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

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

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

#rockbox log for 2005-11-07

00:00:49linuxstbI think I understand the intention of the folder navigation option now - pressing next/prev on the last/first track will take you to the next/previous folder.
00:00:50 Join muesli_- [0] (
00:01:22linuxstbSo I think the patch makes sense. Only crontroversy is the change of the default behaviour.
00:02:35jochen_well, thank you all for your help and of course for the great work you have done with rockbox. it's just amazing!
00:03:28linuxstbGlad it was nothing serious.
00:03:39jochen_yeah, me too. *g*
00:04:26jochen_just the typical "well, i guess it SHOULD work" errors. i love computer science...
00:05:45 Join webguest20 [0] (
00:07:28jochen_gtg, thanks again && gn8
00:08:00 Quit jochen_ ("Verlassend")
00:10:18 Quit webguest20 (Client Quit)
00:12:39amiconnlinuxstb: Yes, (1) it changes the default behaviour and (2) it is hackish
00:13:05amiconn(considering JAJDude's problem)
00:16:32 Part matsl ("ERC Version 5.0 (CVS) $Revision: 1.776 $ (IRC client for Emacs)")
00:17:18 Join matsl [0] (
00:18:07 Part matsl ("ERC Version 5.0 (CVS) $Revision: 1.776 $ (IRC client for Emacs)")
00:18:13 Quit muesli- (Read error: 110 (Connection timed out))
00:19:19preglowok, if i clean away all the sv4-6 cruft, remove the read buffer and sell my soul to satan, i _MIGHT_ be able to fit the entire mpc_decoder struct in iram
00:19:47preglowi don't think it's more than two months ago i swore i'd never touch libmusepack again
00:20:06preglowlinuxstb: how much iram does the ipod have again?
00:20:15Moospreglow: :D thanks for us
00:20:29linuxstbIt's exactly the same as the H1x0 - 96KB IRAM and 32MB DRAM
00:20:44preglowMoos: btw, the file you mailed me _still_ has got id3v1
00:20:49preglowMoos: that's why ape tags doesn't work for you
00:21:13Moosyes I discovered it, and retaged all my collection in APEv2 only
00:21:19preglowMoos: goodie
00:21:24preglowMoos: that should work just fine
00:21:31Moosyes it is
00:21:40Moosthanks for your great works
00:21:47preglowyou're welcome
00:22:33MoosI'm still wondering about the Fredy APE code?
00:22:34 Join ashridah [0] (
00:22:48preglowMoos: the apev2 for mp3 patch?
00:23:04Moosdid you see his mail in the mailing list?
00:23:12preglowyes, i've known about his patch for a long time
00:23:24Moosit doesn't sound good?
00:23:59MoosI think to user using APE yag for mp3 for mp3 gain thing
00:24:32MoosI need to sleep :)
00:24:46preglowsure, it's just that it requires furhter configuration options
00:25:02preglowcurrently we have an option whether you want id3v1 or id3v2 to be used first by rockbox
00:25:16preglowto be consistent, we now need an ape tag option here as well
00:27:28preglowtime for the daily roger whittaker singalong and coldfire assembler coding festivities
00:27:29MoosI how sounds musepack songs, and now the cpu boost is goodie, less than vorbis :), and if you can fit all mpc decoder in iram...
00:30:52preglowi can make musepack faster than it is, sure
00:31:22Moosplease will do :)
00:31:23preglowi'm quite convinced it could be optimised to never boost, but i wont go into that
00:31:38Mooslaborious works?
00:40:11preglowi have no idea what's wrong with this thing
00:40:19preglowit outputs weird glitching static
00:41:55Moostime to sleep, have a good night preglow and at all
00:42:01Moosc you later
00:42:13 Quit Moos ("Glory to Rockbox")
00:50:40preglowthe fun thing about unrolled code is that you need to fix the same code quite a number of times for every bug you find
00:53:14amiconnThe debug first, the unroll ;)
00:53:27*amiconn can't type
00:54:02 Join blacksun [0] (
00:54:22preglowor perhaps there's another way of doing this without unrolling
00:54:26preglowonly adding a branch
00:54:53***Saving seen data "./dancer.seen"
00:55:00preglowalmost like what i do in the new default handler
00:55:15preglowamiconn: what do you guys feel about self modifying code? :P
00:56:36amiconnThere's a method that I'm using in the grayscale lib scroll-up/down-by-n routines for sh1
00:56:50preglowwhich file?
00:57:25amiconnI'm calculating a jump into a chain of shifts (the SH1 doesn't have shift-by-n) once before the loop, then use this in the loop
00:57:36preglowwell, yes, exactly that
00:57:54preglowthat's what i did in the default handler
00:58:06amiconnThat's not self-modifying code... so it's not considered dirrrty
00:58:09preglowthe same can be done here, but with a couple of caveats: an extra branch, and one less register to use for ceofs
00:58:32preglowamiconn: sure, i know, i considered fixing the offset in the lea.l extension words :PP
00:58:38preglowamiconn: so i could save a register, hehe
00:58:48preglowbut that would be dirty
00:59:25amiconnapps/plugins/lib/gray_scroll.c lines 262ff and 447ff
01:00:23amiconnIf you try to understand this code, please note that unconditional branches, jmps, jsrs etc on the SH1 are delayed instructions
01:00:44preglowand how deep is the pipeline?
01:00:51amiconnOne delay slot
01:01:14amiconnThe SH1 is a nice little RISC CPU
01:01:38preglowi've seen dsps that do the same thing for branches
01:02:16amiconnThe SH2 and up have that for conditional branches as well (optional)
01:03:32preglowyou can do quite a few nifty tricks with that
01:03:43amiconnAn SH-DSP also exists, btw
01:05:03amiconnThe delay sot on SH is there solely to cut execution times. It's not needed for flag checking or such. The SH1 never sets the flags implicitly
01:05:12 Nick ashridah is now known as Lost-ash (
01:05:43amiconnYou always have to use explicit cmp or tst instructions
01:05:46XavierGrI can't get the bloody crossfader to work, maybe I am just stupid...
01:06:02preglowXavierGr: like slasheri says, there are some small bugs left
01:06:02 Quit blacksun ("CGI:IRC (EOF)")
01:06:28preglowamiconn: it's usually there do avoid having to flush the pipeline, no?
01:07:13amiconnYes. The slot instruction is already loaded into the pipeline when the branch is decoded
01:07:30amiconnIt saves one cycle on SH1.
01:08:32preglowcould you just have a quick look at the order8 case and see if i'm completely off the bat with my shifting here?
01:09:39 Quit linuxstb ("CGI:IRC")
01:10:45preglowi've got to be blind or something
01:12:21XavierGrHmm I think that having 4 arguments to set for the crossfade option is quite confusing, I will wait a little more for Slasheri's bugfixes before I test it more....
01:13:20preglowXavierGr: just don't touch the delay options
01:14:13XavierGrbut if both delays are set to 0 no crossfade is happening.
01:14:23XavierGrcould you tell me your settings?
01:15:12preglowi don't use crossfade
01:15:24amiconnpreglow: It seems you forgot that the mac output in fractional mode is bits 62..31, not 63..32
01:15:51amiconnApart from that you're wasting registers
01:16:09XavierGrI will try with a clean build.
01:16:51amiconn(1) You can set macsr before loading the initial %d5 -> possible to reuse a reg
01:17:20amiconn(2) You can delay readout of %acc0 until after the second mac.l sequence
01:18:05preglowamiconn: i know i'm wasting registers, i'm currently just trying to make correct sound
01:18:29preglowamiconn: have i forgotten that?
01:18:34amiconnMy first guess is the mac fract thing
01:18:47preglowamiconn: bear in mind that the qlevel parameter had one added to it when the function is called, btw
01:19:18preglowof course, that's in decoder.c, not the asm file
01:19:22amiconnThen that's different from the 16 bit case
01:19:36amiconn..but it doesn't change the point
01:20:04preglowbtw, it's not bits 62 to 31 you're fetching
01:20:19preglowall mac operations are like this acc += (a*b) << 1
01:20:37preglowas far as i understand it, at least
01:20:51amiconnYou *are* fetching bits 62 to 31
01:20:57amiconn...because of the shift
01:21:07amiconnWithout it it would be bits 63..32
01:21:14preglowwell, sure, but does it matter?
01:21:19amiconnYes of course
01:21:23preglowi compensate for the extra shift by adding one to qlevel
01:21:39preglowin decoder.c
01:21:52amiconnThen that's totally wrong
01:21:52preglow lpc_decode_wide_emac(s->blocksize - pred_order, qlevel + 1, pred_order,
01:21:52preglow decoded + pred_order, coeffs);
01:22:07preglowthen there's my error :)
01:22:24amiconnThe fractional chain yields bits 62..31, and the integer chain yields bits 31..0
01:22:34amiconn(yes, bit 31 is calculated twice)
01:22:49preglowyes, i get it now
01:22:56preglowjust ignore me
01:23:05amiconnRemove the +1 from the call, and change the moveq.l #32,%d4 into moveq.l #31,%d4
01:24:18amiconn(and I was probably wrong about the %acc0 readout - most likely it won't work right after switching modes)
01:24:32preglowamiconn: that's why i'm doing it this way right now
01:24:39preglowamiconn: i've yet to find out if that works
01:24:53preglowi think i did an experiment once and found out it didn't work
01:25:32preglowstill no dice, btw
01:25:42preglowthere's some nice scraping static still
01:28:31preglowi wonder if setting unsigned mode even matters
01:28:38preglowi think it's only an issue if you use saturation
01:32:07XavierGrin which function in settings.c, the settings are physically written on disc?
01:32:33amiconnIiuc the second lea.l (-7*4,%a0),%a0 should read lea.l (-6*4,%a0),%a0
01:32:33preglowbut this can't be right, if i specify a zero shift now, there'll still be one bit left in the top bits register when i or it with the lower bits register, which contains all 32 bits still
01:33:37amiconnThe high/low overlapping by one is correct, believe me
01:34:10amiconnAll our combined multiplies work that way, and the work correctly (musepack, mandelbrot, ...)
01:34:11preglowamiconn: no, i'm quite sure -7*4 is correct
01:34:25amiconnThen why is that -6*4 in the 16 bit case?
01:36:11preglowjust tried it with -6, unit hangs
01:37:05preglowamiconn: because i preload one more coef in the coldfire.S case
01:37:15preglowamiconn: i don't do that here
01:37:34preglowthe last mac.l loads the -7*4 one
01:38:12preglowthough i've got enough registers to do that here as well now
01:38:19preglowdepends on how many orders i want to unroll
01:42:04XavierGrSlasheri: The crossfade works very good when tracks change normally (without changing tracks, just wait to finish) but the mix option doesn't crossfade right when I force it to load a new track.
01:44:17XavierGrWhile the crossfade happens (in mix mode) the last track will not "fade out" it will just play as long as the fade out duration is and then stop without first lowering its volume.
01:44:46 Quit actionshrimp ("a bird in the bush is worth two in your house")
01:52:11amiconnHmm, I checked again, I can't find the mistake (apart from the 32->31) :(
01:52:37preglowyes, i see that's correct anyway
01:52:53preglowthe lower and upper bit will be the same if there's no shift, and the or.l will work just fine
01:53:46preglowi'm starting to wonder if there's some emac subtlety going on here
01:57:00preglowperhaps i'll just try doing a mulu version, despite it probably going to be slower
02:00:15XavierGrif I want to append data to a file which open flag is wise to use?
02:01:48preglowO_APPEND? :P
02:02:26XavierGrnice, I just wasn't sure about it.
02:02:52XavierGrwhat O_RDWR does?
02:03:52preglowenable both reading and writing
02:04:14XavierGrthen I will nead this too
02:05:13XavierGrcan I add 3 flags on the open fn? Like open("/battery.txt",O_RDWR | O_APPEND | O_CREATE);
02:07:35preglownice, now i get an illinstr.........
02:07:51 Quit muesli_- (Read error: 110 (Connection timed out))
02:11:17preglowand the offending instruction is a bne.w
02:15:38 Quit ]RowaN[ ()
02:17:18preglowthis is just too blood far fatched
02:17:24preglowbloody at that
02:19:36 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
02:27:15preglowthe reference clearly states the word forms of the Bxx instructions are supported
02:27:29preglowonly the long forms aren't supported in the v2 cores
02:28:02XavierGrWhich will be the easiset way to copy data from a text file an excel graph (2 colums) x - y maybe and then change line?
02:28:35XavierGre.g I want to make an txt archive with the runtime next to remaining battery and then I want to make a graph in excel with those values
02:28:57XavierGrso the best way to print them on the txt is runtime - battery remaining?
02:30:10preglowabsolutely no idea
02:30:19preglowi'm used to doing that sort of thing with matlab
02:30:55XavierGrnothing trivial just asking to make sure I will follow the easiest way.
02:31:33XavierGrI am making this to help the process of bencmarking batteries and make sure that the battery graph is correct.
02:31:37preglowi don't even have excel, heh
02:32:17preglowbut anywho, i need to go to bed
02:32:46XavierGrThat way a txt file should hold last succesfull buffering process withe total runtime and battery changes (estimate) as time changes
02:45:52amiconnXavierGr: The physical settings flush to disk doesn't happen in settings.c. As you might know the settings aren't stored in a file, but in a special sector on disk. This sector is buffered in the ata driver. The ata driver itself flushes this sector to disk if it's got pending changes and another disk access happens
02:51:08 Join lostlogic [0] (
02:54:15XavierGrbut even then I just create a file when settings_save() is called which writes to a physical file remaining battery and current runtime.
02:54:44XavierGrSo as I think of it I don't need the write process of the writing settings on the disk
02:54:56***Saving seen data "./dancer.seen"
02:56:00XavierGrThe txt file will hold remaining battery and current runtime on the last HD successfull write.
02:56:30XavierGrso it will have 0>=30 minutes error.
02:56:48XavierGrI mean when you make a benchmark
03:01:43amiconnIf you write every time settings_save() is called, this will severely affect the runtime which you are trying to log...
03:02:16XavierGrhmm thats a valid point
03:02:40XavierGri will have to do that then on the ata code.
03:02:46amiconnYou can write when the playback engine is rebuffering
03:03:18XavierGryeah that's a good option too.
03:03:34XavierGrwhere that is happening playback.c?
03:04:05amiconnYes it's in playback.c for iriver
03:04:30XavierGrand mpeg.c for archos i think
03:05:08XavierGrbut anyway these will be an optional patch that wants to benchmark his battery.
03:05:30XavierGrand more or less battery issues on archos are known quite well.
03:05:56XavierGrwhat about a plug-in.
03:06:07XavierGrThere was something like that wasn't?
03:08:11amiconnYes there is, but it only simulates activity
03:08:30amiconnThis could in fact be done way better with the tsr plugin mechanism
03:09:05amiconnOne would just load this test plugin, then start playing whatever he wants. The tsr plugin would log the data in the background
03:09:25amiconnThe only situation where it would have to stop is when the user starts another plugin
03:09:45XavierGryes i think i could make that
03:10:21amiconnThat's a plugin that returns to the main thread, but installs an additional thread before. It needs a proper termination function, otherwise loading another plugin would make rockbox crash badly
03:11:42XavierGris it difficult to code?
03:12:02XavierGrsuch plug-in i mean
03:14:32 Join DJDD [0] (
03:29:11 Part Riot-
04:54:07 Join Midgey34 [0] (
04:54:58***Saving seen data "./dancer.seen"
05:02:31 Quit RotAtoR ()
05:02:41 Part Midgey34
05:18:57 Nick Lost-ash is now known as ashridah (
06:07:35 Join _user_ [0] (
06:07:59_user_phaedrus961: the unicode patch doesn't work on the cvs build
06:08:12 Join adiamas [0] (
06:13:50_user_patching file firmware/drivers/lcd-h100-remote.
06:14:05_user_Hunk #1 FAILED at 29.
06:14:10_user_Hunk #2 FAILED at 958.
06:14:17_user_Hunk #3 succeeded at 978 (offset 4 lines).
06:14:23_user_Hunk #4 succeeded at 986 (offset 4 lines).
06:14:27_user_Hunk #5 succeeded at 1022 (offset 4 lines).
06:14:31_user_Hunk #6 succeeded at 1125 (offset 4 lines).
06:14:50_user_2 out of 6 hunks FAILED −− saving rejects to file firmware/drivers/lcd-h100-remote.c.rej
06:18:41 Quit _user_ ("CGI:IRC (EOF)")
06:18:53 Join Lost-ash [0] (
06:39:36 Quit ashridah (Read error: 110 (Connection timed out))
06:54:57 Part adiamas
06:54:59***Saving seen data "./dancer.seen"
07:29:41 Join DJDD_ [0] (
07:29:41 Quit Mxm`Pas`Bien (Read error: 104 (Connection reset by peer))
07:30:03DJDD_Anyone around?
07:31:19DJDD_Are H120's identical to H140s apart from look and HDD size?
07:34:49 Join ender` [0] (i=ychat@
07:36:03 Quit ender` (Client Quit)
07:38:10 Quit DJDD (Read error: 110 (Connection timed out))
07:38:55 Join ender` [0] (i=ychat@
07:45:52 Join B4gder [0] (
07:47:39Lost-ashDJDD_: no, the 120's slightly thinner
07:47:57Lost-ashwhich is why 120's can't take the same disk drives as 140's
07:49:01DJDD_Right, so the HDD in my 140 wouldnt fit in a 120?
07:49:52Lost-ashthat's correct
07:49:55 Nick Lost-ash is now known as ashridah (
07:50:21DJDD_I found 2 ppl willing you hand over their dead in 2 days, both 120s
07:50:37DJDD_Oh wait
07:50:45DJDD_but i dont plan you use a 120 case
07:51:00DJDD_Is the circuitry IDENTICAL?
07:51:15ashridahi have no idea
07:51:17DJDD_eg, the mainboard, LCD, placement of buttons
07:51:35DJDD_hrm, ok. Linus would know wouldn't he?
07:51:35ashridahpossibly, one of the devs would probably know better
07:51:57DJDD_thx anyway
07:58:12 Join einhirn [0] (
08:23:26 Join solexx_ [0] (
08:36:03 Quit solexx (Read error: 110 (Connection timed out))
08:48:17 Join _FireFly_ [0] (
08:55:03***Saving seen data "./dancer.seen"
08:55:25 Join Zagor [0] (n=bjst@pdpc/supporter/sustaining/Zagor)
08:55:33B4gdermorning Z
09:03:15B4gder <= neat project
09:07:53 Nick Lynx_awy is now known as Lynx_ (
09:44:50 Join DJDD [0] (
09:45:41 Join webguest72 [0] (
09:46:46 Quit webguest72 (Client Quit)
10:04:06 Quit DJDD_ (Read error: 110 (Connection timed out))
10:16:12 Join linuxstb__ [0] (
10:18:07 Quit linuxstb_ (Read error: 110 (Connection timed out))
10:21:04 Nick linuxstb__ is now known as linuxstb (
10:38:12 Join Jungti1234 [0] (n=jungti12@
10:41:13Jungti1234What is correct meaning of 'Top-OFF Chg'?
10:55:04***Saving seen data "./dancer.seen"
10:58:19 Join Cassandra [0] (
10:59:00B4gderhowdy ho Cassandra
10:59:11CassandraHas anyone tried loading images in the WPS from a directory other than /.rockbox/? I can't seem to get it to work.
10:59:26CassandraNor can I see anything wrong in the source.
11:02:28CassandraIncidentally, developer builds with the debug option enabled don't seem to be compiling. Has it been superceded by logf?
11:06:14B4gderyes, the debug builds are/were mostly for using gdb
11:07:15CassandraOK - that's fine.
11:07:50CassandraAppears %x is working but %xl not always.
11:18:39 Quit _FireFly_ ("Now if you will excuse me, I have a giant ball of oil to throw out my window")
11:19:17 Join linuxstb_ [0] (
11:26:55Jungti1234Ask manual link about rockbox firmware function.
11:33:51 Join linuxstb__ [0] (
11:34:32 Quit linuxstb (Read error: 110 (Connection timed out))
11:36:55 Join _FireFly_ [0] (
11:38:59 Nick linuxstb__ is now known as linuxstb (
11:44:44 Join ender1 [0] (i=ychat@
11:48:02 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
11:50:23Jungti1234Is no one on seat now?
11:50:34 Quit linuxstb_ (Read error: 110 (Connection timed out))
11:51:06Jungti1234I'm as want to ask a question some.
11:54:09Slasheramiconn: Hmm, am i correct that i could erase flash area 0x000009 - 0x0effff without causing rockbox firmware boot to fail?
11:56:20SlasherI just want to be sure before trying anything as i don't have a bdm yet :)
11:57:47linuxstbSlasher: What are you planning ?
11:59:05amiconnSlasher: I don't know. You should ask Linus about bootloader & stuff
11:59:22 Quit ender` (Read error: 110 (Connection timed out))
12:00:36Slasheramiconn: ah, ok :)
12:01:10Slasherlinuxstb: Just trying something interesting ;)
12:02:59Slasherlinuxstb: I will try removing the original completely from my player and then write some other data to the flash
12:03:39CassandraWhat other data?
12:04:12ashridahJungti1234: we can try. what's the question?
12:04:38amiconnSomeone should try to read the serial eeprom...
12:04:59SlasherCassandra: rockbox firmware for example :)
12:05:54SlasherCassandra: and the little patching to the bootloader so it could also load the firmware from the flash instead of from disk only
12:07:39Slasheramiconn: is there a schematics done about the eeprom chip connections?
12:08:51Cassandraslasher: *nod* I see.
12:09:08CassandraWouldn't really gain you much, if anything.
12:09:36 Join whatboutbob [0] (
12:09:41SlasherYep, just a little faster boot. But it would be interesting to try anyway
12:09:45*Cassandra swears at WPS code and and it's wonderfully transparent failure modes.
12:10:02CassandraWould be an interesting hack, yes. a h120 in the mail today. It might be a bit difficult to revive tho...the sticker on the cpu is brown and crispy. :)
12:12:29Slasherwhatboutbob: hehe, sounds promising :)
12:12:31whatboutbobsurprised that sucker didn't go up in flames.
12:12:42Slasheryou could order a new cpu if you manage to solder it
12:12:59Slasherit costs something less than 20e
12:13:09_FireFly_if only the cpu got burned
12:13:15whatboutbobhrmm...a few of the other bits and bobs in there don't look too happy so i don't think i'm gonna bother.
12:13:22whatboutbobthanks all the same.
12:14:04whatboutbob...and i don't exactly have the hands of a surgeon.
12:15:15Slasheryep, soldering the cpu would be very tricky because it's a bga. I have never tried that but with proper equipments it should be possible
12:15:25whatboutbobhdd's in perfect nick tho, along with the remote and cover etc, so it was worth the $35.
12:17:25whatboutbobSlasher: am i allowed to ask if you've made any headway with recording? (insert annoying kid's voice: "are we there yet?")
12:21:38Slasherwhatboutbob: sorry, but no.. i haven't had any time for that :/
12:23:26whatboutbobno need to apologise. (*sobs* 'why aren't my wishes this complete stranger's #1 priority?!)
12:24:50XavierGrSlasher: Did you found about the bugs in crossfade in mix mode I mentioned?
12:25:12 Join ep0ch [0] (n=ep0ch@
12:25:22SlasherXavierGr: that is intended behaviour so the previous song will not fade out but mix
12:25:41Slasherbut we could disable that when manually skipping tracks as it was before
12:26:10XavierGrbut wait a sec, I though crossfade was to fade out current song and fade in next song.
12:26:23Slasheryes, but not on mix mode
12:26:43Slasher(if you only want "mix" with fade in, just set fade in duration to 0)
12:26:57XavierGrwhy is that? It was the other way around in the last crossfade implementation
12:27:08XavierGrdo you think that this is better?
12:27:13Slasherno it wasn't :)
12:27:35Slasheryou can do the same things (and much more) with the new code
12:27:50Jungti1234ashridah: I asked about function always.
12:28:14Slasherbut it's correct that on manual skipping we should not enable the mix..
12:28:18XavierGrso when in mix mode. I want to fade out last song and fade in the next, can it be done?
12:28:43SlasherXavierGr: yes, but then set the mode to "crossfade", not "mix"
12:29:19XavierGrbut crossfade option can't crossfade via a manuall track change, no?
12:29:29XavierGror you say to wait for the track option?
12:29:57Slasherno, it will crossfade. Curerently the manual and automatic track change applies the same crossfade settings
12:30:25XavierGrso tto get this right, which is the difference between crossfade and mix?
12:31:38Slashercrossfade will fade out the previous song according to crossfade settings, but mix mode will instead play the last song the specified duration without fading at all
12:32:36XavierGrAh ok got it, though with mix while it does what you say, I can hear a slight pop when the music from the last track stops...
12:32:46XavierGrBut better ask other is in this...
12:33:02Slasherthat is a known bug and will be fixed soon
12:33:05XavierGras I have said I think that my player pops more than normal
12:33:11Jungti1234What is 'Trigger' of recording setting menu?
12:33:25ep0chor you have very sensitive ears]
12:33:36XavierGrStarts recording when a set amount of volume is reached, I think
12:33:43SlasherJungti1234: it starts the recording when the volume level reaches the trigger value
12:34:42Jungti1234Ah! Understood.
12:35:27ep0chany one have any musepack sound files i can use to test with?
12:36:11Jungti1234What is 'Presplit gap' of recording setting menu?
12:36:33CassandraMight be a good idea if someone were to put a bunch of test files on the wiki, copyright permitting.
12:38:56Jungti1234Is no there a person to answer?
12:41:23Jungti1234(_) ȭ Ǿϴ.
12:41:23Jungti1234(_) ȭ Ǿϴ ~
12:41:23Jungti1234(_-) !! !! ձ ɾ!! ߿~~
12:41:23DBUGEnqueued KICK Jungti1234
12:41:23Jungti1234(_<)Ū(-.-;).. ȭ ġ Ǿ......
12:41:23Jungti1234(_<)Ū(-.-;;)..(-_Ѫ)...ٴٴ..ٴ~ !
12:41:24***Alert Mode level 1
12:41:24Jungti1234(Ĥ0-)ðű ~~~~~ (-,-)(-.-) ... ~
12:41:31Jungti1234=_= sorry
12:41:31 Quit ashridah (Read error: 110 (Connection timed out))
12:41:40XavierGrPresplit enables you to set a value (time or size), and when the recording reaches this value it will split the recording into another physical file
12:41:50Jungti1234thank you
12:43:43Jungti1234And what is 'For at least' of recording setting menu?
12:43:51 Join ashridah [0] (
12:44:06XavierGrI don't know.
12:44:13XavierGrI don't own an archos
12:44:26XavierGrand iriver recording implementation is pending
12:44:37XavierGrso I guess at what you ask...
12:44:51Jungti1234It is iriver.
12:46:02XavierGrthere is no standard recording for the iriver yet, there is only a debug menu that enables recording but it is featureless and buggy.
12:46:09Jungti1234Is not recording function of iriver perfect?
12:46:29B4gderindeed not
12:46:38B4gderthen it wouldn't be in the debug menu
12:46:40XavierGrit is far from perfect
12:47:06XavierGrbut when it is done, oh boy, I really want to see this :)
12:47:45Jungti1234:) Me too
12:49:54Jungti1234What is 'top off charge'?
12:51:25***Alert Mode OFF
12:53:18 Quit DJDD ("Trillian (")
12:53:53Slasheramiconn: "The longword at address 0 is loaded into the stack pointer and the longword at address 4 is loaded into the PC."
12:54:11Slashernow it seems pretty clear that i can erase the flash area i mentioned earlier :)
12:54:15 Join DJDD [0] (
12:54:31Slasherexcept the first 8 bytes
12:54:56Jungti1234hey, slasher
12:55:06XavierGrSlasher: If a new bootloader comes out, how are you going to flash the unit?
12:55:08***Saving seen data "./dancer.seen"
12:55:17Jungti1234Do not you know about 'Top off charge'?
12:55:19XavierGrIf you replace original firmware?
12:55:37DJDDAnyone here have experience with the insides of both H140 and H120?
12:55:45SlasherXavierGr: Hehe, that is a problem then but i could flash it from rockbox
12:55:47CassandraAnyone have any objections to me increasing the WPS buffer size?
12:55:58CassandraI have a 2k wps file that won't load.
12:56:22amiconnXavierGr: I guess Slasher wants to work on firmware_flash.rock for iriver
12:56:23XavierGrbig WPS
12:56:35CassandraYes, but very pretty.
12:56:39_FireFly_Cassandra have you comment lines in it ??
12:56:53novimon_Cassandra, do you have screen cap of it?
12:56:59Slasheramiconn: Hehe, maybe at some point in the future :)
12:57:00DJDDCass, can you get a screenie of it? :)
12:57:03amiconnSlasher: Take care not to delete the bootloader, and of course you'll give up one of the saftey measures
12:57:05novimon_you got me intrested :)
12:57:11XavierGrah ok then, good luck on this Slasher then, bad boy playing with flash procedures... ;)
12:57:11CassandraI hope to release it soon. It's based on the engineer skin off MisticRiver.
12:57:12B4gderisn't the wps limit 1600 bytes?
12:57:15SlasherBut the first thing i will try is to erase the iriver firmware
12:57:20CassandraI'll do a cap. Hold on.
12:57:30amiconnFlashing isn't all that hard.
12:57:34_FireFly_Cassandra: or try to reduce the length of the filenames of the bitmaps
12:57:52SlasherI will take the little risk of bricking the unit and need to buy a bdm..
12:58:09amiconnAccessing the flash works almost identical to archos, only the address is different and it's 16 bit wide instead of 8
12:58:22XavierGrSlasher: That 's nice motivation!
12:58:23amiconnThis is proved by my flash id read working
12:59:40amiconnThe only non-obvious thing to get it working was that the flash rom waitstates were wrong
12:59:54CassandraWhen did the USB insert code get so flaky?
13:00:01Slasheramiconn: the flash dump feature on the debug menu was quite useful, i used it to analyze the flash structure
13:00:41amiconnYeah, that's what it's for, and it's a rather trivial function btw
13:01:19 Join [IDC]Dragon [0] (
13:02:01CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
13:02:01*[IDC]Dragon was attracted by the trigger word "flash"
13:02:23_FireFly_Cassandra: do you have comment lines in your wps-file??
13:02:29B4gderflash dance!
13:02:29amiconnhi Jrg :)
13:02:55CassandraNo, no comment lines.
13:03:08Slasheramiconn: do you think the flash is fast enough so we could even excecute code directly from it?
13:03:32[IDC]Dragonhi Jens et al
13:04:11amiconnSlasher: A test should tell. We need waitstates at the higher cpu frequencies, but it might still be faster.
13:04:14XavierGrcvs checkout: warning: failed to open //.cvspass for reading: No such file or di
13:04:14CassandraHi Jorg
13:04:14XavierGrrectory <- anyone knows about this error?
13:04:28amiconnThe SDRAM is dead slow compared to our cpu frequency
13:04:28Slasheramiconn: Hmm, nice
13:04:35[IDC]DragonHi Christie
13:04:39B4gderXavierGr: it seems your .cvspass file can't be read
13:05:29XavierGrthat error appears even when I want to download the whole source
13:05:31amiconnEven *with* burst mode and with the CPU running at 45 MHz it's only twice as fast as the archos ram (with 11 MHz CPU clock)
13:05:33preglowCassandra: what font is that?
13:05:40B4gderXavierGr: "download" how?
13:05:54XavierGrcvs -z3 co rockbox-devel
13:06:21B4gderXavierGr: yes, cvs will use .cvspass for that case too
13:06:38B4gderin fact, I bet it might use it for almost any command
13:07:09XavierGryes any command i type into cvs pops up that error.
13:07:09 Join yngwi_away [0] (
13:07:16Slasheramiconn: huh, that sounds really slow to me :/
13:07:24_FireFly_Cassandra: what about my questions ??
13:07:25B4gderXavierGr: so fix the file, or remove it and re-login
13:08:22CassandraErm, namil-8. Works with any 8 pixel high font, including rockbox_default though.
13:09:05_FireFly_Cassandra: what about my questions ??
13:09:08Cassandra_FireFly_, I thought I answered them. What did I miss?
13:09:36_FireFly_hmm i don't get it
13:09:48_FireFly_mybe my personal fw was getting freaky
13:09:52_FireFly_and had blocked it
13:10:05CassandraOr it just got lost in the scroll.
13:10:11CassandraWhat did you want to know?
13:10:25_FireFly_i found it :)
13:10:38_FireFly_can you reduce the filename-length
13:10:59CassandraI can, but I'm not going to.
13:11:03B4gderCassandra: #define FORMAT_BUFFER_SIZE 1600
13:11:17preglowCassandra: could i download that wps anywhere? i never tried a bitmapped one before :)
13:11:34CassandraB4gder, I know. I've increased it. Just wanted to check before I checked it in.
13:11:34[IDC]Dragondo bitmaps work for Archos, too?
13:11:53B4gder[IDC]Dragon: yes they should
13:11:59Cassandra[IDC]Dragon, yeah, they do, but you've got less real estate to play with, of course.
13:12:20Cassandrapreglow, I'll make it available for download as soon as I get the final feature fixed.
13:12:29Cassandra(display of song ratings from rundb)
13:12:33preglowCassandra: great
13:12:36[IDC]DragonI'm a bit disconnected with recent dev
13:12:47 Join amiconn_ [0] (
13:12:47preglowCassandra: you'll post it in the wps gallery?
13:12:54CassandraIDC: me too. RL got very busy.
13:12:56Cassandrapreglow, yes.
13:12:59[IDC]Dragonbut this morning I compiles a fresh cvs, first time since 2.5
13:13:09[IDC]Dragon compiled
13:13:29[IDC]Dragonnice Mandelbrot speedup, Jens
13:16:00 Quit DJDD (Read error: 110 (Connection timed out))
13:18:08Jungti1234Work what are you doing now?
13:18:21ep0chpreglow: do you have any musepack test files i can play with?
13:18:48preglowep0ch: sure, what kind of testing?
13:19:09ep0chsame as before, output quality
13:19:29ep0chwonder if i should also test lossless ones too
13:19:41preglowin this case, btw, you need not worry
13:19:55preglowour musepack output is balls on as accurate as you can get it
13:20:27ep0chah good good
13:23:57preglowep0ch: oh well, i throw away _one bit_ at some points, though
13:24:35ep0chi found some mpc files, i want to test it on rockbox anyway
13:26:48 Join Rikos [0] (
13:26:54 Nick Rikos is now known as ChatCase (
13:27:34ChatCasehey, im sure this has been asked for lots of times, and i know there isnt a rockbox version fit to the Archos gmini xs series, but is there any other user-made firmware i can get for my gmini 202 you guys know of?
13:27:50amiconn_[IDC]Dragon: You should try mandelbrot on an iriver H1x0 ;)
13:27:53whatboutbobnight all.
13:27:57 Part whatboutbob
13:29:26 Quit amiconn (Read error: 110 (Connection timed out))
13:29:27 Nick amiconn_ is now known as amiconn (
13:29:35ChatCaseanyone? :<
13:29:57B4gderChatCase: none
13:30:04ChatCasebah :/
13:30:25ChatCasewhats this GminiPort thingy?
13:30:31B4gderuser made firmwares don't grow on trees
13:30:41B4gderChatCase: there was a rockbox port started for the gmini
13:30:47B4gderit was actually quite far gone
13:31:01B4gderwhen the main developer of it vanished
13:31:19B4gderand no one else took up where he left it
13:31:27ChatCaseshame :(
13:31:30ChatCasethanks anywayz :P
13:31:46B4gderdon't be shy, you can be the one! ;-)
13:32:21ChatCasei'll pass :P for the sake of everyone else
13:35:47 Quit ChatCase ()
13:36:02Jungti1234I do not know meaning of word that is 'Top-off charging' at all.
13:36:04 Quit ansivirus (Nick collision from services.)
13:36:07Jungti1234Do not you have person to answer anyone?
13:36:15 Join ansivirus_ [0] (
13:36:35ep0chyeah musepack output is quite accurate with -3/+1 on the sample range.
13:37:20XavierGrJungti1234: maybe it is just for archos models that have nimh batteries
13:37:28XavierGrJungti1234: Nothing to worry about.
13:37:47 Join Nibbler [0] (
13:37:48 Nick ender1 is now known as ender` (i=ychat@
13:37:54Jungti1234Thank reply.
13:38:04Jungti1234I use iriver.
13:38:17B4gdertop-off means " To fill up (a container), especially when it is almost full to begin with."
13:44:47[IDC]Dragonamiconn: Mandelbrot is about the speed I remember from my Atari ST
13:45:14[IDC]Dragonno wonder, with CPU in the same league
13:45:51[IDC]Dragonmy notebook can run it in the pixel shaders, *that* is fast!
13:46:20[IDC]Dragon(realtime zooming/panning with the mouse)
13:46:24ep0ch[IDC]Dragon: oh url?
13:46:28XavierGris there a mendelbrot simulator for PC windows?
13:47:26[IDC]Dragonep0ch: try google, I don't have it at hand. The version I have is for ATI only, but there may be more today
13:47:30Jungti1234XavierGr, B4gder: Curiosity for 1 week was solved now.
13:47:35preglowep0ch: that's an error i have no idea where comes from
13:47:39Jungti1234Thank you very much.
13:49:26amiconn[IDC]Dragon: mandelbrot uses the EMAC unit on coldfire for the full precision routine. Really fast, even though the h120 lcd has almost 3 times as much pixels as the archos lcd
13:50:07[IDC]Dragonin 2 bit "native" grayscale, I guess?
13:50:17Jungti1234what is 'VOICE NUMBERS'?
13:50:29amiconn[IDC]Dragon: No, using the grayscale lib as on archos
13:50:48[IDC]Dragonso being busy flipping pages, too
13:51:00amiconnJPEG decoding is really fast too (the coldfire has a shift-by-n instruction)
13:51:15[IDC]Dragonsigh, that's really missing on SH
13:51:16amiconnSmaller pictures need less than a second to decode
13:51:21preglownot having a shift-by-n instruction sounds like doing it risc to the last breath
13:51:36preglowunecessarily so
13:51:47[IDC]Dragonsaving gates to the last breath, I guess
13:51:55preglowsure, you don't need barrel shifter
13:52:24[IDC]Dragonwhich disqualifies it from any codec work
13:52:30preglowoh, indeed
13:52:41preglowthe clock rate alone more or less does that
13:53:06preglowthe fastest codec i've seen at the moment runs at 34mhz with tons of iram usage
13:53:09preglowand a heavily optimised lpc routine
13:53:41XavierGramiconn: why don't you utilize the repeat function of buttons in mendelbrot?
13:54:28amiconnpreglow: Not necessarily. There are freaks on amiga that made mpeg-1 *video* run on comparatively slow CPUs
13:54:32Jungti1234Finished H100 firmware Korean translation.
13:55:01amiconnXavierGr: The repeat rate is (way) faster than mandelbrot can redraw, even on h1x0
13:55:25preglowamiconn: yes, yes, but those people are clearly insane :P
13:56:01Jungti1234Who do you send to?
13:56:21XavierGramiconn: Even then cant you make the rate of scrolling lower?
13:57:07Jungti1234Anyone answers. Gloomy. :<
13:57:17preglowJungti1234: you can submit it as a patch
13:57:48Jungti1234Do compile.
13:58:22Jungti1234I can not do compile.
13:58:34preglowyou don't need to compile
13:58:41preglowdoes teh translation work with current rockbox?
13:58:47preglowyou don't need any patches for it to work?
13:58:51amiconnXavierGr: Perhaps. mandelbrot could 'eat' all extra repeat events accumulated during redraw
13:59:01amiconnOtherwise scrolling would lag like hell
13:59:25amiconnpreglow: I guess the korean translation needs the unicode patch
13:59:39preglowso there's no real point in commiting it yet
13:59:42Jungti1234There is lang file.
14:00:33Jungti1234Whom do you give this file?
14:00:44linuxstbIs it just me, or is anyone else against the idea of making "folder navigation" Rockbox's only behaviour?
14:01:06preglowlinuxstb: depends if the behaviour is changed
14:01:11XavierGrJungti1234: You will have to wait before sending this.
14:01:26linuxstbpreglow: What do you mean?
14:01:41XavierGrProper Unicode support must be done to enable Korean Language first.
14:02:14Jungti1234Why is it uproar?
14:02:29preglowlinuxstb: fix it so it doesn't break ordinary playlists
14:02:38preglowJungti1234: you're using unicode patch with rockbox, yes?
14:03:15Jungti1234I have lang file only.
14:03:54Jungti1234Someone must do compile.
14:04:12XavierGrdon't you use a patch that enables you to see korean letters?
14:04:13amiconnlinuxstb: Are you for or against making folder navigation default? I am firmly against it, btw.
14:04:55Jungti1234Did not understand.
14:05:14linuxstbamiconn: I have to be honest and say I don't feel that strongly either way. It's only a default.
14:05:22Jungti1234Do not I do this work?
14:05:23amiconnOr rather, I am firmly against it in its current incarnation
14:05:33preglowi'm firmly against it on the basis that it breaks ordinary playlists
14:05:40preglowif that's fixed, i'm for it being enabled
14:05:45linuxstbBut if it is buggy, then it obviously shouldn't be the default.
14:05:58amiconn...where it would skip folder with the ordinary skip button presses at the first / last track of a playlist
14:06:30amiconnI could live with a separate button (combo) for folder skip, then it could be always-on
14:06:35linuxstbYes - that's the behaviour I don't like (and automatically going to the next folder).
14:06:46amiconnOf course it needs to be fixed to work properly first
14:07:03amiconnThe automatic next-folder thing should stay configurable
14:07:17 Join Paul_The_Nerd [0] (
14:07:28amiconn(and off by default, as long as we don't want different defaults per platform)
14:07:46preglowJungti1234: i want to know if you are using ordinary rockbox or unicode rockbox
14:08:02preglowi love the language barrier
14:08:22Jungti1234unicode rockbox
14:08:43preglowJungti1234: then we can't use your patch yet. ordinary rockbox doesn't have unicode support yet
14:08:54preglowJungti1234: but if you can find markun, you can send your lang file to him
14:08:59preglowmarkun: wake up!
14:09:14preglowJungti1234: he is the one working on unicode rockbox
14:09:18Jungti1234Markun is not on seat.
14:09:19*amiconn prods bagder and whispers "l10n v2" ;)
14:10:11Jungti1234Is in place that he is different now.
14:10:46Jungti1234Does no one have a person to take the place of the work?
14:11:13preglowJungti1234: you can find his email address there
14:12:09Jungti1234His E-mail address knows already.
14:12:10Slasheramiconn: Hmm, do you have the correct wait state value for normal cpu frequency?
14:12:48preglowJungti1234: then send the land file to him
14:13:26amiconnThe flash ROM wait states should be correct for all CPU freqs
14:13:28 Nick paugh is now known as AliasCoffee (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
14:13:58Slasherah, hmm
14:14:31Jungti1234Work of me was wished to go fast.
14:20:14Jungti1234Does rockbox firmware modify everybody?
14:24:37 Join Kohlrabi [0] (
14:25:44XavierGrI am thinking to put the creation proccess of the battery benchmark on an if(filling) statement on playback.c, this will do the job right?
14:28:22Jungti1234It is so late night.
14:28:40XavierGrI guess that there will be night time
14:28:51XavierGrhere it is after noon
14:29:16Jungti1234The South Korea is 10 : 29.
14:29:31Jungti1234I go to sleep. Good Luck.
14:29:47Jungti1234I am appreciative today all. :)
14:31:11Jungti1234Oh! Terrible thunderclap!
14:31:28preglownothing beats a good thunderstorm
14:31:34preglowhaven't witnessed one for ages
14:34:17 Quit Jungti1234 ("Http://www.ZeroIRC.NET Zero IRC Ver 2.8")
14:34:31*amiconn thinks about creating battery_log.rock as a tsr plugin
14:35:29XavierGryeah that would be better.
14:35:53XavierGrI will do a test with that I just made to see if it can make it correctly.
14:36:06preglowamiconn: hah, that's a concept i haven't heard anyone mention for at least a decade
14:36:52XavierGrthe need for it raised as linuxstb started to make battery bechmarks for the new FLAC codec
14:37:05amiconnRockbox has a method for tsr plugins, used e.g. for the alpine_cdc plugin
14:37:13preglowthen i'd say go for it
14:37:16preglowthat would ve very useful
14:37:50amiconnIt's in fact simple: The plugin needs to create an own thread, and register an exit function, then it returns from the main thread
14:38:16XavierGrwow big plugin, comparing to others
14:38:27CassandraRight - my new WPS is sitting in the WPS Gallery. Be aware you'll need a bleeding edge build to install it.
14:38:44XavierGrI could programm it but I don't have a clue about threads and proccesses...
14:38:45linuxstbamiconn: How is the plugin memory handled? Does that mean you can't run other plugins if a plugin has tsr'ed
14:41:45XavierGrI guess that thread creation is happening in void thread(void)?
14:45:25linuxstbXavierGr: If you are talking about the alpine_cdc plugin, then the thread creation is done in the main() function. My understanding is that thread() is the actual main loop of the plugin itself - i.e. the contents of the thread.
14:45:54XavierGryeah now that I look at it you are right
14:45:56preglowCassandra: is the peak meter supposed to work?
14:45:59 Quit _FireFly_ (Read error: 110 (Connection timed out))
14:46:13XavierGrIt is just that I don't know anything about threads, it seems interesting though
14:46:20preglowforget it
14:46:23preglowit's a progress bar
14:46:35*Cassandra swears.
14:46:41CassandraSome icons stopped working.
14:46:50CassandraI'll upload a fixed version in a minute.
14:47:42XavierGr3k WPS? that is a beast!
14:48:36amiconnlinuxstb: If a tsr plugin is loaded and the user starts another plugin, the core calls the registered exit function. This exit function must then terminate the thread, ending the tsr operation
14:52:30 Join Maxime [0] (
14:54:29CassandraOK - that should have fixed it.
14:54:44 Join tucoz [0] (
14:55:10***Saving seen data "./dancer.seen"
14:55:38tucozHi, this is off-topic but.. Does anyone of you know of some simple file format to store a binary search tree to a file. That is in a human readable form.
14:55:41CassandraXavier - does seem to be. But it does quite a lot.
14:55:55Cassandra(Used all but 1 of the 52 image slots.)
14:57:26preglowone bug
14:57:33preglowwhen i turn on hold, the hold icon appears
14:57:37preglowbut it never goes away again
14:57:47CassandraI think that's a Rockbox bug.
14:57:51preglowshould be
14:57:56preglowremote hold is always on
14:57:58CassandraFF or RW will get rid of it.
14:58:00preglowand it's not even plugged
14:58:10linuxstbtucoz: You could write it as XML - XML is a textual representation of a tree.
14:58:42tucozlinuxstb, I've thought of that, but I do think that will be overkill for me.
14:59:09tucozlinuxstb, but there are simple libs for this right?
15:00:22linuxstbFor writing XML, you don't really need any libs. For parsing it, there are many libs available.
15:00:58tucozI might have to think about XML again then.
15:01:40linuxstbWhy does it need to be human readable? If it's just for debugging, you could write a "dump" program to display your binary file.
15:02:20B4gderor fix the human to read hex dumps of it! ;-P
15:03:12tucoznah, it is a tree from an article that is generated by a machine learning algorithm. I do not have the data and will simulate something by using that tree.
15:05:09tucozanyway, what is this folder thing I read about in the dev-mailing list? I never use playlist, only play an album in a folder until finished.
15:06:19tucozAs this is the way I keep my music organized. music->artist->album. Will the new behaviour play the next folder automatically?
15:06:42linuxstbRockbox has acquired a new "folder navigation mode" - so pressing next on the last track in a directory will take you to the first track in the "next" directory. (and similarly for previous at the start of a directory).
15:07:04linuxstbAnd yes, when folder navigation mode is enabled, Rockbox will automatically go to the next directory when it finishes the current one.
15:07:29tucozAh, if it is an option. Then that is great!
15:07:49tucozI got the feeling that it wasn't an option.
15:07:54Slasherhehe, i got the flash working.. soon the iriver fw will be gone ;D
15:08:07linuxstbIt was originally an option, but with the default being enabled. amiconn reversed the default, and the proposal now seems to be to remove the option entirely.
15:08:20ep0chSlasher: ohh good job
15:08:28tucozSlasher, cool.
15:08:42linuxstbSlasher: Nice :). Are you executing code directly from flash?
15:08:50tucozlinuxstb, that should be an option imho.
15:09:06linuxstbtucoz: I agree, and I think most other devs do as well.
15:09:09preglowSlasher: don't fuck up your flash now, we can't use you with a bricked player :>
15:09:14tucozah, good.
15:09:46amiconnlinuxstb: The option should stay, or perhaps renamed to the previous name. The navigation could be always enabled, but only if it's a separate button mapping
15:09:56amiconnAnd of course the feature needs fixing
15:10:18Slasherlinuxstb: not yet, the first stage would be to load it to ram from flash
15:10:36Slasherbut i will try the direct execution too later
15:10:39Slasherpreglow: i hope so :)
15:11:27amiconnCassandra: I don't like your wps buffer increase. At least it should depend on the lcd resolution. A whopping 3KB is a waste on archos
15:11:42XavierGrSlasher: You little devil, and yeah preglow was right indeed, DON'T brick your player... :D
15:12:32Slasheri will test the writing on less important sectors first before touching the critical areas
15:12:48linuxstbAre the sectors you are writing to unused at the moment then?
15:13:02Slasheryes, or they belong to the iriver fw
15:13:24tucozHow large is the flash?
15:13:32Slasher2 MiB
15:13:36XavierGrWhen there is going to be an official realese for rockbox on iriver, will there be an option to remove default fw completely?
15:13:39tucoziirc, the iriverfw is quite big
15:14:12SlasherXavierGr: possibly
15:14:30 Join HET2 [0] (
15:15:14Slashertucoz: yep, it eats almost the entire flash
15:15:17tucozthen again, rockbox is 2.8MiB unzipped on the iriver
15:15:31tucozwith all stash of course
15:15:38Slasherso rockbox has much space on the rom after we erase iriver fw
15:15:57tucozrockbox is not that big, just the fonts etc.
15:16:19ep0chi can think of one big codec
15:16:40amiconnSlasher: Will you implement the combined "Crossfade offset" setting?
15:16:49Slasheramiconn: yes, i will
15:17:22XavierGrand the track change? (sorry I must be killing you)
15:17:29tucozhehe, aac is almost twice the size of rockbox.iriver
15:17:34Slasherprobably on the week end when i have more time for it
15:18:09SlasherXavierGr: trying to remember that also :D
15:19:18tucozIs flash->ram transfer faster than hd->ram on iriver? i.e. when the disk is spinning.
15:21:41preglowwell, i'd say "of course"
15:22:08preglowthe fun thing about having the firmware in flash, is that you can do other things while waiting for the disk to spin up
15:22:11preglowlike init audio interface
15:22:56tucozpreglow, I have no idea of the speed of this. I am just thinking of the flash based players often have usb1.1 wheras hd ones have usb2.2.
15:23:16tucozBut that might be some other type of flash
15:23:35XavierGrwhen I make changes to the source and I want to test it on target, do I have to replace the rockbox folder too or only the rockbox.iriver file
15:23:39linuxstbDoes anyone know if the iriver firmware runs from flash, or if it copies itself into RAM? i.e. could we load the irivier firmware from disk and run it?
15:23:43XavierGr(not plugins of course)
15:24:01B4gderlinuxstb: iirc, it copies itself to ram
15:24:42ep0choh how's the flac without wps testing?
15:24:59linuxstbSo we could potentially rolo into the iriver firmware if we stored a copy in a file?
15:25:03HET2will rockbox run on an archos gmini?
15:25:15linuxstbep0ch: Not started yet - I'll try and do it overnight tonight.
15:25:36B4gderHET2: not now, no
15:26:06HET2B4gder: is it planned? started?
15:26:16B4gderHET2: started, stalled, abandoned
15:26:24HET2B4gder: oh
15:26:34B4gderbut feel free to join in and make it happen
15:26:41HET2B4gder: i need to buy a gmini first :)
15:26:47B4gderits an insane platform
15:27:33linuxstbHET2: Do you own any mp3 players?
15:28:06preglowyes, blasted hardware architecture making it easy for us
15:28:18HET2linuxstb: nope
15:29:35preglowa gmini port will be a nightmare
15:29:39linuxstbHET2: You could buy an ipod and join my Rockbox effort :)
15:30:52HET2linuxstb: i will never again buy an ipod
15:31:01linuxstbWhy not?
15:31:07XavierGripod sucks ;)
15:31:10HET2linuxstb: i've had 3 ipods die on me with hd crashes in the past year
15:31:10 Join Febs [0] (
15:31:54B4gderhd crashes?
15:32:00preglowanyone here have any experience with motorola 56k processors?
15:32:21HET2B4gder: yes, the harddrive would simply stop working
15:32:35preglowseems i'll be working coding 56k for a couple of weeks, but i've never used it :>
15:32:41B4gderwell, most 1.8" players use hard drives of the same brands
15:36:10*[IDC]Dragon reads about flashing Slasher
15:36:56ep0chme looks away
15:37:06[IDC]DragonSlasher: have you thought about ucl compression?
15:37:19HET2B4gder: it's a toshiba
15:37:46 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
15:39:18linuxstbHET2: iriver use Toshiba drives as well (my H140 has a Toshiba MK4004GAH). My 60GB iPod has the MK6006GAH
15:44:23tucozlinuxstb, have you read this old post?
15:44:34XShocKare there any plans on porting rockbox to ipods?
15:44:55tucozhe might be interested in joining you.
15:46:22XShocKpreglow, good. :)
15:47:10XShocKi still have my iriver ihp120, but some ppl here are interested in rockbox on ipod
15:47:15preglowi am
15:48:32preglowi'm probably going to get a nano in not too long
15:49:03crwldoes ipod nano have usb on-the-go support?
15:50:36preglowit's software controlled
15:50:43preglowso it does, if we program support for it
15:50:51preglowor wait
15:50:55preglowyou'd need the photo as well :>
15:51:03crwlum, it, like, has an usb chip that can do about anything?
15:51:09crwl(if someone codes the software, that is)
15:51:23preglowit's more like it hasn't got an usb chip that does anything
15:51:48preglowit's just a dummy passing data from the processor bus format to usb
15:51:48B4gderbut then, ipodlinux does not support USB on any ipod model
15:51:59crwloh, ok
15:52:34preglowweird, really, i'd have thought they would try to port the linux usb drivers to work
15:53:03B4gderit might serve has a hint that it isn't as straight-forward
15:53:19B4gderbut then, there aren't very many people doing lowlevel ipod hacking
15:53:31B4gderbasically one
15:53:45HET2B4gder: i remember it took a long time until linux had useful usb support
15:54:08B4gderHET2: so? they run an existing linux, it has very good usb support
15:54:21B4gderthe reason is rather the support for the ipod hw
15:54:21HET2B4gder: isn't it ucLiunx?
15:54:40*HET2 wouldn't know the difference in the usb drivers anyway
15:55:17*B4gder is gonna do a talk on "linux device driver programming" in two weeks
15:55:19linuxstbThe problem is the low-level USB hardware. It's undocumented so can only be understood by reverse-engineering Apple's firmware.
15:55:39linuxstbOnce that was working, all the joys of Linux USB would be available on the ipod.
15:56:12HET2speaking of proprietary apple hardware, i am wondering how long it will be until the airport driver is reverse engineered
15:56:44linuxstbThe airport driver is already reverse engineered. Do you mean the airport extreme?
15:56:50HET2linuxstb: that
15:57:10HET2linuxstb: as far as i know the original airport is not sold/used any more
15:57:22HET2my lecture is over
15:57:35 Quit HET2 ("This computer has gone to sleep")
15:57:53linuxstbXShock: Who is interested in Rockbox on an ipod? You mention "some people here".
15:58:43XavierGrSlasher: I have found that in normal playback initialize_buffer_fill will access the disk and load the audio buffer into ram
15:59:13XavierGrSlasher: But someties when I fw tracks the disk spins up but this function isn't called.
15:59:35XavierGrthen ather about 1 minute it will rebuffer and access the disk, which function is this?
16:01:02linuxstbtucoz: Yes, I saw that mailing list post. He's right - porting Rockbox to the ipod is by far the easiest port to new hardware that someone could do.
16:02:31preglowlinuxstb: have you shown the ipodlinux the photo of rockbox booting? :)
16:02:43preglowinsert 'people' where fit
16:03:15linuxstbNo, but I plan to offer them my bootloader after I write the button driver. It's a big improvement on their existing bootloader, which doesn't include an ATA driver.
16:03:34preglowoh? how do they do things, then?
16:03:48linuxstbCurrently, upgrading or installing the kernel requires raw writing to the boot partition.
16:04:07linuxstbMy bootloader can load and run a kernel from the FAT32 partition.
16:04:15preglowbut i thought your bootloader was a stripped down rockbox
16:04:20linuxstbIt is.
16:04:29preglowoh, how ironic that will be
16:04:39linuxstbThe Rockbox logo might upset them :)
16:04:42B4gderyou just rock linuxstb
16:05:19XavierGrhaha :)
16:05:30XavierGripodlinux with a Rockbox Logo
16:05:37linuxstbBut I'm seriously thinking about adding the IPL logo to the bootloader - to make it a joint project. After all, they did the hard work as far as the low-level ipod stuff goes.
16:06:07preglowsounds nice
16:06:20XavierGryou shoud write with small letters: Bootloader courtesy of Rockbox ;)
16:08:26preglowlinuxstb: but there'll be the problems of some ipods using whatever file system apple uses
16:09:20linuxstbpreglow: Yes. But Apple users can use FAT32 ipods as well (I use itunes on an Apple).
16:09:48linuxstbIt's a problem, but I don't ever see Rockbox incorporating an HFS driver.
16:10:27linuxstbUnless some enterprising Apple user joins the Rockbox team...
16:11:40SlasherXavierGr: that's audio_fill_file_buffer and audio_load_track (it buffers new tracks partially)
16:11:50B4gderimho, you should post some progress info on your ipod work to the rockbox devel list. It could make someone interested pop up
16:11:59 Quit yngwi_away ("Chatzilla 0.9.68a [Firefox 1.0.7/20050919]")
16:12:07Slasherlinuxstb: very nice idea about loading iriver fw from disk
16:13:05preglowi agree with bagder
16:14:29amiconnSlasher: The idea isn't all that new, in fact that's how rolo was born (on archos): The archos firmware in (stock) flash can load another firmware from disk, that's how rockbox was loaded before there was the possibility to run it from flash as well
16:14:44Slasherhmm, is there any disassembler for coldfire?
16:15:02B4gderobjdump ;-)
16:15:15Slasheramiconn: yep, but it's great if we could implement that on iriver as well :)
16:15:19amiconnRockbox can load the exact same format on archos, so you can load an archos firmware disk version from within rockbox
16:15:35amiconnSlasher: Yes, objdump
16:16:07[IDC]Dragonah, Slasher is back
16:16:14amiconnm68k-elf-objdump −−target binary −−architecture m68k:5249 −−adjust-vma 0x31000000 -D rockbox.bin
16:16:14SlasherB4gder: ah, yes :) but it didn't like the parameters when i tried to pass a binary dump to it without elf headers :/
16:16:30Slasheramiconn: oh, thanks :)
16:16:58Slasher[IDC]Dragon: hehe, partially yes ;)
16:17:15Slashernow i will go eat
16:17:15[IDC]Dragonhow about ucl compression?
16:17:26Slasher[IDC]Dragon: sounds like a good idea
16:17:37linuxstbB4gder: I'm hoping to start committing the start of my ipod work to CVS in the next few days. I just need to clean it up a little more first.
16:17:37[IDC]Dragonare you aware of the Archos booloader?
16:17:39Slasherespecially for targot players
16:17:54Slasher[IDC]Dragon: not very well because i don't have an archos..
16:18:20Slashernice typos
16:18:23[IDC]DragonI made a tiny bootloader capable of handling 2 images
16:18:34Slasheroh :)
16:18:42[IDC]Dragoneach can be either compressed,
16:18:55[IDC]Dragonor uncompressed, to be copied into RAM
16:19:11[IDC]Dragonor uncompressed for in-place execution
16:19:36Slasherhmm, that sounds nice. Are you going to commit it soon?
16:19:43Slasher(back soon ->)
16:19:47[IDC]Dragonwe used to have both the original and Rockbox compressed
16:20:42[IDC]Dragonnowadays, the original got replaced by a small "emergency firmware", to give room for uncompressed in-place Rockbox
16:21:35[IDC]DragonSlasher: not to be committed, we use that bootloader since ages
16:22:03[IDC]Dragonperhaps you should reuse as much as possible from that
16:22:23[IDC]Dragonwould have the extra benefit of reusing the toolchain, too
16:24:38 Quit Zagor ("Client exiting")
16:25:33amiconn[IDC]Dragon: In fact I'm planning to reuse some of the bootloader code, and rockbox.ucl (already created by the Makefile for convenient reuse) for making a self-extracting rockbox.ajz in case we hit the 200KB barrier
16:26:32amiconnIt turned out some time ago that we won't have this problem for a long time on fmr, v2 and Ondio, because these have a 400KB barrier
16:26:36[IDC]Dragonyes, I've seen that
16:26:53amiconnBut now the v1 recorder is really tight, and the player might become tight
16:27:05[IDC]Dragonthe player? really?
16:27:30amiconnNot yet, but perhaps with unicode support and added cyrillic and greek mappings...
16:28:36[IDC]Dragonyou'd have to copy the image around twice, or link to different address?
16:28:42amiconnIdeally the decision is made automatically during the build process. If rockbox.bin is small enough, just scramble it to ajbrec.ajz
16:29:13amiconnIf it's too large, compile an additional selfex module, tack on rockbox.ucl, and scramble that
16:29:40[IDC]Dragonsfx can't happen in-place
16:30:00amiconnYes, the selfex module needs to copy the compressed image to high ram, copy the decompressor to iram, then decompress and start
16:30:20[IDC]Dragonor link to high ram
16:30:31 Join HET2 [0] (
16:31:42amiconnI'd like to avoid that
16:32:11[IDC]Dragonbecause of 2 build configurations?
16:33:46[IDC]Dragonif we grow that much, rombox is in danger again, too
16:34:30[IDC]Dragonluckily, I installed a 512 KB chip ;-)
16:35:47 Join webguest43 [0] (
16:50:33 Join _FireFly_ [0] (
16:54:22 Quit ashridah ("Leaving")
16:55:12***Saving seen data "./dancer.seen"
17:05:22 Quit webguest43 ("CGI:IRC")
17:07:28amiconn[IDC]Dragon: Rombox has a limit of 228KB now, correct?
17:11:39 Part tucoz ("Leaving")
17:16:43 Join muesli- [0] (n=muesli_t@
17:20:52XavierGrhi muesli-.
17:21:01XavierGrDo you remember that old printer I had?
17:21:14muesli-yeah i do..what happened?
17:21:45[IDC]Dragonamiconn: I don't know off the top of my head
17:22:17*[IDC]Dragon checks rombox start address
17:22:24XavierGrI made it to pieces and screws but that's not the point, the very same day my scanner died on me
17:22:49XavierGrWell short story is that the feeder was jammed and I tried to fix this myself....
17:23:54XavierGrI didn't saw 2 hidden screws and I end up braking the carcass that made the scanner bits and screws....
17:24:21XavierGrLesson: Muesli was right I should have auction both of them in ebay :D
17:25:32XavierGrNow I don't have a scanner or a printer and I will pay 180euros for a new multipurpose machine, (those that scan and print)
17:26:21[IDC]Dragonamiconn: 228K for all models but player
17:27:16amiconnThe player is different?
17:27:42[IDC]Dragonyes, bootbox is a page smaller
17:27:44 Quit AliasCoffee ("Leaving")
17:27:59[IDC]Dragonso, 232KB for player
17:29:24amiconnAh ok. I had probably chosen identical start addresses for all models.
17:29:44[IDC]Dragonbootbox could be stripped further, but that'll need a lot of #ifdefs in "innocent" code, or a different code base
17:30:05[IDC]Dragoneg. no disk writing, no subdir support
17:30:44*amiconn found a reproducable bug yesterday :(
17:31:02[IDC]Dragonreproducable is good...
17:31:24amiconnIf you go to the recording screen, but don't start a recording, when the idle timeout hits, the box starts to shutdown...
17:31:33amiconn... then panics with recfile: -1
17:32:28ep0chpreglow: you'll be pleased to know that FLAC output is bit perfect with foobar :)
17:34:13preglowwould have been very surprising were that not the case
17:34:59ep0chto narrow down the issue with mp3, would it help if i looked at the output of mp2?
17:35:09ep0chor are they completely different
17:35:15preglowyes, that would help
17:35:22preglowa bit, at least
17:37:06ep0chanyone have a source of mp2 files?
17:37:42linuxstbep0ch: I have a million MP2 files. I'll upload a short one now.
17:37:58linuxstbThey are 48KHz though - which could confuse the issue.
17:38:20ep0chwill optical out do 48khz without resampling?
17:38:24muesli-amiconn could we have a more accurat sleeptimer? steps of 15mins are too big :o 1m would be very ok imho
17:38:34linuxstbep0ch: Maybe. But Rockbox won't.
17:39:08linuxstbWould it still be useful? I don't think I have any 44.1KHz mp2 files.
17:39:30ep0chok so mp2 can be 44.1 khz
17:39:44ep0chi just need to find away of making them
17:39:46linuxstbI think VCDs use mp2 at 44.1KHz.
17:40:27amiconnJust encode an arbitrary audio track
17:40:48amiconnI've mp2- and mp1-encoded versions of our "reference track" on my webspace
17:40:49ep0chyeah what encoder?
17:41:50ep0chah thanks
17:42:01amiconnThey are encoded with Pegase on Amiga
17:42:14amiconnIt's a pretty good layer 1 + layer 2 encoder
17:42:42ep0chnoooo its the iriver song
17:42:48ep0chsomething else pleaseeee
17:44:41 Quit B4gder ("time to say moo")
17:45:34 Quit HET2 (Read error: 113 (No route to host))
17:47:05preglowi've got some other mp2 files if you're really desperate
17:47:33ep0chnah it's ok i'm quite into it now :s
17:47:53preglowthat's insanity creeping on you're feeling there
17:53:51ep0chmp2: -22/+22
17:55:28ep0chand well foobar09b10 plays mp1 back at high speed
17:56:08preglowep0ch: is this with or without rounding
17:58:24ep0chdon't know. what rounding are you referring to?
17:58:59preglowi told you how to remove the emac rounding yesterday
17:59:00preglowin synth.c
17:59:09preglowhave you still remove the EMAC_ROUND part?
18:00:20*[IDC]Dragon waves
18:00:29 Quit [IDC]Dragon ("CGI:IRC")
18:00:44ep0chi've updated from CVS since, so EMAC_ROUND is present
18:01:05preglowupdating from cvs doesn't overwrite your changes
18:01:50ep0chwell i deleted apps/ before hand
18:02:04preglowthen you have nothing to worry abouyt
18:02:26preglowbut yes, there's so much more processing done in layer3.c that this doesn't surprise me
18:02:27ep0chmade no difference anyway
18:02:36preglowmade no difference for layer2?
18:02:46ep0ch(removing EMAC_ROUND when testing mp3)
18:04:14preglowi thought you said there was quite a difference...
18:04:24ep0chi then changed my mind :)
18:04:43ep0chthere was a problem at the very beginning and very end of one of the files
18:05:10ep0chsomething wrong with my sample trimming
18:06:37 Join webguest43 [0] (
18:07:21webguest43hello! is it possible to "downgrade" to v2.4 (I have 2.5 installed on my player)
18:07:22ep0chso for the mp3 test the actual error is -21/+19 not -57/+60
18:07:35ep0chwhich is more or less the same as the mp2 error
18:08:04preglowahh, yes
18:08:22preglowthen my guess is most of the error happens in synth_full
18:08:27ep0chso i'll try and decode mp1 properly
18:08:56preglowi don't think there's a need, i'm pretty certain you'll find the error is the same
18:09:11ep0chprobably, nice to be sure
18:09:17preglowep0ch: have you disabled dithering in foobar, btw+
18:10:15 Quit webguest43 (Client Quit)
18:11:24ep0chdon't suppose you have foobar 0.83 installed?
18:12:11preglowyes i do
18:12:19preglowit's what i use
18:12:38ep0chcan you try amiconns mp1 and see if it plays back at normal speed please?
18:13:01preglowmpglib doesn't support layer1
18:13:50ep0chi'll decode with mad then
18:14:00preglowyou'll need the libmad plugin to decode mp1 files with foobar, afaik
18:16:02ep0chmp1: -22/+21
18:16:52preglowbetween rockbox and foobar with libmad?
18:17:11linuxstbDoes anyone know if "dd" works under cygwin to read/write from/to a raw disk partition? Or if there is a similar utility available for Windows?
18:17:15ep0chbetween rockbox and MAD frontend
18:17:22preglowep0ch: interesting
18:17:35amiconnlinuxstb: dd works
18:17:38preglowep0ch: this suggests a bug in synth_full somewhere
18:17:51ep0chgood stuff
18:17:54 Join tvelocity [0] (
18:18:01amiconnYou need to know the right device name. Cygwin generates the /dev stuff dynamically
18:18:02preglowbut it might not be a bug as well
18:19:00linuxstbamiconn: Thanks. I need to start writing some ipod bootloader docs, and didn't know how Windows users could do it.
18:19:57linuxstbBasic instructions are: Use dd to read the contents of the boot partition; extract apple firmware from image ; create new image containing bootloader and apple firmware ; use dd to write the new image back to the boot partition.
18:20:37amiconnAll windows disk devices are visible from within cygwin as /dev/sdx, the x is determined by the device number in windows
18:20:53amiconndevice 0 = /dev/sda, device 1 = /dev/sdb etc
18:21:04ep0chpreglow: is there an old version off synth.c you can dig out thats realtime?
18:21:55amiconnYou can find the device id in the disk management console
18:22:45linuxstbThe ipl people have written GUI windows and Mac OS X installers - so I'm sure they could be adapted for Rockbox eventually. The Windows one at least is GPL'd.
18:24:15preglowep0ch: none that are more accurate than the current one
18:24:44preglowmy first attempt was very inaccurate
18:25:22*amiconn is curious how a comparison of iriver firmware output with the reference would turn out
18:25:54ep0chamiconn: mp3 and vorbis were both accurate
18:26:11ep0chrockbox vorbis is accurate too
18:26:24ep0chjust rockbox mp3 is slightly out
18:26:34XavierGris it fixable?
18:26:39preglowof course it is
18:27:08XavierGrsoall this time we listen to slightly altered mp3s?
18:27:14preglowep0ch: how accurate were the iriver ones?
18:27:32ep0choff the top of my head -1/+1
18:27:38preglowrounding errors, then
18:27:42preglowrockbox vorbis?
18:28:12ep0chmaybe -1/+1, i should record these results :)
18:28:33XavierGrwhat the numbers mean?
18:28:49ep0chsample difference between iriver/rockbox and reference
18:29:01XavierGrand how do you test this?
18:29:22ep0chrecord from iriver, record from rockbox via optical out
18:29:44ep0chdecode the same file in foobar, invert reference and mix
18:30:07ep0chmaking sure each file is sample aligned of course
18:30:12XavierGrhmm so recording twice the same song with the same firmware shows 0 differences?
18:31:07ep0chi have two recordings, but haven;t checked against each other yet
18:33:28XavierGrjust test 2 same recordings (same firmware) to be sure that your recording procedure is bug proof.
18:33:38ep0chthey vary by 1 bit in places
18:34:24ep0chwhich is madness
18:34:47XavierGryou mean inaudible right?
18:34:51XavierGr1 bit wow
18:35:03ep0chno i would expect them to be the same
18:35:49XavierGrah you said in places so there is alteration more than once
18:36:04XavierGrmake another one just to be sure.
18:36:24XavierGr2 with the iriver firmware and then another 2 with rockbox
18:36:35ep0chyeah someday
18:41:25 Quit Cassandra (Read error: 110 (Connection timed out))
18:46:57ep0chpreglow: is it worthwhile me removing the optimised version of synth_full. would the other synth_full get used instead?
18:47:17 Join bluebrother^ [0] (
18:49:01ep0chwell i'll try it and see
18:49:43preglowep0ch: just add a "&& 0" to the #if defined(CPU_COLDFIRE) line above it
18:49:58preglowbut i don't think that'll make any difference, you also need to undeclare my mad_f_mul macro
18:50:17preglowor perhaps it's a function
18:51:02 Join _user_ [0] (
18:51:45_user_phaedrus961: does the unicode patch work for the latest build?
18:52:31XavierGrI think not
18:53:01 Quit _user_ (Client Quit)
18:53:35ep0chmad_f_mul is in fixed.h?
18:55:14***Saving seen data "./dancer.seen"
19:20:00amiconnlinuxstb: About the differences in ipod hardware you're mentioning that should be runtime-detected in rockbox - similar differences exist on archos
19:20:09markunyesterday and 2 weeks ago I sat in the train next to people with an iriver H120. Both didn't have rockbox installed so I gave them the URL.
19:21:33XavierGrha come to think.
19:22:09markunI thought all H1xx owners would know about rockbox by now..
19:22:30XavierGrImagine a guy that spend so much money on an iriver only 2 find after (lets say) 3 years that another OS way better than default OS.
19:23:33XavierGrI would be crazy, to have a device like that and not be able to use it in its full splendor.
19:23:35tvelocityyeah rockbox rox
19:23:47markunWhen I bought my H120 over ebay I asked the seller why he wanted to get rid of it. He told me he bought an archos because it runs rockbox and he wanted otf playlists :)
19:23:48Paul_The_NerdImagine a guy who spent a bunch of money on a different brand after looking at an iRiver and deciding on something else instead, only to find that *now* an iRiver does more.
19:24:10ep0chpreglow: i'm not too sure how to remove mad_f_mul
19:25:46XavierGrgod I am even thinking to buy an H300 just to be sure that if my H160 fails I will have the other to use with rockbox.
19:26:13XavierGrOnly thought that keeps me away from this is that H300 maybe will not be ported in the end...
19:26:33XavierGrnah I guess it will be...
19:26:46muesli-just shopped a 60gb drive
19:26:56tvelocityi wish someone would port rockbox on the DS:P
19:27:02muesli-didnt needd it but couldnt resist :D
19:27:32XavierGryeah me too I hadn't the need for 60GB but I bought it anyway.
19:27:54tvelocityhow much does the 60gb drive cost?
19:27:54XavierGrthough 400 is too much for me to spend on another DAP.
19:28:09muesli-i use my iriver as data carrier can never be enough...
19:28:31muesli-got it from ebay..140e incl s&h
19:28:45tvelocitytoo bad it doesnt fit in H120
19:28:46muesli-new 180-190e
19:28:53markunI don't know if I should wait for the 40gb single platter or just buy the 30gb
19:29:30XavierGrI bought my 60GB drive for 210euros from dapstore but that was back in June
19:29:32muesli-markun i guess the first use of single 40gig platter are in ipod
19:29:38markunmuesli-: For a few dollars you can buy a USB case for your old HD
19:29:43muesli-and they dont use them so far
19:29:55XavierGrI can't find an 1.8 case
19:30:13muesli-markun no, will sell my 40gb..will get too expensive
19:31:03muesli-and i dont want to carry cases with me
19:31:13muesli-i want an all-in-one package
19:31:26markunDon't know if it will fit both single and double platter
19:31:47muesli-good point
19:32:07muesli-i guess it works with hitachis only!?
19:32:20muesli-recall myself that ive heard something like this
19:33:22markun"USB 2.0 1.8" External Hard Disk Case for Toshiba"
19:33:27muesli-ok ;) dimensions are given..
19:34:12XavierGrwell I don't think that platters matter in enclosures
19:34:26muesli-2mm can matter ;)
19:35:21markunXavierGr: If it didn't matter I could just buy a 2 platter 40gb..
19:35:32muesli-even 3mm
19:36:20XavierGrwell an encolure should hold both of them, no?
19:36:29XavierGrit is a case not a tight DAP.
19:36:32muesli-it should..
19:36:33XavierGrmaybe I am wrong
19:37:12preglowep0ch: same approach there, just add a "&& 0" to the #if part just above
19:37:19linuxstbamiconn: Yes, I know Rockbox already does some run-time hardware detection. But I'm still not 100% convinced which is better - a single binary that any user can use on any ipod, or multiple binaries tuned to the specific hardware.
19:37:25preglowi'll be gone most of the rest today no, so be quick if you've got further questions
19:37:33preglowlinuxstb: i know which i prefer
19:37:42ep0chpreglow: oops i just cut and paste and different routine into it
19:37:47amiconnlinuxstb: Well, it depends on how different the ipods are
19:39:22amiconnI don't know the ipods, but I would judge from what apple does. Are there different firmwares from apple? Does apple offer firmware upgrades at all?
19:40:17linuxstbYes, they do offer firmware upgrades. I think the actual firmware is hidden inside the installer application - so the user doesn't know what is being installed.
19:40:36amiconnHmm, so the installer is usable for all models?
19:40:57ep0chpreglow: so on line 423 it should be: # elif defined(FPM_COLDFIRE_EMAC) && 0 ?
19:40:59linuxstbI think there is one installer, containing different firmwares for different ipods.
19:41:01_FireFly_afaik during firmware update the content of the disk/flash is ereased
19:41:07_FireFly_by ipods
19:41:17amiconnI.e. is it all the same installer even for classic ipod, mini, photo, nano etc?
19:41:38linuxstbamiconn: Yes, I think so. But I'm not an expert on that side of things either.
19:41:56amiconnSounds similar to our rockbox installer
19:42:01preglowep0ch: yes, that should do
19:42:21amiconnIt contains several rockbox binaries for all our archos platforms (6 in total)
19:42:48amiconnThe differences within each platform are handled at runtime
19:43:02linuxstb_FireFly_: I think you are confusing "upgrade" with "ipod restore". The latter involves resetting the ipod back to the factory settings - installing a fresh firmware and wiping the disk. An upgrade doesn't wipe the disk.
19:43:24 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
19:43:40_FireFly_could be
19:44:04muesli-XavierGr did any problems occured while using rockbox with the 60gb drive?
19:44:37muesli-hi _FireFly_
19:45:02linuxstbWhat does the Archos installer do? I thought installation was just a matter of unzipping the file to the root of the device.
19:45:19amiconnThere are 2 ways of installing rockbox for archos
19:45:59amiconnYou can either download the zip archive for your model and unzip it to the drive, or on windows you can use the installer
19:46:28amiconn...which contains the binaries for all platforms, and you select the platform within the installer
19:46:46linuxstbBut it still means the user needs to know the exact model they own.
19:47:02amiconnYes, 'cause that's not detectable from the PC
19:47:03linuxstbBut I guess that's not too hard.
19:47:18amiconnThe archoses are plain USB MSDs
19:47:54amiconnBoth methods install disk-based rockbox only. They do contain the necessary plugins for flashing, and the rockbox flash versions, but not the binary image necessary for first-time flashing
19:49:25ep0chsomething quite odd happens when you're in USB mode (not boot usb mode), and hold stop the >8 seconds
19:49:31amiconnKnowing the exact archos model is simple for the user. All archoses have printed their model on the front plate,
19:49:46amiconnand the installer lists all known variants to make it really simple
19:49:55 Quit muesli- ("ich will Khe!!!")
19:51:09amiconnE.g. for archos player it says: "Archos Jukebox 5000, 6000, Studio 10, 20" for the recorder it says "Jukebox Recorder 6, 10, 15, 20" etc
19:51:26XavierGrmuesli: no, no problems!
19:51:56amiconnThe installer supports English, Swedish and German :)
19:52:03linuxstbI'm happy enough with the five targets I've listed on the IpodPort page - I think that's a sensible compromise. But it may change if I find larger differences than I first thought between devices - or similarities.
19:52:35 Quit Kohlrabi (Read error: 104 (Connection reset by peer))
19:53:48amiconnThe two largest differences for archos models are (1) Old LCD vs. new LCD for player and (2) Samsung vs. Philips FM radio chip for Ondio FM
19:54:12linuxstbAre those run-time detected differences?
19:54:31linuxstbSimilar to the old/new LCD in the 4G color ipod.
19:55:03amiconnThere are numerous other differences, but they are small (mostly the polarity of certain signals)
19:55:36_FireFly_these should be easy handle on runtime
19:56:33amiconnThe player LCDs are quite different, they use a different command set, and the character set is also different, see
19:57:08amiconnVery early rockbox had two different versions for oldplayer and ewplayer afaik
19:57:24amiconn(insert the missing 'n')
20:05:06Slasherhehe, now i can no longer boot into the original iriver firmware because i have erased it ;)
20:05:17preglowi need a stiff drink
20:05:20ep0chpreglow: when i compile i get a "no FPM selected" error. so i "define FPM_DEFAULT" at the top of fixed.h. but i get no sound when playing
20:05:21preglowwith all this talk of erasing flashes
20:06:07preglowep0ch: i've got no time to look at it currently
20:06:15ep0chthat's alright, cos i've had enough too
20:07:27XavierGrand now what sort of upgrade way will you follow?
20:07:33linuxstbSlasher: Congratulations :)
20:07:39Slasherit even erased exactly the sectors i wanted to, looks good that far :)
20:07:47XavierGralso what happens when you press and hold Play + recrod?
20:07:58amiconnpreglow: If you know what to erase, and have a dump of the original content, I's say it's not too dangerous
20:08:14SlasherXavierGr: it crashes immediately, but rockbox still boots :)
20:08:37XavierGrso there you now dont have lame WMA support!
20:08:55_FireFly_who needs wma ??
20:09:04XavierGrthat's why the lame prefix
20:09:11_FireFly_ah ok :)
20:09:13Slasheri also take dump of the flash contents after each procedure to verify everything went ok. After that i can safely turn off the player :)
20:09:24 Join DangerousDan [0] (
20:09:24XavierGrcan you revert back?
20:09:32SlasherXavierGr: yes, i have the original dump
20:25:05 Quit tvelocity ("Leaving")
20:29:08 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
20:54:29 Join Magistus [0] (
20:55:17***Saving seen data "./dancer.seen"
20:57:57preglowlinuxstb: btw, i'm slightly fed up with the coldfire_wide thingy as of now, so will let it rest for a bit
20:58:05preglowcan't for the life of me find out what's wrong
21:02:01Slashernow my flash starts only with the 8 bytes: "10 01 7F F0 00 1F 2A 30" and then entire flash is empty until to 0x1f0000.. And it still boots!
21:02:21preglowyou're really teasing it, aren't you
21:02:28Slasherit was the moment of horror when i pressed power on ;)
21:03:15amiconnYou should get yourself a new nick ;)
21:03:25Slasherhaha :D
21:05:04preglowso all you've got now is the jump instruction and the bootloader?
21:05:22Slasherpreglow: yes
21:06:23Slasheri think i will next try to put rockbox to start at 0x000008 so i can safely boot it by pressing rec+on (without touching the boot loader)
21:07:26preglowwell, you'll need to create a stub that does copying
21:07:27 Quit ep0ch (Read error: 104 (Connection reset by peer))
21:07:46Slasherpreglow: not necessarely if i try to run it directly from flash at first
21:08:00preglowthen you'll need to link it to run from flash
21:08:08Slasheryes, i do
21:08:17Slasherbut i will test that to measure the flash performance
21:10:09 Join Midgey34 [0] (
21:11:15 Quit linuxstb (Read error: 110 (Connection timed out))
21:12:36amiconnMeasuring flash performance could really be done easier, even without flashing anything
21:13:45preglowwell, yeah, but a practical test is always nice
21:14:39 Join linuxstb [0] (
21:15:05 Quit Magistus (Read error: 110 (Connection timed out))
21:15:46amiconnSlasher: The crossfade time settings use set_int() now, just committed :)
21:16:03Slasheramiconn: ah, nice work :)
21:16:11preglowsome people have requested fractional crossfade times as well
21:16:21preglowmost of them mad
21:16:28Slashereh :D
21:16:39 Join hardeep [0] (i=hardeeps@norge.freeshell.ORG)
21:16:50Slashereven now we can't guarantee that the requested delays and durations will happen
21:17:37linuxstbpreglow: I'm not in any hurry to use your lpc_wide function - the decoder is already faster than realtime for 24-bit files after it was moved into IRAM.
21:17:58linuxstbI've just finished porting the bootloader to the Nano (I think). Now I need an adventurous person to test it...
21:18:05preglowdo find one
21:19:34thegeekthe rockbox bootloader?
21:20:26linuxstbYes - my work-in-progress Rockbox port to the iPod bootloader
21:20:47linuxstbKnow anyone with a Nano?
21:20:55thegeekso it already runs on other 4g ipods?
21:21:07thegeekI dont
21:21:14thegeekand truth be told
21:21:17linuxstbThe bootloader works on the 4G color/photo ipods.
21:21:36thegeekI'm not sure if I'd feel very good convincing anyone to test an alpha bootloader
21:21:42thegeekoh wait
21:21:52thegeekcan you run it without actually flashing anything?
21:21:54linuxstbIt's completely safe - it's just written to the hard disk.
21:22:06thegeekmy brother is thinking of buying one
21:22:12 Join arkascha [0] (
21:22:31thegeekif he had it now I'd make him test it;)
21:22:50linuxstbThe ipod has a hardware reset (hold two keys together for 4 seconds), and then you press another combination of two keys whilst it is booting to enter the flash-based disk mode. You then have access to the disk again.
21:23:03thegeekI actually knew that,)
21:23:16thegeekjust been a while since I read up about the linux for ipod
21:23:16linuxstbSeems like Apple actually want to encourage hacking...
21:23:23 Join Magistus [0] (
21:23:29Magistus<Magistus> Hi all, is there anyone that feels responsible for the Rockbox iAudio M3 Wiki at: ? I just noticed the existence of this page last week and since I saw the images were missing, I made the original scans available on my website again ( So you can link to those if wanted. Hope you guys had some help from the inside of my beloved M3L and can ge
21:24:45Magistus(sorry if I'm posting that twice, got disconnected at the point I tried to post that the first time)
21:46:13 Quit Magistus (Read error: 110 (Connection timed out))
21:48:47linuxstbpreglow: I've found a Nano volunteer on #ipodlinux but he can't test until tomorrow...
21:49:05SlasherHmm, should .bss section go to flash or dram?
21:49:18Slasherah, good
21:49:38Slasherso it contains non-const data?
21:49:51linuxstbYep. It should be initialised to zero as well.
21:50:06amiconncheck - that's the linker script for archos rombox
21:50:24amiconnThe bss zero init is done by crt0.S
21:50:50Slasherhmm, yes :) now i will test it
21:52:44 Join Lear [0] (
21:57:19 Quit markun (Read error: 60 (Operation timed out))
21:57:43preglowlinuxstb: so i see, good news
21:58:09 Join Mxm`Pas`Bien [0] (
21:59:14preglowa _lot_ of the mpc_decoder members use far more precision than they need
21:59:40 Join Midgey31 [0] (
22:06:16 Quit Midgey31 ("Download Gaim:")
22:07:55preglowamiconn: is there any speed penalty involved in using 8 bit values, again?
22:08:15amiconnIt depends on how they are used
22:08:32amiconnThey might require a sign extension or zero extension
22:08:51preglowa cycle, it seems
22:08:53amiconnOtoh, loading them from SDRAM should be faster than 32 bit values
22:09:05preglowyes, sure
22:09:10preglowthe bus is only 16 bits wide
22:09:11amiconn(one access instead of two)
22:09:19preglowbut i'm aiming for iram here
22:09:22preglowso it doesn't matter
22:09:23preglowbut still
22:09:47preglowi think the slightly slower move time doesn't matter as opposed to having the fast ram
22:10:00preglowso i think i'll try to downgrade the size of as many of the mpc_decoder members as possible
22:10:09_FireFly_ @devs: what dou think about to move the var key_locked of the keylock key-combo (the WPS_KEYLOCK define) for the other targets beside iriver to button.c and having only function to aktivated it and automatik deactivate in button_w_tmo when the combo is pressed when key_lock is activ
22:10:57TiMiD_FireFly_: very good idea !
22:11:02TiMiDwould make wps code cleaner
22:11:07amiconnI don't understand fully...
22:11:32amiconnOn archos, *only* the wps has keylock, and the Ondio has no keylock at all
22:11:55TiMiDit would make a global keylock
22:12:04TiMiDI see no problem with it
22:12:09amiconnI do
22:12:34amiconnarchos has no global keylock because there is no dedicated lock switch
22:12:41TiMiDand on the top of that the code that handles if a key should be taken in account or not would be centralized
22:12:47 Quit Maxime (Read error: 110 (Connection timed out))
22:12:49TiMiDI know
22:13:02amiconnKeylock in the browser doesn't make much sense, but it would permanently block a button combo everywhere
22:13:31 Quit Midgey34 (Read error: 110 (Connection timed out))
22:13:32_FireFly_currently only this var is left over be global when i'm finished with the wps-widget
22:13:50TiMiDand this var has nothing to do with wps oO
22:13:58amiconnIt has, at least on archos
22:14:08_FireFly_the keylock would be only activated in wps-button-handling
22:15:48_FireFly_but it would be deaktivated when the combo is pressed again in the general-button-handling insteed in the wps-button-handling function
22:15:52TiMiDto have a global keylock would make sense and simplify the code
22:16:10_FireFly_as a compromise
22:16:40amiconnperhaps it would make sense to have a function in button.c to activate/deactivate the keylock on "soft-keylock platforms"
22:17:04amiconnThe wps code would enable it on entry, and disable it on exit
22:17:27amiconnIf it is enabled, the global combo would engage the lock
22:17:56amiconnBut it would mean a loss of functionality
22:18:16amiconnToday, you get visible feedback if you enable or disable the lock, via a splash
22:18:47amiconnThis is necessary because the keylock combo is a bit tricky to get (otherwise it wouldn't work as a lock)
22:18:55_FireFly_it would be also
22:19:04TiMiDit would be displayed in statusbar
22:19:20_FireFly_because i want only remove the global-var form the wps-sources
22:19:22amiconnThat is done today
22:19:36amiconnGlobals vars aren't that bad, imho
22:21:05amiconnThe enable function for soft keylock would have the advantage that other modules could use it too
22:21:07_FireFly_and replace it with a function to activated it from wps-button-handling-code and deactivaten would be handled in the general button-handling.code
22:21:10amiconn(like the radio screen)
22:21:43TiMiDglobal vars are to avoid
22:21:51TiMiDthey make the code tricky to understand
22:21:57amiconnNot always
22:22:06TiMiDyou don't know where a var can be modified
22:22:10amiconnPlus, they often help to reduce code size
22:22:17amiconnLess parameter pasing
22:22:58TiMiDbut if you do an object like design you have only 1 parameter to pass
22:23:10amiconnSlight extension of my own idea: The enable function in button.c could take an int parameter, defining the button combo to use as a lock
22:23:28amiconnLocking/unlocking would then be done in the button driver
22:23:55TiMiDthat sounds good :)
22:23:58_FireFly_yepp if sof-keylock is activated
22:23:59amiconnPassing zero to the enable function would disable the lock (and unlock if it's currently locked)
22:24:29amiconnbutton_set_softlock(BUTTON_F1 | BUTTON_DOWN);
22:24:40amiconnat wps start
22:24:48_FireFly_or simply use the WPS_KEYLOCK define
22:24:54amiconnand button_set_softlock(0); at the end
22:25:21_FireFly_sounds good
22:26:14 Join markun [0] (
22:27:31amiconnIn order to indicate an engaged lock (like it is done today), the button driver could send special button events when keylock is on and buttons other than the unlock combo is pressed
22:28:39_FireFly_maybe a three button event would be useful for this
22:28:56_FireFly_to indicate is keylock is activated or deactivated
22:29:19amiconnThe status is shown on the statusbar - *if* statusbar is enabled
22:29:31amiconnI mean a slightly different thing though
22:30:29amiconnAs it is today, if you enable the wps softlock on archos, and then press other buttons than the unlock combo, a splash is shown to indicate that the lock is on
22:31:15amiconnWithout that, there would be no indication why the buttons don't work if the status bar is disabled
22:31:29amiconnOn iriver you could still look at the lock switch position
22:32:19_FireFly_afaik won't be a keylock message be displayed on iriver in wps
22:32:45amiconnThat's because the keylock works different, on a much lower level
22:33:03amiconn...and you are always able to tell the lock state, even without a status bar
22:33:28 Quit XavierGr ()
22:36:16_FireFly_i'm maybe blind but hold-states are only showable with a wps-tag but no keylock message is displayed
22:36:19 Join XavierGr [0] (
22:36:35_FireFly_if statusbar is of
22:36:52 Quit XavierGr (Read error: 104 (Connection reset by peer))
22:36:54amiconnOn iriver - of course
22:37:12amiconnThey're not necessary because you can just check the hold switch position
22:37:15_FireFly_that was i 'm saying ;)
22:37:31amiconnWith a softlock, there is no position to check, i.e. on archos
22:37:39amiconnSo the splashes are really useful
22:40:31 Quit ender` (Read error: 104 (Connection reset by peer))
22:41:09amiconnOn iriver, the button driver sends *no* events when the lock is on
22:42:12amiconnOn archos, I suggest a slightly different behaviour. The driver should send a special event (e.g. SYS_BUTTON_LOCKED) for every button press that is not the unlock combo when the lock is on
22:42:42amiconnThis could then be used in the wps for splashing
22:43:45_FireFly_sounds good
22:44:04amiconnJust another case: in the button switch
22:44:11amiconncase SYS_BUTTON_LOCKED:
22:44:20amiconnsplash("key lock is on");
22:44:55amiconnThe queue clearing is just to avoid hundreds of splashes when repeat kicks in
22:45:19amiconnThis case: can of course be ifdefed out for other platformsa
22:45:56_FireFly_similar to this #ifdef WPS_KEYLOCK
22:45:56_FireFly_ /* key lock */
22:45:56_FireFly_ case WPS_KEYLOCK:
22:45:56DBUGEnqueued KICK _FireFly_
22:45:56_FireFly_ case WPS_KEYLOCK | BUTTON_REPEAT:
22:45:56_FireFly_ keys_locked = !keys_locked;
22:45:56***Alert Mode level 1
22:45:56_FireFly_ display_keylock_text(keys_locked);
22:45:58_FireFly_ restore = true;
22:46:01_FireFly_ waitfor_nokey();
22:46:03_FireFly_ break;
22:46:48amiconnThe case WPS_KEYLOCK: is then no longer necessary
22:47:06amiconn(but the WPS_KEYLOCK define is, for calling button_set_softlock() )
22:48:48amiconnAh, yes, the restore = true; is also necessary
22:49:20amiconnWe won't have a splash anymore when enabling or disabling the lock though
22:52:47_FireFly_hmm the button-driver could maybe send another button-event for enableing/disableing key_lock
22:54:09amiconnThis splash might be wanted as an indication when the status bar is off. Alas, the player has no status icon to indicate keylock
22:55:16amiconnToday, the same splash is displayed when enabling the lock and when arbitrary buttons are pressed while lock is on
22:55:21***Saving seen data "./dancer.seen"
22:55:34 Join Kohlrabi [0] (
22:55:42amiconnSo the driver would just need to send SYS_BUTTON_LOCKED once when enabling the lock
22:55:57***Alert Mode OFF
22:55:59amiconn...and perhaps SYS_BUTTON_UNLOCKED when disabling it?
22:56:24_FireFly_why not so we could reuse the display_keylock_text fn
22:56:55 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
22:58:42amiconnHmm, and waitfor_nokey() could go away (independent of your changes)
22:59:07amiconn...and replaced by button_clear_queue() *within* display_keylock_text()
22:59:14amiconnSaves some bytes
23:01:55CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
23:01:55*amiconn wonders how many more places there are which use cumbersome constructs easily replaceable by a simple fn call
23:02:14_FireFly_maybe many
23:03:47linuxstbIs anyone familiar with tools/scramble.c ? There is what looks like a small bug at line 323
23:04:40amiconnTiMiD: Something is wrong with your new listbrowser on the player.
23:05:03linuxstbThat line contains the statement "strncpy(header,MY_FIRMWARE_TYPE,9);" and MY_FIRMWARE_TYPE is the string "Rockbox" (i.e. a string containing 7 chars and a trailing \0)
23:05:03amiconnWith the old browser, .ucl files had a 'mirrored question mark' icon
23:05:34amiconnNow they have no icon, although they are still displayed as supported (which is correct)
23:05:51Bagderlinuxstb: strncpy() zero pads
23:06:26linuxstbAh. OK.
23:06:43linuxstbI should RTFM.
23:06:47hardeepspeaking of the new listbrowser, one thing that's been annoying me in the tree browser is that when you go back to the parent directory, your cursor location has changed
23:07:01*amiconn checks that
23:07:01Bagder"In the case where the length of src is less than that of n, the
23:07:02Bagder remainder of dest will be padded with nulls."
23:07:05_FireFly_i gonna go good night everybody
23:07:34amiconnhardeep: It doesn't do that here...
23:07:51hardeepamiconn: i was seeing it on my iRiver
23:08:06hardeepamiconn: happens when you're in the middle of the list somewhere
23:08:14 Quit _FireFly_ ("Leaving")
23:08:21linuxstbI get lots of warnings (char and unsigned char being incompatible) when compiling scramble.c I assume this is because I'm running gcc4
23:08:22hardeepthis was with a build done on friday so maybe something has changed
23:08:30Bagderlinuxstb: yes
23:08:35Bagderwe should fix them though
23:08:45amiconnIf I enter a dir "Rock", then enter a sub-dir of it, then go 2 levels up, I am again on the "Rock" dir
23:09:17hardeepamiconn: yeah, the cursor is still point to the correct directory, it's just moved to a different location on the screen
23:09:40amiconnAh ok, didn't notice that
23:09:41hardeeper, pointing
23:09:47amiconnThis was different before?
23:10:18hardeepyeah, it used to go back to the exact same location
23:10:49amiconnIt seems that isn't important for me, as I didn't even notice it
23:11:53hardeepi only noticed it when i was trying to find a particular song and was browsing a lot of directories... an annoyance more than a bug
23:13:21 Quit hardeep ("BitchX: to the rescue!")
23:13:31 Join hardeep [0] (i=hardeeps@norge.freeshell.ORG)
23:13:34amiconnIndeed, when going up one level, the browser prefers the one-below-top cursor position, if it's possible with the selected dir
23:14:21 Join LinusN [0] (
23:16:50preglowi had a swedish lecturer that persisted in using that expression
23:16:52preglowvery charming
23:17:08Bagderbork bork bork
23:17:48 Quit arkascha ("Konversation terminated!")
23:18:02amiconnLinusN: I have a question: I noticed ata_standby() is an unused function since the call to it before entering usb was removed
23:18:09amiconnIt could be removed...
23:18:17linuxstbBagder: I assume that adding new targets to the configure script won't break the automatic builds?
23:18:29Bagderlinuxstb: correct
23:18:46amiconnLinusN: I could do that...
23:18:51LinusNdo that
23:19:00Bagderlinuxstb: as long as you don't change the meaning of anything existing
23:19:17amiconnBagder: Speaking about the build system... I noticed a bug today
23:19:33linuxstbBagder: No, nothing I've done should break anything existing.
23:19:33Bagdernoooo ;-)
23:19:40amiconnIt's not possible to build (archos) debug builds - it complains that max. firmware size is 200KB
23:19:51preglowi noticed that quite a while ago
23:20:10amiconnIt used to not do that when I last checked
23:20:13Bagderdoes anyone use debug builds?
23:20:30amiconnI tried it today when Cassandra mentioned debug builds
23:20:47Bagderyes, but do you use them?
23:20:48amiconnI actually used them back when I worked on the player lcd init
23:21:07LinusNi use them
23:21:09BagderI don't think I've ever tried to build one myself
23:21:15amiconn...and would probably use them again if only I could Linus' oldplayer lcd working ... :-(
23:21:37amiconn(insert get somewhere)
23:22:18LinusNi used the recorder debug build two days ago
23:24:03amiconnI mean, how did you build it?
23:24:24LinusNi only use the .elf file
23:24:39amiconnAh, yes, that one is built before the error msg
23:24:44LinusNand i didn't need the plugins
23:25:00Bagderit is "scramble" that outputs the warning and stuff
23:25:06Bagderit has no knowledge of debug builds
23:25:09LinusNin fact, the debug build shouldn't attempt to build the ajz file
23:25:12amiconnYes, but why does scramble get called?
23:25:23Bagderbecause it builds a .bin file
23:25:28amiconnThat didn't happen back then
23:25:35Bagderback when?
23:25:52amiconnNot sure... before one of the build system changes
23:25:59BagderI believe it did
23:26:07Bagderjust that no one cared
23:26:12LinusNshouldn't be too hard to fix
23:26:31LinusNi'll do that the next time i run the debugger
23:26:31BagderI'll fix
23:27:11amiconnIt surely didn't one year ago, when I did the player lcd inits
23:27:35BagderI think it did
23:28:07preglowSlasher: player still working? :P
23:30:04*preglow applaudes the first ipod related commit!
23:31:14*LinusN does a little dance
23:32:58 Join ehntoo [0] (
23:33:30 Quit ehntoo (Client Quit)
23:33:36 Join ehntoo [0] (
23:34:54amiconnBagder: Could you change the daily build table a little, so that the date-time string is also a link for a currently running build?
23:35:00amiconnSometimes I am curious ;)
23:35:24Bagderlemme check
23:36:30 Quit markun (Read error: 113 (No route to host))
23:36:41Bagderbut I agree, I miss that too at times
23:38:01 Join markun [0] (
23:38:06Bagderwe'll see at next build round
23:38:33*amiconn wonders about the comment in ata.h
23:39:19amiconnLinusN: Isn't the ata_spindown() value always in seconds, since rockbox handles it itself?
23:39:47LinusNit's 5 seconds per unit in the ata spec, iirc
23:42:02amiconnvoid ata_spindown(int seconds)
23:42:02amiconn sleep_timeout = seconds * HZ;
23:42:02DBUGEnqueued KICK amiconn
23:42:17amiconnCompare that to the comment at the top of ata.h ...
23:44:01LinusNamiconn: iirc, the ata_spindown() function was once a direct ATA_STANDBY command
23:44:23LinusNand we evidently forgot to change the comment in ata.h when we changed the function
23:44:28amiconnProbably... but it no longer is
23:44:35amiconnI'll remove the comment
23:45:03amiconnOne more change - I'll change the return type of ata_sleep() to void
23:45:24amiconnIt always returns zero...
23:46:16LinusNBagder: the debug build doesn't need to build rockbox.bin and rombox
23:46:32 Quit ehntoo ("Leaving")
23:46:47 Join ehntoo [0] (
23:46:53LinusNand not rockbox.ucl
23:47:06LinusNonly rockbox.elf is needed
23:47:29Bagderdoes it build the .bin now?
23:47:57amiconnlinuxstb: ipco ??
23:47:58LinusNOBJCOPY rockbox.bin
23:48:07Bagderyes, obviously
23:48:08LinusNUCLPACK rockbox
23:48:28Bagderoh well, you may fix that if you feel its important
23:48:30preglowamiconn: ipod color?
23:48:35LinusNi will, some day
23:49:37LinusNpreglow, amiconn: sounds like AOL-speak for ipod color
23:49:54 Quit Lear ("Chatzilla [Firefox 1.5/undefined]")
23:50:30LinusNu know, us aol users spk lik diz all da time
23:50:44LinusNmakes us feel c00l
23:51:07markuny0u hax0r..
23:51:44LinusNgotta fly, cu
23:51:54 Part LinusN
23:52:35amiconnd12 5h0uLd r34lly m4k3 0n3 ph33l c00l ? ;-)
23:56:33preglowall the preceding model names are four characters, it probably felt natural to continue in that fasion :)
23:56:37 Quit ehntoo ("Leaving")
23:57:13amiconnHmm, but 'nano' doesn't match the mange pattern, shouldn't it be 'ipna' then?
23:57:15Bagderdarn, the stamp was no link, another fix attempt
23:57:38Bagdersomeone needs to commit something! ;-)
23:57:47amiconnI'll do so soon

Previous day | Next day