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

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

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

#rockbox log for 2007-02-25

00:00:27jhMikeSSoap: there's quicks that the x5 picked up while mine was serviced that a version of rb from right before it was sent out doesn't have...and I don't know why yet
00:04:22jhMikeSbut I see from my own battery bench it didn't hit 3.70v until 6:38:57...crap...not good
00:09:52Soaplinuxstb: maybe "Error -1" should have been changed to "You skipped section 2.2.2 - read the instructions again" ;)
00:10:07jhMikeSCowon is in for it and they are going to replace the player free of charge and that's the bottom line...maybe I'll change my mind and say don't but an x5 cause if it needs service they'll mess it up for you.
00:10:16jhMikeSs/but/buy/
00:10:51jhMikeSIn fact on that basis I'm going to officially recommend noone purchase another goddamn thing from Cowon
00:11:06 Join Hub441 [0] (n=Hub441@pD9FA1255.dip.t-dialin.net)
00:12:45jhMikeSwhich is too bad because it's otherwise a very nice player ... I'm so pissed if anyone can't tell already
00:14:40 Quit mightybrick ("CGI:IRC (EOF)")
00:15:04linuxstb_Soap: You think that would help? :)
00:15:42Soapno. :(
00:16:07Soapbut a verbal taunt from the bootloader, while being oh so wrong...would feel oh so good.
00:16:43scorche"I cant let you do that Dave"
00:17:50 Quit idnar (Nick collision from services.)
00:17:52 Join idnar_ [0] (i=mithrand@unaffiliated/idnar)
00:22:45 Quit Mouser_X (Nick collision from services.)
00:22:54 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
00:25:21jhMikeSpreglow: hope I didn't mess things up in there too much for you :o
00:25:30Hub441linuxstb_: thx for the link to the fixed rockbox (nofreqscaling)
00:26:04Hub441can i hide compitaltions from the regular artists list?
00:34:16Soapyou can cludge it with the AlbumArtist tag.
00:34:43SoapThere should be a forum discussion on that, I am not a Database man myself.
00:38:08 Join Ashet [0] (i=55e6abf0@gateway/web/cgi-irc/labb.contactor.se/x-c5a55300b373429a)
00:38:10scorcheapparently bascule is >_>
00:38:20Ashethey guys, sorry for interrupting and etc...
00:38:21 Quit amigan_ ("leaving")
00:38:21 Quit amigan ("leaving")
00:38:32 Join amigan [0] (i=dcp1990@unaffiliated/amigan)
00:38:36Ashetbut I was just wondering if there are people working on makeing this work with the ipod nano 2nd gen?
00:38:46scorchenot actively
00:39:04Ashettoo bad :(
00:39:25 Quit Mouser_X (Read error: 60 (Operation timed out))
00:39:29scorchefeel free to workon it yourself and tell us your progress
00:39:46AshetI would've loved too, if I had the experience
00:40:01scorcheread books
00:40:06Ashettakes time.
00:40:15scorcheand it wouldnt for us?
00:40:38Ashetsome people are better at things than others
00:40:44redbrevahurah! I have now got a working 'Menu image' option working on the rockbox-themes... basically if there is a second image that is usefull to display the theme, a mouse over shows it... some examples on http://www.rockbox-themes.org/index.php?res=240x320x16.
00:41:14 Quit Arathis ("Bye, bye")
00:44:06Domonokyhm which one does have a second image ?
00:44:23redbrevaBeatMP
00:44:43Domonokyah.. nice..
00:45:30redbrevathere are quite a few in the the 240x320 page... I jusr need to start getting some for other pages now
00:46:10redbrevaI think it works quite well, and my PHP scripting has moved on to another level in the process!!
00:47:20 Join Thundercloud [0] (n=thunderc@82.153.198.79)
00:47:21Domonokyand you fixed the rbutil.php script :-) (ats half ready)
00:47:24markunAshet: the nano 2g has very different (but equally undocumented) components compared to all the other ipods. Also, they did a good job of encrypting everything.
00:47:30 Quit Ashet ("CGI:IRC")
00:47:48redbreva:-) Is that working Ok for your needs?
00:48:10SoapWhatever happened to "My Humps" being the official song to be played during WPS screenshots?
00:48:11Soap;)
00:48:27Domonokyjeah, its good, but i dont need the file size.. description would also be good in the response :-)
00:48:27redbrevaGot fed up of that one :D
00:48:47Shaidthere was an official song?
00:48:52Shaidand it was that?
00:48:55Shaid...
00:49:06scorchehe was joking
00:49:19redbrevaI thought filesize may be usefull - some are pretty big so as a warning (3 to 4 MB)
00:49:50Domonokyoki that would be good
00:50:15Domonokybut including the notes would also be good (preferrbly at the end of the line)
00:50:23Shaidphew.
00:50:31redbrevaThe desctiption could be added pretty easy (I think) - Just may have to add some 'Delimiter' text as there are sometimes commas in the text
00:50:44 Quit x1jmp (Remote closed the connection)
00:51:09Domonokyjeah, or change the delimiter
00:51:15redbreva48,000 hits and counting so far this month!
00:51:45redbrevaI could try tabs...?
00:52:54Domonokyjeah or some other seldom used chars like |
00:55:15Hub441Soap: yupp, i searched the forum and found a little bit, but nothing so far that would read the compilation-tag from mp3-files
00:57:24Soapcopy Artist to AlbumArtist on all albums.
00:57:42SoapSet AlbumArtist to whatever you want for the comilations.
00:57:51SoapThen display AlbumArtist in the database.
00:58:07Soapwith a bulk tagger it should take you only a couple of minutes.
00:58:43Hub441yes, that would do it, but it's just a work around and not the feature i'm looking for
00:58:46Soapor, can you have the database display Artist only if AlbumArtist doesn't exist? If you can do conditionals in the Database like that.
00:58:58 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
00:59:00SoapWell, the compilation tag isn't as widely supported as AlbumArtist.
00:59:12Soapand Database just got support for AlbumArtist.
00:59:33Hub441do ogg-files have a compilation tag?
01:00
01:00:06markunI think you can define whichever tag you want
01:00:07Soapvorbis comments are free-form, but I believe AlbumArtist is considered a standard, don't quote me on that.
01:00:12 Quit Thundercloud_ (Connection timed out)
01:03:36redbrevaDomonoky: Hmm, have alook at the rbuil.php page now... That's using | as a delimeter, and free text t end of line...
01:04:06 Join Vyrus001 [0] (n=Vyrus001@adsl-75-5-5-52.dsl.irvnca.sbcglobal.net)
01:05:53Domonokygood, i will adapt my code
01:07:23redbrevaCool, Ok I'm off to bed - Goodnight
01:07:33DomonokyGoodnight
01:07:41 Part redbreva
01:07:44 Join EvilG [0] (n=evilg123@S0106001346a401fd.ed.shawcable.net)
01:11:10EvilGHello, I'd like to introduce myself and ask for Wiki write permisions
01:11:16 Join decayedcell [0] (n=decayed_@ppp79-106.lns1.mel3.internode.on.net)
01:13:35markunEvilG: are you Gray Man?
01:15:13 Quit ender` (Read error: 104 (Connection reset by peer))
01:15:22 Quit funky (Read error: 113 (No route to host))
01:15:39EvilGHello, I'd like to introduce myself and ask for Wiki write permisions, my name on the wiki is Gary Light
01:17:12markunEvilG: Sorry for asking, but that is your real name, right?
01:19:00***Saving seen data "./dancer.seen"
01:20:09markunEvilG: still there??
01:22:39EvilGyes
01:22:41EvilGand yes
01:22:49EvilGno problem, sorry for being away
01:23:10 Part nls
01:23:29markun:)
01:24:17EvilGso, about write permissions?
01:24:50markunyou have them now
01:24:59markunhave fun in the wiki!
01:24:59EvilGThanks! :)
01:25:27 Quit EvilG ()
01:28:38 Quit Mouser_X (Nick collision from services.)
01:28:49 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
01:31:18 Quit Shaid (Read error: 104 (Connection reset by peer))
01:32:16 Quit decayedcell (Read error: 113 (No route to host))
01:33:46 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
01:34:52 Join lookup [0] (n=ajogaowe@c-68-45-79-127.hsd1.nj.comcast.net)
01:34:57lookuphey guys i need some help
01:35:05lookupim using rockbox on my ipod nano 1st gen
01:35:07 Join Shaid [0] (n=adam@203-214-30-35.dyn.iinet.net.au)
01:36:01lookupi have used itunes and the tunes are in iPod_Control, how do i get the database option to show?
01:39:19 Join printfXh4 [0] (n=pseudo@ppp98-196.lns1.bne1.internode.on.net)
01:39:20webtazhttp://www.rockbox.org/twiki/bin/view/Main/IpodFAQ#How_to_control_Rockbox_on_the_iP
01:39:54webtazi use iriver so i can't tell you how to manage it
01:40:08 Quit fwopenXh4 (Read error: 113 (No route to host))
01:43:44webtazlookup: try: Press Menu in WPS, to get in to the Main Menu -> File View -> Database ->
01:44:12webtazThere you should initialize
01:44:19webtazthen
01:44:22webtazrestart
01:45:17webtazand with holding menu down go into the quick menu and at Show Files switch to Database
01:46:43 Nick idnar_ is now known as idnar (i=mithrand@unaffiliated/idnar)
01:47:06 Quit linuxstb_ ("CGI:IRC (Ping timeout)")
01:47:09 Join Thundercloud_ [0] (n=thunderc@82.153.194.254)
01:49:13 Quit Mouser_X (Nick collision from services.)
01:49:22 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
01:52:40lookuptrying that now webtaz
01:55:22lookupwebtaz tried it, but like the directories and such don't show the folders/song tags
01:55:28 Join Hub441_ [0] (n=Hub441@pD9FA1092.dip.t-dialin.net)
01:56:03webtazdie you restart ?
01:56:09webtazwith the iriver
01:56:30webtazit's fitting
01:56:33webtazsry
01:56:42webtazi can't help you more
01:56:50webtazbut search at www.rockbox.org
01:56:52webtazmanual
01:57:01webtazthere you should find how it works
01:58:25 Join bonbonthejon [0] (n=jon@cpe-65-27-173-68.cinci.res.rr.com)
01:59:53 Quit lookup ()
02:00
02:00:23 Quit Nimdae (Read error: 110 (Connection timed out))
02:01:41 Join Thundercloud__ [0] (n=thunderc@82.153.192.148)
02:04:02 Quit entheh ("^~")
02:04:39 Quit hcs ("Download Gaim: http://gaim.sourceforge.net/")
02:06:43 Join hcs [0] (n=hcs@rockbox/contributor/hcs)
02:10:12 Quit Thundercloud (Read error: 110 (Connection timed out))
02:11:11 Quit Hub441 (Read error: 110 (Connection timed out))
02:15:01 Join rotator [0] (n=e@rockbox/developer/rotator)
02:15:27 Join Thundercloud [0] (n=thunderc@82.152.98.21)
02:19:10 Quit Hub441_ (Read error: 110 (Connection timed out))
02:22:23 Join TeaSea [0] (n=thunderc@82.153.198.77)
02:24:27 Quit Thundercloud_ (Read error: 110 (Connection timed out))
02:27:12 Quit Mouser_X (Nick collision from services.)
02:27:21 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
02:33:13 Quit Thundercloud (Read error: 60 (Operation timed out))
02:34:57 Quit desowin ("use linux")
02:37:05preglowjhMikeS: i'll be ok
02:38:47 Quit Thundercloud__ (Connection timed out)
02:42:25 Quit anathema (Read error: 110 (Connection timed out))
02:46:54 Quit webtaz ("( www.nnscript.de :: NoNameScript 4.02 :: www.XLhost.de )")
02:48:21 Join Roan [0] (n=JustJohn@pool-151-196-120-87.balt.east.verizon.net)
03:00
03:00:48 Join Kittt0s [0] (i=Kaa@84.94.96.252.cable.012.net.il)
03:12:43 Quit Mouser_X (Nick collision from services.)
03:12:51 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
03:12:52 Join webguest94 [0] (i=802446b7@gateway/web/cgi-irc/labb.contactor.se/x-799af5441ff4fd13)
03:13:57 Quit Kitt0s (Read error: 110 (Connection timed out))
03:14:27webguest94hi I was wondering if anybody could help me with my rockbox install?
03:14:56RoanHi, For anyone interested, I've posted some proposed changes to the Rockboxed theme for the Gigabeat in the WPS forum at http://forums.rockbox.org/index.php?topic=9001.0
03:15:31RoanWhat problems are you having with your install webguest94
03:16:25webguest94when I try to create a bootpartition.bin it tells me there is a readerror (I am installing on a 4G greyscale)
03:17:25webguest94I have used both diskdump and ipodpatcher, and both do not work - with diskdump I get an error message "bad boot sector signature"
03:18:24RoanIs your iPod formatted for use with a Macintosh or Windows computer?
03:18:28webguest94Macintosh
03:19:01***Saving seen data "./dancer.seen"
03:19:11Domonokyit has to be Windows format aka Fat32 i think..
03:19:36RoanHave you looked at this page: http://www.rockbox.org/twiki/bin/view/Main/IpodInstallationFromMacOSX
03:20:02webguest94Yeah I have used diskdump and ipod_fw
03:20:09webguest94Do you think it will help if I format it into FAT32 first?
03:20:35RoanYes, it has to be formatted for FAT32. That is the only filesystem that rockbox understands
03:20:45webguest94Ok I will try that. Thank you
03:25:16Soapin bold, third paragraph "IMPORTANT NOTE Rockbox only works with Windows format iPods (FAT32). In order to use a Mac format (HFS) iPod, you will need to convert it to FAT32. This can be done by using the iPod Updater on a Windows computer. After this conversion, you can continue to use your iPod on your Mac. Or if you dont have acces to a Windows computer, follow the "Convert an iPod from HFS to FAT32" steps"
03:25:47webguest94Hmm I see. Sorry I was under the impression that I could convert it to FAT32 at any point during the install
03:27:41preglowwrong impression
03:29:09 Part Roan
03:29:59XavierGrjhmikes: H100/H300 is inferior to X5 in audio?
03:32:24 Quit printfXh4 (Read error: 113 (No route to host))
03:32:40 Quit Domonoky (Read error: 104 (Connection reset by peer))
03:34:45preglowjhMikeS: using emac for just halving a number is a bit overkill, i think. it's also almost guaranteed to be slower. talking karaoke chan mode assembler here
03:37:51 Quit Mouser_X (Nick collision from services.)
03:38:00 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
03:44:29Seedhi preglow
03:44:53webguest94Thanks for the help guys - I've been trying to make a bootpartition for the past 10 minutes now, the file size seems stuck at 16MB when according to the guide it should be 40MB...is this normal?
03:48:08 Quit TeaSea (Read error: 104 (Connection reset by peer))
03:48:48 Join Thundercloud [0] (n=thunderc@82.153.38.89)
03:49:27preglowSeed: hiya
03:50:47 Quit webguest94 ("CGI:IRC (EOF)")
03:52:00Seedpreglow: I'm very happy with RB on the H320.. just wanted to tell everyone
03:52:26Seedbought one for my friend too.. excellent..Musepack and Wavpack work flawlessly
03:56:51preglowtoday you should be even happier! dsp just got blazingly fast
03:57:11preglowyou should be able to play musepack with full 5 band eq and crossfeed at under 45mhz now
03:57:15Seedis it on the CVS?
03:57:17preglowyes
03:57:38Seedsoon you'll be able to work on SV8 code
03:57:40preglowmusepack doesn't work flawlessy yet, it doesn't resume properly
03:57:43preglowoh?
03:57:48preglowdidn't think i'd see the day :)
03:58:03SeedUnix can already play SV8 tracks.. but since quality isn't improved, only size, most won't convert.. there's a util to convert SV7 to it
03:58:34Seedlook at the SVN when you find time
03:58:56 Join saratoga [0] (i=9803f034@gateway/web/cgi-irc/labb.contactor.se/x-77eeea7d9e2db707)
03:59:14saratogathey released SV8?
03:59:32saratogai didn't realize mpc was still being worked on
03:59:37saratogathought frank gave up years ago
03:59:52Seedam I to assume http://build.rockbox.org/dist/build-h300/rockbox.zip is what I need?
04:00
04:00:06Seedsaratoga: a French dev is working at an amazing rate on it
04:00:22SeedFrank only helps us with advice.. he's way too busy to code
04:01:05preglowi thought he couldn't code on it
04:02:05saratogathats just amazing
04:02:16saratogai spent years hoping someone would pick up mpc
04:02:21saratogaits such an awesome codec
04:02:27saratoganow it even has rockbox support
04:02:38 Join Thundercloud_ [0] (n=thunderc@82.153.194.91)
04:03:43preglowanywho, it's bedtime
04:03:43preglowlater
04:03:47Seedthe current coder improved huffman stuff.. if you convert SV7 to SV8 you gain 2.5% in size
04:03:51Seedbye, preglow
04:04:40 Quit Nico_P (Remote closed the connection)
04:04:59saratogahow is rockbox support?
04:05:05saratogathey've got the seeking stuff worked out?
04:07:47 Join aliask [0] (n=chatzill@c210-49-190-113.eburwd8.vic.optusnet.com.au)
04:07:55JdGordonhey aliask
04:08:28aliaskhey there jd
04:08:37JdGordonexcited for tomorow ? :)
04:08:42aliaskugh
04:08:46 Quit Vyrus001 ("Leaving")
04:08:47 Quit Mouser_X (Nick collision from services.)
04:08:49JdGordonstupid uni :'( im alergic to it... ive cought a cold
04:08:49aliaski still havnt bought my books
04:08:56 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
04:09:00aliaskme too
04:09:07JdGordonhaha
04:09:10aliaskit's hit me like a ton of bricks
04:09:14 Join webguest31 [0] (i=415fb8e9@gateway/web/cgi-irc/labb.contactor.se/x-7e470b687698f073)
04:09:28webguest31when is rockbox for 5g coming out(rockbox 3.1
04:10:06aliaskwebguest31: There is a working 5g version at the moment, but 3.0 isn't even released so 3.1 is quite a while away
04:10:06webguest31???
04:10:29webguest31but it says that it was out nov 06
04:10:32saratogarockbox has been running on the 5G for over a year
04:10:33aliaskAll 5g ipods work except for the 80gb version
04:10:48webguest31no 3.1 rokcbox
04:10:53webguest31rockbox*
04:10:58Seedsaratoga: sure. seeking is good
04:11:00aliaskThat's old info
04:11:05saratogaSeed: thx
04:11:26saratogawebguest31: 3.1 will be released whenever someone feels like releaseing 3.0, and then doing another release
04:11:43webguest31wow you guys are taking a long time
04:11:46saratogai don't think there are any plans at the moment for another release
04:11:58aliaskwebguest31: If you download a 5g rockbox install it'll work.
04:12:01 Quit webguest31 (Client Quit)
04:13:15aliaskWhat exactly is needed for 3.0? There was the playback stuff, but IIRC that's been fixed hasn't it?
04:13:30 Quit saratoga ("CGI:IRC (EOF)")
04:15:39JdGordonpretty much... I think we have decided to never really bother with releases...
04:15:41aliaskOnly thing I can think of is the way pause is implemented on SWCODEC... voice won't work while music is paused.
04:15:53 Part hcs
04:22:32 Quit Thundercloud (Read error: 110 (Connection timed out))
04:26:02 Quit Mouser_X (Nick collision from services.)
04:26:11 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
04:45:45 Quit JPMaximilian ("I Shall Return")
04:49:59 Join Thundercloud [0] (n=thunderc@82.153.140.61)
04:50:48JdGordonanyone got a laptop with an XP cdkey under it and the first like of the sticker says "Windows XP Professional 1-2CPU" ?
04:53:15 Quit Mouser_X (Nick collision from services.)
04:53:26 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
05:00
05:01:21 Quit DataGhost (Nick collision from services.)
05:01:27 Join DataGhost [0] (i=dataghos@ip3e832ea5.speed.planet.nl)
05:04:15 Join SirFunk_ [0] (n=Sir@cpe-74-71-205-222.twcny.res.rr.com)
05:05:25 Quit SirFunk_ (Read error: 104 (Connection reset by peer))
05:05:44 Join SirFunk_ [0] (n=Sir@cpe-74-71-205-222.twcny.res.rr.com)
05:07:27 Quit Thundercloud_ (Read error: 110 (Connection timed out))
05:08:01 Quit SirFunk (Read error: 110 (Connection timed out))
05:15:24 Quit JdGordon (Read error: 110 (Connection timed out))
05:19:04***Saving seen data "./dancer.seen"
05:25:49 Join CpuWhiz [0] (n=Eric@cpe-66-87-222-218.co.sprintbbd.net)
05:25:58 Part CpuWhiz ("Leaving")
05:27:04 Quit Thundercloud (Read error: 110 (Connection timed out))
05:35:43 Quit aliask ("Chatzilla 0.9.77 [Firefox 2.0.0.2/2007021917]")
05:48:58 Quit Mouser_X (Nick collision from services.)
05:49:07 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
05:49:40 Part ALTiDEL
05:50:14 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
06:00
06:13:45 Quit Mouser_X (Read error: 110 (Connection timed out))
06:14:42ShaidJdGordon: I do.
06:14:55Shaidfor your laptop question, that is
06:15:13JdGordonhave you got a xp cd that goes with it? or only a recovery disk from the maker?
06:15:21Shaidcd.
06:15:30Shaidand a second disk for the laptop makers stuff
06:17:30 Quit Nibbier (Read error: 110 (Connection timed out))
06:18:11 Join Nibbier [0] (n=sven@e181065124.adsl.alicedsl.de)
06:30:07 Quit CriamosAndy ("( www.nnscript.de :: NoNameScript 4.03 :: www.XLhost.de )")
06:37:38 Quit SirFunk_ (Read error: 110 (Connection timed out))
06:40:06 Join SirFunk_ [0] (n=Sir@cpe-74-71-205-222.twcny.res.rr.com)
06:44:42 Part toffe
06:55:48 Quit rotator ()
07:00
07:19:07***Saving seen data "./dancer.seen"
07:23:24XavierGrShaid grats on fixing your apostrophes! :P
07:23:56Shaid:')
07:28:30 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
07:33:03 Join midkay [0] (n=midkay@rockbox/developer/midkay)
07:59:05 Join printfXh4 [0] (n=pseudo@ppp98-196.lns1.bne1.internode.on.net)
07:59:21daurnanyone have Pink Floyd, Obscured By Clouds in lossless?
07:59:45printfXh4No.
08:00
08:00:15JdGordonnot i
08:00:44daurnmy track 9 (say) is corrupt
08:01:08JdGordonsurely you can rerip for the cdd you own? :p
08:01:22daurnna, borrowed it off a friend :S (shh)
08:01:33daurnwell, one of my mums friends
08:06:34jhMikeSXavierGr: yes, I would say the iRiver doesn't have the playback sound quality but that iRiver as far better service so at least you'll have a device that works again if it breaks.
08:08:52 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
08:30:39 Quit Mouser_X (Nick collision from services.)
08:30:48 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
08:38:25 Quit Mouser_X (Nick collision from services.)
08:38:34 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
08:43:43XavierGrjhMikeS: Don't be so sure on that
08:43:59XavierGrI had major troubles with the USBOTG on my H300 (new unit)
08:44:11XavierGrsent it twice to germany for repairs and still didn't fix it
08:44:41XavierGrat the end in a soldering frenzy I fixed it alone (by pure luck I must say)
08:45:20XavierGrthough there are cases that iRiver performed very generously to other users
08:46:43XavierGrbut really I am wondering how an X5 sounds like, though my ears are far from perfect I believe that H100/H300 is pretty good in audio quality.
08:47:55jhMikeSXavierGr: Well, LinusN told me he heard good things about iRiver but the only experiece he heard about from Cowon was me and was just going wtf? It had what seemed to be a good battery going out and after it was sent back the seconds time (They broke the USB), it now has a battery that dies in a couple hours.
08:48:53jhMikeSIf they didn't fix something, that's one thing...if they break something else when servicing thing...that just incometent.
08:48:58 Join Rob222241 [0] (n=Miranda@p54B166E4.dip.t-dialin.net)
08:49:11jhMikeSincompetant even
08:49:33XavierGrwell yes it is true that I have heard many good stories about iRiver RMA policies, though my own case was a disaster
08:49:46jhMikeSXavier: Not to plug for them but I find the x5 sound to far much clearer and richer. It kind of night and day really.
08:49:57XavierGrbut in the end you can change the battery right?
08:50:24jhMikeSIf I can find one I'd rather do it myself. This player was still under warranty when originally sent.
08:51:43jhMikeSWill a lithium battery just crap out all of a sudden if it was just starting to go a short time before? Something seemed suspicious in that a battery bench would end too soon.
08:52:15XavierGrso its battery is hard to find? Not the same as with ipods and irivers?
08:53:09*XavierGr reminded himself his flacky USBOTG H300 and goes to test it once again just to be sure it works
08:53:25jhMikeSI don't know. I suppose availability depends on when customers start commonly experiencing the battery failure after I think 3yrs. At which point I'm sure it will be easier.
08:55:17jhMikeSI did do a little searching anyway and I _might_ be able to find one if I dig around alot but it's not a common item yet.
08:55:36XavierGrphew! it works just fine :)
08:55:59XavierGrwell, what capacity and voltage is it?
08:57:25 Quit Rob2222 (Read error: 60 (Operation timed out))
09:00
09:01:03jhMikeSNominal voltage? I think 3.95V norminal. It's 950mAh. Charge peak is 4.20V (it drops from this peak rather quickly though to the nominal).
09:02:14XavierGrnominal for ipods and irivers I think is 3.7 though it starts on 4.20 too on rockbox
09:02:32XavierGrH100/H300 battery is 1300mAh
09:03:07XavierGrso you could take the size of the battery and see if an iriver/ipod battery can fit in there
09:03:29jhMikeSsounds about right. it should stay around 3.95 for quite awhile. I'm seeing it drop to 3.70V really fast and my bb attempt didn't show that until 6hrs just before it was sent in
09:03:45 Join etrask [0] (n=etrask@ip68-99-126-250.sd.sd.cox.net)
09:04:50jhMikeSIf it hangs at 3.7V and stays under load it would work since the cutoff is supposed to be I think 3.55V
09:05:39XavierGrwell it could be a shame if you can shove there a 2300mah bettery and not take advantage of it
09:06:49jhMikeSThen I can't tune the x5 battery curves up for rb which was what I was doing before the joystick busted (such a simple repair too).
09:07:09XavierGrmy h300 has a 1900mah battery and I managed to get 30 hours with 160kbps mp3
09:07:42XavierGrhow do you tune the curves? from bb data?
09:08:05jhMikeSWould be nice but it's a 60GB with a dual platter. The X5L version has something like 2300mAh but I preferred storage capacity and 10-14h was adequate.
09:08:37XavierGrwell my h300 has a 60GB drive too, but it is rather thick I must say
09:08:49XavierGrI know that X5 is quite thin compared to an H300
09:09:22jhMikeSThe x5 is about the same size as an H120. I've never actually had an H300 in hand but it appears a bit larger.
09:09:37XavierGrH300 is a bit thicker than H100
09:09:45XavierGrbut a bit shorter too
09:09:49 Quit Mouser_X (Nick collision from services.)
09:09:58 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
09:10:13XavierGrH340 and H140 that is
09:10:25jhMikeSIf fact...putting them side by side the H120 is exactly the same size as the x5
09:10:46XavierGrbecause H340 = H320 in fact the same in thickness (iirc)
09:11:01XavierGrbut H120 != H140 in thickness
09:11:40XavierGreg it is not possible to put a dual platter disk in an H120
09:11:44daurn[19:10:08] <XavierGr> because H340 = H320 in fact the same in thickness (iirc) <−− yo sure?
09:12:02XavierGrthough there are cases that H320 users managed to put a double platter disk
09:12:10XavierGrdaurn: no I might be wrong
09:12:40XavierGrbut there are H320 users with double platter disks (60-80GB)
09:12:46jhMikeSthe x5 20/30GB is thinner than the x5 60GB and the X5L models are the same thickness as the 60gig. So you pick between big hard drive or big battery since both won't fit.
09:13:15XavierGrwell petur will be the best for now
09:13:28XavierGrwith 2300 mah battery and 80GB disk :)
09:14:34jhMikeSSounds like the 300's must be on the big side. They also have a nice large color lcd iiuc.
09:14:52 Join darksoulk [0] (n=adam@124-168-142-179.dyn.iinet.net.au)
09:15:04 Quit Shaid (Nick collision from services.)
09:15:08 Nick darksoulk is now known as Shaid (n=adam@124-168-142-179.dyn.iinet.net.au)
09:15:40XavierGryup
09:15:44XavierGrvery clear
09:18:09jhMikeShrm...x5 said it was fully charged so I started it up and it's already down to 3.85v in a few minutess...bad
09:19:10***Saving seen data "./dancer.seen"
09:19:15 Join perplexity [0] (n=joust@dxb-as87281.alshamil.net.ae)
09:22:31jhMikeSperhaps they broke the pcf50606 which I'm sure they could pull off
09:29:05JdGordonXavierGr: the 320 is a few mm thinner, but without the plastic padding, a dual platter disk will fit in it
09:32:16 Join aliask [0] (n=chatzill@c210-49-190-113.eburwd8.vic.optusnet.com.au)
09:33:01 Join Ribs [0] (n=ribs@91.84.8.218)
09:39:57 Quit etrask (Read error: 110 (Connection timed out))
09:40:54 Quit Galois ("Leaving")
09:41:29 Join Galois [0] (i=djao@efnet-math.org)
09:43:10XavierGrjhMikeS: if it isn't the battery to blame then ouch!!
09:43:42JdGordondoes anyone know if stopping audio, allocating some memory, then calling audio_init() again would kill playback?
09:49:16 Quit JdGordon ("Leaving.")
09:49:57*amiconn will hopefully soon have a H180 :)
09:51:33 Join illriginal [0] (n=illrigin@c-67-191-48-2.hsd1.fl.comcast.net)
09:52:52 Quit printfXh4 (Read error: 113 (No route to host))
09:53:08 Join printfXh4 [0] (n=pseudo@ppp98-196.lns1.bne1.internode.on.net)
09:56:55 Join redbreva [0] (n=chatzill@host86-133-125-60.range86-133.btcentralplus.com)
10:00
10:00:52 Quit printfXh4 (Remote closed the connection)
10:01:07 Join printfXh4 [0] (n=pseudo@ppp98-196.lns1.bne1.internode.on.net)
10:02:46 Join nls [0] (n=nils@nl104-202-175.student.uu.se)
10:03:07 Part dewdude ("Leaving")
10:07:11 Quit thegeek ()
10:08:41 Quit Mouser_X (Read error: 110 (Connection timed out))
10:08:56 Join ackbahr [0] (n=ackbahr@d83-181-113-109.cust.tele2.ch)
10:11:36 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
10:12:07 Join JdGordon [0] (i=dced3920@gateway/web/cgi-irc/labb.contactor.se/x-279aba796023010f)
10:12:27JdGordonShaid: done, thanks alot
10:16:02Shaidand did it work fine?
10:21:50jhMikeSXavierGr: indeed, especially given who did it. I'm actually watching the voltage sit at 4.14V while plugged...down from 4.16V
10:22:35jhMikeSJdGordon: You use audio_get_buffer to do that, then call buffer_alloc
10:22:39 Quit redbreva ("Chatzilla 0.9.77 [Firefox 2.0.0.2/2007021917]")
10:22:56 Join GodEater1 [0] (n=bryan@host-83-146-13-191.bulldogdsl.com)
10:23:10jhMikeSJdGordon: after that, audio will reinitialize the buffer when restarted
10:23:23JdGordonso its perfectly safe?
10:23:34JdGordonshaid: checking now
10:23:49jhMikeSbut don't stop the audio, audio_get_buffer will do it all for you
10:24:09 Quit Entasis_ (Read error: 104 (Connection reset by peer))
10:24:17jhMikeSso if you use the buffer functions as intended, yes
10:24:20JdGordonso why dont we use this for stuff like dircache and buffers and such so we dont have to reboot?
10:25:22JdGordonsaid: it worked, thanks a milion :)
10:25:48 Join fulhack [0] (n=E@81-232-42-95-no37.tbcn.telia.com)
10:25:55jhMikeScould be done...it is done. buffer_alloc will permanently remove that memory from the pool so save the pointer and don't call it repeatedly. there used to be a leak in talk.c that always grabbed a new buffer.
10:26:35fulhackHey, just checking to see if there's any progress on the E10 rockbox port? Does any dev. have an E10?
10:26:36 Join ikkidrow_ [0] (i=58050584@gateway/web/cgi-irc/labb.contactor.se/x-33bd217ebd194514)
10:26:50fulhackThat's Iriver E10.
10:27:09JdGordonaudio_get_buffer() returns the beging of he current audio? so each time you call it you get a new pointer? or the beging of the whole buffer so it breaks every other aloc?
10:27:18 Quit ikkidrow_ (Client Quit)
10:27:22JdGordonfulhack: not to my knowledge
10:28:03*JdGordon apoogiss for bad typos
10:28:07JdGordonfuck :p
10:28:10Mikachuhahaha
10:28:12jhMikeSaudio_get_buffer, no it always returns the start location of the buffer as it is with the option to grab voice resources. buffer_alloc chips away at the start and increments audiobuf
10:28:15fulhackJdGordon, Okay, thanks :)
10:28:47jhMikeSit's also automatically longword aligned by buffer_alloc so no worry there either
10:28:50 Join webguest86 [0] (i=553d125a@gateway/web/cgi-irc/labb.contactor.se/x-cad2779ec6f92ddb)
10:28:53 Part webguest86
10:29:43 Quit midkay ("Leaving")
10:29:55 Join midkay [0] (n=midkay@rockbox/developer/midkay)
10:30:29JdGordonhang on, im confusd a bit... I want to grab some buffer from the begining of he audio buffer, but keep any other buffers that allcated already... we can do this?
10:30:58jhMikeSJdGordon: ones allocated already are in front of audiobuf. look in buffer.c
10:31:13jhMikeSor maybe that's behind it :)
10:31:20*JdGordon doesnt have access to the code atm :(
10:31:38Mikachuviewsvn?
10:31:52 Join redbreva [0] (n=chatzill@host86-133-125-60.range86-133.btcentralplus.com)
10:32:20 Join ack [0] (n=ackbahr@d83-181-113-109.cust.tele2.ch)
10:32:20JdGordonbbs
10:32:46jhMikeSyou just call audio_get_buffer to get the pointer and size (or just ignore it in this case), then call buffer_alloc. buffer_alloc allocates the permanent buffer for you. after that, worry not about doing anything else.
10:32:56 Quit midkay (Client Quit)
10:34:05 Join midkay [0] (n=midkay@rockbox/developer/midkay)
10:35:40 Join thegeek [0] (n=thegeek@s026b.studby.ntnu.no)
10:35:51amiconnjhMikeS: Are you really sure this will work?
10:36:10amiconnI'm quite sure this won't work on hwcodec at least
10:36:38amiconnThe comments in buffer_alloc say that you must not call it after audio is initialised
10:36:45amiconn...iirc
10:37:20 Quit GodEater_ (Read error: 110 (Connection timed out))
10:37:51jhMikeSamiconn: I'm sure it will work for swcodec since that was my intent from the start.
10:39:08jhMikeSYou will have to grab voice resources to use buffer_alloc safely as well.
10:41:19ShaidJdGordon: You're welcome!
10:41:26jhMikeSPerhaps the comments should be updated to reflect all that. Why won't it work for hwcodec? you either have voice or audio there so if mpeg.c can be told the buffer is trashed just like swcodec is it should know to reinit.
10:45:41nlsJdGordon: I found a small menu "bug", the first entry in the Replaygain menu should be "Enaable Replaygain" and it si currently just "Replaygain"
10:45:50nls-a
10:46:48 Join ackbahr_ [0] (n=ackbahr@d83-181-113-109.cust.tele2.ch)
10:47:46 Join JdGordo1 [0] (n=jonno@c220-237-57-32.smelb1.vic.optusnet.com.au)
10:48:58jhMikeSamiconn: I guess I could look at mpeg.c and see what's involved in having them both able to do it. Suppose it's better to have them have similar behavior.
10:50:08amiconnI know what would be still better: playback engine unification...
10:50:23JdGordo1with inbuilt support for m0b
10:53:50jhMikeSamiconn: this is true, but least other part of rb won't have to know if they're unified or not
10:55:17JdGordonnls: i probably wont get a chance till tomorow night to look at it. so could yo mention it in the thread so i dont forget please?
10:56:00nlsJdGordosure
10:56:14nlsumm, JdGordon: sure :-)
10:57:41 Join ender` [0] (n=ender@84-255-206-8.dynamic.dsl.t-2.net)
10:58:07nlsJdGordon: ha it was already mentioned by PaulJam on feb 21, guess you missed it...
10:58:34JdGordo1probably :p
10:59:10 Quit JdGordon (Nick collision from services.)
10:59:17 Nick JdGordo1 is now known as JdGordon (n=jonno@c220-237-57-32.smelb1.vic.optusnet.com.au)
10:59:25 Join funky [0] (n=repulse@unaffiliated/funky)
10:59:29 Join entheh [0] (n=purr@88-106-153-112.dynamic.dsl.as9105.com)
11:00
11:00:40JdGordonoh crap.. went over my download limit :'( stuck at 64kb/s download till the end of the month
11:01:03ShaidJdGordon: Sorry. :(
11:01:21JdGordonhehe dw, it was worth it
11:01:41 Quit illriginal ("Leaving")
11:01:58 Quit ack (Read error: 110 (Connection timed out))
11:06:56 Quit ackbahr (Connection timed out)
11:08:21jhMikeSmpeg.c is so much closer to the hardware than playback.c...and the procedures quite different. I'm wondering what level of unification is desired or possible.
11:09:54 Join linuxstb_ [0] (i=5343d4aa@rockbox/developer/linuxstb)
11:09:55jhMikeSthere does seem to be some common code though...probably a good place to start
11:10:31linuxstb_dan_a: Any progress with rolo?
11:15:41amiconnjhMikeS: I think that adapting the swcodec engine for hwcodec would be the best approach, dropping mpeg.c and introducing a new low-level mas handling
11:15:46amiconnbbl
11:15:49 Quit amiconn (" HydraIRC -> http://www.hydrairc.com <- The future of IRC")
11:18:51 Quit ackbahr_ (Success)
11:19:12***Saving seen data "./dancer.seen"
11:22:25 Join Hub441 [0] (n=Hub441@pD9FA120E.dip.t-dialin.net)
11:22:28Hub441hi!
11:22:50Hub441which tool should i use to set the albumartist tag on my compilations? (linux)
11:26:23 Quit Mouser_X (Read error: 110 (Connection timed out))
11:29:57 Quit ender` (Read error: 104 (Connection reset by peer))
11:31:33 Quit JdGordon ("Leaving.")
11:32:35nlshmmm, it seems replaygain is not applied when rockbox starts... I have to enter and exit the replaygain menu to make it apply...
11:34:32jhMikeSn1s: since yesterday maybe?
11:35:29jhMikeSnls, pardon (that's not a 1 :)
11:35:40nlsjhMikeS: I havent tried with older builds yet, just saw it happend with todays build...
11:36:26jhMikeSI don't doubt it was introduced yesterday in my commit...I will have a look pronto
11:36:35nlshtanks :-)
11:39:32nlsjhMikeS: seems you are right, it works fine with a build from Feb 23
11:41:05 Join Hub441_ [0] (n=Hub441@pD9FA13CE.dip.t-dialin.net)
11:41:07jhMikeSI had a feeling there might be some quirks with that. the dsp is much more conservative about calling functions now.
11:41:29markunfulhack: jungti1234 has a E10, but he's not a dev, only does korean translations
11:41:57fulhackmarkun, Ah, that sucks :\ It's a really great device.
11:42:27markunbut it probably is very similar on the inside compared to the other new irivers (and possibly the nano 2g)
11:42:56markunI have access to a iriver T30, perhaps I can start with that
11:44:52 Quit Hub441 (Connection timed out)
11:45:01fulhackI noticed there were some irivers currently supported, with color screens and all. Maybe they differ a lot from the E10, though.. I have no idea :\
11:45:29markunfulhack: the screen is probably not the biggest problem
11:45:39 Quit Presence (Read error: 104 (Connection reset by peer))
11:45:49nlsfulhack: yes the currently supported iriver players differ a lot from the E10
11:46:06fulhackAha.
11:46:37markunfulhack: also the Meizu M6 has a similar CPU, so perhaps we can add support for that one as well in future.
11:46:50 Join JdGordon [0] (i=dced3920@gateway/web/cgi-irc/labb.contactor.se/x-f3924bdfffe94c9e)
11:46:57 Quit linuxstb_ ("CGI:IRC (Ping timeout)")
11:48:05fulhackThat meizu device looks like a lot of fun, I bet rockbox would be cool on that
11:48:20fulhackI have an E10.. What would the first step be to port rockbox onto it?
11:48:48markunfulhack: the first step has been done: http://www.rockbox.org/twiki/bin/view/Main/IriverE10Info
11:49:13nlsfulhack: find out how to run your own code on it :-)
11:49:15fulhackAh, brilliant
11:49:20fulhacknls, :P
11:49:42markunyes, a good start it to try and understand how the firmware upgrade works
11:49:56markunso we can exploit it to run our own code
11:50:56fulhackSounds like an easy way to brick the device, huh?
11:51:46markunyes, it could be
11:52:11markunbut decrypting the firmware files is useful anyway. Here you can find a lot of them: http://nyaochi.sakura.ne.jp/iriverupdate/
11:52:34markunI hope the later models (E10, U10, Clix etc) all use the same encryption for the firmware files
11:52:53markunI tried the ifp_decode program, but that didn't work
11:53:33markunbtw, also the H10 Jr is part of this new generation and not really related to the original H10
11:54:33 Join inversions [0] (n=none@cpc3-bele3-0-0-cust660.belf.cable.ntl.com)
11:55:59fulhackOnce the firmware is decoded, you end up with a hex file?
11:56:23Mikachuhex is not a filetype
11:58:16fulhackI guess you could read anything as a "hex file".. heh :) What do you end up with?
11:58:32Shaida binary file
11:58:42markunfulhack: it will be a binary with the ARM instructions and a lot of strings
11:58:53markunand the images etc
11:59:43fulhackAha
12:00
12:00:40markunMaybe we should make a page for firmwares that still need to be decrypted. The Kenwood HD20GA7 would be nice for example.
12:03:46XavierGrwhat are the methods to decrypt a firmware file anyway?
12:03:50XavierGrbrute force?
12:08:52 Quit JdGordon ("CGI:IRC (EOF)")
12:08:54markunXavierGr: I'm not an expers, but with the Kenwood firmware for example Bagder and I found out there is a 16 byte pattern by looking at the hexdump
12:09:13fulhackI saw the source for the iriver decoder was available.. http://www.rockbox.org/twiki/bin/view/Main/IfpCryptanalysis
12:09:23fulhackfor XavierGr :)
12:09:30markunfulhack: yes, but that one didn't work with the newer irivers
12:10:03fulhackAh :\
12:10:08jhMikeSnls: I'm attempting to verify but my H120 seems to apply it if I start the file from the tree right after boot. How are you going about starting playback? Just want to make sure if you updated your whole build cause that might be a reason if not.
12:10:45 Quit aliask ("Chatzilla 0.9.77 [Firefox 2.0.0.2/2007021917]")
12:11:04nlsjhMikeS: I use resume on startup, will test more...
12:11:59jhMikeSnls: will give that a try
12:12:53markunXavierGr, fulhack: I found something interesting, in the E10 firmware there is 0x6d6d6d6d6d6d at many places at regular intervals..
12:13:37XavierGrwhy do they encrypt their firmwares anyway :P
12:14:17markunAnd the same pattern is in the U10 firmware!
12:14:28fulhackThat's good :D
12:15:11jhMikeSnls: that's working for me too with flac which shouldn't matter.
12:15:53*jhMikeS is thinking old codecs
12:16:17markunfulhack, XavierGr: not in the H10_JR and T10 firmwares..
12:16:40fulhackhas the U10 been decoded?
12:16:58markundon't think so
12:17:13markunbut if you find anything let us know
12:17:27fulhackyeah
12:19:50nlsjhMikeS: Ok, tried with the latest build frpm rockbox.org, all settings reset except replaygain, no resume on startup, it still happends, but maybe I was unclear before, I have to enter a setting in the replaygain menu and then just back out, no need to change anything and after about 1-2 seconds I hear a quite large drop in volume (because the replaygain kicks in)
12:23:12 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
12:24:09jhMikeSnls: yes, old codecs shouldn't load so that's out. I tried it both ways myself and my flac starts up amplified as it should so atm I'm a bit puzzles. What format?
12:24:22nlsogg vorbis
12:24:56nlsmy mp3's and flac's does not have replaygain tags, but I will try to add some....
12:25:24jhMikeSI'll make an vorbis and see with that
12:25:41Mikachumetaflac −−add-replay-gain for .flac iirc
12:26:55 Join Benni [0] (n=chatzill@cc829402-a.groni1.gr.home.nl)
12:27:42jhMikeSnls: should have asked, which player (did you say already?)?
12:28:35nlsh320, dunno if I said it :-)
12:29:48 Join Feral_Kid [0] (n=FeralKid@red-corp-201.170.31.219.telnor.net)
12:30:21Feral_KidIf I change out the firmware on my Ipod 4G, will I lose my music as well?
12:30:33jhMikeSthat's ever creepier since they should be basically the same
12:30:40markunfulhack: no
12:30:45nlsjhMikeS: ok, tried with flac and the same thing happends...
12:32:01ShaidFeral_Kid: No, but you'll want to read up on http://www.rockbox.org/twiki/bin/view/Main/DataBase to get it to index your music
12:32:03fulhackmarkun, "no"?
12:32:26markunfulhack: sorry, that was for Feral_Kid :)
12:32:32fulhackah, np :P
12:32:40Feral_KidShaid> Thanks, I will go look at it right now... I am hoping the with the change over it will play nicely with Amarok...
12:32:44markun(auto completion :)
12:33:08ShaidFeral_Kid: with the change over you can just drop music onto the ipod wherever you feel like.
12:33:19Shaiddon't need any 3rd party software to add/remove songs
12:33:25 Quit Benni ("Chatzilla 0.9.77 [Firefox 2.0.0.2/2007021917]")
12:34:39 Quit linuxstb (Read error: 113 (No route to host))
12:45:27 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
12:47:32 Join JazzBone [0] (n=JazzBone@cc829402-a.groni1.gr.home.nl)
12:48:37 Part LinusN
12:52:09Feral_KidI was trying to download the 64-bit version of ipodpatcher, but it is not downloadable... Is there somewhere else I can find it?
12:52:09 Quit redbreva ("Chatzilla 0.9.77 [Firefox 2.0/2006101023]")
12:53:09Feral_Kidhttp://download.rockbox.org/bootloader/ipod/ipodpatcher/linux64amd64/ipodpatcher
12:53:28Feral_KidForbidden
12:53:28Feral_KidYou don't have permission to access /bootloader/ipod/ipodpatcher/linux64amd64/ipodpatcher on this server.
12:53:37Bagderhm, me check...
12:54:25Bagderretry now
12:54:52Feral_KidBagder> Thanks...
12:55:35 Join amiconn [0] (n=jens@p54BD4D5D.dip.t-dialin.net)
12:55:36Feral_KidBagder> Worked like a champ...
12:56:34Bagderit had bad permissions
12:57:52Feral_KidSo all I need to do now is unmount the ipod and reboot it and it will be good to go?
12:58:13linuxstb_Did you install rockbox.zip and rockbox-fonts.zip ?
12:58:16Feral_KidYes
12:58:23linuxstb_And you ran ipodpatcher ?
12:58:25 Join Obsys [0] (n=chatzill@p11223-ipcd01-1hon.tokyo.ocn.ne.jp)
12:58:26Feral_KidYes
12:58:31linuxstb_Then you should be good to go...
12:58:51Feral_Kidlinuxstb_> Thanks... Let's see if this is a go!!!
13:00
13:01:33Feral_KidWell, that isn't good... It said it couldn't load rockbox.ipod and now it seems that I can do select+play to get back to disk mode...
13:02:34markunFeral_Kid: but you do have rockbox.ipod in your root?
13:03:16Feral_KidNo, I just had the rockbox.zip and the rockbox-fonts.zip on the root partition...
13:03:22dan_alinuxstb_: No progress on Rolo, I'm afraid. Using the last patch I posted, you can run dual-core Rockbox, disk mode and diag mode, but trying to run the OF reboots the iPod
13:03:34nlsFeral_Kid: unzip :-P
13:03:38linuxstb_dan_a: That's some progress then.
13:03:40 Quit Hub441_ (Read error: 110 (Connection timed out))
13:03:55dan_aNone since the last I reported in here, though
13:03:57linuxstb_dan_a: Do you still have the URL, or shall I search the logs?
13:04:14MikachuFeral_Kid: you press menu+select, then select+play, just as the message says
13:04:28linuxstb_dan_a: Maybe it's worth posting to flyspray... Does it rely on the kernel_on_cop patch?
13:04:33Feral_KidMikachu> Does seem to be working...
13:04:33 Quit JazzBone ("Trillian (http://www.ceruleanstudios.com")
13:04:35amiconndan_a: Does that mean both disk mode and diag mode run on greyscale ipods?
13:04:59Feral_KidI missed that bit about unzipping the file on the device..
13:05:21dan_alinuxstb_: I don't still have it. It's separate to kernel_on_cop, but the patch is against a kernel-on-copped Rockbox
13:05:48dan_aamiconn: Disk mode doesn't work, but that's believed to be because of Rockbox messing with the usb
13:05:55 Join desowin [0] (n=desowin@avc146.internetdsl.tpnet.pl)
13:06:03 Join ubuntu [0] (n=ubuntu@c220-237-57-32.smelb1.vic.optusnet.com.au)
13:06:08Feral_KidMikachu> I meant to say that it is not working...
13:06:16linuxstb_dan_a: This the latest? http://www.pastebin.ca/370288
13:07:02 Nick ubuntu is now known as JdGordon (n=ubuntu@c220-237-57-32.smelb1.vic.optusnet.com.au)
13:07:02MikachuFeral_Kid: you have to press the second combo before the apple shows up again
13:07:08 Join ender` [0] (n=ender@84-255-206-8.dynamic.dsl.t-2.net)
13:07:09 Join bluebrother [0] (i=HK0RbU4w@rockbox/staff/bluebrother)
13:07:25 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
13:07:40Feral_KidMikachu> It has not rebooted, it is still sitting there bright as day with nothing going on...
13:07:48dan_alinuxstb_: It needs "COP_CTL = PROC_WAKE;" adding in at line 149 - apart from that, yes.
13:07:54Mikachumake sure you turn off hold
13:08:04dan_a(And it looks like I did do it against SVN rockbox)
13:08:11Feral_KidMikachu> It is off...
13:08:33Mikachuand menu+select doesn't make it reboot? you're holding it down for 5 seconds?
13:08:36 Join JazzBone [0] (n=JazzBone@cc829402-a.groni1.gr.home.nl)
13:09:50Feral_KidMikachu> Ok, it worked after the 5 seconds... So when it boots back up, immediately press select+play when I see the Apple?
13:10:11Mikachupress select+play as soon as you see it starts to reboot
13:10:16linuxstb_dan_a: And you said the 4g bootloader works fine with gcc 4.1.1 ? Is that with any patches at all, or just plain SVN? And does it install OK with the latest ipodpatcher?
13:10:19 Quit Rob222241 ()
13:13:54Mikachuwhen the 2g nano hibernates, does it encrypt the hibernation image?
13:14:10Feral_KidMikachu> Thanks... Thought I bricked it... Ok, so I need to go in the /media/ipod and do a unzip rockbox.zip and rockbox-fonts.zip, correct?
13:14:17Mikachuyes
13:14:28dan_alinuxstb_: SVN bootloader won't boot the OF. If I take out the memcmp check, then one built with GCC 4.1.1 boots OF correctly, one built with GCC 4.0.3 does not
13:14:51dan_aI had no problems installing any of these with the latest ipodpatcher
13:15:02linuxstb_OK, so ipodpatcher doesn't seem at fault?
13:15:24linuxstb_Although ipodpatcher also fails to install the very old bootloader?
13:15:38Feral_KidMikachu> Ok, now should I do a ipodpatcher again?
13:15:42dan_aIpodpatcher seems fine
13:15:45MikachuFeral_Kid: no need
13:16:00Feral_KidMikachu> And I can now delete the *.zip files, correct?
13:16:06Mikachuyes
13:16:14linuxstb_dan_a: Where should the COP_CTL = PROC_WAKE go? line 149 of rolo.c looks an odd place...
13:16:31dan_aSorry, I meant 149 on the patch
13:16:56dan_aIt's just after CPU_MESSAGE = COP_REBOOT
13:17:01linuxstb_OK.
13:17:53linuxstb_dan_a: Are you planning on looking at the bootloader some more, or do you think it's worth releasing your 4.1.1 build?
13:18:05amiconnlinuxstb_: I'm about to move the ipod lcd code to the target tree. Do you think the 'ipod' part in the source file name should be kept?
13:18:54linuxstb_Does anyone else sometimes get "arm-elf-gcc: /home/dave/rockbox/rockbox-devel/build-ipod/apps/codecs/codec_crt0.o: No such file or directory" when building with "make -j 5" ?
13:19:16***Saving seen data "./dancer.seen"
13:19:26jhMikeSnls: I also forgot to ask what rpg mode you're using (not sure you said that either)
13:19:34nlslinuxstb_: yes when building with make -j i get those quite often, both for m68k-elf-gcc and native
13:19:45 Join Rob2222 [0] (n=Miranda@p54B166E4.dip.t-dialin.net)
13:20:01linuxstb_amiconn: All the other lcd-*.c files seem to do that. Or are you just thinking "lcd-color_nano", "lcd-video" etc?
13:20:02nlsjhMikeS: track if shuffling, but I can try others
13:20:16jhMikeSI was testing with track gain
13:20:28JdGordonamiconn: did you see my message on the ml re the buttton bar? I nuked my linux setup accidently, but once i get that going again and an answer I can fix itt
13:20:39amiconnlinuxstb_: I was about to split the code into 3 files. The video code already is separate, so it'd be moved as-is
13:20:57amiconnAnd lcd-ipod.c would be split into a -color and a -gray part
13:21:03Shaidlinuxstb_: That's possibly because it hasn't made that file yet, since you've got it doing 5 tasks at once maybe it got carried away and went ahead to far...
13:21:22linuxstb_Shaid: Yes, but it means there's a missing dependency in the Makefile.
13:21:30amiconnJdGordon: On the recorders there's a global option to enable or disable the button bar
13:21:51amiconnAll screens which can hve a button bar should obey this setting
13:22:11linuxstb_amiconn: Sounds good to me. For the -color and -gray split, will you need to make two copies of the file in SVN, followed by deleting half of each?
13:22:37dan_alinuxstb_: I'm not happy with just leaving it - but I don't know enough to be able to fix it.
13:22:38amiconnHmm, that'd probably be the best solution, in order to keep history for both
13:24:03nlsjhMikeS: Tested with all 3 different type settings, same thing, the prevent clipping setting didn't affect it either.
13:24:11JdGordonamiconn: right, but does the bar actually do anything in the menu? (well, did it before the rework?) because it does nothing now
13:24:44amiconnIt just showed '<<<' for F1, to indicate that you can go back
13:25:04JdGordonok, so ill make it do that again then
13:27:37dan_alinuxstb_: Loading apple_os.ipod from a file works with a 4.1.1-compiled bootloader too. Am I right in thinking that the only time the memcpy check would fail is if someone has completely removed the OF from the boot partition?
13:28:01 Quit JdGordon ("Konversation terminated!")
13:28:54dan_as/memcpy/memcmp/
13:29:12jhMikeSnls: I will check the x5 shortly ... H120 just runs fine. Then I'll check the source code and review other recent commits.
13:31:41nlsjhMikeS: thanks, it is really strange that it doesn't happend on the h120,,,
13:31:55linuxstb_dan_a: Yes.
13:32:57Mikachui had to remove all the return code checks to get the bootloader to boot my apple_os.ipod
13:33:04linuxstb_dan_a: Have you looked at a disassembly of the bootloader? Maybe gcc is doing something odd.
13:33:05dan_aAnyone who has gone to the trouble of doing that, who hasn't created an apple_os.ipod and who tries to boot into the OF probably deserves some weird behaviour
13:33:08Mikachubefore that it loaded the file and printed the checksum, but then bailed out anyway
13:34:15dan_alinuxstb_: There was nothing obviously wrong, but I find disassemblies a bit hard to follow!
13:34:38Mikachumaybe the -keep-temps files are easier to follow?
13:34:47Mikachuiirc they sometimes have some sort of comments in them
13:35:18amiconnlinuxstb_: Do we want a separate folder for ipod video in the target tree?
13:35:22*amiconn thinks we do
13:35:57linuxstb_Yes, I think so too. But I think up to now, there haven't been any video-specific files.
13:36:20linuxstb_Most ipod stuff is shared between at least two targets.
13:36:24amiconnyes
13:36:26dan_aMikachu: I'll give that a try
13:41:20Mikachuah, it's -save-temps
13:43:07 Join webguest88 [0] (i=45e09984@gateway/web/cgi-irc/labb.contactor.se/x-a3c064957b2a8bd2)
13:43:23preglowjhMikeS: tested dsp yesterday, and it works great
13:43:38preglow~220 kbs mp3 with crossfeed, stereo width, and 5 band eq with no boost whatsoever
13:43:41preglowthat's just bloody great
13:44:28webguest88hey guys
13:44:31 Join pixelma [0] (n=Marianne@p54BD4D5D.dip.t-dialin.net)
13:44:32webguest88hello
13:44:39dan_apreglow: Tell me that was on ARM!
13:44:46dan_awebguest88: Hello
13:44:54 Quit webguest88 (Client Quit)
13:45:44preglowdan_a: no, but i hope to start optimising for arm in not too long
13:45:52preglowdon't hope to see anything like that on arm
13:45:57preglowjust mp3 alone boosts alot there
13:47:12markunpreglow: time to lower the non-boosted frequency?
13:48:22 Join lee-qid [0] (n=liqid@p549666FE.dip.t-dialin.net)
13:49:39 Join MobiLex [0] (n=MobiLex@85-156-7-144.elisa-mobile.fi)
13:49:43Shaidegh.
13:50:19ShaidI should never have bought expensive headphones
13:50:36Shaidalot of my music actually sounds lossy now
13:50:43ShaidI think I'm becoming a flac convert...
13:50:52MobiLexSennheiser cx 300 are good
13:51:03ShaidSenn hd212pro's here
13:51:17MobiLex:)
13:51:19Shaidhad them for ages.
13:51:40 Join Nico_P [0] (n=nicolas@jau31-3-82-239-20-145.fbx.proxad.net)
13:52:06MobiLexThese suck with ipod but with psp... <3
13:52:22Shaidyeah, thank god for rockbox.
13:52:36 Quit JazzBone ("Trillian (http://www.ceruleanstudios.com")
13:52:42Shaidwhy aren't they good on ipod, though? need amping?
13:53:27MobiLexThey will go cracky with high volume
13:53:38Shaidoh, nasty.
13:53:44MobiLexAt least on apple os
13:55:07MobiLexBut psp's unique-equalizer takes so care of that problem
13:57:26markunthe EQ takes care of it?
13:58:04 Join anathema [0] (n=none@cpc3-bele3-0-0-cust660.belf.cable.ntl.com)
13:58:19 Quit jhMikeS (Read error: 110 (Connection timed out))
13:58:41MobiLexIt gives a nice bass and reduces high sounds
13:58:55 Quit inversions (Read error: 113 (No route to host))
13:59:06 Join inversions [0] (n=none@cpc3-bele3-0-0-cust660.belf.cable.ntl.com)
13:59:08 Quit Nico_P (Remote closed the connection)
13:59:41 Join Nico_P [0] (n=nicolas@jau31-3-82-239-20-145.fbx.proxad.net)
14:00
14:00:05 Join jhMikeS [0] (n=jethead7@adsl-75-46-7-182.dsl.sfldmi.sbcglobal.net)
14:01:49preglowmarkun: perhaps
14:02:13preglowjhMikeS: using emac for just halving a number is a bit overkill, i think. it's also almost guaranteed to be slower. talking karaoke chan mode assembler here
14:05:13 Join Hub441 [0] (n=Hub441@pD9FA1096.dip.t-dialin.net)
14:09:47jhMikeSpreglow: Take a look at what the compiler generates and you won't be so convinced. Don't forget that it's coercing shifts (instead of using divs.l) to do rounding toward zero. This introduces forward branches to be taken 50% of the time (if no DC offset) and extra rounding code to executed the other 50% of the time in the negative case.
14:10:49jhMikeSTests also indicated emac even for this trivial operation to be faster...and I can probably speed it up more by pipelining things better.
14:11:13jhMikeSAnd besides (l - r) = -(r - l) :)
14:14:50linuxstb_dan_a: Your rolo patch isn't working at all on my Color - not even rockbox.ipod works.
14:15:46dan_alinuxstb_: Does it hang at "Executing", or does it reboot?
14:15:56linuxstb_Hang at Executing
14:16:58linuxstb_With SVN on my Color, rockbox and diagmode will rolo, but diskmode doesn't. On my Video, rockbox and diskmode will rolo, but diagmode won't...
14:17:20 Quit anathema (Read error: 113 (No route to host))
14:17:31linuxstb_I'm just compiling your patch for my video now.
14:18:17Shaiddoes rockbox support album art stored in mp3 id3v2 tags?
14:18:32Shaidin the patch, whichever
14:18:36linuxstb_No.
14:18:41ShaidThat's a pity.
14:18:51*Shaid adds that to his to-look-into list
14:19:31linuxstb_The general view is that it's better to extract those jpeg images once on your PC, rescaling using a high-quality algorithm, and storing them as bmp files, rather than making Rockbox try to do that every time you play the file.
14:19:56Shaidmaybe I'll write a program to convert the itunes format into something rockbox likes...
14:20:09linuxstb_The AlbumArt wiki page has links to tools to do that automatically for you.
14:20:17linuxstb_(I think...)
14:20:30Shaiddamnit I feel like coding something...
14:20:33Shaidoh well
14:20:40Shaidthe feeling will probably pass. :/
14:20:42dan_alinuxstb_: It looks like I've made changes since that patch. Fresh, up to date version coming in 2 minutes...
14:21:52preglowjhMikeS: i wouldn't consider the round towards inf behaviour at the last bit to matter the least
14:22:14preglow-inf, at least
14:22:27dan_alinuxstb_: Try http://www.pastebin.ca/371557
14:23:53preglowjhMikeS: remember that movclr right after a mac has a two cycle penalty too
14:24:46 Join theli_ua [0] (n=theli@197-244-124-91.pool.ukrtel.net)
14:24:55 Quit theli_ua (Read error: 104 (Connection reset by peer))
14:25:01 Join JazzBone [0] (n=JazzBone@cc829402-a.groni1.gr.home.nl)
14:26:25 Quit MobiLex (Connection timed out)
14:26:51linuxstb_dan_a: On my video, that patch starts diskmode (I haven't tested it though), the apple_os freezes at Executing, and the diagmode code immediately turns off the backlight, then seems to freeze on Executing.
14:28:02 Join Hub441_ [0] (n=Hub441@pD9FA1337.dip.t-dialin.net)
14:28:54 Join theli_ua [0] (n=theli@197-244-124-91.pool.ukrtel.net)
14:28:58 Quit theli_ua (Read error: 54 (Connection reset by peer))
14:29:03 Join theli_ua [0] (n=theli@197-244-124-91.pool.ukrtel.net)
14:29:10 Quit theli_ua (Client Quit)
14:29:17 Join theli_ua [0] (n=theli@197-244-124-91.pool.ukrtel.net)
14:29:23 Quit theli_ua (Read error: 104 (Connection reset by peer))
14:30:17linuxstb_dan_a: On my Color, both diagmode and diskmode work, but apple_os.ipod freezes on Executing. And diskmode seems to be working if I rolo it, and then attach to my PC.
14:30:27dan_aThe apple_os freezes at executing for me too, but all the others work.
14:30:47dan_aDiagmode makes a nasty hard drive click when I run it though
14:31:06linuxstb_Yes, I get that too.
14:31:58 Quit Hub441 (Connection timed out)
14:35:09Mikachudiskmode worked the first time for me, but not since then
14:35:23Mikachui get things like this in dmesg, [809346.549104] usb 1-6: device descriptor read/64, error -71
14:35:52Mikachuit displays the lightning bolt in the battery when i plug it in though
14:37:01 Quit JazzBone ("Trillian (http://www.ceruleanstudios.com")
14:37:14 Join JazzBone [0] (n=JazzBone@cc829402-a.groni1.gr.home.nl)
14:39:53dan_aThis is a pretty stupid thought, I know, but... I wonder how successful roloing the Rockbox bootloader would be
14:40:43Mikachuseems to work fine
14:43:12Mikachuseems hard to get it to boot the of though
14:43:19Mikachuit loads rockbox almost instantly
14:43:59linuxstb_Anyone know what the IP register is used for on ARM? Looking at the first line of a disassembly of apple_os.bin, the first thing it does is "mov ip,r0" - so am I right in thinking that it is expecting something interesting in r0?
14:44:03linuxstb_See here - http://www.pastebin.ca/371575
14:47:20Mikachuhttp://sources.redhat.com/ml/binutils/2000-06/msg00242.html
14:47:35Mikachu"The only difference between the standard set and the gcc is {r10, r11, r12} = {sl, fp, ip}. Under Linux I know that the got pointer is in sl, fp is the frame pointer and ip is a scratch register, most commonly used in function prologues. "
14:47:52linuxstb_Yes, I found that as well...
14:48:06Mikachuah
14:48:09 Quit XavierGr ()
14:48:21Mikachuit would suggest that it's just a random register
14:48:42Mikachui guess preglow would know
14:48:52Feral_KidIs there some way to synchronize my files on Amarok to my ipod?
14:50:16linuxstb_It's probably worth seeing what r0 contains when the Apple bootloader starts the Rockbox bootloader...
14:50:30dan_aFeral_Kid: Yes - set the iPod up as a media device in Amarok. That's off topic for #rockbox, though
14:50:55Feral_KidFor some reason Amarok is able to see all of my files, but Ipod doesn't see the names? So I am thinking that if I can synchronize the files, then the ipod database would be restored properly...
14:51:30 Part JazzBone
14:51:44 Quit pearldiver (Read error: 60 (Operation timed out))
14:53:39linuxstb_That startup code in apple_os.bin also seems to save the values of r2 and r3...
14:54:05amiconnlinuxstb_: ip is just r12 according to the instruction encoding
14:54:12markunBagder: what do you think about having a bot here to anounce SVN commits?
14:54:19Feral_KidDoes that sound reasonable or am I missing some trick to get the files to show up on the Ipod?
14:54:35MikachuFeral_Kid: are you talking about rockbox or the ipod firmware?
14:55:19Feral_KidThe rockbox... Thanks to your guidance is seem to be loaded correctly, but I can't seem to find any of my files...
14:55:40MikachuFeral_Kid: they're in Ipod_Control if you still have them as managed by the itunes system
14:55:52Mikachuyou'll need to enable the database or 'view all files'
14:56:19markunFeral_Kid: this is a useful document: http://www.rockbox.org/twiki/bin/view/Main/IpodFAQ
14:57:33 Quit Hub441_ (Read error: 110 (Connection timed out))
15:00
15:05:40 Join Thundercloud [0] (n=thunderc@82.153.143.201)
15:06:22Feral_Kidmarkun> Thanks, reading it now...
15:09:06 Quit kubiix (Read error: 104 (Connection reset by peer))
15:09:17 Join kubiix [0] (n=Miranda@mos-81-27-201-28.karneval.cz)
15:18:22jhMikeSpreglow: well, doing -inf does cause a tiny DC error. I've also experienced that doing something useful between the movclr.l and the use of the destination register (or reuse of the accumulator) can be when makes the difference. Using all four in a line loop like the sample writing does is the best way to go really. movclr.l from an accumulator from a long ago completed mac than the one you just initiated the mac on will not stall the pipelin
15:19:16preglowjhMikeS: it will
15:19:20***Saving seen data "./dancer.seen"
15:19:20preglowa coldfire bug
15:19:53preglowdoesn't care what accumulator you touched last, the movclr will stall if called too soon
15:25:13preglowamiconn: got time to test another sw tone controls patch soon? think i'll commit it soon
15:26:02preglowme likes reusing "soon"
15:30:03preglowjhMikeS: the dc error will be tiny indeed, and since there is no further amplification after channel processing, it'll never be used in the output
15:30:31preglowhell, even if a full 24 db lowshelf filter came right after it, the dc error would still not be seen in a full 24 bit quality output
15:30:52preglowbut it is nitpicking, performance shouldn't matter THAT much in karaoke mode channel processing...
15:31:03linuxstb_dan_a: If I've done it right, on my video, the Apple bootloader is starting our bootloader with r0=0x40000050, r2 = 0x000000AA and r3 = 0x80000000. On my Color, r0=0x40000050, r2 = 0x00000000 and r3 = 0x80000000. These are the registers the apple_os seems to be preserving on entry.
15:31:48preglowr2 = 0xaa? that's the cpu id code for cop
15:34:55linuxstb_Here's my patch which saves and displays them - am I doing something stupid (I'm not confident in my ARM skills...) ? http://www.pastebin.ca/371623
15:35:46preglowlooks good
15:36:10linuxstb_I'm pretty sure the end of IRAM isn't being used by the bootloader.
15:36:45linuxstb_BTW, that comment is wrong, I'm storing them 98.5KB into IRAM.
15:36:53preglowit's used as stack in diag mode, i know
15:37:08preglownothing i've seen touches that
15:37:18Mikachuin case you were wondering, the diag mode is identical in the 1.0 and 1.3 apple firmwares
15:37:23Mikachui checked the md5sums
15:37:30linuxstb_For the Nano I assume?
15:37:32Mikachuyeah
15:37:44linuxstb_What about the diskmode code?
15:37:49Mikachuit's different
15:38:18preglowit should be
15:38:22preglowit get rids of several bugs
15:38:23Mikachuyeah, since it works now
15:38:27linuxstb_preglow: One flaw in my plan - that code will be executed by both CPUs...
15:38:39preglowlinuxstb_: then do a quick check for what you are
15:38:50 Quit midkay ("Leaving")
15:39:08linuxstb_Yep.
15:39:23amiconnpreglow: I'm currently busy splitting lcd-h100-remote.c
15:39:36preglowamiconn: sure, no rush
15:39:37Mikachudoes anyone want the flash contents from 1.0 nano?
15:39:46preglowMikachu: wouldn't hurt
15:40:36Mikachuhttp://mikachu.ath.cx/nano-1.0_internal_rom_000000-0FFFFF.bin
15:40:55preglowis itunes the only official way of updating an ipod now?
15:41:21 Quit Mouser_X (Read error: 110 (Connection timed out))
15:41:22Mikachuif you're asking me i have no idea, i just wrote Firmware-xxxx to sda1
15:41:28Feral_KidI was just reading about the codec... Is MIDI possibly on the Ipod or was that something only for the iriver?
15:41:40linuxstb_preglow: I believe so, yes. With itunes 7 the standalone updaters disappeared from Apple's website.
15:43:33 Quit Obsys ("Chatzilla 0.9.77 [Firefox 2.0.0.2/2007021917]")
15:46:21 Quit pixelma (" HydraIRC -> http://www.hydrairc.com <- State of the art IRC")
15:47:42preglowjhMikeS: btw, unfortunately the uda datasheet does not elaborate on its definition of cutoff frequency. the ordinary defintion is -3 dB point, but the datasheet can't mean that, since such a cutoff at 200hz would interefere massively with mid freqs
15:51:19 Quit YouCeyE (Read error: 104 (Connection reset by peer))
15:52:08 Quit miepchen^schlaf (Read error: 113 (No route to host))
15:52:24 Join miepchen^schlaf [0] (n=hihi@p57B96876.dip.t-dialin.net)
15:53:41jhMikeSpreglow: about the stall, I can only dig up info from Freescale re: the mcf5282 and the incorrect stall and that V2+
15:54:28jhMikeSpreglow: right it seems a bit odd, but that's all the info it gave
15:54:35preglowi'll link you up
15:55:23preglowhttp://www.freescale.com/files/dsp/doc/errata/MCF5249E.pdf
16:00
16:02:08 Join midkay [0] (n=midkay@rockbox/developer/midkay)
16:03:32jhMikeSwhatdya know.. same paper as for the 5282. Well, I suppose I can rearrange things in several places. I searched and dont see one for the 5250 though.
16:03:49preglowi can almost guarantee you it's the case there as well
16:03:55preglowthe best thing would of course be to measure
16:05:19jhMikeSsuppose the insertion of tpf instructions between mac and movclrs, then after movclrs following macs would tell. there should be speed difference then
16:06:28jhMikeSI also don't know why we have noops in the fat driver for a delay when for coldfire tpf is the real noop.
16:08:58 Join akaidiot [0] (i=not@c-521d70d5.020-58-6b736411.cust.bredbandsbolaget.se)
16:09:08markunamiconn: what do you think about removing the LRU accounting from the font cache and just remove a glyph randomly when the cache is full?
16:13:44 Join ackbahr [0] (n=ackbahr@d83-180-237-125.cust.tele2.ch)
16:16:36 Join petur [0] (n=petur@rockbox/developer/petur)
16:19:46markunCan anyone take a look at my ideas for a improved tagcache? http://130.89.160.166/rockbox/newfontcache.txt
16:20:46preglowtagcache? :>
16:21:05markunsorry, fontcache of course :)
16:21:33preglowi don't even know how it works now, so i don't think i have much to contribute
16:22:15markunpreglow: I believe that it now searches through a linked list to find a glyph and does LRU accounting for every glyph lookup
16:22:35markunI will check the code again..
16:24:23markunI also have an idea to make better use of the available memory for variable width fonts
16:26:19 Quit Nico_P (Remote closed the connection)
16:27:43markunpreglow: ah, an my idea also makes multiple font support easy!
16:28:04preglow\o/
16:29:55 Quit ender` (Read error: 104 (Connection reset by peer))
16:34:16 Quit Seed (Read error: 145 (Connection timed out))
16:35:46 Join Seed [0] (i=ben@bzq-84-108-237-178.cablep.bezeqint.net)
16:36:21 Quit ackbahr ("Chatzilla 0.9.75 [Firefox 1.5.0.10/2007021601]")
16:36:55markunpreglow: I've update the document
16:41:40preglowdoes the lru accounting have much overhead?
16:43:01markunsome, you need to change some pointers every time you access a glyph and it takes up memory
16:44:01markunbut we can add it to the lookup table as well to test it
16:45:24markunsimply add a more-recenty-used and less-recently used pointer
16:48:49markunbut it will increase every lookup entry by 8 bytes and you need to change the pointers of the glyph every time it is looked up. I think random removal will work better in the end.
16:50:30preglowi'd do some benchmarking, i think, unless the performance impact is easily predictable
16:50:37preglowbut i don't know if i like the sound of random removal
16:50:45preglowsome deterministic function would be better, if at all possible
16:51:32peturmarkun: why not put this in the wiki?
16:51:48markunpetur: I can do that
16:52:20amiconnmarkun: I would expect lru to be much better than random
16:52:30markunamiconn: why?
16:52:35amiconnrandom removal will cause more dik accesses
16:52:44amiconn*disk
16:53:39markunbut it also saves more bytes which can be used for bitmaps..
16:54:12markunI will implement both with some ifdefs and then we can test it easily, ok?
16:54:56preglowgood point, bitmaps are pretty small
16:56:05peturhow about a counter? if space is needed, kill the one with the smallest counter. increase counter on access. not lru but more efficient as random
16:56:45peturI admit it has some flaws
16:57:12amiconnHmm. Removing the one with the smallest counter requires checking all loaded glyphs when one needs to be removed
16:57:31peturtrue
16:57:31markunwith random you don't need to check any glyphs ;)
16:58:00peturand you'd need to keep all counters for more efficiency...
16:58:19*petur goes minding his own business
16:58:24markunI'll just do it and then we can try out different stragies, ok?
16:59:39amiconnDoes lru currently work using a linked list?
17:00
17:00:37markunyes, looks like it
17:00:43markun(firmware/lru.c)
17:01:00markunlook at lru_touch
17:01:20Feral_KidWell, two hours later and finally rockbox became usable... For 1 1/4 songs and then is locked up... Now I can't even get it to shut down...
17:02:08markunFeral_Kid: I've never seen anyone with so many problems getting rockbox running ;)
17:02:22*amiconn thinks lru might be possible with a single-linked list and 2 pointers, without needing to walk the whole chain when one entry needs to be moved
17:02:37amiconn2 global pointers I mean, not per entry
17:02:49Feral_KidNope, strike that, I was able to shutdown and reboot... But I need to figure out why it just locked up...
17:03:16amiconnHmmm......
17:03:20 Join Rondom [0] (n=Rondom@p57A97E7D.dip.t-dialin.net)
17:03:27 Join XavierGr [0] (n=xavier@ppp91-203.adsl.forthnet.gr)
17:03:42Feral_KidThe first issue that I ran into was going into charging mode without it going into drive mode...
17:04:52Feral_KidIf I hit the menu button, and then plugged in the cable, I would still get the USB graphic. After about 2 minutes, it would mount...
17:05:25Soapwhich target again Feral_Kid?
17:05:58Feral_Kid4G Ipod
17:06:13Soapare you using a no-scaling build?
17:06:43Feral_KidUh, I am using what ever was available for current download
17:06:55Feral_KidI am not sure what it is...
17:07:05SoapI'll take that as a no. Check out mikeage's builds in the "unsupported builds" forum. They will fix your problem.
17:07:22SoapYou want to try the no-frequency-scaling one I believe.
17:07:40preglowwhy is scaling enabled when it's known to be buggy?
17:08:19Feral_KidSoap> Where do I find that?
17:08:23 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
17:08:35Soapin the "unsupported builds" forum
17:09:33preglowamiconn: btw, do you think i should use the new prescaling hooks to add prescaling also for the eq?
17:10:36Feral_KidOh, in the forums that doesn't seem load right this minutes? :)
17:10:41preglowthey load fine here
17:11:16 Join pearldiver [0] (n=say@cpe-66-65-88-127.nyc.res.rr.com)
17:11:57 Join Thundercloud_ [0] (n=thunderc@82.152.200.20)
17:12:19Feral_Kidpreglow> Ok, they are not loading on my side of the world... :)
17:12:35 Join redbreva [0] (n=chatzill@host86-133-125-60.range86-133.btcentralplus.com)
17:12:58Feral_KidAh, there it goes... :)
17:13:02preglowi believe this is what you want anyway:
17:13:05preglowhttp://mikeage.net/content/rockbox/index.php
17:15:43 Quit Thundercloud (Read error: 113 (No route to host))
17:19:21***Saving seen data "./dancer.seen"
17:20:04Feral_Kidpreglow> Ok, I now have a file rockbox-bootloader-ipod4g.ipod ... Do I make that executable and install it on the ipod or do I change it back to rockbox.ipod and then load it?
17:20:14preglowthe last
17:20:22preglowjust copy it to the ipod as rockbox.ipod
17:20:31preglowbut i think you'll also need to update the plugins and stuff
17:21:16Feral_Kidpreglow> Ok, how do I go about doing that?
17:21:24 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
17:21:26preglowwouldn't know, haven't tried any custom builds
17:21:35preglowthey should be bundled with the custom build
17:22:49Feral_Kidpreglow> Hmmm, the only thing I saw on that page was the rockbox-bootloader-ipod4g.ipod ....
17:22:58preglowoh, bootloader
17:23:00preglowforget what i said
17:23:02preglowi misread
17:23:10preglowgive me a sec and i'll check it out
17:23:51jhMikeSpreglow: I'm thinking I want to test the mac; movclr, Rx <do something> <use Rx> sequence vs. mac; movclr Rx <use Rx> <do something> as well. I witnessed a 5% boost drop using the first over the second when first messing with the spc codec. I switched it back and forth several times to verify. It seems to suggest not all is documented.
17:24:35preglowFeral_Kid: you see those select boxes almost at the top? choose build name "noscaling" and the other fields appropriately, then press "download selected build"
17:24:44 Join Criamos [0] (n=Criamos@p54931DF6.dip0.t-ipconnect.de)
17:25:04preglowFeral_Kid: then just unzip the zip you download directly to your ipod, and you're good to go
17:25:22preglowjhMikeS: no surprise there
17:25:27preglowjhMikeS: with regards to the documentatin
17:25:38amiconnjhMikeS: If you delay the use of Rx, you can as well delay the movclr to be on the safe side
17:25:58 Join Hub441 [0] (n=Hub441@pD9FA13C1.dip.t-dialin.net)
17:25:58 Join web-taz [0] (n=taz@p5081B87B.dip0.t-ipconnect.de)
17:26:08Feral_Kidpreglow> Thanks... What is the deal on COPS vs unscaling?
17:26:17preglowFeral_Kid: i don't really know
17:26:25preglowi haven't followed the discussion too closely
17:26:28amiconnOh, and btw, the somewhat strange instruction order in lcd_write_yuv420_lines() for X5 and H300 is due to this emac stall...
17:26:51jhMikeSamiconn: ok...so it's for real and it really the use of Rx that can cause the stall
17:27:22Feral_Kidpreglow> I am reading the comments below, but I guess I will stick with the unscaling one for now...
17:27:32Feral_Kidpreglow> Again, thanks...
17:27:34crashdhi kids!
17:27:39preglowFeral_Kid: you're welcome
17:27:51preglowhello uncle crashd!
17:27:56crashdheh
17:27:58Soapthe problem with crashes will be reduced with a COP build, Feral_Kid, but the problem will be eliminated with a noscaling build.
17:28:00crashdwhat's the status of rb on the sandisks?
17:28:03jhMikeSamiconn: never mind...misread what you intended. I'll delay it as much as possible if possible.
17:28:06preglowcrashd: good, no sound
17:28:10Soapcrashd: no audio
17:28:12crashd:| booo
17:28:13preglowcrashd: with sound support possibly on the way
17:28:16crashdwhat's the stumbling block ?
17:28:25preglowcrashd: and audio driver exists, but is written under nda so can't be released
17:28:27amiconnjhMikeS: 2 instructions between the last mac and the first mov(clr) are enough
17:28:32preglows/and/an
17:29:26 Join YouCeyE [0] (n=YouCeyE@unaffiliated/youceye)
17:30:23jhMikeSamiconn: because of the 3 stage pipeline, correct? It takes three clocks from one end to the other.
17:30:37preglowanyone know how far off from langv2 cleanup commit we are?
17:30:45crashdahh, hmm
17:30:48preglowjhMikeS: because of the emac pipeline
17:31:16preglowit computes a single multiply and accumulate over several cycles, with each part of the pipeline being usable to new instructions each cycle
17:31:18jhMikeSright...I just wanna keep it filled and not stall it. Though I do want to test out things
17:31:31preglowyou wanna keep it filled until you're done, at which point you want it to empty
17:31:42preglowwithout you having to wait for it
17:31:52preglowtesting is encouraged, the motorola docs are often ambiguous
17:31:54jhMikeSpreglow: I figured that's what piplinling is for :)
17:33:37jhMikeSbut as far as the exact behavior, thing still seems ambiguous and some results seem to suggest that it's use of the register by another instruction within the pipline time. If we have two ways to avoid it, things are more flexible for us.
17:39:29Feral_KidSoap> Will this also rid me of the issue of not going into drive mode when I hold down the menu button before inserting my cable?
17:39:55jhMikeSthis is the sort of thing the profiling is for. I have no idea how to use it atm :) Better get crackin' and write a test plugin.
17:40:55jhMikeSWell, _I_ will write a test plugin, if what I said might be interpreted as _You_ write one :)
17:41:20preglowheh
17:41:31preglowi am not susceptible to such notions :>
17:41:41preglowyou interested in testing another sw tone controls patch, btw?
17:41:50 Join Nico_P [0] (n=nicolas@jau31-3-82-239-20-145.fbx.proxad.net)
17:42:13jhMikeSoh why not? link me up
17:42:43*jhMikeS just realized his H120 has been usb connected sans charger for awhile...oops
17:43:15preglowdiffing now
17:44:56preglowhttp://www.pvv.org/~thomj/rockbox/sw_tone_controls.patch
17:45:05preglowonly x5 has it enabled by default
17:46:13jhMikeSpreglow: sure there's nothing in dsp.c that would qualify me as a code butcher (since you seemed a bit put off at the start)? do tell if anything sound be done better.
17:46:23preglowi'm not put off, i think it looks great
17:46:45preglowi was a bit depressed at times yesterday so may have seemed a bit glum, but nothing meant by it
17:47:55jhMikeSok, cool. hls was having problems with replay gain being applied after boot on H300 but H120 and x5 seem to be ok and I can't figure why it should only affect H300s. A build from the day before doesn't have the problem for him.
17:48:09preglowweird
17:48:17preglowi'm not that familiar with it yet to be able to help you much there
17:48:34jhMikeSBefore, it was confoundation with recording and H300s :\
17:48:49jhMikeSH300s hate me
17:49:56preglowdo tell if something in this patch could be done more cleverly, btw
17:51:04jhMikeSIt'll be a bit gotta get a source tree updated and build it
17:52:03preglowsure, no worries
17:54:13jhMikeSmore fuzz in eq.c and eq.h :\
17:55:10jhMikeShmmm...power is flickering here. I hope it doesn't go out. It's freezing rain.
17:55:35markunpreglow, amiconn: we can add a "bool found" which we can set to false if a glyph was not found, avoiding to access the disk again next time we look it up.. what do you think?
17:55:42 Join Domonoky [0] (n=Domonoky@p549AE541.dip.t-dialin.net)
17:56:17 Quit bonbonthejon ("Konversation terminated!")
17:56:30 Quit Mouser_X (Nick collision from services.)
17:56:39 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
17:57:02 Join toffe [0] (n=chatzill@adsl-70-137-198-225.dsl.frs2ca.sbcglobal.net)
17:58:23preglowjhMikeS: i don't get where the hell that fuzz is coming from
17:58:32preglowi do svn up, make patch, and it's still bloody there
17:59:09jhMikeSI wonder if it's because I saved through ie and it's stripping trailing crs. Can't think of anything else if it's made from latest cvs.
17:59:25preglowi get the fuzz even locall
17:59:27preglowy
17:59:28preglowso noi
17:59:28jhMikeSoh
17:59:30preglowbhragh
17:59:34preglowkeyboard hell
18:00
18:00:00jhMikeSok...make zip and should be running
18:00:49preglowreplaygain isn't asm yet?
18:00:58preglowapply_gain, that is
18:01:09jhMikeSpreglow: no...wanted to get to that soon
18:01:22preglowyep, it deserves it
18:01:28preglowif we want to use prescaling for eq too, then it'll be needed
18:01:32jhMikeSso, did you test an mp3 with all the crap running? :P
18:01:36preglowi did, no boost
18:01:40pregloweven ~220kbps vbr
18:01:47jhMikeSAll mine are like that
18:02:03preglowthings are really tight
18:02:07preglowthat's a lot of shit running
18:02:10jhMikeSx5 might get a little boost but that's the damn slow lcd
18:02:40amiconnhrmph!
18:02:51jhMikeSflicking the backlight off stops it since the display is off
18:03:12pregloweven 320kbps with crossfeed, stereo width, and 5 band eq doesn't boost
18:03:15preglowthat's _really_ tight
18:03:16jhMikeSamiconn: what do you want from the stupid 9-bit bus? :P
18:03:18amiconnWhy does including lcd-remote-target.h from lcd-remote.h (for targets) cause sysfont.h not to be created??
18:03:42amiconn'make' has a strange logic...
18:03:52jhMikeSpreglow: I can do better...and then my dap might time warp
18:04:14amiconnThe X5 lcd isn't that slow at all
18:04:17jhMikeSI saw more optimization
18:04:30amiconnAll ipod colour lcds are slower, especially the video
18:05:17jhMikeSamiconn: Well, then I feel kind of bad for them. Ever since using 16-bit only it gets 65.5 fps but it's still a fair load and visible on the boostometer
18:06:06*preglow pats himself on the back for eq_cf.S
18:06:34amiconnEven the greyscale ipods only get roughly the same fps...
18:06:42amiconnThe H100 lcd is the fastest lcd by far in rockbox
18:06:59jhMikeS?? H100 lcd is like it's not even there and then was gonna mention the H300 :)
18:07:01preglowa nice lcd for a nice dap
18:07:10amiconn(and hopefully soon the M5, which has the same lcd controller)
18:07:20jhMikeSexecpt it gives me nightmares
18:08:19amiconnThe H300 lcd is fast-ish with dma, but still not very fast
18:09:07jhMikeSpreglow: I've got the controls but they don't seem to be having any effect :\
18:09:14jhMikeSx5
18:09:15 Join norbusan [0] (n=norbusan@host88-222.pool8175.interbusiness.it)
18:10:02preglowjhMikeS: bah, i know why
18:10:53jhMikeSwell, patch me up again then or is it a simple edit?
18:11:45preglowit needs entries in sound.c sound_settings_Table
18:11:47preglowtable
18:12:01preglowyou'll see tlv320 has no bass and treble entires
18:12:18jhMikeSI imagine it doesn't :D
18:12:23preglowshould be easy to fix
18:12:45preglowthat's an issue i'm wondering about, i wonder if people should be able to hook up any target with sw tone controls
18:12:56preglowvia editing config-*.h
18:13:04preglowjhMikeS: but no, just put in some standard bass and treble entires
18:13:09jhMikeSThat I don't know
18:13:16preglowthe details shouldn't matter, only the callback function
18:15:10jhMikeSwell, anything without tone controls period should be hooked up but maybe ones with inferior hw ones could be switchable ... hate to see uses be blocked from using what they find best for them.
18:15:44preglowindeed, like ipod
18:15:47preglowthey adjust volume...
18:16:12 Quit web-taz ("( www.nnscript.de :: NoNameScript 4.02 :: www.XLhost.de )")
18:16:29jhMikeSpreglow: num_decimals doesn't matter?
18:16:42preglowjhMikeS: that probably does
18:16:47preglowbut steps, max and min don't
18:16:52preglowthey get overrided in settings.c, i think
18:16:54preglowor sound_menu.c
18:16:59jhMikeSI meant steps yes
18:17:36 Quit Eimann (Read error: 101 (Network is unreachable))
18:17:50jhMikeSI just copied the entries from uda and used -24 to 24 instead
18:18:48preglowyup
18:18:54jhMikeSpreglow: what are the callback names? it can't find a reference to sound_set_bass/treble
18:19:37 Join KoCb0 [0] (i=KoCb0@217.79.65.189)
18:20:26nlspreglow: regarding langv2 cleanup, FS #6652 needs to be closed before it can be committed and lng files as well as voice files should get target specific headers (and strings in enlgish.lang should be grouped more nicely)
18:21:48preglowjhMikeS: there's just one callback, and i call that via the sound_set_bass/treble functions in sound.c
18:21:52preglowso just make sure those are called
18:23:12jhMikeSpreglow: you forgot to remove the #ifndef TLV_320 :D
18:23:19preglowjhMikeS: where?
18:23:33jhMikeSin sound.c around sound_set_bass/treble
18:23:38preglowjhMikeS: shit...
18:23:55preglowbut this is why i have testers with x5! :)
18:24:28jhMikeSI think that got it :)
18:24:44 Nick Everybody is now known as Everybody|away (n=everybod@harpo.demon.co.uk)
18:25:13jhMikeSbarely with x5 anyway...should give me about 10min of battery life now to test things :P
18:25:19 Quit Hub441 (Connection timed out)
18:28:33jhMikeSThe bass sounds good. The treble doesn't hit voice enough now though...vocals sort of disappear into the hi hats. What changed? I thought you lowered the high shelf cutoff.
18:28:55preglowthey're 200/5000
18:29:06preglowwhat did it use to be? 6000 ?
18:29:08preglowif so, i did
18:29:23jhMikeSweird...it just comes out different sounding on x5
18:29:25preglowprobably can be lowered somewhat
18:31:16jhMikeScrap...now we need a new pcm buffer. I can't stand the delays.
18:31:32preglow:D
18:33:43jhMikeSOk, what's the minimum you've got it set up to be safe at for the high?
18:34:41preglowyou mean until the coefs explode?
18:35:10jhMikeSyes
18:35:17linuxstb_Another ARM question - if I do "ldrh r0, [r1]", then the high 16 bits of r0 are set to zero?
18:36:10 Join ender` [0] (n=ender@84-255-206-8.dynamic.dsl.t-2.net)
18:36:21*jhMikeS guesses the high 16 bits are set by that from the mnenonic...just guessing though
18:36:35preglowlinuxstb_: by no means
18:36:42preglowlinuxstb_: i mean yes :P
18:37:03*preglow needs to think before he answers
18:37:11jhMikeSldrh doesn't mean "load register high"?
18:37:24 Join ackbahr [0] (n=ackbahr@d83-181-102-131.cust.tele2.ch)
18:37:28preglowjhMikeS: load halfword
18:37:49preglowlinuxstb_: ldrh zero-extends, ldrsh sign-extends
18:37:59jhMikeSah...ok like move.w (Ay), Rx
18:38:05jhMikeSoh, so not
18:38:10preglowmove.w doesn't touch the upper bits
18:38:16preglowafaik
18:38:24jhMikeSunless the destination is an address reg
18:38:38linuxstb_So why would gcc do this (when reading a short from memory)? http://www.pastebin.ca/371814
18:38:39jhMikeSso it's more like move.w (Ay), An
18:38:45amiconnhrmph
18:38:58amiconnHackish sim code ahead :/
18:39:27preglowlinuxstb_: only the lord knows
18:39:43jhMikeStrying 3500 hz
18:39:56 Quit norbusan (Read error: 110 (Connection timed out))
18:40:50amiconnlinuxstb_: Because it's stupid?
18:41:02preglowi'm with amiconn here
18:41:05jhMikeSpreglow: well, that exploded it...
18:41:18preglowjhMikeS: give me five mins and i'll verify with my app
18:41:22*linuxstb_ just wanted to make sure he wasn't being stupid
18:42:41jhMikeSI'm sure I could binary search it in about log2(4096) or so iterations
18:44:07amiconnldrh and ldrsh are both "version 4 and above" so ldrsh should work on our arm targets
18:44:20jhMikeSbut I'm too lazy so just use the app ;)
18:45:06preglowhmm
18:45:26preglowmin coef is -15.8 for that
18:45:29jhMikeSwell, now they're working
18:45:31preglowwhich should fit nicely in our format
18:45:37jhMikeSI got blasted when I resumed
18:45:46preglownot only should it fit nicely, we would be able to cut a bit...
18:46:16jhMikeS3500 seems to be operating
18:46:26preglowthen how come the white noise?
18:46:52jhMikeSIt happened when I resumed at startup...then it cleared out
18:47:08preglowwell, that's not good...
18:47:14preglowvoice enabled?
18:47:32preglowamiconn: well, it does use ldrh, it just seems to assume it doesn't clear the upper bits
18:47:51amiconnpreglow: In this case it should just use ldrsh ...
18:48:12jhMikeSpreglow: no voice
18:48:17preglowamiconn: true
18:48:33preglowjhMikeS: brb, gotta get more wood
18:48:35preglowplace's freezing
18:48:39jhMikeSbut a voice file present...will try again...with headphones _near_ ears
18:51:06jhMikeSpreglow: It blasts until I go lower the treble gain, then I can raise it all the way up to the top after that np.
18:53:31jhMikeSjust tried with no voice file present at all...same thing
18:55:30preglowhmm
18:55:55preglowso, you stop, resume, then it goes off?
18:56:08preglowbass has nothing to do with it?
18:56:34preglowamiconn: if gcc knows we have ldrh, it should know we have ldrsh too
18:56:34jhMikeSStop, power off, power on, resume. But I am using 3500Hz.
18:56:44preglowi'll check it out now
18:56:47amiconnpreglow: Tell that to gcc ;)
18:57:00preglowbloody stupid gcc
18:57:35preglowjhMikeS: btw, didn't it happen at lower cutoffs?
18:57:40jhMikeSIt also puts in unneeded cmp.l instructions on cf
18:57:49jhMikeSOh, there was no bass involved btw
18:58:13jhMikeSLower? I lowered it from 5000. If you want lower I'll check it
18:58:28preglowah
18:58:30preglowhigher :)
18:59:00jhMikeSI don't recall...will set it back and recheck. but why would it pick up the ball again later and be fine?
18:59:35preglowsHIT
18:59:36preglowhaha
18:59:40preglowit blows up, alright
19:00
19:00:21preglowdon't even need to power off
19:00:44 Nick Everybody|away is now known as Everybody (n=everybod@harpo.demon.co.uk)
19:00:50jhMikeS3500 or 5000?
19:01:00 Join miepchen^schlaf [0] (n=hihi@p57B95786.dip.t-dialin.net)
19:01:03preglow3500
19:01:12preglowi suspect it doesn't matter'
19:01:20jhMikeSabout to find out
19:01:38 Quit ackbahr (Remote closed the connection)
19:01:55jhMikeS*earphones _near_ ears and go!*
19:02:01preglownot too close...
19:02:11preglowbass does the same
19:02:19jhMikeShehe...yep that was too close :)
19:02:25preglowah, hmm
19:03:11jhMikeSisn't set_tone_controls called at init?
19:03:35preglowtone_set_controls is called whenever the prescaler changes in sound.c
19:03:47preglowwhich should be whenever volume, treble, bass or balance is set
19:03:59preglownow the filter history has locked completely up
19:04:00preglowall i get is dc
19:04:51preglowamiconn: oh, and btw, why should it do lrsh? that code zero-extends, no sign-extends
19:06:50amiconnThat's true... but the both shifts are still superfluous
19:06:57amiconns/the/then/
19:08:00preglowoh, indeed'
19:09:28preglowthere's no denying that
19:09:35preglowbut gcc being stupid is... unsurprising
19:10:37preglowjhMikeS: weird bug, this, think i've found the cause, but god knows why it does what it does
19:11:08jhMikeSwhat is it you think?
19:11:41preglowif i remove the sound_set_volume in audio_play_start(), it works when stopping and resuming
19:11:48preglowam investigating it
19:13:03 Quit Mouser_X (Nick collision from services.)
19:13:12 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
19:13:22jhMikeSso there's nothing in there that depends on the correct codec being present?
19:13:54preglowcorrect codec?
19:14:04jhMikeSappears to not depend
19:14:31preglowshouldn't ever care
19:14:36jhMikeSoh never mind...I have no voice file so there's no voice codec to be loaded
19:14:54jhMikeSand I see nothing that depends on that...it's all global
19:15:07jhMikeSso...umm.odd
19:15:12preglowi
19:15:19preglowi'll add that volume call back and see if it reappears
19:15:24preglowthen i'll do some logfing
19:16:36preglowyeah, it's that, time for logf
19:17:04jhMikeScould be a deeper bug with memory useage or something now showing up. :\
19:17:53 Join Hub441 [0] (n=Hub441@pD9FA12F2.dip.t-dialin.net)
19:19:10BHSPitMonkeywhy is the WPS implementation of the progress bar so weird?
19:19:15jhMikeSpreglow: couldn't we cache the result of (bass != 0 || treble != 0) into tone_enabled so we don't have so much logic in the loop?
19:19:24***Saving seen data "./dancer.seen"
19:20:29preglowsure
19:20:59preglowwell, logf says all params tone_set_controls get are perfectly valid
19:21:06jhMikeSactually, doing if (bass | treble) would eliminate branches
19:21:21 Nick redbreva is now known as redbreva_away (n=chatzill@host86-133-125-60.range86-133.btcentralplus.com)
19:21:33pregloweliminate one at any rate
19:22:52jhMikeSbut just if (tone_enabled) would only load one variable
19:22:56preglowso keeping a tone_enabled would save us one fetch and one or
19:23:08jhMikeSyup
19:23:15preglowwhich kinda makes it a bit redundant
19:23:26preglowdsp_process can stand one fetch and one or
19:23:50preglowbut i don't get this
19:24:04preglowthe filter created should be in order
19:24:11preglowi can also reset the history to be sure
19:24:12jhMikeSyes, but it's two fetches and one or :)
19:24:38Mikachudon't you usually not write != 0, (not that it matters)
19:25:11jhMikeSoften times yes
19:25:19preglowMikachu: depends on preference, i tend to write != 0 if i'm actually checking for the value 0
19:25:21jhMikeSdifferentiates it from a bool
19:25:28preglowdoing logic i do if (value)
19:26:57jhMikeSI do that for if (enabled) but for values if (value != 0)
19:27:32preglowfilter history didn't matter
19:27:47preglowso i really have no idea what's causing this, the filter set is correct
19:28:56preglowtried with the eq?
19:29:12jhMikeShmmm...I used to get blasted by the eq like that from time to time when the config block changed and I didn't reset.
19:29:19jhMikeSwill do.
19:29:53jhMikeSthat's all history though
19:32:10jhMikeSeq seems fine
19:37:06jhMikeSthis is weird, now my x5 battery is showing 4v...something is misbehaving...or it's rising from the grave and about to eat my brains
19:40:43preglowhmm, filter coefs don't seem to be right at all, no
19:40:57preglowbut they should be stable
19:41:02jhMikeSprelglow: they're right if you want a fancy noise generator
19:41:22jhMikeSa tentitave one anyway
19:42:39preglowwhy isn't anything simple?!?!?!
19:43:35jhMikeSthat question answers itself
19:44:02preglowjhMikeS: this does not happen on ipod
19:44:15pregloware you sure you don't leave the emac in some weird state somewhere?
19:44:48 Quit Ribs ("eh eh ehhhh!")
19:44:48jhMikeSwell...then it's a stupid bug and not deliberate if so
19:45:14jhMikeSpreglow: for the heck of it, try setting everything explicitly in the function realy quick and see if it goes away
19:45:46preglowfirst, i need to eat somethingh
19:49:01 Quit Hub441 (Connection timed out)
19:49:47 Join x1jmp [0] (n=x1jmp@p57B09F14.dip0.t-ipconnect.de)
19:50:05jhMikeSI'll look for any mistakes
19:54:57jhMikeSpreglow: the only emac routine that I added that should be running is samples_output_stereo and it appears ok. It saves/restores macsr and uses movclr.l to fetch results so I dunno.
19:55:03 Quit Nico_P (Remote closed the connection)
19:55:03 Join Thundercloud__ [0] (n=thunderc@82.152.218.83)
19:58:03 Part Domonoky
19:59:06jhMikeSpreglow: I'm thinking if anything called by the audio thread uses emac then that could be it since it might be in the wrong mode for the dsp
20:00
20:00:08 Join Nico_P [0] (n=nicolas@jau31-3-82-239-20-145.fbx.proxad.net)
20:07:11preglowjhMikeS: it would pretty much have to be set to integer mode for this to happen
20:07:38preglowvery few things does that
20:08:19jhMikeSummm...but FRACMUL is called in filter_tone_coefs which ends up being called by the audio thread
20:09:07jhMikeSthen once it leaves sample_output_stereo everything should be borked for good
20:09:55preglowghaaah
20:10:03preglowwait a minute now
20:10:09preglowforget that
20:10:15preglowi'll check it out when i've eaten
20:13:10jhMikeSpreglow: I set macsr explicity in filter_tone_coefs to the dsp mode and the problem went away
20:14:04preglowbaah
20:14:08preglowthis callback was bound to bring misery
20:14:22preglowbut ok, it can be alleviated by setting macsr in dsp_callback
20:15:20preglowbeing called from outside of course means macsr can be unpredictable
20:15:56jhMikeScould set it there too sure
20:17:46preglowwhy too? why not only set it there?
20:18:10 Join illriginal [0] (n=illrigin@c-67-191-48-2.hsd1.fl.comcast.net)
20:18:20jhMikeSI meant "as in a place other than..."
20:20:12preglowright :-)
20:20:15preglowfoodery
20:20:23jhMikeSI just checked my suspicion by following things and plopped the test code in there since FRACMUL is used there. Create thread doesn't set macsr for the new thread so it's whatever it was on the first context switch
20:20:43 Quit Thundercloud_ (Connection timed out)
20:20:51 Join Eimann [0] (i=eimann@chaconne.etherkiller.de)
20:26:19 Quit amiconn (" HydraIRC -> http://www.hydrairc.com <- IRC for those that like to be different")
20:26:34jhMikeSeh, but everyone knows that so I shan't babble
20:40:16 Quit Feral_Kid (Read error: 104 (Connection reset by peer))
20:42:43 Join Feral_Kid [0] (n=FeralKid@red-corp-201.170.31.219.telnor.net)
20:47:19preglowyeah
20:47:29preglowi forgot that completely
20:47:33preglowbut ok, then that's solved
20:47:44preglowi'll stuff a macsr set/reset in dsp_callback
20:48:41 Join Hub441 [0] (n=Hub441@pD9FA12E0.dip.t-dialin.net)
20:48:44Hub441hi!
20:50:21 Join darksoulk [0] (n=adam@203-214-30-35.dyn.iinet.net.au)
20:51:52Hub441i just tried to compile a plain rockbox (no changes to the source-tarball from the website) on a gentoo-host but got a bunch of error messages: http://phpfi.com/208055
20:52:23Hub441the only difference from the howto: i used gcc 3.4.6-r2 and not r1
20:53:06bluebrotheriirc you need gcc 4 for arm targets
20:53:06preglowi've seen those messages plenty of times, but never on a default build
20:53:13preglowyeah
20:53:53scorche4.0.3 for arm targets
20:58:12 Join DX_ [0] (i=DX_@cpe-67-11-226-113.satx.res.rr.com)
20:58:36DX_hey
20:59:32 Join linuxstb [0] (i=5343d4aa@gateway/web/cgi-irc/labb.contactor.se/x-0c81d0459156c12d)
21:00
21:01:20preglowjhMikeS: but you can think of no wiser solutions than that, no?
21:01:28preglowstuffing macsr set/restore in dsp_callback
21:01:51DX_okay before i screw this process up a second time
21:01:54 Join Hub441_ [0] (n=Hub441@pD9FA1305.dip.t-dialin.net)
21:01:58DX_say i have a newer version of the bootloader
21:02:08DX_(for the ipod version of rockbox)
21:02:16DX_how would i go about installing it instead of the one included with ipodpatcher
21:02:30jhMikeSpreglow: how are you going to guarantee dsp compatible macsr but to make sure it's compatible by default for every thread ?
21:03:26preglowjhMikeS: well, by setting it explicitely like i suggested
21:03:45jhMikeSin the scheduler code?
21:03:52preglowin dsp_callbacj
21:03:58preglowi just do like i do in dsp_process
21:04:06preglowset macsr to something dsp compatible, do my stuff, restore it
21:04:07jhMikeSyes, why do you not want to now?
21:04:22preglowi want that! just wondered if you had a more clever solution in any way
21:04:48 Quit pearldiver (Read error: 110 (Connection timed out))
21:05:05jhMikeSthere is none except to make it default or so something weird by setting it in audio_thread, codec_thread and anything else that could call it.
21:05:23DX_:<
21:05:24jhMikeS*shrugs*
21:06:00preglowi think each part that needs macsr set to something specific should do so instead of relying on it, so we'll just do it this way, then
21:06:32 Quit Hub441 (Read error: 60 (Operation timed out))
21:06:39preglowwow, arm-gcc generates an amazing amount of code for shortarray[index] if you don't optimise
21:07:34jhMikeSpreglow: It doesn't show up for the eq since I suppose dsp_set_eq_coefs is called by the main thread only and that has a default macsr right/
21:07:34jhMikeS?
21:08:32preglowyeah, it does
21:08:45preglowi set that in system.c
21:09:34 Quit Shaid (Read error: 110 (Connection timed out))
21:09:35 Nick darksoulk is now known as Shaid (n=adam@203-214-30-35.dyn.iinet.net.au)
21:09:54 Quit illriginal ("Leaving")
21:09:58 Quit Mouser_X (Nick collision from services.)
21:10:07 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
21:10:50preglowlinuxstb: did a test now fetching a halfword, and gcc does what's right, using just a ldrsh
21:11:24linuxstbDX_: You install/replace a bootloader with "ipodpatcher -a bootloader-ipodxxx.ipod"
21:11:24x1jmpCould I get write access to the wiki? My account is TimoHorstschaefer there.
21:11:52DX_aha
21:11:54DX_thanks, linuxstb
21:12:00DX_hopefully third time will be a charm
21:12:25 Join darksoulk [0] (n=adam@124-168-142-179.dyn.iinet.net.au)
21:12:27linuxstbDX_: See alos http://www.rockbox.org/twiki/bin/view/Main/IpodPatcher
21:12:32jhMikeSpreglow: I suppose the default macsr for any thread should be dsp compatible then. could set it in the scheduler
21:19:25***Saving seen data "./dancer.seen"
21:19:26preglownot a bad idea, i guess
21:19:36 Join Hub441 [0] (n=Hub441@pD9FA12F1.dip.t-dialin.net)
21:19:37preglowat least dsp can forget that problem for good, then
21:19:43 Quit Feral_Kid (Read error: 110 (Connection timed out))
21:19:49 Join Thundercloud_ [0] (n=thunderc@82.152.218.83)
21:19:56 Quit Shaid (Read error: 60 (Operation timed out))
21:19:57 Nick darksoulk is now known as Shaid (n=adam@124-168-142-179.dyn.iinet.net.au)
21:20:23DX_okay so i put the bootloader on there
21:20:33*x1jmp wants write access to the wiki...
21:20:43DX_and now it just sits there displaying "loading original firmware..." and rebooting
21:20:51DX_and so on and so forth
21:22:11linuxstbDX_: What bootloader did you install, and what ipod?
21:22:45DX_http://mikeage.net/content/rockbox-custom-builds/ the 4G grayscale bootloader on a 4G grayscale ipod
21:23:10linuxstbThose bootloaders are known not to work with the latest ipodpatcher.
21:23:18DX_hm.
21:23:26DX_well.
21:23:31DX_that would explain a lot :<
21:26:43preglowlinuxstb: in what context did you make arm-gcc spit out that silly code, btw?
21:28:08 Quit jhMikeS (Nick collision from services.)
21:28:14 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
21:28:33linuxstbI was trying to get the dbg_flash_id() function in debug_menu() to work. The lines that generate the code are where flash[0] and flash[1] are read at the start of the function.
21:28:52linuxstbFor ARM, flash[] is defined the same as coldfire - unsigned short*
21:29:06jhMikeSpreglow: well, it can't in codec thread context. we already have an integer mode codec. I'm not sure where to put it in thread.c yet myself.
21:29:16linuxstbBut looking at random other .o files, I saw similar madness.
21:29:22linuxstb(although not always)
21:29:49linuxstb^that should be debug_menu.c
21:30:06preglowwhat arm-elf-gcc version do we recommend these days, btw?
21:30:11preglowi see rockboxdev still used 4.0.3
21:30:17preglowi use 4.1.0 with much success
21:30:31linuxstbI think rockboxdev is the definition of recommended now.
21:31:09preglowthink i'll try out 4.1.2
21:31:20*x1jmp asks again for write permission on the wiki
21:31:26preglowjhMikeS: codecs already have to set their own macsr stuff
21:32:31linuxstbx1jmp: Are you a spammer?
21:32:31Soapbluebrother: did you get your mini yet?
21:33:25jhMikeSpreglow: right, but dsp mode can be different from the codec's mode...so it's not free of everything. that's what I mean.
21:33:42preglowtrue
21:33:44x1jmplinuxstb: no
21:34:09 Quit redbreva_away ("Chatzilla 0.9.77 [Firefox 2.0/2006101023]")
21:34:17linuxstbx1jmp: OK, then you're added.
21:35:20x1jmplinuxstb: thanks
21:35:21bluebrotherSoap, no. Didn't went for that auction, but won another yesterday :)
21:35:52Soapahh, for I think I can score one for $75, and was curious if you still wanted one.
21:35:53 Join illdred [0] (n=illdred@S010600055dff1af6.wk.shawcable.net)
21:36:09x1jmp... have there been bots on IRC asking for wiki permission or why was that question?
21:36:36 Join matsl [0] (n=matsl@1-1-4-2a.mal.sth.bostream.se)
21:37:11bluebrothertoo bad ... mine was more expensive :(
21:37:25bluebrotherbut thanks for asking anyway. They seem to be quite higher here in germany
21:38:09 Quit Hub441_ (Connection timed out)
21:39:40 Quit Thundercloud__ (Connection timed out)
21:40:56 Quit Hub441 (Read error: 110 (Connection timed out))
21:43:27preglowwhy does rockboxdev.sh download the entire gcc source?
21:43:30preglowjust the core will do
21:44:31 Quit Mouser_X (Read error: 110 (Connection timed out))
21:45:00Soapbluebrother: I haven't gotten word back yet, might only be a 1st generation one.
21:46:49 Join fasmaie [0] (n=yohann@c-71-233-79-102.hsd1.ma.comcast.net)
21:48:17jhMikeSpreglow: I don't know but I had to clean up my debian image a lot to get it and it takes about 500MB free to work
21:49:35preglowi've got more than enough free, just thinking it's a waste to download the vastly bigger gcc source as opposed to the small and good gcc-core which just does c
21:50:22jhMikeSpreglow: my suggestion is http://www.pastebin.ca/372074
21:50:32 Quit KoCb0 ("• IRcap [8.23] • www.ircap.com/int •")
21:50:39 Quit Thundercloud_ (Connection reset by peer)
21:51:27jhMikeSor maybe a macro that is empty if a cpu doesn't need that and is defined the set special data if it needs it :\
21:53:03preglowjhMikeS: the pastebin thing is clean and sweet
21:53:44illdredhello =) I'm wondering if anyone is using rockbox on their iaudio x5 today
21:53:48preglowa code cleanup measuring in at 55 bytes plus average? :P
21:53:52preglow555
21:54:31 Join Thundercloud_ [0] (n=thunderc@82.153.10.45)
21:54:40jhMikeSpreglow: I suggest though adding a thread_cpu_init in there instead by the save/restore_context stuff. At the end, if not already defined, define it empty.
21:54:58jhMikeSwhat you talkin' about?
21:55:47preglowcommits
21:56:26preglowi wonder why it added so much, doesn't look like much code was added
21:56:31jhMikeSjust looked...lol
21:59:39 Join Plouj [0] (n=Plouj@dsl-207-112-70-4.tor.primus.ca)
21:59:40Ploujhi
21:59:46 Nick SaLoMoN is now known as SaLo^ToT (n=SaLo@salooo.org)
22:00
22:01:19 Quit Nico_P (Remote closed the connection)
22:01:24 Join SaLoMoN [0] (n=SaLo@salooo.org)
22:01:42 Part SaLo^ToT ("Verlassend")
22:01:52 Join Thundercloud__ [0] (n=thunderc@81.5.171.229)
22:02:21PloujI'm looking for some kind of a layman's changelog for rockbox builds. I'd like to teach my bro to update his player if there are any fixes that he might want to see but the problem is the cvs messages are so confusing that even I cannot undertand them.
22:03:07preglowthere just was a thread about this on the ml, i think
22:03:40SoapPlouj: Either the MajorChanges wiki page, or just update weekly.
22:04:26Ploujah, not bad
22:04:28Ploujthanks Soap
22:05:16Soapthere is little reason not to update weekly is all.
22:05:36Ploujheh
22:05:44Ploujthat's why I wrote a script that he can run to update his firmware
22:09:19Ploujgah
22:09:30Ploujwhat's with this twiki's complicated access control?
22:09:38 Join bospaadje [0] (n=bospaadj@ip82-139-84-12.lijbrandt.net)
22:09:42PloujI just wanted to make some text more clear
22:09:48preglowcomplicated access control?
22:09:58jhMikeSpreglow: you gonna add that with your patch or should I do it up real quick. testing it for bugs should be rather quick.
22:10:01 Quit fulhack (Read error: 104 (Connection reset by peer))
22:10:07preglowjhMikeS: add it separetly
22:10:11Ploujpreglow: I tried to edit a page and it directed me to: http://www.rockbox.org/twiki/bin/view/TWiki/TWikiAccessControl
22:10:18peturPlouj: to keep spammers out
22:10:45Plouj"Access check on Main.MajorChanges failed. Action "change": access not allowed on web." :(
22:10:49preglowPlouj: that's not particularily complex, is it? you don't have write access
22:11:20Ploujpreglow: I'm not even gonna waste my time reading that page. I just want to make a small page. I don't have time to learn all these made up AC rules.
22:14:06 Join Nico_P [0] (n=nicolas@jau31-3-82-239-20-145.fbx.proxad.net)
22:14:52 Quit fasmaie (Remote closed the connection)
22:15:20linuxstbPlouj: Are you registered as a user on the wiki?
22:15:22preglowahh, made up rules, as opposed to those who aren't made up
22:15:28Ploujlinuxstb: just registered
22:15:47linuxstbThen you just need to mention your username here, and someone will enable write access for you.
22:15:47Ploujpreglow: I actually meant useless, heh
22:15:54SoapThe Divine Right of Kings isn't made up - it just is.
22:16:01Soap;)
22:16:08linuxstbIf you had seen the spam we had in the past, you wouldn't call them useless....
22:16:45preglowPlouj: oh, they're not useless
22:16:53Ploujok, maybe they aren't
22:16:55preglowa fine addition they are indeed
22:17:04Soapeven the forum spam appears to have dropped off. Rockbox must have been know as an easy target.
22:17:20preglowthe page telling you you can't write could possibly have been a bit more helpful, but
22:17:32linuxstbPlouj: You're added now.
22:17:37 Join funky_ [0] (n=repulse@81.202.255.93.dyn.user.ono.com)
22:17:38 Quit funky (No route to host)
22:17:41Ploujlinuxstb: oh, thanks
22:22:17 Join bonbonthejon [0] (n=jon@69.61.203.3)
22:25:59 Quit Thundercloud_ (Connection timed out)
22:26:11 Join Thundercloud_ [0] (n=thunderc@81.168.42.92)
22:26:50ze~/math 1272-256
22:27:59linuxstb1016
22:28:12zeheh thanks :p
22:28:23linuxstbYou're welcome.
22:32:04jhMikeSpreglow: it all works with that quite nicely
22:32:14preglowcool
22:32:41preglowcommit?
22:35:41jhMikeSme yes in a couple minutes, you?
22:35:57preglowgoodie
22:35:57preglowme not yet
22:36:08Mikachuare you trying for a simultaneous commit?
22:36:12preglowhaha
22:36:20preglowwe're trying to crash svn
22:36:24 Quit Thundercloud__ (Connection timed out)
22:37:25 Quit miepchen^schlaf (Read error: 113 (No route to host))
22:37:28*jhMikeS better be careful he doesn't commit the tone controls by accident :P
22:37:47 Join miepchen^schlaf [0] (n=hihi@p57B95786.dip.t-dialin.net)
22:38:11preglowrockbox.ipod gained 5kb with 4.1.2 :/
22:40:03 Join Roan [0] (n=JustJohn@pool-151-196-120-87.balt.east.verizon.net)
22:40:07linuxstbCompared to 4.1.0?
22:42:05preglowyes
22:42:18preglowcompared to 4.0.3 i don't know, can try an unmodified build later
22:43:05 Join Thundercloud__ [0] (n=thunderc@82.152.214.92)
22:44:18RoanHi I've run into an error compiling rockbox firmware. Compiled fine on Feb 22, but today I ran the 'svn update' command; then changed into my build directory and ran 'make clean', then 'make' and I get a bunch of similar errors to this one "export/config.h:220:6: warning: "MEMORYSIZE" is not defined". It looks like a macor variable is not being defined somewhere.
22:44:46Mikachutry rerunning configure
22:45:19 Join amiconn [0] (n=jens@p54BD69C9.dip.t-dialin.net)
22:46:23jhMikeSpreglow: I guess svn up now
22:46:24 Join pixelma [0] (i=pixelma@82.193.235.34)
22:46:43preglowyup, see it
22:47:09RoanMikachu: Nope still getting the same thing
22:47:18 Quit printfXh4 (No route to host)
22:47:51preglowbeh, need to do some dishes before i do anything else
22:48:13amiconnda bineater is active again :\
22:48:42MikachuRoan: no 3rd party patches applied?
22:48:52preglowamiconn: what about that commit should increase the bin size so much?
22:48:58preglowdoesn't look like much code is added
22:49:42Mikachu tagcache.c | 193 ++++++++++++++++++++++++++++++++++++++−−−−−−−−−−−−−−−−−−−−−−-
22:49:45Mikachu 1 file changed, 123 insertions(+), 70 deletions(-)
22:50:34RoanMikachu: No, this is in my plain Rockbox build branch.
22:51:21MikachuRoan: what does the line it complains about look like?
22:51:27linuxstbRoan: Did you svn update all directories? (apps, tools and firmware)
22:51:32Mikachuhere, it's config.h:#if ((defined(MEMORYSIZE) && (MEMORYSIZE > 8)) || MEM > 8) && \
22:54:27Roanlinuxstb: yes I applied svn update from the root of my rockbox source code
22:54:48 Join tick [0] (i=95e13cb2@gateway/web/cgi-irc/labb.contactor.se/x-8922dda8741da7ad)
22:55:00linuxstbWhat target are you building for?
22:55:18RoanMikachu: I have that line in my ./firmware/export/config.h file
22:55:35Roanlinuxstb: building for an ipod nano 1st gen
22:55:36MikachuRoan: yes i know what file it is, what does the line say?
22:56:12 Join asinine [0] (n=chatzill@82-40-85-101.cable.ubr01.cast.blueyonder.co.uk)
22:56:13RoanMikachu: #if (MEMORYSIZE > 8 || MEM > 8) && !defined(BOOTLOADER)
22:56:17amiconnlinuxstb: Ooops, I always forget FILES
22:56:25Mikachuwell then your svn isn't updated properly
22:56:31linuxstbamiconn: No problem, so do I.
22:56:43tickWho can create a new category for FS patches? I think it would be good if we had a category for code cleanup and improvements without feature change
22:57:06Mikachuwhy not just select the category as what it does cleanup of?
22:57:13linuxstbtick: I read your request earlier - why not just s.....
22:57:18RoanMikachu: Okay, I will try updating again.
22:57:18Mikachuheh
22:57:20asininecan any one tell me if rockbox works on creative zen
22:57:20linuxstb(what Mikachu said)
22:57:55bluebrotherasinine, no
22:58:08bluebrother(i.e. it doesn't work)
22:58:27bluebrotherRB works on the targets stated on the front page. For all others see the New Ports forum
22:58:42asinine@bluebrother, cheers all i needed to know
22:59:23tickBut if the change is just... like fix a typo in a comment (I've seen a few) I wouldn't say it's (e.g.) drivers (e.g. if the typo is in fat.c −− just a hypothetical example)
22:59:46Mikachui would
22:59:51*linuxstb too
22:59:51bluebrotherwhy shouldn't it?
23:00
23:00:27amiconnlcd-h100-remote.c split seems to work :)
23:01:34tickok then. Hm... after a thought I think you are right. fat.c (to stick with the example) is a driver after all −− with all its parts, comments included
23:02:41pixelmaanyone who is checking the root menu too around - and could try something for me to see if it's not only an Ondio problem?
23:03:47 Quit bonbonthejon (Remote closed the connection)
23:03:58tickpixelma: can't this be checked with the sim?
23:04:17bluebrotherpixelma, can do
23:04:21pixelmahmm.. maybe
23:04:35bluebrother(h120 here)
23:06:06bluebrotherbtw, have you enabled the "show path" setting? Your browser problem sounds much like the problem I experienced earlier
23:06:12dan_aIs there any way to tell the build system to add GCC flags to individual files?
23:06:24pixelmabluebrother: ah... can you try what happens when you go to (for example) the themes browser while playing music - and then cancel it with going left?
23:06:59Mikachudan_a: i don't know exactly how, but if you find the right Makefile you should be able to copy the generic .o target and make a specific one
23:07:02 Quit Thundercloud_ (Connection timed out)
23:07:05 Quit Rondom ("Ex-Chat")
23:07:17bluebrotherhehe, got it reproduced
23:07:17Mikachusometimes it's extra tricky if the generic target is auto generated
23:07:20 Join Thundercloud_ [0] (n=thunderc@82.152.174.90)
23:07:22bluebrotherexactly as you described.
23:07:41pixelmabluebrother: so it's not just my imagination...
23:07:53dan_aMikachu: I think the generic target is autogenerated
23:08:30linuxstbThere is generic stuff in tools/make.inc
23:08:43 Join Brunellus [0] (n=luigi@unaffiliated/brunellus)
23:08:57pixelmabluebrother: I noticed that it works alright without playing music, seems a bit weird
23:09:00Mikachui want to try compiling rockbox by giving all .c files to gcc at once, but i haven't tried it yet
23:09:05RoanMikachu: Thanks for the help. Don't know what I was doing wrong with the 'svn update' command, but wiping the directory and starting over with a fresh checkout is working
23:09:14Mikachui did it with vorbis and got a .codec that was 300 bytes larger :)
23:09:21MikachuRoan: okay, np
23:09:23 Quit tick ("CGI:IRC (Ping timeout)")
23:09:29bluebrotherpixelma, just did a few tests. I have "startup screen" set to file browser and can always reproduce the bug.
23:09:42bluebrotherbut with the startup screen set to the RB menu I can't reproduce it
23:09:47 Quit asinine ("Chatzilla 0.9.77 [Firefox 2.0.0.2/2007021917]")
23:11:36preglowbeh
23:12:03dan_aMikachu: Is that to try -fwhole-program -combine?
23:12:06pixelmabluebrother: ah, well my startup screen is the root menu - but when entering the menu from WPS I have to go via the file browser... guess that's related
23:12:14Mikachui was going for -funit-at-a-time
23:12:23 Join elinenbe_ [0] (n=elinenbe@207-237-225-94.c3-0.nyr-ubr1.nyr.ny.cable.rcn.com)
23:12:25bluebrotherpixelma, just commented on the FS entry.
23:12:38bluebrotherlet's see how much JdGordon will hate me for this ;-)
23:12:43pixelmathank you! :)
23:13:20bluebrotherI left the ondio specific findings to you due to the lack of an ondio :D
23:13:40bluebrotherbut I'm pretty sure it's related, if not even the same thing
23:14:08 Quit Thundercloud__ (Connection timed out)
23:14:27 Join Thundercloud__ [0] (n=thunderc@82.153.13.103)
23:14:56preglowwell, that could have sounded better
23:15:47 Quit miepchen^schlaf (No route to host)
23:17:32preglowjhMikeS: both eq and tone controls sounds like merzbow cds now
23:17:51 Quit petur ("Zzzzz")
23:18:24Mikachuamiconn: did you forget FILES again?
23:18:38amiconnNo new folder this time...
23:18:49Mikachuah, i forgot it uses globbing
23:19:18preglowjhMikeS: does the main thread have macsr set correctly after your update, though?
23:19:19*amiconn thinks it would be a nightmare if FILES listed each and every single file
23:19:20 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
23:19:26***Saving seen data "./dancer.seen"
23:19:31preglowmy guess is no, since you removed the system.c macsr init
23:20:42 Quit elinenbe (Read error: 145 (Connection timed out))
23:20:43 Nick elinenbe_ is now known as elinenbe (n=elinenbe@207-237-225-94.c3-0.nyr-ubr1.nyr.ny.cable.rcn.com)
23:20:43desowincan anyone review my latest patch for http://www.rockbox.org/tracker/task/5495 ? if something isn't good, let me know
23:21:35jhMikeSpreglow: yes...it's set when it is added to the list
23:21:44preglowi just get white noise here
23:22:18 Quit Brunellus ("Ex-Chat")
23:22:19preglowprobably doing something stupid, then
23:22:22 Part Plouj ("thanks a lot!")
23:22:28jhMikeSwell, that's just plain odd cause I didn't at all
23:22:36 Join Feral_Kid [0] (n=FeralKid@red-corp-201.143.151.175.telnor.net)
23:23:26preglowcodecs work nice, both eq and tone contriols white noise
23:23:29preglowi'll build from scratch
23:24:00 Quit bluebrother ("time for some sleep")
23:24:06 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
23:24:18jhMikeSyou didn't take the the macsr setting out of dsp.c did you?
23:24:22LinusNdesowin: i have no comments - if it solves the problem, we should commit it
23:24:41jhMikeSin dsp_process?
23:25:05preglowjhMikeS: i did first, then i added it back
23:25:15preglowjhMikeS: should it matter anyway? i thought all threads had frac mode in macsr now
23:25:21 Quit x1jmp (Remote closed the connection)
23:25:27jhMikeSyes, but not all codecs
23:25:35preglowvery, very good point
23:25:40preglowso i'll leave the one in process, then
23:25:42jhMikeSremember spc.codec is integer mode :)
23:25:46preglowand just hope something funny happened with the build
23:25:47desowinLinusN: I'll do test while going to school tomorrow, and I'll let you know how it behaves under 'real-time situations', not just 'fast-tests'
23:25:54preglowjhMikeS: i didn't run spc either :/
23:25:56 Join miepchen^schlaf [0] (n=hihi@p57B95786.dip.t-dialin.net)
23:26:14jhMikeSI'll check again with my build from cvs and the patch
23:26:14 Quit Mouser_X (Nick collision from services.)
23:26:16LinusNdesowin: great
23:26:23 Join Mouser_X [0] (n=mouser_x@67.110.120.159.ptr.us.xo.net)
23:26:32 Join Rocco [0] (n=Rocco@71-80-129-250.dhcp.hspr.ca.charter.com)
23:27:17 Part nls
23:28:16jhMikeSpreglow: working fine here
23:28:40jhMikeStesting H100
23:28:47jhMikeSsoon
23:28:59Roccohey, i know this is like a really annoying questiong... but do you guys have any idea if/when rockbox will be avalibe on zune? if its even possible
23:29:04*dan_a thinks he may be able to shave 100k off the core for ARM
23:29:22dan_aRocco: The short answer is we have no idea
23:29:41Roccooh... no developers own a zune yet?
23:29:53BagderRocco: rockbox ports don't just happen, it takes a lot of sweat and dedicated hackers
23:30:05Rocco ya i understand
23:30:23Roccobut i got one for xmas and its kinda lame but im not rich... so
23:30:41Bagderand I would say that there are reasons to believe that zune might be the hardest nut to crack so far in this aspect
23:30:58dan_aRocco: There are a couple of people looking at the Zune and the Gigabeat S (which is similar) but I don't think they're looking very hard, and running our own code on them will be very difficult
23:31:02amiconnBagder: I noticed a slight problem with the latest changes table and the svn diff links: A rename might include a diff as well, which can't be checked because [rename] isn't a link
23:31:17amiconnHappened in both of my commits today...
23:31:29preglowdan_a: how are you going to do that?
23:31:33Bagderyou mean when you rename _and_ change contents?
23:31:44amiconnyes
23:31:45Roccoi figured its gona be real hard to do it, but once you get ability to run your own program its gonna be really easy
23:31:50dan_apreglow: -mthumb
23:31:54 Quit desowin ("use linux")
23:31:59amiconnHappens when splitting stuff, to keep history for both parts
23:32:00preglowRocco: is that so
23:32:07BagderRocco: if you think that, then prove it
23:32:12Mikachudan_a: will that not require everyone to install new compilers? ie arm-thumb-elf-gcc
23:32:16preglowdan_a: have you mad eit build?
23:32:24Roccoif i could run my own code on it i would, did i state i could?
23:32:26preglowMikachu: no, all arm gccs can do thumb
23:32:37preglowRocco: getting code to run on it is only a small part of a port
23:32:41dan_apreglow: Yes - I just don't know if it works yet
23:32:41preglowyou'll just have to trust us on that
23:33:01preglowdan_a: why not?
23:33:05Mikachui do
23:33:50dan_apreglow: Because I've not copied rockbox.ipod over yet - I wanted to boast about the size reduction first!
23:33:58preglowhahaha
23:34:01Bagderhaha
23:34:05preglowhere ten saying it doesn't work
23:34:08preglowhere's
23:34:12 Part Rocco
23:34:14preglowgcc is bound to have been playing tricks on you
23:34:16Bagderwe want the youtube video first!
23:34:42linuxstbAnd you need a logo and new graphics before you can release it.
23:34:59Bagderand a cool site
23:35:11dan_aHahaha - it hangs with the bootloader telling me "Rockbox loaded"
23:35:27preglowsurprise!
23:35:35preglowis the bootloader thumb aware, though?
23:35:44preglowdoes it use bx to branch? :)
23:35:59 Quit Thundercloud_ (Success)
23:36:04 Join Thundercloud_ [0] (n=thunderc@82.152.168.96)
23:37:04dan_apreglow: It will be branching into crt0 - which is ARM. I'm also compiling main.c as ARM
23:39:00preglowjhMikeS: can't get rid of the white noise :/
23:40:43jhMikeSI'm having no trouble at all with it.
23:40:48preglowjhMikeS: the current build also vomits white noise
23:40:53jhMikeSwhat?
23:41:02preglowit does here
23:41:19preglowthe moment i switch on the eq and tweak a band, i get white noise
23:41:20jhMikeSI just tested two different players...now I'm freaked
23:42:12preglowwhat codec do you test with?
23:42:29markundan_a: and the interrupts will be executed in ARM as well
23:43:16preglowthey're in crt0 so shouldn't matter
23:43:39jhMikeSpreglow: can test a bunch really
23:43:39 Quit Thundercloud__ (Connection timed out)
23:43:49dan_amarkun: That might be it - I've set system.c to no-thumb now, and I'll see what happens
23:44:16preglowmpc, vorbis and mpa all spout white noise here
23:44:29preglowcould anyone else on h1x0 please test a current build to see if eq thrashes out noise?
23:45:04*jhMikeS 's players have magical powers :)
23:47:15jhMikeSplaying with eq with spc codec seems ok
23:47:49markunFor ARM coders: The book "ARM Architecture Reference Manual" can be downloaded from altera: http://www.altera.com/literature/lit-exc.jsp
23:48:09crwlpreglow, does here
23:48:16preglowcrwl: it does make white noise?
23:48:19crwlyep
23:48:30markunbe aware that it's 13MB big..
23:48:33preglowjhMikeS: i'd revert that rather quickly, then, unless you've got an idea what's wrong
23:49:17jhMikeSyes...and I don't know cause I'm not experiencing anything...will recheck it cause it really should work
23:49:25preglowin any case a clear notice of the fact that the build might damage your ears would be nice :)
23:49:56jhMikeSheh...np
23:49:57amiconn?
23:50:12*amiconn wonders how white noise might damage your ears
23:50:13markundan_a: chapter A6 might be interesting to read
23:50:21preglowamiconn: have a guess
23:50:26crwlheh, for some reason my old WPS says it's playing FLACs when I'm playing Ogg Vorbis
23:50:31amiconnToo high a volume?
23:50:31crwlI wonder since when this has happened...
23:50:34preglowamiconn: correct
23:50:55dan_amarkun: Thanks - I'll have a glance, though I much prefer hardcopy books ;)
23:50:59jhMikeScrwl: since recording was updated
23:51:25amiconnWell, volume can't be significantly higher than what is set as listening volume...
23:51:43jhMikeSpreglow: I'll just readd the emac init to the old place for now...the rest shouldn't harm it. I'll figure it out.
23:52:07preglowamiconn: if you're listening at around -6 dB then full scale white noise will be very loud indeed
23:52:12markundan_a: how do you switch to THUMB mode? The manual says that bit[0] of the target address needs to be 1
23:53:03amiconn-6 dB??
23:53:11Mikachuwhite noise contains much more energy than music
23:53:17preglowdan_a: how did you deal with kernel.c thread context switch?
23:53:20*amiconn thinks that would damage the ears even without white noise
23:53:28dan_amarkun: I've just let GCC do it. Basically, if you've got a function which is at 0x40, you do "bx 0x41" and the processor knows to switch to thumb mode
23:53:30preglowamiconn: it all depends what you're listening to
23:53:46preglowdan_a: just rewrote them in thumb asm?
23:53:51dan_apreglow: Anything with ASM in it I have just compiled as ARM
23:54:01 Join tim66 [0] (n=tim_pick@82-69-41-162.dsl.in-addr.zen.co.uk)
23:54:09markundan_a: that's nice of GCC, didn't think it was that smart :)
23:54:10 Nick SirFunk_ is now known as SirFunk (n=Sir@cpe-74-71-205-222.twcny.res.rr.com)
23:54:49preglowdan_a: are you sure it's handled correctly, though? the context switch stuff relies heavily on inlining
23:55:12 Nick funky_ is now known as funky (n=repulse@81.202.255.93.dyn.user.ono.com)
23:55:16 Nick Everybody is now known as Everybody|away (n=everybod@harpo.demon.co.uk)
23:55:32markunah yes, the inlined asm..
23:56:16dan_apreglow: Not 100% sure, but IIRC the inlines are only used within thread.c, and the arm/thumb choice is done on a per-file basis (not per-function)
23:57:34 Quit Thundercloud_ (Connection timed out)
23:58:32 Join Febs [0] (n=chatzill@207-172-204-33.c3-0.rdl-ubr4.trpr-rdl.pa.cable.rcn.com)
23:58:49preglowdan_a: oh
23:58:59preglowmakes sense

Previous day | Next day