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 2005-09-27

00:15:44***Saving seen data "./dancer.seen"
00:26:26 Quit Moos (Read error: 104 (Connection reset by peer))
00:43:19 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
00:57:30 Quit ender` (Read error: 113 (No route to host))
01:04:04 Nick _CoCoLUS is now known as CoCoLUS (
01:26:56 Join Strath [0] (
01:27:23 Quit dpassen1 ()
01:36:14 Quit zeekoe (Remote closed the connection)
01:41:53 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
01:56:31 Join fuzzie [0] (
02:15:46***Saving seen data "./dancer.seen"
02:36:28 Quit _aLF ("Leaving")
02:43:30 Quit t0mas (Read error: 104 (Connection reset by peer))
02:47:15 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
03:04:01*fuzzie makes a note to never trust other people's code to be right, ever
03:14:01 Join courtc_ [0] (
03:14:07 Quit courtc_ (Read error: 104 (Connection reset by peer))
03:15:31 Quit tvelocity ("Leaving")
03:24:39 Join tvelocity [0] (
03:28:33 Join bagawk [0] (i=1000@unaffiliated/bagawk)
03:28:57 Quit goa (Read error: 110 (Connection timed out))
03:47:37 Join goa [0] (
04:03:06 Quit bagawk ("Leaving")
04:05:17 Join QT_ [0] (i=as@madwifi/users/area51)
04:05:24 Quit cYmen ("zZz")
04:15:47***Saving seen data "./dancer.seen"
04:15:51 Quit QT (Read error: 110 (Connection timed out))
04:27:08 Quit tvelocity ("Leaving")
04:27:08 Quit goa (Read error: 110 (Connection timed out))
04:32:26 Join Coldtoast [0] (
04:34:50Coldtoasthi. any chance you can have, like, a crash counter so every time Rockbox crashes on he same file, it ups the counter by 1 when Resume On Startup is enabled? I have an .mp4 on my player and when Rockbox tries to play it, it crashes and Rockbox has to be reset but cos I have Resume On Startup enabled, it keeps trying to play the file and crashes with an illegal instruction and has to reset
04:35:01Coldtoastso it's a constsnt crash-boot loop
04:38:41 Join goa [0] (
04:40:43 Part Coldtoast
04:40:49 Quit rasher (Remote closed the connection)
04:52:38 Quit JoeBorn ("")
04:53:07 Join ashridah [0] (
04:59:43 Quit Strath (Read error: 104 (Connection reset by peer))
05:03:22 Join Strath [0] (
05:13:09 Quit Sucka ("a bird in the bush is worth two in your house")
05:21:13fuzzieRight, I got AAC decoding working without the mp4ff code.
05:21:16fuzzieI am vaguely proud of myself.
05:21:23fuzzieNext step, try it on an actual H120.
05:46:11 Join webguest78 [0] (
05:47:01 Join Ismo [0] (
05:53:40fuzzieit fails on a bunch of files, which seems to be a known faad2 issue, argh.
05:54:32fuzziebut it is playing some of my other albums wonderfully, not even with the hiccups i normally get from the simulator.
05:54:40fuzzie[this still isn't on a H120 :P]
05:54:46fuzziesleep now, i guess.
05:59:21 Part webguest78
06:15:51***Saving seen data "./dancer.seen"
06:58:31 Join amiconn_ [0] (
07:16:58 Quit amiconn (Read error: 110 (Connection timed out))
07:16:58 Nick amiconn_ is now known as amiconn (
07:36:13 Quit Rob- (Read error: 110 (Connection timed out))
07:42:19 Join Xombie [0] (
07:42:37 Part Xombie
07:59:50 Join LinusN [0] (
08:15:53***Saving seen data "./dancer.seen"
08:25:53 Join ender` [0] (i=ychat@
08:34:02 Join einhirn [0] (
08:34:55linuxstb_Coldtoast: I'm guessing you mean ".m4a", and not ".mp4". If this is the case, are you using the very latest build?
08:35:47linuxstb_fuzzie: Congratulations on the AAC decoder. Please post the code somewhere.
08:46:52 Join Vladoman_ [0] (
08:47:04 Join ender1 [0] (i=ychat@
08:53:26 Quit ender` (Nick collision from services.)
08:53:30 Nick ender1 is now known as ender` (i=ychat@
08:56:22linuxstb_rasher: Thanks for the metadata.c fix.
08:57:16 Part Vladoman_
09:00:27 Quit Vladoman (Read error: 110 (Connection timed out))
09:00:59 Join B4gder [0] (n=daniel@
09:02:20 Join Vladoman [0] (
09:12:05 Join Paul_The_Nerd [0] (
09:25:22 Part Vladoman
09:36:47 Quit Paul_The_Nerd (Read error: 110 (Connection timed out))
09:41:53 Join Vladoman [0] (
10:03:26 Part stamppot ("Kopete 0.10.2 :")
10:07:11 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
10:13:05 Quit gromit` (Remote closed the connection)
10:15:01 Join gromit` [0] (n=gromit`
10:15:56***Saving seen data "./dancer.seen"
10:16:11 Join [IDC]Dragon [0] (
10:37:56 Nick QT_ is now known as QT (i=as@madwifi/users/area51)
11:16:10 Quit paugh ("Leaving")
11:16:23 Quit ashridah (Read error: 110 (Connection timed out))
11:23:37 Quit lImbus (" word")
11:32:55 Join preglow [0] (
11:33:54*linuxstb_ curses Debian for not compiling things with SSL
11:37:25preglowwoot, aac!
11:37:44linuxstb_preglow: Yes, but only in the Sim. I'm very curious to see how fast it is on the H1x0
11:38:10linuxstb_I'm expecting it to be very slow - given how slow libmad was to start with.
11:40:02 Join rasher [0] (
11:40:04preglowoh, yes
11:40:16preglowso faad has problems with some files?
11:40:53linuxstb_Maybe the problems were fixed with later versions. Also, only part of the librariy has been integer-ised.
11:40:56preglowbut hmm
11:41:08preglowany more codecs that are way too bad, apart from mpc?
11:41:17 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
11:41:42linuxstb_How is Tremor working? I don't use it.
11:42:08preglowtremor is doing pretty well
11:42:56preglowi dont want to touch it before i can measure the improvements
11:43:00linuxstb_Have most of the optimisations been through good use of the IRAM or the EMAC? I'm thinking about the work needed to be done on the ipod (which also has 96KB of IRAM)
11:43:18preglowmost has been done with both emac and iram
11:44:12 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
11:44:55preglowone of the last obvious places to optimise would be to replace the imdct, but that too might give too subtle results for me to do it before i can measure properly
11:45:16preglowi think i'll just have a quick go a doing some iram work on libmusepack
11:45:21preglowthe math i wont touch
11:46:05linuxstb_ALAC was working about the same speed as musepack before I optimised the IRAM usage.
11:46:19preglowi can get musepack going realtime easily with iram
11:46:24linuxstb_Does MPC use any iram at the moment?
11:46:32preglowfor the filter bank
11:46:37preglowproblem is the main struct is too large for iram
11:51:00preglowthe data struct uses all the available iram in just one field...
11:59:33linuxstb_Is the iriver charger connector positive on the inside and negative on the outside?
11:59:43linuxstb_(I've confused my Nokia and iriver chargers again)
12:00:02linuxstb_That's why I'm asking :)
12:00:14HCli'm at college so i can't check.
12:00:23HCldoesn't it say on the iriver unit itself?
12:00:34 Join ashridah [0] (
12:00:38rasherHang on
12:00:40HClwhy don't you simply throw the one labeled nokia away?
12:01:15amiconnlinuxstb_: "Charger spec 5V 2A center +"
12:01:31HClyea, it does say it on the unit itself
12:01:38HClindeed, center positive
12:02:21preglowi think they look pretty different
12:02:25linuxstb_Thanks both of you. Yes, the unit itelf identifies the polarity, so I've found the right charger.
12:02:27preglownever managed to confuse them
12:02:50linuxstb_preglow: I've also lost my Nokia charger. I was 99% sure I had the right charger, but wanted to double-check.
12:03:03linuxstb_I must start labelling my chargers.
12:03:20rasherI keep them seperate. Iriver charger always connected to the same plug, nokia charger moving around.
12:03:51ashridahi keep them in separate rooms
12:04:08rasherwell, I don't have seperate rooms
12:04:13rasherapart from the kitchen.
12:04:20preglowlibmusepack is so incredibly memory hungry
12:05:53preglowfor such a simple codec
12:09:43linuxstb_preglow: A simple possible optimisation - could the loop in mpc.c that copies the data from sample_buffer to output_buf do the conversion in-place?
12:10:19linuxstb_And then call pcmbuf_insert directly on sample_buffer.
12:10:53preglowwell, should definitely be possible
12:10:56preglowi'll try
12:11:13linuxstb_That's one change I made for ALAC. It seemed to give a noticable improvement.
12:11:28preglowoy, what the hell
12:11:31preglowit converts to 16 bit numbers!
12:11:34preglowthat's not right
12:12:01linuxstb_So you should remove that conversion completely?
12:12:02preglowi'll whip me up some food before trying here, stomach's growling at me
12:12:05preglowoh yes, i shall
12:15:04 Join Greek-Boy [0] (n=grb@
12:15:54 Part Greek-Boy
12:15:59***Saving seen data "./dancer.seen"
12:27:30rasher(mostly) Good news about aac
12:28:23 Join uncledrax [0] (
12:31:27linuxstb_I'm curious to know which AAC files do and don't work. e.g. do "standard" 128kbps iTunes-encoded files work?
12:33:04preglowgod knows, i'd love some juicy details
12:33:24linuxstb_.. and a copy of the source.
12:33:59linuxstb_We still need to sort out the problem of two codecs using .m4a
12:34:36linuxstb_But I'm happy to rename my alac files to .alac if we can't think of an easy solution.
12:34:37B4gderoh my goodness...
12:34:37preglowoh yes
12:34:47*B4gder visits MR for the first time in a week or so
12:35:01*preglow hands B4gder a pair of dark sunglasses
12:35:05linuxstb_B4gder: "Horrific" is the adjective normally applied.
12:35:33B4gderwhen expressing it in a mild way
12:35:38linuxstb_But I think you can revert back to the old style if you log in.
12:36:14LinusNyes you can
12:36:58LinusNthere is a "quick style chooser" in the bottom left corner, select "test" from there
12:49:43amiconnZagor: The daily installer builds are still broken...
13:00:06 Join thegeek [0] (
13:00:42elinenbeLinusN: an unrelated quick question... have you had any time to work on the H300 series?
13:02:12elinenbethanks... just wondering.
13:02:20elinenbegood luck when you do get a chance.
13:02:29elinenbeI know it takes a good amount of time.
13:22:30 Join tucoz [0] (
13:23:11preglowmy oh my, the codec sources need cleaning
13:25:26tucozJust a small question. Radio plays when connected to usb (which is ok), but when disconnected it stops. Is this the wanted behaviour? It is the irver port I am talking about.
13:25:27 Quit uncledrax (Read error: 104 (Connection reset by peer))
13:26:23tucozI mean it _keeps_ playing when connected.
13:28:00LinusNwe haven't really thought that much about the desired behaviour in that case
13:28:17LinusNit works the same on the archos though
13:28:43rasherin my eyes it'd make sense to either a) stop playback on insert or b) not stop ot at all
13:28:58LinusNi agree
13:30:09tucozWell, it does b) now, but it is the disconnect that is not really doing what I expect :)
13:31:07rasherWell I meant, b) stop it on neither insert nor remove
13:31:52tucozthen we agree
13:33:18tucozIf I wanted that behaviour (b). Is this done by checking the status of SYS_USB_DISCONNECTED in radio.c?
13:37:36LinusNadd this after line 478 in radio.c:
13:37:38LinusN keep_playing = true;
13:38:04tucozok, thanks :)
13:38:15LinusNit will still leave the radio screen though
13:38:24rasherDoes this also work if you're not in the radio screen?
13:39:04LinusNiirc, if you leave the radio screen with the radio playing, it will keep on playing even after exiting usb mode
13:39:15rasherah, okay
13:40:02LinusNin fact, there is no need to leave the radio screen when leaving usb mode
13:40:30rasherindeed, it does keep playing
13:40:35LinusNbut the general rockbox behaviour has always been to go back to the browser
13:40:39preglowlinuxstb_: hmm, i might need to shift the numbers some anyway, i don't think there's enough dynamic headroom left in musepacks original format to allow for a big replaygain without clipping
13:46:32tucozbtw, did you see that someone got mp3 encoding working? Antonius Hellman iirc.
13:46:44preglowyes, i noticed
13:46:50preglowwhich encoder was he talking about?
13:46:52preglowshine or the other one?
13:46:55rashershine, I think
13:46:58rasherthat's how I read it at least
13:47:11preglowwell, then i think he should commit it :)
13:48:05tucozI think so too. For HQ recordings, people would use wavpack, and for other stuff, like lectures one could use mp3 as he said the quality is so so.
13:48:26rasherHow about speex encoding - any idea how complex that is?
13:48:32preglowi think it might work
13:49:06tucozIt is floating point still
13:49:20preglowit's not
13:49:27rasherah. hurray, then
13:49:28preglowjust parts of it
13:49:31tucozit isn't, I checked last week
13:49:39preglowi think a complete encoder or decoder can be compiled in fixed point mode
13:49:44preglowat least i managed to have the decoder compile
13:49:45tucozand then it was still floating point
13:49:57preglowtucoz: you're talking about the stable or the dev branch?
13:50:12preglowthey've even ported it to blackfin, and that's impossible without fixed point support
13:50:26tucozstable I guess, I just saw that they needed help with the fixed point version
13:50:28preglowthat's even 16 bit fixed point
13:50:32preglowstable has no fixed point support
13:50:32preglowdev has
13:50:57rasherSo speex on iriver is very possible? Encoding might be?
13:51:02preglowyes, i'd say so
13:51:14preglowdecoding should certainly work, unless speex is more complex than i'm aware of
13:51:22tucozI understood that they are working on fixed point. Didn't realize they were quite there yet though.
13:52:16preglowonly one way to find out!
13:54:25tucozhehe, you go ahead. I didn't even manage to get it to compile under rockbox.
13:57:28preglowpeople using that little whitespace needs to use more whitespace :>
13:58:06preglowthis isn't c64
13:58:08rasherMaybe his space key was broken
13:58:24preglowi've been there, i just started pasting spaces instead
14:00:00*preglow pokes Slasheri
14:03:38tucozah, now my radio is following plan b)
14:03:57LinusNdoes it leave the fm screen?
14:04:17tucozIt returns to fm screen when unplugged
14:04:30paughpreglow: any progress with flac?
14:04:45LinusNtucoz: then we need top decide what to do with the presets
14:04:59tucozhmm, what do you mean?
14:05:16preglowpaugh: nothing more than what i did some weeks ago, no
14:05:17LinusNwhat if you entered usb mode to change the preset file?
14:05:18tucozah, if someone updates the preset file
14:05:26amiconntucoz: Hopefully that doesn't interfere with the recording handling...
14:05:43LinusNamiconn: it won't enter usb mode if recording
14:05:52tucozNo, that is checked before
14:05:55paughpreglow, okies
14:05:59tucozentering usb-mode
14:06:52preglowpaugh: i wont to any more non-obvious optimising until i can measure my improvements, which wont be until either i or someone more clever makes a wav writer
14:07:15tucozI just need to know what done=true does in the radio. done means done with the radio right?
14:07:26tucozradio screen that is
14:07:46LinusNtucoz: line 277
14:08:05LinusNdone=true leaves the loop
14:08:25tucozI get it
14:08:30tucozI am lazy
14:09:46LinusNyou could add this after usb mode:
14:10:02LinusNpresets_loaded = false;
14:10:31Slasheripreglow: hi
14:10:39LinusNthat will reload the presets
14:11:35preglowSlasheri: had any time for preparing some commits yet? :)
14:11:41tucozok, cool
14:11:46preglowi'm anxious to try your buffering
14:12:02rasherI'm anxious to try recording
14:12:42Slasheripreglow: not yet and currently i am little sick :/
14:13:01Slasheripreglow: but i will prepare the dir cache for review on this week
14:13:18tucozThe buffering came to mind when I read about the remote handling. Threaded or not. Won't buffering cause problems or issues if the remote is in another thread?
14:13:29preglowwhy should it?
14:13:46tucozI don't know
14:14:07preglowSlasheri: and why does pcmbuf_insert take char * when pcmbuf_insert_split takes void * ?
14:14:18preglowhistorical reasons? ;)
14:14:38tucozthat is why I am asking. I do not know how the buffering works.
14:15:09tucozI meant caching of course
14:15:17Slasheripreglow: Hmm, yes. Or maybe the reason was that the samples may be 16 or 32 bits that are passed to that function
14:15:25 Join Rori [0] (
14:15:30RoriLOL @
14:15:31SlasheriSo using char or short wouldn't be right either
14:16:03***Saving seen data "./dancer.seen"
14:16:48RoriBTW I saw the new Sony 20GB player the other day. Small and swish looking and can take a knock or two as demonstrated by them dropping it onto the counter several times.
14:16:58Slasheribut yes, char* might be better
14:17:21RoriShame the screen is not full colour and no gapless though
14:18:21tucozRori: sounds like a demonstration in the lines of infomersials where they cut tin boxes with their fantastic knives.
14:19:10rasherHmm.. interesting
14:19:12tucozBut I do like the looks of Sony's mp3 players
14:19:29preglowisn't interleaved audio supported yet?
14:19:53preglowoh yes
14:20:19tucozrasher: fun
14:23:25 Join [0x00] [0] (
14:26:12*B4gder bugs his friend to covert his sudoku site to a .ss generating one
14:26:20 Part [0x00]
14:29:11linuxstb_B4gder: What does it produce at the moment?
14:29:26linuxstb_Ah. We don't want to parse that.
14:29:43B4gderit should be very easy for him to make .ss as well
14:30:25preglowlinuxstb_: well, it's ALMOST realtime now
14:30:33preglowbut i've done something wrong
14:30:36preglowsounds pretty bad
14:32:17tucozmaybe a bad patch, but it works. Returns to radio-screen on usb disconnect. also handles changes in the preset file.
14:32:46Rorihey guys what is wrong with the bass and treble controls? if I turn bass up the treble gets quieter and if I turn up treble the bass ends just drops. no additional bass or treble on either control.
14:33:34rasheryou probably have the volume way up?
14:33:37Roriit's better with both off
14:33:45RoriI have it 100% yeah heh
14:33:52preglowRori: if the volume is at 100 and you use the eq, the main volume will be lowered automatically to avoid clipping
14:34:51Roriyeah noticed pre-amp does not work either when at 100%
14:35:08rasherhow could it?
14:35:18rasherwell, except for clipping
14:35:27preglowworks for me
14:35:51LinusNtucoz: no need to set screen_freeze twice
14:35:52preglow100% volume and 12db preamp, sounded really, really bad
14:36:05tucozLinusN: oh, ok
14:36:27rasherLinusN: how do you feel about the "enable/disable statusbar wps tags"?
14:36:41LinusNi think it's a good idea
14:36:57LinusNbut i'm not sure i like the way it is implemented
14:37:09LinusNi haven't looked at the latest patch though
14:37:14preglowSlasheri: is 32 bit interleaved stereo audio supported by pcmbuf_insert?
14:37:19rasherHe didn't change a lot
14:37:25preglowSlasheri: to my untrained eye it doesn't look like it's handled at all
14:37:33rasherLinusN: Pretty much just fix the bug I pointed out
14:38:02LinusNi'll have another look one of those days
14:38:25Slasheripreglow: yes should be when dsp is enabled
14:38:49LinusNtucoz: you don't need to set keep_playing=true since you don't leave the radio screen
14:40:13tucozLook what I know. Then it's easier than I thought.
14:40:34tucozNow I understand the code, didn't 5 minutes ago
14:40:57tucozand also no need to freeze the screen either
14:43:17preglowSlasheri: and conversion from arbitrary sample depths are handled correctly?
14:43:37Slasheripreglow: yes, that's handled by the dsp
14:43:38tucozThen, is this ok?
14:44:11rasherBy the way, m4a supports per-sample samplerate - can the dsp handle that?
14:44:17LinusNtucoz: looks ok
14:44:20rasher(lord knows why you'd do that though)
14:44:48preglowper sample sample rate?
14:44:52preglowthat doesn't make sense
14:45:12rashereh.. frame?
14:45:23preglowlike in mp3, then?
14:45:27preglowif so, yes, we already handle that
14:45:27Slasherirasher: that depends but should be possible
14:46:00linuxstb_preglow: How do we handle that?
14:46:25preglowmpa.c line 236
14:47:37linuxstb_OK. But is the track duration calculated correctly?
14:47:52preglowwell, afaik, that is handled by id3 data
14:47:56preglowso it's out of our hands
14:48:32preglowor in the case of frame counts from xing headers or the like, no, it's probably not
14:48:36linuxstb_Or even the current location in the file - that's normally calculated by dividing by samplerate.
14:48:41preglowwe'd need to keep a running estimate thebn
14:48:55preglowand i don't think we should do anything more complicated than that
14:49:07preglowswitching sample rate in the middle of a file doesn't make sense
14:49:11preglowwe just make sure we handle it gracefully
14:49:15linuxstb_mpa.c should probably be changed to keep the current position indicator in ms, not samples.
14:49:38preglowin which case you will quickly run into precision problems
14:49:52linuxstb_I thought each frame was an integer number of ms in duration?
14:49:54preglowunless you let it be a 64 bit fixed point int or something
14:50:17preglowno, not as far as i know
14:50:37preglowin fact, i can pretty much guarantee it's not
14:51:12linuxstb_Maybe it's just MP2 then - I don't really use MP3.
14:51:31preglowSlasheri: length passed to the insert functions is bytes per channel, yes?
14:51:42LinusNiirc, 26.12ms for a 44.1kHz layer3 frame
14:51:43preglowi can pretty much guarantee it's not the case there either
14:51:50preglowLinusN: that's correct
14:51:56preglowLinusN: and even that's an approximation
14:52:09preglowlinuxstb_: the number of samples per frame is independent of sample rate
14:52:13Slasheripreglow: yes i think so :)
14:52:34preglowlinuxstb_: for some combinations it's bound to be a non-integer
14:52:48linuxstb_Ignore me then. I only have 48KHz layer 2 files, which I think are exactly 24ms per frame.
14:53:17preglowlinuxstb_: that is correct
14:53:46preglowbut not for 44100 which is far more common
14:56:26SlasheriLinusN: btw, if i would like to have a global HAVE_DIRCACHE define is firmware/export/config-h100.h (and h120) the right place to put that?
14:57:38 Join webguest02 [0] (
14:58:25tucozadded patch to the tracker
14:58:55 Quit webguest02 (Client Quit)
14:59:52 Part tucoz
15:00:25LinusNSlasheri: yes, or define it based on the ram size
15:00:33 Join tucoz [0] (
15:00:58Slasheriok, good
15:01:07LinusN#if MEM > 8
15:01:08rasherThe latter sounds more reasonable
15:01:16LinusN#define HAVE_DIRCACHE
15:01:26LinusNor sth like that
15:01:32Slasheriwhat is then the right file to put that define?
15:01:43B4gderconfig.h probably
15:01:48Slasheriah, yes :)
15:01:50LinusNconfig.h, after including the various special config files
15:01:59 Part tucoz
15:02:17preglowthe right channel is all noise whatever i do
15:02:35LinusNgotta go, cu all
15:02:38 Part LinusN
15:03:06Slasheripreglow: hmm.. what are you trying to do?
15:03:41 Join tucoz [0] (
15:05:24preglowSlasheri: mpc.c was converting the numbers to 16 bits before passing to pcmbuf_insert, i'm making it pass the native format instead
15:05:31preglowSlasheri: which seems to be 32 bit interleaved stereo
15:05:38 Quit tucoz (Client Quit)
15:05:50Slasheriah, that sounds nice
15:06:10preglowthe left channel is music at the original tempo, but at half pitch and the right channel is pure high amplitude noise
15:06:35Slasherido you think it could be then possible to play some mpc files realtime?
15:06:48preglowno, but it's not far away
15:06:53Slasheripreglow: have you set up the dsp correctly?
15:06:56Slasheriok :)
15:07:00preglowwanna have a look at the source?
15:07:05Slasherihmm, maybe
15:07:46preglowi think i have
15:07:48 Join uncledrax [0] (
15:10:05Slasherihard to say but there might be some bugs in the dsp code if you really have interleaved stereo (i think that's still untested)
15:10:32preglowyes, so do i
15:11:57linuxstb_Don't shoot me for asking this, but when the tag database was being designed, was using the ITunes database format considered?
15:12:25linuxstb_I'm not just thinking about Rockbox on the ipod - but there are lots of tools available to manipulate such databases.
15:12:27preglowis the format even described anywhere?
15:12:31preglowahh, ok
15:12:57linuxstb_Plus all the open source tools that use it - e.g. gtkpod
15:13:02B4gderlinuxstb_: we wanted a slim and mean format for less capable devices too
15:13:15B4gderthe itunes db doesn't seem to match that criteria
15:13:28linuxstb_That's all I was asking - so you did look at it?
15:13:34B4gderyes, briefly
15:14:10linuxstb_A file browser interface is pointless on an ipod that's been populated by itunes - filenames are hashed.
15:14:25Slasherilinuxstb_: i would like to sync the rockbox with amarok's database
15:14:27linuxstb_e.g. F01/IUEQ.m4a
15:15:59linuxstb_And directory "F01" doesn't relate to an album - it's just an arbitrary way of splitting the files into directory.
15:16:29linuxstb_So it seems each file is given a 7 character hash, the first 3 are the directory name, the last 4 are the filename.
15:16:51linuxstb_The metadata is still in the .m4a file though.
15:17:09rasherobfuscation at its finest
15:17:21rasher"no sire, you can't move files off your ipod"
15:17:27rasher"See, they're not there!"
15:18:44linuxstb_All the music files and the iTunesDB are stored in a directory called "iPod_Control". This has the system/hidden bits set on FAT32 so Windows users can't normally see it.
15:21:27linuxstb_Any ideas for how Rockbox could support such a scheme?
15:22:03rasherReading the db, I guess
15:22:43rasherI don't see how else you'd use such a player
15:24:00rashercan't itunes be told not to mangle filenames though?
15:24:13rasherI was under the impression that you could use it for regular ums players as well
15:24:33 Quit B4gder ("time to say moo")
15:25:21linuxstb_rasher; I don't think you can tell anything to iTunes. Maybe other tools like gtkpod could.
15:26:11rasherI'm pretty sure you can use iTunes with other players
15:26:17rasherin some way or other
15:26:28rasherI'd expect it not to mangle filenames in that case
15:29:45*preglow curses people who use interleaved data
15:31:15linuxstb_rasher: I would be surprised if you could use iTunes with other players.
15:31:40rasherwell, I'll try
15:32:00linuxstb_iTunes stored the files on your PC in a sensible Artist/Album/NN-TrackName.m4a structure. But not when you use it to sync files to the ipod.
15:32:22rasherBut that's because they know the ipod to read the music db thing
15:35:25 Join linuxstb [0] (
15:35:34linuxstb - the "BadApple" plugin
15:36:00linuxstbSo it seems that standard itunes can't do that.
15:36:24linuxstbExplains it all.
15:37:13preglowdeinterleaver code looks ok to me
15:37:36linuxstbSeems that BadApple only syncs MP3 files to generic players - it ignores other formats :(
15:37:39rasherMaybe that's what people have been talking about, and I just thought it was something iTunes could do
15:39:24VladomanHi, I just built the Iriver X11 simulator, when I start it I get something like: We open the real file 'archos/.rockbox/codecs/mpa.codec',
15:39:29 Join ]RowaN[ [0] (
15:39:38Vladomandoes that mean I need to create this folder and put a codec there?
15:39:46]RowaN[guys is the iriver hd in fat32 or ntfs format?
15:39:52linuxstbVladoman: Did you do a "make install"
15:39:53rasher]RowaN[: fat32
15:40:31]RowaN[i was going to get a "usbextreme" bootdisc for my playstation2.. it lets u run games from a usb hd, but has to be ntfs aparently
15:40:44linuxstbVladoman: Then type "make install" from your build directory :).
15:41:00linuxstbThat will update the archos/.rockbox/ directory.
15:41:16VladomanI just did, it crashes now with: Floating point exception./rockboxui
15:41:26Vladomanwhen I tryx to play an MP3 file
15:41:41linuxstbIs this the X11 or win32 simulator?
15:42:02 Join tvelocity [0] (
15:42:36linuxstbUnder Linux or Cygwin (or something else?)
15:42:49Vladomanpure Linux
15:42:59linuxstbOK, let me test it as well.
15:43:12Vladomanit's 2.5
15:44:05linuxstbForgot to ask - which target are you simulating?
15:44:10]RowaN[if i formatted my river to ntfs.. ? it would no longer boot?
15:44:22rasher]RowaN[: no
15:44:29linuxstb]RowaN[: Yes, neither Rockbox or the iRiver firmware will recognise it.
15:44:52Vladomaniriver H120/H140
15:45:12linuxstbVladoman: OK, compiling now.
15:46:40 Join cYmen [0] (
15:49:23linuxstbVladoman: Works fine for me under Linux (Debian).
15:50:33linuxstb I don't know what to suggest apart from starting again with a clean cvs checkout and a clean build directory.
15:51:53VladomanI'm running it in DDD now
15:52:11 Quit linuxstb_ (Read error: 110 (Connection timed out))
15:52:18VladomanProgram received signal SIGFPE, Arithmetic exception.
15:52:19Vladoman[Switching to Thread 16384 (LWP 17764)]
15:52:19Vladoman0x0806620b in peak_meter_draw (x=0, y=56, width=160, height=8) at recorder/peakmeter.c:847
15:52:39rasherTiMiD[despair]: patching
15:53:15Vladomanlook like it is this line:
15:53:17Vladomandb_scale_lcd_coord[i] =
15:53:17Vladoman (i * (MAX_PEAK / 10) - peak_meter_range_min) *
15:53:17Vladoman meterwidth / pm_range;
15:55:08VladomanGotcha, it works if I disable the peakmeter!
15:55:37linuxstbThat's strange - the WPS in my simulator contains a (non-working) peakmeter.
15:55:57linuxstbi.e. the peakmeter displays, but doesn't work.
15:56:27Vladomanmaybe pm_range has a random value and mine is 0?
15:56:39linuxstbVladoman: Possibly.
15:57:41Vladomannow, if I knew the keys to use...
15:59:23rasherTiMiD[despair]: Good job!
15:59:30*rasher browses on his remote
16:00:32rasherQuite cool
16:00:58preglowxavier patch?
16:01:07rasherTiMiD patch
16:01:11 Join webguest41 [0] (
16:01:16amiconnVladoman: (1) even if pm_range is zero, you shouldn't get a floating point exception
16:01:17 Quit webguest41 (Client Quit)
16:01:25amiconnRockbox doesn't use floating point
16:01:54amiconn(2) The keyboad layout is printed to the console you started the simulator from
16:02:19Vladomanthx, me = blind :-)
16:03:53 Quit uncledrax ()
16:05:13rasherI so need to add this to my patched-up build
16:05:17*[IDC]Dragon unlurks
16:05:58[IDC]Dragonamiconn, last time I didn't manage to d/l your gcc
16:06:43amiconnYeah, probably because the file wasn't called as announced here. There was a stray dot in the filename
16:07:05[IDC]Dragondo you still have it?
16:07:20 Join DangerousDan [0] (
16:08:01amiconnNo, I would need to zip it again, and my apache isn't running atm
16:08:11 Quit rasher (Remote closed the connection)
16:08:28]RowaN[ah now i read that usbextreme requires fat32, NOT ntfs
16:08:59[IDC]Dragonno hurry
16:09:26[IDC]DragonI'd just like to compile a bootbox with 2.5 someday
16:09:46 Join rasher [0] (
16:10:37Vladomanamiconn: I just added a "return" in peak_meter_draw() and then the crash disappeared
16:11:41 Join XavierGr [0] (
16:13:16rasherTiMiD[despair]: oi, just got an IllInstr after plugging usb with your patch(es)
16:13:26rasherTiMiD[despair]: while in wps
16:15:43preglowmpc is now aaaaaaalmost 100% realtime for this standard encoded file i have
16:16:04***Saving seen data "./dancer.seen"
16:16:14preglowtoo bad the sound doesn't sound like shit
16:18:38preglowi mean it sounds like a bag of shit
16:18:47preglowto use the new terminology
16:19:20*rasher is confused
16:19:57rasherI thought firefly on the forums was timid, for some reason... but that is not the case, is it?
16:22:10rasherAnyway, this remote support *looks* right
16:22:16rasherfrom the user's viewpoint
16:22:46rashermenus and filetree is displayed correctly
16:23:46*amiconn is confused
16:24:35preglowbut i gotta go
16:26:09rasheramiconn: about what?
16:26:40amiconnConfusion solved, I just didn't remember how to login to cvs with a different user name that my local one
16:27:16]RowaN[offtopic: does anyone know if theres a gfx card which can output (im talking TV-out) resolutions lower than 640x480?
16:30:01 Quit linuxstb (Read error: 110 (Connection timed out))
16:30:18amiconnVladoman: I'm just trying to build the h120 x11 sim under linux
16:44:54 Join linuxstb [0] (
16:47:09amiconnI don't receive a floating point exception here
16:48:15 Quit XavierGr ()
16:48:49rasherI have a rockbox with remote support build if anyone wants to test and don't want to apply the patches:">
16:49:14rasherThis is the patch from
16:51:30linuxstbrasher: Have you tested the patch with the remote unplugged?
16:52:35rasherNot yet
16:52:56rashera quick test doesn't show any problems
16:53:14rasher(ie. remove remote, change menus, replug)
16:53:34ashridahrasher: it apply against currnt cvs HEAD?
16:53:47linuxstbIt seems a _very_ large patch. (160KB)
16:54:10rasherashridah: It depends on two other patches - included in the package
16:54:24rasherThis is not my work by the way, if anyone was under this impression
16:54:48rasher(the author posted the forum thread)
16:55:04rasherlinuxstb: it touches a lot of files
16:55:45rasherIt even patches rolo, so I get "Rolo loading" on the remote
16:55:45 Quit tvelocity ("Leaving")
16:56:59rasherI think this is the sort of patch that should be acted upon quite quickly
16:57:07rasheror it'll end up being impossible
16:57:40rasherTiMiD[despair]: I was confusing you for a forum poster
16:58:16linuxstbMy first impression after a brief look at the patch is that it just seems to duplicate everything done on the mail display to the remote - thereby doubling the amount of code.
16:58:42linuxstbBut I'm not sure how much that can be avoided.
16:59:11rasherShould be possible to avoid in some places with the inclusion of a list-widget I guess
16:59:29linuxstbSome functions will need to be made more general - and called twice (once for main LCD, once for remote).
17:00:08rasherAnyway, this is at least better than XavierGr's patch, since it correctly takes screen size into account
17:00:09ashridahit works.
17:00:24rasherBut, I haven't looked at the source, to be honest.
17:00:42Vladomanamiconn: thx for the help, it's not really important.
17:00:43*ashridah is just happy to not have to take out his player to poke at it
17:01:10TiMiD[despair]if you don't accept this patch, I'm working on a totally different approach
17:01:18rasherTiMiD[despair]: oh, it *was* you?
17:01:23rasheror wasn't
17:01:28rasherI get it now.
17:01:32*rasher is unconfused.
17:01:42TiMiD[despair]I didn't posted this patch
17:01:57rasherI think your approach (the list-widget) is more likely to be accepted
17:02:01TiMiD[despair](Copy-paste code is boring :p)
17:02:17TiMiD[despair]mine it's working pretty good :p
17:02:21rasherBut of course, some places will have to be fixed like this patch is doing
17:02:33rasherSo those parts of this patch can be used
17:02:34TiMiD[despair]you can display synchronized lists on both display in 3 lines :)
17:02:47ashridahthe scrolling settings don't seem to have an immediate effect on the remote
17:02:57TiMiD[despair]you can also hot-switch from a display to another
17:03:28TiMiD[despair]but it need more work (and it would involve to modify more deeply the existing code)
17:03:47 Join linuxstb_ [0] (
17:04:02ashridahhas issues drawing the marker when scrolling through the file list as well
17:04:02 Quit ]RowaN[ ("CGI:IRC (EOF)")
17:04:18TiMiD[despair]by the way I have a question
17:04:32rasherAny places that aren't menus or filetrees though, will benefit from this patch even if we want to go with TiMiD[despair]'s approach
17:04:49rasherI think I'll try to split those parts out from this patch
17:04:57TiMiD[despair]the lcd_putc on the player takes an 'unsigned short' for the character to display
17:05:07rasherie. stuff like error messages and such
17:05:38TiMiD[despair]but to display the cursor, the value is '#define CURSOR_CHAR 0x92'
17:05:49TiMiD[despair]and 0x92>127, so wtf ??
17:07:14ashridahisn't a short 16 bits long?
17:07:38TiMiD[despair]rasher: I intend to uste the same 'gui list' code in both menus and filetree (since it's possible that one day menus have icons too ...)
17:07:52TiMiD[despair]you are right :)
17:08:14ashridahand if it's unsigned, 0x92 wouldn't be a problem for an 8 bit datatype
17:08:37rasherTiMiD[despair]: makes sense
17:08:43TiMiD[despair]the problem is that when I try to display a char > 127, it doesn't work
17:08:56TiMiD[despair]17:07 < ashridah> isn't a short 16 bits long?
17:09:06TiMiD[despair]lcd_putc(0, 1, 65426 '0xff92')
17:09:31*rasher stares at 4800 lines of patch
17:09:49TiMiD[despair]that's what I get after calling screen->putc(x, y, icon); (which is a pointer to lcd_putc)
17:10:02ashridahrasher: heh, that's what you get when you have a month long freeze :)
17:10:15TiMiD[despair]when icon=CURSOR_CHAR (0x92)
17:10:16rasherashridah: No, that's just the remote-patch
17:10:37linuxstb_rasher: Does it use the same wps file for both the remote and the player?
17:10:46rasherlinuxstb: the remote wps is hardcoded
17:10:50 Nick TiMiD[despair] is now known as TiMiD (n=TiMiD[
17:10:52ashridahrasher: i know, but given that it wasn't able to be piece-wise merged... (if you wanted to merge it) because of the freeze... :)
17:11:07rasherAh, like that
17:11:14rasherI guess it would be nice to have it split up
17:11:29rasherlike "this patch adds filetree browsing on the remote"
17:11:42rasher"this one adds debug menu"
17:11:44rasherand so on
17:12:11rasherI wonder if there are any patch-editing gui tools
17:12:20TiMiDrasher: kdiff :)
17:13:06rasherWith me using ubuntu/gnome
17:13:42rasheranyway, I think I'll just cut'n'paste a lot
17:18:03 Quit linuxstb (Read error: 110 (Connection timed out))
17:21:19linuxstb_There seem to be lots of places in the code where a call to lcd_?? is followed by an identical call to lcd_remote_??
17:22:32rasherI don't see how that can be avoided
17:23:08linuxstb_By having one function do both.
17:23:21rasheroh, an identical function
17:23:37rasheridentical call*
17:24:06linuxstb_Yes, the identical parameters. Not quite sure why that works, but it seems to be quite widespread in the patch.
17:24:13 Join Moos [0] (
17:24:30rasherwhy wouldn't it work?
17:25:28 Quit xen` (Read error: 110 (Connection timed out))
17:25:33TiMiDsince all lcd_ and lcd_remote have the same prototype, you can do a fn pointer that will call one or the other depending on your need
17:25:52TiMiD(that's how I decided to handle it)
17:27:12TiMiDwith this approach you can handle as may screens as you want :)
17:37:31 Join muesli- [0] (
17:38:13rasherSay, how likely is it that some guy will be called "Rubber Glove"?
17:57:55 Quit paugh ("Leaving")
18:03:17 Quit ashridah ("Leaving")
18:05:13 Quit muesli- (Read error: 113 (No route to host))
18:05:33 Join muesli- [0] (
18:10:05 Quit DangerousDan (Read error: 110 (Connection timed out))
18:16:05***Saving seen data "./dancer.seen"
18:16:57 Nick Lynx_ is now known as Lynx_awy (
18:18:53 Join dpassen1 [0] (
18:28:59 Quit muesli- (Read error: 110 (Connection timed out))
18:30:54fuzziestatus of AAC: libfaad ICEs when crosscompiling. more soon!
18:30:57fuzziei mean, gr. :P
18:31:19Mooscongrates :)
18:38:37 Join Sucka [0] (
18:38:51 Quit [IDC]Dragon ("CGI:IRC")
18:44:49 Join tvelocity [0] (
18:52:07 Join webguest47 [0] (
18:57:10 Join Lear [0] (
19:07:35rasherGood work on the gcc4 changes
19:08:02rasherAny ideas if it's noticably faster?
19:08:13rasherI guess it's hard to tell without a wav writer
19:08:17fuzziegcc4 changes?
19:08:20fuzziedoes it work with gcc4, now?
19:08:30fuzzieit seems silly trying to workaround ICEs if it does
19:08:35 Join muesli- [0] (
19:09:05rasherLear has done changes to make iriver compile with gcc4
19:09:14fuzziechecked in?
19:09:15LearI haven't really tested yet. Could play a Q10 ogg without skipping, that had had problems before, but it was a while ago I tested that. I was planning to check that file again with a GCC3 build.
19:09:43LearI haven't encountered ICE:s at least...
19:10:04rasherfuzzie: that is assuming that faad builds with gcc4-m68k
19:10:14rasherwhich may not be the case
19:10:19fuzziewell, it seems worth a try
19:10:24Learrasher: you working on that port?
19:10:28rasherNo, fuzzie is
19:10:31fuzzieit sure works with gcc4 x86/ppc
19:10:38fuzzieeven though that means nothing
19:10:44rasherI just test, since fuzzie doesn't have an iriver
19:10:47fuzziemy recent experiences with non-rockbox m68k toolchains have been < *
19:11:05 Join DangerousDan [0] (
19:11:16fuzzieeg, half of debian failing to compile
19:11:19rasherGcc4 for anything that isn't x86 seems half-baked
19:11:35rasherBut if it builds rockbox, it's worth a try
19:11:46rasherCan be made to build rockbox, that is
19:11:46fuzzieso, is this stuff checked in?
19:12:35rasherNot a lot of changes to be done, really
19:15:55 Join linuxstb [0] (
19:16:14linuxstbfuzzie: Which version of libfaad are you using?
19:16:25fuzziethe one you handed me
19:16:42rasherlinuxstb: which gcc do you use
19:16:44fuzziethe errors i was encountering might be issues with it being an older version or issues with me not initialising it properly right now
19:17:01linuxstbThe version I gave you compiles fine with 3.4.4
19:17:10fuzziewith an iRiver target?
19:17:16rasherWeird. That's what I use.
19:17:38rasherAnd I got the ICE
19:17:58linuxstbI'm using the CVS binutils from August 2005.
19:18:17linuxstbGNU assembler 2.16.91 20050813
19:18:41fuzziei guess i need to build a cvs version then
19:19:07linuxstbfuzzie: Did you change anything in the config.h ?
19:19:27fuzzieit looked okay as it was?
19:19:38fuzziefuzzie/libfaad.txt"> is all i've changed to make it build
19:19:51linuxstbWhich file does the ICE appear in?
19:20:06fuzziesbr_dec.c:611: internal compiler error: in reload_cse_simplify_operands, at post reload.c:391
19:20:39 Join webguest92 [0] (
19:21:08linuxstbDo you want me to try and compile it?
19:21:48*rasher builds a snapshot binutils
19:22:06 Quit pike (Read error: 110 (Connection timed out))
19:22:58fuzziemight be an idea. need to finish dinner/rl life stuff first, on the laptop right now
19:25:47fuzziethere isn't very much code of mine involved .. most of the actual work so far has been working out why the metadata of all my files failed [hence the read_uint32be fix] :/
19:26:09linuxstbfuzzie: Thanks for that.
19:26:23rasherYeah, wasn't my fix
19:26:24linuxstbI obviously never tested that function.
19:26:57fuzzierasher: well, you wrote it..
19:27:14rasherWriting it was the easy part
19:27:17linuxstbfuzzie: Do you deal with ALAC files, or have you disabled ALAC?
19:27:31fuzzieat the moment, i just changed alac to .alac
19:27:42linuxstbFair enough.
19:28:19linuxstbI only own one AAC file, so I guess I will do the opposite - until we fix it properly.
19:29:42 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
19:29:42 Quit paugh (Client Quit)
19:29:51 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
19:29:59fuzzieit ought to be simple to add detection to probe_file_format in a very inefficient manner
19:30:22linuxstbprobe_file_format isn't allowed to spin the disk up....
19:30:33fuzziei was about to ask if that'd be possible.
19:31:02amiconnrasher: binutils version shouldn't have any effect on ICEs
19:31:14*amiconn received 3 nice CDs today :)
19:31:19fuzziethe disk has to be spun up to examine the files, so i'm not sure how to handle that [merge alac/faad codecs to do different things depending on filetype?]
19:31:25fuzzieoh, also, is increasing CODEC_SIZE a problem?
19:31:30amiconnExpect some shiny new voice files sonn :)
19:31:39fuzziebecause i'm pretty sure libfaad is going to exceed the limit.
19:31:58fuzzieit apparently works on iriver, but i don't know the possible consequences
19:32:25linuxstbI would say to increase it for now, and that is something to look at when we start optimising it.
19:33:05 Quit muesli- (Read error: 110 (Connection timed out))
19:33:16amiconnFile type detection either has to be done at buffering time, i.e. as part of the metadata handling,
19:33:23 Join |Lupin| [0] (
19:33:37linuxstbI think the real fix for the ".m4a" problem is to call get_metadata() at a different point in time - i.e. before the codec is loaded, at the same time as buffering the file.
19:33:50|Lupin|Hello, there.
19:33:50amiconnOtherwise all files with the same extension would need to be handled by one (multiformat) codec
19:33:54linuxstbget_metadata() can then set the codec type.
19:34:09fuzziesounds reasonable, if it can be done
19:34:19amiconnI.e. a combined alac/aac codec for .m4a, a combined vorbis/speex codec for .ogg etc
19:34:35linuxstbI don't see why not - but I don't know the playback.c code well enough.
19:34:47amiconnI think this would soon become unmaintainable, if not impossible
19:34:56amiconnrasher: Ah yes, forgot about flac
19:35:03linuxstbCombining codecs will cause headaches when trying to use IRAM efficiently.
19:35:12amiconnFunny thing would be a universal .wav codec...
19:35:13linuxstbpreglow already wants to split the libmad codec
19:35:23linuxstb(I think)
19:35:25amiconnlinuxstb: Yes I know
19:35:44amiconnThat's why I'd prefer the first method (detection as part of the metadata handling)
19:35:53fuzziei think that's the only realistic option
19:36:00|Lupin|I' sorry if this is the wrong place, but I'm wondering if someone here could suggest a good mp3 player / recorder for a blind person ?
19:36:23t0masiriver H1x0 ?
19:36:38linuxstbDo you want a hard disk or flash based player?
19:36:40 Join muesli- [0] (
19:37:19|Lupin|linuxstb: Well, I have no idea of the consequences of such a choice...
19:37:58|Lupin|What I know is I'm looking for a very good quality recording, with a big capacity ifpossible.
19:38:47linuxstbIf you want big capacity, then it will have to be a hard disk based player. They are larger and more fragile than flash players.
19:39:18linuxstbThe iriver H1x0 is a good choice - if you installed Rockbox onto it.
19:39:33 Join vorbisgain [0] (
19:39:35|Lupin|ahah, I see. Thanks. What'sthemaximal capacity one can expect with a flash-based player ?
19:40:05|Lupin|linuxstb: and with rockbox it can speao, right ?
19:40:07vorbisgainhello all, is here some vorbis developer or someone in charge of the project?
19:40:24linuxstbTypically no more than 1GB or 2GB. But flash players are getting larger every week.
19:40:56amiconn...and some flash players are extendable with memory cards
19:41:26linuxstbBut price per gigabyte is much cheaper for hard disks.
19:42:01|Lupin|ok, thanks.
19:42:23vorbisgaintwo questions: first, how can I contact the rockbox staff? anyone knows?
19:42:25linuxstbAnd yes, the iriver H1x0 can speak when using Rockbox. But Rockbox is still in beta status for the iriver.
19:42:52vorbisgainsecond question: H340 doesn't understand replayagain tags, do you know a way to apply the gain to the data?
19:42:52rashervorbisgain: What do you want? There's no "rockbox staff" as such
19:43:16linuxstbAll the rockbox developers read the IRC logs, and some are around at the moment.
19:43:24vorbisgainrasher: I know there is no "staff" but well, I could perhaps offer help some way
19:43:49vorbisgainI have seen there is dire need of electronics for decoding hardware
19:44:01vorbisgainsorry, electronic engineers
19:44:09|Lupin|What's about the iRiver H340-SE ? Is itsupported by rockbox ?
19:44:23amiconnNot yet
19:44:25vorbisgainno, H340 is not supported
19:44:36rasher|Lupin|: Not right now, but it is planned. It will probably happen faster than the h1x0 port
19:44:36vorbisgainwell, I am teacher in a polytecnic school
19:44:52Moosohh :)
19:44:54vorbisgainI could have students do their projects on something
19:45:05vorbisgainthey are electronics students
19:45:16vorbisgainand they MUST do a final project
19:45:23vorbisgainwhy not for rockbox :)
19:45:42 Quit webguest92 ("CGI:IRC")
19:45:42linuxstbThey could disassemble MP3 players and draw schemetics and locate documentation for the chips....
19:46:03linuxstb... and then port rockbox.
19:46:07vorbisgainyes, that's the problem, dissasembling means necessarily destroying them? :)
19:46:25vorbisgainit means money thrown to the trash
19:46:59linuxstbYou could say it is money invested...
19:47:03vorbisgainha ha
19:47:22vorbisgainwell, I should check which student wants to "invest" :)
19:47:55vorbisgainup to my knowledge, rockbox is for players with a screen, isn't it?
19:48:07vorbisgainI mean, kind of H340 or iaudio X5
19:48:10linuxstbIt doesn't have to be - it has a voice UI.
19:48:13|Lupin|so is there one product in the iRiver series for which the support is "stable", with Hard-Drive, good microphone ?
19:48:25linuxstbBut all targets so far have had a screen.
19:48:39vorbisgaina voice UI??? you can talk to rockbox?
19:48:48Maxime`no, but rockbox talk to you
19:48:49rashervorbisgain: voice output
19:49:06vorbisgainwell, my H340 has a microphone, it could be!
19:49:07MoosLupin: h1xx
19:50:05vorbisgaindissasembling one 128M flash player is one thing (around 30euro) but a H340 is arond 400 euro, its another thing ;)
19:50:28|Lupin|It seems the only one proposed onthe website I am is the H10. If this is OK, that could be great.
19:50:55tvelocitywhy not port rockbox on the nintendo DS? :PPPpPPppp
19:51:00Maxime`or PSP
19:51:06vorbisgainwell, another question please
19:51:14MoosLupin: mp in french
19:51:19vorbisgainI am now involved in converting flac files into vorbis for my H340
19:51:29tvelocityPSP doesnt have a mic or a touch screen
19:51:32vorbisgainbut H340 doesn't understand replayagain tags
19:51:51vorbisgaindo you know how can I appy gain to the files in linux?
19:51:57|Lupin|Moos: mp ?
19:52:03rasherI don't think there's a way to do that easily with vorbis
19:52:09linuxstbI know that metaflac can do it for the FLAC files.
19:52:14MoosLupin: private msg
19:52:31vorbisgainin linux there are several tools: metaflac inserts replaygain tags
19:52:44vorbisgainthere is a normalize tool which has a gain parameter
19:52:57vorbisgainand a normalize-ogg counterpart
19:53:03vorbisgainor a normalize for wav
19:53:30vorbisgainI tried normalize using the gain, and I obtained a less audible file :(
19:57:59linuxstbHave you read the description of the algorithm at
20:01:54vorbisgainnot the algorigthm why?
20:04:28 Quit muesli- (Read error: 110 (Connection timed out))
20:04:53vorbisgainI think I made it well, I first calculated the gains with vorbisgain, and then I applied the gain with normalize.
20:05:12vorbisgainI am now calculating the replaygain again, and it is near zero in album mode
20:06:38vorbisgainwell, I think I made it, thanks anyway
20:06:43 Part vorbisgain
20:14:38 Quit paugh (Read error: 110 (Connection timed out))
20:15:51 Join XavierGr [0] (
20:16:07***Saving seen data "./dancer.seen"
20:22:13 Join linuxstb__ [0] (
20:22:28 Join tucoz [0] (
20:22:55tucozHi, how do I enable a debug build for the simulator?
20:23:19rasherPress S at the first question
20:23:22rasheror second
20:23:35tucoztype of build
20:23:59tucozand it is logf to send stuff to that?
20:24:35rasherlogf works in logf builds
20:24:45rashernot sure what that does on sim though
20:25:01tucozhmm, I think I have to take a look at an example
20:26:31amiconnIirc, simulators are always debug builds,
20:26:39amiconnso DEBUGF() should work
20:27:06tucozdebugf("output") or DEBUGF("output")
20:27:27tucozI have seen both by searching google for debugf
20:28:24Learlogf puts stuff in a buffer that can be viewed/saved from the debug menu.
20:28:44Learalso displayed on the remote, if the target has one (on the simulator too).
20:29:25amiconnDEBUGF() prints to the console (simulator)
20:29:35XavierGrUnfortunately the simulator is way too buggy.
20:29:45tucozok, great. Do I need to include debug.h for that?
20:29:49XavierGrSometimes it crashes while on the real target it works
20:30:05XavierGrAt least the windows version
20:31:07rasherThere's no debug menu in the sim
20:31:14rasherthat I can find
20:31:54rasherthere is however a logf dump menu
20:31:55Learit's directly on the info menu on the simulator.
20:32:08Learprovided you have a logf build...
20:32:47tucozok, when choosing a (D)ebug build in the config. Do I choose (D) or (S) then for DEBUGF output to show?
20:32:56tucozin the next step
20:32:58 Join yosemite [0] (
20:32:59tucozok, thanks Lear
20:33:05LearFor a simulator.
20:33:10linuxstb__Lear: Do you have a built with gcc4 that you can put somewhere for testing?
20:33:11amiconnrasher: There is no debug menu in the sim because all its entries don't make sense for the sim
20:33:14LearFor debug on target, you must have serial out...
20:33:14 Quit linuxstb ("CGI:IRC (Ping timeout)")
20:33:21amiconn...erm, for the archos sim
20:33:26yosemitehi, I just installed rockbox on my iriver this morning, and it now seems to have locked up hard
20:33:31amiconnPerhaps we should add it for iriver
20:33:44tucozyosemite: what is wrong?
20:33:49yosemitepressing reset or play seems to have no effect
20:34:09yosemiteit says I00: at 40d82000
20:34:17tucoztry holding play for a while
20:34:27yosemiteI believe it said illegal instruction before that
20:34:38yosemiteok holding
20:34:55yosemitehow long should I hold it
20:35:37tucozUntil it shuts down
20:35:50yosemitewell, it hasn't yet
20:36:20rasheryosemite: pressing reset will always reset
20:36:22tucozlike 10 seconds. if that doesnt work, insert a paperclip on the bottom
20:36:28 Join solexx_ [0] (
20:36:28rasheryosemite: Are you perhaps not hitting it right?
20:36:28amiconn40d82000? hmm....
20:36:41yosemiterasher: troll
20:36:45rasheryosemite: it's perfectly possible to miss the actual button inside the hole
20:36:50rasheryosemite: no, really!
20:36:53tucozin the small hole next to the usb connector
20:37:00yosemiteboth of those don't appear to be working right
20:37:08Learlinuxstb__: I have a build...
20:37:09yosemiteI'll try the reset again
20:37:14yosemitemaybe I don't have it lined up
20:37:14tucozthe reset is a hardware reset
20:37:32rasheryosemite: use a paper-clip or something.. you'll most likely miss if using a needle or similar
20:37:55 Quit linuxstb_ (Read error: 110 (Connection timed out))
20:37:57yosemiteok, I'm using a mechanical pencil
20:38:01yosemiteI'll find something bigger
20:38:16rasherNice design choice of iriver, this.
20:38:25 Join linuxstb [0] (
20:38:30amiconnrasher: oh yes
20:38:32rasherwhy not just have the button be *larger* than the hole, I ask
20:38:42amiconnI much prefer the archos method here
20:38:42 Part tucoz
20:38:59amiconn(having a hardware poweroff with a long timeout)
20:39:20rasherWell, if they had just used a real button instead of this dinky thing
20:39:31amiconnNo need for an extra button, just hold OFF for 10 seconds and it will power off regardless what the software is doing
20:41:30SlasheriNow the directory caching patch is available for a review:
20:42:11rasherand me without a clean sourcetree
20:42:19ghode|afkany chance of a build as well? :)
20:42:33fuzzierasher: it works apart from the linking?
20:42:45rasherfuzzie: Hm?
20:43:23fuzzieyour tree
20:43:52rasherwell it gets as far as the error I showed you
20:44:24rasheryosemite: See!
20:44:26yosemiteI got it jammed in there just right apparently
20:44:39rasherOf course, the crash is curious
20:44:42yosemiteit booted to iriver firmware
20:44:54rasherIt does that if it thinks something went wrong
20:44:54yosemiteI think it was trying to crossfade an mp3
20:45:19yosemitecan I just reboot back into rockbox then?
20:45:24rasherWell that certainly works. Must be more involved than that
20:46:30rasherCould be caused by the next file in the playlist being problematic
20:46:42yosemiteyeah that was my guess
20:46:48yosemiteI have it shuffling all files
20:46:58yosemiteand I know some of my files are a bit damaged
20:47:16rasherWith a bit of luck you should be able to resume the same playlist
20:47:21rasherand find out which is the next song
20:47:24XavierGrSlasheri: dircache? Sounds sweet when are we going to see this comitted?
20:48:22 Quit solexx (Read error: 110 (Connection timed out))
20:50:56 Quit linuxstb__ (Read error: 110 (Connection timed out))
20:51:30SlasheriXavierGr: i hope this week :)
20:51:43Slasheriunless somebody finds some bugs from the code
20:53:19 Quit Maxime` (Read error: 104 (Connection reset by peer))
20:53:23 Join Maxime` [0] (n=flemmard@
20:53:36 Join tucoz [0] (
20:54:17tucozSlasheri: how does it work? I select cache, it tells me to reboot, but when I check the menu after a reboot it says cache -> no
20:54:54XavierGrSlasheri: Sweet. Always you suprise us pleasantly.
20:55:10Slasheritucoz: weird. Don't shutdown the player before you have completely exited the menus
20:55:32Slasherirockbox may not save the settings if you are browsing the menu same time you shutdown the unit
20:55:36SlasheriXavierGr :)
20:56:03tucozok, now it's caching
20:56:12Slasheritucoz: and the first boot with cache enabled will be slow but future boots are fast
20:56:24tucozit took only like 3 seconds extra
20:56:35Slasherihehe, nice
20:56:47tucozhave to wait for the disk to spin down now
20:57:31tucozhehe, this is cool
20:57:49preglowLear: took the cowards way out in synth_full did you? ;)
20:58:18LearYeah, didn't feel like rewriting it all to assembler, which would be one option...
20:59:16LearOr do you have a better option? :)
20:59:22tucozSlasheri: seems to be working great. One issue though. It tells me Boot changed, reboot now when I move out of a folder
20:59:39XavierGrSo this patch will enable us to see a folder (say from exiting the WPS) without waiting for the disk to spin?
20:59:58Slasheritucoz: oh, can you find a way to reproduce that?
21:00:05Leartucoz: normal with a new build
21:00:08tucozI'll try
21:00:13tucozit's not a new buld
21:00:20Slasherigood :) that will be definately fixed before the commit
21:01:20XavierGrGod if this patch does what I am thinking it is "magnificent"! I was always annoyed by this little HD spin delay.
21:01:40SlasheriXavierGr: yes, it will remove completely the delay
21:01:53linuxstbSlasheri: What's the cost in terms of RAM to enable the cache?
21:01:57tucozSlasheri:Ok, this is reproducable. I boot, enter some subdirectories and move back out to the root. It tells me that boot has changed
21:02:02Slasheriand it's much faster also
21:02:24tucozLike, music->artist->album->disk1 and move back out
21:02:25Slasheritucoz: interesting.. that doesn't happen on my player. But i will try to do some tests soon
21:02:28preglowLear: no, no, that's the exact other option, the one i was hoping you'd take...
21:02:43XavierGrSweet I am going to include this patch on my build ASAP.
21:02:46Slasherilinuxstb: depends on how many files you have. With quite full player it's around 400 kB
21:02:48tucozI do not play anything. It is the first thing I do after a boot
21:03:08linuxstbHow is the memory allocated then? I thought it would have to be a fixed size.
21:04:38Slasherilinuxstb: on first (slow) boot the required ram size is calculated + some reserve buffer for cache live updating. Future rebuilds use the previous cache size as fixed buffer and transparent cache rebuild is possible
21:04:42tucozstill very cool Slasheri.
21:04:43Slasheriback soon ->
21:05:32 Quit tucoz ("CGI:IRC 0.5.4 (2004/01/29)")
21:05:43linuxstbSo the long delay first time you boot is to calculate the buffer size needed?
21:07:05preglowyes, and do some buffering, unless i'm mistaken
21:07:08*amiconn wonders what will happen if the reserved cache size is used up
21:07:38*linuxstb was wondering the same.
21:07:48amiconnPerhaps I should put this under stress test: a nice little plugin creatings thousands of files...
21:08:31amiconnI'm still not convinced at all that directory caching is a good idea
21:10:10preglowi'm not convinced it's the ultimate answer
21:10:13preglowbut i know two things:
21:10:27preglow2. the rest of the rockbox code wont end up more complex because of this
21:10:37preglow1. lots of people, me included, will use this
21:10:44preglowexcuse the reverse ordering
21:10:56preglowdespite it not being perfect
21:11:02amiconnFor one, it adds complexity
21:11:31amiconnThen it eats up RAM that would otherwise be used as audio file buffer
21:11:47rasherThat's negligeble out of a 32mb buffer
21:11:53rasheror 30mb or whatever we end up with
21:12:01 Quit linuxstb (Read error: 113 (No route to host))
21:12:20rasherIf the alternative is to keep the disk spun up the entire time, I'll sacrifice half a meg for that
21:12:38 Join linuxstb [0] (
21:13:08preglowamiconn: it's an option, i hope
21:13:25amiconnYes it is
21:13:28preglowamiconn: i'd happily sacrifice 400kb for my player
21:13:31preglowamiconn: i browse a lot
21:13:34amiconnThe complexity stays, however
21:13:59preglowamiconn: how? slasheri says the rest of the code isn't touched much, and all the caching is in one file
21:14:04amiconnFor me, it's one more option to disable immediately
21:14:15amiconn(like replaygain, ...)
21:14:16preglowamiconn: and for me it's not
21:14:36XavierGryeah I agree with preglow. I think it is a great feature.
21:15:09amiconnI'm browsing for minutes (if at all), then playing for hours...
21:15:39XavierGryes but doesn't this little dealy gets on your nerves every now and then?
21:15:39rasherSo it's different uses. Is it hard to imagine that one of them might benefit from caching?
21:15:53preglowamiconn: i'm not
21:15:57preglowamiconn: i browse a lot
21:16:00rasherOr should we just use the player like you do?
21:16:19XavierGrI remember how thoughtfull I was to browse on the iriver firmware.
21:16:37amiconnrasher: certainly not
21:16:42XavierGrif you browse then the HD would spin up untill the buffer was reloaded again.
21:17:06XavierGron the iriver firmware
21:17:11rasherWell, I don't see the last argument then.
21:17:23rasherAbout code-complexity, I know nothing.
21:17:25linuxstbIs the tag database cached in RAM?
21:17:40XavierGris there a difilter for the fmr (presets files)?
21:18:07amiconnNo, since they aren't supported for browsing
21:18:39amiconnBut now I'll check out the newly installed voices :)
21:18:50preglowwhat new voices?
21:19:12XavierGris there anyway to instruct the rockbox_browse function without the dirfilter argument?
21:19:13amiconn3 shiny new at&t natural voices to generate voice files with :)
21:21:07linuxstbXavierGr: You mean with SHOW_ALL ?
21:23:50XavierGrwell I mean to use a string filter for the time being. But as soon as the dirfilter argument is an integer I don't see how I can do this.
21:24:22Slasheriamiconn: when the reserve buffer size is consumed (i doubt this will rarely if ever happen), then the cache will be just disabled
21:24:23XavierGronly way is to use another SHOW_FMR filter, but I don't know how many devs will like this.
21:24:44amiconnSlasheri: And what will happen at the next boot then?
21:24:56Slasheriamiconn: a full (slow) cache rebuild
21:25:56linuxstbXavierGr: I don't see a problem with that.
21:26:37amiconnSlasheri: What will happen if the reserved cache buffer size is *almost* used up, and then the box is rebooted?
21:27:25amiconnWill it reserved the same amount of RAM as before, or will it adjust the size on every boot so that the reserved size is large enough?
21:28:32|Lupin|Sorry to interrupt, guys, but just one additional question:
21:29:10|Lupin|I'm about to order an iRiver H140. Once Rockbox will be installed on it, will I really be able to record voice thanks to the internal microphone ?
21:29:18|Lupin|And which quality of sound can I expect ?
21:32:44Slasheriamiconn: iirc the size allocated at next bootup will be the last cache size + reserve size
21:32:53XavierGryes. record is on its infancy in rockbox but you can make voice recordings.
21:33:04Slasheriso there should be always free space after boot up
21:33:17amiconnSlasheri: Is the reserved size a constant amount of entries, or a percentage of what is already there?
21:33:18rasherSounds sane
21:33:45Slasheriamiconn: currently it's just constant (64 KiB)
21:34:21amiconnHow many entries is that?
21:36:05|Lupin|XavierGr: What do you mean by "on its infancy" ?
21:36:30XavierGrwell on the iriver recording is not fully enabled.
21:36:50XavierGryou can record in wav with limited options
21:36:54|Lupin|XavierGr: What is missing, precisely, please ?
21:37:00XavierGralso you can record from the radio
21:37:16rasher|Lupin|: currently it only records to wav
21:37:24MoosLupin: you can use the original fw for this in waiting rockbox
21:37:27XavierGrOn archos models you can do pretty much everything.
21:38:04XavierGrPrerecording, timesplit and size split, quality, mp3 encoding, triggered recording e.t.c
21:39:04|Lupin|Then I wonder if I really should buy an iRiver
21:39:08|Lupin|Since I'm blind
21:39:22|Lupin|I guess the original firmware might not b very useful.
21:39:24XavierGrRockbox is perfect for blind users/
21:39:34|Lupin|Actually I thought rockbox woul do it...
21:39:47XavierGrbut it is a matter of time since all these will be implemented.
21:39:48amiconnXavierGr: Yes, but not the recording feature on iriver (yet)
21:40:01amiconnIt's a matter of time, definitely
21:40:03|Lupin|XavierGr: I'm a blind user planning to use the mp3 reader mainly as a _recorder_ ...
21:40:03rasher|Lupin|: recording will be possible through the same interface as on the "old" archos models. But lossy encoding will probably take a while
21:40:23Slasheriamiconn: that depends on the filename length but it's for sure hundreds of entries
21:40:23fuzziedidn't someone get shine working for encoding?
21:40:50Slasherinight, cu ->
21:40:52amiconnSlasheri: A cache entry isn't constant size?
21:41:01Slasheriamiconn: no
21:41:13Slasheriamiconn: only that much space is used as necessary
21:41:41Slasherithe cache is basically a linked list (take a look at the code)
21:41:42|Lupin|So what do you guys recommend to me, inthe end ?
21:41:50|Lupin|(regardingThe model I should buy...) ?
21:42:16XavierGrI don't know but do you have any alternatives?
21:42:33|Lupin|XavierGr: They are limited in France.
21:42:41XavierGrare there any other fullfledged blind userfriendly mp3 players?
21:42:52|Lupin|XavierGr: But I'd rather look first at which model would be the best, and then find a placeto buy it...
21:43:14*amiconn found a menu option that isn't voiced :(
21:43:16|Lupin|XavierGr: This is precisely what I don't know...
21:43:47|Lupin|btw guys: I'm myself a programmer. So if I can help with improving voicesupport, I'll do so with pleasure.
21:43:51XavierGramiconn do you know of mp3 players that support blind interface other than Rockbox ones?
21:44:07amiconnNo I don't but then I'm no expert on that
21:44:50amiconnI only know that rockbox is blind user friendly, and that it runs on a number of mp3 players which unfortunately aren't made anymore
21:44:52XavierGrI don't know either, so I think Lupin that still an iHP-100 (H-100) is your best chance
21:44:59amiconneBay is an option to get one
21:45:18rasher|Lupin|: for Rockbox, I'd say get an iriver h1x0 and wait for the kinks to be worked out. The problem is that these players aren't produced anymore. None of the players rockbox supports are. Unfortunately.
21:45:22XavierGrI think Jeff on has one for sale.
21:45:28amiconnRockbox on archos is more mature, but iriver support is catching up
21:46:21 Join muesli- [0] (
21:46:53amiconnThe archos recorders record directly to mp3 format, but otoh they don't support other formats than mp3 and mp2, the latter for playback only
21:47:18amiconnThat might change in the future, I hope to get wav support running
21:47:46amiconnThe iriver can play a whole bunch of different formats, but recording is only possible to wav format atm
21:48:06rasher|Lupin|: what are you planning to record?
21:48:06amiconn(and not in a blind friendly way, since it's still experimental)
21:49:42|Lupin|it's to record courses
21:49:56|Lupin|yoga sessions, for instance
21:50:54rasherI'm guessing the shine mp3 encoder will be perfectly fine for that
21:51:08|Lupin|what's that please ?
21:51:29XavierGrthis can be done currently, in wav format though and in a not so blind friendly way.
21:51:58|Lupin|Well, perhaps I can contribute in making this more blind-friendly ?
21:52:00rasher|Lupin|: Well, I was mostly talking to myself, it's the mp3 encoder software that is probably going to be used in Rockbox. It's not the best quality encoder, but for such things it'll be fine I guess
21:52:37|Lupin|well the encoding can certainly be done on a laptop, usinglame, say, no ?
21:52:38Learpreglow: almost sounds like you had tried to compile with gcc4... :)
21:53:21rasher|Lupin|: if recording to lossless and then using lame is fine, I see no reason why you shouldn't buy a used iriver-h1x0
21:53:35rasher|Lupin|: It's just not quite ready now, but will be in the not so distant future
21:54:28rasherI'm starting to think that recording to high-quality mp3 might be a problem
21:54:36|Lupin|rasher: Ok, so let's follow this path and I'll make my bestt contribute.
21:54:44|Lupin|That's an exciting task...
21:58:46yosemiterasher: an mp3 locked it up this time, less hard than last time though
21:59:04rasheryosemite: is it repeatable?
21:59:20yosemiteI'll find out in a second
22:00:18yosemitedid it again
22:00:59rasheranything special about it? weird tags? other strangeness?
22:01:01XavierGrok strange behaviour here. I call rockbox_browse to certain folder. All is fine excpet that I can't exit the browser. I call it exactly the way setting menus do and there you can exit with the left button.
22:01:35yosemitethere might be a tagging issue, I think I got the file from the interweb
22:01:37rasherI guess you need to get it to someone who knows stuff
22:01:49rasheryosemite: id3v2 -l on it?
22:02:17rasherI guess it could be broken in a multitude of ways
22:02:23yosemitehmm, no wait this was an ogg that I ripped
22:02:39rasherSlasheri: Dir cache feels great
22:02:43yosemitejust trying to play it doesn't work either
22:03:05rasherSounds terminally broken
22:03:13Learyosemite: chained ogg?
22:03:23yosemiteI dunno
22:03:31yosemiteis there an easy way to find out?
22:03:45Learwait, if you ripped it, you ought to know. :)
22:04:58yosemiteLear: it was a long time ago
22:05:08 Join Bagder [0] (
22:05:43yosemiteUser comments section follows...
22:05:44yosemite title=Korn Ring Finger
22:05:44yosemite artist=Captain Beefheart and His Magic Band
22:05:44DBUGEnqueued KICK yosemite
22:05:44yosemite album=Safe as Milk
22:05:44yosemite REPLAYGAIN_TRACK_PEAK=1.01288915
22:05:46yosemite REPLAYGAIN_TRACK_GAIN=-2.53 dB
22:05:47Learfor it to be chained, you need to concatenate two or more oggs...
22:05:55yosemitethere's only one stream
22:05:59 Join matsl [0] (
22:06:04rasherThat looks peachy
22:06:11rasherIs replaygain turned on?
22:06:41yosemiteah I thought that was unimplemented looking at the wiki
22:06:49rasherAnd if yes, does it lock up if you turn it off
22:07:11rasherWhere was that? (needs to be corrected)
22:07:13CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
22:07:13*Lear uses replaygain on oggs all the time...
22:07:30rasherLear: Yeah, me too.. it was all I could think of
22:08:15yosemiteI don't remember where I saw it now
22:08:20yosemiteno replaygain was off
22:08:34LearWell, with a copy of the file, there's at least the chance of it getting fixed...
22:09:12LearIf it is a problem in rockbox, and not in the file. :)
22:09:26rasherRockbox shouldn't crash
22:09:32rasherUnless it was simply a "codec failure"
22:09:38rasherThat's acceptable on broken files.
22:09:44rasherCan't do much better.
22:09:51yosemiteyou'd think it'd just skip it then right
22:10:15rasherHow did it crash?
22:10:18yosemiteyeah still not good
22:10:31yosemitewell now it has the drive light on hard and nothing is happening
22:10:36yosemitenothing responds
22:10:38rasherAh, well that's broken
22:10:45yosemitenow, it might not be rockbox
22:10:46rasherI don't know who you should send it to
22:10:51yosemitethis is an abused iriver
22:10:52Maxime`yosemite is your played fully chaged?
22:11:02rasherAh, hm.
22:11:17Learyou can send the file to me...
22:11:18yosemiteit's at 3/4 according to the little graphics bar
22:11:27yosemiteI'm going to try something first
22:11:33yosemiteI have a theory
22:11:47yosemitethat this may just be my disk brokeness
22:16:08***Saving seen data "./dancer.seen"
22:17:26yosemitehmm the iriver firmware seems to handle the file fine
22:22:54yosemitewhat is the format of the .playlist_control file?
22:23:43yosemitenm looks like I'm safe just nuking it
22:25:39 Quit Moos ("Glory to Rockbox")
22:30:19yosemiteyeah it's the file's fault
22:31:53yosemiteLear: <−− here's what breaks it
22:32:21LearThanks, I'll give it a go.
22:33:54 Quit thegeek (Read error: 104 (Connection reset by peer))
22:36:27 Join thegeek [0] (
22:41:47 Join thegeek_ [0] (
22:41:47 Quit thegeek (Read error: 104 (Connection reset by peer))
22:41:48 Quit thegeek_ (Read error: 104 (Connection reset by peer))
22:43:46 Quit matsl (Remote closed the connection)
22:45:35 Quit |Lupin| ("'dnight, all")
22:47:08LearCrashes the simulator at least...
22:48:40amiconnAh, no heck, just some missing voice: entries in deutsch.lang
22:49:10 Quit muesli- (Read error: 110 (Connection timed out))
22:50:29LearHm.. Any easy way to map a backtrace address to an address in for the simulator?
22:51:17LearHm... Wait, looks like that address is allocated; some addresses do match...
23:00:10linuxstbLear: I think that Ogg file crashes somewhere in the Codec. get_metadata() parses the file OK.
23:00:42LearIndeed. Codec setup is okay, it's in tremor, invalid access (0) in simulator...
23:01:02LearAnd yes, simulator runs out of memory.
23:01:26rasherThe vorbis codec doesn't work in sim (too big)
23:01:41Learworks quite well, thank you very much... :)
23:01:51rasherOh.. crashed for me yesterday
23:02:06*Lear listens to 48 khz ogg in simulator...
23:02:22XavierGrhmm just created a fmr filetype is that ok?
23:02:51XavierGrnow I must write the actual code that will initialize the radio and load the presets from the file...
23:03:15LearThis file is a little unusual (I think), in that it uses floor0.c...
23:04:03rasherold encoder?
23:05:10rasherWe open the real file 'archos/.rockbox/codecs/vorbis.codec'
23:05:10rasherSegmentation fault
23:05:22rasherThat's what happens when I try to play anything vorbis
23:05:57LearCould be... Looks like the problem I see here. Tried the same file on target?
23:06:11rasherI don't have the file
23:06:14rasherthis is a random vorbis file
23:06:23Learthat's the file I was refering to...
23:06:41rasherNow I'm confused.
23:06:59fuzzieall my vorbis files die in the simulator
23:07:05fuzziebut, i have no real hardware to test them on
23:07:17fuzzieit gets further if you increase CODEC_SIZE [because the simulator vorbis.codec exceeds the default]
23:07:17rasherthis file _works_ on target, but not in the sim
23:07:21fuzziebut, still dies
23:07:46rasherI wonder how Lear is listenint to oggs in sim
23:07:57LearOuch, that file allocates least 400 kB, if my (very) quick calculations are right...
23:09:30*amiconn wonders why there are separate settings for crossfade and crossfade duration
23:09:57rasherHow else would it work?
23:10:00amiconnImho this could be simplified by having a crossfade duration of zero
23:10:06elinenbequick question...
23:10:12rasherThere are two crossfade types
23:10:37elinenbedoes anyone here know who is responsible for this?
23:10:50elinenbeit looks good.
23:10:58 Join Paul_The_Nerd [0] (
23:11:09amiconnrasher: Ah ok. Didn't think of it as I never use crossfade...
23:12:23rasherOne fades both in and out, the other only fades the new song in
23:12:30rasherand lets the current song play
23:12:57LearOh well, time to leave for me. Maybe "real" memory allocation functions could make this one work; need to know more about the allocs to tell...
23:13:39elinenberasher: any idea on that link?
23:13:45rasherNot at all
23:15:20elinenbeafter watching the video it looks like it uses a buffer overrun error on the "get info" function for an unplayable mp3.
23:15:48elinenbe... an mp3 that only has the code for the sms emulator
23:16:32 Quit Lear ("Chatzilla [Firefox 1.4/undefined]")
23:17:51 Join muesli- [0] (
23:17:59linuxstbThe alloca implementation seems broken for the sim - it allocates from the main memory buffer, not the stack. Or am I misunderstanding how alloca should work?
23:20:33 Join webguest39 [0] (
23:21:53XavierGrhmm the patch utility of the rockbox devkit isn't working!!!
23:22:11rasherWhat's the problem?
23:22:39XavierGrpatching file apps/filetree.c
23:22:39XavierGrAssertion failed: hunk, file ../patch-2.5.9-src/patch.c, line 340
23:22:39XavierGrThis application has requested the Runtime to terminate it in an unusual way.
23:22:39DBUGEnqueued KICK XavierGr
23:22:39XavierGrPlease contact the application's support team for more information.
23:23:12 Quit gromit` (Read error: 104 (Connection reset by peer))
23:23:13XavierGrand I got that same error with one of my patches so...
23:23:33XavierGrI tried other -p handlers too but then it asks for the files manually.
23:24:01XavierGrcrap and I needed that dircache build.
23:24:13rasherI have one
23:24:24 Quit webguest39 (Client Quit)
23:24:25XavierGrI was just going to ask you that :)
23:24:25rasherif you just want a clean dircache build?
23:24:40XavierGryes please!
23:25:27 Join gromit` [0] (n=gromit`">
23:25:55XavierGrthanks rasher! :D
23:29:00XavierGrany option to enable it or it should work right away?
23:29:28preglowlinuxstb: you're perfectly right
23:29:33preglowlinuxstb: alloca allocates from the stack
23:29:47preglowlinuxstb: freeing is automatic like that
23:29:49XavierGrnevermind foundit
23:29:59preglowlinuxstb: i'
23:30:18preglowlinuxstb: i've already noted that that is a bug, but i guess lear had a good reason for doing like it is
23:30:44XavierGrsweet!!! it is working!
23:31:10XavierGrNow I have to go... thanks rasher for the build and Slasheri for the implementation of it.
23:31:12 Quit XavierGr ()
23:33:39CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
23:33:39*rasher adds dircache to his list of patches
23:35:14preglowi think this is great
23:35:28preglowit's even frigging faster than the iriver cache
23:36:03rasherit boots relatively fast
23:36:16rasherEven when it has to rebuild, that is
23:36:50preglowit just gave me a false positive for a rockbox.iriver change
23:37:01rasherYeah, someone else had that
23:37:03preglowthe background rebuild ran, then suddenly it asked if i wanted to reboot
23:37:46amiconnTalking about complexity....
23:38:01Paul_The_NerdHow does dircache handle when you create a new file of any sort?
23:38:08rasherIt updates the cache
23:38:17rasherAt least, that's my understanding.
23:38:22rasherIt just adds the file
23:38:46rasheramiconn: most new code has bugs
23:38:50Paul_The_NerdAlso, did you ever resolve the "what happens when the cache gets close to full" question? Didn't see it in the log but I may have missed it.
23:39:31amiconnrasher: The thing is that it influences an area that isn't even touched by the new code, merely by doing something in the background that isn't expected
23:40:41rasherPaul_The_Nerd: I think the answer was that Slasheri was unsure, but pretty certain that the extra-buffer would be the same size (ie, total buffer would be larger) on next reboot
23:41:13amiconnWho messed up all the voice: lines in english.lang? :/
23:41:17preglowi take it the cache routines consult the actual fat table when the cache isnt readily built?
23:41:23preglowi dont see how bugs can be triggered if this is the case
23:42:10amiconnWell, not all lines, but quite a number...
23:42:48rasheramiconn: Not me!
23:43:01rasher :-\
23:43:31rasheroooh, the mp3 encoder as a plugin
23:44:23amiconnrasher: Okay, so I have to blame HCl, Lear and Slasheri
23:44:44preglowin one file
23:45:42preglowoh well, kudos to him, it's even emac optimised
23:45:44rasherI'm hurt that he suggests overwriting my plugin
23:45:46preglowin _integer_ mode, no less
23:47:45*rasher tries it out on iriver.wav
23:47:53rasher(yes, *that* one)
23:48:35preglowthe only one
23:48:40 Join thegeek [0] (
23:48:40*ghode|afk likes the new dircache feature
23:48:55preglowyes, get rid of the rockbox.iriver false positive and i'm game
23:49:38ghode|afkdo you have show all files on?
23:51:51 Join tucoz [0] (
23:53:30tucozI am doing something wrong. If I use fdprintf, what format should I use to get for instance 122.444:blablabla where 122 and 444 are separate integers
23:54:03tucozthat is integer dot integer ...
23:54:35preglow%d.%d ?
23:54:44tucozNo, I get an error then
23:55:01tucozBut I thought that it was that format
23:55:29preglowwell, i think it is
23:56:28 Quit tucoz (Client Quit)
23:56:33 Join tucoz [0] (
23:56:55rasherAnd again, tucoz' browser has an argument with the webclient
23:57:04preglowit's losing
23:57:30tucozI guess, but %d.%d:%s produces a compiler error (syntax error before ':' token)
23:57:39preglow_compiler_ error?
23:57:56 Join arkascha [0] (
23:57:57rasherSounds like you have something else broken to me
23:58:22preglowgcc's been drinking again
23:58:32preglowlet me see the actual line
23:58:33tucozah, I mean I have done something wrong.
23:58:46tucozand the compiler tells me so :)
23:58:53tucozrecorder/radio.c:648: error: syntax error before ':' token

Previous day | Next day