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).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2009-06-30

00:00:07kugelit does?
00:00:23saratogawe had this happen before due to alignment issues, where random changes in the code would fix or break the problem between builds
00:00:23kugelthat doesn't make sense though, it should be strafing ;(
00:00:33kugelor turning at least
00:01:19kugelfunman: if you assign a drive name in windows, linux will recognize it
00:01:47funmansaratoga: yeah that's why I think about alignement and looked at that sd_select_banjk
00:02:03funmanunfortunately i have never seen fs corruption on my fuze since fs#10048 was committed
00:06:42kugelwhy don't we define NEED_CACHE_ALIGN btw=
00:07:23 Quit DarkDefender ("Leaving")
00:07:41funmanwhy would we ?
00:08:01funmanI see that some code depends on it in mpegplayer for example, and I didn't understand what this code did
00:10:31saratogais there a windows binary of mkamsboot?
00:12:03LloreanIt sounds like 3.3 may be pretty broken. There seem to be an awful lot of bug reports from people coming from 3.2 relative to what I remember from 3.1->3.2
00:14:56funmansaratoga: all 1.0 binaries were removed, i can give you a binary if you wish
00:15:17saratogait still lacks the md5 check right?
00:15:28funmanSVN has the check
00:16:07kugelFlynDice: what was the fix for the white screen?
00:16:26 Quit PaulJam (".")
00:17:23saratogado you know who compiled them originally?
00:17:39 Quit evilnick ("Page closed")
00:17:50FlynDiceput this right before the ascodec read ascodec_write(AS3514_ADC_0, 4<<4); /* ADC Source = CVDD */
00:18:38funmansaratoga: me
00:20:34funmansaratoga: kugel is working on mkamsboot, and i would like to check if everything is allright before we release mkamsboot again
00:20:35kugelFlynDice: before the while loop?
00:20:47saratogasounds good
00:21:04 Quit petur (Remote closed the connection)
00:21:11FlynDicekugel: yes right before the while loop
00:22:16funmanare released binaries packed with upx ? that can save a bit of bandwidth/server storage
00:25:22kugelplain zip
00:26:00kugelFlynDice: seems good, I added an entry to the hw info debug page and it shows the expected values
00:31:33|niko|# chmod 777 .rockbox/
00:31:33|niko|chmod: changing permissions of `.rockbox/': Read-only file system
00:31:43|niko|I ran that as root, and it is vfat and clearly mounted as rw
00:31:56|niko|I can copy stuff to and from the filesystem, just not the .rockbox dir
00:32:11rasher|niko|: err, cat /proc/mounts
00:32:34|niko|rasher: /dev/sdf2 /mnt/ipod vfat ro,fmask=0022,dmask=0022,allow_utime=177777,codepage=cp437,iocharset=utf8 0 0
00:32:39funman|niko|: and verify the filesystem before anything
00:32:40JdGordon|kugel: do you want to comment on the viewportmanager stuff?
00:32:50funman|niko|: "ro" clearly shows readonly
00:32:51*JdGordon| is hoping it didnt get drowned out by the current thread
00:33:01rasher|niko|: what funman said
00:33:10|niko|Why does the mount command show:
00:33:17|niko|this: /dev/sdf2 on /mnt/ipod type vfat (rw)
00:33:17rasher|niko|: on read/write errors, it will get remounted read-only
00:33:55rasher/proc/mounts is more authoritative than the mount command output I believe
00:35:05 Join Lazatar [0] (
00:35:46kugelJdGordon|: maybe :)
00:35:56 Quit Lazatar (SendQ exceeded)
00:36:21notlisteningrasher i am trying to get the voice list back from the command line and it I am not quite sure what this cod is doing...
00:36:27 Quit CIA-71 ()
00:36:28 Join Lazatar [0] (
00:36:30kugel|niko|: like I said, linux tends to remount in read-only mode if it finds FS issues
00:36:43ForumJunkieIs there supposed to be an m when I load up a patched Rockbox in the firmware (ex. rXXXXX-XXXXXm)?
00:36:43kugelthat doesn't show up in the mount commands
00:36:44|niko|mkfs.vfat turns up errors that I want to fix
00:36:54 Quit BeChris (Ping timeout: 180 seconds)
00:36:59JdGordon|ForumJunkie: yes
00:37:05JdGordon|the m means modified
00:37:06|niko|and I do fix but, at the end of me telling it what to do:
00:37:08|niko|Leaving file system unchanged.
00:37:08|niko|/dev/sdf2: 2717 files, 2065459/4873343 clusters
00:37:10 Nick fxb is now known as fxb__ (
00:37:31funman|niko|: rtfm, use fsck.vfat -a
00:39:37 Quit bertrik ("De groeten")
00:39:53 Join Thundercloud [0] (
00:42:20rashernotlistening: you are more or less free to do it however you like
00:42:36rashernotlistening: as long as it's proper shell code, and you put it in the options
00:43:31 Quit martian67_ (Connection timed out)
00:44:23ForumJunkieI have two versions of source code in Cygwin, I can't apply a patch, what can I do? I think my patched version is still there but I deleted it myself. (There's revision XXXXX and "working copy")
00:45:09saratogawhat is your question
00:45:36ForumJunkieWell, I can't patch it, but when I go to make zip
00:45:44ForumJunkieIt gives me the simple source code unedited.
00:46:06saratogaare you asking how to patch?
00:46:08ForumJunkieWhen I go to patch it, it prompts me to patch two files.
00:46:15ForumJunkieI know how to do that, I did it before.
00:46:17notlisteningrasher, does not help then i am running a compiled and a non comiled version :D one with the changes and the other without :D school boy errors
00:46:55 Join CIA-69 [0] (n=CIA@
00:47:02 Quit funman ("free(random());")
00:47:42 Join at0m [0] (
00:48:22 Quit saratoga ("Page closed")
00:49:46***Saving seen data "./dancer.seen"
00:52:12JdGordon|kugel: I wrote that email with your patch in mind, so if you have done alot of it already thats good...
00:53:25 Quit faemir (Read error: 60 (Operation timed out))
00:56:07 Quit matsl (Read error: 110 (Connection timed out))
00:58:48 Join barrywardell [0] (
01:07:12*kugel wishes we can get back to the state were confused happy users aren't our target audience but ourselves
01:07:32 Join AndyIL [0] (i=AndyI@
01:11:55kugelI really can't understand that those confused happy users, which don't even contribute the features in question (or at all), are the reason we can't implement things we'd like to see
01:12:19notlisteningrasher, I want to include a "?" in a command line from the configure script is there a trick to do that?
01:12:27 Quit linuxstb (Remote closed the connection)
01:12:33Mikachunotlistening: \?
01:12:40Mikachuor indeed "?"
01:12:51rashernotlistening: Try escaping it I guess. It's hard to tell without more specifics
01:12:57pixelmakugel: what if I tell you that I wouldn't want it? ... ;)
01:13:08kugelthat's another thing
01:13:27Mikachukugel: if you want to develop things only you yourself use, feel free?
01:13:34Mikachuno point in committing them then either
01:14:08kugelthat's not practical in the long term, I tried it (before I could commit) :P
01:14:23Mikachuhm, i've had custom patches against rockbox for >4 years
01:14:28 Quit |niko| ("KVIrc Insomnia 4.0.0, revision: , sources date: 20090520, built on: 2009/06/06 12:36:26 UTC")
01:14:41Mikachui even stopped looking at svn for 2 years and rebased, almost no conflicts :)
01:15:16kugelI had dozens of conflicts after ignoring SVN for just a month
01:16:07kugelI'm not arguing about features where I'm obviously the only one who wants it. I'm arguing against "helpless users" being an excuse for rejecting things
01:17:36 Join faemir [0] (n=faemir@
01:18:27 Quit AndyI (Read error: 110 (Connection timed out))
01:18:41 Join robin0800 [0] (
01:19:27*kugel hopes he doesn't forget this topic for DevCon2010
01:20:52 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
01:22:41 Join Hillshum [0] (n=chatzill@unaffiliated/hillshum)
01:30:28 Part domonoky1
01:31:04 Quit tessarakt (Read error: 110 (Connection timed out))
01:31:25 Join tessarakt [0] (
01:32:11 Quit Thundercloud (Remote closed the connection)
01:32:54lilltigerI just reformated both my player and the microSD, readded the rockbox firmware and the .rockbox folder. I also added a few songs to the microSD card. But now the microSD wont show up in rockbox at all, and when I tried to build an database it found 388 items then seems to freeze.
01:34:18Hillshumlilltiger: what player, x5?
01:34:23 Quit tessarakt (Client Quit)
01:34:25 Part JdGordon|
01:34:27 Join JdGordon| [0] (n=Miranda@rockbox/developer/JdGordon)
01:34:30kugelprobably not
01:34:40rasherlilltiger: Try removing half the files, then re-initializing. If it works, try with the other. Keep halfing the collection until you figure out which file causes the freeze.
01:34:59*JdGordon| *REALLY* hates that solution
01:35:04kugelpossibly broken tags or something
01:35:13JdGordon|can someone just put a damn logger in the database builder pleasE?
01:35:15MikachuJdGordon|: it's a hell of a lot better than trying them one at a time :P
01:35:40lilltigerHillshum: fuze
01:36:17kugelwhat revision are you running?
01:36:29rasherJdGordon|: I don't even understand why it'd be so difficult. Just keep track of the last opened file. It doesn't usually crash Rockbox, does it?
01:36:53Mikachuor just display the currently processing file on screen, that way you have plenty of time to write it down when it dies :)
01:36:56JdGordon|yeah, it just needs someone to do it
01:37:10Mikachu(i guess optionally, or it would slow it down a bit)
01:37:19Mikachu(or can't run in the background)
01:37:29linuxstbDo we have such bugs though? i.e. do people end up identifying a specific troublesome file?
01:37:47rasherlinuxstb: They don't, because it's too much work..
01:38:04Mikachuyou just need to try ~8 times
01:38:09kugelevery time I heard of such an issue it turned out to be broken tags
01:38:31linuxstbkugel: Then we need to get these people to post bug reports...
01:38:32Mikachuyou can optimize the process a bit over what rasher said
01:38:39Mikachutry one half, if that works, try _half_ of the other half
01:38:45Mikachuyou already know all of the other half will crash
01:38:56kugellinuxstb: We have people reporting bugs based on broken tags
01:39:02linuxstbdatabase.ignore should help - no copying of files, just moving them around...
01:39:19kugelonce the tag is identified, it's closed as not a bug or something
01:39:24rasherThis is all too much work for someone who just wants the Database working
01:39:35Hillshumlilltiger: this is to be expected on a Fuze
01:39:40Mikachuit's obviously a bug if it hangs the scanning
01:39:41rasherkugel: seriously?
01:39:42linuxstbkugel: Well, that's wrong...
01:39:43kugelHillshum: not really
01:39:56kugelI didn't do it myself, but I've seen it
01:40:02kugelat least once
01:40:09Hillshumkugel: Oh?
01:40:31lilltigerkugel: is rockbox picky on special characters in tags?
01:40:56kugelwe are trying to iron out the last serious issues to get the fuze supported, today we've had a commit which should resolve some or even all storage problems
01:41:11rasherkugel: if you can find such a closed bug, it should be re-opened
01:41:14kugellilltiger: possibly, I don't know exactly
01:41:24rasherA crash or hang is *always* a Rockbox bug, no matter how broken the tags are
01:41:34 Quit tvelocity ("Αποχώρησε")
01:41:46 Join Blue_Dude [0] (
01:42:22Blue_DudeHi. Just posted FS #10400 - Consolidate all fixed point math routines in one library.
01:43:21 Quit robin0800 ("Leaving")
01:44:53lilltigerr21540-090628 is the revision
01:46:01lilltigerand adding a few albums enabled rockbox to build the database, see like it stalls/fails to finish the database building when there is no music avaible
01:46:08kugelshould be new enough
01:46:45lilltigerBut i wonder why it cant see the microSD any longer
01:46:51kugelyea, the database fails if there's no music at all. weird bug
01:47:11kugelI don't know
01:47:28kugelcan you compile?
01:49:13kugelyou could try to add some mci_delay() in sd_init_card() in firmware/targe/arm/as3525/ata_sd_as3525.c
01:49:29kugellike where the other 4 are or somewhere else
01:49:50kugelwe've been adding delays there over the time, some cards just don't work
01:50:16 Part toffe82
01:50:21lilltigerit workt fine before the format, could i have formated it to an fat version that rockbox disslikes?
01:50:55kugelwhat revision have been running before?
01:50:56kugelthe same?
01:51:03lilltigeryes the same
01:51:14kugelthat's weird then
01:51:52rasherHow did you format the card?
01:52:05rasherMore to the point, did you partition the card?
01:52:26lilltigerrasher: yes I did partion it using fdisk
01:52:42rasherI'm not sure that's smart
01:52:56rasherDoes rockbox even look at partitions on sd cards?
01:53:00kugelisn't that mkfs' job?
01:53:06lilltigerkugel: fdisk is an linux tool for partition managment
01:53:26rasherkugel: mkfs makes the filesystem, fdisk makes the partitions
01:53:30kugelrasher: probably only at the first one, we need to poke gevaerts to finish off his multivolume work
01:53:43lilltigerkugel: mkfs just formats the drive dosent repartion it
01:53:54Mikachumultivolume works fine on my nano, fwiw :)
01:53:56rasherlilltiger: try doing the mkfs on the whole disk
01:53:57kugelyou don't need to repartion it
01:54:32kugelMikachu: I mean multi as in more than 1 or 2 volumes
01:54:59lilltigeri wonder what the OF of sansa does, strange that it found it
01:55:20kugelthe OF does magic w.r.t. to filesystems
01:55:25lilltigeri accidently repartioned it to an Linux partition
01:55:25JdGordon|Mikachu: you've enabled multivolume on your nano?
01:55:32JdGordon|wtf for?
01:55:42MikachuJdGordon|: yes, i made a 32MB partition in the space where the OF used to put it's hibernation image
01:55:46krazykithow do you accidentally manage that?
01:56:08Hillshumlilltiger: mkdosfs /path/to/device
01:57:17lilltigerkrazykit: fdisk sts it by default and i forgott to override the default
01:57:56lilltigersets, -f
01:58:31lilltigerand now rockbox finds the microSD
01:59:05kugelgood thing you didn't use fdisk for the internal storage?
01:59:14kugelalso I think that's recoverable
01:59:56 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier.")
02:00:39lilltigerkugel: as I dident know if the internal storage included the bootsector or not I just used the in player format for it, but I havent found an in player way to format the microSD.
02:03:57lilltigerI just triggered the "update database" to see if it adds the 3 albums i have on the microSD, it said that updating started in the background, but now it has elapsed several minutes and I dont see any information of it's progress. How to know when it has finished updating the database?
02:05:05 Quit faemir (Remote closed the connection)
02:06:07 Join mc2739 [0] (
02:08:02mc2739kugel: should FS #10157 be closed now since it was committed with r21570?
02:08:32rasherlilltiger: There's a tiny disk activity indicator in the top right corner if your theme has the statusbar enabled
02:08:57rasherlilltiger: Once that goes away, it's done. Should be over in a few seconds with just 3 albums
02:09:15lilltigerrasher: right of the time?
02:10:11lilltigerand when it's gone I need to restart of the update to be avaible?
02:12:51lilltigerfound em, althought it made an duplicate of what seems to be the same albumartist name, have to check the tags
02:15:05 Quit Hillshum (Read error: 60 (Operation timed out))
02:17:00 Join JdGordon| [0] (i=209958e6@gateway/web/freenode/x-6cda93af9a868aa3)
02:20:27 Quit JdGordon| (Ping timeout: 180 seconds)
02:21:27kugelmc2739: I thought I did, thanks for notifying me
02:22:16mc2739kugel: you're welcome
02:22:53kugelhow's your e200v2 going?
02:24:57mc2739just installed latest svn - not seeing any problems yet
02:25:44 Quit efyx ("Quitte")
02:36:05 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
02:41:53 Quit kugel (Remote closed the connection)
02:43:39 Quit barrywardell ()
02:49:48***Saving seen data "./dancer.seen"
02:57:02 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
03:09:36 Join perrikwp1 [0] (
03:13:45 Join patmulchrone [0] (
03:19:50 Quit perrikwp1 ("Leaving.")
03:21:57 Quit perrikwp (Read error: 110 (Connection timed out))
03:25:59 Quit notlistening ("Leaving")
03:50:06 Quit mc2739 ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060215]")
03:51:37 Quit pixelma (Nick collision from services.)
03:51:39 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
03:51:47 Quit amiconn (Nick collision from services.)
03:51:51 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
03:51:56 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
03:51:57 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
04:14:59 Quit Llorean ("Leaving.")
04:16:21 Quit patmulchrone (Read error: 110 (Connection timed out))
04:27:16 Quit Sajber^ (Read error: 104 (Connection reset by peer))
04:35:49 Quit Rob2222 (Read error: 110 (Connection timed out))
04:49:49***Saving seen data "./dancer.seen"
04:53:03 Quit ForumJunkie ()
04:59:05 Join dlenski [0] (
05:02:51 Join Rob2222 [0] (
05:12:57 Quit Blue_Dude ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060215]")
05:20:06lilltigerrasher: the problem reapered, but it's only present in the main storage device and it seems thay the fuze automaticly creates a second FAT table after about 2/3 are filled or so. And rockbox seems not to be able to access this second fat table correctly.
05:28:34 Join patmulchrone [0] (
05:45:14 Quit dmb (Read error: 113 (No route to host))
06:05:41 Quit Horscht ("Verlassend")
06:07:01 Quit courtc (Read error: 60 (Operation timed out))
06:08:38 Join evilnick_home [0] (
06:12:35 Quit Strath ("Leaving")
06:13:20 Join ademille [0] (
06:20:55 Part ademille
06:23:09 Join __lifeless [0] (n=lifeless@
06:23:55 Quit _lifeless (Read error: 104 (Connection reset by peer))
06:27:00 Quit evilnick_home1 (Read error: 110 (Connection timed out))
06:27:17 Quit _Auron_ ("Infinity repeatedly denies rumours of plotting with zero to bring down the Universe.")
06:34:26 Quit Zarggg (Read error: 104 (Connection reset by peer))
06:37:01 Quit __lifeless (Remote closed the connection)
06:37:09 Join __lifeless [0] (n=lifeless@
06:37:12 Join _Auron_ [0] (n=DarkAuro@
06:39:26 Quit bzed (Read error: 104 (Connection reset by peer))
06:39:29 Join bzed [0] (
06:49:53***Saving seen data "./dancer.seen"
06:50:04 Join Zarggg [0] (
07:06:59 Join Gigawatts [0] (n=chatzill@
07:24:20 Join webguest28 [0] (n=443a7fba@gateway/web/cgi-irc/
07:24:44webguest28okay so can i ask whats going on with the creative zen vision m rockbox?
07:26:00 Quit webguest28 (Client Quit)
07:27:30 Join newnick [0] (n=443a7fba@gateway/web/cgi-irc/
07:27:49newnickcan anyone help me
07:28:08 Join einhirn [0] (
07:28:16newnickcan someone help me
07:28:35newnicki am wondering what is going on with the creative zen vision m rockbox
07:31:35 Quit newnick (Client Quit)
07:32:24safetydanso... this whole patience thing? I guess we've run out
07:34:32 Nick fxb__ is now known as fxb (
07:38:44 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
07:47:24 Quit dlenski (Read error: 110 (Connection timed out))
07:51:09 Join ZZeik [0] (
07:51:18 Quit alexbobp (Read error: 104 (Connection reset by peer))
07:53:20ZZeikon the rockbox utility it auto dectects my player as a Sansa e200v1 series, i have a e250, guess it dont matter that its a e250 since its a e200 series?
07:54:38 Join alexbobp [0] (
07:55:05 Nick fxb is now known as fxb__ (
07:55:18 Quit kachna|lappy (Read error: 113 (No route to host))
07:57:36JdGordonZZeik: the e250 is the e200 series
07:58:08ZZeikthought so :D, thank you, 1 more question if i may
07:59:07ZZeikright now i have Build Date: 2006.11.03Build Number: (Build 32.10) cpchan branch, can i use the program to update it?
07:59:23ZZeikjust reinstall a new rockbox or would i have to uninstall it first?
07:59:31 Quit Zarggg (Read error: 113 (No route to host))
07:59:33JdGordonjust reinstall
07:59:41JdGordonthats a bloody old build
08:00:05ZZeikya it is, its been a while since i used it
08:01:11ZZeikwill it mess up anything if i hit complete install even tho i already have rockbox installed?
08:04:13ZZeiknice, it even has a backup option
08:11:54ZZeikthank you JdGordon for your help
08:11:59 Quit bmbl (Read error: 104 (Connection reset by peer))
08:13:38 Join courtc [0] (n=court@unaffiliated/courtc)
08:16:02 Quit Cory` ()
08:17:56 Quit ZZeik ()
08:18:41 Quit Gigawatts (No route to host)
08:19:53 Join Zarggg [0] (
08:20:47 Join flydutch [0] (
08:25:24 Join ender` [0] (
08:38:31 Quit safetydan ("Leaving.")
08:42:35 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/freenode/x-7f1b3e3573159793)
08:49:54***Saving seen data "./dancer.seen"
08:51:00 Join Rob2223 [0] (
08:55:02 Join Grahack [0] (
09:03:58 Quit alexbobp (Read error: 104 (Connection reset by peer))
09:05:23 Join Fnargl [0] (
09:05:24 Quit Fnargl (Client Quit)
09:06:33 Join alexbobp [0] (
09:06:54 Join Plerg [0] (
09:09:19 Quit Rob2222 (Read error: 110 (Connection timed out))
09:10:38 Join matsl [0] (
09:27:53 Join dlenski [0] (
09:28:14 Join petur [50] (n=petur@rockbox/developer/petur)
09:32:28 Quit mt (Read error: 113 (No route to host))
09:32:56 Join Thundercloud [0] (
09:34:44 Quit BHSPitMonkey (Remote closed the connection)
09:38:09 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
09:38:10 Join efyx [0] (
09:40:01 Join jfc [0] (
09:51:22 Quit Thundercloud (Remote closed the connection)
09:56:49 Quit jfc^3 (Read error: 110 (Connection timed out))
10:01:12 Quit Kopfgeldjaeger (Read error: 60 (Operation timed out))
10:01:39 Join Kopfgeldjaeger [0] (
10:01:53 Quit freqmod_qu (Read error: 60 (Operation timed out))
10:01:57 Join freqmod [0] (
10:14:41 Quit bmbl ("Woah!")
10:15:38 Join ForumJunkie [0] (
10:15:47ForumJunkieAnyone know why I can't install AAFreak theme off the site?
10:16:00ForumJunkieUsing the utility or manually
10:16:28ForumJunkieIt says "extract failed"
10:16:40GodEaterthe zip file is broken ?
10:16:54ForumJunkiewell the utility + the manual download doesn't work
10:17:07ForumJunkieAnd I've been reloading RB everyday
10:17:20ForumJunkieAnd it's worked.
10:17:47 Quit dlenski (Read error: 110 (Connection timed out))
10:22:18ForumJunkieAny idea what it could be?
10:22:32ForumJunkiea zip file doesn't usually go corrupt; it's the link that goes broken.
10:24:52 Quit ForumJunkie ()
10:31:09 Join barrywardell [0] (
10:41:26GrahackHi everyone, could someone try to explain to me how the constant BUTTON_REPEAT is given a value ? I'm still struggling to have control on buttons detection with the lua plugin and BUTTON_REPEAT and BUTTON_REL are still missing (though I found a RB_CONSTANT(BUTTON_REPEAT); in apps/plugins/lua/rocklib.c that is cryptic to me).
10:45:07 Join s1mple_m4n_ [0] (
10:45:55 Nick zitune[afk] is now known as zitune (n=zitune@bearstech/zitune)
10:49:46 Join kugel [0] (n=kugel@rockbox/developer/kugel)
10:49:55***Saving seen data "./dancer.seen"
10:54:35 Join robin0800 [0] (
10:56:18 Join kachna|lappy [0] (
10:57:11linuxstbGrahack: Those values are bitwise or'ed with the value of the button - e.g. "BUTTON_MENU | BUTTON_REPEAT" (in C, I don't know Lua).
10:58:40 Quit cool_walking_ ("Page closed")
11:00:02kugelGrahack: you generally need to prepend "rb." for rockbox functions and constants
11:19:15Grahackthanks linuxstb and kugel, I already knew that. Sorry for not being clear. The creator of the Lua plugin managed to expose BUTTON_* constants to Lua, I managed to use bitwise operations on this constants (using rb.buttons.*) but BUTTON_REL and BUTTON_REPEAT are not automaticly created by the Lua plugin. I have to add manually those codes: rb.buttons.REL = 0x02000000 and rb.buttons.REPEAT = 0x04000000 that I get on my iPod. My questio
11:20:10 Nick Lazatar is now known as martian67 (
11:21:25kugelint btn = rb.button_get(rb.ZH, false); if (btn == rb.button.BUTTON_XX|rb.button.BUTTON_REPEAT); .....; end [or whatever the correct lua syntax is)
11:23:18 Join AndyI [0] (n=pasha_in@
11:25:54 Quit AndyIL (Read error: 104 (Connection reset by peer))
11:26:14 Join n1s [0] (n=n1s@rockbox/developer/n1s)
11:31:04GrahackSorry again kugel, I know how to do that. Gonna try to be clear. exposes few BUTTON_* constants to Lua *except* BUTTON_REL and BUTTON_REPEAT. processes the button-target.h file. But button-target.h files don't define BUTTON_REL and BUTTON_REPEAT. Where in the core rockbox code are BUTTON_REL and BUTTON_REPEAT defined
11:32:38 Join PaulJam [0] (
11:32:50kugelbutton.h I think
11:34:45CIA-69New commit by 03mcuelenaere (r21573): Lua: also expose BUTTON_REL, BUTTON_REPEAT & BUTTON_TOUCHSCREEN
11:35:39CIA-69New commit by 03mcuelenaere (r21574): Correct comment
11:36:56Grahackyou are right, thanks a lot, and thanks mcuelenaere !
11:42:15kugelI'm wondering why we make lua scripts so target dependant
11:43:30kugelCan't there be a way to expose BUTTON_TOUCHSCREEN so that it's possible to have runtime-detection of touchscreen? I'm not really a fan of a scripting language which only runs on 1 target
11:51:50 Part kugel ("return(0);")
11:52:02 Join kugel [0] (n=kugel@rockbox/developer/kugel)
11:58:31GrahackI don't understand (yet) why you say that Lua scripts are target dependant. And by the way which Lua scripts ?
12:07:24 Quit matsl (Read error: 110 (Connection timed out))
12:09:53 Quit n1s (Read error: 60 (Operation timed out))
12:10:13 Join n1s [0] (n=n1s@rockbox/developer/n1s)
12:12:39kugelGrahack: I mean you can't have touchscreen handling on non-touchscreen
12:13:04kugelthe script will be simply broken. I'd rather have some sort of exception handling
12:17:30Grahackbut defining BUTTON_TOUCHSCREEN doesn't mean HAVE_TOUCHSCREEN will be true, mcuelaenere will then have to expose the latter, as well as lots of others, but I don't need them actually :)
12:26:34 Join AndyIL [0] (n=pasha_in@
12:26:50 Quit PaulJam (Nick collision from services.)
12:26:57 Join PaulJam_ [0] (
12:27:10kugeldoes anyone know why tuner.c has only target-specific code?
12:31:12 Quit CaptainKwel (Read error: 104 (Connection reset by peer))
12:31:34 Join CaptainKwel [0] (
12:32:40 Join dmb [0] (n=dmb@unaffiliated/dmb)
12:33:24 Join Sajber^ [0] (
12:33:31 Nick fxb__ is now known as fxb (
12:34:09Torneok, the TFAT stuff needs atomic sector writes, that's why it needs a magical driver
12:34:19*GodEater was just going to say that
12:34:21Tornewhat is the storage on the beast?
12:34:27GodEaterATA disk
12:34:34Torneoh, that's ok then :)
12:34:40Torneata writes are already expected to be atomic generally
12:34:52GodEaterunless someone's done a CF mod to one I'm unaware of
12:34:58TorneCF should be atomic too really
12:35:01GodEaterbut that'll still be over an ATA interface I expect
12:35:02Torneand if not there's nothing we can do about it
12:36:07Torneso yah, using fat1 as a temp copy to do versioned objects is not too had
12:36:16Tornebut i'm somewhat puzzled about this whole directory claim
12:36:46Tornethis is pretty different to what our rugged fat system does :)
12:36:53 Quit s1mple_m4n_ ("Ex-Chat")
12:37:03TorneTFAT avoids the need for the boot time disk scan on unclean unmount
12:37:19GodEaterthat would appear to be the idea, yes
12:38:48Tornebut evidently they've lost 100% interoperability somewhere, with htis directory issue
12:39:26GodEaterThis one is peculiar too : "If a directory is created on the removable media using a Windows desktop OS, and then used with TFAT in Windows Embedded CE, operations on files in that directory may not be transaction-safe."
12:40:45Torneyah, that's what i meant
12:41:02Tornethat seems really odd, nothing else in the description implies directories ar ehandled differently
12:41:23 Quit AndyI (Read error: 110 (Connection timed out))
12:41:25Tornei was assuming they update directory entries by rewriting hte entire cluster containing the entry being modified to a new location and rewiring fat1 to point there
12:41:31Tornethen committing it to fat0 when it's done
12:41:42Tornebut that wouldn't need any particular starting layout on disk to work
12:41:56GodEaterit says something like that further up doesn't it ?
12:42:07GodEaterbut for modifiying files, not directories
12:42:12Tornethat's how it journals file updates i fyou turn that on, yes
12:42:19Torneand directories in FAT are no different to files :)
12:42:25Torneas far as cluster allocation/etc goes
12:42:47GodEaterwe should write our current assumptions down in the wiki I think
12:42:52GodEaterif there isn't a TFAT page already
12:42:59Torneprobably :)
12:43:55GodEaterit's a shame linuxstb already installed Rockbox on it actually
12:44:03GodEaterI wonder if he backed up the partition table
12:44:06GodEaterthat's peculiar too
12:44:19Torneit needs *something* on it to be able to dump anything useful, though, no? :)
12:46:29GodEaterbut just installing the RB bootloader doesn't fiddle with the partition table
12:46:30Tornewell anyway the first step fr me is getting the thing to power on at all
12:46:33GodEaterso you could have dumped that
12:46:40GodEaterwell yes, there is that :)
12:46:41Torneand being able to run a build of RB on it :)
12:46:58GodEaterI suspect it already has one of those on it
12:47:13GodEaterproposal for Wiki page name = TfatDetails
12:47:16GodEateranyone other offers ?
12:48:15Torneyah, but i mean 'an arbitrary build'
12:48:18Tornei.e. one i've compiled with changes :)
12:49:11Tornei'm sure someone somewhere has a beast that's not been touched, anyway
12:49:25Tornei guess they don't have a convenient restore process like the ipod? :)
12:49:40GodEaternot nearly as convenient no
12:49:46GodEaterbut it's not awful
12:49:54GodEaterI only know of one dev who managed to brick one
12:49:56***Saving seen data "./dancer.seen"
12:50:10GodEaterand that was before we had our MTP firmware loader written
12:50:19GodEaterso you had to use the japanese windows things, which was horrific
12:50:32GodEater(unless you a) read japanese and b) have a windows machine somewhere)
12:50:41Tornei run windows mostly :)
12:50:55Tornebut my japanese is strictly amateur :)
12:52:16kugelGodEater: TFATDetails maybe?
12:52:34kugelthe FAT in IpodConversionToFAT32 is all upper case too
12:52:42*GodEater has to run to a meetings
12:56:27linuxstbGodEater: It's impossible to back things up from the S before installing Rockbox...
12:57:18Torne(...well you could pull the disk...)
12:58:12linuxstbs/impossible/easy/ ....
13:11:34 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
13:13:59PaulJam_kugel: i just tried the latest custom_ui_vp patch (v20) on my h300 and i get an empty screen with just the statusbar. this is with reset settings. do you have an idea what's wrong?
13:15:13 Quit kachna|lappy (Read error: 110 (Connection timed out))
13:16:04Tornedoes someone have an image of the gigabeat firmware btw?
13:17:17Tornenevermind, i'm blind, sorry :)
13:30:08 Quit kugel (Read error: 113 (No route to host))
13:52:50 Join wincent [0] (
14:02:23 Quit FlynDice (Remote closed the connection)
14:09:47 Join intrados [0] (
14:11:58Tornethe registry dump from the beast rom doesn't set any of the settings MS's docs say you should set for hard drives :)
14:12:15Torneso either toshiba fail it or they are very sure that their hard drives write sectors atomically
14:13:29 Quit Rob2223 ()
14:13:45 Join Rob2222 [0] (n=Miranda@
14:14:48Torneanyone know if the eboot.bin bit of the firmware image is what gets flashed, or if that's the second stage bootloader after the code in flash has already done whatever it does?
14:24:41*linuxstb looks at the screenshot here and wonders if Pacbox is really rotated on the e200?
14:31:44 Quit freqmod (Read error: 60 (Operation timed out))
14:33:49 Join notlistening [0] (
14:35:29notlisteningrasher, I have a solution to the problem i am having getting bakc a list of voices availalbe for my speech engine but it might not be a great one
14:35:32CIA-69New commit by 03dave (r21575): Correct the screenshot for Pacbox on 176x220 targets (E200) - the screenshot was showing it at 50% size and rotated, whereas it is 75% and not ...
14:37:11notlisteningIt involves sending the fullfy nice output list to stderr and the read values to stdin is that acceptable?
14:38:53notlisteningrasher, this is because my programming knowledge is not up to scratch in this alrea ;)
14:41:16 Join kugel [0] (n=kugel@rockbox/developer/kugel)
14:44:02Tornedo we have the imx31 manual somewhere? :)
14:44:55gevaertsit's somewhere on the freescale website
14:45:05Tornei hate their website
14:45:15gevaertsI'm sure there's a link on the wiki somewhere :)
14:46:38Torneaha, there we go
14:46:46Tornei am excessively blind today
14:47:59 Join LambdaCalculus37 [0] (i=44a0430d@gateway/web/freenode/x-7a8e25df990a0803)
14:49:59***Saving seen data "./dancer.seen"
14:52:14 Join Jaykay [0] (
14:57:58 Quit kugel (Read error: 113 (No route to host))
14:58:59 Join kugel [0] (n=kugel@rockbox/developer/kugel)
15:16:49 Quit kugel (Read error: 110 (Connection timed out))
15:18:06 Quit PaulJam_ (".")
15:21:41 Join kugel [0] (n=kugel@rockbox/developer/kugel)
15:36:46 Quit n1s (Read error: 60 (Operation timed out))
15:37:08 Join n1s [0] (n=n1s@rockbox/developer/n1s)
15:41:27 Quit at0m ("it is now safe to turn of your machine.")
15:42:48 Join at0m [0] (
15:47:36 Quit kugel (Remote closed the connection)
15:48:12lilltigerAnyone arund that is runing rockbox on the sansa fuze with 8GB storade and the latest OF firmware, I have noticed that the OF now creates two FAT tables (probebly the sansa way to handle alot of files). But id like someone to confirm this.
15:50:46Torneer, what? :)
15:51:03Torneyou mean two copies of the FAT?
15:51:24Torneor two partitions?
15:52:23lilltigerTorne: two tables, the OF dosent seem to create any proper partition at all with the latest firmware
15:52:36Tornetwo copies of the FAT is normal for most disks
15:52:41Tornethey are identical
15:52:50Tornehas nothing to do with handling more files
15:53:10Tornesome formatting tools only bother to create one copy to save space
15:53:51TorneYah. That's a corrupt filesystem.
15:54:10TorneIf the two copies of the FAT are different the FS is corrupt, fsck is asking which one to trust.
15:54:31Torneyou probably lost power during a write or similar :)
15:54:39lilltigerno write went perfect
15:54:52Tornewell the filesystem is corrupt now :)
15:54:58Torneso *something* has gone wrong at some point
15:55:09lilltigerand that happend yesterday as well, so i reformated and recopied (other files) and the same thing happend
15:55:23lilltigerand in the OF it's all working well
15:55:28 Join evilnick [0] (i=0c140464@gateway/web/freenode/x-7ff5a8df68d25f26)
15:55:35Torneyah, the FATs being different doesn't make the disk unusable or anything
15:55:45Tornein the DOS era chkdsk didn't even bother to check if they were the same
15:55:54Torne99.99% of software will just use the first copy, always
15:56:06Torneyou're supposed to write to both copies when you change anything
15:56:12Tornebut nothing requires you to actually read the second copy.
15:56:15Torneonly filesystem checking tools do.
15:56:43Tornebut it's still FS corruption
15:57:16lilltigery-day when this also happend, i told it to use the first one, but that changed nothing except for the output from fdisk, the files still howed up curopted in rockbox but correct in OF
15:57:48Tornei would be surprised if rockbox bothers to read the second fat either
15:57:58Torneso it should really work in either firmware :)
15:58:07Tornebut it's still corrupt, so all bets are off
15:58:38lilltigerstrange that it starts getting corrupt at the same breakpoint again
15:59:42 Join zeveroare [0] (n=zeveroar@
16:00:12Tornewell, the alternative is that the OF is doing it on purpose
16:00:42 Join aaron424 [0] (
16:00:45Tornewhich is not impossible (beast does it for complicated reasons, but only temporarily)
16:00:50lilltigerTorne: wich is what i suspected
16:00:54Tornebut seems unlikely, because it will make chkdsk and similar fall over
16:01:04Tornewhich is not a good idea for players that expose their contents via MSC
16:02:07Tornethe transaction system the beast has is a good reason, but even that only does it temporarily and finished writes are supposed to have the FATs being identical again.
16:02:21Torneso it's more likely something going wrong. :)
16:02:22lilltigerTorne: I updated to the latest OF and formated the disk, after that fdisk says that there is no partitio on the drive
16:02:37Tornewell, maybe there isn't
16:03:21Tornei know nothing about the fuze specifically, btw, but generally you aren't required to have a partition table on a MSC device
16:03:50lilltigerTorne: and the info about the F says that it needed to be reformated to allow for higer file count. So I suspect that they create a second FT tale to handle alot of files but this is just a guess
16:03:55 Join BryanJacobs [0] (n=bryanjac@
16:03:59TorneThat's not how the FAT works
16:04:07 Join Gigawatts [0] (
16:04:15Torneit almost certainly had two FATs before
16:04:21Torneit's very rare to have any number other than two
16:04:50Torneit may have reformatted with a different cluster size, or similar
16:05:01Tornehard to say without comparing the BPB before and after
16:05:20Tornebut the number of FATs has nothing to do with how many files you cna fit: the FAT refers to the space on the device, not the files
16:05:32Tornein fact having more copies of the FAT takes up more space and thus reduces the usable size
16:06:47 Quit Gigawatts (Client Quit)
16:06:56lilltigerquite strange that I ran into the issue twice in a row then
16:07:21lilltigerill update rockbox and try again
16:08:18 Nick zitune is now known as zitune[afk] (n=zitune@bearstech/zitune)
16:08:34Torneoh, you're writing to the disk with rockbox?
16:09:04lilltigerIm using the OF fr it
16:09:46lilltigerso I have a hard time to relize why it gets corrupted and only rockbox seems to care about the corruption
16:10:39 Nick zitune[afk] is now known as zitune (n=zitune@bearstech/zitune)
16:11:25Tornerockbox doesn't read the second fat, though we do update it in sync with the first one
16:11:49Tornewhich is normal behaviour (same as dos/windows/linux)
16:12:04Torneso if rockbox thinks it's corrupt then there must be something else wrong as wel :)
16:12:20lilltigerand when i trasfered music to the player I did it in chunks of 2GB, and it's only the last 2GB of files that shows the corruption in rockbox (while the are fine both in OF and when connected to the usb in the OF)
16:12:35Torneyou're doing this over MSC?
16:12:58Tornei dunno, anyway. there's a million and one reasons for filesystems to get corrupted ;)
16:13:06lilltigerMSC is the UDP mode right, then yes
16:13:28 Join jgarvey [0] (
16:13:37notlisteningWill rockbox ever be writing to the MicroSD without me specifically asking it to? e200v2
16:13:42lilltigerand it only occures to the players internal memory, the microSD workd fine
16:13:53rasherlilltiger: You are writing to the disk with Rockbox.
16:14:05rasherCreating the database writes a few megabytes to the disk, for one
16:14:23rasherMight be enough to mess things up
16:14:52lilltigerrasher: the issue occures without an update to the database
16:15:08notlisteningahh humm got MicroSD corruption now not the internal disk :P
16:15:42notlisteninglilltiger, when did you last update rockbox?
16:16:19 Join toffe82 [0] (n=chatzill@
16:16:28lilltigernotlistening: sun 28th
16:17:26notlisteningahh a commit yesterday should solve your issues
16:18:21lilltigerat least i think it was :) ill try to update then
16:18:45notlisteningplus it was recommended to do a full format first
16:19:51notlisteningthis is leave the fs in a state where it will report curroption but running a disk check on it should sort that out then go for rockbox on it again and there should be not more issues
16:20:46notlisteningand you may want to give in internal disk a label/name again as i find that useful
16:21:35lilltigernotlistening: Should I reformat it from within the player or using the linux tools for it?
16:22:51lilltigerAlso should I update the bootloader as well?
16:24:00notlisteningPersonally i have had that for a few months now and there has not been an issue i know of but i am not the guru
16:24:36notlisteningI reformatted using the player, was not sure what it would do doing it from linux
16:27:02 Nick LambdaCalculus37 is now known as ZombieBillyMays (i=44a0430d@rockbox/staff/LambdaCalculus37)
16:27:20 Nick ZombieBillyMays is now known as LambdaCalculus37 (i=44a0430d@rockbox/staff/LambdaCalculus37)
16:29:24lilltigerrasher: when upgrading rockbox, to try to fix the corruption issue, should I update the bootloader/firmware as well
16:31:16rasherNo idea. I don't *think* it'd be necessary
16:33:14lilltigerrasher: so just the "make fullzip" stuff then?
16:35:17 Join {phoenix} [0] (
16:38:15lilltigerI just formated it using the OF format option, but it still reports two FAT's
16:39:02Torneplease listen to what you've been told :)
16:39:10Tornethere are (nearly) always two FATs
16:39:17Tornethey are supposed to be identical
16:39:21Torneif they are not then something is broken
16:40:50 Join sumit [0] (i=7aa7265c@gateway/web/freenode/x-4df2632542ad6721)
16:40:54lilltigerTorne: yes but why does fdisk only report it for the fuse drive, not any other USB drive I have or any other fat partition I create
16:41:24Tornebecause it only has to ask you which one to use when they are not identical!
16:41:32Torneif they're the same then amazingly it doesn't matter :)
16:41:36 Quit BryanJacobs (Remote closed the connection)
16:41:43Torneit has *detected an error* and is telling you about it
16:41:51Tornethe reason it doesn't tell you for other disks is the other disks are *not corrupt*
16:42:02sumithello all, i am a newbie here. Have been using rockbox for last 6 months on my ipod video. Wanted to contribute, can i help anyway? technically or non-technically
16:42:20lilltigerTorne: but why is it still corrupt when I just formated it
16:42:32Tornei would guess that whatever you formatted it with got it wrong :)
16:42:35lilltigerTorne: do i need to do something more then format it?
16:42:46Torneyou formatted it with the OF?
16:42:49Torneand then did nothing else?
16:43:13notlisteninglilltiger, the OF makes what are seen to be corruptions
16:43:14Tornethen the OF's format tool is either intentionally doing some obscure nonstandard thing that we don't know about (and will probably break on all dos/windows/linux machines regardless of rockbox)
16:43:19Torneor the OF's format tool is broken
16:43:49notlisteninglilltiger, what operating system you using
16:44:33notlisteningsudo fsck/dev/sansainternaldisk -p
16:45:19sumitcan anyone help me with this? "[20:14] freenode-connect [freenode@freenode/bot/connect] requested CTCP VERSION from sumit: "
16:45:23lilltigernotlistening: it's fsck that is reporting the the formating did something wrong
16:45:48Tornesumit: that's not on-topic and is also not a problem. freenode is checking what client you are using, this is normal.
16:46:21Tornelilltiger: yes, so let fsck fix it :)
16:46:21notlisteningyeah this is the OF and a problem is seems to cuse running fsck with the -p option will fix it forever more
16:47:02sumitok, Torne thanks. can you pls answer to my previous question above
16:47:16notlisteningand then you just need to worry whn you donot umount it ;) Then that's user error
16:47:37 Join spaztik [0] (n=none@
16:47:41Tornesumit: how you can help depends what skills you have or are interested in learning :)
16:47:44lilltigerTorne: just fixing it should be enought, or should I reformat it with the linux tools insted?
16:48:08sumiti can program in C, java. can act as tester. or anything else that i can do
16:48:21Tornesumit: report bugs, or issues with the manual or other documentation. help people out on the forums. Test the occasional official test builds that come up on the forums.
16:48:37Tornerun the current versoin instead of the stable release, if you want to test stuff, also ;)
16:48:47Tornethough standard caveats about maybe eating all your music apply
16:49:13Tornejust using the current version, updating frequently, and reporting any problems you find with it is a pretty good help ;)
16:49:52sumitTorne: ok, sure. i will update today to current version n use that.
16:50:01sumitthanks Torne !
16:50:03***Saving seen data "./dancer.seen"
16:50:12notlisteningsumit, if you speak another languages other than english you can help on translation
16:50:15Tornegood for you. just as long as you don't complain if something *does* go wrong and eat all your music ;)
16:50:20Tornekeep backups :)
16:50:53sumitI do have backups on my pc. unless it does something wrong with my ipod, i am ok
16:51:11sumitI know "Hindi" language. not sure that may help.
16:51:16Tornerockbox can't permanently damage the ipod, no
16:51:28Torneyou can always restore the disk
16:51:41lilltigerlucky ipods :p
16:51:51sumitallright, I am fine then. yes!
16:51:54Tornelilltiger: indeed. apple have a USB disk mode implemented in rom
16:52:05Tornelilltiger: so even if you erase the entire hard drive you can still boot that and restore it
16:52:06Jaykay"everything might just explode and there's noone but yourself to blame" (or similar :) )
16:52:28Tornesumit: there is a hindi translation
16:52:29LambdaCalculus37sumit: If you're looking to contribute, perhaps you're interested in helping improve the Hindi translation for Rockbox?">
16:52:30 Part spaztik
16:53:26sumitLambdaCalculus37: i will install the hindi one also on my ipod n find if there are any bugs/improvement scope
16:53:43 Join robin0800_ [0] (
16:53:49 Quit n1s (Read error: 110 (Connection timed out))
16:53:53lilltigerBtw if anyone got an 8GB fuze it would be intresting if they could get the latest OF and formate the drive and check it with fsck to see if it's just my fuze that is acting wierd or if it's the OF:
16:53:55sumitohh, that needs a lot of improvements/additions!
16:54:13Tornesumit: yup, so it'd be great if you could help update that
16:54:40sumitoh yes, my pleasure. will start work on that.
16:54:43 Join n1s [0] (n=n1s@rockbox/developer/n1s)
16:55:04sumitif I need any help, can i come and ask here? or there is some better place to ask for that
16:56:41evilnickHere is the right place to ask for help, or the forums.
16:56:54 Join kugel [0] (i=kugel@rockbox/developer/kugel)
16:56:59 Join BryanJacobs [0] (n=bryanjac@
16:57:14kugellilltiger: the two fat's are fine
16:57:40lilltigerkugel: i ment that they are differing so fsck asks about wich to use
16:57:51kugelyea, use the second
16:57:56Torne..the second?
16:58:17kugelTorne: the OF isn't doing proper things with the storage
16:58:31kugelit's all magic
16:58:31Tornei gathered :)
16:58:37sumitok, thanks :)
16:59:13lilltigerkugel: so fsck /dev/sansadev -p will fix it in the wrong way?
17:00:25kugelin fact, it doesn't even expose the first few 60K (the firmware sits there)
17:00:41kugelRockbox skips that, and USB doesn't see it
17:00:51 Join _lifeless [0] (n=lifeless@
17:01:27kugelfor the rest of the storage, there doesn't seem to be a proper FS. although creating one with windows or mkfs, as well as checking with chkdisk, works without problems
17:01:45Torneit sounds like their format code is just busted then
17:02:06kugellilltiger: read the most recent forum post before doing any changes. it appears fsck will always report errors for the folders created by the OF
17:02:12rashersumit: Unfortunately, I believe rendering of the hindi text (I'm not aware of the correct name) is not very good on Rockbox, but getting the translation updated might provide a better incentive to fixing the text rendering
17:02:16Tornewhich is not uncommon, there's a wide and exciting range of ways to get FAT wrong. :)
17:02:30kugelindeed :)
17:03:35lilltigerkugel: ahh ok, so the seemingly disk curruption i experienced both today and yesterday shouldent be related to that?
17:03:36 Nick zitune is now known as zitune[afk] (n=zitune@bearstech/zitune)
17:04:07kugelif the fsck reports match to what notlistening posted in the forum, then there should be no problem
17:05:18 Quit sumit (Ping timeout: 180 seconds)
17:05:23 Nick zitune[afk] is now known as zitune (n=zitune@bearstech/zitune)
17:08:45 Quit robin0800 (Read error: 110 (Connection timed out))
17:08:53 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
17:08:55 Join Rob2223 [0] (
17:08:57notlisteningas far as i can tell after fixing the FS issues created after an OF format everything seems fine
17:12:14lilltigernotlistening: but it's important to choose the seconfd FAT? i used -p this time, should i reformat?
17:12:31kugeljust don't reformat if you don't have problems
17:12:40kugelI didn't even care about fixing those OF folders
17:13:52lilltigerkugel: I have the data corruption issues, and did reformat (without fixing it) and the data corruption reappered
17:14:07kugelwhat sort of data corruption?
17:14:43lilltigerkugel: the file browser in Rockbox is just squares and other strange stuff
17:15:01lilltigerbut only for a few albums
17:15:13lilltigerwhile in OF it all sseems fine
17:15:34kugelare they having weird characters?
17:15:34 Quit __lifeless (Read error: 110 (Connection timed out))
17:15:56kugelnot all fonts in rockbox can display all characters
17:16:19kugelsquares are shown then
17:16:44lilltigerkugel: yes, to the point of unreadeble not a single character is of an human language
17:16:48kugelmake sure to use unifont to rule out encoding issues
17:16:52 Quit fyre^OS (Read error: 104 (Connection reset by peer))
17:17:05lilltigeralthought the filename is in utf8 and all english characters
17:17:10 Join PaulJam [0] (i=Paule@
17:17:11 Join fyrestorm [0] (
17:17:25 Quit fyrestorm (Read error: 104 (Connection reset by peer))
17:17:38lilltigerand the font shows the other titales just fines eaven if they contain swedish characters
17:17:38 Join fyrestorm [0] (
17:17:46rasherHuh, has the iaudio m3 not been released?
17:19:01 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
17:19:06kugellilltiger: I regulary see squares for umlauts depending on the font
17:20:15lilltigerkugel: yes but there isent a single roman character it's all squares and strange symbols, if it was the character set some characters would be readable
17:20:44kugelhow do you format?
17:20:53lilltigerkugel: and it only appered on the albums i added last, while thire foldername was correct non of the filenames was
17:21:09kugelI always use mkfs.vfat w/o problems
17:21:39kugeldid they appear instantly in the first boot of rockbox after putting the albums onto it?
17:22:12lilltigerkugel: did you use fdisk to create an partition first, cos mkfs says it cant format it cos there is no partition after an OF format
17:22:16kugelRockbox hardly writes to the disc (not at all actually until you changed a setting or did filemanagement which needs write access)
17:22:37kugellilltiger: no, I never did that
17:22:50kugeland it definitely worked for me after a OF format
17:23:10lilltigerkugel: first boot was fine, up till i aded 6GB of music it was fine, just the last 2GB that causes issues
17:23:13kugelbut there isn't really a readable partition anyway (parted fails too)
17:23:15Torneyou just need -I to force mkdosfs to let you format an unpartitioned volume
17:23:54kugeldid you try to just "mkfs.vfat -F32 -I /dev/sdX"? (-I overrides problems IIRC)
17:23:55lilltigerok, was unsure if i dared to use the -I D
17:24:09Torne-I just turns off that warning
17:24:22kugelnah it's the way to go
17:24:34kugelyou can't format without I, due to the strange OF
17:25:27lilltigermmm as the OF dosent make an proper partition
17:25:30 Quit Rob2222 (Read error: 110 (Connection timed out))
17:29:09AlexPrasher: I think manual was blocking it (because of the problems with it mainly using the remote and the manual up till very recently not handling that).
17:29:24AlexPrasher: However we decided at devcon that the manual shouldn't be a blocker
17:29:49AlexPrasher: So not only should hopefully we have a manual in not too long, but it can be released anyway AFAIK :)
17:30:01rasherAlexP: ACK
17:34:26 Quit robin0800_ (Read error: 110 (Connection timed out))
17:34:40 Quit flydutch ("/* empty */")
17:49:13 Quit HellDragon (Connection timed out)
18:11:02 Nick zitune is now known as zitune[afk] (n=zitune@bearstech/zitune)
18:13:04 Join mcuelenaere [0] (
18:17:55 Join atrus [0] (
18:19:10 Quit n1s ("Lmnar")
18:21:04mcuelenaerekugel: I only exposed BUTTON_TOUCHSCREEN when HAVE_TOUCHSCREEN is defined because that button is totally unnecessary on non-touchscreen targets (+ there should be an other way to detect what target you're running)
18:21:34kugeldoes lua have exceptions?
18:25:25Tornelilltiger: if you don't know the answer please don't jus tlink to the first google result
18:25:31Torneunless you know for sure it's relevant :)
18:25:55mcuelenaerekugel: yes
18:25:57Tornenow i just pasted your url
18:26:19lilltigerTorne: I read throught it first, and I thought it was what he asked about
18:26:59mcuelenaereAFAIK exceptions in Lua go something like this: if(pcall(func) == 1) { −− OK } else { −− Exception }
18:26:59Tornelilltiger: the page is talking about a nonstandard patch to lua :)
18:27:00*JdGordon doesnt like exceptions.... a single error code in the return value for all errors ftw!
18:27:52mcuelenaerealthough /me doesn't really like the way how exceptions work in Lua..
18:28:13Tornethey're the minimum amount of functionality you cna possibly have and it still be an exception, yah :)
18:28:39Torneit has the big advantage that it's trivially implementable in terms of longjmp() :)
18:29:27lilltigerI implemented Lua into an game a few years back, after that I started to disslike anything that has with Lua to do, imo phyton is superior
18:29:43kugelmcuelenaere: better than nothing, if we can get lua not be target dependant this way (not that everything works on all targets, but at least have some basic handling when something doesn't work)
18:29:53Tornepython is also much, much bigger :)
18:29:57mcuelenaerekugel: why use exceptions for that?
18:30:08mcuelenaerelilltiger: try implementing Python for Rockbox then ;)
18:30:13lilltigerTorne: indeed
18:30:30Tornemcuelenaere: it owuldn't be very difficult, really :)
18:30:35Tornemcuelenaere: but it'd have to steal the audio buffer ;)
18:30:45Torne(and you'd need to provide it with a malloc())
18:30:47kugelit seems logical to me, but if you have a better idea
18:30:48mcuelenaereTorne: Lua does that now too (if it runs out of memory)
18:31:06kugeland it has a malloc() :)
18:31:10Torneyah, but my minimal build of python on ARM needs several kb of heap just to make it to the interpreter prompt with no modules loaded
18:31:15Torneseveral hundred kb
18:31:34mcuelenaerethere are targets with >=16MB RAM ;)
18:31:47Torneyah, i just mean it would *always* need to steal the buffer :)
18:32:28mcuelenaerenot if you upgrade the plugin buffer :)
18:32:34 Quit PaulJam (Nick collision from services.)
18:32:40 Join PaulJam_ [0] (
18:33:51Torneand you need about 1.1MB just for .text/.data/.bss of the interpreter
18:34:18Torneand this is with no builtins at all even the ones that are normally supposed to be mandatory (zipimport, warnings, etc)
18:34:42mcuelenaerehmm I suppose plugins need some kind of dynamically linking infrastructure?
18:34:54Tornenot really
18:35:11Tornebuiltin modules and extensoin modules are the same thing really :)
18:35:19Torneyou cna compile them into the interpreter instead of having them as .so if you want
18:35:37lilltigerstorage space isent much of an issue i guess, but ram needs is probly to high for phyton
18:35:56Torneso, yah. there's no particular reason why not other than the fact that it would be memory hungry and even slower than lua :)
18:36:37Tornesince nobody has written any cool and exciting lua scripts for rockbox yet i think it's a bit premature to go for an even vaster embedded language
18:36:50Torneif people use lua for interesting things and find it limiting then maybe it's worth it ;)
18:37:21mcuelenaereI think exporting the Rockbox API is probably the hardest part currently limiting Lua
18:37:34 Join faemir [0] (n=faemir@
18:37:45lilltigerif players would have wifi there is alot of cool stuff one could do but i guess wifi onn mp3 players is about 5yers or more away
18:37:46Torneit would be a very similar amount of work to do so in python :)
18:38:03AlexPlilltiger: The Sansa Connect and the Zune both have wifi
18:38:22GrahackI'm writing a set of lua libs to be able to code things that will be compatible with Lua on Rockbox and Lua on a Nintendo DS. It will be very basic but useful.
18:38:23lilltigerAlexP: neat
18:41:31 Quit petur ("work->sports->beer :)")
18:41:36 Join bertrik [0] (
18:45:44lilltigerTrasfear now done, no corruption detected
18:46:21 Join Hillshum [0] (n=chatzill@unaffiliated/hillshum)
18:48:28 Quit aaron424 ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060308]")
18:50:07***Saving seen data "./dancer.seen"
18:53:13 Join perrikwp [0] (
18:54:40HillshumOn FS #10290, should I upload a patch with the changed langfiles?
18:55:12 Quit fyrestorm (Read error: 104 (Connection reset by peer))
18:55:13 Join fyre^OS [0] (
18:55:48bertrikHillshum, no the langtool command lines should be enough
18:56:29HillshumOk, is it ready to commit then?
18:56:34bertrikthank you for posting them, by the way
18:56:52bertrikI guess so, but I wonder if there may be other strings that we overlooked
18:57:24rasherThe fuze => fuze* is not enough
18:57:37bertrikother than LANG_CONFIRM_WITH_BUTTON and LANG_TIME_SET_BUTTON I mean, like strings for buttons to cancel things
18:57:47rasherthat won't change "e200,fuze,c200" into "e200,fuze*,c200" - only "fuze" into "fuze*"
18:58:02rasherjust do sed -i 's/fuze/fuze*/' *lang
18:58:25rasherIT needs to match on the *entire* target string
18:58:52Hillshumshould that be done on the clip too?
18:59:07rasherWell ifyou want clip to be clip*, then yes
19:03:26 Quit Jaykay (Read error: 110 (Connection timed out))
19:10:47 Join Lear [0] (i=chatzill@rockbox/developer/lear)
19:12:48Hillshumrasher: sed gives me
19:14:32rasherHillshum: then you didn't do what I told you
19:15:26Hillshumor it might be I'm on a mac?
19:16:00rasherI guess all bets are off then
19:16:29HillshumI uploaded a new set of commands
19:17:50rasherOh hah, that's not good
19:18:22rashersed -i 's/clip/clip*/' *lang messes up strings such as "AGC clip time"
19:18:40rasherhang on while I come up with something better
19:19:36rasherAh, there's no clip in english.lang right now
19:19:40rasherso just skip that line
19:19:49Mikachuyou would probably need something on the order of s/([^:]*)clip/\1clip*/
19:20:13Mikachu^ too :)
19:20:15rasherMikachu: Yeah, but not needed. The target "clip" is not present anywhere in the translations :)
19:20:54rasherHillshum: Just remove that line (sed -i 's/clip/clip*/' *lang)
19:21:37 Join Cory [0] (
19:21:44 Nick Cory is now known as Cory` (
19:22:02LearIsn't there a good way to tell gcc to never inline a function?
19:22:33agaffney-O0? :P
19:22:46LearJust one specific function, please... :)
19:22:47MikachuLear: /noinline
19:22:54 Join JdGordon| [0] (i=209958e6@gateway/web/freenode/x-c59daaa1d6018845)
19:22:57agaffneyLear: ah, you mean something in the code
19:23:49LearMikachu: Ah, yes, now lets see if it is available in 3.4.x too...
19:24:04Mikachuit should be, since i googled up the 3.4 docs and changed the url to 4.4 ;)
19:24:16LearAnd I changed it back and found it. :)
19:26:21*Hillshum posts that change
19:26:54 Quit mcuelenaere ("Dinner time!")
19:31:27 Join FlynDice [0] (n=FlynDice@
19:32:36 Quit JdGordon| (Ping timeout: 180 seconds)
19:34:21 Join robin0800_ [0] (
19:36:48 Nick zitune[afk] is now known as zitune (n=zitune@bearstech/zitune)
19:39:07CIA-69New commit by 03learman (r21576): Small change to significantly reduce stack usage during database scanning. This should fix FS #10396.
19:39:41HillshumDo all references to buttons in langs end in "_BUTTON"?
19:39:45 Join JdGordon| [0] (n=Miranda@rockbox/developer/JdGordon)
19:40:18 Quit bambu ("When two people dream the same dream, it ceases to be an illusion. KVIrc 3.4.2 Shiny")
19:40:49 Join flydutch [0] (
19:46:05MikachuLear: you could just have put a brace pair around the code i think
19:46:25 Quit PaulJam_ (Nick collision from services.)
19:46:30 Join PaulJam [0] (
19:46:35Mikachu(and put the variable declaration inside it)
19:48:44LearMikachu: I'm pretty sure the stack allocation is done for the whole function...
19:49:20Mikachuwell, calling another function still increases the stack usage
19:49:32Mikachui assumed the point of the exercise was to have it be reduced after that function call?
19:50:31LearOnly temporarily, for that function call. check_dir is recursive, so having a new newpath allocation on the stack for each level is pretty significant.
19:54:18Mikachui mean something like this,
19:54:20 Join Jaykay [0] (
19:54:29 Quit PaulJam (Nick collision from services.)
19:54:35 Join PaulJam_ [0] (
19:54:57BagderMikachu: gcc typically allocate stack space once anyway
19:55:00Bagderfor the entire function
19:55:07Mikachuhm, that's pretty silly
19:56:25CIA-69New commit by 03FlynDice (r21577): FS #10344 - AMSSansa Dynamically adjust core voltage to extend playtime. ...
19:56:53 Join Blue_Dude [0] (
19:56:54 Quit PaulJam_ (Nick collision from services.)
19:57:02 Join PaulJam__ [0] (
19:57:19 Quit Jaykay (Client Quit)
19:57:25Blue_DudeHey, which targets use the Coldfire CPU?
19:57:54HillshumH100, H300, and the iAudios (i think)
19:58:40HillshumWhat key is used on the fuze to cancel database init?
19:59:07Blue_DudeIt seems the asm multiplication routine only works for small bit shifts with Coldfire. Is it worth trying to keep the asm with such a limitation?
20:00:51HillshumLANG_BUILDING_DATABASE might need work
20:01:22 Quit PaulJam__ (Nick collision from services.)
20:01:26 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
20:01:28 Join PaulJam_ [0] (
20:03:03bertrikI would have liked to stay FS #10344 out of SVN while there are still issues with the ams sansas
20:03:55FlynDicebertrik: Sorry bout that funman wanted it committed soon and I didn't know you had concerns
20:04:22Hillshumit hasn't been committed yet has it?
20:04:32 Join Thundercloud [0] (
20:04:33FlynDice10 mins ago
20:04:57*Hillshum missed that
20:05:14 Quit BryanJacobs ("Java user signed off")
20:05:49kugelLear: since when do we have database.unignore again?
20:06:18MikachuDate: Mon Feb 4 21:20:04 2008 +0000
20:06:19Mikachu add support for database.unignore files (adds dirs to the database which would be skipped~
20:06:38AlexPthat is indeed ages
20:07:03kugelyea, but wasn't that reverted?
20:07:20Mikachuit's the only commit in history that has "database.unignore" in the commit message at least
20:07:43kugelI'm farily sure it was..
20:07:59Hillshumsearch for whatever revision number that wasa
20:08:20Mikachuif you want to know if it was reverted, it's much easier to see if it is in svn now ;)
20:08:23Mikachuwhich it is
20:08:37kugelmaybe the revert wasn't complete
20:08:41Mikachuapps/tagcache.c:4063: /* check for a database.unignore file */
20:08:41Mikachuapps/tagcache.c:4064: snprintf(newpath, MAX_PATH, "%s/database.unignore", dirname);
20:08:49kugelI saw that, hence I asked ;;
20:09:08Mikachuif it's there, it wasn't reverted :P
20:09:44kugelread what I write
20:09:52pixelmaif I remember correctly it wasn't reverted
20:10:25AlexPI don't think it was - there was argument then it was left
20:10:46kugelsomehow I was sure it was reverted
20:10:50kugelI still think it should
20:10:52Learkugel: Again? It hasn't changed since r16213, which was just a couple of months after database.ignore was added...
20:10:52Mikachucomparing the code as it is now with the commit that introduced it, it is still there
20:11:41Mikachuwhy don't you want it, to avoid the recursion into the dir?
20:12:02JdGordon|amiconn: pixelma: hows would you feel about the buttonbar being removed?
20:12:38kugelMikachu: yes, because it also speeds up tagcache noticeably
20:12:49Mikachuyou could have a database.norecurse ?
20:13:03 Quit PaulJam_ (Nick collision from services.)
20:13:10 Join PaulJam [0] (
20:13:53pixelmathe Ondio doesn't have a button bar so I don't have a strong opinion, but I thought you saw some use for it on touch screen targets? (just curious)
20:15:34kugelMikachu: I think that should be possible
20:15:45 Join DarkDefender [0] (
20:15:59kugelLear: I just thought it was reverted, but it seems I was wrong :/
20:17:29 Quit PaulJam (Nick collision from services.)
20:17:36 Join PaulJam_ [0] (
20:17:55 Quit _lifeless (Read error: 110 (Connection timed out))
20:18:02kugeldoes anyone see a good name?
20:18:23Mikachukugel: this should do it
20:18:37Mikachuexcept you can correct my typo in the comment ;)
20:19:01kugelnorecurse isn't a good name alone, IMO
20:19:10Mikachuyou can change the name too
20:19:18kugelyea, I don't know which though
20:19:30Mikachuit's basically a revert of the commit that introduced .unignore, but without removing the + lines :)
20:19:52JdGordon|pixelma: ARG yeah I forgot about that... touchscreen targets could beenfit from it, but it its current form its more annoying and wouldnt be missed
20:19:53kugelalso, you're not on the latest reversion :)
20:20:00Mikachui know
20:21:00kugeldatabase.ignore.recursive? database.force.ignore?
20:21:13kugelor database.recursive.ignore
20:21:23 Join freqmod [0] (
20:21:54JdGordon|ok all these databse. files are retarted... cant we use lua or something do script the whole proccess properly?
20:22:10kugelare you serious?
20:22:14gevaertsdo we really need the complexity? I think that if database.unignore support is thought to be too expensive, we should just drop it and only keep database.ignore
20:22:33kugelI'd be happy with it too
20:22:41rasherPeople just need to be more careful about where they place .ignore files (and use more of them)
20:23:09JdGordon|imo the whole system is too convoluted
20:23:09kugelmost people seemed to be in favor for reverting when I read the logs
20:23:18gevaertsexactly. Complex setups may be a bit harder, but they can still be done
20:24:23*gevaerts proposes a poll on the forums :)
20:24:32AlexPI don't use the database so don't really care either way, but without unignore there would be no way to ignore a top level but not a sub dir right?
20:24:46kugelput ignore into the subdirs
20:24:59TorneAlexP: not for files directly in te top level, no
20:25:09AlexPkugel: That is the opposite of what I said
20:25:28kugeloh, I missread
20:25:39rasherAlexP: I find it unlikely that you'd seriously need that
20:25:39AlexPTorne: Or in any directory that is a higher level than a lower one that you want the files in
20:25:51kugelyes that's true, but I can't imagine anyone trying to achieve that
20:26:04gevaertsAlexP: true. I'd consider that a messy layout though
20:26:08AlexPrasher: As I say, I don't care/don't know - just trying to think of why they are needed
20:26:17*Torne can't think of a sensible layout that needs unignore at all, so hey :)
20:26:36*gevaerts can think of a reasonably fast way to support this :)
20:26:45AlexPgevaerts: Sure, but just because we consider something messy isn't the be all and end all
20:26:57 Nick zitune is now known as zitune[afk] (n=zitune@bearstech/zitune)
20:27:58gevaertsHow about a .rockbox/database.roots file, and database.ignore? If database.roots is not there, scans start from the root and ignore anything that has a database.ignore in it. If it is there, scans start from each directory listed there, with the same usual database.ignore rules
20:28:40gevaertsThe only thing I'm not sure about is how to handle "roots" that are within eachother
20:28:50AlexPI wouldn't object to just dumping .unignore, I'm just playing devil's advocate here :)
20:29:12kugelgevaerts: I thought you wanted to make it simpler
20:29:19TorneAlexP: i think the people who are likely to care about exactly what goes in the datbaase or not have probably arranged their files sensibly ;)
20:29:31Blue_DudeAll: are there any targets with CPU's other than Coldfire and ASM? Where can I find out which is which?
20:29:31AlexPBut then I wouldn't object to dumping the database, so take what I say with a pinch of salt
20:29:46AlexPTorne: People that have organised their files sensibly don't need the database :)
20:29:50gevaertskugel: I think that's reasonably simple :)
20:29:57gevaertsBlue_Dude: arm, coldfire, sh and mips
20:30:01TorneAlexP: now that's just flamebait :)
20:30:31 Join _lifeless [0] (n=lifeless@
20:30:44Blue_Dudearm, not asm. sorry. I have asm on the brain.
20:30:55*kugel thinks we should talk to markun again before doing anything
20:31:26 Join Llorean [0] (
20:32:17Blue_DudeI'm still trying to consolidate math functions. Any objections to implementing functions as macros, even if they're not capitalized? Trying to be consistent.
20:32:18gevaertsAlexP: aren't there things you can do properly with the database that (currently) aren't that clean with the file browser? I'm thinking about random folder advance and similar things
20:32:58Torne"finding songs without remembering who performed them" is kinda the easy one :)
20:33:05AlexPgevaerts: Probably true - I was being flippant. Personally I just tend to play albums (or create playlists) which = play folders
20:33:33gevaertsTorne: that's what the wikipedia plugin is for :)
20:33:33Tornei use the db so i can pick random albums without having to hit forwards a lot :)
20:33:55Tornealternate between hitting <random> and back until i find one i want :)
20:34:53 Quit FlynDice (No route to host)
20:36:18 Quit bertrik (Read error: 54 (Connection reset by peer))
20:36:30 Join bertrik [0] (
20:36:54kugelbertrik: hey, have you looked at the freeze when changing the region on AMS?
20:37:43bertrikyes, I'm pretty sure it tries to retune when changing the region, this causes it to hang in the wait-for-tune loop when the fm chip has not powered up yet
20:39:27kugelI also noticed it doesn't play music when resuming after reboot
20:39:40kugelyou have to change the station back and forth once
20:40:07 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
20:40:16kugelplus, I wondered my tuner.c contains almost exclusively target dependant code. all the structs could be in the drivers
20:40:34JdGordon|gevaerts: random folder advance works better in the file browser playlists than with the db
20:41:09gevaertsJdGordon|: I think it probably doesn't work well in either
20:41:09HillshumWhere would I look to find what button is used on the Fuze for reverting from within the time set screen?
20:41:23 Join n00b81 [0] (n=taylor@unaffiliated/n00b81)
20:41:40gevaertsJdGordon|: random folder advance is not nice at all for battery life
20:41:55 Part n00b81 ("Leaving")
20:41:57JdGordon|how so?
20:42:00 Join n00b81 [0] (n=taylor@unaffiliated/n00b81)
20:42:08PaulJam_gevaerts: how does it affect battery life significantly?
20:42:15 Part n00b81 ("Leaving")
20:42:27bertrikkugel, yes, I'd like to have non-target dependent fm regions too
20:43:28 Join mirak [0] (
20:43:54*JdGordon| has a partial patch in the tracker somewhere (I hope) to move some of the tuner stuff out of apps into firmware if someone wants to look at that
20:44:03JdGordon|if you're playing with the radio anyway :)
20:44:04bertriklike having one definition for the fm regions and let the fmchip drivers just ignore the fm region setting they don't support
20:44:11gevaertsJdGordon|, PaulJam_: (correct me if I'm totally wrong) : it decides where to jump to after finishing the current playlist, so you often don't get a full audio buffer
20:44:30JdGordon|and the db is different?
20:44:50gevaertsno idea. "I think it probably doesn't work well in either" :)
20:44:55PaulJam_gevaerts: but you have the same when you manually select a new album after the current one finishes, no?
20:45:29AlexPDid people playing with radio notice the bug report that changing volume in the radio screen doesn't work anymore?
20:45:44*JdGordon| blames kugel for that one :)
20:45:45gevaertsPaulJam_: sure, but I already know that I don't care what it picks next. A proper album shuffle would be a lot better
20:45:59bertrikand have some mechanism to more easily add/edit fm regions
20:46:12kugelAlexP: It works on my fuze. I didn't test 3.3 on it though
20:46:24kugelI can't really imagine what would break the controls
20:46:29AlexPkugel: I haven't tried it on anything - will do in a bit
20:46:42JdGordon|gevaerts: yeah but thats a general playlist problem... it should try getting the next playlist (whatever it is) on the last low buffer...
20:47:13bertrikfm radio volume control on my clip with r21464 works fine
20:47:20gevaertsJdGordon|: somehow I think that doing this at playlist generation time will work better
20:47:35gevaertsI'm not sure though
20:47:40JdGordon|yes, but that doesnt happen ever untill the playlist finishes
20:47:48kugelmy e200 doesn't have a radio so I can't test
20:48:01JdGordon|you mean add a heap of folders onto the playlist at the start of creation?
20:48:17kugelJdGordon|: I rather think it was gevaerts change to make ascii sort default, he was so unsure if he managed it :)
20:48:51LloreanJdGordon|, gevaerts: Isn't the m3u format supposed to support folders
20:48:57gevaertsJdGordon|: yes. The same way that "insert shuffled" works now
20:49:05LloreanLike if a line appears in it that's just a folder name, it should play the whole content of the folder.
20:49:11JdGordon|Llorean: I dunno, maybe
20:49:16JdGordon|but still how many would you add?
20:49:21JdGordon|5? 10?
20:49:24JdGordon|every folder?
20:49:38LloreanSo we could have "Insert Folders Shuffled" and "Insert Folders" as options, then if Shuffle is Off you get what you're looking for
20:49:42LloreanFOlders shuffled then played in order.
20:49:49LloreanIf we supported that aspect of M3U that is
20:49:56gevaertsWhat I'd like is just all my music, shuffled, but with albums together in the right order
20:50:09***Saving seen data "./dancer.seen"
20:50:18LloreanThere's no reason not to insert every folder within the folder the user is "Inserting" - random folder advance more or less doesn't care how many folders are there either, it just keeps going.
20:50:31Lloreangevaerts: Which is basically what I just described, if your albums are individual folders.
20:50:52LloreanYou can have a playlist of albums, instead of a playlist of songs, and shuffle it.
20:51:06JdGordon|if you do that then the next dir doesnt work
20:51:21JdGordon|also you would have to recreate the entire playlist when you add/remove files
20:51:28gevaertsCurrently random folder advance is the way to do that, but (a) it's not good for battery life, and (b) it somewhat interferes with bootmark on stop and recent bookmarks (which I have for spoken material), in that you get a bookmark per album
20:51:40JdGordon|but I agree that the playlist would be better created while music is still playing
20:51:52bertrikkugel, I'm also seeing the problem of no fmradio after boot on my clip, it seems the frequency is set, but the radio hasn't tuned to it
20:52:21 Quit Hillshum (Read error: 60 (Operation timed out))
20:52:23PaulJam_Llorean: but the current random folder method has the advantage that you can exclude folders and also give certain folders a higher chance of being played, this wouldn't be possible with just inserting your whole music directory.
20:52:30kugelmaybe si4700_init() isn't doing correctly?
20:52:37gevaertsJdGordon|: maybe a tsr plugin could do this. Watch the playlist and the buffer, and if the playlist gets too short add bits to it
20:52:57JdGordon|we dont really do tsr plugins...
20:53:00 Join Hillshum [0] (n=chatzill@unaffiliated/hillshum)
20:53:04gevaertsnot many, no
20:53:11JdGordon|excalty 2 i think?
20:53:12Lloreangevaerts: TSR plugins conflict with the playlist viewer, I think.
20:53:17JdGordon|but yeah. that might actually work
20:53:28*gevaerts wants to do MTP as a tsr plugin written in lua :)
20:53:35LloreanPaulJam_: We don't really give weighted playlisting for anything else either, I'm not sure that's something that "must" be preserved. It's an anomaly.
20:53:48JdGordon|its a feature :)
20:54:18JdGordon|gevaerts: wouldnt that actually end up being worse on the battery life? seen as its an additional thread running?
20:54:21gevaertsJdGordon|: or just a plugin that takes the random folder advance config file and dumps a huge playlist...
20:54:41JdGordon|the currne tplugin could be made to do that already
20:55:03gevaertsindeed. I think that's actually the best plan in the short to medium term
20:55:09LloreanThat's probably the simplest/cleanest way
20:55:29PaulJam_gevaerts: i assume you wouldn't be able to skip an album that way?
20:55:30JdGordon|does that mean we have a volanteer to do the work? :)
20:55:31LloreanCouldn't you then take the random folder advance code out of core, and clean up the playback engine a little bit, once that happens?
20:55:53JdGordon|not worth it
20:55:58gevaertsJdGordon|: on huge memory targets with low bitrate files and reasonably short albums, you currently spin up twice as much. I doubt if an extra thread can use that much battery :)
20:56:00*JdGordon| would veto that anyway
20:56:14LloreanJdGordon|: I thought the random folder code caused problems basically any time someone worked on playback.
20:56:43JdGordon|I'm not sure where you got that idea from.... that specific feature is tiny compared to the rest of playback.c
20:56:52gevaertsJdGordon|: you mean you want me to do plugin/UI work?
20:56:52LloreanNot size-wise. Complexity wise.
20:57:13LloreanI just remember it a lot of complaints about it being a dirty hack, and similar.
20:57:38JdGordon|the whole playback<->playlist bit is nasty... removing that wont make it better by enough
20:57:57LloreanBy enough for what?
20:58:03JdGordon|to make it worth it...
20:58:12LloreanWorth what? I'm confused.
20:58:31gevaertsno, but maybe (NB: I've never looked at it) removing it would make the whole thing just a bit less daunting to clean up?
20:59:00LloreanPeople have talked about getting rid of random folder advance anyway. If there's a replacement that covers most of it, and offers better battery life while it works, it seems like it makes arguments for not removing it a lot smaller.
20:59:42JdGordon|im giong by memory here.. and obviously cant actually look at the code now to be sure, but the whole random folder bit is a hack ontop of the current regular next folder hack which is a hack ontop of the current end-of-playlist handling
20:59:48 Join saratoga_ [0] (i=9803c6dd@rockbox/developer/saratoga)
20:59:50JdGordon|so the actual issue is 4 or 5 levels up
20:59:52LloreanIt *could* cover all of it if the "Folder Skip" key combination within a playlist skipped to the next track in a different folder as well.
21:00:24LloreanThen we could get rid of the "next folder" hack by telling people to just use playlists. =p
21:00:36JdGordon|hell no
21:00:49LloreanWell that was an informative statement as to why this is a bad idea...
21:01:02gevaertsJdGordon|: that also means that it's one of the blockers to clean up the lower hacks
21:01:40saratoga_from the manual: "Some old “floor 0” files may crash Rockbox."
21:01:51JdGordon|I disagree.... I agree that its a hack, but the problem is the playback playlist interaction... there is no point cleaning this up untill that interaction is fixed completly
21:01:53saratoga_really? I thought any attempt to malloc too much resulted in an error and track change
21:02:06LloreanJdGordon|: Why the "hell no" though?
21:02:22 Quit Thundercloud (Remote closed the connection)
21:02:23JdGordon|thats my way of saying I strongly object to removing that feature
21:02:36LloreanIt's not removed.
21:02:48JdGordon|but yes, allowing that action to work in .m3u playlists would be nice
21:02:49 Quit lilltiger (Read error: 54 (Connection reset by peer))
21:02:51LloreanYou just access it differently.
21:03:05LloreanIn fact, from the WPS it would feel exactly identical.
21:03:08 Join lilltiger [0] (n=lilltige@
21:03:13LloreanIt would just entail slightly different behaviour starting playback.
21:03:23JdGordon|and a buigger delay on startup
21:03:27JdGordon|music startup that is
21:03:37*gevaerts thinks that we should take all these hacks out. JdGordon| obviously feels strong enough about them to implement them properly then ;)
21:03:42Learsaratoga_: Tremor doesn't really check for null returns from malloc. Should be fixed by the longjump hack I added though. :)
21:04:05LloreanJdGordon|: A couple seconds, and only if you don't already have a playlist in place.
21:04:07JdGordon|gevaerts: you seem to forget that I have all these grandious plans for like the entire code base, but not enough time to do any of them :p
21:04:08saratoga_we should just kill floor0 support entirely
21:04:35Lloreansaratoga_: I thought we didn't have floor0 support. I read that line as "old files that use floor0"
21:04:41 Join Jaykay [0] (
21:04:43LloreanRather than "some files that use floor0, but not all"
21:04:56saratoga_we have floor0 support, it just uses a lot of memory and can potientially exceed the available codec buffer
21:05:15saratoga_its basically non-existent in the wild though
21:05:28LloreanSo fail graciously with a clear message, rather than it being unknown whether it'll work until tried?
21:05:43saratoga_Lear: did you see my patch in FS #9866?
21:05:49saratoga_i'd like to commit that
21:06:01JdGordon|gevaerts: its been discussed a few times... I tihnk everyone agrees that playback should have nothing to do with playlist at all.. it should only request a track to buffer, and let the playlist know which has just finished
21:06:32JdGordon|What I'd love to happen is that the playlist gets told the buffer is low, and the current playlist has been buffered, so go and build a new one
21:06:40saratoga_in my testing it fixed a nasty crash in vorbis and saved us some codec memory as well, but i'm not 100% sure its safe since I don't understand vorbis's bitstream
21:06:43JdGordon|I guess keeping the current tracks in the playlist as "queued"
21:07:07 Join __lifeless [0] (n=lifeless@
21:07:41Learsaratoga_: Don't you need to seek past the comments for that to work?
21:08:37 Join kachna|lappy [0] (
21:09:23gevaertsJdGordon|: that could work, with some care to make sure that the playlist code doesn't spin up the disk too often
21:09:33JdGordon|if DCW has a codeing session and enough interest, I'd love for this to be worked on then
21:10:07JdGordon|well, the playlist should never spin up the disk... it would be notified when to get more tracks
21:10:16*gevaerts wants a DCsomewhere with all playback specialists in one room :)
21:11:10Learsaratoga_: Hm, maybe it skips to the next packet automatically. Never really learned what those ogg functions did... :)
21:11:42 Join Xqtftqx [0] (
21:11:48XqtftqxWhats up people?
21:11:49PaulJam_Llorean: well, my personal usage pattern for the random folder feature is to start playback on a specific album i want to hear and have random folder enabled so it continues playback after the current album has finished. so for me it is important at which album the random folder playback starts. i imagine with the playlist idea it would start with a random folder.
21:12:06JdGordon|gevaerts: that number is very small... I'd settle for anyone who has a blame line in playback.c ro buffering.c or really anyone who is masochistic enough :)
21:12:16XqtftqxHey who said they porting the sim to lua
21:12:17LloreanPaulJam_: Why? You can just browse to the folder in the playlist rather than the folder tree.
21:12:19Xqtftqxfor a ds
21:12:27*JdGordon| 100% agrees with PaulJam_
21:12:47JdGordon|Xqtftqx: im pretty sure noone said that
21:12:51saratoga_Lear: I tried a bunch of files with various comments and they all played with it
21:13:07saratoga_my assumption is that theres some ogg packet just for comments and its automatically skipped by the parser anyway
21:13:09XqtftqxJdGordon|: I just looked at the log for a sec, and somebody said they were
21:13:09gevaertsLlorean: it could be a *long* playlisy
21:13:10PaulJam_Llorean: do you know how long it takes to scroll through a 3000 entries playlist?
21:13:21saratoga_but i don't use vorbis much so i wasn't 100% sure it was safe
21:13:24LloreanPaulJam_: Not very long at all for me. i don't know how long it takes you.
21:13:36LloreanPaulJam_: You have a page up and page down button, since you're on a player that can folder skip.
21:13:37AlexPXqtftqx: Link?
21:13:57AlexPSpecific time?
21:14:03XqtftqxGrahack: Can i pm you
21:14:11AlexPP.S. You can link to the time...
21:14:15XqtftqxAlexP: 18:38
21:14:23PaulJam_Llorean: for paged scrolling i need both hands, and sometimes i have only one hand free.
21:15:02AlexPXqtftqx: I don't see anyone there even suggesting porting the sim to Lua
21:15:09GrahackXqtftqx: go on
21:15:24 Join Hillshum_ [0] (n=chatzill@unaffiliated/hillshum)
21:15:35JdGordon|Llorean: its also this... if you are browsing the tree and come up to a folder you want to listen its 1 press to get it going and then leave it for a few hours if you want to... your way requireds starting the playlist (probably with a long select) followed by searchign the (potentially very long ) list
21:15:37LloreanPaulJam_: This is going to be one of those "it has to be exactly the same as it was before, or I won't accept it" situations, isn't it?
21:16:07JdGordon|yes, because the issue here isnt the UI.... its the backend
21:16:28LloreanJdGordon|: Yeah, but the UI is being used as an excuse to not replace the backend.
21:16:35LloreanSo the issue is the UI.
21:17:11gevaertsLlorean: I think that's a valid use case to keep RFA in core. That doesn't mean it needs to be the same code of course
21:17:12 Join tessarakt [0] (
21:17:16LloreanIf you've got random folder inserting playlist generation code already in place, then you could make th existing UI create the same effect anyway
21:17:39LloreanThe problem is switching it from "playback effects" to "playlist generation effects"
21:17:58 Quit Hillshum (Nick collision from services.)
21:18:05 Nick Hillshum_ is now known as Hillshum (n=chatzill@unaffiliated/hillshum)
21:18:24LloreanAnd having the "Next folder" command work within playlists allows the UIs to match once playback starts, making the issue boil down to making playlist generation follow the old UI for when random advance is wanted.
21:18:30 Join Hillshum_ [0] (n=chatzill@unaffiliated/hillshum)
21:19:11PaulJam_it doesn't have to be exactly the same as before, but with the suggested solution it gets really overcomplicated: first i have to create the playlist and launch it, then enter the playlist viewer then i assume the folder order in this playlist is shuffled, so i'll have to search for a while to find the desired album that i want to hear now. opposet to just select a song from the filebrowser.
21:20:11LloreanPaulJam_: And yet you're preparing for hours of playback, so a few moments of prep time relative to it is hardly a significant amount. And as I've said, the UI could be fixed
21:20:19LloreanAs JdGordon| has pointed out, the issue is backend.
21:20:22JdGordon|Llorean: no, you are arguing the wrong argument.... the UI is fine, the backend is the issue.. what you want to do is jus a band aid fix for no real gain
21:20:37LloreanJdGordon|: As I've _said_ the UI could be made to match the current UI.
21:21:10JdGordon|the current UI is fine
21:21:17JdGordon|I wouldnt have a problem adding your way
21:21:24LloreanAre you just ignoring what I'm saying?
21:21:54 Quit _lifeless (Read error: 110 (Connection timed out))
21:22:03LloreanWhat part of 'the UI could be made to match the current UI" suggests that there's something wrong with the current UI?
21:22:25JdGordon|what you want to do does nothing but removes the feature for some people like me and PaulJam_
21:23:00LloreanIf the UI was made to match the current UI, people like you and PaulJam_ would be physically unable to tell the difference unless you manually entered the playlist viewer and saw it was more populated than usual.
21:23:08PaulJam_Llorean: maybe i have expressed myself badly. it is not that i really prepare for hours of playback, it is mainly that i want the player to continue playing albums after an album has finished, so i dont have to manually select a new one.
21:23:12LloreanYou really are just ignoring what I'm saying, right?
21:23:33JdGordon|Llorean: you miss the case where I'm at the folder i want to play now.. why do i need to go searching through the list to find it?
21:23:45LloreanThis is one of those "let's argue for the sake of arguing, but not actually read what's being responded" moments where you're missing the fact entirely that I've said it should be possible to address _all_ of your complaints from the UI
21:24:03LloreanJdGordon|: No, I didn't miss that. When I said "the UI could match the current one" I didn't mean in special cases.
21:24:07LloreanI meant in all cases
21:24:22LloreanThere's already a "random folder advance" option, you just make it change playlist generation instead of playback.
21:24:44LloreanWhen it's on, and you do what you just described, it iterates through folders randomly inserting them, so you get a giant playlist, rather than doing stuff once playback starts.
21:25:23LloreanThe problem with creating a giant playlist now is that once it's playing, the UI would be different.
21:25:34LloreanWhich is why I suggested a way to fix the in-playback UI.
21:26:49*Llorean gives up.
21:26:58JdGordon|ok I aprrently missed something
21:27:04JdGordon|that sounds fine... but it doesnt solve anything
21:27:06*Llorean is saying "you could have what you wanted without the folder advance hack in place" and is still being argued with.
21:27:13LloreanJdGordon|: What doesn't it solve?
21:27:23JdGordon|the underlying issue that playback<->playlist is a mess
21:27:26LloreanIt means the hack no longer needs to be in the playback system, because the playback system is just dealing with whole playlists.
21:27:31JdGordon|there are bigger fish to fry there
21:27:44LloreanAren't these hacks also in the way of frying those fish?
21:27:45JdGordon|playback shouldnt deal with *any* playliosts
21:28:20JdGordon|they would all be moved in one step
21:29:23*gevaerts thinks that random_folder_advance_config needs to be renamed. That name is at least twice as long as the next longest plugin name...
21:30:08Lloreangevaerts: It's always seemed odd to me that we've had that plugin, but use database.ignore files for the database
21:30:16LloreanIt seems like two different solutions to essentially identical problems.
21:30:37JdGordon|if someone wants to get their hands dirty, my suggestino would be to create a minimalistic playlist handler which allows the playback to get a filename to buffer, listens for the "track finished/changed" event, accepts a track change command from the WPS/UI and nothing more
21:30:41 Quit GodEater (Read error: 113 (No route to host))
21:31:26JdGordon|that plugin is there because it was deemed that the code to generate the file it works off was too big for core
21:32:18 Join GodEater [0] (n=nnyeahri@rockbox/staff/GodEater)
21:32:22 Quit GodEater (Remote closed the connection)
21:32:52 Join GodEater [0] (n=yeahrigh@rockbox/staff/GodEater)
21:35:50LloreanHow specific do these configurations usually get?
21:35:56LearNo .map files on
21:36:11LloreanI mean, is it just to make sure only your music folders get played, or are there often cases where you have specific albums you just never want to hear?
21:36:28LloreanTo which one?
21:36:47JdGordon|its up to the user i guess... everyone will be different
21:37:01JdGordon|the conf file isnt plain text which makes thing difficult
21:37:12JdGordon|rbutil could easily generate the file, and present a much nicer ui
21:37:25LloreanIt seems silly to have core functionality in a plugin, period.
21:37:44LloreanCould most uses of it be addressed by simply being able to set an "upper" bound for random folder advance? Like "Advance, but never leave the 'music' folder"?
21:37:51 Join raphi [0] (
21:38:05gevaertsis there some sort of shuffle function that's available to plugins?
21:39:01 Quit AndyIL (No route to host)
21:39:14PaulJam_i have a samples&demos folder and a "various artists" folder that contain unorderes single tarcks in my main music dir. i dont want those folders to be played in RFA mode. but i want them to be included when i listen to the whole music collection in normal shuffle mode.
21:39:44LloreanPaulJam_: So you could have /Music/Albums/everything and /Music/Singles/those other ones?
21:39:57JdGordon|Llorean: the issue is finding the folder to play if you dont generate a full listing on the fly
21:39:57LloreanThen bind it to the "Albums" folder
21:40:17*gevaerts also would like random folder advance to never kick in in the first place if the currently played track isn't in the RFA config
21:40:34LloreanJdGordon|: Huh?
21:40:44LloreanI'm not sure I understand how that relates to what I was suggesting.
21:41:01JdGordon|I guess I was trying to stop your next question
21:41:14LloreanWhat next question?
21:42:10JdGordon|why the config is needed at all
21:43:20LloreanDoesn't random folder advance work without the config already?
21:43:27LloreanIt just requires it if you want to constrain it to a subset of folders?
21:44:06PaulJam_Llorean: yes, splitting the collection into music/albums and music singles/ would work. i didnÄt think of that
21:45:27LloreanMaybe I'd like it better if the plugin wasn't more or less hidden among the others.
21:46:22rasherPerhaps it should be called from the menu, like credits?
21:46:59JdGordon|it does require the config
21:47:11JdGordon|which is nothing more than a \0 seperated list of folders
21:47:14JdGordon|and the count
21:51:10 Quit AlexP (Remote closed the connection)
21:51:31PaulJam_are you talking about the same file? i thought Llorean was referring to the folder_advance_dir.txt and not the folder_advance_list.dat
21:52:10 Join AlexP [0] (n=alex@rockbox/staff/AlexP)
21:52:13JdGordon|I was talking about the .dat requirement
21:54:15*JdGordon| dreams of a minimalistic playlist.c
21:54:29*JdGordon| fears playlist might be just too ingrained into everything to do it though
21:54:41 Part robin0800_
21:55:29 Join perrikwp1 [0] (
21:56:15 Join perrikwp2 [0] (
21:56:41*gevaerts suspects that using dircache from a plugin may not work well
21:56:53 Join robin0800_ [0] (
21:57:15JdGordon|why not?
21:57:46JdGordon|dircache just adds a layer above the actual fat/dir level
21:57:51gevaertswell, at least not the way playlist.c uses it, by borrowing context from the file browser
21:58:13JdGordon|apparently the dircache pointer in the playlist struct is not dircache related
21:58:34 Quit LambdaCalculus37 ()
21:58:37Jaykaya question about core voltage... can you set it as you want on every target (or some of them)?
21:58:43gevaertscould be. Maybe an older thing?
21:58:47 Join Thundercloud [0] (
21:58:47JdGordon|I tinhk so
21:58:57JdGordon|thats tripped me up just about every time i look at that file
21:59:15JdGordon|its a seperatly allocated buffer iirc
21:59:42kugelJaykay: no, not on all
22:00:40 Part robin0800_
22:00:41JdGordon|hmm... should a tr5ack change command go directly to playback? or through playlist?
22:00:51CIA-69New commit by 03alle (r21578): Improve the highscore related functions in plugin lib (FS #10350)
22:01:03Jaykaywhat would a too-low-voltage cause? a crash?
22:01:40 Quit {phoenix} (Remote closed the connection)
22:03:02*Hillshum thinks he has FS #10290 done
22:03:55kugelJaykay: you just can't change it
22:04:18kugeland where you can set it, yes it could happen
22:06:25 Join robin0800_ [0] (
22:06:26Jaykaykugel: on targets where you can change it... did somebody do test where the minimum voltage for boosted/unboosted is?
22:06:58Hillshumrasher: I posted another file to FS #10290. Anything missing?
22:07:24kugelJaykay: I think the AMSes are the first targets actually set set the voltage
22:07:40Jaykayok :)
22:08:05kugelthe gigabeat S can do it, maybe F/X too, maybe the creative and mr:robe 500 too
22:08:21kugelthe beast doesn't do it yet, I have no idea about the others
22:08:43AlexPI don't know if the F?X can, but it doesn't
22:08:47kugelJaykay: for the AMSes, we basically stick to what the OF does. Going by the datasheet, the setting isn't even allowed
22:08:53rasherHillshum: I don't really know much about what's needed
22:09:03AlexPThe F/X doesn't do clock scaling as runtime went down slightly when clock scaling was turned on
22:09:03LloreanI think I remember talk about doing it on the iPods so it may also be possible there.
22:09:04rasherHillshum: but I don't see any *errors*
22:09:12gevaertsCan you get at filetypes.h info from a plugin?
22:09:22Hillshumrasher: Should I just wait awhile then?
22:09:22saratoga_kugel: The V1 sansa let you change it too
22:10:05kugelwell, the as3514 does, but we don't know if the CPU supports it (I was going by the CPU)
22:10:12saratoga_we left it at 1.2v though since we didn't know if lower is safe
22:10:34kugelhas anyone tried lowering?
22:10:34 Join mcuelenaere [0] (
22:10:43saratoga_yeah i ran at 1.05v for a while
22:10:50kugelI can't imagine going a bit lower harms anything, particilarly for unboosted
22:11:05saratoga_it wont' break anything but it may cause crashes
22:11:26Jaykaybut the runtime should be longer right?
22:11:29saratoga_anyone who uses vorbis feel like testing a patch?
22:11:41saratoga_yes power scales with the square of voltage
22:12:45*kugel would be quite interested in a battery bench
22:13:00gevaertsWhat happens if an m3u playlist has a file that's not of a playable type?
22:13:06kugelJaykay: feel like risking your player? :P
22:13:23kugelwell, you can't really brick hardware by undervolting
22:13:25JdGordon|gevaerts: it should get skipped
22:13:32 Quit perrikwp (Read error: 110 (Connection timed out))
22:13:40gevaertsok. So I can just ignore the problem?
22:14:13 Join Jaykay_ [0] (
22:14:46JdGordon|depends on the problme?
22:14:56bertrikI'm seeing occasional hard hangs on my clip with SVN from a few hours ago
22:15:16 Quit perrikwp1 (Read error: 110 (Connection timed out))
22:15:20 Quit Jaykay (Nick collision from services.)
22:15:25 Nick Jaykay_ is now known as Jaykay (
22:16:40Jaykayone of the last questions :) is it easy to change the voltage?
22:16:43 Quit robin0800_ ("away")
22:16:53gevaertsJdGordon|: the playlist code use filetree things to see if a file is playable. That's not possible (at least now) from a plugin, so right now the RFA playlist generator just includes all files
22:17:25JdGordon|why isnt that possible from the plugin?
22:17:49 Quit perrikwp2 ("Leaving.")
22:18:01JdGordon|why dont you just the the add dir to playlist function in playlist.c?
22:18:05 Join perrikwp [0] (
22:18:10JdGordon|or whatever/wherever it is
22:18:25JdGordon|it might be in filtree.c even
22:18:40JdGordon|thats another thing which sucks.... filetree/tree/dirtree.c
22:18:50gevaertshm, good point
22:19:06kugelJdGordon: that doesn't suck at all
22:19:14JdGordon|it does!
22:19:24JdGordon|well... on a much smaller scale than other suckages
22:19:25 Join robin0800 [0] (
22:19:32gevaertsJdGordon|: because I've never looked at this code before! ;)
22:19:52kugeltagtree and filetree are divded were needed, and combined were possible in tree.c
22:20:05JdGordon|be careful... first you look.. then get involved... next thing you know your reworkignthe entire thing
22:20:46Hillshumrasher: if the specified target string does not exist, does langtool just skip it?
22:21:29JdGordon|Jaykay: if you have to ask these questions.... you really should be asking yourself why you want to change the values
22:21:58kugelJaykay: it's quite easy, yes
22:22:00rasherHillshum: It should, and I'm pretty sure it does.
22:22:05 Quit evilnick ("Page closed")
22:22:18JaykayJdGordon|: because i want to look how big the increase of the runtime is...
22:22:21Hillshumthat would be that problem then
22:22:39Hillshummaybe it should give an error message?
22:22:56rasherThat would be complicated
22:23:11rasherNot sure it's worth it. Just get the target string right!
22:23:19kugelJaykay: we're using the same code for v1 and AMS sansas when it comes to "ascodec"
22:26:48 Join Blue_Dude_ [0] (
22:27:41Jaykaywell the core voltage change for the ams sansas seems to be not easy enough for me.... maybe i'll annoy you another time in the next few days
22:32:01 Join merbanan [0] (
22:33:46Grahackgood night everyone
22:34:28 Quit Grahack ("Leaving.")
22:35:56 Join Ubuntuxer [0] (
22:37:04 Quit perrikwp (Read error: 113 (No route to host))
22:39:29 Join perrikwp [0] (
22:42:34 Quit Jaykay ("ChatZilla 0.9.84 [Firefox 3.5/20090624025744]")
22:42:41 Quit Blue_Dude (Read error: 110 (Connection timed out))
22:50:12***Saving seen data "./dancer.seen"
22:58:26 Quit mirak (Read error: 104 (Connection reset by peer))
23:00:29 Quit perrikwp (Read error: 110 (Connection timed out))
23:01:57kugelfml: does your latest commit break existing highscores?
23:02:11kugelI'd hate to lose my rockblox scores :(
23:05:10Ubuntuxerno, I tested it, it won't destroy your highscore
23:06:11 Join funman [0] (n=fun@rockbox/developer/funman)
23:06:56funmanHillshum: hi, i noticed a lot of updates to fs#10290 , is it ready to be committed or are you still working on other changes?
23:07:11HillshumI think its ready
23:08:30HillshumI havn't readied some of the recording entries, that will need to be done when recording is ready on AMS
23:09:17 Quit efyx (Remote closed the connection)
23:09:39HillshumBut it's fine for now
23:10:23gevaertsAre things like playlist_insert_directory() supposed to be safe to call while something else has grabbed the audio buffer?
23:11:37JdGordon|they should be
23:11:56JdGordon|playlist and filetree dont use the audio buffer
23:12:24*gevaerts adds more splashes
23:13:02 Join efyx [0] (
23:15:24CIA-69New commit by 03funman (r21579): fs#10290: Sansa AMS language fixes by Bertrik Sikken and Hilton Shumway
23:15:45funmanHillshum: thanks!
23:16:24Xqtftqxis CIA-69 a bot?
23:16:33 Quit Blue_Dude_ ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060215]")
23:16:58 Quit barrywardell (Remote closed the connection)
23:17:05funmanXqtftqx: yes, from
23:20:49 Quit Horscht ("Verlassend")
23:29:30 Quit raphi ("leaving...")
23:32:41CIA-69New commit by 03learman (r21580): Second attempt at fixing FS #10396. Turns out that gcc for coldfire choose to inline a function with a big appetite for stack. The previous fix ...
23:39:18 Quit merbanan (Remote closed the connection)
23:41:56bertrikhi all, working on DMA and PCM for s5l8700, doesn't look too complicated so far, I'm reusing the framework of the ams sansas for dma and pcm
23:42:04gevaertskugel: at some point you were working on a way to get plugins to "exit" to the wps. Did you get anywhere back then?
23:45:19 Quit Lear ("ChatZilla 0.9.85 [Firefox 3.5/20090624025744]")
23:45:54 Quit PaulJam_ (".")
23:47:08 Quit J-23 (Read error: 104 (Connection reset by peer))
23:47:11 Join J-23_ [0] (
23:52:55 Quit jgarvey ("Leaving")
23:54:09 Quit funman ("free(random());")
23:55:55kugelgevaerts: I got it to work
23:56:22kugelbut there were some issues with plugins using the simplelist
23:56:41kugelor rather, plugins executed via a simplelist
23:56:53 Part toffe82
23:58:34 Join FlynDice [0] (

Previous day | Next day