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

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

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

#rockbox log for 2007-07-18

00:00:51 Quit jhMikeS (Nick collision from services.)
00:00:57 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
00:06:32 Join Blackbrother [0] (i=7d19bc03@gateway/web/cgi-irc/
00:06:50 Quit datachild ("- nbs-irc 2.35 - -")
00:07:26 Quit obo ("bye")
00:07:51BlackbrotherHow do i iinstall rock box on my 5g?
00:09:15 Quit Lear ("ChatZilla [Firefox]")
00:09:49merbananBlackbrother: by reading the install instructions on the webpage
00:10:38 Quit ender` (" I think I remember an episode of MacGyver where he overthrew a violent dictator with a rubber band, 2 bottle caps, and some )
00:10:47Blackbrothermerbanan: Well i just gotta extract the files on my ipod right? what about back ground pics?
00:10:55 Join krazykit [0] (
00:11:44merbananwell I know nothing more then what's on the webpage
00:11:54LloreanBlackbrother: Please simply read the manual
00:12:49 Quit Guile`` ("KVIrc 3.2.0 'Realia'")
00:12:55BlackbrotherLlorean : Can th player now REW or FF?
00:13:16LloreanRockbox has had FF and RW since long before the iPod was supported...
00:13:23dan_aBlackbrother: That's in the manual ;)
00:15:02Blackbrotherdan_a: lol i don't see the point why you won't tell whether the player can FF or REW movies ... sinmple yes or no answer lol
00:15:41Blackbrotherlinuxstb: thanks
00:16:17linuxstb_But mpegplayer is very poor on the 5g, so I wouldn't recommend it.
00:16:18dan_aBlackbrother: I didn't tell you whether the player can FF or REW movies because I didn't know that is what you wanted to know
00:17:05LloreanBlackbrother: All you asked is if the player could ff/rw.
00:17:17Blackbrotherwell if anyone can help me ... how do i remove my old bootloader (now using ipl but and want to use rockbox bootloader) .... ok i see dan_a
00:17:49 Join Wiwie [0] (
00:18:23linuxstb_How did you install your ipl bootloader? Was it with ipodpatcher?
00:19:34 Quit MySic ("Miranda IM! Smaller, Faster, Easier.")
00:20:09 Part maffe
00:22:37 Quit Blackbrother ("CGI:IRC (Ping timeout)")
00:23:11bluebrotherscorche: what do you think about converting the svg images of the players for the overview table?
00:24:43 Quit godzirra (Read error: 104 (Connection reset by peer))
00:28:30*linuxstb_ now recalls the last time blackbrother was around...
00:30:39peturwow, even I remembered
00:32:19 Join webguest17 [0] (i=18d0504e@gateway/web/cgi-irc/
00:32:46 Join maffe [0] (
00:33:17 Quit webguest17 (Client Quit)
00:41:11 Quit MonoPete ()
00:42:26 Quit petur ("*poof*")
00:42:30 Quit Siltaar (Remote closed the connection)
00:45:02 Quit maffe (Remote closed the connection)
00:46:20 Join Blackbrother [0] (i=7d19c492@gateway/web/cgi-irc/
00:49:06 Quit Blackbrother (Client Quit)
00:49:22 Join Blackbrother [0] (i=7d19c492@gateway/web/cgi-irc/
00:49:56 Join SirFunk [0] (
00:49:56 Quit Zagor ("Client exiting")
00:52:05BlackbrotherWhere can i get background images for the rockbox , and I read the support and i couldn't find a link
00:54:00linuxstb_People don't tend to share backdrops separately, only as part of themes.
00:54:08bluebrotherthere is no page with backdrops around. Themes utilizing a backdrop usually ship them
00:54:43Blackbrotherahh I see
00:55:51 Quit merbanan ("Leaving")
01:00:48 Join My_Sic [0] (
01:01:19 Quit My_Sic (Client Quit)
01:01:36 Join sarixe [0] (
01:01:48 Join datachild [0] (
01:04:14BlackbrotherHow do we add album art to the songs?
01:04:26 Join donutman25 [0] (n=chatzill@
01:05:07 Join [0] (
01:05:32linuxstb_We don't - that's not supported by Rockbox. But there are unofficial builds which handle it - see their forum threads.
01:05:43***Saving seen data "./dancer.seen"
01:06:24Blackbrotheris it in the rockbox forums?
01:06:50linuxstb_Yes, see the "Unsupported Builds" forum. Any questions about them should be asked in those threads.
01:08:22 Quit Robin0800 (" HydraIRC -> <- \o/")
01:09:31 Join safetydan [0] (i=cbca159f@rockbox/developer/safetydan)
01:09:46BlackbrotherOh, I see .... well thanks
01:14:51 Join saratoga [0] (i=9803c6dd@gateway/web/cgi-irc/
01:18:05linuxstb_saratoga: Hi. Any WMA news?
01:20:02pixelmalinuxstb: beat me to it (re. 80GB Ipods...) :)
01:23:08BlackbrotherHow come my themes i loaded isn't displaying?
01:25:15linuxstb_Do you mean not displaying in the "Browse Themes" menu, or not changing the display after you select them?
01:26:44BlackbrotherI extracted the themes but in rockbox nothing is being displayed
01:27:40PaulJamWhere did you extract them to, and did the .zip files contain a .rockbox directory?
01:27:42saratogalinuxstb: yes I'm getting close on the low bitrate files issue
01:28:20saratogai've got the noise coding stuff working (at least it gives identical output to the fp decoder), however very low bitrate files change a number of table sizes which breaks a lot of the nasty assumptions i made when I got rid of the malloc calls
01:28:41saratogafirst step will be to commit some more sensible static memory allocation code, which i'm about how half way done with
01:29:05BlackbrotherI extracted the backdrops, themes, and the wp folder but there isn't a font folder in my .rockbox folder
01:29:13 Nick Tr1ckY^Gone is now known as Tr1ckY (
01:29:36linuxstb_Then create one. You may also want to install the standard Rockbox fonts (see the manual)
01:30:12saratogai've also been meaning to ask, is there someway to specify in ARM that a multiply should only use the 2 most significant bytes of a word? I'd like to speed things up by not doing full multiplies but I don't think it'd be worth it if I have to keep adding =& 0xFFFF000 after every op
01:31:03linuxstb_I don't think so - some ARMs have 16-bit multiplication instructions, but our ones don't.
01:31:15Blackbrotherlinuxstb_: I have seen those fonts before .... thanks anyway
01:31:48saratogalinuxstb: well they do 16 bit multiplies if theres 0s in the last 2 bytes of the second operand, so you'd think there'd be a way to specify that
01:31:50saratogabut maybe not
01:32:03linuxstb_Blackbrother: If you're saying you don't want those fonts, then fine, but don't complain when lots of themes don't work as you expect.
01:32:04 Part toffe82
01:34:29Blackbrotherlinuxstb_: I already have them installed .... but what can i do if the fonts and the background picture is the same color .... is there a way to reset the background?
01:34:38PaulJamconcerning the "stkov main" problem on Hxxx players: wouldn't it be possible to increase the main stack temprarily in SVN until someone fixes the problem?
01:35:10PaulJamBlackbrother: turn on the hold switch when rockbox boots. this resets all settings.
01:35:14saratogaI think the idea was to movtivate someone to fix the problem by not increasing the stack size
01:35:40BlackbrotherPaulJam: Thanks
01:35:41PaulJamok, that makes sense.
01:39:12BlackbrotherWhaen i start up my ipod i automatically enter rockbox but how can i go to the normal ipod menu?
01:40:25linuxstb_Turn the hold switch on immediately after your ipod boots (before Rockbox starts).
01:40:44linuxstb_and before the backlight comes on.
01:41:33Blackbrotherlinuxstb_: Thanks
01:42:25 Quit Wiwie ("Miranda IM! Smaller, Faster, Easier.")
01:42:29linuxstb_saratoga: Do you know how much memory the WMA codec is using now?
01:42:30BlackbrotherHave anyone found a way to reduce the usage of batter for the ipod video
01:45:25BlackbrotherThe source code isn't known i guess
01:45:52saratogalinuxstb: is there some easy way to find that in the map file? or do I just look for the highest 0x4... address?
01:46:42 Quit Blackbrother ("CGI:IRC (EOF)")
01:47:02saratogaerr rather i mean 0x01f address
01:48:47linuxstb_Yes, looking for the highest address is the easiest.
01:48:52saratogalooks like 257Kb if I'm reading the file right
01:48:56saratogaplus IRAM
01:49:19saratogai have some ideas to decrease that further, but they'll have to wait for the low bitrate fixes
01:49:49 Quit ("Leaving")
01:51:37linuxstb_Yes, that looks right. Over 200KB of that seems to be BSS (i.e. runtime data)
01:53:08 Join JdGordon [0] (
01:53:26linuxstb_And about 42KB of IRAM?
01:53:43 Quit Tr1ckY ("( :: NoNameScript 4.02 :: )")
01:54:58*linuxstb_ wonders if tomal has got WMA working on the iFP
01:55:21 Quit secleinteer (Read error: 104 (Connection reset by peer))
01:55:40saratogalinuxstb: that looks about right
01:55:41 Quit sarixe ("Peace")
01:55:47JdGordonhey Nico_P
01:55:47saratogathe IRAM stuff is still fairly sloppy
01:55:58saratogai need to play with it more
01:56:22 Join mbishop [0] (n=martin@unaffiliated/mbishop)
01:58:11mbishopHow well supported is the sansa e200 line of players? any problems?
01:58:49 Join secleinteer [0] (
01:59:03saratogambishop: pretty well, though its only been working for a few months
02:01:34 Quit spiorf_ (Remote closed the connection)
02:06:37 Quit linuxstb_ ("CGI:IRC")
02:07:59mbishopHmm, ok, I might end up trying it
02:09:11 Part mbishop ("Konversation terminated!")
02:19:32 Quit atsea- (Remote closed the connection)
02:20:27 Quit entheh ("^~")
02:21:25 Join Tr1ckY [0] (
02:23:26 Join atsea- [0] (i=ariel@gateway/tor/x-1952dab87fe16479)
02:27:21 Part pixelma
02:37:27 Join DefineByte [0] (i=5751c305@gateway/web/cgi-irc/
02:38:09 Quit DefineByte (Client Quit)
02:38:24 Quit XavierGr (Read error: 110 (Connection timed out))
02:38:27 Join DefineByte [0] (i=5751c305@gateway/web/cgi-irc/
02:38:53DefineByteanyone seen killer's post in plugins/viewers?
02:39:24DefineBytewrong place. wrong title tut tut
02:39:30DefineBytestill not really his fault
02:39:58DefineBytehow is a new user supposed to know that music isn't played back via plugins? >.>
02:40:55 Quit jhMikeS (Nick collision from services.)
02:41:01 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
02:41:07DefineByteidlers in IRC are far more annoying. xD
02:41:11 Part DefineByte
02:43:14saratogaDefineByte: hes got 40 posts so I wouldn't call him a new user
02:44:35 Join toffe82 [0] (
02:49:49 Quit kfazz ("Leaving")
02:57:13 Quit PaulJam (".")
03:02:40 Join MournBlade [0] (i=4401597c@gateway/web/cgi-irc/
03:05:45***Saving seen data "./dancer.seen"
03:09:00 Join natror [0] (i=d4026320@gateway/web/cgi-irc/
03:11:35natrorHello, I just want to upload a .wps and a theme for the iPod 5 gen.
03:12:20natrorwill you give me (GrzegorzMurzynowski) write right fot twiki?
03:13:39midkaynatror: that should be it.
03:14:18natrorthanks. Hope someone will like that theme and wps.
03:14:40midkaywhich device is it for?
03:15:10natrorit's for iPod 5th Gen (iPod Video).
03:15:40midkaycool, same player here. i'll have a look at it when it's up.
03:24:31 Quit gotthardt ("ChatZilla [Firefox]")
03:24:42 Nick Tr1ckY is now known as Tr1ckY^Gone (
03:27:25 Join webguest4884 [0] (i=415fada3@gateway/web/cgi-irc/
03:27:37 Quit safetydan ("CGI:IRC (Ping timeout)")
03:27:40webguest4884rockbox 3.1 coming out anytime soon?
03:28:15scorchebluebrother: not quite following you
03:28:52webguest4884why not?
03:30:02scorchebecause the goals of the project have changed a bit...we no longer care about releases much
03:30:25scorchewe still plan to release a 3.0 release, but i cant really give you a timeline on that
03:30:43 Quit dandin1 ()
03:33:03webguest4884fuck this then
03:33:05 Quit webguest4884 ("CGI:IRC")
03:33:53iambeni never even thought about what the current version # is
03:34:06iambenmy version is "2 hours ago"
03:34:51scorchewell, version numbers entail a release
03:35:00scorchewe havent had a release in quite a while
03:35:35saratogai actually favor just pulling 2.5 entirely and doing away with releases altogether
03:36:06saratogatheres really no reason anyone should be using the releases we have, and it may be ages until we have another
03:36:18saratoga they're somewhat misleading
03:36:41saratogaah good
03:36:45saratogamakes sense really
03:37:57saratogais anyone closely following that Zen thread?
03:38:18saratogai kind of want to weed out the off topic posts, but they're so hard to parse Im not sure what can be deleted
03:39:57 Join jhulst [0] (
03:40:01Battousaidelete the whole thing and see what gets posted again
03:50:08 Quit Thundercloud (Remote closed the connection)
03:53:04 Join skyfaller [0] (n=nelson@wikipedia/Skyfaller)
03:55:03skyfallerhey folks, sorry for the stupid question, but I can't quite figure it out from your website... does the latest iPod video work with Rockbox?
03:55:08skyfalleris 5.5 the current generation?
04:17:52 Join Febs [0] (
04:30:58 Quit donutman25 (Read error: 104 (Connection reset by peer))
04:32:56 Join Journey [0] (i=Journey@
04:33:55JourneyI'm having trouble with my IPod. It crashes constantly and I'm forced to reset it several times before it works again. Could this have anything to do with rockbox?
04:34:04Journeyit's a 5.5g video IPod
04:34:39 Part skyfaller ("Leaving")
04:35:53 Quit Febs (Read error: 104 (Connection reset by peer))
04:38:35chrisjs169|brbI've tried downloading the source archive three times so far, and each time when trying to compile Rockbox for the Sansa, I get "export/font.h:30:21: error: sysfont.h: No such file or directory"
04:43:23saratogachrisjs169|brb: just built the sansa from SVN with no issues
04:43:29saratogahave you tried downloading it that way?
04:45:25 Join mrobe5001 [0] (i=4792942c@gateway/web/cgi-irc/
04:46:13 Join nick89 [0] (
04:46:28 Quit nick89 (Read error: 104 (Connection reset by peer))
04:46:30chrisjs169|brbsaratoga: i'm thinking I'm going to need to
04:46:33mrobe5001anyone here? i from the mrobe fourm
04:46:58mrobe5001im here for about gettin rockbox for the 500i
04:48:18 Quit Journey ()
04:48:34 Join aliask [0] (
04:48:51natrorthe two files attached to by me (GrzegorzMurzynowski) are there by mistake, please delete them.
04:49:14 Quit mrobe5001 (Client Quit)
04:51:39 Quit mbr (Read error: 110 (Connection timed out))
04:54:05 Join relaxed [0] (n=relaxed@unaffiliated/relaxed)
05:02:15saratoganatror: i think you can just delete them yourself
05:03:19natrorno, i dont'have right
05:04:10 Quit jhulst (Remote closed the connection)
05:05:03 Quit Soap (Read error: 110 (Connection timed out))
05:05:49***Saving seen data "./dancer.seen"
05:17:43 Join homielowe [0] (n=chatzill@
05:18:06homielowenewbiezhunter: I suggest you go on the IRC when some interested developers are on
05:19:16 Part homielowe
05:20:04 Join Soap [0] (n=Soap@rockbox/staff/soap)
05:35:02 Part natror
05:36:51 Join sarixe [0] (
05:38:26 Quit relaxed ("")
05:39:37 Quit krylon ()
05:44:32 Quit Nico_P (Remote closed the connection)
05:48:31 Quit saratoga ("CGI:IRC (EOF)")
05:50:35 Join relaxed [0] (n=relaxed@unaffiliated/relaxed)
05:58:21 Join karma [0] (
05:59:25karmahey, I have a 30G 4th gen ipod that I used dd to copy a 20G 4th gen partition table to (the one on the wiki)...
05:59:48karmaso now I have an sdb1 and sdb2 which is 20Gs, and I need it to be 30
06:00:32karmaI ran mkfs.vfat /dev/sdb2 and now there's a ~20G partition at sdb2
06:00:52karmawhat should I use to make sdb2 expand by 10G
06:07:09karmaJdGordon: Should I wipe sdb2 completely then recreate a new sdb2 as vfat?
06:07:35 Quit relaxed ("")
06:08:26aliaskkarma: Try parted, or gparted (if you use gnome)
06:08:45 Join xNibbler [0] (
06:08:58karmaI tried running gparted aliask
06:09:06karmaIt doesn't see anything at ALL on sdb!
06:09:21aliaskHrm, maybe it doesn't like doing it over USB.
06:09:25karmaIt just stupidly says that all of sdb is 27.94 GB of unallocated space
06:10:10karmawhich worries me...
06:10:36karmathere exist both sdb1 and sdb2 files though and I can mount sdb2 as vfat
06:10:45karmait's just not large enough
06:11:08TiMiD[FD]would be nice to have a format plugin
06:11:28TiMiD[FD]could be dangerous though
06:11:33aliaskTiMiD[FD]: Bad idea if you ask me.
06:12:04TiMiD[FD]for example for people who put >120Gb hdd in their players
06:12:17aliaskEspecially for an OS like rockbox which isn't flashed (on most players)
06:12:23scorchevery bad idea imo...
06:12:33TiMiD[FD]who are obliged to remove the disk and put it in the computer to format the >120gb partition
06:12:49scorcheplus, why would you even want one on a device?...just us a computer...
06:13:07TiMiD[FD]even if rb gets wiped you can still reinstall it
06:13:28scorcheon ipods maybe...
06:13:53scorcheTiMiD[FD]: the reason why you have to remove the disk is because the device itself cant access higher sectors, so a format plugin wouldnt do a bit of good...
06:14:17TiMiD[FD]oh I read that it could, but that the usb chip couldn
06:14:46TiMiD[FD]so that you had to remove the disk in order to make the partitions and format
06:15:00scorchewell, yeah...the USB to ATA bridge, but still...
06:15:18scorchethere is no reason to format on the device itself...that is just asking for trouble
06:15:39TiMiD[FD]yes, but I'm idle :)
06:15:39karma: v
06:15:39karma19631428 unallocated sectors
06:15:53karmathat's what fdisk says is on my ipod
06:15:57TiMiD[FD]And I don't know what to program ...
06:16:05TiMiD[FD]I need something fun
06:16:18TiMiD[FD]to enjoy my work hours
06:16:19scorcheplenty of feature requests to wade through if you wish =)
06:16:20karmawhy doesn't fdisk or gparted recognize that I have an sdb1 and sdb2 partition
06:16:32TiMiD[FD]scorche: ytes, but they seem so boring ...
06:16:33scorchepartition table...
06:16:51TiMiD[FD]and also most of them require to have sound or a device
06:16:59TiMiD[FD]which I havn't at work ...
06:17:12TiMiD[FD]now that the remote port is almost finished
06:17:35TiMiD[FD]I would make another useless demo plugin
06:18:00 Quit Nibbier (Read error: 110 (Connection timed out))
06:18:38TiMiD[FD]or study for jlpt
06:25:38Davo_DinkumTiMiD[FD]: What level are you studying for?
06:30:14MournBladeIf I have a problem do I ask in here or some other place?
06:31:06aliaskMournBlade: Here is fine
06:31:16TiMiD[FD]I'm thinking about taking the test for level 3
06:31:24TiMiD[FD]I don't know my level actually
06:31:31MournBladeseems I have a problem with database mode Sansa
06:31:46Davo_DinkumDon't you have to take the tests in order?
06:31:46TiMiD[FD]I live here so I learn what is usefull to survive :)
06:31:55MournBladehangs on scanning
06:32:15TiMiD[FD]I bought the test application form yesterday
06:32:20TiMiD[FD]I'll see
06:32:24MournBladedoes OK in file mode
06:32:33TiMiD[FD]I think my level is somewhere between 2 and 3
06:32:36*scorche pushes the offtopic conversation to a PM or to #rockbox-community
06:32:40Davo_DinkumTiMiD[FD]: Everyone lives "here"
06:32:51TiMiD[FD]ah righ
06:32:56aliaskMournBlade: Try deleting all the .tcd files in the .rockbox directory, and reinitializing the database
06:32:59 Quit lazka (Remote closed the connection)
06:33:17TiMiD[FD]I meant in the country where they speak the language of the jlpt :)
06:33:24Davo_Dinkumah ok
06:33:26Davo_DinkumTiMiD[FD]: join offtopic
06:33:29Davo_Dinkumwhat part though?
06:33:40MournBladekk thanks, will try that, thought I did but who can tell with me
06:35:43 Join relaxed [0] (n=relaxed@unaffiliated/relaxed)
06:38:14 Join ptw419 [0] (
06:43:30MournBladealiask: deleted 1 .tcd file then rebooted. On bootup when I went to database I had to initialise. I did and it reported so many items and finished, asking for a re-boot...I did and when I tried to enter database it said that Database is not ready...initialise now?
06:43:41MournBladeI did and it hung on 0 items
06:44:20aliaskHrm, sounds like you might have some bad metadata in one of your files, but I'm certainly no expert on the database.
06:45:16MournBladeI am looking at the new tcd file now
06:45:20 Quit kubiix ("Miranda IM! Smaller, Faster, Easier.")
06:46:18MournBlademaybe I should just bite the bullet and move music folder off and add 1 directory at a time back
06:48:59MournBladeis the filename database_tmo.tcd correct?
06:49:16MournBladetmp sorry
06:49:34amiconnscorche: An fdisk+format plugin would be a *very* useful thing for archoses with >128GB harddisks. And it wouldn't be more dangerous than a delete folder function
06:50:13TiMiD[FD]I knew it :)
06:50:20scorcheas long as you had it plugged in the entire time
06:50:48amiconnYou can always reinstall rockbox
06:51:19amiconnEithe you have rockbox flashed, then it's obviously no problem at all (just moaning at boot that .rockbox is mising
06:51:29 Join n17ikh|Lappy [0] (i=n17ikh@dial-181.r2.ncbrvr.InfoAve.Net)
06:52:00amiconnOr it's not flashed/flashable, then the archos firmware starts (and only sees the first 128GB)
06:52:19scorchei suppose worst case scenario for those is just plugging itno a computer to re-write the table/format
06:52:26amiconnEither one will give you usb access so you can make a new partition and reinstall
06:52:36scorchebut in terms of dangerousness i was thinking more of the irivers
06:52:56amiconnThe bootloader also provides usb access
06:53:29SlasheriMournBlade: hi, it seems that many users have problems using the db with the sansa. it has been reported that on sansa created db files can get corrupted
06:53:46amiconnIt's just that the current ones have a bug that they hang when there is no partition
06:54:03MournBladeahhhhh I shall just be patient then, Thanks
06:54:06amiconnThis needs to be fixed (LinusN wanted to fix that months ago...)
06:55:11 Quit sarixe ("Peace")
07:02:29 Join kubiix [0] (
07:04:26 Quit aliask ("ChatZilla [Firefox]")
07:05:08 Quit n17ikh|Lappy ("asfalk")
07:05:51***Saving seen data "./dancer.seen"
07:11:01 Join davina [0] (
07:22:07TiMiD[FD]does anyone here uses the h1x0 remote ?
07:41:36 Quit midkay ("Leaving")
07:41:50 Join midkay [0] (n=midkay@rockbox/developer/midkay)
07:43:26 Join Transience [0] (
07:48:48 Join Farpnut [0] (
07:48:49 Quit Farp (Read error: 104 (Connection reset by peer))
08:03:30 Join merbanan [0] (n=banan@
08:06:34 Part toffe82
08:07:29 Quit JdGordon ("Konversation terminated!")
08:09:30 Join My_Sic [0] (
08:16:51karmadoes rockbox support album art on the ipod color?
08:18:35 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
08:18:35 Join GodEater [0] (
08:18:59 Join miepchen^schlaf [0] (
08:19:16 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
08:22:18 Part Transience
08:24:04 Quit karma ("KVIrc 3.2.4 Anomalies")
08:36:42 Join ender` [0] (
08:40:52 Join Rob222241 [0] (
08:57:41 Quit ptw419 ()
08:59:18 Quit Rob2222 (Read error: 110 (Connection timed out))
08:59:51 Join kaaloo [0] (
08:59:59 Part kaaloo
09:05:52***Saving seen data "./dancer.seen"
09:06:01 Join pondlife [0] (
09:09:32 Join JdGordon [0] (
09:09:36 Join petur [0] (n=petur@rockbox/developer/petur)
09:10:15 Join Nick_Brackley [0] (i=7dff0332@gateway/web/cgi-irc/
09:10:46 Quit JdGordon (Read error: 104 (Connection reset by peer))
09:12:19 Join JdGordon [0] (
09:15:45 Quit Nick_Brackley ("CGI:IRC (Ping timeout)")
09:16:41 Join spiorf [0] (
09:17:03 Join desowin [0] (
09:18:51 Join XavierGr [0] (
09:31:45 Quit BHSPitMonkey (Read error: 110 (Connection timed out))
09:33:46 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
09:35:24 Join bdgraue [0] (
09:41:42 Part dan_a
09:41:44 Join aliask [0] (
09:42:29*petur discovers ohloh is now updating its database quite quickly (every 2 days?)
09:43:04petur( )
09:44:15 Join ptw419 [0] (
09:48:10 Quit linuxstb (Read error: 110 (Connection timed out))
09:51:39 Join zacs7 [0] (
09:55:21 Quit ptw419 ()
09:58:53 Quit JdGordon (Read error: 104 (Connection reset by peer))
10:01:12 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
10:01:28JdGordonhey all
10:04:59 Nick petur is now known as all (n=petur@rockbox/developer/petur)
10:05:04allhi JdGordon
10:05:10 Nick all is now known as petur (
10:05:52 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
10:07:47 Join Nick_Brackley [0] (i=7dff0332@gateway/web/cgi-irc/
10:13:48 Join linuxstb__ [0] (
10:20:26 Join obo [0] (
10:20:42JdGordonamiconn: you round?
10:21:34 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
10:21:40JdGordon.... I was just reading the logs and you mentioned some extra stuff hwcodec needs to do...
10:22:06JdGordonspecifically storing extra data for whats in the buffer, why does it need that?
10:22:37 Join safetydan [0] (n=dan@rockbox/developer/safetydan)
10:23:29JdGordon... also, to keep the playback engines as close as possible, couldnt we setup a simpler pcmbuf for hwcodec which automatically does the bitswapping before going into that buffer and then out to the dac?
10:27:29 Quit linuxstb_ (Read error: 110 (Connection timed out))
10:28:27pondlifeJdGordon: I *think* it was so that the data might be indentified as WAV or MP3 format, and little- or big-endian. This would indicate whether a bitswap was needed and whether the MAS would need reprogramming for WAV output.
10:29:01pondlifeIf we had a second "PCM" buffer (although it would still contain bitswapped MP3 most of the time) then that may not be needed.
10:29:01JdGordonok, I'm a bit worried about what nico said, I'd like to keep all that out of the buffering code
10:29:10 Quit darkless (Read error: 113 (No route to host))
10:29:24JdGordonthat stuff would ideally go in the metadata block for the track
10:29:29JdGordonwhich playback would control
10:29:46pondlifeYes. Maybe you could just provide a "userdata" field associated with the handle, but not do anything with it?
10:30:08pondlifei.e. add it to the structure, then see if it's actually used in the final version?
10:30:36pondlifeJust a word or dword should do, with API calls to get/set it.
10:30:44 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
10:30:48JdGordonno, I want to keep the struct for generic buffering only
10:31:05JdGordonplayback will have a seperate struct with the handles to the differnt bits so it wont get lost
10:32:58 Join darkless [0] (
10:34:07amiconnpondlife, JdGordon: The format indication is necessary so the enigne knows what to feed to the mas
10:34:39amiconnBut in fact that's not different from swcodec, as swcodec also buffers various formats in the main buffer and needs to remember which codec must be used to decode it
10:34:44JdGordonok, but buffering is just to store the files/data, playback controls the actuall sending music out bit
10:35:02amiconnThe bitswap on SH1 targets is necessary for _all_ formats
10:35:16 Quit miepchen^schlaf ("Verlassend")
10:35:24pondlifeBut it's probably useful to have a data type (or userdata) field for buffering in general.
10:35:29 Quit linuxstb__ (Read error: 110 (Connection timed out))
10:35:41amiconnpondlife: It's not only useful, it's necessary
10:35:44pondlifeThe actual values stored are dependent on the caller, not buffering.
10:36:01*JdGordon disagres
10:36:19amiconnThe buffering needs to call filetype identification _when buffering_, and buffer the necessary codec along the metadata
10:36:51pondlifeThe codec is just another file to buffer as far as buffering is concerned, no?
10:37:01JdGordonno, the playback engine tells buffering which files to buffer in the order it wants them, the codec is the first to be buffered for a track
10:37:03amiconnJdGordon: The reason why we might not want a separate "pcm" buffer on hwcodec isn't due to it being hwcodec, but due to ram constraints
10:37:12JdGordonbuffering only reads from disk and dumps from ram
10:37:21amiconnThis buffer should be small, and making it small requires a guaranteed low latency handling
10:37:38pondlifeIs the SH1 fast enough for that..?
10:37:59JdGordonwhen does the bitswap happen atm?
10:38:07pondlifeOn disk read?
10:38:17amiconnThe advantage of a separate buffer would be that it allows some (simple) real codecs (like adpcm)
10:38:30pondlifeAnd helps unification.
10:38:36amiconnCurrently, bitswap runs behind disk read
10:38:38 Join linuxstb__ [0] (
10:38:53JdGordonafter metadata is read I assume?
10:39:05amiconnThat is, the buffering first fills the buffer, spins down the disk, and then bitswaps
10:39:18amiconnThis is to keep spinning time down
10:39:41amiconnJdGordon: Bitswap has nothing to do with metadata
10:39:51amiconnMetadata does never enter the main buffer atm
10:40:14pondlifeBitswap could be handled by the "codec" that copies to a PCM buffer then quite nicely.
10:40:15JdGordoni know... does the metadata parser needs the file bitswapped?
10:40:29amiconnBitswap is interleaved with disk read in case the playable buffer gets really low
10:40:41amiconnJdGordon: Of course not
10:40:49pondlifeamiconn: is a bit useless, but would you expect enabling scaling to shorten battery life noticably?
10:40:59pondlife(Sorry to jump topics.)
10:41:12amiconnThe bitswap is _only_ due to the SH1 SCI, and the MAS is connected via SCI
10:41:37JdGordonpondlife: close it i rekon
10:41:37amiconn(for playback)
10:42:07amiconnpondlife: Scaling is disabled atm on sansa
10:42:22pondlifeHmm, I am confused, ignore me.
10:42:29 Quit linuxstb (Read error: 110 (Connection timed out))
10:42:35JdGordonthe sh1 is too slow to bitswap as its being fed to the mas?
10:43:43amiconnTheoretically not, but SCI runs via DMA
10:44:04amiconnAnd the DMA engine obviously can't bitswap
10:44:47JdGordoncan it be done before dma starts?
10:44:55JdGordonor is it continuous?
10:45:29pondlifeIt must be diskread/bitswap/DMA for each block...
10:45:54pondlifeWith the diskread done asap to minimise spinup.
10:46:17JdGordonanyway, for the moment I dont want to think about this :p its going to be hard enough to merge our code with swcodec
10:46:51 Part zacs7 ("Gonversation terminated!")
10:47:11pondlifeJdGordon: Sounds like you've made good progress to me.
10:47:33JdGordon1/3 threads possibly done
10:48:09amiconnDMA is run block by block. Swapping a block when the old block ended would be too late
10:48:30amiconnSo we need at least one block swapped in advance
10:48:59amiconnThe biggest problem would be to guarantee low latency as we need to keep extra buffers small
10:49:03pondlifeAgain, sounds like a "codec" in the SWCODEC model.
10:49:26JdGordonwe can either use a dedicated thread for it, or put it in the playback thread... | bitswap, send to dma, bitswap the next one, dma..
10:49:31pondlifeOnly question is would the SH1 keep up, and what would the battery life hit be?
10:49:47JdGordonare block sizes static?
10:49:51amiconnOverall cpu load won't change much
10:50:14 Quit linuxstb_ (Read error: 110 (Connection timed out))
10:50:18amiconnDMA block size can be chosen arbitrarily, upper limit is 65536 bytes
10:50:46JdGordonwhat size is it set to in svn?
10:51:10amiconnA battery life hit would result from having to reduce main buffer size when making room for an extra buffer
10:51:10JdGordonhwcodec "pcmbuf" can just be 3 or 4 of these...
10:51:44amiconnJdGordon: Only if we can make sure that they get refilled fast enough (i.e. low latency)
10:52:15JdGordonis there any reason to think it wouldnt keep up?
10:52:25JdGordonas long as the data gets into ram fast enough it should be fine
10:52:52JdGordonrecorders go for $150 usd still!
10:53:11pondlifeMine went for 45.
10:53:19pondlifeWish I hadn't sold it now!!
10:53:48 Nick linuxstb__ is now known as linuxstb (
10:54:04amiconnJdGordon: Threading.
10:54:30amiconnThe thread might just get postponed too long
10:54:55JdGordonthats possible now anyway...
10:58:35linuxstbI'm coming late to the conversation, but made some quick notes about what I understand the swcodec and hwcodec engines need to do -
10:59:02amiconnDMA chunk size in svn is 8KB (for mpeg audio data)
10:59:23linuxstbIt seems to me that having a second buffer on hwcodec will just be a waste of valuable RAM, especially when big enough for PCM data.
10:59:26amiconnThe last block of a file is shortened as necessary
10:59:49amiconnlinuxstb: Yeah, but then there are advantages as well
10:59:57amiconnI'm undecided...
11:00:45linuxstbI'm thinking that MAS "codecs" are similar to (non-existing at the moment) "loader plugins" for swcodec - plugins responsible for loading data from disk into the audio buffer before processing by the main codec.
11:01:11pondlifei.e. like the metadata parser?
11:01:24amiconnlinuxstb: That would be a waste for anything but pcm...
11:02:03linuxstbpondlife: Yes, they could also parse the metadata.
11:02:05JdGordonwe would just need to handle a few dma blocks at a time, not pcm size blocks..
11:02:25amiconnwavplay uses 16KB dma chunks
11:02:42pondlifeAt least 2 DMA buffers, so 32KB lost from the main buffer?
11:03:01amiconnI think that 3..4 8KB chunks should be sufficient, but then we need to make sure they get refilled fast enough
11:03:57amiconnOne 8KB chunk can be consumed in 41.6 milliseconds worst case
11:04:06amiconn(48kHz 16 bit stereo pcm)
11:04:15pondlifeIs that the limit of pitch-up?
11:04:19JdGordonworst case beign the slowest or fastest?
11:05:05amiconnpondlife: No, pitch up can still cut this time in half
11:05:24pondlife21ms isn't much.
11:05:56***Saving seen data "./dancer.seen"
11:05:57amiconn(but I wouldn't care about pitch up for pcm on mas)
11:06:07JdGordondinner time
11:06:14aliaskAnyone familiar with ARM assembly - how can I set the VE bit of the cp15 register?
11:06:48aliaskSo many places refer to it, but none say WHICH bit the VE bit is.
11:10:38 Join The-Compiler [0] (
11:12:45markunaliask: I could take a look at the arm docs I have at home, but no time now
11:13:05aliaskEven the arm docs I have read simply refer to it as the VE bit...
11:14:12aliaskMaybe I'm just looking in the wrong place.
11:14:59aliaskAh, there it is. Bit 24 if anyone was wondering :)
11:16:44aliaskThanks markun
11:17:28markunyou're welcome :)
11:23:13 Join webguest78 [0] (i=c023110b@gateway/web/cgi-irc/
11:28:01 Quit amiconn (Read error: 110 (Connection timed out))
11:32:41MournBladealiask: after a long nap I moved all the music folders off my sansa and putting it all back in 1 dir at a time...seems to be working so far....can use database mode
11:36:30aliaskGood to hear
11:36:47aliaskBut it's odd that it didn't work in the first place.
11:38:12MournBladeI know......maybe it was me....messing with things I shouldn't. I tend to do that with new a bull in a china shop
11:39:33JdGordonLlorean: can I ask why there is such an anti-feature-requests vibe in the forums?
11:39:42JdGordonFS really isnt good for actual discussions
11:45:03linuxstbI don't think there's a problem with discussions of potential features, but most threads of that kind tend to be simple feature requests without any discussion aspect.
11:50:55 Join amiconn [0] (i=3e088e42@rockbox/developer/amiconn)
12:00:14 Quit JdGordon ("Konversation terminated!")
12:00:39TiMiD[FD]is there any reason why the action api is disabled for ipod 3g and 4g in some plugins ?
12:01:46 Join JdGordon [0] (
12:01:50TiMiD[FD]erm JdGordon just left .. he was the one who wrote the action api
12:02:36JdGordonnice timing :p
12:02:43TiMiD[FD]ah right
12:02:54TiMiD[FD]I was wondering about something
12:03:15TiMiD[FD]the action api is disabled for ipod 3g and 4g in some plugins, but is there any reason for that ?
12:03:49TiMiD[FD]for example maze.c
12:03:58TiMiD[FD]line 38
12:04:13JdGordonim in windows atm so dont have the code...
12:04:20JdGordoni dont know why they wouldnt use it
12:05:15TiMiD[FD]maybe they didn't wanted to use the standard actions mappings and didn't knew how to make new actions with the action api ...
12:05:31TiMiD[FD]so I can use this one in my plugin without problem I suppose
12:05:42amiconnThat reminds me... some plugin actions are still broken, e.g. metronome for player
12:06:33JdGordonis there a FS taks?
12:08:18 Join My_Sic [0] (
12:13:49 Join klbostee [0] (
12:15:36klbosteeHi. I am trying to find the right code to play an mp3-file from a plugin. Is there any documentation available about this or so?
12:16:25TiMiD[FD]there is maybe the libmad used in the mpegplayer ...
12:18:03klbosteesome of the plugins apparently use mp3_play_data, but the data is always an array defined in the source itself
12:18:27klbosteeso basically I would like to know how to read an mp3 from disk
12:19:39TiMiD[FD]open() ? :p
12:19:59linuxstbmp3_play_data is only for Archos devices I think.
12:20:28linuxstbThere is currently no simple way to play compressed audio from a plugin (uncompressed audio isn't straightforward either...)
12:20:45TiMiD[FD]would be nice to have an standardized playback api
12:20:54KasperleJdGordon: hurray for websvn :)
12:21:19pondlifeTiMiD[FD]: We're working on it...;)
12:21:47klbosteewell, let me rephrase my questing then
12:21:51JdGordonKasperle: : i suggest you ask dionoea why he did it that way
12:22:22TiMiD[FD]JdGordon: this hack is present in other plugins as well ...
12:22:41klbosteeI am a PhD student working on audio similarity and related things, and one of the things we worked on recently is dynamic playlist generation ...
12:22:41*JdGordon doesnt know
12:22:44KasperleJdGordon: i didn't ask about it. i just posted it because you said you didn't have the code handy
12:22:58JdGordonoh, woops wrong person
12:23:05Kasperleno worries :)
12:23:22linuxstbklbostee: So you want to generate dynamic playlists in Rockbox?
12:23:38klbosteebasically the idea is to choose the next song to be played next intelligently by trying to find a song that is similar to the songs already listened to, and not similar to the skipped ones
12:24:03klbosteeand I wanted implement this on rockbox
12:24:06pondlifeSimilar based on tags?
12:24:14pondlifeOr more complex?
12:24:20scorchelike pandora?
12:24:20klbosteeno, based on the audio content
12:24:34GodEaterlike modlogic
12:24:36linuxstbklbostee: One immediate problem that comes to mind is that Rockbox buffers many tracks in advance (up to around 28MB worth)
12:24:59linuxstbklbostee: But I guess you can work with that...
12:25:02klbosteewell, actually it could be based on anything, but we used an audio similarity measure in our experiments
12:25:08pondlifelinuxstb: Yes, but the decision would only change if the user skipped, right?
12:25:28linuxstbklbostee: Isn't that very CPU intensive? Or do you perform the audio analysis in advance?
12:25:32klbosteethe similarity values would be calculated in advance (using matlab or so) though
12:25:46pondlifeSo they could be stored as tags...
12:26:57klbosteewhat I wanted to do is put a file file that contains the similarity matrix (ie the values that depict the degree of similarity between each pair of songs) on the player, and then use that info to determine the next song to be played
12:27:13linuxstbklbostee: I don't think that's possible from a plugin, I think it would need to be integrated into the core, probably in the database code (so the tags are all available).
12:27:22klbosteeso I was planning to write a simple player plugin that did this
12:27:56linuxstbSo basically a kind of weighted shuffle?
12:28:12klbosteesomething like that yeah
12:28:45linuxstbHow do skipped tracks factor into it?
12:28:46 Quit My_Sic (Read error: 104 (Connection reset by peer))
12:29:24pondlifeSome kind of negative weighting against that track's matrix values..?
12:29:30klbosteewell, what we do is start from a seed song
12:29:51klbosteeand then we assume that a suggested song is accepted when the user listens to it
12:30:06klbosteeand that it is bad (gets rejected) when the user skips it
12:30:08linuxstbSo you're using "skipped" as a sort of user rating of a song?
12:30:41 Nick The-Compiler is now known as The-Linker (
12:31:00klbosteekind of, the heuristics basically try to find songs that ar similar to the accepted (ie played) ones and not similar to a rejected (ie skipped) one
12:31:01 Quit jhMikeS (Nick collision from services.)
12:31:07 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
12:31:27 Quit The-Linker (Remote closed the connection)
12:31:43 Join The-Compiler [0] (
12:31:51 Join Entasis [0] (
12:31:54klbosteelinuxstb: so why would I need the tags then?
12:32:21pondlifeSeems like a sensible place to store the matrix details?
12:32:38pondlifeRather than another file that needs to be transported along with the files.
12:32:57klbosteehmm, a similarity value applies to a pair of songs
12:32:57pondlifeAssuming these are details based purely on the audio content.
12:33:05Kasperleyou couldn't really store the matrix in there, but you would need to store characteristics
12:33:07klbosteeso using tags might be a bit tricky I guess?
12:33:22pondlifeYes, I see.
12:33:45linuxstbklbostee: You wouldn't need the tags - I wrote that before you mentioned the matrix file...
12:33:52pondlifeI thought you might be able to pre-analyse each file and tag it, then compare tags for a similarity value.
12:34:15pondlifeBut I'm probably over simplifying lots.
12:34:16Kasperleklbostee: how hard is it to quantify the characteristics you base your similarity analysis on? and how much data would that amount to
12:34:45klbosteewe could store the feature vector for each songs (we compare those vectors to obtain the sim values)
12:35:04Kasperleah. how big are those feature vectors?
12:35:05klbosteebut calculating the sim values on the fly would be too expensive
12:35:11klbosteeI think
12:35:44pondlifeRockbox has some optimisation experts around... :)
12:36:13pondlifeWMA playback was optimised up to realtime pretty quickly.
12:36:34klbosteeand you would have to recalculate the sims a lot then anyway
12:36:45 Join maffe [0] (
12:36:45 Quit maffe (Remote closed the connection)
12:36:49pondlifeAfter each skip?
12:37:22klbosteeeach time the next song has to be chosen
12:37:37pondlifeSo it would need to be on-the-fly, right?
12:37:43klbosteewell, not all of the sims, but certainly a lot of them
12:37:52pondlifeThere are many permutations available!
12:38:21klbosteenono, the sim values stay the same, but you need most of them each time a next song has to be chosen
12:38:30klbosteeso it would be better to calculate them all once
12:38:52Kasperlebut maybe you could find a nice tradeoff if you don't want a perfect match
12:39:18klbosteemaybe yeah
12:39:40klbosteeso at which code files should I be looking then?
12:39:51Kasperlehow many features do you extract, i.e. what is the dimension of a feature vector? and are all features equally important?
12:40:15klbosteethe number of features can be tuned
12:40:17markunklbostee: I worked on exactly what you are suggesting as a university project
12:41:33markunusing a neural network to try to learn the music taste of the listeren based on the audio features of the listened and skipped files
12:41:36Kasperleklbostee: how many are normally used?
12:42:01klbosteeour work is based on / an extension of the heuristics described in
12:42:23klbosteeKasperle: I don't know by heart how many we used
12:43:44klbosteewe worked on the heuristics, not the underlying audio sim measure
12:44:29klbosteebut what I really want to know is where I should start when trying to implement one of our heuristics in rockbox... :)
12:45:29 Nick Seedy is now known as Seed (i=ben@
12:46:26klbosteemaybe I should look at the shuffle code, since that is about the closest functionality
12:47:32markunklbostee: but shuffle 'preshuffles' the playlist, it doesn't pick a song at random
12:47:49pondlifeRockbox is based on playlists
12:47:56klbosteebah, not good :)
12:48:07 Quit JdGordon (Read error: 104 (Connection reset by peer))
12:48:30pondlifeTo play a track, you'd insert it into the playlist (as next track).
12:49:00pondlifeOf course, you'd have to be very quick when the user hit skip!
12:49:01 Join JdGordon [0] (
12:49:12klbosteecan plugins append tracks to the playlist then,
12:49:27pondlifeShould be able to, don't know if the API is exposed, but it seems reasonable.
12:50:24klbosteeso I could write a plugin that responds to a skip action, modifies the playlist, and then skips the current song?
12:50:35TiMiD[FD]klbostee: not at the moment
12:51:41klbosteealso not if the plugin defines the "skip action"? so for instance a plugin that shows a simple menu, and "skip" is an item in that menu
12:52:17klbosteewhen the user clicks on the skip item in the menu, the plugin modifies the playlist and then makes the player skip to the next song somehow
12:52:23pondlifeI don't think there's anyway to notify the plugin of a skip.
12:52:45klbosteethat wouldn't be necessary then
12:52:49TiMiD[FD]the plugin cannot really modify the playlist I think
12:53:04markunklbostee: by spinning up the HDD every time you will decrease the runtime a lot, but I guess there is no escaping that
12:53:31 Join Thundercloud [0] (
12:53:42klbosteeanyway, I really gotta go now
12:53:45pondlifeIdeally you could calculate many tracks ahead, but given the calculation speed....
12:53:59markunklbostee: you could also split it in 2 phases, learning and playlist generation
12:54:22klbosteeI'll come back later to get some more info when needed :)
12:54:26 Quit klbostee ("Leaving")
12:55:49 Join maffe [0] (
13:00:12 Quit Nick_Brackley ("CGI:IRC (EOF)")
13:00:36linuxstbmarkun: Was your university project successful?
13:01:02 Join Nick_Brackley [0] (i=7dff0332@gateway/web/cgi-irc/
13:02:13markunlinuxstb: I tests I did were very artificial
13:03:47markunif you choose metal as music to be liked and trance as music to be disliked it worked very well for example
13:04:03markunbut on more subtle differences it became very hard
13:04:41markunand there were some pitfalls: if you never play the songs which initially (randomly) get skipped you will not know if the user did in fact like them
13:05:11markunso used the result from the neural network as the chance for the song to be played
13:05:58***Saving seen data "./dancer.seen"
13:10:44 Join Nico_P [0] (
13:12:45 Quit desowin (Remote closed the connection)
13:28:18 Quit amiconn ("CGI:IRC (EOF)")
13:38:21 Join amiconn [0] (i=3e088e42@rockbox/developer/amiconn)
13:42:30 Quit safetydan ("Ex-Chat")
13:50:09 Join webguest53 [0] (i=50c8f8c8@gateway/web/cgi-irc/
13:50:20 Quit webguest53 (Client Quit)
13:54:18 Nick Tr1ckY^Gone is now known as Tr1ckY (
13:59:56 Quit JdGordon (Read error: 104 (Connection reset by peer))
14:03:31 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
14:04:33 Quit pondlife ("disconnected has pondlife")
14:18:46 Quit xNibbler (Read error: 104 (Connection reset by peer))
14:28:42 Quit aliask ("ChatZilla [Firefox]")
14:38:03 Join xNibbler [0] (
14:41:37 Quit webguest78 ("CGI:IRC")
14:48:36 Quit darkless (Read error: 113 (No route to host))
14:51:11 Join ptw419 [0] (
14:55:35JdGordondoes anyone know excatly which are the bare minimum functions in the codec api that need implementing to get sound?
14:55:43 Join Febs [0] (n=chatzill@
14:55:49JdGordons/sound/decoding working
14:56:07JdGordonis it just read_filebuf() ?
14:58:04 Quit elinenbe (Read error: 110 (Connection timed out))
14:59:00 Quit Nick_Brackley ("CGI:IRC (Ping timeout)")
14:59:44 Quit daurnimator (Read error: 104 (Connection reset by peer))
14:59:51 Join daurnimator [0] (n=fake@unaffiliated/daurnimator)
15:05:28 Quit atsea- (Remote closed the connection)
15:06:00***Saving seen data "./dancer.seen"
15:10:11 Join atsea- [0] (i=ariel@gateway/tor/x-e5385aaf4bacb2cc)
15:16:55 Join basscade1 [0] (
15:18:48linuxstbJdGordon: grep for ci-> in whichever codec you wish to test with. The only ones I think you can avoid are the seeking ones.
15:19:42JdGordonim having trouble reading the codec into the buffer atm so not quite ready for actual testgin yet
15:20:03 Join webguest17 [0] (i=c023110b@gateway/web/cgi-irc/
15:20:33 Quit SirFunk (Connection timed out)
15:20:54 Join Arathis [0] (
15:25:39 Join elinenbe [0] (n=elinenbe@
15:29:06 Quit sneakums (Read error: 110 (Connection timed out))
15:30:55 Join Nick_Brackley [0] (i=7dff0332@gateway/web/cgi-irc/
15:40:22JdGordonwhats the simplest codec from a testing POV?
15:41:34JdGordoneven wav is a PITA to get working
15:41:48 Join darkless [0] (
15:42:04GodEaterI can't imagine what's going to be easier though
15:47:01 Quit Nick_Brackley ("CGI:IRC")
15:47:32Nico_PGodEater: hi, have you tried reading the MoB wiki page again today ?
15:47:46GodEaterah no - not yet
15:47:50GodEaterhave been dstracted
15:48:28 Join Jens [0] (i=Jens@pdpc/supporter/active/Jens)
15:48:58linuxstbJdGordon: I think they're pretty much all the same.. As I said, a grep to see how many (and which) ci-> functions each codec uses might show you a simple one.
15:49:28*JdGordon needs a small wav file to test with
15:55:05Nico_PGodEater: distracting ideed
15:58:15 Quit relaxed ("")
15:59:16 Join SirFunk [0] (
16:11:13*JdGordon having some very odd problems with this stupid code
16:11:22*JdGordon obviously too sleepy to get anything working
16:15:06*petur walks off to another bloody meeting :/
16:16:16 Quit elinenbe (" HydraIRC -> <- IRC has never been so good")
16:20:05 Join indro_ [0] (
16:20:11indro_hi all
16:22:38indro_I updated rockbox to newest version. Now I have a problem with playing music. After almost 1:11 playing a track rockbox display an error and freezed, so I need to reboot the ipod.
16:22:50indro_Anyone has an idea whats wrong?!
16:23:21oboHow old is your bootloader?
16:23:41 Join linuxstb_ [0] (i=5343d4aa@rockbox/developer/linuxstb)
16:24:04indro_Almost more than half year...
16:24:43oboGrab the latest one (linked from the manual) and see if that fixes it
16:24:47*linuxstb_ tries to translate that to months... ;)
16:25:12indro_its the linux bootloader, because I have ipodlinux on ipod (but actually dont use it)
16:25:47indro_obo: bootloader and playing tracks has not much common or?
16:25:57indro_bootloader is only for booting rockbox, or i am wrong?
16:25:58 Part maffe
16:26:11Lloreanindro_: The bootloader does some hardware initialization, older ones don't handle things quite right
16:26:24linuxstb_Older bootloaders (both the IPL one and official Rockbox one) didn't initialise the coprocesssor correctly for Rockbox.
16:26:25 Join maffe [0] (
16:26:29 Quit maffe (Remote closed the connection)
16:26:49indro_okay, thanks, is it possible to boot linux with rockbox bootloader?
16:26:50 Join maffe [0] (
16:27:15linuxstb_Yes, copy the IPL kernel to the root of your FAT32 partition, call it linux.bin, and hold PLAY as you boot.
16:27:41linuxstb_Although that feature of the Rockbox bootloader isn't well tested...
16:28:00indro_thanks, will try that!
16:28:13LloreanIt worked last time I tried it, but that was... well... quite some time ago.
16:28:21linuxstb_The current versions of the IPL bootloader _should_ work, but it's hard for us to help if you have problems - we didn't write it.
16:29:26indro_yep, I see...
16:30:41linuxstb_Llorean: I've been half-following all these "my ipod won't power on" threads - do you think there's a chance of a problem or are people just letting the battery drain completely and having the usual issues associated with that?
16:31:40Febslinuxstb_: There are times when my ipod does not power on normally, even when charged. (For example, it happened when we were in the pub last week.)
16:31:55FebsA reset always solves the problem, however.
16:32:15Lloreanlinuxstb_: My nano has shown a glitch where if it shuts down, when I attempt to power up (even immediately after) it shows the low battery Apple screen
16:32:26LloreanThen a hold on/off, menu+select can recover it, but nothing else seems to.
16:32:47LloreanI've read, somewhere, that this can even happen in the apple firmware if it goes into deep sleep
16:33:05LloreanAFAIK, it's a problem with 5/5.5G and Nanos.
16:33:38LloreanIn fact, someone suggested updating to the newest Apple firmware resolves it, though I haven't tested that since for me at least it's a rare occurrence.
16:33:39indro_I have the same problem btw.
16:34:06 Join jgarvey [0] (
16:34:15LloreanIf it's a problem with Rockbox, it has to be something we leave when we shut down that's tripping up the Apple bootloader. Is that possible?
16:34:35LloreanIt could be that we make the bug more common, somehow.
16:34:51indro_I needed to do a reset with usb cable plugged in, since then it do power on...
16:34:51LloreanAn "we didn't make the bug, but we trigger it more" sorta thing that could be fixed.
16:35:12*dionoea never had that kind of problem with a 5G 30GB
16:35:19dionoeaIs that specific to 5.5G?
16:35:42indro_I have 5Gen 60GB
16:36:03linuxstb_Most reports seem to be the 5.5G, but that could just be because it (I expect) is one of the most widely used ipods.
16:36:29indro_5.5G is that what I have?
16:36:45JdGordonok, its possible im asleep and so missing something obvious, but where is the *ci initialised in the plugins? the codec_start function doesnt take any params...
16:36:46linuxstb_But yes, it's possible that the power-off we do is confusing the Apple bootloader.
16:37:27amiconnI also get this low-bat thing on my mini G2 occasionally
16:37:38linuxstb_JdGordon: I assume you mean codecs instead of plugins?
16:37:43amiconnJust happened this morning. reset usually works
16:37:49JdGordonlinuxstb_: yeah
16:38:15JdGordonah, hmm... codec_start is the entry point, not codec_main
16:38:53FebsJust to be clear: I have experienced the low battery issue, though it is very rare. The more common issue is that Select does not cause the ipod to turn on. In either instance, reset works. I don't know if these are two manifestations of the same problem or not.
16:39:15amiconnI think it's just that the voltage doesn't rise fast enough depending on how the button is pressed
16:39:48amiconnThe apple firmware doesn't power down, only goes to deep sleep, so it's a non-issue there
16:39:59Lloreanamiconn: After a day or two, I thought it did.
16:40:36JdGordonok, well this is wierd... in codec_start() ci is 0x80bd740 but in codec_main() its 0!
16:40:38amiconnMaybe it does after some extended time, but then I'd expcet to see this low bat effect with the OF as well
16:41:02Lloreanamiconn: As I said, at one point I found out that some people have seen it when their iPod shuts down fully using retail firmware
16:41:04linuxstb_JdGordon: I can't see how it works either...
16:41:04*amiconn never actually used the apple OF
16:41:14LloreanI was using "Deep Sleep" to mean the full shutdown. I may have gotten terms confused.
16:41:26LloreanI thought "Sleep" was the low power, screen off, and "Deep Sleep" is what Apple called "shut down"
16:41:48JdGordonci is set in codec_crt0.c, but for some reason its being set to 0..
16:41:52JdGordonall very wierd..
16:41:56linuxstb_I wouldn't expect many Rockbox users to leave their ipods in the Apple firmware for long enough for the full power-off to happen anyway. So maybe it does happen there as well...
16:42:31 Quit Entasis (Read error: 104 (Connection reset by peer))
16:42:40indro_Installing bl on 0 should be good, or?
16:43:11linuxstb_indro_: How are you installing the bootloader? ipodpatcher should take care of it automatically.
16:43:28indro_ipodcatcher, it ask me where I want install the bl.
16:43:36JdGordonbtw, is it possible for RB to put the pp's into deep sleep mode?
16:43:56*JdGordon assumes its a pp thing and not an ipod thing
16:44:06KasperleLlorean: i've had that happen on my 5.5g as well
16:44:14linuxstb_indro_: It shouldn't ask that - it should autoscan for your ipod, and just ask you to enter "i" to install.
16:44:21 Quit Jon-Kha (Read error: 145 (Connection timed out))
16:44:24indro_linuxstb_: oh I thought it ask me but I only need to press I.
16:44:32indro_linuxstb_: right!
16:44:41amiconnSleep mode is a PP thing, but we can't use it yet
16:44:59JdGordonamiconn: we dont know how? or oter reasons?
16:44:59 Quit webguest17 ("CGI:IRC (EOF)")
16:45:18amiconnSetting sleep mode involves setting the CPU clock to 32kHz, as well as some other hardware configuration
16:45:28amiconnThe latter is what we don't know yet
16:45:38linuxstb_JdGordon: "Deep sleep" is what Apple refer to as the power-off that Rockbox does already. (IIUC). Normal sleep is what amiconn just described.
16:45:59JdGordondeep sleep is where I should be
16:46:16bluebrotherobo: around?
16:46:20amiconnOkay. This 'deep sleep' vs. 'sleep' is confusing because there's also the normal cpu (core) sleep
16:46:28 Join fm2 [0] (i=d9b9677a@gateway/web/cgi-irc/
16:46:42JdGordonsleep is what i meant when i asked, not deep sleep
16:46:46*JdGordon -> zzzzzzz
16:46:46linuxstb_That's just a brief nap...
16:46:47obobluebrother: yup
16:46:52 Quit JdGordon ("Konversation terminated!")
16:48:01indro_I get now the error right after playing a track...
16:48:03fm2GodEater: Hello. I checked what folder shortcuts are and you are right −− they do exactly what I intended with the "folder bookmarks"
16:48:15indro_Prefetch abort at ....
16:49:11fm2But if I understand correctly the shortcuts is not a part of the FAT spec but are a windows specific feature, no? So there is little to no chance for them to be supported by RB.
16:49:58bluebrothershortcut files are simply files with the extension ".lnk"
16:50:01indro_Rebootet and tried again, now I get:
16:50:12Lloreanindro_: And this is with the bootloader in iPodpatcher downloaded from Rockbox's site, and an up to date SVN build?
16:50:17linuxstb_fm2: I guess anything could be implemented if it's seen as generally useful.. I'm sure supporting .lnk files has been discussed before, but I can't remember the arguments...
16:50:22indro_Undefined function at ...
16:50:32indro_after 0:12 minutes
16:51:23fm2linuxstb: they have been discussed on IRC? I'll check then. But am I right in the assumption that they are not a part of FAT?
16:51:24markunindro_: and you didn't have this problem with your previouw build?
16:51:25indro_svn build from 12. july
16:51:47markunfm2: not a part of the FAT spec no
16:52:04indro_markun: yes, but it was an old version.
16:52:18indro_before the 32mb/64mb split.
16:52:29linuxstb_fm2: Flyspray seems to have four closed tasks about them -
16:52:38pixelmaindro: maybe if you've updated from an old build... is there a rockbox.ipod in the root of your ipod still?
16:52:47markunindro_: and now you are using the 64MB version?
16:52:58indro_pixelma: i repladced it with the one in the new svn build
16:53:06markunWell, if you only have 32MB the 64MB version wouldn't even work
16:53:10indro_markun: yes, using 64mb version.
16:53:32indro_I have 5gen 60gb...
16:54:20indro_thats 64mb afaik
16:54:48indro_should I test 32mb version?
16:55:02markunyou could
16:55:06pixelmaindro: it shouldn't cause that error but the right place for it is inside the .rockbox folder - you should leave it there and delete the one in the root (or move it back)
16:55:30pixelma(that has changed some time ago)
16:55:41 Quit amiconn ("CGI:IRC (Ping timeout)")
16:55:48Lloreanindro_: Was it ever refurbished?
16:56:12LloreanI've heard rumors of refurbished 60GB iPods with 32mb of RAM
16:56:24indro_pixelma: okay, mom...
16:56:56fm2linuxstb: thanks for the references! Very interesting. All four have been closed almost without any discussion (in FS, I haven't cheked IRC yet).
16:56:58 Join lost|X40 [0] (
16:57:02 Quit lostnihilist (Read error: 104 (Connection reset by peer))
16:57:14Lloreanindro_: Being impolite to respected members of the community isn't going to make you more well liked here.
16:58:20pixelmait would just help you with the update process...
16:58:23linuxstb_fm2: Yes. seems Bagder is the most against it - there is a patch here which is closed without mentioning a reason -
16:58:38indro_Llorean: okay, then I will try 32mb...
16:58:47fm2And if we wouldn't support .lnk files at the file system level (which is unnecessary IMHO) we could support them via viewers
16:59:06indro_Llorean: hm, why do you thing im impolite? because of mom?
16:59:33 Join amiconn [0] (i=3e088e42@rockbox/developer/amiconn)
17:00:36linuxstb_fm2: I think the file browser would need to be able to support them. I personally can't think of a reason to use them (I have no problems navigating my music as it is), but wouldn't object if others wanted them - assuming the implementation is small and simple.
17:01:42Lloreanindro_: Yes.
17:01:53linuxstb_fm2: i.e. selecting a .lnk file would just take you to that folder. The issue is what do you do when you go back one level - do you go back to where the link was, or back to the real parent directory.
17:02:07 Join dandin1 [0] (
17:02:38fm2linuxstb: I'd expect the latter.
17:03:37fm2Since (for me) the shortcut is just a way to get to a certain folder. One I'm there ... I'm there. History doesn't matter anymore
17:03:46indro_Llorean: thats a mistake, mom is shortcut for "moment pls" ;)
17:03:47linuxstb_But I guess it can be done with a viewer plugin - you need to spin up the disk to load the .lnk file anyway.
17:04:01indro_Llorean: I didnt want to be impolite. ;)
17:04:18Lloreanindro_: Ah, there's a figure of speech, "yes mother" or "okay mom" that usually means "thank you for giving me instructions I don't feel I need". Sorry 'bout that.
17:04:41indro_Llorean: yes I see, no prob.
17:04:45linuxstb_fm2: The viewer plugin just then needs a way to set the current path in the file browser - which may not be in the plugin API at the moment, so would have to be added.
17:04:52fm2linuxstb: yes, this is the less intrusive way. And it wouldn't be frown upon by bagder and amiconn :-) since it doesn't increase the core bin size!
17:05:26 Join miepchen^schlaf [0] (
17:05:30fm2linuxstb: exactly my thought (about extending the plugin API) −− but you type faster than I
17:05:36linuxstb_Adding the required API may increase the core size a little - depending on what needs to be done. But I wouldn't expect it to be very much.
17:05:37amiconn.lnk support would be something I wouldn't object to, if it's done nice & simple
17:06:05***Saving seen data "./dancer.seen"
17:06:10fm2amiconn: i.e. without increasing size? :-)
17:06:30fm2linuxstb: is there a ready-to-be-used func that just needs to be exported via API?
17:06:35amiconnI don't think that's possible, but there's increase and increase
17:06:50linuxstb_fm2: Probably not. You would need to look at the file browser code.
17:07:01fm2amiconn: that would be the second type :-)
17:07:16linuxstb_And see where it would normally go to when exiting a viewer plugin - and change that behaviour.
17:07:44linuxstb_But I can't see it being very complex at all.
17:08:09amiconnI think that .lnk support without history should be doable with little extra code
17:08:31amiconnThe most difficult part is probably getting the path from the .lnk file. It's not documented
17:08:58linuxstb_The alternative could be our own (.link ?) format simply containing a text string.
17:09:41indro_hmm, how I get the current build in 32mb version?!
17:09:43fm2linuxstb: ok, I'll look into it when I get time. But not in the next couple of weeks. I'm going to leave.
17:10:04indro_doesnt work...
17:10:34linuxstb_Remove "32mb" I think.
17:10:37Lloreanindro_: The link on the build page works fine for me.
17:10:44fm2linuxstb: that would work too. Even better.
17:10:54indro_Llorean: on the build page is only 64mb.
17:11:07indro_if I see corret.
17:11:14markunlinuxstb_, fm2: ac.yu/extra/fileformat/windows/lnk/shortcut.pdf">
17:11:41linuxstb_markun: Don't point me to it... ;)
17:11:52markunlinuxstb_: try to resists!
17:15:25 Join desowin [0] (
17:16:44indro_Llorean: yes but its the 64mb version.
17:17:00indro_pixelma: I get Can't load rockbox.ipod
17:17:04Lloreanindro_: I don't understand, the 32mb version is available on the download page too.
17:17:07indro_When its in .rockbox
17:17:35Lloreanindro_: What version number does your bootloader report?
17:17:43indro_Llorean: on
17:17:50indro_Verison 1.1
17:17:52Lloreanindro_: Yes, see the one for the 30gb iPod?
17:18:03 Quit fm2 ("CGI:IRC 0.5.7 (2005/06/19)")
17:18:30indro_I see only one build for 60gb and thats 64mb.
17:18:48LloreanAs I said, the 30gb iPod.
17:18:55indro_but that works,
17:19:08indro_I have the 60gb.
17:19:12LloreanYes, that's the link the 30gb iPod links to
17:19:16LloreanYour disk size doesn't matter.
17:19:25LloreanWe named them that way because most people don't realize that the real difference is the RAM
17:20:20indro_allright, my fault then. ;)
17:21:20LloreanSo the Rockbox bootloader is saying it can't find rockbox.ipod?
17:22:57amiconnmarkun: http timeout...
17:24:51Lloreanindro_: Alright, try removing the .rockbox folder entirely, and just extracting a current build straight to the root of your player.
17:27:53markunamiconn: I only checked the google cache
17:28:34markunamiconn: the search query was "lnk shortcut format"
17:32:44GodEaterwhy do we have to reimplement Microsoft's version of .lnk ?
17:32:49GodEatercan we not supply our own ?
17:33:12GodEaterotherwise they're only of use to windows owners
17:34:17LloreanGodEater: The advantage of our own format goes further than that.
17:34:26LloreanOur plugin could parse multiple lines. could have six entries, so you see a list when you launch it, then the one you click is where you end up
17:35:22GodEatersorry - I saw markun and amiconn trying to hunt down the spec and wondered what on earth for if they were not trying ti reimplement it
17:35:43LloreanWell, the advantage of windows links are that windows users can create them without any typing.
17:35:54GodEaterlucky them
17:35:57LloreanWhich is better than "no users can create them without any typing"
17:36:25GodEatersurely this is a context menu thing ? Long hold on a folder, and one of the options would be "Add to favourites"
17:36:31GodEaterno typing involved there then
17:36:46LloreanGodEater: I thought the plan was to keep it as much out of the core as possible and just use a viewer plugin
17:36:58 Part maffe
17:37:09GodEateramiconn seemed to indicate he doesn't mind if it ends up in the core
17:37:16peturLlorean: you can do that with a plugin too (see file/dir properties)
17:37:18 Join maffe [0] (
17:37:39Lloreanpetur: Open With and choose the "Create a .link" plugin?
17:38:11peturLlorean: no, just run the plugin in the core code
17:38:19LloreanAh, right.
17:38:28petur... (see file/dir properties)
17:38:46LloreanI didn't realize that was a plugin
17:38:51peturit is
17:40:00linuxstb_A viewer plugin would be more flexible as well - it could present a list as Llorean described, or parse actual Windows .lnk files, or anything else...
17:40:33 Join Lear [0] (i=chatzill@rockbox/developer/lear)
17:40:33peturif we create our own format, it must have another file extension anyway
17:41:05linuxstb_But we could associate our own (e.g.) .link files, and .lnk files to the same viewer.
17:41:19bluebrotherhow about using .desktop?
17:41:33peturand folder.ini :)
17:41:41GodEaterfilthy windows speak!
17:41:45*linuxstb_ slams the can of works shut
17:41:59linuxstb_^I meant worms...
17:42:01*GodEater holds up his tux effigy
17:42:48linuxstb_I'm sure Gnome and KDE have comparable kludges on top of the filesystem...
17:43:26GodEaterI think that's what the .desktop type is
17:43:28dionoeathey have symbolic links too (well, the FS has)
17:43:38GodEaterand hard links too
17:43:55petursuch as a ton of annoying .hiddendir stuff in your home directory :/
17:44:01GodEaterthough I think .desktop is a standard, rather than being gnome / kde specific
17:44:21linuxstb_What's .desktop anyway?
17:44:24*GodEater awards himself a gold star
17:44:35bluebrotherand it's plain text, so maybe it's better than some binary file?
17:44:57GodEaterpaste alert!
17:45:12GodEater[Desktop Entry]
17:45:12DBUGEnqueued KICK GodEater
17:45:12GodEaterName=Wine Windows Emulator
17:45:12GodEaterName[de]=Wine Windows-Emulator
17:45:13***Alert Mode level 1
17:45:13GodEaterExec=wine %f
17:45:22GodEaterthey look like that
17:45:33LloreanGodEater: Not all of them
17:45:36 Quit iamben (Read error: 104 (Connection reset by peer))
17:45:39 Join iamben_ [0] (
17:45:45 Join mpeccorini [0] (
17:45:52LloreanThe line that says "Exec" can also say "URL" instead if you make a "File" one instead of a "Executable" one
17:45:56linuxstb_GodEater: Yes, but what is it?
17:45:58indro_anyone know what that mean?
17:46:02CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
17:46:02*linuxstb_ never leaves bash...
17:46:11Lloreanlinuxstb_: Basically a .lnk for Gnome, I think
17:46:14indro_I wanted delete something.
17:46:52bluebrotherI just created a shortcut link from KDE. It basically has a Name= entry with the title and a URL= entry with the target path
17:46:53Lloreanindro_: Know what you mean? You haven't asked a new question since my last response to you as far as I can see.
17:47:00linuxstb_indro_: It seems you have a bad filesystem. When you repartitioned to add your Linux partition, did you reformat your FAT32 partition?
17:47:19LloreanOh, link
17:47:44 Quit Rick ("I don't need to be here.")
17:48:05indro_linuxstb_: im not sure, that was a while ago.
17:48:08bluebrotherindro_: have you tried running chkdsk /f / fsck.vfat on that device? Sounds like filesystem trouble.
17:48:20indro_no, but i will try now.
17:48:21linuxstb_indro_: That would be my guess...
17:48:35linuxstb_(that you didn't reformat)
17:48:54indro_linuxstb_: could be...
17:49:03indro_but i get other strange errors too
17:49:10indro_so i think fsck is a good idea.
17:49:26linuxstb_That may not help - you may just have to reformat it.
17:51:31 Join sarixe [0] (
17:51:51indro_hm, what i should choose?
17:52:10*linuxstb_ shrugs
17:53:05 Join Rick [0] (
17:53:46bluebrotherindro_: I would try chkdsk /f on windows −− afaict it look like that can correct more errors than linux fsck.vfat
17:53:57 Join pixelma_ [0] (i=pixelma@rockbox/staff/pixelma)
17:54:18indro_thx but i have no windows.
17:55:14***Alert Mode OFF
17:55:18linuxstb_indro_: If I was you, I would just reformat and copy the files again.
17:57:14 Quit thegeek_ (Remote closed the connection)
17:57:29 Join thegeek [0] (
17:57:49 Join Domonoky [0] (
17:58:37 Part mpeccorini
17:59:29indro_i unzipped the build again, still get Cant load rockbox.ipod: File not found
17:59:40Learbluebrother: my (very limited) experience is that fsck.vfat is better at finding and fixing errors.
18:00:29linuxstb_indro_: You fixed your filesystem errors?
18:01:30indro_havent test it, just unzipped
18:01:44indro_that worked
18:02:37indro_FAT: Directory bread(block 61388753) failed
18:02:51indro_after deleting something. :/
18:03:02indro_scsi 11:0:0:0: rejecting I/O to dead device
18:03:29indro_but can still read and write
18:06:47*linuxstb_ would reformat...
18:10:19indro_yes, probably better. ;)
18:10:52indro_*create backup
18:11:47 Quit pixelma (Read error: 110 (Connection timed out))
18:11:47 Nick pixelma_ is now known as pixelma (i=pixelma@rockbox/staff/pixelma)
18:12:51 Quit obo ("bye")
18:15:44 Quit petur ("work->home")
18:17:51 Quit linuxstb_ ("CGI:IRC")
18:19:47*Nico_P added another thread to his test plugin to add new files to the buffer
18:19:51Nico_Pmuch more fun that way
18:19:58Nico_Pand it still seems to work almost ok
18:24:04indro_I recreated the FAT partition with cfdisk
18:24:07indro_Now I get this:
18:24:15indro_Partition table entries are not in disk order
18:25:30indro_is that important?
18:25:48linuxstbThere was no need to repartition - just reformat.... You've now lost the firmware partition.
18:26:54indro_whats now?
18:27:02 Quit sarixe ("Peace")
18:28:10 Quit enyc (SendQ exceeded)
18:29:00linuxstbLooking again at your original parition layout (, that had problems anyway - the firmware partition and linux partition were overlapping.
18:29:44 Join My_Sic [0] (
18:30:09indro_hmm, okay..
18:30:31indro_is 5.5g = 5g with 60gb?
18:30:37indro_or do i need take only 5g?
18:31:40linuxstbDoes ipodpatcher say your sectors are 512 or 2048 bytes?
18:31:52linuxstb512 = 5g, 2048 = 5.5g
18:33:32Kasperleindro_: 5.5g is the "improved" ipod video with a brighter screen and longer-lasting battery, iirc
18:33:42Kasperlecomes with 30gb and 80gb
18:33:44 Quit amiconn ("CGI:IRC")
18:34:07linuxstbKasperle: Some people seem to have 60gb ipods with 2048-byte sectors - that's the important issue.
18:34:29linuxstbThere was at least one person like that in the original 5.5g forum discussion IIRC.
18:34:35 Join amiconn [0] (i=3e088e42@rockbox/developer/amiconn)
18:34:39 Quit amiconn (Client Quit)
18:34:56indro_hm, ipodpatcher seems do not found my ipod anymore.
18:35:02Kasperlei think i recall someone saying he got one like that after it was repaired
18:35:13Kasperlebut i don't really know
18:35:13linuxstbindro_: Then use fdisk to find outt.
18:35:23linuxstb(or scroll up if it's still in your history...)
18:35:50linuxstb"fdisk -l /dev/sdX" will tell you (replace sdX appropriately)
18:37:11linuxstbIf you see something like "Units = cylinders of 16065 * 512 = 8225280 bytes", then you have 512-byte sectors.
18:37:49indro_okay, then I have 512. ;)
18:39:39indro_is the partition now right?
18:41:10linuxstbIf you correctly written the right MBR, then yes...
18:46:06 Quit Lynx_ (Read error: 104 (Connection reset by peer))
18:46:34 Join alberink [0] (
18:50:43indro_I did point 1-4, now fdisk -l /dev/sdb show nothing?
18:51:23indro_oops, ipod is not in disk mode. ^^
18:53:21 Quit My_Sic (Read error: 104 (Connection reset by peer))
18:58:04 Join linuxstb_ [0] (i=5343d4aa@rockbox/developer/linuxstb)
19:00:54indro_hm i loaded rockbox and shutted down with play button.
19:01:10indro_now i cannot powered it on again. :/
19:01:34linuxstb_Try holding menu+select for a few seconds.
19:01:59indro_that works. :)
19:06:09***Saving seen data "./dancer.seen"
19:06:51 Join jac0b [0] (
19:07:27jac0bis there any gigabeat owners in here
19:08:56Lloreanjac0b: Did you have a specific question?
19:09:33jac0bwell it is about getting a scratch out on a gigabeat
19:09:56LloreanAh, well this is #Rockbox. It's for questions relating to the software.
19:10:11LloreanYou might want to try #Gigabeat, or a Gigabeat specific forum somewhere.
19:11:00 Part jac0b
19:11:16 Join Xerion [0] (
19:11:55*linuxstb_ reads the #gigabeat logs and wonders why they don't chat here, where other devs can help them...
19:17:27 Join dan_a [0] (n=dan_a@
19:17:35 Part maffe
19:17:38 Join maffe [0] (
19:18:10Lloreanlinuxstb_: When I first talked with them, it seemed like they weren't certain they were welcome here because Rockbox wasn't running yet (regarding the F) but I don't really know.
19:18:36 Join My_Sic [0] (
19:23:27 Quit My_Sic (Read error: 104 (Connection reset by peer))
19:23:35Nico_PLlorean: have you seen my messages from earlier ? I've added a thread to my plugin to simulate files being added at random times
19:23:55Nico_Pnow I have several files on buffer at a time, and it works fine
19:24:27 Join GodEater_ [0] (
19:24:39Nico_Pand it made me find and fix two bugs in the process
19:25:04LloreanGood to hear.
19:25:26 Join donutman25 [0] (n=chatzill@
19:26:00linuxstb_Does your code include the MOB concept?
19:26:16Nico_Plinuxstb_: no, but it's close
19:27:05Nico_PI can buffer and play a file back while there is another one waiting there doing nothing
19:27:12Nico_Ptypically what would happen with MoB
19:28:42Nico_Pin fatct with the current code I could say any file gets max. 10 MB of the buffer and has to stay within this. It'd work fine
19:30:30 Join enyc [0] (
19:35:34 Join hashier [0] (
19:36:08GodEater_good afternoon :)
19:36:50 Quit tedr0ck ()
19:37:24hashierI just got a "Creative NoMad Zen NX". Is it worth to invest some time or is it nearly impossible to geht rockbox working on it?
19:37:52linuxstb_It depends on your skill, detemination, and how much free time you have...
19:38:17hashierNever done that before
19:38:26linuxstb_This may help -
19:39:15linuxstb_I don't know if all the Zens are the same, but there is also some discussion on the Rockbox forums about porting Rockbox to the Zen. I'm not sure if any real progress has been made though.
19:40:35hashierI found that page... but not verry much information on that.
19:41:19linuxstb_You mean the NewPort page?
19:41:31hashierThe page about ZEN
19:42:01hashierI look on the dev page, there wasn't that "porting" page linked
19:47:28hashierA verry verry long way
19:47:37indro_linuxstb: fs seems to be clean now, read/write/delete works without problem. ;)
19:48:08markunhashier: are there pictures of the insides of the nomad somewhere?
19:48:26indro_now filling it with sound, so I can test if that playing works too...
19:48:27hashiermarkun: Just got it a few hours ago.
19:48:41markunI opened my Gigabeat a few minutes after I got it :)
19:49:04hashiermarkun: Got it at work. Just come home and joined here (=
19:49:09markunhashier: and are there firmware updates available?
19:49:10hashierI open it
19:56:11 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
20:01:15hashierAh! gnomad2 works
20:03:01 Quit Febs (Read error: 104 (Connection reset by peer))
20:05:06GodEater_what does MEM_FUNCTION_WRAPPERS do ?
20:10:33*GodEater_ reads the source instead of being lazy
20:13:13hashiermarkun: Yes, it exist a new firmware
20:13:54 Join My_Sic [0] (
20:14:28 Join Elazar [0] (
20:14:56ElazarAnyone know if work on the Creative Zen: Vision M port would work on or lead to a working port for the Creative Zen: Vision W?
20:15:01 Join petur [0] (n=petur@rockbox/developer/petur)
20:16:20GodEater_where did all the Creative fanboys come from all of a sudden? ;)
20:16:39GodEater_Llorean: don't think that guy has throught through his text adventure suggestion much do you ? ;)
20:16:52LloreanGodEater_: I think he played much simpler text games
20:17:10LloreanOnes that didn't involve an inventory, dialogue, or anything beyond exploration from the sound of it.
20:17:18ElazarGodEater_: If I was a fanboi, I'd be content with their firmware. :P I got the player as a gift, it's the best player I've got, and I don't like the stock firmware very much.
20:17:46LloreanElazar: Well, compare the hardware in yours to the hardware listed on the Vision: M page and that could give you a clue as to how similar they are
20:18:33GodEater_Llorean: I like the idea of SCUMMVM on a DAP though ;)
20:18:43LloreanI've liked it for quite some time, it's just an awful lot of work.
20:18:44GodEater_it's nearly an ideal thing to play on a DAP if you're going to play anything at all
20:19:24LloreanYeah, even the one death (in Monkey Island 1) takes 10 minutes to accomplish, so moving a pointer around with directional controls isn't going to be a real hindrance (dependent on player patience).
20:19:26GodEater_were the discworld games done in Scumm or did they just look the same ?
20:19:30LloreanJust look the same
20:19:48GodEater_I think you could die in Indiana Jones too
20:19:51LloreanThe Lucasarts games all were (they created SCUMM), and then some childrens games (Putt Putt, Pajama Sam, etc)
20:19:58LloreanYou could lose fights with Nazis.
20:20:06LloreanIt didn't result in death though, if I recall. Just a setback
20:20:07 Part Elazar ("WeeChat 0.2.3")
20:20:13GodEater_and the end sequence too - you had to complete it in a certain as I recall
20:20:27GodEater_insert the word "time" in there somewhere appropriate
20:20:54*GodEater_ gets all nostalgic over the games he played at the start of the 90s
20:21:52LloreanSomeone should just port DosBox, make it interact with the virtual keyboard, and then say "There, anything VGA should work, good luck for speed."
20:23:02*ender` 'd like to see scummvm ported :)
20:23:24LloreanI think many of us would. ;)
20:23:41ender`(and at least it wouldn't have keyboard problems :)
20:25:49 Join tucoz [0] (
20:26:11 Join Siltaar [0] (
20:26:16tucozhi, rockbox
20:27:36 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
20:27:51tucozanyone knows why the manual fails with make: /home/martin/src/dev/rockbox/manual/platform/ Command not found
20:27:51tucozmake: *** [info] Error 127
20:28:03peturdoes anybody know if the iHP100 provides mic power on the line-in connector? The h120/140/... do that, but somebody bought an 100 that doesn't and wonders if it's by design or a defect...
20:28:21*ender` 's had a hard drive die in his server today...
20:28:34*tucoz sends his condolences
20:28:55*petur checks his backup schedule
20:29:14ender`i didn't loose any data, but i'll keep it off until i get a replacement, just to be on the safe side
20:29:57tucozI once knocked gently on a dead hard drive, which made the power supply emit sparks and and poof
20:30:27MournBladeI love sparks and smoke
20:30:29Lloreantucoz: I get the same error. But it still creates a manual?
20:30:43tucozyep. I wonder what that file is supposed to do
20:30:51pixelmatucoz: that's a flaw in the make file atm... - it doesn't fail just gives this message. And if you type "make manual" that doesn't happen...
20:31:09tucozaha. ok. all good then
20:31:23ender`i was rescuing data from a client's hard drive yesterday, and it only worked while cold
20:31:23tucoz..well almost good
20:31:47 Join stripwax [0] (
20:32:00ender`solution: move the computer downstairs where the fridge is, put it on the fridge, get long enough cables to have the disk running in the freezer
20:32:19ender`it didn't look pretty, but i managed to salvage most of the data :)
20:32:52 Quit linuxstb_ ("CGI:IRC (Ping timeout)")
20:33:14MournBladethings like that NEVER look pretty
20:33:28 Join linuxstb_ [0] (i=5343d4aa@rockbox/developer/linuxstb)
20:34:40*petur yawns
20:35:22*bluebrother spots tucoz
20:35:32tucozhi bluebrother :D
20:35:51 Join amiconn [0] (
20:36:01bluebrotherhaven't seen you quite a while.
20:36:44 Quit ptw419 ()
20:37:16tucozyeah, i've not been too active this year. i think i've become a pure user of rockbox, as opposed to contributor
20:37:33*petur checks ohloh
20:37:46 Quit datachild (Connection timed out)
20:37:47bluebrotherlet's see how much time I'll have once I found a job ... finished uni recently
20:37:58tucozoh, congratulations
20:38:08tucozi really want to restructure the manual though
20:38:27tucoz..but i do not know where to start
20:38:45*amiconn experienced a really annoying rockbox bug y'day and today
20:39:24tucozpetur, hehe. they keep track of me :)
20:39:26peturamiconn: do you know if the iHP100 also provides plug power on the line-in (like h1x0/h3x0)?
20:39:50amiconnNeeds more investigation, but rockbox didn't save any sound setting when only adjusting them from the menu
20:39:52 Join My_Sic [0] (
20:40:20amiconnThat is, with many spinups between changing the settings and the final shutdown, but no volume adjustment from wps
20:40:42*petur just read a complaing about rockbox not saving rec settings at idle shutdown - related?
20:40:45bluebrothertucoz: you still have more commits than me ;-)
20:40:47amiconnpetur: I didn't even know the h1x0 and h300 provide power there. What for?
20:40:55peturI thought the settings saveing bugs were out
20:41:01peturmic power
20:41:08petur3.7V I think it is
20:41:28tucozbluebrother, i was quite busy in the early days of the manual. but as you might remember, i committed quite a few patches from you and nils back then
20:41:33*amiconn wonders which mic would need power there
20:41:47peturmost mics do, amiconn...
20:41:49bluebrotheryeah. Just noticed the number ;-)
20:41:55amiconnSurely dynamic mics don't need it, and condenser mics have their own power source + preamp
20:41:58*bluebrother looks at ManualStructureTalk
20:42:03amiconnI only have dynamic mics
20:42:37bluebrotheronce rbutil is finished the quick start chapter will get much nicer.
20:42:49peturall my mics use it, but I use a pre-amp most of the time and that feeds them, but I can use them directly too
20:43:00peturso I guess you don't know
20:43:11tucozbluebrother, yes. rbutil looks really nice.
20:43:15peturamiconn: do you have an iHP100?
20:43:17amiconnCheap electret capsules need that power iirc, but who uses those?
20:43:34tucozbluebrother, i think we should move the main menu part from chapter 5, to chapter 4
20:44:07petureverybody over at taperssection uses mics that need power (most 48V), noe of them are cheap
20:44:17bluebrotherlike you added in "Proposal sandbox"?
20:44:23amiconnYes, but 3.7V definitely isn't 48V
20:44:38peturnot all mics need 48V
20:44:55tucozi think i was more motivated back then, in terms of changes. but i think so
20:45:07peturmost are happy with 7-12V, 3.7V gives a bit less range spl-wise
20:45:11tucozi mean, just to move it there as a start.
20:45:38amiconnBut the high quality mics either need no power (dynamic) or higher voltages like 48V (condenser) afaik
20:45:42peturamiconn: do you know anybody with an iHP100? LinusN maybe?
20:45:53 Join obo [0] (
20:46:19*tucoz checks out the old manualstructuretalk
20:46:19pixelmaI believe XavierGr has one (or an H115) or so...
20:46:33*petur pings XavierGr
20:46:37bluebrothershouldn't we start advertising rbutil? It isn't completed yet, but so is Rockbox ...
20:46:39amiconnAfaik rasher has one, but he isn't around much these days
20:46:57peturrasher's been more around lately
20:47:13amiconnpetur: You see, I'm not much interested in recording, and even less in recording w/ external mic...
20:47:31amiconnEither internal mic (for voice notes) or line in (sometimes)
20:47:54*petur wonders if amiconn visits concerts
20:47:57amiconnI have an external mic (dynamic) which I never used for real...
20:48:06amiconnI do, but why should I record them??
20:48:21peturto enjoy them again and again
20:48:21amiconn(Okay, I visit concerts _rarely_)
20:48:49amiconnBelieve it or not, but I prefer studio recordings over live recordingds
20:49:18peturmost of my favourite bands sound better live than on CD, those are the true artists, putting more into a song live than in studio (IMHO)
20:49:20*linuxstb_ generally prefers live recordings, but is happy for others to make them
20:49:38*linuxstb_ hugs bittorrent ;)
20:49:47*petur buys the CD and listens to live versions
20:50:10 Join Febs [0] (n=chatzill@
20:51:20 Part maffe
20:52:21 Join maffe [0] (
20:56:34 Join datachild [0] (
21:00:51*GodEater_ prefers his studio recordings too
21:01:10 Join dandin1x [0] (
21:01:37 Quit dandin1x (Client Quit)
21:06:12***Saving seen data "./dancer.seen"
21:09:42 Quit amiconn (" Fensteraufdatum...")
21:15:41petur[21:09] <−− amiconn has left this server (" Fensteraufdatum...").
21:16:04peturFenster = window(s), auf = up, datum = date
21:16:05GodEater_translation ?
21:16:20GodEater_I got the windows bit but didn't know the rest
21:16:33Kasperleengrish at its worst
21:16:40tucozhehe. will force you to reboot for sure
21:16:42indro_is there a tool available for translate the rockbox lang file?
21:16:45*petur likes literal translation jokes
21:17:22peturindro_: yes
21:18:02 Quit dandin1 (Read error: 110 (Connection timed out))
21:20:56MournBladeyay !! I found out what was wrong with my database.......huge comment tags. Whoever tagged em used the whole Amazon review and pasted it in the comment field
21:21:40MournBladeshould have looked there first
21:21:58peturalmost as bad as stuffing album art in it :)
21:22:03linuxstb_Rockbox still shouldn't choke on them though...
21:22:08indro_petur: genlang?
21:22:21peturindro_: yes
21:22:48MournBladeit sure did though.....could not use database mode cause it hung up
21:22:51indro_ perl genlang -e=english.lang -t=test.lang -b=test.lng
21:22:57linuxstb_MournBlade: I'm guessing they're id3v2 tags?
21:23:00indro_doesnt work, know whats wrong?
21:23:11peturindro_: new language?
21:23:12MournBladedamn, I should have looked
21:23:17indro_petur: yes
21:23:27*GodEater_ is taking his first stab at "proper" code for rockbox by implementing a "Favourites" plugin, rather than submitting noddy patches to ipodpatcher and rbutil.
21:23:28peturmuch work!
21:24:01linuxstb_As long as you call it favourites ;) (with a U)
21:24:23linuxstb_I assume that's the favourite folder idea discussed earlier?
21:24:27MournBladelinuxstb: wma tags
21:24:31GodEater_linuxstb: naturally - none of this poor abused language from the colonies ;)
21:24:37GodEater_linuxstb: yup
21:24:41linuxstb_MournBlade; Ah, then it's my fault - I wrote that code...
21:24:44peturindro_: I have no idea, I use genlang on linux and it just works. Make sure you execute it in the lang folder fault
21:25:15indro_petur: it display me the help text, so the syntax must be wrong.
21:25:19MournBladecan I get a refund? <g>
21:25:21indro_but dont see whats wrong...
21:25:36GodEater_MournBlade: certainly, a 200% one if you like
21:25:40linuxstb_MournBlade: No refunds, but if you can give me one of your troublesome WMA files, I'll try and fix it. :)
21:25:48MournBladecoolness lol
21:26:06*petur summons perl user to help indro_
21:26:09indro_linuxstb_: I get the error again. :/
21:26:31MournBladereally dont know why I am fooling with wma since now I have a choice
21:26:42indro_linuxstb_: fs error
21:26:46linuxstb_indro_: Do you disconnect cleanly? i.e. unmount/eject before pulling the plug?
21:26:51bluebrotherGodEater_: favourites plugin? We had one once ...
21:27:15indro_linuxstb_: I was still unzipping sound to ipod
21:27:33 Join amiconn [0] (
21:27:50GodEater_bluebrother: what happened to it ?
21:28:08linuxstb_Replaced by the playlist catalog(ue) IIRC
21:28:09bluebrotherit got removed in favor of the playlist catalog.
21:28:23indro_Extracting Patrice/Videos/Chiemsee Reggae Summer 2002.mpg 48%
21:28:24indro_Read error in the file Patrice (Reggae) Discographie Complete 1998-2006 Full Albums,Lives,Maxi,BonusTracks,Pochettes,Lyrics,VideosClips & Concert... By GanjaCrew.rar [R]etry, [A]bort R
21:28:44bluebrotherthere is a wiki page listing old plugins −− OldPluginIndex iir.
21:28:50linuxstb_indro_: ?
21:29:20GodEater_bluebrother: does that detail what it actually did ?
21:29:22linuxstb_That's what I'm asking - what were those two lines?
21:29:32MournBladeLinuxstb: oops I just deleted the tag
21:30:18indro_I dont plugged it off, I was unzipping when the fs got readonly.
21:30:20GodEater_bluebrother: a - this isn't the same sort of favourites ;)
21:30:41linuxstb_indro_: Your paste talked about a read error - which means problems on the source disk IIUC.
21:30:54 Join hannesd [0] (
21:31:03indro_source/dest is the same.
21:31:10indro_i copied the zip file to ipod
21:31:17 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
21:31:24bluebrotherwhat kind of favourites plugin are you working on?
21:31:32indro_then did unp *
21:31:42GodEater_bluebrother: be easier if you read back through todays logs
21:31:53indro_*zip files
21:33:39linuxstb_indro_: Then it could be a problem with your USB connection - it's not uncommon for people to report disk errors on ipods which are fixed with different USB cables, or different USB ports. Also, if you're using a hub, connect directly to your PC.
21:34:08 Join austriancoder [0] (
21:35:03indro_hum, im using a hub.
21:35:39austriancoderpetur: ping
21:36:13bluebrotherah, for shortcuts. Didn't remind that by "Favourites".
21:37:29XavierGrpetur: yes I have an iHP-115, I am not sure about the mic power on it. Is there a test I can do to find that out? For now I have to go but tell me if I can help and I will do it later
21:37:59peturaustriancoder: plonk
21:38:32austriancoderpetur: time to do some debugin?
21:38:37peturXavierGr: could you plug a cable in the line-in and measure any voltage between ground and L/R
21:38:49peturaustriancoder: yup
21:39:13peturaustriancoder: same state still?
21:39:22austriancoderpetur: fine :) tx problem is still here...
21:39:30indro_is actually eject/unmounting really necessary?
21:39:45peturdoes e200tool do transmission?
21:39:45austriancoderpetur: so we need to find out, what I am doing wrong/different
21:39:53GodEater_bluebrother: well shortcuts was the initial request, Llorean came up with the idea for just having one file that keeps lots of shortcuts in it.
21:39:59austriancoderpetur: it should... but I have never testet it
21:40:14indro_if yes, can I only unmount or do I need eject too?
21:40:28XavierGrpetur: will do when I come back
21:40:32peturindro_: I only unmount
21:40:48peturXavierGr: thanks!
21:41:06austriancoderpetur: have a look at e200toll source and one of mine candidates: 1) my normal code 2) like_mrh_but_not_working.diff -> merged e200tool source into my source
21:41:19indro_hum okay, what could be happen if plugged out without unmount?
21:41:25peturaustriancoder: I will do so now...
21:41:44peturindro_: disk corruption
21:42:01 Join stripwax [0] (
21:42:05peturif lazy-write is active
21:42:35bluebrotherdelayed writing.
21:42:38peturwhen data is written in background
21:43:02indro_so when I writing something to ipod and plugged it off?
21:43:25bluebrotheryou might corrupt your disk if you don't unmount.
21:43:49bluebrotherf.e. the FAT might not be updated yet, as that needs to be changed quite often upon write.
21:44:11bluebrothercaching the FAT (and writing it later) makes accesses faster.
21:44:29indro_is then hardware failure or just partition corrupt?
21:44:46amiconnthe latter
21:44:49bluebrotherthat's not a hardware failure, as you just broke the data on disk.
21:44:51 Join saratoga [0] (i=980398fe@gateway/web/cgi-irc/
21:45:13indro_okay, thanks for info!
21:45:23indro_for the info ;)
21:45:52indro_I heard about that it damage the hardware..
21:46:38 Part tucoz (""another time..."")
21:48:17peturaustriancoder: where's that e200tool code again?
21:49:31linuxstbpetur: google knows...
21:49:42*bluebrother just wondered half an hour why this stupid thing isn't working and it turned out to be a typo :(
21:49:50*petur asks google
21:50:13peturtoo late...
21:50:26peturgoogle was faster :p
21:51:36*austriancoder gets something to drink
21:52:33*petur discovers he already had the code and kicks his brain
21:53:23 Join webguest03 [0] (i=557f7025@gateway/web/cgi-irc/
21:55:52amiconnI can confirm that _no_ settings are saved unless some special locations are used to change at least one of them
21:55:59amiconn(e.g. volume adjustment in wps)
21:56:24amiconnNo standard settings that is... didn't check browser-based settings like font & wps yet
21:56:31petursure? I just tested with recording and they got saved
21:56:48amiconnDid you change them from the menu _only_?
21:57:04peturwell no I went into recording
21:57:28peturand I probably changed gain
21:57:30amiconnSo the recording screen is another such special location
21:58:03peturwe've had this bug before and I'm sure it was fixed :/
21:58:31amiconnLooks like it was reintroduced
21:59:28 Part Llorean
22:01:05 Quit webguest03 ("CGI:IRC")
22:01:06*amiconn suspects JdGordon's July 11 commit
22:01:14indro_humm, I get still error when playing a tracdk. :/
22:01:23indro_Data abort at ....
22:01:57pixelmaah... had that with ajusting volume from an audio playback menu in a plugin (and idle power-off kicking in)
22:02:50indro_I have now the rockbox bootloader, rockbox.ipod on the right place, reformated the partition...
22:02:54Learpetur: Tried to measure on my H140, but I get strange values. Maybe I'm doing something wrong; it isn't really my area of expertise... :)
22:03:06 Join Llorean [0] (n=llorean@rockbox/administrator/Llorean)
22:03:20indro_32mb version of rockbox...
22:03:41peturLear: h140 should be fine (somewhere between 3 and 3.7V)
22:04:35indro_so any more idea what could be wrong?
22:05:29Learpetur: Ah, better readings now (wrong setting on the meter apparently). Got 3.3.
22:05:32 Join Rondom [0] (
22:05:33bluebrotherindro_: do you have a recent bootloader?
22:05:42indro_bluebrother: version 1.1
22:06:13bluebrotherhmm. Then no quick idea.
22:06:26peturLear: I really wanted to know what iHP100 gives (10GB disk)
22:06:42indro_my previous version worked..
22:06:58Learpetur: I see.
22:07:25bluebrotherold bootloaders with new builds usually give data aborts when playing music.
22:07:39indro_yeS? hmmm
22:07:46bluebrotherbut 1.1 is a recent bootloader. You could of course try to reinstall it.
22:08:03bluebrotheryes, old bootloaders didn't initialize the coprocessor correctly.
22:10:40indro_do I need root access for installing bootloader?
22:11:20indro_It worked without but its better as root?
22:11:38austriancoderpetur: so.. how shall we proceed to find and fix tx problem?
22:11:54indro_could that be the error?
22:12:40peturaustriancoder: well I'm comparing code. If e200tool works and your version not, there must be a difference...
22:13:24austriancoderpetur: okay..
22:13:24peturare we sure there are no other external pins to configure/set?
22:13:37indro_bluebrother: ?
22:14:34austriancoderpetur: dont know... but if so, e200tool does it too
22:17:07peturaustriancoder: nevermind, the comment says it is just toggeling a led
22:18:03*petur removes cache_flush from list as well
22:20:26 Quit saratoga ("CGI:IRC (Ping timeout)")
22:20:37linuxstb_indro_: Do you get the same problems on all tracks? Do the same tracks play fine on your PC (when you mount your ipod and try to play them directly from your ipod's disk)?
22:21:03*amiconn wonders whether there are diodes involved ;)
22:22:31*petur doesn;'t think that's funny :p
22:23:44indro_linuxstb_: yes, I tried different tracks.
22:23:52indro_I try now if they play on my pc.
22:25:34bluebrotherindro_: ipodpatcher needs to access the disk directly, so you need root access.
22:27:00indro_bluebrother: strange, it loads rockbox even I installed bootloader as user..
22:27:29linuxstb_bluebrother: It depends on permissions of the device, and which groups the user is a member of. I can write to removable disks with my normal account.
22:27:51bluebrotheroh −− never seen such a configuration.
22:28:06bluebrotheror at least I can't remember having seen such a configuration ;-)
22:28:09*amiconn thinks that's quite common for workstations
22:28:18indro_Starting playback...
22:28:18indro_A: 152.2 (02:32.1) of 225.0 (03:45.0) 1.2%
22:28:22amiconnOtherwise things like burning a cd would also require root
22:28:43indro_seems to work when im playing with mplayer...
22:29:21bluebrotherwell, a CD writer is a different thing than a removable drive
22:29:48indro_Exiting... (End of file)
22:29:50indro_it works...
22:30:05indro_I try now the same file with rockbox.
22:30:20linuxstb_bluebrother: In my Debian install, removable drives are in the "floppy" group, and so am I.
22:30:46linuxstb_"permanent" drives are in the "disk" group.
22:31:07bluebrotherwell, on my box it's the "disk" group.
22:31:26bluebrotheranyway, my setup works for me ;-)
22:31:27indro_bootloader reinstalled as root
22:31:28*amiconn wonders why vmware workstation doesn't like kernel 2.6.21 :(
22:31:56 Quit SirFunk (Connection timed out)
22:32:53indro_0:30 - freezeee
22:33:14 Quit donutman25 ("ChatZilla [Firefox]")
22:33:38 Quit hcs (Read error: 110 (Connection timed out))
22:33:40indro_but this time without an error..
22:33:58indro_just freeze :/
22:36:32Nico_Plinuxstb_: was it you who asked if a handle data buffer could be expanded when there was new free space ?
22:36:48LloreanI asked that too.
22:37:21Nico_Pwell I've implemented something like that
22:37:31Nico_Pbut I have doubts on how to do it
22:37:47indro_so what I should do now? :/
22:39:07indro_can I show if which version im using?
22:39:13indro_32mb or 64mb?
22:39:15LloreanNico_P: I've no clue how best it should be done, but it does seem somewhat essential.
22:39:30Nico_Pit does
22:39:38linuxstb_indro_: I don't know... Maybe try an older build of Rockbox - e.g. the oldest one here -
22:39:50Nico_Pthing is that you can't expand the buffer without moving data if it's wrapped
22:40:21linuxstb_Nico_P: No, it wasn't me. I'm not sure I understand what you mean.
22:40:25Nico_Pand it'd mean basically moving all of the data within the buffer so not a good idea I suppose
22:40:28 Join hannesd_ [0] (
22:41:18Nico_Plinuxstb_: oh right. I mean when a track is freed the previous one can take the space that was freed to have a bigger buffer
22:41:42linuxstb_I thought that's how ringbuffers worked...
22:42:20 Quit The-Compiler (Read error: 104 (Connection reset by peer))
22:42:25Nico_Pyes but here we have ringbuffers in the ringbuffer :)
22:42:29linuxstb_What do you mean by "previous one" ?
22:42:48linuxstb_I would have thought that if a track is freed, then the previous one would have been freed long ago.
22:44:19Nico_Phmm yeah that's true... In fact it'd happen when a previous track is freed
22:44:33Nico_Pbut there still is free space to exapnd the buffer
22:45:26indro_1:00 :)
22:45:35Nico_Pactually when I think of it, the only useful case is when the first handle is closed. then the last one might have additional space to use
22:46:18linuxstb_I'm not following what you mean with "first" and "last". Do you mean "track N" and "track N+1" ?
22:46:25indro_yay! 2:00, seems to work! :D
22:46:57indro_great, thanks guys! :)
22:46:59indro_2:30 ;)
22:47:21Nico_Plinuxstb_: first and last handle in the linked list
22:47:22linuxstb_indro_: Not really - it means something has been broken...
22:47:35indro_yes, seems so...
22:47:38 Quit RaRe (Read error: 104 (Connection reset by peer))
22:47:56 Quit Rondom (Read error: 113 (No route to host))
22:47:57linuxstb_indro_: Next step would be to work out the most recent build that still works fine.
22:48:06indro_I got 2007-06-18.
22:48:58indro_linuxstb_: kk
22:49:18 Join datachild` [0] (
22:49:30Nico_PLlorean: am I missing something ?
22:50:28linuxstb_Nico_P: Is there something I should read on the wiki? I'm not sure I understand what you're doing...
22:51:22 Join ac [0] (n=austrian@
22:52:20 Quit GodEater_ ("User pushed the X - because it's Xtra, baby")
22:53:18 Quit Lear ("ChatZilla [Firefox]")
22:54:19LloreanNico_P: Yes, I was thinking that if say, 5 tracks are buffered, the last is a long flac that expands beyond the buffer, you don't want to do ringbuffer-within-ringbuffer until you've reached it's own beginning/metadata in the external ringbuffer. That's all I was referring to in regard to it.
22:54:27 Quit Domonoky (Read error: 104 (Connection reset by peer))
22:55:08 Quit hannesd (Read error: 110 (Connection timed out))
22:55:26 Join RaRe [0] (
22:55:31Nico_PLlorean: I hadn't though of it like that :)
22:55:53 Join saratoga [0] (i=9803c6dd@gateway/web/cgi-irc/
22:56:09LloreanSo, "expanding forward" until using a nested ringbuffer is necessary, basically
22:56:09 Nick hannesd_ is now known as hannesd (
22:56:15indro_linuxstb_: 2007-07-18 works!
22:56:37Nico_Pso I can break the buffering before wrapping occurs to allow possible expanding
22:56:47linuxstb_indro_: That's today.... What version were you using before?
22:56:56indro_recent build
22:57:09 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
22:57:20linuxstb_indro_: Do you mean "Current Build" ?
22:58:08Nico_PLlorean: a possible problem I'm thinking of (not completely related, though), is the codec. if it's stored on the main buffer it will just block it
22:58:38linuxstb_The space used by the codec should be freed as soon as the track starts playing.
22:58:49Nico_Poh ok so no problem
23:00:38linuxstb_I'm still not grasping why you want to treat a long track as file-in-ringbuffer-within-ringbuffer, instead of just file-in-ringbuffer like everything else.
23:01:07Nico_Pbecause the metadata needs to be there too
23:01:33Nico_Pand it's better not to have to move it all the time
23:01:33bluebrotherwhy can't you just put the metadata at the beginning of the file?
23:01:48Nico_Pbluebrother: because the beginning of the file will end up being overwritten
23:02:09bluebrothersure, so will the file.
23:02:25linuxstb_But I still don't see that as a ringbuffer-within-a-ringbuffer. If anything, it's split the main buffer into two buffers.
23:02:49amiconnbluebrother: Yes, but the metadata has to be available until the file is fully played, unlike the beginning of the file
23:03:10Lloreanlinuxstb_: Isn't that functionally equivalent anyway?
23:03:13Nico_Plinuxstb_: basically yes, you end up with two buffers
23:03:34 Join GodEater_ [0] (
23:03:38amiconnNico_P: As for buffering the codec, I'd like to see the new code behave like the old one, i.e. _only_ put the codec on buffer if it's different from the previous track
23:03:44bluebrotherso this means that with this ringbuffer in ringbuffer thing if a file doesn't fit into the buffer it will get buffered all the time once the buffer is full?
23:03:55Nico_Plinuxstb_: but you need to manage one of those two as another ringbuffer, don't you ?
23:04:00Lloreanbluebrother: What do you mean "buffered all the time"?
23:04:16Nico_Pamiconn: of course
23:04:18bluebrotherwhy not just put the metadata along the raw data and just block the metadata block if the file is still playing?
23:04:31linuxstb_Nico_P: I would have thought that you still have one ringbuffer, but with a hole in it you need to take special care of.
23:05:19amiconnAnd the codec is fetched when the track starts playing, i.e. its buffer area can be freed by then
23:05:21bluebrotherLlorean: well, if I think of a ringbuffer for the file, and the file doesn't fit into that buffer I'd need to rebuffer. But as the (internal) ringbuffer is limited (and not necesarily as large as the whole buffer) you end up with a (much) smaller buffer
23:05:29bluebrotherthus you need to buffer much more often.
23:05:30Nico_Plinuxstb_: I don't see how to manage a hole in a ringbuffer
23:05:33 Join FOAD [0] (
23:05:37Lloreanbluebrother: "Much" smaller?
23:05:43 Quit datachild (Connection timed out)
23:05:45Lloreanbluebrother: Its size would cap out as "Total buffer - Metadata size"
23:05:57bluebrotherwell, depending on the point in the main buffer.
23:06:02Nico_Pbluebrother: no
23:06:16***Saving seen data "./dancer.seen"
23:06:20bluebrotherthen I don't get the point.
23:06:24Lloreanbluebrother: That's why I was asking Nico_P if the data area (the nested ringbuffer) could expand as area in the main buffer was freed.
23:06:29LloreanIt doesn't need to ring at that point yet
23:06:44LloreanIt only needs to ring once it meets its own metadata (or a block that hasn't been freed, but that should be the only one)
23:07:15bluebrotherbut what is the ring-in-ring thing then? Is it not just some blocked part of the main buffer?
23:07:20Nico_PLlorean: very true. I think I'm close to having this behaviour
23:07:38 Quit desowin ("use linux")
23:07:58bluebrotheri.e. the metadata block?
23:08:02Lloreanbluebrother: It has pretty much the same results, iiuc.
23:08:17 Join stripwax [0] (
23:08:19Nico_Pbluebrother: pretty much, but the metadata block can be anywhere in the buffer
23:08:21bluebrotherwell, then I don't understand what the buffer-in-buffer should behere.
23:08:39bluebrotherNico_P: I'm assuming the metadata block can be anywhere in the buffer.
23:09:36Lloreanbluebrother: One way is "have the main ringbuffer skip over the metadata", the other way is "allocate all the buffer other than the part the metadata uses as a new ringbuffer"
23:09:37amiconnWell, if the buffer wraps, you could also move the metadata block
23:09:40LloreanThey both do the same thing.
23:09:46bluebrotherso as far as I understood by now it's just music data interleaved with metadata, and that metadata gets a special protection to get freed only if the track it's assigned to has ended playing. the physical start of the ringbuffer, and then ring behind that, until the metadata block is freed
23:10:57bluebrotherringbuffer for the metadata?
23:11:16 Quit ompaul ("Leaving")
23:12:05Nico_Pamiconn: to me that seems harder to manage
23:12:44 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
23:12:46amiconnI would think that's less complex than managing holes one way or another
23:13:03bluebrotherLlorean: and what is the exact difference between both? I really don't get that new ringbuffer thing.
23:13:20bluebrotherisn't it just a protected block? What's the ring for the metadata?
23:13:26LloreanThere's no ring for the metadata.
23:13:35Nico_Pbluebrother: it's essentially the same
23:13:46linuxstb_amiconn: But it also moves the complication to other parts of Rockbox - that may be assuming the metadata isn't moving. Although maybe that is managable (I don't know those parts...)
23:14:09bluebrotherhmm. So why call it ringbuffer-in-ringbuffer? I don't see something like that (or I'm too tired to understand right now=
23:14:29LloreanBecause the whole playback buffer is a ringbuffer, just by definition.
23:14:29Nico_Pbluebrother: a long track's buffer will become a nested ring buffer
23:14:50LloreanThe data area of a very long track can become a nested ringbuffer if necessary so that the metadata at the start of the track does not need to be overwritten.
23:14:57Nico_Pbluebrother: have you seen the picture on the MoB wiki page ?
23:15:01bluebrotherwhy nested?
23:15:16bluebrotherNico_P: yes. Checking it again ...
23:15:22LloreanIt's only nested in the sense that it's partitioned off from the original ringbuffer.
23:15:42bluebrotherwell, so it's more like a fragmented buffer, right?
23:18:44linuxstb_One problem I see is when it comes to buffering the next track after a long track. If the low watermark is 1MB, then when 1MB (minus metadata size) of the long file is left in the buffer, you want to fill the buffer with the upcoming tracks. But you still can't overwrite the metadata...
23:19:30 Part Llorean
23:20:51*linuxstb_ wonders if the current track's metadata should always be moved to be at the tail of the data for the track.
23:21:35*Nico_P was pondering head or tail
23:21:57linuxstb_If it's head, then there will always be a hole.
23:22:18 Quit FOAD_ (Read error: 110 (Connection timed out))
23:22:19Nico_Pyeah with tail the problem you mentioned above shouldn't happen
23:22:39bluebrotherwhy can't you just skip the metadata and use the space afterwards? But as I understand the term "ringbuffer" then there is the ringbuffer of the currently playing file which you can't overwrite.
23:23:18Nico_Pbluebrother: once it stops being a ring it can become shorter and shorter to make room for new tracks
23:23:19bluebrotherwhich makes it somewhat problematic for me to understand this ring-in-ring thingy
23:23:46*bluebrother wouldn't call this a ring anymore
23:24:05Nico_Pyes, there's a point where it stops being a ring
23:24:05bluebrotherit's more like a snake that crawls through the buffer
23:24:22bluebrotherand why does it need to be a ring in the first place?
23:24:32Nico_Pit's a snake carrying a ringbuffer
23:24:47bluebrotherwhy does it carry a ring?
23:25:08Nico_Pbecause it can't always be as long as the file
23:25:20bluebrotherand what's the purpose of that ring the snake (audio) is carrying?
23:29:33Nico_Pbluebrother: I don't see whay you mean...
23:30:13bluebrotherwell, the audio data is getting loaded, so it kinda crawls like a snake through the buffer.
23:30:23bluebrotherand it has the metadata attached.
23:30:38bluebrotherbut you said that this is a ring too. Why?
23:32:10 Quit Jens ()
23:33:03Nico_Pwhat you call the snake is actually the concatenation of all the data in the main buffer... the head is the writing position and the tail is the reading. this snake carries little snakes on his back, each of them crawling through a part of the big snake's back
23:33:51stripwaxso the snake itself doesn't really move?
23:33:55bluebrotherok ... but why are those little snakes moving?
23:34:24Nico_Pmost of them aren't, but there might be a big one that will need to move
23:34:39Nico_Pstripwax: the big one moves forward too
23:34:50Nico_Pbut only because it follows the small ones
23:35:17bluebrotherso ... the moving big-file-snake moves the metadata around with itself?
23:36:21Nico_Pno, it moves in the part the metadata left him
23:36:54 Join donutman25 [0] (n=chatzill@
23:36:55jhMikeSsnakes, little snakes? this IS serious. what do they eat?
23:37:09Nico_Phehe :)
23:37:14bluebrotherso if there is some metadata block left it just cycles between that block (end) and the place it started?
23:37:36bluebrotherso it might become a much smaller portion of the big world?
23:37:36*petur wonders if they'll eat album art :)
23:37:46bluebrotherjhMikeS: they eat metadata ;-)
23:38:44*bluebrother thinks this would be a good talk for DevConPub or similar ...
23:39:52markunNico_P: can you take a look at this?
23:40:02markunIt is more or less what you had in mind?
23:40:27markunIf yes, do you have an answer to the question?
23:40:59Nico_Pmarkun: I don't see how to read the diagram at the bottom
23:41:12Nico_Peach line is a state ?
23:41:34Nico_Pok I see now
23:41:44markunthe first line in the empty buffer
23:41:53markunthe second when starting playback
23:42:04markunthe others after rebuffering was needed
23:43:09Nico_Pmarkun: in the third line, BM wouldn't replace A3 like that
23:43:28markunWhy not?
23:43:35markunIt has been played
23:43:54markunWhere should BM be stored after buffering all of A?
23:44:46Nico_Pwell that's the problem with storing the metadata before the audio data
23:45:04markunSo where do you suggest to store it?
23:45:13bluebrotherwouldn't there be a similar problem when storing the metadata after the audio data?
23:45:28bluebrotheryou still have metadata block somewhere in the way.
23:45:33Nico_Pno, because the end of the audio and the metadata are basically freed at the same time
23:45:48Nico_PI'll try doing a diagram similare to yours
23:45:56markunok, great
23:46:09bluebrotherthen you need to move the metadata around upon every rebuffering?
23:46:09jhMikeSthe metadata is just really any bytes you want associated with a particular track?
23:46:13markunI'll try to come up with a fix as well
23:46:21Nico_PjhMikeS: yes
23:46:38bluebrothermaybe drawing before buffering -> after buffering state helps.
23:47:12markunNico_P: where would you place AM for example?
23:47:30Nico_Pright after A4 in the second line
23:47:34markunthe end of the file cannot be buffered yes because it's too big for the buffer
23:47:59markunok, let me see
23:48:16bluebrotherand then you get |A5|A6|B1|BM|AM| in the next step
23:49:34 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
23:49:38bluebrotherbut if A has a length of 5, you'll get |A5|BM|B1|B2|BM|AM|
23:50:06bluebrotherhow do you buffer in the next step? I.e. if AM is freed, you you continue buffering at the AM position or at A5?
23:50:14markunshouldn't you buffer when A4 is playing?
23:51:02*linuxstb_ invites everyone to London next Tuesday to discuss this over a beer...
23:51:14linuxstb_(with a large sheet of paper)
23:51:24*petur hates these kinds of triggers
23:51:36*bluebrother would love to come as that is definitely a much better way of discussing it.
23:51:37markunlinuxstb_: I hope to be there some time
23:51:55 Join rotator [0] (n=e@rockbox/developer/rotator)
23:51:55pixelmalinuxstb: you pay the flight? ;)
23:52:14bluebrotherI wanted to visit London for a couple of days since a bunch of years now. One day I'll manage ...
23:52:18*linuxstb_ hides
23:52:20*jhMikeS asks the same ;)
23:52:57Nico_PI have an example coming...
23:53:00markunbluebrother: do you live near Berlin?
23:53:13bluebrothermarkun: no (at least not yet) −− Karlsruhe.
23:53:28markunbluebrother: not yet?
23:53:41linuxstb_Nico_P: What do you think of the idea of moving the metadata? It seems it may simplify the buffering...
23:53:51bluebrotherI just finished uni, so let's see where I find a job. Want to leave southern germany ;-)
23:53:57linuxstb_But I'm not sure how much moving would need to be done...
23:54:13bluebrotherisn't the metadata quite small anyway?
23:54:22Nico_Pthe metadata could be quite big
23:54:44bluebrotherin which cases? Images?
23:54:48saratogaLlorean: I deleted that off topic thing in the E200R thread
23:54:50Nico_Pfor example
23:54:53saratogahope you were done with that guy
23:54:53bluebrotherin the 4th line, where is A6 gone?
23:55:00markunNico_P: it looks a bit inefficient
23:55:11markunbecause you don't buffer as much as you can
23:55:37markunbluebrother: indeed, it's overwritten..
23:55:58bluebrotherthat's about what popped into my head when I heard about that ring-in-ring thing.
23:56:19markunMaybe we should put these examples in a wiki so everyone can chip in :)
23:56:22bluebrotherwell, we could assume A to be only of size 5 for now ;-)
23:56:25Nico_Perr yes I obviously got confused
23:56:36bluebrotherthis _is_ quite confusing ;-)
23:56:52*linuxstb_ suggest b33r mats on a table...
23:57:14 Join My_Sic [0] (
23:57:43*Nico_P is starting to think moving the metadata around might be a better idea
23:57:52bluebrotherwell, how big might the metadata in the worst case?
23:57:53*stripwax is still lost
23:58:14Nico_Pbluebrother: the current worst case would be AA + cue + ID3
23:58:40bluebrotherok, but how much will this get in bytes? Do you have some numbers for me?
23:58:52 Quit hannesd (Read error: 110 (Connection timed out))

Previous day | Next day