderthe script that continues after this is done
13:09:40pondlifeDid you not kill the entire process though?
13:09:43B4gderI can see this happens when the last thread is done
13:11:17B4gderit is probably the waiting for all threads to die that somehow locks up
13:12:58B4gder...starting over
13:13:30B4gderno, killing script
13:13:30 Quit mcuelenaere ()
13:13:52B4gderwhich then detects the broken situation and then starts another build round
13:14:00pondlifeheh, I was hoping to use that line about "Only Windows needs reboots."... I guess I can anyway :)
13:14:14pondlifeAh, a parent script?
13:16:34B4gderthe futex() conclusion was a bit premature
13:16:47B4gderit seems the "root" pid always hangs on that until the child threads are done
13:20:24 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
13:26:49MTeeis rockbox planning to participate in gsoc next year or not ?
13:27:00pondlifeB4gder: Hmm, still hanging?
13:27:37B4gderMTee: probably we are
13:27:59B4gderif they run it and they accept us and we have mentors volunteering
13:28:34MTeeis it rockbox's first time in gsoc ?
13:28:44 Join Thundercloud [0] (
13:28:58BigBambiMTee: last year and this year
13:30:01MTeegreat .. for the development of rockbox, do I have to learn assembly ? I already know C and am going to delve into embedded programming this summer
13:30:16BigBambiOnly for speed critical areas
13:30:21B4gderMTee: for most things C is fine enough
13:30:27BigBambiMost is in C
13:30:33 Join Bensawsome [0] (n=upirc@
13:31:31MTeeBigBambi, B4gder : Thanks a lot :)
13:33:32 Quit K4rP4D (Read error: 104 (Connection reset by peer))
13:33:42Nico_PMTee: you don't need to wait for the next SoC to contribute patches ;)
13:34:12gevaertsActually, starting to contribute now will probably increase your chances
13:34:15 Part Bensawsome
13:34:32 Join K4rP4D [0] (n=KrPD@unaffiliated/krpd)
13:35:34 Quit Jayden (Read error: 104 (Connection reset by peer))
13:36:16MTeeNico_P, gevaerts : Yeah, I've already planned to do so, but my finals won't end till 12/6 :(
13:39:10 Quit BigBambi (" ajax IRC Client")
13:39:51 Join BigBambi [0] (i=86ceaf40@rockbox/staff/BigBambi)
13:40:10 Quit [CBR]Unspoken|w (Remote closed the connection)
13:40:31***Saving seen data "./dancer.seen"
13:41:58B4gderit seems there's one thread that hangs on a read() which prevents them all from terminating and thus continuing
13:42:53*gevaerts sends that thread to school, so it can learn to read properly
13:43:23B4gderif I only could figure out why
13:44:00 Quit Zarggg (Read error: 104 (Connection reset by peer))
13:44:17 Join Zarggg [0] (
13:46:12pondlifeB4gder: What's it reading? Response from a server?
13:46:41B4gderI don't know what, it seems to be something internal that the script isn't doing
13:46:51B4gderthe freeze happens after all server threads are complete
13:48:27B4gderat line 586
13:49:47 Quit Nico_P (Remote closed the connection)
13:50:38 Quit bughunter2 ("Leaving.")
13:51:11B4gderbut the script hasn't changed in a long time
13:51:22B4gderand now I get the problem 100% repeatable
13:51:32 Join bughunter2 [0] (
13:51:37B4gderso atm we just can't get the build table done
13:52:00*gevaerts wonders
13:52:01PaulJampondlife: concerning your latest changes, could you check if test_codec.c needs additional changes i get a warning when enabling it (test_codec.c:767: warning: passing arg 1 of `test_track' discards qualifiers from pointer target type).
13:52:19gevaertsDid anything change in the environment ? (kernel, perl version,...)
13:52:41pondlifePaulJam - yes, that ill need updating, I should probably try building all of the test_ plugins.
13:53:55B4gderit uses debian testing with 5.8.8-12 which has been around since 2007
13:54:02B4gderperl version
13:56:36*gevaerts doesn't know much about perl unfortunately
13:57:02 Join Shaid [0] (
13:57:29 Join ali_as_ [0] (
13:57:47 Quit MTee (Read error: 110 (Connection timed out))
13:59:13*pondlife thinks the test_ plugins should be built normally, even if we then delete them from the release...
13:59:54preglowme talks to his sd card
14:00:14preglowand also at the same time learns how important sd card states are
14:01:46preglowB4gder: that damned flash is so troublesome i just opted to start out with sd card support :>
14:02:27gevaertsWho needs flash if you have sd ?
14:02:46preglowglad you see my point :P
14:03:00preglowtalking to the flash is piece of cake, understanding telechips' translation layer is not
14:03:09preglowand sd cards deal with stuff like that in their controller
14:03:22pondlifePaulJam: Fixed test_codec.c now committed.
14:03:34PaulJamthank you
14:04:03 Quit CaptainSquid ("Miranda IM!")
14:05:22 Join nplus [0] (n=nplus@141.25.Globcom.Net)
14:07:39 Quit K4rP4D (Remote closed the connection)
14:09:05 Join K4rP4D [0] (n=KrPD@unaffiliated/krpd)
14:15:51*B4gder sighs
14:17:22JdGordontime to completly rebuild the build system?
14:18:13B4gderyeps, in 20 minutes before I leave for home
14:18:38 Join hd [0] (i=jd@unaffiliated/helldragon)
14:19:14 Quit HellDragon (Nick collision from services.)
14:19:16 Nick hd is now known as HellDragon (i=jd@unaffiliated/helldragon)
14:19:32pondlifeB4gder: Did you try a reboot?
14:19:41*preglow gets his first crc check failed
14:20:26B4gderpondlife: no
14:20:43B4gderI don't see how a reboot will help
14:20:58 Quit K4rP4D ("Leaving")
14:21:07pondlifeGot to be worth a shot, no? Or would you rather preserve the situation so you can analyse further as you get time?
14:21:24JdGordonreboots dont fix problem in *nix
14:21:32 Join swimmer [0] (
14:21:41B4gderI don't think a reboot will help, but sure I might give it a desperate attempt later on
14:22:06swimmerhi - anyone knows the battery capacity of an Ipod Video 80GB?
14:22:21swimmerI mean the settings in system ...
14:22:47GodEaterswimmer: it won't make much difference, the calibration isn't done for the ipod yet
14:23:16swimmerGodEater: thank you ... atm I get some unrealistic 22h if fully loaded ;-)
14:23:33swimmerwas just wondering which setting would be realistic for the calculation
14:23:54swimmerbut if it's not done I don't mind any further ;-)
14:23:55 Quit ali_as (Connection timed out)
14:24:02 Part austriancoder
14:24:56GodEaterI just go on the battery % left, and work from a very rough guide that battery life is about 12 hours
14:25:20swimmeryep - I'll do the same
14:25:41JdGordonif you just want to reset the value to the default you can hold menu in the setting screen for it
14:25:46JdGordonthat should set it back
14:25:47swimmerdo you know by any chance whether "Autochange Directories" is working?
14:25:59JdGordonit does... sometimes..
14:26:01*gevaerts suspects desperate attempts
14:27:33GodEaterworks for me
14:28:05GodEaterswimmer: you have to have repeat set to "Off" for it to work though
14:28:44JdGordonand be playing a directory instead of a database search or playlist file (.m3u)
14:28:51GodEaterthat too
14:32:49swimmergnaaaa - stupid me ... I had not yet changed the (default?) setting of 'Repeat=All' :-/
14:32:57 Join K4rP4D [0] (n=KrPD@unaffiliated/krpd)
14:33:00swimmerthanks for kicking my a** in this direction ;-)
14:33:00GodEateryep that's the default
14:33:51swimmercan't tell you how happy I am with rockbox 8^D
14:34:07swimmerit made me even buy an ipod ;-)
14:36:29*JdGordon too :p
14:36:44JdGordonexcept im a moron and broke it :'(
14:37:05swimmereasy to break?
14:37:15GodEaterthey are if you drop kick them
14:37:22swimmerwhat kind of earphones do you guys use btw?
14:37:30swimmerGodEater :)
14:37:31GodEaterShure E2C at the moment
14:37:37JdGordonnot usually...broke a ribbon cable inside it.. need to wait a week for a bloody replacement now :(
14:37:54swimmerthose are in-ear aren't they?
14:38:00GodEaterthey are
14:38:05GodEaterI wouldn't use anything else
14:38:11GodEateryou don't get the isolation from anything else
14:38:38swimmerit's not necessary - I'm afraid for my ears with those devilish things ;-)
14:38:51GodEaterafraid for your ears
14:39:18JdGordonI changed to in ears at the begingi of the year and love them... I can have the volume MUCH quieter even in loud areas which is great
14:39:32GodEaterthat's why I love them too
14:39:41GodEaterno need for ear shattering volume with them
14:40:10swimmerthere were some tests in NL which said that in-ears are by far the most damaging headphones ...
14:40:19swimmerand I read it
14:40:26GodEaterI guess they might be if you rammed the volume up through the roof
14:40:27swimmer's hard to get them in ...
14:40:32GodEaterbut there's no need to
14:40:44JdGordonthey are a bit bass-y though, and sound is wierd when I talk with them in.. but other than that, love em
14:40:46GodEaterhard to get them in?!?!
14:41:03*GodEater has never struggled with that
14:41:08GodEaterwhat sort of weird ears you got ?
14:41:17swimmerGodEater: just what I heard and read in 1 day browsing the internet ..
14:41:28*gevaerts suggests not to try using out-of-ear headphones in-ear
14:41:31GodEaterswimmer: well I've been using them since 1999, and I've not damaged my ears yet...
14:41:38swimmerGodEater: excuse my ignorance I'm still collecting information
14:41:40GodEatergevaerts: hahahaha - amen to that
14:41:55swimmergevaerts :)
14:42:18GodEaterswimmer: I suggest getting a cheapish pair and trying them out
14:42:29GodEatereven Sony MDR-EX71s are pretty good
14:43:05swimmerGodEater: no sennheisers?
14:43:09GodEaterside splittingly funny pondlife
14:43:23GodEatersennheisers are good too - but tend not to be as cheap as the Sonys
14:43:26pondlifeBut poor timing - I was away at the critical moment
14:43:36GodEaterit depends on how much you're prepared to spend for a trial ;)
14:43:55GodEaterif you're super rich, get some Shure E5cs ;)
14:44:08*GodEater tried out Febs ones of those - and was suitably impressed
14:44:26GodEaternot going to drop £300 on headphones though, no matter how good they are
14:45:08swimmergood lord!
14:45:11JdGordonholy crap!
14:45:28JdGordonwhat are they? bloody gold plated?
14:45:34pondlife = £500 !!
14:45:45pondlife(with free delivery, natch)
14:45:59gevaertsDo they have built-in valve amplifiers ?
14:46:00GodEaterpondlife: I thought they were $500 - I must have misread
14:46:09swimmerI found the e2c already expensive ;-)
14:46:13pondlife£265 here -
14:46:25GodEaterthat's more like it
14:46:29GodEaterbut even so ;)
14:46:36markunswimmer: I have Creative EP630 in-ears. Quite cheap and not too bad. Comparable to the Sennheiser CX300.
14:46:51pondlifeWacky price variation, huh?
14:46:53GodEateryep - my g/f has those
14:47:21GodEaterpondlife: Febs' earphones were awesome though. The isolation was fantastic. I couldn't hear *anything* from the surrounding pub with them in, even with no music going.
14:47:33GodEaterwhich meant you could have the music playing very quietly indeed to enjoy it
14:47:33pondlifeNot sure I'd like that, to be honest
14:47:39 Part B4gder
14:47:46GodEaterhorse for courses.
14:47:50pondlifeI rather like knowing when I'm about to be run over
14:47:51 Join Bagderr [0] (n=daniel@gateway/web/cgi-irc/
14:48:31*GodEater uses his eyes for this function
14:48:45*gevaerts might prefer not to know in advance
14:49:29GodEateranother positive point for them then ;)
14:49:49pondlifegevaerts: Perhaps you could enable ROCKBOX_USB before you next cross the road...
14:49:49Bagderreep, the buildmaster completed
14:50:08pondlifeWoo, no yellow
14:50:09Bagderrand I'm off
14:50:12 Part Bagderr
14:50:27JdGordonpondlife: enabling it is the easy bit....
14:50:46pondlifeheh, but it presupposes the other issues are fixed.
14:50:54JdGordonisnt it only the sansas with issues now?
14:51:07pondlifeHigh speed got fixed on hubs etc?
14:51:18pondlifeOr was that just amiconn's hub that was playing up?
14:52:29gevaertsThat issue is still there, and it's not just amiconn's hub
14:52:44pondlifeThat's what I thought.
14:54:18*JdGordon decides the rec trigger menu wont be as nice as it could be... but still better (code and looks wise) than svn
14:54:27 Join kharo [0] (
14:56:30markunJdGordon: why not make it as nice as it could be?
14:57:08JdGordonwell.. it will be as nice as the current code will allow
14:58:18 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
15:00:00 Quit crwl (Read error: 110 (Connection timed out))
15:05:29 Join LambdaCalculus37 [0] (n=44a0430f@gateway/web/cgi-irc/
15:12:11swimmerok guys - thanks for all the info and the nice chat ;-)
15:12:56 Quit thegeek_ ("( :: NoNameScript 4.2 :: )")
15:13:11 Part swimmer
15:16:58 Quit Galois (Read error: 104 (Connection reset by peer))
15:19:10 Join Galois [0] (
15:24:35 Join obo [0] (n=522e52e0@gateway/web/cgi-irc/
15:25:51 Part LinusN
15:25:53oboBagder: you may want to disable my host for awhile - I don't know if it's causing problems with the build process or not, but dmesg is full of CPU soft lockups and backtraces
15:29:07gevaertsI seem to remember (but I can easily be wrong) that it was your server that had a gcc internal compiler error a few days ago (last weekend ?)
15:31:22 Join jgarvey [0] (
15:31:24oboNo idea - I wasn't around much last weekend
15:40:32***Saving seen data "./dancer.seen"
15:44:59 Join magmaniac [0] (
15:45:07 Join kugel [0] (n=kugel@unaffiliated/kugel)
15:45:12 Part magmaniac ("Leaving")
15:48:07pixelmaLambdaCalculus37: your editing the patch file directly wasn't very successful. No biggy because I could just take out the text and implant it but it took me a while to figure out why patching failed... ;)
15:48:32LambdaCalculus37pixelma: Which patch was that?
15:48:50pixelmathe description of the c200
15:49:24 Quit Shaid (Read error: 110 (Connection timed out))
15:49:51LambdaCalculus37Sorry about that. :)
15:55:36 Nick fxb__ is now known as fxb (
15:58:25 Quit goffa (Read error: 110 (Connection timed out))
16:11:30 Quit Ragnor (Nick collision from services.)
16:11:40 Join Ragnor [0] (
16:11:42pondlifeGodEater: According to TFL's wonderful website, I should get a bus from Aldgate to get to Wapping.
16:11:54pondlifeEven better, the suggested route from Tower Hill is :
16:13:05pondlifeSorry, this should be in -community, just mispasted...
16:28:45 Join MethoS- [0] (
16:29:57 Join MethoS_ [0] (
16:35:19 Join toffe82 [0] (
16:35:38JdGordonfor the rec builds which file gives the better bin estimate? rockbox.bin or rockbox.elf?
16:38:31JdGordonin that case... my patch is nearly -2k delta :)
16:39:13n1snice :)
16:39:28JdGordonwell.. 1.3k coz I cant do simple maths at 1am :p
16:41:35 Join goffa [0] (n=goffa@
16:49:21 Quit MethoS- (Read error: 113 (No route to host))
16:50:24 Join Spiffelajt [0] (
16:50:55SpiffelajtUhm, dunno if it should be asked in here.
16:51:05SpiffelajtBut I got a small problem with the installation of rockbox.
16:51:23LambdaCalculus37Ask your question.
16:51:42SpiffelajtWhen I press Complete Installation (3rd Gen).
16:51:52SpiffelajtStarting bootloader isntallation, searching for ipods
16:51:54Spiffelajtno ipods found
16:51:58Spiffelajtand It's connected.
16:52:18gevaertsWhat model exactly do you have ?
16:52:23LambdaCalculus373rd gen? You mean a 3rd generation iPod?
16:52:26SpiffelajtiPod Nano.
16:52:41*Spiffelajt nods
16:52:58LambdaCalculus37No, the 3rd gen nano does not work with Rockbox.
16:53:05Spiffelajtaw, :-/
16:53:49LambdaCalculus37The only nano Rockbox runs on is the 1st gen.
16:54:04SpiffelajtOK, thanks anyways.
16:54:08 Quit Spiffelajt (Client Quit)
16:56:29 Join Spiffelight [0] (
16:56:54Spiffelight:D is it possible that there is coming out a version that works with 3rd gen nano?
16:56:59Spiffelightor is my dreams crushed?
16:57:15scorche|shno one is working on it, so that is quite unlikely
16:57:30*Spiffelight melts
17:01:57 Join [CBR]Unspoken|w [0] (n=cbr@
17:02:07*JdGordon seems to have misplaced some drawing code :(
17:02:26 Join MethoS- [0] (
17:02:49*petur hopes JdGordon will put his progress on the tracker
17:03:05peturnot that I have any time tonight :/
17:03:17JdGordonyeah, i will...
17:03:23JdGordonalthough i tihnkim done for the night
17:04:22 Quit goffa (Read error: 110 (Connection timed out))
17:05:52 Quit MethoS_ (Read error: 110 (Connection timed out))
17:08:46 Quit obo ("CGI:IRC (Ping timeout)")
17:13:34JdGordonfound it :)
17:16:17JdGordonpetur: 8989
17:18:02 Join MethoS_ [0] (
17:18:30 Join goffa [0] (n=goffa@
17:20:56 Nick JdGordon is now known as JdGordon|zzz (i=jonno@rockbox/developer/JdGordon)
17:30:11 Quit amiconn (Nick collision from services.)
17:30:17 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
17:30:56 Join hcoal [0] (n=hcoal@unaffiliated/hcoal)
17:35:47 Quit hcoal ("Ciao")
17:38:01 Quit MethoS- (Connection timed out)
17:38:09 Join MethoS- [0] (
17:39:53 Quit petur ("bow for the religion of bad-min-ton")
17:40:36***Saving seen data "./dancer.seen"
17:40:50 Join nuonguy [0] (
17:41:05 Quit BigBambi (" ajax IRC Client")
17:42:47 Quit goffa (Read error: 110 (Connection timed out))
17:52:14 Quit MethoS_ (No route to host)
17:52:22 Join MethoS_ [0] (
17:59:06benimedoes anyone know of a part number or an appropriate replacement for the line out jack on an iriver h1x0? the plastic casing on mine cracked and I need to replace it :/
17:59:34LambdaCalculus37Have you tried Googling for "iriver Replacement parts"?
18:00:06benimeyes, most replacement parts consist of screen, battery, hard drive
18:00:15benimeor accessories
18:00:46benimeso I figure if anyone knew of a part number I could look up at digikey or mouser, that'd be great
18:06:17 Quit Mathiasdm ("Yuuw!")
18:07:01linuxstbbenime: Have you tried searching the forums at ?
18:07:42benimeyes, and the rockbox forums, closest I got was this
18:07:48 Join Mathiasdm [0] (
18:08:17benimesomeone suggested replacing their lineout with that, but then never followed up whether or not it worked, or whether or not they even went ahead with it
18:08:37 Quit MethoS- (Read error: 110 (Connection timed out))
18:08:54 Join MethoS- [0] (
18:12:05 Quit gevaerts ("work->home")
18:15:44 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
18:17:00 Quit amiconn (Nick collision from services.)
18:17:03 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
18:17:12 Quit mak3 (Read error: 104 (Connection reset by peer))
18:17:39 Part pondlife
18:30:36 Quit MethoS_ (Read error: 113 (No route to host))
18:30:36 Join [1]Falafel [0] (
18:30:56 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:31:20 Quit SirFunk (Read error: 104 (Connection reset by peer))
18:31:26 Quit LambdaCalculus37 (Read error: 104 (Connection reset by peer))
18:36:43 Join LambdaCalculus37 [0] (n=44a0430f@gateway/web/cgi-irc/
18:36:53*LambdaCalculus37 kicks the IRC web client
18:42:37 Join herrwaldo [0] (
18:46:45 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
18:47:45 Nick fxb is now known as fxb__ (
18:48:00 Quit Falafel (Read error: 110 (Connection timed out))
18:48:01 Nick [1]Falafel is now known as Falafel (
18:48:34 Join Nibbler [0] (
18:49:48 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:53:05 Join miepchen^schlaf [0] (
19:02:07 Join dabujo_ [0] (
19:03:18 Join peddy [0] (
19:07:11 Quit mcuelenaere (Read error: 60 (Operation timed out))
19:08:45 Quit dabujo_ (Read error: 104 (Connection reset by peer))
19:09:34 Quit peddy ("Miranda IM! Smaller, Faster, Easier.")
19:19:36 Quit dabujo (Connection timed out)
19:25:02 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
19:28:05 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:28:37 Join fehmicans [0] (n=canavar@
19:28:44 Quit homielowe (Read error: 104 (Connection reset by peer))
19:29:16 Join homielowe [0] (
19:32:26 Quit DerDome ("Leaving.")
19:35:59 Join DerDome [0] (
19:36:19 Quit DerDome (Client Quit)
19:36:33 Join DerDome [0] (
19:37:19 Join MethoS_ [0] (n=clemens@
19:37:57 Join culture [0] (n=sam@
19:38:56 Quit inversions (Connection timed out)
19:40:39***Saving seen data "./dancer.seen"
19:42:30 Quit Horscht ("")
19:45:13 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:52:09 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
19:54:27 Quit MethoS- (Read error: 110 (Connection timed out))
19:54:53 Quit Nibbler (Read error: 110 (Connection timed out))
19:55:38 Join Nibbler [0] (
20:00:59 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
20:01:18 Quit homielowe (Read error: 110 (Connection timed out))
20:01:53 Quit MethoS_ (Connection timed out)
20:03:52 Quit benime ()
20:04:37 Join bertrik [0] (
20:04:38 Join MTee [0] (n=MT@
20:04:49 Join einhirn_ [0] (
20:07:24 Quit XavierGr ()
20:09:16 Quit perrikwp (" ajax IRC Client")
20:09:24 Join dabujo [0] (
20:09:33 Join bluebroth3r [0] (
20:09:44 Quit bluebrother (Nick collision from services.)
20:09:47 Nick bluebroth3r is now known as bluebrother (n=dom@rockbox/staff/bluebrother)
20:11:52 Join spiorf [0] (
20:15:29 Join c0uch [0] (
20:18:02c0uchhey, i was wondering if there are any stated goals for rockbox supporting car interfaces for ipods? specifically i just got a Panasonic CQ-C7403U which has an ipod hookup.
20:19:09bluebrotherc0uch: no. But I imagine that once AAP is working most car interfaces will do
20:19:36c0uchsorry, what's aap? (or better yet, do you know where i can read up on it?)
20:19:47 Join homielowe [0] (
20:19:47 Join dan_a [0] (n=dan-mirc@
20:19:55 Join perrikwp [0] (i=98212bb4@gateway/web/ajax/
20:19:59 Join Lear [0] (i=chatzill@rockbox/developer/lear)
20:20:24linuxstbc0uch: The Apple Accessory Protocol - the standard way ipods talk to things
20:20:58c0uchah alright. thanks for the help =)
20:32:32 Quit Horscht (Nick collision from services.)
20:32:44 Quit c0uch ()
20:32:50 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
20:37:31 Join lee-qid [0] (
20:38:08 Quit linuxstb (Read error: 110 (Connection timed out))
20:44:52 Quit Horscht ("I am root. If you see me laughing, you better have a backup")
20:44:59 Join desowin [0] (n=desowin@atheme/member/desowin)
20:45:20 Quit nuonguy (Remote closed the connection)
20:46:55amiconnjhMikeS: Eh, unaligned stores? Is that an arm v6 special?
20:47:21 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
21:00:10 Join FOAD_ [0] (
21:01:01preglowi believe so
21:01:43preglowarm cpus aren't very minimalistic these days
21:06:14 Join inigomontoya [0] (n=621cdcfd@gateway/web/cgi-irc/
21:06:58 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
21:09:34jhMikeSamiconn: yes, and choosing the endianness of loads and stores on the fly
21:11:40 Quit FOAD (Read error: 110 (Connection timed out))
21:11:40 Nick FOAD_ is now known as FOAD (
21:14:59bluebrotheraustriancoder: for the logs, bass and treble works fine now on m:robe
21:18:15 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
21:18:31amiconnjhMikeS: I was wondering because armv4 and amv5 (optionally) also support "unaligned" accesses
21:19:13amiconn...but unlike true unaligned accesses, those accesses are rounded down to proper alignment, and then the loaded data is rotated
21:19:22amiconnNot quite what you'd expect...
21:19:25jhMikeSamiconn: It is more efficient for stream parsing than byte reading and shifting
21:20:02jhMikeSon arm v6 it it's behaves like m68k
21:21:22 Join nedd1 [0] (n=john@nat/sun/x-898301e79ea3ea91)
21:21:23amiconnIf it behaves like that then it's what I'd expect, but I fail to see how this is more efficient (unless arm manages to do this without additional memory cycles). Otherwise it just simplifies things which aren't time critical
21:21:27jhMikeSbut the bit for must be set in the SCC or else it does as you described
21:21:49preglowi believe armv6 uses one extra cycle for unaligned loads
21:22:11jhMikeSI've found it to be faster if the CPU handles it through bus cycles and not extra instructions
21:22:56jhMikeSeven on coldfire it was faster just to read unaligned than assemble words from pieces in software
21:22:57preglowlinuxstb: do you know what relationship the cpu clock and bus clock need to have on telechips cpus?
21:23:43 Join raph1 [0] (
21:24:16jhMikeSof course it's still better to produce alignment if possible in any case
21:24:40preglowcode still has to be aligned, yeah?
21:24:42linuxstbpreglow: No, I haven't got that far yet. I've just set up the clocks in the same way as the OF does.
21:25:05jhMikeSpreglow: indeed
21:25:17*preglow can't get the crc to check out on his sd card command replies :/
21:26:13*bertrik is only familiar with SPI mode for sd cards where you can get away with two hard-coded CRCs
21:27:51jhMikeSobviously v6 has DSP aspirations so they added the stuff that's helps out. all the swapxx routines we use only need one instruction each.
21:28:08 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
21:29:20amiconnjhMikeS: On coldfire it's massively faster to read aligned and assemble the data if you're dealing with sdram
21:30:06*amiconn thinks that with all these extra core gimmicks adding to the insane clock speed of the beast, he should *demand* realtime ape -c5000 support
21:31:11*preglow couldn't care less about ape
21:31:20preglowadd flac encoding instead :>
21:31:42 Quit Mouser_X (Nick collision from services.)
21:31:45jhMikeSamiconn: The SPC codec wasn't aggreeing so much (where it reads two 16-bit samples and interpolates). It turned out faster to load unaligned from the BRR cache and diassemble the data into words via emac.
21:31:51 Join Mouser_X [0] (
21:32:08raph1can someone tell me, what the 'button-codes' in 'firmware/target/coldfire/iaudio/button-target.h' means? the scancode is actually mentioned in the X5. I'm asking, because I need to remap some keys (down-key is dead.. ;-))..
21:32:23jhMikeSwhat the fastest ape -c5000 benchmark?
21:33:10LambdaCalculus37The Gigabeat F, IIRC.
21:33:19LambdaCalculus37And even that couldn't handle it in realtime.
21:33:21jhMikeSElephants Dream can play well over 80fps at 528MHz without any armv6 optimizations :)
21:33:27bluebrotherraph1: you most likely will only need to adjust the button maps −− see apps/keymaps
21:34:28 Quit ompaul (Client Quit)
21:34:48*Llorean had his Gigabeast crash when trying to play a video, but it may have been due to data corruption.
21:34:52 Quit desowin ("KVIrc 4.0.0 Insomnia")
21:35:02LambdaCalculus37jhMikeS: linuxstb performed a codec test on the S:
21:35:08raph1bluebrother: loool :-D thanks! sometimes the solution would be so near...
21:38:02jhMikeSpreglow: 320x240. only problem was I found out later I left the SW eq enabled :p
21:38:05preglowno exatly awe-inspiring
21:38:12Lloreanpreglow: Note that's at half clock.
21:38:14preglowjhMikeS: sw eq for mpegplayer? :/
21:38:31preglowgigabeast has instructions that will benefit ape decoding as well
21:38:37jhMikeSpreglow: sure. all the DSP effects can be enabled or forced off there.
21:38:46preglowjhMikeS: in svn too?
21:39:09jhMikeSactually you can run it in PP targets and coldfire targets hardly slow down using the DSP fully.
21:39:32jhMikeSpreglow: yes, it's been setup that way for a few months already
21:39:42*preglow didn't notice that
21:39:45preglowbut that's nice
21:39:48preglowmake radio work that way as well :)
21:39:54 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
21:40:01 Quit spiorf (Remote closed the connection)
21:40:22*jhMikeS wants the hardware EQ UI worked out so support can be completed on the S as well
21:40:38amiconnjhMikeS: Is that data located in iram?
21:40:41***Saving seen data "./dancer.seen"
21:40:56jhMikeSamiconn: not the BRR cache - it's too big for that
21:40:58amiconnIf it's not, you should try to read whole lines and disassemble into 8 words...
21:41:54jhMikeSthe problem is that it has random alignment so reads that happen to be aligned suffer too if extra code exists
21:41:56 Join Buschel [0] (
21:42:11*shotofadds sees preglow has been playing with a D2 SD driver. how's that going?
21:42:53amiconnjhMikeS: Can byte alignment happen, or is it at least 16 bit aligned?
21:43:03 Quit Mouser_X (Nick collision from services.)
21:43:11 Join Mouser_X [0] (
21:43:25amiconnAnd is it just reading & processing, or does it involve writing (with different alignment)?
21:44:55jhMikeSamiconn: it's reading two 16-bit samples (16-bit aligned) and interpolating but no memory write
21:45:17jhMikeSstarting spc_dsp.c line 809
21:45:49preglowshotofadds: ok enough, some troubles
21:46:15preglowshotofadds: haven't gotten past init yet, card never claims it has powered up, and the reply for the command i use to find that out fails the crc check
21:46:22preglowshotofadds: but i can talk to it and get coherent replies
21:46:28jhMikeSalot of experiments were done on that code since it can be executed (average rate) 320K times/second
21:46:50jhMikeSer...256K times/s
21:47:11preglowjhMikeS: well, my take on hw eq is definitely that i don't want two separate screens for sw and hw eq. does gigabeast have a proper hw eq too? (wm chip)?
21:47:12 Join Makuseru [0] (
21:47:16Buschelhi folks
21:47:38 Quit Makuseru (Read error: 104 (Connection reset by peer))
21:47:44amiconnjhMikeS: Hmm, so it's just 2 samples and no longer streaks...
21:47:52Buschelany decisions or opinions about enlarging iram for PP5022/24?
21:48:10preglowBuschel: what's the issue?
21:48:10Buschelor about re-partitioning iram for other targets?
21:48:25preglowjust give the extra ram to codecs and be done with it :>
21:48:31jhMikeSamiconn: right
21:48:32Buschelpreglow: we do not use the 128KB of PP5022/24 −− only 96KB
21:48:50preglowi know, you're just wondering about allocation of the rest?
21:49:04*amiconn would still split the iram in 2 equal parts
21:49:15preglowamiconn: core/codec?
21:49:17amiconnBut it's not very important imo
21:49:22Buschelno, e.g. m5/x5 use 64KB for core and plugin each.
21:49:24preglowamiconn: codec needs iram more than core, if you ask me
21:49:43preglowsure, but in general i think that is true
21:49:57amiconnMaybe someone will implement a sane resampler one day :cough:
21:50:08preglowahh, sure, dsp also likes iram
21:50:19*preglow forgot about dsp for some reason :/
21:51:10preglowi've been looking at one resampler scheme lately, and it actually requires tons of memory
21:51:17preglowmore than iram can hold, potentially
21:51:21preglowbut it should be easy to code
21:51:36 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
21:51:44Buschelwell, 64KB of iram is enough to speed up mpc :o)
21:52:02preglowmpc is already fast enough, stop thinking about it! :)
21:52:09*Buschel is quite egoistic with mpc-performance
21:52:16 Quit tedrock (Remote closed the connection)
21:52:27Buschelit's _my_ runtime on _my_ player ;o)
21:52:27 Join tedrock [0] (
21:53:05 Join low_light [0] (i=c730180a@gateway/web/ajax/
21:54:08Buschelif the additional iram can be used for significantly speeding up mp3 we should use the iram for this. i guess the mp3-decoder is one of the most used codecs and should bring the most gain in runtime
21:54:26 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
21:54:37preglowit's already quite fast, but i'd love it faster
21:54:54jhMikeShmmm -c4000 is definitely within reach on the S and -c5000 looks like a challenge even there
21:55:20Buschelmaybe the better runtime of the iPod-OF is not caused by weird PP-settings but by much higher codec efficiency.
21:55:41LambdaCalculus37Or a combination of both.
21:55:42Buscheli've read about mp3-decoding on arm7 at 20-25MHz
21:55:47Nico_Phow much better is the OF now?
21:55:57jhMikeSBuschel: the disable of the ata lock hack hasn't caused any issues for you now?
21:56:13LloreanjhMikeS: What about that "insane filter in IRAM" note that took the coldfire from 2% to 12%? How effective might that be on the Beast?
21:56:15LambdaCalculus37Nico_P: ~3 hours.
21:56:23BuscheljhMikeS: i've got a 30GB one −− never had issues with it
21:56:49jhMikeSBuschel: ah, I somehow I though you had the 80Gig. :)
21:56:55preglowBuschel: it's a little bit of both, i think
21:57:10LambdaCalculus37Nico_P: According to Apple's claims, the OF should get 14 hours of runtime with audio.
21:57:13preglowBuschel: apple runs their codecs much tigher, afaik they actually clock the cpu very tightly by what each codec needs
21:57:28jhMikeSLlorean: I don't know. Those ARMs have very good caches so I'm don't think it would accelerate it by nearly that much.
21:57:53Buschelpreglow: might be... nevertheless i get >14h runtime with a 30GB 5.5G (mpc ~170kbps). apple claims 14h for 128kbps aac.
21:58:07 Quit inigomontoya ("CGI:IRC (EOF)")
21:58:07 Quit LambdaCalculus37 ("CGI:IRC 0.5.9 (2006/06/06)")
21:58:12LloreanjhMikeS: So it's more or less down to "how much can the DSP instructions help out"?
21:58:15preglownot bad at all
21:58:16LloreanOr DSP-like
21:58:22preglowbut of course mpc is very easy to decode compared to aac :-)
21:58:52Buschelpreglow: this shows how the codec efficiency scales battery runtime
21:59:14Buschelpreglow: optimize mp3 to be running @24MHz and it's done :)
22:00:49Buschelback to my question: putting the additional iram for PP5022/24 to the plugin-iram is the agreed solution?
22:01:37 Quit jhMikeS (Nick collision from services.)
22:01:43 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
22:01:53preglowplugin iram == codec iram, yes?
22:02:13preglowreserving some for dsp would not be sutpid
22:02:18preglowbut that can be done later, of course
22:02:32jhMikeSLlorean: pretty much
22:03:01Buschelpreglow: for src?
22:03:16*Buschel meant sample rate conversion
22:03:30preglowamong other things
22:03:38preglowbut yeah, that would be an important thing
22:04:17Buschelthe iram of all other targets is split into halfs for core and plugin/codec
22:05:00Buschelbut the core-iram of 5.5G is only filled to about 38KB as far as i can see... adding 16KB would be a waste
22:05:09preglowif there is already some free, sure
22:05:25preglowhave you tried out allocating more iram to mpc, btw?
22:05:29preglowahh, yeah, there was a patch...
22:05:40Buschelyep +6.6% speed ;o)
22:05:49 Quit linuxstb (Read error: 113 (No route to host))
22:06:23 Join mf0102 [0] (n=michi@
22:06:31preglowdata, i assume?
22:06:38Buschelso, at least no-one bothers about the partitioning. then i will commit the "small" solution which only touches
22:06:52Buschelpreglow: yes, the subband sample buffer
22:06:56 Quit tedrock (Remote closed the connection)
22:07:03 Join tedrock [0] (
22:07:04Buschelpreglow: Y_L, Y_R
22:07:07preglowwell, i care, but if there is nothing extra to use it on in the core: use it for the codecs
22:07:29pregloweasier to find places to use it there
22:07:36*gevaerts sees his personal 32KB of IRAM for experimental use vanish
22:07:41Buschelpreglow: you've got a M5 or X5 target?
22:08:16 Quit dan_a (Read error: 110 (Connection timed out))
22:08:35Buschelthe additional iram-use of mpc could also be set for those targets. testing and measuring it would be interesting...
22:08:40preglowno, my only coldfire target has 96kb ram
22:08:52preglowbut iram for data is Much more important for coldfire than for pp
22:09:02 Quit Mouser_X (Nick collision from services.)
22:09:11 Join Mouser_X [0] (
22:09:46*Buschel doesn't like to submit the patch for non-tested targets
22:10:01shotofaddspreglow: apologies for asking a question and then disappearing... sounds like you're making good progress. I was intending to split the current ata-nand-tcc78x.c into an ata-tcc78x.c that calls into separate ata-nand-tcc78x.c and ata-sd-tcc78x.c implementations, but it looks like you might get there first ;)
22:10:10Buschelanyone with a M5/X5 around to test some mpc-patch?
22:10:17preglowshotofadds: i think i will, but yeah, that's what i had in mind too
22:10:35preglowshotofadds: you know how the bus freq and cpu freq are related, btw?
22:11:38toffe82somebody interested with a shematics of a based pp5022 player ?
22:12:18Bagderh10 ?
22:12:34 Quit perrikwp (" ajax IRC Client")
22:12:47preglowBuschel: plenty of people have x5, amiconn has m5, don't know if anyone else do
22:13:11toffe82Bagder: philips
22:13:20toffe82it seems to be made by Mitac
22:13:28BagderI'm interested
22:13:36toffe82email ?
22:14:12Bagdersure, daniel at
22:14:24shotofaddspreglow: as i understand it the CPU freq can be scaled from bus freq (divided PLL) up to system freq (PLL). the comments in set_cpu_frequency() attempts to explain this.
22:14:29 Join stripwax [0] (
22:14:32toffe82actually the hdd1630, I am downloading the hdd6330 , will have it in a moment
22:14:52amiconnpreglow: I have an X5. pixelma has an M5
22:15:00*Buschel open another Beck's
22:15:03preglowamiconn: ah, that's right
22:15:41amiconnAnd btw, while iram is sometimes more useful for codecs, it's easier to make use of it in the core
22:15:50Buschelamiconn: would you try to add a mpc-patch (which must pimp in one line of code)?
22:16:20preglowamiconn: easier to make use of it, yes, but not always to very great effect
22:16:28preglowfinding spots where it matters is pretty easy in codecs
22:16:28amiconn...since the core's iram isage is actually different between targets, while the codecs are usually identical within an architecture
22:16:32preglowsince the code gets called all the time
22:16:48amiconnAnd changing that makes the codecs rather ugly
22:17:04amiconnOn PP502x, iram usage isn't veryimportant anyway
22:17:20amiconnIt has an efficient (and non-broken) cache
22:18:05amiconnBuschel: I doubt it would be useful. I have no mpcs other than 1 or 2 test files
22:18:23 Quit Buschel (Read error: 104 (Connection reset by peer))
22:18:34 Join Buschel [0] (
22:18:37Lear Any buffering.c experts around? :)
22:18:55Buscheldamned wlan adapter
22:19:35Buschelamiconn: 1 file is enough for a quick check via test_codec
22:20:33preglowamiconn: it will make them rather ugly, but efficiency isn't always pretty
22:20:44preglowamiconn: and we can't just throw better runtimes out the window
22:21:30 Quit tedrock (Remote closed the connection)
22:21:41 Join tedrock [0] (
22:22:00 Join linuxstb__ [0] (
22:22:15Buschelamiconn: FS #8987, the mpc-patch. pleas hard-code a "#if 1" instead of the PP-ifdef-ing in musepack.h
22:22:15*amiconn thinks the main problem regarding PP runtime i sthe lack of ide dma support in rockbox
22:22:38preglowdo we transfer data so often that that really matters?
22:23:19jhMikeShd's burn alot of power though when they do run
22:23:52preglowdo we know enough to do pp ide dma?
22:24:00*Buschel is really interested in see-ing the difference in performance (if iram-usage scales much better on coldfire than on PP)
22:24:31preglowi'd expexct the performance difference to be greater on coldfire
22:24:33 Join dan_a [0] (n=dan-mirc@
22:24:34jhMikeSpreglow: I don't think so but having a UDMA driver would be a nice start since I'd like to use it on the beast
22:24:47LearNico_P: In buffering.c, shouldn't update_data_counters grab the list mutex? I can trigger data aborts from that code, it seems.
22:25:16*preglow knows really little about ide
22:25:29 Quit Buschel ()
22:25:31Nico_PLear: it seems you're right
22:25:40Nico_Plet me look a bit deeper
22:25:40 Join Buschel [0] (
22:25:44jhMikeSMDMA isn't any faster in principle but DMA usage would allow core sleeping during transfers too
22:25:55amiconnThe PP datasheets mention that the ide controller is udma capable
22:26:34Nico_PLear: do you have a fix, or should I commit one?
22:26:41toffe82Bagder: check your mail
22:26:41LearNico_P: Mind you, only seem to happen on my e200 with dircache and albumart usage (and possibly certain file formats).
22:26:57amiconnAnd given that we don't reach OF runtime on PP, while we exceed OF runtime on coldfire (where the ide controller can't do dma), this is rather likely the reason...
22:27:09Nico_PLear: I've seen some weird things on the ipod, which I suspect could come from there.
22:27:17LearNico_P: Not anything tested yet. About to do that. Triggering can be tricky though...
22:27:17Bagdertoffe82: thanks!
22:27:19toffe82Bagder: there is the hdd1630 and the hdd6330 schematics, they are really complete
22:27:29jhMikeSThe imx is UDMA5 capable but that requires some sort of clocking trick to the controller I'm not clear how to setup without messing up ipg_clk for other peripherals. UDMA4 can do 66MB/s and UDMA5 100
22:27:54amiconnRaw interface speed, I know
22:28:03LearNico_P: If the data abort is in the low range (I get it at about 0x9000, but that's with some stuff disabled), it could possibly be that...
22:28:06amiconnIirc the PP should be capable of UDMA2
22:28:20Nico_PLear: I haven't had data aborts though
22:28:39Buschelamiconn: i'd also not expect saving in battery runtime with DMA for IDE. we would only save boosting while buffering then. this a few seconds boosting each hour.
22:28:42 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
22:29:04amiconnEven MDMA would be better than PIO, because it lowers CPU load
22:29:05jhMikeSof course that's raw. right now I'm getting slower reads than writes on the beast even though _I think_ the timings should be setup properly for PIO4
22:29:32amiconnBuschel: We would lower boosting during buffering *and* shorten the buffering times
22:29:50amiconnSlower reads than writes seem to be normal
22:30:19amiconnAnd it's not just a few seconds an hour
22:30:20jhMikeSreally? other's benchmarks on the test_disk page don't seem to aggree irrc
22:30:25Buschelamiconn: hmm, you're right
22:30:33toffe82Bagder: I would like to have schematics like this for all the plauers :)
22:30:43*Buschel forgot about lowring loading time
22:30:50Bagderoh yeah, that'd be great
22:30:54amiconnI observed buffering cacles taking 10..30 seconds on 32MB ipod
22:31:20Bagdertoffe82: we got a lot of the iriver ones as you might've seen
22:31:40Buschelamiconn: i see about 10-12s
22:31:40LearNico_P: Still data abort, in find_handle this time...
22:31:41jhMikeSgiving buffering a priority bump lowered buffer times on my H10 by 5 seconds
22:31:42amiconnBuschel: Use a heavy wps, and/or a cpu intensive codec, or some dsp...
22:31:46preglowwell, someone do a test with no loads whatsoever and see how drastically it changes
22:32:03 Join petur [50] (n=petur@rockbox/developer/petur)
22:32:12amiconnI my first tests on PP5002 I've even seen buffering times in excess of one minute
22:32:15Buschelamiconn: ok, i of course use my well known codec
22:33:13 Quit linuxstb_ (Read error: 113 (No route to host))
22:33:13jhMikeS <= /me seem he was the only one that did the test without dircache besides Barry Wardell
22:33:35toffe82I found also the ipod testing procedure, don't know really what it is , but downloading it now
22:34:28toffe82Bagder: we can't put these on the wiki
22:35:06jhMikeSI think the gigabeat S will use UDMA if PIO isn't forced since the ATA controller seems to be the USB/ATA bridge and is DMA mode capable.
22:35:25Nico_PLear: you put the mutex_lock call after the find_handle call, did you?
22:35:34amiconnjhMikeS: Well, with vs. without dircache only changes Open and Dirscan speeds substantially, and Create and Delete (a little bit, and in the opposite direction!)
22:36:06stripwaxBuschel - OF seems to fully buffer in about 5s on 64MB 5g video
22:36:37amiconnThere are also some measurements done by Godeater with are affected by the overflow bug...
22:36:53jhMikeSI can set timings for read/write invidually so maybe I'll play with that stuff at some point
22:37:16*amiconn can't type :\
22:37:24LearNico_P: Yes, right before "m = first_handle" and the while loop
22:37:29Buschelstripwax: i never used the OF −− but that's pretty impressive
22:37:35jhMikeSbut who can complain about buffering 62MB in about 9 seconds anyway (that's what actually happens during playback)
22:37:38Nico_PLear: ok, I did the same.
22:37:56Buschelseed: you can see my response?
22:38:30LloreanBuschel, stripwax: Have you confirmed the OF buffers 62mb though?
22:38:38toffe82anybody knows about the built in diagnostic on the ipod ?
22:38:50preglowwhat about it?
22:39:18stripwaxLlorean - not officially.. I had the ipod held up to my ear for about forty mins and after the initial buffering it didn't rebuffer.. but didn't hold out for the full hour-and-a-bit
22:39:33toffe82preglow: built in tools to check your ipod
22:39:42jhMikeSI'm sure the beast will benefit greatly from DVFS since scaling voltage and frequency together make it operate cubically
22:39:52stripwaxwould UDMA explain the difference between OF 32MB and 64MB runtime?
22:39:52preglowwell, yeah, left + select when you boot, but i thought everyone knew that
22:40:00Lloreanstripwax: Well it also depends on the bitrates of tracks buffered.
22:40:14toffe82preglow: I found the manual for this
22:40:22toffe82with all the function
22:40:23stripwaxLlorean - ~130kbps average
22:40:30jhMikeShow does e200 do with runtime?
22:40:31amiconnstripwax: The 64MB ipods also have larger batteries
22:41:24stripwaxamiconn - unh, true.. but what I guess I meant was that the difference between rockbox 32mb vs 64mb isn't as great as the difference between OF 32mb vs OF 64mb.
22:41:24amiconnMore RAM gives *slightly* longer runtimes on hdd based daps, if you let it play continuously (i.e. no heavy skipping that causes rebuffering)
22:42:05amiconnThe effect probably becomes less noticeable with increasing ram sizes
22:42:13jhMikeShmmm....e200 seems to do very well
22:42:24Buschelbut maybe the larger hdd also needs more energy to speed up? the 64MB vs. 32MB may compansate this
22:42:44Buschellike dual-disk instead of one (just an idea)
22:42:54amiconn...for 2 reasons. (1) More ram also needs a bit more power. (2) More ram makes the disk spin longer, lowering the effect of spinup time
22:43:00LloreanBuschel: Well, we don't see significant runtime differences between H120 and H140 afaik
22:43:04jhMikeSM5 is just sick :)
22:43:16amiconnjhMikeS: M5*L*
22:43:22*Nico_P just times buffering at ~30 secs on the 80 GB iPod
22:43:45stripwaxthat's after dircache was initialised?
22:43:54Nico_Pyes, it had dircache
22:44:02*stripwax just realised his timing probably included all the seeking at startup to build the dircache
22:44:12jhMikeSamiconn: of course, my bad :)
22:44:26amiconnI know that going 2MB -> 8MB (for MP3 on archos) extends runtime by ~22%
22:44:30stripwaxI didn't try timing a full rebuffer after stopped playback in an idle rockbox
22:44:31Nico_Pstripwax: it also had been on for long enough for the dircache scan to be over
22:44:44stripwaxyep, cool. so that's roughly equivalent to what I saw too then
22:44:50amiconnThat is with all other factors kept equal (same disk, same batteries, same music)
22:46:00 Join Deying_Gravity [0] (
22:46:18amiconnNico_P: ~30 sec on 64MB ipod means ~15 seconds on 32MB ipod - about what I'd expect too atm
22:46:23Nico_Pamiconn: is disk spinup expected to take longer with your recent fix?
22:47:03Nico_PjhMikeS: what about disabling the hack? spinup seems to take longer now. I guess I should perform a real test
22:47:07jhMikeSgigabeat S does pretty well too (if you can't tell by now, I hadn't really explored those pages before)
22:47:12amiconnThe disk starts spinning up as soon as power is applied. The extra wait is shorther than the spinup takes
22:47:28Nico_Pbuffering on the S is mighty fast
22:47:41 Part Deying_Gravity ("Leaving.")
22:47:41*jhMikeS keeps typing "S" when he means "FX"
22:47:59 Quit mcuelenaere ()
22:48:31*amiconn should really re-run test_disk on those targets where only old data is available
22:48:51jhMikeSNico_P: maybe the extra attention is slowing your perceptions of time? I get about 15seconds buffering on H10 which is 32MB so if you're getting that now it sounds reasonable.
22:48:54LearNico_P: Seems like my data aborts are caused by something else, unsure what. If I resume a particular track it crashes. If I first resume another track, the first track works fine. (And regular playback of the first track is fine, it's just resume that crashes.)
22:49:08*amiconn thinks that with svn, mentioning the revision once is enough
22:49:37Nico_PLear: strange. have you tried on sim? and on other targets?
22:50:18 Quit low_light (" ajax IRC Client")
22:51:22Nico_PLear: I committed the mutex locking fix btw
22:51:53 Quit einhirn_ (Read error: 113 (No route to host))
22:52:06LearNot much on sim, and no other targets yet. Not anything else with an arm on it at least.
22:53:05*amiconn started a rebuild cycle
22:53:31preglowshotofadds: btw, can i get away with upgrading the firmware without running the "0" part of it, which i assume is the one that deletes everything?
22:54:07 Nick linuxstb__ is now known as linuxstb (
22:54:13LearNico_P: Btw, isn't the buffering thread stack a bit larger than necessary?
22:54:45Nico_PLear: from what I've seen, it is, yes
22:54:53 Join adiamas [0] (n=adiamas@
22:56:43 Quit domonoky (Read error: 104 (Connection reset by peer))
22:57:09shotofaddspreglow: yeah, just the '1' and '2' folder should be fine (but check the file date on the '0' bin file)
22:57:36Buschelamiconn: i gotta go now. if you could test the patch (performance wise) i've described above on your x5 this would be great :o) i am reading the logs!
22:57:40preglowshotofadds: for what?
22:57:41Buschelsee you!
22:57:45 Quit Buschel ()
22:59:40preglowshotofadds: yeah, unchanged from last update
22:59:41 Quit leox (Read error: 104 (Connection reset by peer))
22:59:59 Quit Falafel (" GoobyeEveryone")
23:04:24 Part raph1
23:04:33shotofaddspreglow: 24/8/07? if so it's the same as the very first DAB version (v4.51) so no need to apply it
23:05:56preglowshotofadds: aye
23:06:07 Quit bluebrother ("leaving")
23:06:08preglownew firmware now, and i'll just upgrade to a new bootloader while at it
23:06:49*shotofadds is still on a rockbox blue bootloader...
23:07:04preglowi still get debug output on mine
23:08:16preglowbtw, what would i do if i for some reason managed to screw up my firmware image?
23:08:24shotofaddspreglow: btw. if your SD work is successful (hopefully it will be!) I strongly believe that the SD card should be the default drive until we have flash write support
23:09:29preglowshotofadds: then i'll have to get an sd card with more than 32 megs! :)
23:09:47preglowshotofadds: i have no doubt the sd work will be successful, i just need to figure out the bugs
23:10:11shotofaddsif you screw up your flash (unlikely, even I haven't managed that) you should be able to load firmware 0 using tcctool, then flash 1 again using the normal method
23:10:27preglowshotofadds: bah, of course :)
23:10:32preglowwe don't set up the 32 bit rimer now do we?
23:10:46shotofaddsno, but I was hoping to use it for the user timer
23:11:50preglowah, so no using it for quick timing info, then
23:11:59preglowwould be nice to use for delay loops here and there
23:12:08 Quit bughunter2 ("Leaving.")
23:13:40 Quit fehmicans (Remote closed the connection)
23:13:45shotofaddswould one of the other timers suffice (as in the lcd driver, which uses TIMER2 or something as there's no tick in the bootloader)?
23:14:17preglowwell, can you just read their current value from a register?
23:14:23preglowit'd need to be a microsec timer
23:15:26preglowargh, i set my d2 to some gibberish language :/
23:15:45preglowquick info on where the lang setting is, please :)
23:15:53shotofaddshold on a sec...
23:15:56 Quit jgarvey ("Leaving")
23:16:11preglowa right to left language, on top of everything
23:16:15preglowhaha, which bugs the d2 os out, i see :P
23:16:24 Join leox [0] (
23:16:27preglowshotofadds: i figured it out
23:16:34preglowit was hebrew, heh
23:17:04shotofaddsyeah it's quite simple, System (cog thing) -> Display (screen icon) -> first item
23:17:34preglowbut yeah, back to what i asked about the timers
23:20:26shotofaddsyeah, the counters can be read from a hardware register (or a flag is set when a ref value is reached). timers 4 and 5 are 20 bit, if that helps
23:21:42shotofaddsthe timer source clock is set to 2Mhz in svn, and you can divide from that
23:24:55 Quit lee-qid ("aufwiederbyebientotsayonara")
23:27:18 Quit tvelocity ("Αποχώρησε")
23:27:35preglowshotofadds: well, 2mhz and divide by two should yield a microsec clock, no?
23:27:52 Quit BitTorment (Read error: 110 (Connection timed out))
23:29:01shotofaddsdiveide by two is the minimum, btw
23:29:02preglowthen that should be good enough for most timing needs
23:29:28shotofaddshence mychoice of 2mhz...
23:29:39preglownice thinking :)
23:29:47 Quit Lear ("ChatZilla 0.9.82 [Firefox 3.0pre/2008051206]")
23:29:57*amiconn got rather odd timing results on ipod G2
23:30:03shotofaddssorry for being a bit slow today. i'm really still not thinking straight
23:30:12*shotofadds needs more time off
23:31:00preglowno worries, i've got the time :)
23:31:48 Nick FOAD is now known as wax (
23:31:53 Nick wax is now known as FOAD (
23:32:03preglowwow, that was pointless
23:33:37preglowshotofadds: i thought you removed the debug info from the bl, i still get it
23:34:12 Quit bertrik ("bye")
23:34:19preglowshotofadds: oh, and btw, the problem that nand access sometimes doesn't seem to work when tcctooling and image persists
23:34:29preglowshotofadds: sometimes i get full theming, sometimes a plain black rockbox with 8x8 font
23:37:33 Quit Zarggg (Connection reset by peer)
23:37:37 Quit adiamas ("ChatZilla 0.9.82 [Firefox]")
23:38:04 Join Zarggg [0] (
23:40:44***Saving seen data "./dancer.seen"
23:43:18 Quit XavierGr ()
23:53:11 Join Xqtftqx [0] (i=40e9e3d2@gateway/web/ajax/
23:53:18XqtftqxHey guys
23:53:54shotofaddspreglow: the debug text is still there. what I removed was the ability to switch into the old GPIO/ADC/etc bootloader debug screen.
23:54:03shotofaddsno point now that that's available in the main image
23:54:15XqtftqxIm having a problem, whenever i turn on my gigabeat S, rockbox says "Undefined instruction at 0000051c
23:54:44 Quit MTee ()
23:54:45preglowshotofadds: didn't even know about that :>
23:55:19 Quit toffe82 (Read error: 104 (Connection reset by peer))
23:56:40shotofaddsthe nand errors will persist until I get time to figure out the last details of the translation scheme
23:57:34shotofaddsthe missing theme is literally because the backdrop and/or font or icons failed to load. usually works for me, though!
23:57:35jhMikeSXqtftqx: a recent commit compells a bootloader upgrade (adding write buffering)
23:57:53preglowshotofadds: just weird that it sometimes happens, sometimes not, even with no changes to the flash
23:58:18shotofaddsdid you boot into the OF between runs?
23:58:20 Join BlakeJohnson87 [0] (
23:58:27preglowshotofadds: i don't think s
23:58:51XqtftqxjhMikeS can you send me the bootloader? Mines not compiling corrects
23:58:54shotofaddsweird. it should be consistent unless something has altered the flash.

Previous day | Next day