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 2005-08-25

00:02:25 Quit ashridah ("uni")
00:03:44NekoNoNiaowany idea of how to tweak the configure settings so mac os x is detected and gemoss is compiled with XP_MACOSX defined ?
00:05:24Strathi'm sure there is
00:06:11NekoNoNiaowcertainly, but i don't know a bit about configure :)
00:23:49 Quit TCK- (Read error: 110 (Connection timed out))
00:25:36 Join Sucka [0] (
00:28:01NekoNoNiaowcool, everything compiles, even the testbed
00:29:53 Quit Aison ("( :: NoNameScript 3.72 :: )")
00:29:54NekoNoNiaowi'll try to run this tomorrow
00:30:12NekoNoNiaowneed to sleep :)
00:36:41 Part NekoNoNiaow
00:41:22 Quit Moos (" HydraIRC -> <- The future of IRC")
00:48:02 Join TCK [0] (
00:57:25 Quit matsl (Remote closed the connection)
01:08:50 Quit Sucka ("a bird in the bush is worth two in your house")
01:11:43***Saving seen data "./dancer.seen"
01:15:12 Join stripwax_ [0] (
01:23:00 Quit ender` (Read error: 110 (Connection timed out))
01:37:28 Quit hicks (Remote closed the connection)
01:38:15 Quit Strath ("Client closed")
01:42:03 Part stripwax_
01:49:37 Join radagast_ [0] (
02:14:44 Part radagast_
02:15:51 Join radagast1 [0] (
02:21:10 Quit _DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
02:34:20 Quit noC|andY`fRa (Read error: 104 (Connection reset by peer))
02:37:17 Part radagast1
02:39:21 Quit preglow ("leaving")
03:08:48 Nick paugh is now known as AliasCoffee (n=pete@2001:5c0:8fff:ffff:8000:0:3e03:6822)
03:11:15 Join phaedrus96 [0] (
03:11:46***Saving seen data "./dancer.seen"
03:27:39 Quit phaedrus961 (Read error: 110 (Connection timed out))
03:35:37 Join webguest67 [0] (
03:35:47 Quit webguest67 (Client Quit)
03:42:01 Nick phaedrus96 is now known as phaedrus961 (
04:06:05 Join QT [0] (i=as@madwifi/users/area51)
04:18:30 Quit QT_ (Read error: 110 (Connection timed out))
04:38:54 Join Leperkawn [0] (
04:39:14LeperkawnHuzzah for the h3x0!
04:39:25LeperkawnOnce again, as a non programmer, is there anything I can test/do?
04:43:07 Quit Leperkawn (Client Quit)
05:05:22 Join JoeBorn [0] (
05:11:49***Saving seen data "./dancer.seen"
05:33:06 Join Paul_The_Nerd [0] (
05:41:11 Quit phaedrus961 (Read error: 145 (Connection timed out))
05:43:11 Join phaedrus961 [0] (
06:36:09 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.6/20050716]")
06:40:30 Join LinusN [0] (
06:43:07 Part LinusN
06:43:20 Join LinusN [0] (
07:04:58 Join Noel_sad_song [0] (
07:11:53***Saving seen data "./dancer.seen"
07:12:32 Join B4gder [0] (
07:17:15LinusNmorning B4gder
07:20:25LinusNit's cold in here, must be the freeze
07:20:44B4gderno such jokes at this hour
07:21:01B4gderhaven't had my first coffee yet
07:21:05B4gderstill brewing
07:21:44B4gdermy new RAM I installed yday seem defective
07:22:11B4gderone of those days
07:26:16amiconnmorning rb
07:26:25LinusNB4gder: defective?
07:27:11LinusNamiconn: mooooning
07:27:15Bgermorning :)
07:27:33B4gderLinusN: yes, my machine seems to have become... "weird" and unstable with it
07:27:53LinusNouch indeed
07:28:03LinusNwhich mb?
07:28:15B4gder2 x 512MB, pc3200
07:28:20LinusNwhich mb?
07:28:52B4gderah, asus A7V8X-X
07:30:40BgerB4gder: memtest86 ?
07:30:54B4gderI haven't had the time to investigate yet
07:31:29Bgeralso, are settings taked out from the SPD ...
07:35:19 Quit Bagder (Read error: 110 (Connection timed out))
07:45:13 Quit Noel_sad_song ("CGI:IRC (EOF)")
07:48:10*Bger is very happy because of the BDMing
07:55:25B4gderBDM is short for "Bring Da Magic", right?
07:55:37Bgerright :P
07:56:00Bgeror "Break Da Machine" :)
07:56:45Bgerbut in the case of Linus's BDMing i'm sure it'll be the first :)
07:59:06LinusNdon't be so sure :-9
08:09:04 Join ender` [0] (
08:36:06 Join tvelocity [0] (
08:51:59 Quit AdNauseam (Read error: 110 (Connection timed out))
08:59:49 Quit tvelocity ("Leaving")
09:11:55***Saving seen data "./dancer.seen"
09:14:40 Nick Lynx_awy is now known as Lynx_ (
09:15:26BgerLinusN: u said 2-3 days ago (maybe) that the default Power management of the hdd when the iriver is connected to USB is written in the 2kbit eeprom chip
09:15:54Bgeris there a way to reprogram it ?
09:16:14LinusNyes, but you can't do it without hardware modification
09:16:29Bgeroh, right
09:18:26Bgeri think it's bad (at least for the battery and probably for the drive) to be constantly spinning when it's in usb mode for some hours
09:18:38B4gderthe rc5-72 compo is 4 days away from 1000 days
09:18:52B4gder0.252% complete
09:18:52LinusNyes, but that should really be taken care of by the operating system, imho
09:19:08LinusNB4gder: how encouraging :-)
09:19:15B4gderyeps, really close now
09:19:29B4gderhit 100% in 386,798 days at yesterday's rate"
09:19:39BgerLinusN: say this to Billy...
09:27:49 Join Febs [0] (
09:28:53 Join AdNauseam [0] (
09:30:46 Nick Febs is now known as Febs_away (
09:40:46SlasherHmm, eeprom chip can't be reprogrammed without hardware modification or is there some protected sector in the chip?
09:41:49BgerSlasher: this is the second EEPROM chip
09:42:00Bgerconnected directly to the USB->ATA bridge
09:42:39Slasherah :)
09:43:55Bgerand it contains the default settings for the bridge
10:03:02 Quit webguest64 ("CGI:IRC (Ping timeout)")
10:22:56 Join DangerousDan [0] (
10:31:44 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
10:36:44LinusNi just made a nice change in the wps display code
10:36:50LinusNnot committed yet
10:36:56Bgerwhat change ? :)
10:36:59LinusNit allows for enums in the ?<> conditionals
10:37:09LinusNthe repeat mode tag now looks like this: %?mr<Off|All|One|Shuffle>
10:37:15Bgergood :)
10:37:17LinusNmuch nicer
10:37:26LinusNand %mr without condition returns the enum value (1 is the lowest)
10:37:35LinusN%mm is the playmode
10:37:40BgerLinusN: can you add and clock tag for rtc units ? O:-)
10:37:51LinusNdon't push it :-)
10:38:48LinusNsorry, repeat mode is %mm
10:38:59LinusN%mr is the remote hold
10:39:32Bger blah :(
10:39:40LinusNthe enums combined with the preloaded bitmaps will be really cool
10:39:42CoCoLUSand why exactly is it mr and not remotehold? :)
10:40:02LinusNCoCoLUS: you don't know why?
10:41:14CoCoLUSbecause senseless abbreviations are much cooler then wasting a few letters?
10:42:01LinusN1) because the parser becomes way more complicated
10:42:16LinusN2) because you will need a much larger buffer for the wps storage
10:43:03CoCoLUShow often do you code the parser, how often do users create new wps?
10:43:25LinusNCoCoLUS: you do the parser then, and send us the patch
10:43:37LinusNit's my spare time
10:45:15LinusNon the archos, both data and code space is crucial
10:45:23CoCoLUSwith all the talk about making it useable for the user (the runtimedb), it seems inapt do have wps files you can only read with a seperate tag-table
10:46:28LinusNeven if the wps tags were plaintext, wps programming would still be difficult
10:47:06CoCoLUSi don't think so - understanding a more complex wps from someone else is a lot easier if you don't have to look up a tag
10:47:18LinusNi agree that the wps syntax is intimidating
10:47:26CoCoLUSfear-inducing :)
10:48:49CoCoLUSwell, wasn't the general consensus that the wps code needs a rewriting, anyway?
10:49:27LinusNbut not mainly because of the syntax
10:50:00LinusNmore because the general, line-based approach is limited
10:53:46CoCoLUSyou'd need a full blown box-model... ugh.... reminds me of css... text floating around boxes (images) etc.
10:55:27LinusNprobably, yuck
10:56:33 Quit epl ("leaving")
10:56:46amiconnLinusN: Does your enum patch reduce code size?
10:57:21amiconnCoCoLUS: The box based approach for wps would be way simpler. Just absolute positioning with absolute sizes
10:57:25 Quit HET2 ("This computer has gone to sleep")
10:57:45amiconn(expressed in pixels on bitmap displays, and in characters on the mplayer)
10:58:18CoCoLUSand if a defined textbox is too small for the text (eg, a very long song title)?
10:58:34amiconnThen it will scroll
10:58:46amiconn...if scrolling for that element is enabled
10:59:49LinusNamiconn: i think so
11:00:11 Join webguest35 [0] (
11:01:31 Nick AliasCoffee is now known as paugh (n=pete@2001:5c0:8fff:ffff:8000:0:3e03:6822)
11:02:15CoCoLUSamiconn, a model as you describe it would have a pretty big disadvantage though, you'd need different wps files for each model/display?
11:02:51amiconnThat's also the case with today's approach
11:02:55CoCoLUSlike some websites have differnet stylesheets for different resolutions
11:03:21amiconnYou can use recorder wps files on the iriver with no problem today, but it won't work on the player
11:04:07Bgeryep, but today's approach is only different between char and bitmap displays ...
11:05:06Bgerhm, in fact the result will be nearly the same
11:06:14amiconnOne disadvantage with a boxed layout is that it depends on the font size
11:06:18CoCoLUSif the new model would use relative positioning (like "bottom right" or "center"), the problem would be solved, wouldn't it?
11:06:52CoCoLUSadd a automatically reduced font size for smaller displays...
11:10:24BgerCoCoLUS: the problem is different
11:10:35Bger5pt point is 5pt on all bitmap displays ...
11:11:24*Bger thinks that the box layot will have a great advantage in things that can be done ...
11:11:52Bgerbut if the box is 6pt high and the font is 8pt
11:11:58***Saving seen data "./dancer.seen"
11:13:06CoCoLUSthen the font needs to be switched for a smaller font
11:13:31BgerCoCoLUS: currently rb supports *only* 2 fonts : system & UI
11:13:31amiconnRelative positioning still makes sense, at least positioning relative to left border/center/right border, and top border/center/bottom border
11:14:01Bgerfor sure
11:21:17Bgerthis way WPS made for example for recorder can look pretty good on iriver
11:24:19 Join ashridah [0] (
11:25:27HClahh! people! o.o
11:25:37Bgerah, HCl
11:25:48HClhi :p
11:26:32Bgerhow's runtimedb's documentation :p
11:26:50HClthe runtime database is pretty well documented
11:26:55HClmaybe slightly out of date
11:27:04HClits mostly the searchengine / databox that needs docs
11:34:09 Quit paugh ("zzzz")
11:55:29 Join Sucka [0] (
12:23:24 Join hicks [0] (
12:28:41 Join preglow [0] (
12:36:57 Quit webguest35 ("CGI:IRC (EOF)")
12:49:18 Join webguest00 [0] (
12:50:07webguest00i saw on the news on your website that linus managed to run bdm on his h320, what does that mean ?
12:50:24LinusNit means that i can run a debugger on it
12:50:41LinusNand develop a boot loader
12:51:22webguest00so it near that we soon have the new firmware?
12:51:40ashridahit means it's progressing
12:52:03LinusNa working h300 firmware is still quite far away
12:52:39ashridahLinusN: do you have a partially decoded firmware for the 3xx like you did for the 1xx series?
12:53:00 Nick Febs_away is now known as Febs (
12:53:12LinusNashridah: i can disassemble the h300 firmware if i want
12:54:17*LinusN found a bug in the multi-align wps code
12:55:42preglowLinusN: congrats!
12:55:50preglowLinusN: how did you solve the power issue?
12:56:02LinusNi haven't solved it
12:56:08LinusNbut i found a workaround
12:56:30preglowwhich is?
12:56:40LinusNif i hold Play when the bdm connects, the power survives the reset
12:57:03preglowwell, that's not the worst hack ever
12:57:12 Quit ReKleSS (Nick collision from services.)
12:57:25LinusNthe power control chip is a really complicated beast
12:59:15 Join webguest20 [0] (
13:01:04LinusNthe string merge in the multialign code is naive to say the least
13:01:51LinusNand it fails miserably if the align tags aren't in left to right order
13:12:02***Saving seen data "./dancer.seen"
13:17:14 Join HET2 [0] (
13:17:28 Quit webguest00 ("CGI:IRC (Ping timeout)")
13:35:38 Join tvelocity [0] (
13:57:29 Quit webguest20 ("CGI:IRC (EOF)")
14:01:05 Join webguest01 [0] (
14:13:06 Quit Febs (" HydraIRC -> <- The future of IRC")
14:14:43 Quit phaedrus961 (Read error: 113 (No route to host))
14:32:34 Quit B4gder ("Lämnar")
14:39:34 Join Psy-Dead [0] (
15:05:33 Join henrikb [0] (
15:06:19 Join webguest42 [0] (
15:06:40 Join courtc [0] (
15:07:14webguest42emm does anyone know what new features will be on the h300 when made
15:07:19 Join phaedrus961 [0] (
15:08:08 Nick webguest42 is now known as mmm (
15:08:19 Nick mmm is now known as Mongey (
15:08:27LinusNMongey: no, we don't know
15:08:35LinusNwe're not psychic :-)
15:09:29Mongeydo you know if there could be a really better feature
15:09:31BgerMongey: it'll start probably with about all that's working on iriver h1x0 now ...
15:09:58Bgerin the beginning it'l be b/w only
15:10:12Bgerno USB host at all ...
15:10:27Bgerno video ..
15:10:49Bgerbut you can start iriver fw if you want any of these
15:11:29henrikbHi LinusN , Interested in a patch that moves credits.c to a plugin?
15:11:30Mongeyis it the first DAP yous have done with a colour screen
15:11:41LinusNhenrikb: sure
15:11:43Bgerit's not done
15:12:06***Saving seen data "./dancer.seen"
15:12:16Bgerthe other candidate for this "title" is iaudio x5, but it's very quiet on this front recently
15:13:01 Quit phaedrus961 ("Leaving")
15:13:51Mongeybut will everyone on RB finish the h1x0 then if the bootloader was done by then move to the x5 or h300
15:14:02Bgerbtw, Linus, the similarities between h1x0 and h3x0's hardware will help you in writing the bootloader, won't they?
15:14:11LinusNMongey: "move to"?
15:14:17LinusNBger: yes
15:15:03LinusNrockbox is never "finished", not on any platform
15:15:17Mongeythats really cool
15:15:39LinusNwe still work on the oldest archos devices
15:15:40Bgersorry for the silly question, but what *exactly* you do through the BDM ?
15:15:51Bgeri know what it is for
15:15:55LinusNthe bdm is a special debug port on the cpu itself
15:16:07Bgeryes, and you run code on it
15:16:13LinusNso i can read/write the registers and memory
15:16:19 Join diway [0] (
15:16:22LinusNso i connect the debugger via bdm
15:16:28LinusNgdb in this case
15:16:38LinusNi can then download and run code on it
15:17:05Bgerwhat code
15:17:13LinusNthe boot loader, for instance
15:17:29henrikbLinusN: mail or patch tracker
15:17:32LinusNor small test programs
15:17:37LinusNhenrikb: patch tracker
15:18:34diwayI've a "little" pb with my iriver 120. I succesfully flashed it with rockbox some months ago and now I've a 'loading firmware: result -1' :/ Any idea ?
15:18:48 Join phaedrus961 [0] (
15:18:55LinusNdiway: did you install rockbox?
15:19:18diwayit was disabled bur I instelled it
15:19:48diwayeverything was fine until today :/
15:20:08LinusN-1 means that it can't find the "rockbox.iriver" file
15:21:58diwayyep, I deleted it use the original firmware, but I did that some months ago and it still worked until today
15:22:34LinusNdiway: so reinstall it
15:22:38diwayand now, I can't even connect to my iriver to upload that file
15:22:42LinusNand do a scandisk
15:22:50diwayno access :/
15:22:54BgerLinusN: some months ago ....
15:22:54LinusNcan't connect?
15:23:11LinusNwhich boot loader version?
15:23:12Bgerthe bootloader's USB mode is more recent iirc
15:23:35diwayan old on
15:23:44diwaymay I would say
15:24:01Bgerdiway: any guesses about the date u've installed it ?
15:24:04LinusNit should say which boot loader version it is when it starts
15:24:32diwayis there a version number displayed ?
15:24:38LinusNshould be
15:24:39diwayI don't think there is on mine
15:24:42diwaywait, checking...
15:24:50LinusN"Rockboot <something>"
15:25:39Mongeyis wat next linus Write a GDB stub for the serial port
15:26:09LinusNMongey: next is to find out enough about the hardware to make a boot loader
15:26:39Bgerthis is more near to mine question
15:26:48diwayLinusN, nope, nothing, sorry
15:27:06Mongeyoh and what are the things that you dont know with the hardware
15:27:25ashridahdiway: does it get to irivers firmware if you hold record while pressing play?
15:27:35diwayand btw, I can't swich it off, I've to wait the battery is dead or reset it
15:28:54diway"starting original firmware"
15:28:58diwayand nothing...
15:29:21 Quit Mongey ("CGI:IRC")
15:29:41diwaymy hdd may be dead :/
15:30:00ashridahthe original firmware usually gets a bit further than that if the hd's dead
15:30:26ashridah(at least, iirc, i don't know all the modes of failure for the firmware)
15:31:01ashridahcrap. 11:30pm.
15:31:05 Quit ashridah ("skeep")
15:33:52LinusNdiway: did you charge your batteries?
15:34:24diwaypowser cable is plugged atm
15:34:36Psy-Deadat both ends?
15:34:48diwayhum, yep :)
15:35:08LinusNdiway: how long has it been charging?
15:35:28diway1-2 minutes
15:36:18LinusNwait half an hour and try again
15:36:33BgerLinusN: and what does the code you "download" to the coldfire do mostly?
15:36:52diwayoh, ok, thx!
15:37:10LinusNBger: tests, like try to access the hard drive etc
15:37:13Bgertrying to "communicate" with other components?
15:37:45LinusNone benefit of the bdm is also that i can program the flash
15:38:10Bgerand test "prealpha" versions of the bootloader
15:38:41LinusNnah, only if something goes wrong, so i can't use the firmware upgrade in the orig firmware
15:39:22Bgerdo you already have an idea about the "unknown" chip ?
15:39:28Bgerwhat is it for
15:39:34LinusNwhich one?
15:40:13LinusNno i don't
15:40:56 Quit tvelocity ("Leaving")
15:41:05 Quit courtc (Connection timed out)
15:49:12 Part LinusN
15:52:04 Quit dwihno_ (Read error: 110 (Connection timed out))
15:53:59 Join dwihno [0] (n=dw@
16:00:26 Quit henrikb ("CGI:IRC")
16:22:24 Join adiamas [0] (n=adiamas@
16:25:30 Nick adiamas is now known as adi|home (n=adiamas@
16:35:29 Quit JoeBorn (Read error: 110 (Connection timed out))
16:42:15 Quit phaedrus961 (Read error: 113 (No route to host))
16:45:55Mode"#rockbox +o adi|home " by ChanServ (ChanServ@services.)
16:51:50Bgerbye all
16:52:00 Quit Bger ("gtg")
16:53:24webguest01Hi all. A strange thing happened... I was using firmware 2005-08-16 and today I wanted to upgrade to 2005-08-25. By accident I have also removed the patched ihp_120.hex file. So, I have decided to do the whole process once again. Everything is ok now, but there is no .rockbox nor ihp_120.hex displayed on the screen; there is only rockbox.iriver. I'm sure they are there because I've tried to delete all the files connected with rockbox: .rockbox; ro
16:57:34webguest01god damn how stupid I am; I didn't check show all files in the settings, sorry for that; nevermind :)
17:00:30 Nick adi|home is now known as adiamas (n=adiamas@
17:02:02amiconnwebguest01: You don't need to keep ihp_120.hex on your box once you flashed it
17:04:43 Join linuxstb [0] (
17:06:58adiamasgod i love seeing this channel this full.
17:10:12amiconnadiamas: Linus or Bagder wanted something from you iirc
17:10:27adiamasyup... already took care of it :)
17:11:16adiamasit's amazing to me.. this channel is a little bit more then 6 months behind the project being created, and the channel has been around 3.5 years
17:12:08***Saving seen data "./dancer.seen"
17:16:22HClwho is adiamas ?
17:16:53adiamasadiamas is the channel owner
17:17:09adiamaslong ago developer... been MIA for just about 18 months
17:17:13adiamasmaybe a bit longer
17:19:57 Join webguest44 [0] (
17:19:57 Quit webguest01 ("CGI:IRC (EOF)")
17:22:41 Join phaedrus961 [0] (
17:28:59 Join phaedrus96 [0] (
17:33:07 Quit linuxstb ("CGI:IRC (EOF)")
17:34:44 Quit HET2 ("This computer has gone to sleep")
17:35:26 Join HET2 [0] (
17:45:30 Quit phaedrus961 (Read error: 113 (No route to host))
17:46:47 Join elinenbe [0] (
17:46:51 Part elinenbe
17:47:15 Join amiconn_ [0] (
17:49:33 Join Lear [0] (
17:55:41 Quit amiconn (Nick collision from services.)
17:55:41 Nick amiconn_ is now known as amiconn (
17:58:36 Quit adiamas ("Chatzilla 0.9.68a [Firefox 1.0.6/20050716]")
17:59:44 Quit diway ("Leaving")
18:03:41 Join bagawk [0] (
18:21:40 Nick phaedrus96 is now known as phaedrus961 (
18:32:47 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
18:38:53 Join courtc [0] (
18:39:11 Join lost|X40 [0] (
18:48:58 Quit Nibbler (Remote closed the connection)
18:49:26 Join Nibbler [0] (
18:51:01 Quit Nibbler (Client Quit)
18:51:48lost|X40so, i recently installed rockbox on my iHP, on the 21st. yesterday, i booted into the stock firmware to record a conference, and after 15 minutes, it stopped recording. Now it won't record more than 6 seconds. I tried defragging and it said there was corruption. So i ran badblocks, and it didn't find anything. So I went ahead and formatted it completely. HTat didn't work. Any ideas on why this may be happening?
18:53:03bagawklost|X40, no idea
18:54:30lost|X40what are the chances this has something to do with the hacked firmware/bootloader (not rockbox itself, since it has been wiped out at this point)
18:56:23Learlost|x40 what do you mean by format didn't work? still recording problems?
18:56:50Learfull or quick format?
18:57:04lost|X40full format, still recording problems
18:57:19lost|X40format went through fine
18:57:55 Join Nibbler [0] (
18:58:46lost|X40it seems to me that the firmware change makes this unlikely, except that it's the only major change to the thing since it stopped working fully
18:59:39LearI guess something could have gone wrong in the flash/patching, but it does seem unlikely, as you say...
18:59:47LearThe battery is decently charged?
19:00:02lost|X40yea, full charge
19:01:43lost|X40i'll try reflashing iwth a couple different verions of the bootloader and report back, just wanted to see if there was maybe something obvious i wasn't seeing
19:01:48LearCould be worth trying a new patch and flash round, double checking that the right files are used... That's what I'd do anyway.
19:03:05lost|X40yea, ok, thanks for thinking about it
19:11:44 Quit bagawk ("Leaving")
19:12:12***Saving seen data "./dancer.seen"
19:14:31 Quit dpassen1 (Read error: 110 (Connection timed out))
19:27:53 Quit solex (Read error: 104 (Connection reset by peer))
19:33:01 Join solex [0] (
19:36:15 Join Moos [0] (
19:37:35 Join DangerousDan [0] (
19:39:12 Join paugh [0] (n=pete@2001:5c0:8fff:ffff:8000:0:3e03:6822)
19:55:45 Join linuxstb [0] (
19:56:19preglowlinuxstb: any codec news?
19:56:56linuxstbALAC is working nicely, but I've done nothing else codec-wise.
19:57:16preglowahh, thought you started working on aac
19:57:44linuxstbI can't link my laptop to the net, so can't download anything.
19:57:46Moosor MPC :D
19:57:57preglowMoos: mpc is a mess i wont touch for a while
19:58:12preglowi was thinking about perhaps looking at speex again
19:58:18linuxstbBut my mp4 parser works perfectly with the AAC files I've tested with it. AAC gapless is going to be hard.
19:58:30Moospreglow: very good plan :)
19:58:43preglowlinuxstb: how come? no standard?
19:58:45linuxstbSame problems as MP3, but no lame-type tags to "fix" it.
19:58:46Mooslinuxstb: good luck
19:58:58Leardoesn't itunes or nero support gapless aac?
19:59:06linuxstbI added a link to a HydrogenAudio forum thread about it on the SoundCodecs wiki page.
19:59:11preglowlinuxstb: someone is bound to come up with a winamp style gapless postfixer, we'll just use that for aac as well
19:59:36linuxstbReading that thread, it seems Apple may be planning something for the next version of iTunes, so the concensus was to wait and see what they do.
19:59:51Learmp4 supports custom data somehow ("atoms"), so there should be a way... :)
19:59:57linuxstbBut we have to get the codec working first.
20:00:50linuxstbpreglow: Is speex floating point? I had a quick glance at it, and it seems that way.
20:00:56preglowlinuxstb: it's both
20:01:06Learpartially fixed-point at least.
20:01:10preglowlinuxstb: parts still are floating point only, but it's very possible to make it work with fixed point only
20:01:15 Join _DangerousDan [0] (
20:02:18linuxstbI can't think of many missing codecs now - apart from AAC and Speex. We now have the difficult work of getting them all working perfectly.
20:03:03preglowi'll try out speex now
20:03:38linuxstbI would like to try AAC, but I'm back at work next week, so not sure if I will have any time. I'll mention it here if I try anything.
20:03:58preglowi'd like to try to make faad2 work, but i'll give speex a go first
20:04:19*linuxstb remembers WMA is also "missing"
20:04:37lost|X40hmm, so flash to stock firmware didn't work to fix my recording problem, any other ideas?
20:04:46Moospreglow: hehe voice UI could use this speex
20:04:52preglowMoos: yes, indeed
20:06:12linuxstbI'm curious about the H3x0 port - do any active devs actually own one (apart from Linus)?
20:06:39Moosmaybe Bger
20:07:06Moosnot a lot of h300 devs currently :)
20:07:25preglowall iriver devs are h300 devs as well
20:07:32preglowsince most of the code will be the same
20:07:39Mooshehe tue
20:07:51preglowdevs for the h300 specific parts will come when the bootloader is up
20:08:10linuxstbEverything in "apps" - but not the firmware side. But it seems Linus is keen atm.
20:08:40amiconnpreglow: While I'd like to see speex support, I think the existing codecs that aren't really up to the task yet should be optimised before adding new ones
20:09:01Moosoh yeah :)
20:09:12amiconnI think the current status of flac and mpc make iriver rockbox look rather half-baked
20:09:36linuxstbWe just need to be clear which codecs are "production quality" and which are not.
20:10:01amiconnSomeone should apply and test the .wav patch. Cvs .wav support is really limited with no seek capability
20:10:11preglowamiconn: i'm looking for a bit lighter work now
20:10:34preglowflac and mpc i don't care about anymore
20:11:07preglowthe flac code is a bloody nightmare, and the mpc implementation is flawed
20:11:16linuxstbI'm motivated to work on FLAC, but not immediately.
20:12:20linuxstbpreglow: I agree. After seeing the ALAC decoder in one nice clear .c file, I can't understand why libFLAC is so complicated.
20:12:34preglowbecause he tried to write c++
20:12:38preglowbut in c
20:12:42preglowand he took it to extremes
20:12:51linuxstbMaybe we will need to rewrite it in C then.
20:13:10preglowthere are mallocs all over the place as well :/
20:13:15preglowa rewrite would be very nice
20:13:21preglowbut that will take lots of time
20:13:30 Quit DangerousDan (Read error: 110 (Connection timed out))
20:13:37preglowafter getting my mits on wavpack, i personally dont care about flac anymore
20:13:46preglowsure, support would be nice and all, but i can't summon the motivation
20:14:33amiconnMy point is that flac must be either optimised to at least work correctly for all compression levels, or we need to drop flac for release
20:14:41amiconnDoes the iriver fw support flac?
20:14:55linuxstbWould it be sensible to mark some codecs as "experimental", so they are not compiled by default?
20:15:10linuxstbamiconn: No - no lossless codecs at all.
20:16:16preglowyes, sure, flac is not ready for release
20:16:19preglownor is mpc
20:16:26paughpreglow: does the flac codec just need optimising? i would have thought it to be one of the lightest..
20:16:28preglowand both libmad and tremor needs minor stroking
20:16:36preglowpaugh: well, then you're wrong
20:16:39linuxstbI would hate to see FLAC support completely removed from Rockbox - it's very close.
20:17:09preglowi think i've got one opt for flac lined up, btw
20:17:16Learpreglow: libmad can't need much at least.
20:17:17preglowi just need to benchmark it to see if it even counts
20:17:20linuxstbInterestingly, the ALAC decoder author said that FLAC was simpler to decode than ALAC - but his ALAC decoder runs significantly faster than libFLAC on the H1x0.
20:17:25preglowLear: oh no, just some layer1/2 work
20:17:26amiconnflac -8 skips like hell on some tracks
20:17:35paughyeah i see that
20:18:00preglowflac needs to use iram
20:18:08linuxstbamiconn: I think your experiences are worse than most, but it's definitely an issue.
20:18:10preglowbut thanks to the way its coded, that's not trivial
20:18:15paughlocks the player on funky sample rates too.
20:18:54linuxstbpreglow: What's the problem with iram?
20:19:09amiconnI wonder why lossless should be harder than lossy
20:19:28linuxstbMore input data to process?
20:19:43amiconnI wouldn't mind flac support being removed, as I won't use it except for experimenting with it
20:19:53amiconnI don't see the real point in lossless compression
20:19:59preglowlinuxstb: it mallocs everything
20:20:00paughfor what it's worth, flac is very important to me.
20:20:01amiconn...for audio
20:20:26linuxstbpreglow: How about writing an iram-malloc set of functions that allocate parts of the iram?
20:21:01preglowwell, if it always mallocs the same amount of space, that can work
20:21:14amiconn(1) If I rip my CDs, I have the lossless original anyway (the CD). (2) lossless compression doesn't really gain much, so either I use uncompressed PCM if I can afford the diskspace, or if I can't afford the diskspace I use lossy compression
20:21:30preglowamiconn: cds detoriorate really fast :OP
20:21:41linuxstbamiconn: The main purpose is to create a lossless backup of your CDs.
20:21:50preglowamiconn: last argument isn't really good, if you ask me, wavpack for instance can compress really well
20:21:59preglowa lot better than half size, in some cases
20:22:09preglowand flac isn't far behind
20:22:26linuxstbBut that's still 700 kbps
20:22:27amiconnI found no track that any of flac or wavpack could compress to less than half the original size
20:22:49paughamiconn, depends on the content
20:22:50amiconnGot 70% original size for my test album
20:23:00linuxstbIf they've been recorded at a low volume, then they do - many live bootlegs are like that.
20:23:21paughclassical music packs down a long way for example
20:24:02amiconnRe deterioration of CDs: I only had exactly one CD that failed in my whole collection, and that was replaced by the german distribution of the label at no cost, after >6 years (!)
20:25:03linuxstbIt's also a pain to have to re-rip your collection if you want to re-encode to a different lossy format. It's a slightly easier task if you have FLAC versions.
20:25:06amiconnlinuxstb: ABBA - More Gold <= no live recording
20:25:40amiconnlinuxstb: re-ripping is easy...
20:26:13paughamiconn, but time consuming
20:26:16Learstill, a lot of shuffling of cd:s, so he does have a point...
20:26:50preglowa lot of my cds go below 50%
20:26:53preglowhaven't tried that many
20:27:29linuxstbLarge hard drives for PCs are now cheap, and I'm sure in a few years portables will be large enough to store an average CD collection losslessly. I just don't see any future in lossy compression.
20:27:33amiconnlinuxstb: I don't think it's easier to get the album from a lossless encoding than re-ripping. I won't store a backup of my CDs on the PC for sure
20:28:01linuxstbamiconn: You can store 8-10 FLAC'ed CDs on a single-layer DVD-R
20:28:41amiconnI can't. No DVD writer
20:28:47paughamiconn, also, my pc is my jukebox. my cds gather dust in the rack and have done for a long time.
20:29:01Learwith a program like foobar2000, transcoding is very convenient...
20:29:08amiconnHehe, my jukebox is my jukebox, much more convenient than the PC
20:29:33amiconn..and certainly better sounding than the crappy laptop sound chip
20:29:42amiconn...even with lossy compression
20:33:59preglowno, lossy compression is just fine for me
20:34:29linuxstbI've just updated the SoundCodecs wiki page to remove the claim that FLAC is real-time :(
20:36:29linuxstbALAC is running with about 60%-70% boost on my test CDs - some recent "alternative" studio CDs. This is without any real effort at optimisation - I'm sure the EMAC could help a lot there.
20:39:50preglowyes, it can
20:39:52preglowi had a look
20:40:03preglowit's got the same kind of filter inner loop flac has
20:40:08preglowusing any iram?
20:40:41linuxstbYes - the temporary 32-bit output buffer is in IRAM - it is exactly 32KB.
20:41:03linuxstbI did tests with various buffers, and that seems to make the biggest difference.
20:41:51linuxstbI've also started work on unrolling and optimising the two special cases that the author refers to in alac.c - but haven't progressed beyond the obvious yet.
20:42:22linuxstbIt's those that should be EMAC'd I think.
20:42:34preglowseems speex might need libogg
20:43:17linuxstbThat's fun...
20:43:37linuxstbWhat does Tremor use?
20:43:50linuxstbCould you borrow parts of Tremor for Speex?
20:44:31amiconnHmm, maybe we could give more IRAM to codecs, but that requires lowering the main stack
20:45:05amiconnCurrently the core uses ~60 KB of its 64 KB, 32 KB of that is the main stack...
20:45:59Learlinuxstb: tremor uses a kind of integrated libbogg. it's only two .c files or so...
20:46:52linuxstbDo we know the maximum usage of the main stack?
20:51:16linuxstbAre all the new targets (iaudio, H320) the same clock speed with the same amount of IRAM? Do they all have EMAC units?
20:51:33linuxstbi.e. are future targets going to be easier or harder than the H1x0?
20:52:17 Quit TCK (Read error: 110 (Connection timed out))
20:53:12lost|X40so, any chance that anyone here can help me with this recording problem? (stops recording ater 6 seconds)
20:53:20amiconnlinuxstb: H3x0 is identical, iAudio X5 has a bit more IRAM (128KB instead of 96)
20:54:13amiconnAll the capabilities you mentioned depend on the CPU. H1x0, H3x0 and some iAudios have MCF5249
20:54:24amiconniAudio X5 has SCF5250
20:56:36linuxstbAnother quick question - do either of the sims support colour yet? i.e. is there a H320 sim?
20:57:54linuxstbI've been toying with the idea of writing a wxWidgets simulator - so we can have a single sim for win32/X/MacOS
20:58:10linuxstbI mean a Rockbox simulator using wxWidgets.
20:58:42amiconnHehe, no wxWidgets simulator using rockbox? ;)
20:59:04linuxstbThat's another idea...
20:59:49linuxstbAre you familiar with wxWidgets?
21:00:02 Join christel [0] (i=christel@freenode/staff/christel)
21:00:11 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
21:00:47amiconnlinuxstb: Not at all
21:02:02linuxstbThat's probably the problem. I've only used it a little, but the nice thing is that it compiles to native widgets on all targets - win32, GTK and Mac OS X (plus others)
21:02:30Learit borrows the design from MFC though...
21:04:07linuxstbAnyway, I don't think I understand the current sims enough to do that job quickly. But I think it would be useful.
21:04:40linuxstbGotta go. Bye.
21:05:00 Quit linuxstb ("CGI:IRC (EOF)")
21:08:25preglowamiconn: more codec iram would be great, then i can optimise layer1/2 a bit more
21:08:36preglowamiconn: what does the core use the iram for, currently?
21:08:56preglowand is it only me, or is 32kb main stack a bit overkill?
21:10:15amiconn32kb main stack is almost surely overkill
21:10:47 Join JoeBorn [0] (
21:11:17amiconnIt was increased from the archos value (8KB) for early versions of rockboy
21:11:47preglowand btw, what do you think of letting the plugins share iram space with the codecs?
21:11:51amiconnI checked it with current rockbox. Debug->View OS stacks says the main stack is only 12% used
21:12:00amiconnThey do...
21:12:13***Saving seen data "./dancer.seen"
21:12:14preglowyes, but not changing it as well
21:12:30preglowhow's that handled now? what happens if a plugin uses the iram?
21:12:30amiconnI think we should leave that as-is
21:12:50preglowi think plugins using iram are probably too intensive to be used while playing music anyway
21:13:01amiconnFortunately the only plugin using IRAM is rockboy, and that needs the audio buffer anyway so it stops playback
21:13:38amiconnThat was the problem with rockboy causing crashes when started while music was playing - it tried to copy its IRAM part before stopping music...
21:14:06amiconnWe should simply establish this as a rule: Plugins must not use IRAM unless they stop playback beforehand
21:14:23preglowyes, speexdec.c uses libogg :/
21:14:44amiconnI think we can reduce the main stack to at least a half (16KB), perhaps even to the same size as on archos (8KB)
21:14:55preglowi think 8kb should be enough
21:15:04preglowbut 16kb extra iram for codecs would be a blast
21:15:34amiconnI'm not sure whether this falls under bugfixing though...
21:16:37Moosit's not a nex feature :)
21:16:40preglowhrpmh, and tremor doesn't have all the ogg_ functions it needs
21:17:26amiconnOPtimising layer2 playback would be sweet
21:17:42preglowthere are plenty of tables that should be in iram in layer12.c
21:17:47preglownone of them very large
21:17:54preglowmainly gain and quantisation tables
21:17:58 Join phaedrus96 [0] (
21:29:05 Quit phaedrus961 (Read error: 113 (No route to host))
21:31:50preglowman, layer 1 really is simple
21:32:05 Join stamppot [0] (
21:33:32*preglow just ran across his old imdct36 hack :/
21:34:13 Join adiamas [0] (n=adiamas@
21:34:37amiconnpreglow: Something bad?
21:36:14 Join tvelocity [0] (
21:39:58 Join NekoNoNiaow [0] (
21:39:59preglowvell, yes, it's incredibly ugly
21:40:10preglowi think i'll try to implement the cacheseg hack again
21:40:16NekoNoNiaowi'm ugly ?
21:41:37adiamas'incredibly ugly'
21:41:39adiamaswas the statement
21:41:52NekoNoNiaowgood, that's better :)
21:42:48preglowperhaps i should spend some more time with libmad
21:42:54preglowsynth_full isn't exactly pretty
21:44:07 Join banan_ [0] (
21:44:53 Quit merbanan (Read error: 104 (Connection reset by peer))
21:45:17 Quit phaedrus96 (Read error: 113 (No route to host))
21:45:49 Nick banan_ is now known as merbanan (
21:53:36 Join midk [0] (
21:59:38 Join edx [0] (
22:13:32 Join miner49er [0] (
22:14:43 Join Bagder [0] (
22:19:29 Join amiconn_ [0] (
22:19:45 Quit amiconn (Nick collision from services.)
22:20:02 Nick amiconn_ is now known as amiconn (
22:21:58 Part Lear
22:25:16 Join DangerousDan [0] (
22:26:43 Quit Bagder (Read error: 104 (Connection reset by peer))
22:32:37 Quit paugh ("preblom")
22:34:41NekoNoNiaowdoes anyone know approximately how much time qt takes to compile ?
22:34:50NekoNoNiaowon a 1 ghz machine
22:38:20ender`i'd guess about an hour
22:38:52 Join matsl [0] (
22:40:02 Quit HET2 ("This computer has gone to sleep")
22:40:23NekoNoNiaowi feared that :)
22:40:47 Join courtc_ [0] (
22:41:35 Quit courtc (Read error: 110 (Connection timed out))
22:42:19preglowamiconn: would you say tremor is relase quality?
22:43:32 Quit _DangerousDan (Read error: 110 (Connection timed out))
22:50:21 Join thegeek_ [0] (
22:51:24 Quit DangerousDan (Read error: 110 (Connection timed out))
23:05:25 Quit thegeek (Read error: 110 (Connection timed out))
23:05:38 Nick courtc_ is now known as courtc (
23:08:17 Join Strath [0] (
23:12:17***Saving seen data "./dancer.seen"
23:17:48 Join muesli- [0] (
23:22:34 Join HET2 [0] (
23:23:30NekoNoNiaowtime to sleep for me
23:23:35NekoNoNiaowhi and bye ;)
23:23:46 Quit NekoNoNiaow ()
23:30:51 Quit muesli- (Read error: 104 (Connection reset by peer))
23:32:17 Quit Moos (" Try HydraIRC -> <-")
23:35:14 Quit tvelocity ("Leaving")
23:55:05 Quit miner49er ("ChatZilla 0.9.61 [Mozilla rv:1.7/20040514]")

Previous day | Next day