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 2008-10-20

00:01:20MelaGoso i'm able to run the complete tutorial to get it to FAT32. But when I try to install after that using the program, i get the RW-error...When i replug the ipod, itunes gives a message to restore it, which is good i guess, since it is in fat32-structure. However, even after this replug and the shutting down of itunes, i get the same RW-error
00:01:49linuxstbGoing back a step, after you converted it to fat32, does the ipod work as normal? i.e. does the apple firmware still work?
00:02:11MelaGono, when i unplug it
00:02:16MelaGoit says to connect it to a pc
00:02:25MelaGoand then gives me the instruction to restore it, which i ofcourse don't do
00:03:05linuxstbThat's the first problem you have to solve
00:03:34linuxstbAre you sure it's a "5g", and not the "5.5g" ?
00:03:37 Quit mcuelenaere ()
00:04:01 Join Thundercloud [0] (
00:04:09 Quit HBK (Read error: 110 (Connection timed out))
00:04:25MelaGopretty sure yeah. No search function, and 512 byte sector
00:05:12linuxstbOK, 512-byte sectors is the important point.
00:09:14 Quit bluebrother ("leaving")
00:14:54amiconnAny manual expert around?
00:15:38MelaGoSo, linuxstb, when i get it to fat32, my apple firmware should still be working?
00:16:11linuxstbMelaGo: Yes
00:21:32MelaGonot working...Keep on auto-getting it in disk mode and itunes telling me to restore it...
00:22:50 Quit jhulst (Read error: 110 (Connection timed out))
00:24:12 Join Seed [0] (
00:24:45 Join markun [50] (n=markun@rockbox/developer/markun)
00:24:49linuxstbMelaGo: And after you restore with itunes, it works normally?
00:25:46linuxstbCan you describe the steps you are doing? Copy and paste any commands you're typing.
00:27:58MelaGoi download the
00:27:58MelaGo30GB Video (512-byte sectors) mbr-video30gb.bin file and put it on my desktop
00:28:06MelaGoopen up terminal
00:28:13MelaGoand disk utility
00:28:52MelaGocheck the disk-info in disk utility for the ipod, which is: disk2s3
00:29:03MelaGodeactivate the ipod in disk utility
00:29:31 Quit Nibbl (Read error: 113 (No route to host))
00:30:41MelaGotype in terminal: dd if=/Users/<myname>/Desktop/mbr-video30gb.bin of=/dev/disk2
00:30:55MelaGothen i get: 1+0 records in
00:30:55MelaGo1+0 records out
00:30:55MelaGo512 bytes transferred in 2.509447 secs (204 bytes/sec)
00:31:33MelaGothen type: newfs_msdos -F32 -v iPod /dev/rdisk2s2 to get it in fat32
00:31:51linuxstbDo you remove the ipod between the those two steps?
00:32:06MelaGoI tried it once
00:32:11MelaGogave the same thing
00:32:14MelaGoshould i remove it?
00:33:06linuxstbI can't remember - it seems other people have changed those instructions since I first wrote them.
00:33:59MelaGoalright, so now when i go to disk utility
00:34:12linuxstbNo, it doesn't seem to be required to remove and reattach.
00:34:19MelaGoi got my ipod with the name "disk2s2" still on deactivated mode
00:34:31MelaGostructure is: MS-DOS (FAT32)
00:34:47MelaGonow, in order to install rockbox with the program, I have to activate it
00:35:18linuxstbI've no idea what that means - I never had to do anything else.
00:35:36linuxstbIf you now eject your ipod, does the Apple firmware work (do nothing else after the newfs_msdos command)
00:35:39MelaGowell, when I'd go to rockbox utilty now, it wouldn't find the ipod
00:35:43MelaGosince it's deactivated...
00:35:56MelaGoit gives me an error...So I have to activate to get the program to locate it
00:36:08MelaGosec, i'll eject
00:36:20MelaGohaven't run rockbox yet btw
00:36:43linuxstbYes, I know - you need to confirm you have a working winpod first.
00:36:49MelaGowhen i eject, it reboots and
00:36:55MelaGoYES, it works on apple firmware
00:37:16MelaGowhen connecting it to itunes, it asks me to name it and all that...
00:37:45MelaGowow, and it got a windows structure too
00:37:48MelaGofirst time i see this
00:37:51MelaGoalright, so what's next?
00:38:26MelaGonow i have to run rockbox utility right?
00:38:28linuxstbSimply run rockbox utility - it should work now.
00:39:53 Quit Tetracomm (Read error: 110 (Connection timed out))
00:40:21 Join Tetracomm [0] (n=nicholas@
00:43:38 Quit bughunter2 ("bye")
00:43:44 Quit ender` (" Theorem: a cat has nine tails. Proof: No cat has eight tails. A cat has one tail more than no cat. Therefore, a cat")
00:45:17 Quit havien ("Konversation terminated!")
00:47:32*sarixe skdjf
00:48:24 Join havien [0] (
00:48:24*sarixe dies
00:49:31 Quit kharo (Read error: 60 (Operation timed out))
00:51:57*sarixe hi
00:52:36MelaGoalright, so i got a succesfull install of Rockbox, but another problem arises :p...When rebooting Rockbox, i get a "Can't load rockbox.ipod" on my screen...
00:54:35 Quit Schmogel (Read error: 104 (Connection reset by peer))
00:57:29linuxstbSo it wasn't a successful install...
00:57:47linuxstbDid you install both the bootloader and Rockbox itself?
00:57:58 Join kharo [0] (
00:58:08pixelmain the Rockbox Utility which option did you use? I've heard that there unfortunately are still problems with the "Complete installation" from the "Quick start" tab on MacOS...
00:59:23MelaGoused complete installation
00:59:26 Quit einhirn (Read error: 104 (Connection reset by peer))
00:59:43MelaGoso i'll manually install the bootloader and rockbox individually?
01:00:19 Join Munkie [0] (
01:00:45Munkiewhere can i download plugins?
01:00:46 Join syn4pse [0] (
01:01:07linuxstbMelaGo: It looks like the bootloader install worked - you just need to install Rockbox itself.
01:01:24linuxstbMunkie: All plugins are included with Rockbox. is failing for me in 64bit ubuntu, any ideas?
01:01:46pixelmaMelaGo: no, bootloader is already installed now (that's the one that put "Can't load..." on the screen). You should be able to just install the Rockbox build from the "Installation" tab in the utility (second option)
01:01:48Munkiei want to get the Gameboy emulator. it is not included
01:01:58***Saving seen data "./dancer.seen"
01:02:18linuxstbMunkie: Yes it is - please read in the manual about how to use it (and also about "viewer" plugins)
01:02:24pixelmaMunkie: the emulator itself is incuded
01:02:27syn4psespecifically it won't compile the gcc cross compiler (arm), there is an error in the number of arguments in the script.
01:02:29MelaGok, i manually installed rockbox (second option in the utility), still get the same error...
01:02:31pixelma*included too
01:02:37Munkiethank you
01:02:38linuxstbMunkie: Although what mp3 player are you using Rockbox on?
01:02:48MelaGo"can't load rockbox.ipod
01:02:50MelaGofile not found"
01:02:58Munkieipod video 30gig
01:03:07MelaGoeven though install went perfect
01:03:08linuxstbMunkie: Then yes, it's included.
01:03:55linuxstbMelaGo: Maybe your ipod wasn't mounted - sometimes I experienced OS X leaving ghost "/Volumes/[ipod name]/" folders
01:04:01Munkieoh and is sound supported in doom?
01:04:32linuxstbThere are sound effects, but no background music.
01:04:38Munkieok cool
01:05:35MelaGoIt says it's activated/mounted. Got it running in disk mode now right now
01:05:54MelaGoHowever, when opening the ipod in Finder, I got no Rockbox directory, don't know if that's ok...
01:05:55 Quit culture (Connection timed out)
01:07:00linuxstbMelaGo: It's ".rockbox" (note the dot) - which is hidden by default in Finder. You can check in the terminal by typing: ls -al "/Volumes/ipod name/"
01:07:40MelaGoI know it's hidden
01:07:46MelaGoI can see Hidden Files
01:07:50MelaGodefaults write AppleShowAllFiles TRUE
01:07:51MelaGokillall Finder
01:08:02MelaGono Rockbox though...
01:08:42amiconnI've rewritten the archos flashing manual chapter. It would be nice if a native english speaker could have a look at it...
01:09:59linuxstbMelaGo: OK. You could just install the manually - download it from the website, and type "unzip -d "/Volumes/ipod name/"
01:10:03linuxstbamiconn: Sure
01:11:31 Join evilnick [0] (i=60e8ce0d@gateway/web/ajax/
01:11:50amiconnThe relevant chapter is "Advanced Topics->Rockbox in Flash" in and−−r18829M-081019.pdf
01:13:01 Quit linuxstb (Remote closed the connection)
01:13:16amiconnThey're slightly different. There are also and , but there are only there for completeness. They only contain the same text blocks as the first 3, just in partially different combinations
01:14:32amiconnErr, the Ondio FM link should be of course
01:17:26 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
01:18:18amiconnHere's the diff as well: . Practically everything has changed though...
01:18:42linuxstbOK, I think I'll just look at the PDFs...
01:22:32 Join BlakeJohnson86 [0] (
01:22:59 Quit MelaGo ()
01:25:23 Quit Thundercloud (Remote closed the connection)
01:25:48 Join Thundercloud [0] (
01:26:33 Quit Thundercloud (Remote closed the connection)
01:26:41 Quit Nico_P (Remote closed the connection)
01:28:27linuxstbamiconn: I would write "three to five" instead of "3..5", (I was always taught that digits less than 10 should be written as words). Also, you say "several", but then only list two reasons. You could rephrase it to say something like "The main reason to change this is to increase the startup time of your player. The Archos bootloader is rather slow, and by putting Rockbox in flash you can decrease the startup time to around three to f
01:28:27linuxstbive seconds. Furthermore, you..."
01:28:41 Join gartral [0] (
01:28:45linuxstb(it also says "2 files" later on)
01:28:54amiconnThere are 3 reasons when rombox is available, hence "several"
01:29:03gartralgood evening everyone
01:29:08amiconn(Player, Ondio SP)
01:29:17 Quit MethoS-- (Read error: 104 (Connection reset by peer))
01:29:24amiconnI'll change the small numbers to works
01:30:15 Quit Munkie ("Leaving")
01:30:32linuxstbamiconn: OK, but the manual I'm reading only lists two... (ondiofm)
01:30:33gartralhas anyone built and released an updated rbutil? my build system has a dead hard drive atm
01:31:21amiconnlinuxstb: Yes, two or three, depending on the availability of rombox.
01:31:38linuxstb"it’s gonna program" -> "it is going to program"
01:31:43 Join nuonguy [0] (
01:31:54 Quit syn4pse ("ChatZilla 0.9.83 [Firefox 3.0.3/2008101315]")
01:32:13amiconnThat was laziness... (reusing parts of the old text) thanks.
01:32:34gartralhas anyone built and released an updated rbutil? my build system has a dead hard drive at the moment
01:35:09amiconnlinuxstb: Hmm, *increase* startup time? That would be a reason against changing it ;)
01:35:29linuxstbamiconn: Just testing that you're still awake...
01:35:47linuxstb(i'm obviously not)
01:36:55gartrallol, nvm me, i was blind and wasnt reading the "b" above the 1.0 entry in the rbutil page
01:37:17gartral(gives everyone a fresh hot cup of coffee)
01:37:59saratogalinuxstb: how did you calculate the codec memory usage for flac?
01:38:11saratogajust dig through the map looking for the largest addresses?
01:38:13linuxstbsaratoga: I just manually looked at the .map
01:39:06linuxstbamiconn: No more comments on the ondiofm manual - it reads fine to me.
01:41:27 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
01:41:28 Nick fxb is now known as fxb__ (
01:41:36linuxstbamiconn: Recorder manual seems fine.
01:43:25linuxstbamiconn: Player seems fine as well.
01:43:46*amiconn found a grammar mistake in the extra Ondio warning for Ondio SP
01:44:13 Part gartral
01:45:07linuxstbamiconn: "These versions..." ?
01:45:44amiconnI had a mini-\opt there, but this was obviously not enough
01:45:48linuxstbAnd "they are" in the next sentence.
01:46:35linuxstbIs it possible to detect Archos fw versions? Maybe with a crc?
01:46:49amiconnYou mean in RoLo?
01:47:08linuxstbI guess I do...
01:49:42amiconnDoesn't look like it
01:49:47saratogais there a "no plugins" make option ? It'd be nice to not have to wait for them to build and then copy over
01:50:14amiconnWell, a crc might work, but then the rockbox binary would need to know these "bad" CRCs
01:50:36amiconnsaratoga: 'make bin'
01:50:36linuxstbsaratoga: "make bin"
01:50:53*linuxstb lost by a microsecond...
01:50:55pixelmaamiconn won ;)
01:51:10*pixelma loses
01:51:22linuxstbsaratoga: "make help" should show you the options
01:51:45saratogaok thanks
01:52:06saratogafaad only mallocs 36KB for me on a 3 minute AAC-LC file
01:52:27amiconnlinuxstb: If rockbox wouldn't have grown so big, we could have used the size. The flash-update archos firmwares are significantly larger than the normal on-disk firmware files. But rockbox is now closer in size to the bad ones...
01:52:27linuxstbfaad or the mp4 parser?
01:52:32JdGordonlinuxstb: any idea how much of that e200v2 code works? does it build a working bin at all?
01:53:13linuxstbJdGordon: It compiles.. But it's still very incomplete - someone (maybe me...) needs to finish reading an OF disassembly and implementing the funcitons.
01:53:37amiconnThe CRC method may be an idea for the RoLo rework
01:53:40*amiconn sighs
01:53:53saratogalinuxstb: 36KB is as large as the mem_ptr in the malloc buffer gets
01:54:24saratogai just edited the malloc,calloc,realloc functions to print the mem_ptr whenever they're called and started playing files
01:55:46 Join [i][B][e][n] [0] (n=Bensawso@unaffiliated/bensawsome)
01:56:13linuxstbI wouldn't call 36KB "only" - it should grow linearly with the size of the mp4 file. So a 30-minute file would be 360KB, and a 2-hour file would be trouble...
01:58:29saratogai still don't understand why such large seek tables are needed for MP4
01:59:04saratogai started looking through the mp4 specs, but the format is surprisingly complicated
02:00:29linuxstbYou're surprised an MPEG standard is complex? ;)
02:01:43amiconnlinuxstb: I've now committed the reworked flashing chapter. Thanks for checking!
02:01:54linuxstbYou're welcome.
02:02:00JdGordonare these mallocs used for the life of the track?
02:02:34*amiconn will commit it (with an extra remark regarding the 8MB mod) to the 3.0 branch tomorrow, for building the 3.0.1 manuals
02:03:03linuxstbJdGordon: Yes, I'm pretty sure they all will be.
02:03:08saratogaJdGordon: most of them are seek tables which could be needed whenever someone seeks, the rest are maybe 16KB or so of internal decoder buffers
02:03:47saratogadoes ogg not malloc?
02:04:50JdGordonif the seektables arnt available it can stll seek? it just has to do alot of extra work?
02:05:26saratogaah Ogg duplicates the codeclib malloc functions
02:05:34saratogahow handy
02:06:28amiconnWell, ogg does this for a reason, iirc
02:06:31 Quit kharo (Read error: 110 (Connection timed out))
02:06:51amiconntomal implemented a malloc in the codec ram for ogg, in order to make it work on the iFP
02:07:08amiconnThe iFP doesn't have enough ram for the extra malloc buffer
02:07:18saratogaJdGordon: I've been trying to figure that out
02:07:27*amiconn guesses that with today's change, no codec would work anymore on the iFP
02:08:34 Join kharo [0] (
02:09:09 Quit sarixe (Read error: 110 (Connection timed out))
02:11:59 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
02:13:29linuxstbamiconn: Today's commit was a synced version of a patch by tomal
02:13:52amiconnI doubt he increased the codec buffer to 1MB though
02:14:14linuxstbNo, but the point is that it's now a unified buffer - that's what tomal wanted.
02:14:49amiconnYes, but 1MB would be too big for the iFP, the same way as 512KB codec buffer + 512KB malloc buffer are too big
02:15:09linuxstbYes, but I'm sure it's not 1MB there
02:16:00linuxstbPrior to today's commit, the codec malloc function only used the 512KB malloc buffer - as far as I could see, the unused part of the codec buffer was never used.
02:16:47linuxstbTomal's patch means that malloc uses the codec buffer - which we increased (for now) to 1MB on the large-mem targets, and got rid of the malloc buffer.
02:20:57 Join axionix_ [0] (
02:25:12linuxstbsaratoga: IIRC, the information about the duration and sizes of frames is stored in the STTS chunk. Looking at the ALAC codec (alac.c), the decode_frame() function doesn't return a "bytes_consumed" value - that's taken from the STTS chunk. AAC also uses the STTS chunk values when decoding frames.
02:25:20linuxstbSo there are two issues - 1) Can the decoding happen without the info in the STTS chunk; 2) How do we seek without it?
02:25:37 Join webguest21 [0] (n=18aaaecf@gateway/web/cgi-irc/
02:27:26saratogalinuxstb: its hard to imagine they require you to have a huge lookup table to seek through an audio file
02:27:35saratogaperhaps some of the atoms in the stream have time stamps or something similar
02:28:36 Quit webguest21 (Client Quit)
02:30:31 Join webguest44 [0] (n=18aaaecf@gateway/web/cgi-irc/
02:31:09 Quit webguest44 (Client Quit)
02:31:48 Quit axionix (Read error: 110 (Connection timed out))
02:32:28 Quit kachna (Read error: 113 (No route to host))
02:44:34amiconnlinuxstb: Actually an .ajz file has both its size and a checksum (no real crc, just a sum) in the header
02:45:51amiconnSo checking both should be sufficient for detecting the bad archos firmwares. Problem is how to detect whether rockbox is flashed or not...
02:51:44 Quit Tetracomm (Read error: 104 (Connection reset by peer))
02:52:53linuxstbamiconn: That makes that part easy - and I see those values are already read by rolo...
02:53:29 Join Tetracomm [0] (n=nicholas@
02:55:57 Part pixelma
02:56:22 Join pixelma2 [0] (n=marianne@rockbox/staff/pixelma)
03:02:01***Saving seen data "./dancer.seen"
03:09:56 Join sarixe [0] (
03:17:50 Quit [i][B][e][n] (Read error: 104 (Connection reset by peer))
03:17:51 Join Bensawesome [0] (n=Bensawso@unaffiliated/bensawsome)
03:18:56 Quit saratoga ("CGI:IRC (EOF)")
03:20:23 Quit perrikwp (" ajax IRC Client")
03:20:39 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
03:21:00 Join reacocard [0] (n=reacocar@
03:25:26 Quit Bensawesome (Read error: 104 (Connection reset by peer))
03:25:58 Join Bensawesome [0] (n=Bensawso@unaffiliated/bensawsome)
03:30:40 Quit soap (Remote closed the connection)
03:40:11 Quit intrados (Connection timed out)
03:40:17 Join soap [50] (n=soap@rockbox/staff/soap)
03:41:42 Nick Bensawesome is now known as [i][B][e][n] (n=Bensawso@unaffiliated/bensawsome)
03:42:26 Quit axionix_ (Read error: 104 (Connection reset by peer))
03:44:51 Quit Tetracomm ("Visit:")
03:46:54 Join EspeonEefi [0] (i=espeonee@STRATTON-SIXTY-EIGHT.MIT.EDU)
03:49:39 Quit [i][B][e][n] ("t3h 1337 h4X0R sT4t80t h42 QuiT")
03:51:49 Quit DerDome (Nick collision from services.)
03:51:50 Join DerDome1 [0] (
03:52:02 Nick DerDome1 is now known as DerDome (
03:52:18 Quit Strife89 ("Leaving")
03:59:18 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
04:19:49 Nick HBK- is now known as HBK (
04:20:47 Join blkhawk- [0] (
04:23:46 Join massiveH [0] (
04:29:06 Join kharo1 [0] (
04:29:35 Join miepchen^schlaf [0] (
04:30:22 Quit Kopfgeldjaeger (Read error: 104 (Connection reset by peer))
04:35:19 Quit miepchen^schlaf_ (Connection timed out)
04:36:24 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
04:38:10 Quit blkhawk (Read error: 110 (Connection timed out))
04:41:05 Quit kharo (Read error: 110 (Connection timed out))
04:49:08 Quit obo ("bye")
04:57:31 Join gkffjcs [0] (n=john-cha@
04:57:43 Quit JdGordon ("Konversation terminated!")
04:59:01gkffjcsIf I make a playlist on my ipod, under rock box, how do I save it? I get to the nameing dialog, and then give it a name, but then I cannot save it, there is no option to save, and all the buttons just return me to the previous menus.
04:59:54 Join JdGordon [0] (
05:00:36scorchethe manual should go into this...
05:00:45 Quit mc2739 ()
05:02:05***Saving seen data "./dancer.seen"
05:04:08 Join rocky_ba [0] (i=74e1722a@gateway/web/ajax/
05:04:10 Join mc2739 [0] (
05:10:28cool_walking_gkffjcs: PLAY/PAUSE confirms those naming dialogs.
05:10:32 Join Tetracomm [0] (n=nicholas@
05:12:04 Part toffe82
05:12:14 Part mc2739
05:12:49 Join mc2739 [0] (
05:13:06 Join someone972 [0] (n=someone9@VDSL-151-118-8-114.DNVR.QWEST.NET)
05:13:49 Quit mc2739 ()
05:14:22 Quit someone972 (Client Quit)
05:17:54 Part rocky_ba
05:18:11 Quit Tetracomm (Read error: 104 (Connection reset by peer))
05:24:19 Join devslashnull [0] (
05:25:38 Quit devslashnull (Client Quit)
05:25:56 Join devslashnull [0] (
05:26:08gkffjcsthank you cool_walking_
05:28:59cool_walking_gkffjcs: More info:
05:29:03 Nick blkhawk- is now known as blkhawk (
05:29:21 Quit devslashnull (Client Quit)
05:35:33 Quit HBK ()
05:35:39 Join HBK [0] (
05:40:18XavierGrJdGordon: ping
05:50:02XavierGrJust updated the scroll padding patch
05:50:12XavierGrwow I feel stupid that I didn't think of prefix/suffix
05:50:44 Quit Horscht ("User was distributing pornography on server; system seized by FBI")
05:50:53XavierGrafter hours fighting to handle properly quotes and strings the solution was much more simple
05:51:44JdGordonhehe, sorry i didnt tihnk of it sooner also :p
05:52:15XavierGralso about the length of the padding. 10-20 can be quite small for large screens with small fonts
05:52:26XavierGrthat's why I ended up on LCD_WIDTH / 4
05:53:01XavierGralso it muches the logic of scroll_engine.h (before the patch)
05:53:02JdGordonyeah, but if they are all spaces then its very obvious where the line ends anyway
05:53:18XavierGrmuches = matches
05:54:20XavierGrSCROLL_LINE_SIZE is defined by LCD_WIDTH even if the array holds characters and not pixels
05:54:36XavierGrby the way why is this array so big, I didn't have to change it and the padding still fits
05:57:56XavierGrmy logic is that on a huge screen (say ipod 5.5) LCD_WIDTH / 4 is 60 characters, which on all fonts will cover the extreme preferrence of one whole screen of padding.
05:58:31XavierGrof course hardcoding to 60 characters won't make sense for ondio which will need quite fewer characters to fill in the screen
05:58:47JdGordonwhich array is that?
05:59:14XavierGrin scroll_engine.h scroll_info.line
05:59:41JdGordonthats not the array to hold the full text of the line being scrolled is it?
06:00:42XavierGryes that's it
06:01:05XavierGrthat array is then put on the screen with a different offset (scrolling mechanism)
06:01:07JdGordonthen it needs to be at least MAX_PATH long
06:01:39XavierGryeah and it needs another half screen for the next append of the string
06:01:40JdGordonoh, or it is the subpart of the string being displayed?
06:01:53XavierGrthe part I don't get is the 3*LCD_WIDTH/2
06:01:58XavierGrand without the /2 for the player
06:02:09*JdGordon doesnt know this code much
06:02:10XavierGrJdGordon: AFAIU is the whole string
06:02:24 Quit Seed ("cu, Andre")
06:02:31JdGordon... im also in windows atm so dont even have the code here
06:02:34JdGordonlemme reboot....
06:02:46 Quit JdGordon (Read error: 54 (Connection reset by peer))
06:05:17 Join JdGordon [0] (
06:05:20 Quit HellDragon (Client Quit)
06:05:25XavierGrthe scrolling mechanism takes that string and types it to the screen with an offset (and that is how scrolling is achieved)
06:06:21XavierGrJdGordon: so it is the full string plus half of the screen which some part of the string is reappended to make scrolling look normal
06:06:36*JdGordon wonders if scrolling could be redone with viewports so we could use a pixel spacing between start and end
06:11:38ameyerspeaking of :
06:11:49ameyerdarn keyboard
06:14:25XavierGrJdGordon: the first version of the patch pads spaces calculated from pixels/font width
06:14:49XavierGrbut then some people said that the best would be for the user to select its own padding string
06:15:08JdGordonyeah, I sort of followed the discussoin
06:15:28 Join HellDragon [0] (
06:16:18XavierGrso what do you think, is it commitable? Any more comments?
06:18:23JdGordon... by overkill I meant that its wasting too much space... imo adding any more than half a dozen spaces doesnt make it any *more* readable... so having such a high limit is a waste
06:19:32XavierGrWell, I for one, like at least half the screen with spaces. So I took as base the extreme of one screen (beyond that it isn't visible)
06:19:35JdGordonbut apart from that it looks ok
06:22:14XavierGrokay then if you think it is good to go and you are comfortable with commiting it I would be glad. Right now though I have to go to sleep. Thanks for the tip again. :)
06:25:29 Quit Zarggg ()
06:28:12 Quit XavierGr ()
06:29:05 Join intrados [0] (
06:36:34 Join kushal_12_27_200 [0] (n=kushal@
06:41:00 Join Tetracomm [0] (n=nicholas@
07:02:09***Saving seen data "./dancer.seen"
07:28:57 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
07:36:17 Quit BHSPitMonkey (Remote closed the connection)
07:38:31 Quit nelek ("shutdown -h 0")
07:43:59 Quit n17ikh|Lappy ()
07:44:37 Join HBK- [0] (
07:46:18 Quit Tetracomm ("Visit:")
07:52:00 Join num1 [0] (n=brian@unaffiliated/num1)
07:53:19 Quit miepchen^schlaf ()
07:57:22 Nick num1 is now known as num1_ (n=brian@unaffiliated/num1)
08:01:54 Quit HBK (Read error: 110 (Connection timed out))
08:03:27 Quit kushal_12_27_200 ("Leaving")
08:03:47 Quit DerDome ("Leaving.")
08:05:29 Quit BigBambi (Remote closed the connection)
08:06:50 Join goffa [0] (n=goffa@
08:07:48 Quit agaffney (Read error: 104 (Connection reset by peer))
08:13:47 Quit joe2371 (Remote closed the connection)
08:14:02 Join joe2371 [0] (
08:14:16 Quit bmbl (Read error: 104 (Connection reset by peer))
08:15:18 Join Zagor [0] (n=bjorn@
08:16:09 Join agaffney [0] (n=agaffney@gentoo/developer/
08:19:34 Quit goffa_ (Read error: 110 (Connection timed out))
08:27:41 Quit jhulst (Remote closed the connection)
08:29:00 Join Rob2223 [0] (
08:34:35 Nick fxb__ is now known as fxb (
08:44:21 Join ender` [0] (
08:47:44 Quit Rob2222 (Read error: 110 (Connection timed out))
08:48:55 Quit GodEater (" ajax IRC Client")
08:56:16 Join Bagderr [241] (n=daniel@rockbox/developer/bagder)
08:56:39 Nick Bagderr is now known as B4gder (n=daniel@rockbox/developer/bagder)
08:57:56 Quit advcomp2019 (Read error: 104 (Connection reset by peer))
08:58:47 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
09:02:14***Saving seen data "./dancer.seen"
09:02:49 Join reacocard_ [0] (n=reacocar@WL-48.CINE.HMC.Edu)
09:06:39 Join lasser [0] (
09:06:58 Join petur [50] (n=petur@rockbox/developer/petur)
09:08:32 Quit reacocard (Read error: 110 (Connection timed out))
09:14:56 Nick reacocard_ is now known as reacocard (n=reacocar@WL-48.CINE.HMC.Edu)
09:23:03 Quit Bagder (Read error: 110 (Connection timed out))
09:23:48 Join kachna [0] (
09:26:50 Quit pixelma2 ("-")
09:26:56 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
09:29:47 Quit joe2371 ("Lost terminal")
09:32:15 Join Thundercloud [0] (
09:46:36 Join Nibbl [0] (
09:50:26 Quit JdGordon ("Konversation terminated!")
09:53:40 Quit kachna (Read error: 113 (No route to host))
10:07:47 Join culture [0] (
10:11:32 Join MethoS [0] (
10:13:24 Join Slack_ [0] (
10:22:32 Quit MethoS (Read error: 104 (Connection reset by peer))
10:27:54 Quit Thundercloud (Remote closed the connection)
10:44:17 Join GodEater [0] (i=c2cbc962@gateway/web/ajax/
10:51:06 Quit reacocard (Read error: 104 (Connection reset by peer))
10:52:13 Join kachna [0] (
10:52:24 Quit cool_walking_ (" ajax IRC Client")
10:52:34 Quit nuonguy ("This computer has gone to sleep")
10:54:56 Quit Slack_ ("Ex-Chat")
11:02:15***Saving seen data "./dancer.seen"
11:11:21B4gderwith _some_ targets getting a 3.0.1 the download/release page gets a bit tricky
11:11:29B4gderI mean, how to write on it to describe the situation
11:12:44Zagordo we even have to?
11:13:05 Quit culture (Connection timed out)
11:13:26B4gderI think so, since the users who experienced problems with 3.0 might want to know that it is now 3.0.1
11:14:56linuxstbProbably too late to suggest this, but couldn't we just bump every target to 3.0.1 and have a changelog showing what changed - i.e. that it is only different to 3.0 for certain devices.
11:15:20B4gderwe can still do that, if that's what we really want
11:15:28Zagor"UPDATE: An update, version 3.0.1, has been released to fix problems with xxx and yy on targets zzz and nnn."
11:15:43Zagoruh double update looks a bit silly, but you get the idea
11:16:04scorchei would rather have that than "why havent you guys released 3.0.1 for my ipod yet? I dont want to be behind!"
11:16:08GodEaterI think that's the best plan
11:16:34Zagorscorche: well the other side of that coin is "what is the difference between 3.0 and 3.0.1?"
11:16:53linuxstbZagor: The answer is a link to the changelog
11:16:54scorcheZagor: and the changelog or notice should take care of that
11:16:59Zagoror worse: "3.0.1 sounds much worse than 3.0 on my ipod. can I have 3.0 back please?"
11:17:17GodEatercan't wait to get one of those ;)
11:22:21 Quit kachna (Read error: 110 (Connection timed out))
11:23:08 Join kharo [0] (
11:24:50 Quit kharo1 (Read error: 110 (Connection timed out))
11:24:51pixelmaboth sound like a BoS anyways...
11:29:30B4gdersome people also obviously believe in secret development somewhere not publicly visible...
11:31:23LloreanB4gder: As can be seen by how many people thought 3.0, the first time, was gonna have a bunch of new features.
11:31:57B4gderI guess I just never really learn to expect this
11:37:43LloreanWell, I think part of it is that people expect it to be similar to some other softwares (take GCC) where you have the 3.x and the 4.x branches.
11:38:06LloreanSo when we did Rockbox 3.0 they expected all the previous versions had been on the 2.5 branch, and maybe they just didn't know where to get 3.x test builds.
11:38:45linuxstbLots of projects do that though - have a "new features" branch.
11:39:24linuxstbMaybe we should call "trunk" the "3.1" branch to make it clearer.
11:40:31 Join vitja_ [0] (n=vitja@
11:53:56 Join balou_ [0] (
11:55:21balou_is there a hw platform that is currently sold which is supported by rockbox?
11:55:54balou_ipods don't work anymore, i dunno when iriver quit producing its players, and many other players seem to only work in older releases...
11:56:00B4gdernope :(
11:56:16B4gderbut there is work in progress on several current models
11:56:18Zagorbalou_: only used/refurbished ones
11:57:42Zagorbalou_: if you're in the US, you can get rockbox-compatible sansas from ... uh what are they called again?
11:58:07amiconnbalou_: What do you mean, only seem to work in older releases?
11:58:10balou_yeah.. but ipod videos are getting scarcer and scarcer
11:58:13Zagorscorche: ah, thanks
11:58:21balou_amiconn, old hardware revisions
11:58:31balou_btw, I'm not in the us
11:58:32linuxstbbalou_: There are many better devices for Rockbox than the ipod video
11:59:18balou_what would you suggest? perhaps?
11:59:36linuxstbIf you want a hard-disk player, then the Gigabeat F is very good value.
12:01:11balou_B4gder, good link :)
12:01:40*Zagor removes the "available" column
12:02:27balou_It's a pity... I think there _would_ be a market for dap with flac support and for example digital output
12:02:50markunI just update unifont: "As of 20 June 2008 (by coincidence the Summer Solstice), the GNU Unifont has a glyph for every printable code point in the Unicode Basic Multilingual Plane (BMP)"
12:02:51balou_it just seems that no manufacturer got that yet...
12:03:36markunbalou_: the iriver h1xx has digital out
12:04:53markunand with rockbox it plays back flac of course
12:05:08balou_I know, but that player has been out of production for a few years now
12:05:38linuxstbThat shouldn't stop you buying it if it has the features you want.
12:07:10markunbalou_: and if you want to do some DIY work you could probably make a digital-out for the Gigabeat F dock connector :) (it has the I2S signal)
12:08:17balou_sure, but I'd still like a brand new digital-out sporting rockbox-by-default hires display player more than an old player bought off ebay
12:08:58GodEatersadly no such thing exists - so you have to compromise
12:10:18balou_...or petition a big dap manufacturer
12:10:36scorchepetitions dont work
12:10:36GodEatergood luck with that
12:11:00B4gderI'm sure they just hadn't thought about it so a quick call will fix it
12:11:32linuxstbbalou_: Have you found such a device at all? i.e. one without Rockbox support?
12:12:09 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
12:12:38balou_linuxstb, no, except if you count portable dvd players with digital out
12:12:41 Join DerPapst [0] (
12:12:59markunlinuxstb: according to this the Cowon Q5 will
12:13:03GodEaterwe don't count them
12:13:38 Join moos [0] (
12:14:49linuxstbmarkun: Hmm, a 5" LCD makes it hard to think of it as a portable audio player...
12:14:57balou_wow... a new dap that will actually have digital out
12:15:42linuxstbAnyone know what an "Alchemy" processor is?
12:15:53amiconnmarkun: That was a bad commit...
12:16:05markunamiconn: what happened?
12:16:07*linuxstb googles and sees it's a MIPS based CPU
12:16:08 Join BigBambi [0] (i=86ceaf40@rockbox/staff/BigBambi)
12:16:09balou_okay... a bit big..
12:16:28amiconnmarkun: Check the revision log: r18342...
12:16:34 Join M0rtus [0] (
12:16:55amiconnUnifont *was* already updated to 20080820, but with the unnecessary junk stripped, which you now re-added
12:16:59B4gderfunny how the MIPS arch seems to still survive
12:17:15markunamiconn: oops, sorry :(
12:17:35markunI'll revert at once
12:17:36B4gderQ5 is claimed to run windwos CE too
12:17:51linuxstbSo it's a tablet computer...
12:18:28B4gderyes, with wifi and all
12:19:01B4gderquite an n810 competitor I'd say
12:19:17*linuxstb wonders if we'll ever see hard-disk based portable _audio_ players again...
12:19:38markunamiconn: I just noticed that it had more glyps. I'll be more careful next time.
12:21:10M0rtushello all
12:21:12amiconnSure it had more glyphs than our version... but only ones which you'll never get to see for any valid unicode text. I took 20080820 and stripped those, using a perl script
12:24:29M0rtusguys im a right in thinking that the USB does nothing on the sansa still?
12:24:48BigBambiIt should restart to the Sansa firmware
12:24:52BigBambiAnd then connect
12:26:09ameyeremphasis on "should"
12:26:33*ameyer glares at whatever build server is screwing up portalplayer builds
12:28:44M0rtuslast time i tried rockbox it did just that........... and then would go back to rockbox without a reinstall :(
12:29:04pixelmaameyer: what do you mean?
12:29:32BigBambiM0rtus: It should reboot to the Sansa firmware to let you copy files, then when you restart, start Rockbox like normal
12:30:21M0rtusmight give it another try............ last time it would restart to sansa f/w then copy files......... reboot and sansa firmware still.
12:30:37BigBambiThat shouldn't happen
12:31:01BigBambiAnother way is to boot to the Sansa firmware, then insert USB (manual tells you how to dual boot), or plug in USB from off
12:31:18pixelmathere's a bug with the automatic reboot though that appears in one build and vanishes in the next, but you can always get the OF for USB by either plugging from off state or start the OF manually
12:31:30ameyerI mean some official builds freeze on usb connect but I've never had a build I built myself that didn't restart to OF on usb connect
12:32:13BigBambiAnd you have verified that all the non-working ones come from the same build server?
12:32:37ameyerI'm not entirely sure how I'd do that
12:32:52BigBambiOn the build details page you can see which server built which build
12:32:59BigBambiOtherwise it is pure speculation
12:36:42markunamiconn: git really doesn't like these big file commits :(
12:39:35ameyerthe current mini2g build seems to have the issue
12:39:56ameyerassuming the ipod reboot failure is the same as the sansa reboot failure
12:40:11BigBambiOK, so note down which build server and then try the next few as well to see if it correlates
12:40:20linuxstbameyer: Can you compile that same mini2g build yourself, and test that?
12:40:30soapFront page still needs fixed to explicitly mention that Rockbox does not work on the 4th gen Nano.
12:40:30pixelmaI've experienced myself that using a downloaded build had the problem while compiling my own from the same revision did not. I didn't think it's a buildserver's fault, more like memory alignment or what's it called
12:40:50ameyerlinuxstb: sure
12:41:26 Join Seed [0] (
12:41:28linuxstbpixelma: But you've also experienced the problem with builds you've compiled yourself? i.e. some of your own builds work, some don't?
12:41:46pixelmareason: I have the problem on my c200 with self-compiled... yes
12:44:26ameyererm, false alarm on the mini2g, it works right now
12:44:55preglowwhat's the reason for the codec memory rearrangement?
12:45:41linuxstbpreglow: IMO it's cleaner, and hopefully will mean we can reduce the combined size.
12:45:46 Part M0rtus
12:46:10linuxstbIIUC, prior to that commit, the unused part of the main codec buffer remained unused.
12:48:36ameyeris it in any way possible the sansa charging patch could fix the mystery sansa usb reboot issue?
12:49:05ameyer(and adds a nasty 100% reproducible stkov)
12:49:28preglowlinuxstb: but the total amount, which is the same, still remaions unused...
12:49:32 Join J-23 [0] (
12:49:40preglowwell, not unused
12:49:42preglowbut occupied
12:50:00preglowi hate that malloc buffer, so few codecs now need it
12:50:10ameyerbecause that could also explain why my builds always seem to work
12:50:29ameyererm, seem to always work
12:53:04linuxstbpreglow: I agree - but technically there is now no longer a malloc buffer, just the unused part of the codec buffer. So if we can reduce the amount of malloc'ing the guilty codecs do, we can easily reduce that buffer size.
12:53:31linuxstbi.e. this is just the first step to reducing the memory usage of codecs
12:54:00linuxstbpreglow: I have started a wiki page on this topic here if you have time to contribute -
12:54:49 Quit DerPapst (Read error: 104 (Connection reset by peer))
12:54:57linuxstbameyer: Ah, so you run patched builds?
12:57:20ameyerjust the sansa charging patch
12:58:23ameyerthough, that shouldn't affect the mini2g builds, and I *think* that I've seen official builds work and my build not work on the same revision
12:58:28preglowlinuxstb: ah, i actually collected some data on this, i'll try to update it today
12:58:57linuxstbB4gder: Would you be interested in writing a script to analyse the codec .map files and output the memory usage (code, rodata, bss, icode, irodata, ibss) ? I'm guessing (hoping) it's similar to what you've done for the main builds...
13:00:16 Join kachna [0] (
13:01:14preglowyeah, that would be awesome, the wiki data would be out of date really fast...
13:01:59B4gderthat's a good idea, I'll look into it
13:02:20***Saving seen data "./dancer.seen"
13:04:04linuxstbpreglow: What are the problem codecs? IIUC, mp4 codecs have problems with the seektable info, and Tremor needs to allocate (and init) some large lookup tables, especially for floor0 files. Are there any others?
13:05:01*ameyer wonders how well flake -12 files would work on rockbox
13:05:50*scorche didnt know we had a codec for flake
13:05:52linuxstbameyer: The Rockbox flac decoder only supports the "flac subset" - I've no idea what "flake -12" produces. The main issue is the maximum frame size is something like 4608 samples.
13:06:11linuxstbscorche: It's a third-party flac encoder
13:06:32ameyermaximum frame size could be an issue
13:09:47preglowlinuxstb: vorbis is the major one, afaik
13:09:56preglowlinuxstb: and it's hard to work around, the codec has very few hard limits
13:10:07ameyerapparently at higher compression levels it uses 16k blocks, whatever that means.
13:10:16preglowbut of course, only way to handle it is to find the common limits used in encoders and hard code for that
13:10:52pregloweven the flac author himself says higher block sizes seldom matters
13:11:13linuxstbameyer: That's probably not going to work then - it increases the memory requirements for decoding, meaning those buffers could no longer be in the fast IRAM
13:11:26linuxstb(which is why Rockbox limits to the subset)
13:11:57ameyermakes sense
13:12:52preglowand i really don't think we should fix that
13:12:56linuxstbI can't recall anyone ever complaining either, which would seem to prove that non-subset files are very rare in reality.
13:13:07preglowcodec authors should be aware that there are limits, if they want portable playback
13:13:31preglows/codec authors/encoder authors/
13:14:27ameyerlinuxstb: and lossless/lossless transcoding isn't as big of a deal as lossy/lossy
13:14:39linuxstbameyer: Indeed.
13:15:11ameyerafaik, Trent Reznor is the only person who actually uses flake
13:15:43linuxstbIt could be interesting to port it to Rockbox - I'm sure it's far more understandable than libFLAC...
13:16:00linuxstbBut IIRC, it uses floating point...
13:16:53preglowwell, so does flac, for lpc afaik
13:17:04linuxstbHmm, variable block size encoding sounds interesting... (reading the flake hompage...)
13:18:14preglowsounds like a weird lossless encoder feature
13:18:40preglowhard to imagine why you would want to use shorter blocks, unless there is some kind of switchable mode
13:19:55preglowlinuxstb: i'll update the wiki with my malloc findings right now, some of it might be inaccurate (don't think it is, tho...), but it beats nothing
13:20:07crwlflake seems to use block size of 4608 in all of its presets (at least −−help of flake 0.11 indicates that)
13:20:24crwler, 4608 or lower
13:23:02preglowlinuxstb: didn't ape use to malloc a seek table?
13:24:41ameyererm, what I said about Trent Reznor apparently isn't true
13:26:48linuxstbpreglow: Ah yes, it does...
13:26:55ameyerI have 0.10 and flake claims to onlu use up to 4608
13:28:00preglowdoes? i couldn't find it :/
13:28:04linuxstbpreglow: I don't remember the details, but I expect that can be changed to a sensible sized static buffer, similar to FLAC. I'll investigate.
13:28:18linuxstbIt's in apps/codecs/demac/libdemac/
13:28:43preglowoh, there's another dir in there...
13:28:58preglowbtw, is it "malloc free" or "malloc-free"? as in free of mallocs
13:29:40linuxstbI meant to say it there is no use of malloc or free.
13:30:10ameyerpreglow: flake -h shows a table with equivalent command lines for -0 through -12
13:30:12preglowoh, i meant which of the two alternatives are correct english :P
13:30:53linuxstbpreglow: Probably best to just say "Does not use malloc"...
13:31:12 Join LambdaCalculus37 [0] (
13:31:14preglowyeah, but i'm curious now :)
13:31:29linuxstbI would use a hyphen, but I don't know if that's correct.
13:35:54 Quit kachna (Read error: 110 (Connection timed out))
13:36:25preglowlinuxstb: i'll just put in entries for all the codecs that don't use malloc. they'll have to be added anyway
13:41:04linuxstbpreglow: So AAC itself uses malloc, in addition to the mp4 parser?
13:41:23preglowlinuxstb: it allocates buffers all over the place
13:41:32preglowlinuxstb: asf doesn't touch any malloc, no?
13:41:44linuxstbAs far as I can remember, no.
13:42:22preglowfaad mallocs a lot in ifdef places, and i can't really see which we use
13:42:29preglowneeds to be clarified anyway
13:42:37linuxstbDo speex and vorbis share an ogg parser?
13:42:46preglowdon't think so
13:43:03preglowogg can be handled free of mallocs, but i think it requires some api changes
13:43:14preglowlibogg2 is supposed to deal with that
13:43:25linuxstbDoes that exist?
13:44:16 Join legendsohai [0] (n=legendso@
13:44:18preglowin some form, yeah
13:44:26preglowi believe tremor uses an early version of it
13:44:30 Quit moos ("Rockbox rules the DAP world")
13:44:53 Join pvbcharon [0] (n=charon@
13:45:37linuxstbThe whole Ogg thing seems such a mess...
13:49:10preglowwhat codecs use libm4a? aac? alac?
13:49:38 Join nplus [0] (n=nplus@141.25.Globcom.Net)
13:50:00preglowhow do i make subheadings? how do i link to one? don't have much time, or i'd read up...
13:51:29linuxstbMore "+" signs for subheadings
13:51:47linuxstbYes, aac and alac are the only mp4/m4a codecs
13:53:00 Quit LambdaCalculus37 ("Ka-chunka")
13:53:45 Quit Nibbl (Read error: 110 (Connection timed out))
13:56:44 Join HBK [0] (
13:57:23B4gdermy first codec size overview
13:57:30B4gderfrom an X5 build
13:58:16 Part legendsohai
13:58:45preglowbah, can't make anchors work
13:59:12preglowB4gder: looks sweet
13:59:38linuxstbB4gder: Interesting - looks like nothing is anywhere near the 512KB limit...
13:59:49preglowmalloc use hides a bit
14:00:05preglowa fair bit, in the case of vorbis and aac
14:00:40 Quit HBK- (Read error: 60 (Operation timed out))
14:00:47linuxstbYes, but that's always been in the separate malloc buffer.
14:01:17preglowoh yeah, just meant to say it isn't a good way to see what memory the codecs do need
14:02:06*linuxstb wonders how that info can be presented for the various architectures in a useful way...
14:02:25B4gderthere's a lot of numbers...
14:02:33linuxstbB4gder: Is const data included anywhere in those numbers?
14:02:57preglowargh, isn't using anchors just going like [[#AnchorName][link here] and then putting #AnchorName in the spot you want to link to?
14:03:08B4gderconst would be rodata won't it?
14:03:50preglowlinuxstb: each codec entry would probably just have a bunch of links to files that get updated per target per build?
14:04:13B4gderright, rodata is part of the Text size in that table I believe
14:04:14preglowonly way i can see it work and still be updated automatically
14:04:54linuxstbMaybe the whole page needs to be generated dynamically, and include some kind of figure (or textual description) for malloc usage (taken from a static file)
14:05:51amiconnlinuxstb: Do all malloc()s in question never free() before the codec exits?
14:06:14linuxstbI think the entire malloc buffer is reset after each track.
14:06:17preglowamiconn: we purge the malloc buffer after each track
14:06:19linuxstbi.e. there's no free()
14:06:23preglowso it doesn't matter
14:06:23amiconnIf there are some free() calls, it _might_ help to implement a proper free()
14:06:42amiconn...e.g. dbestfit or whatever that's called
14:06:50preglowit _might_ help to implement a proper "malloc" too, but let's just remove the mallocs instead shall we? :)
14:07:16*B4gder is on the remove-them train
14:07:34*linuxstb lives on no-malloc street
14:07:36amiconnThat would be much preferred. The question is whether it's possible without a major rewriet of the codecs in question.
14:08:25preglowi think it would be very possible
14:08:38preglowi do not believe the codecs to malloc/free on a per frame basis, it's all init/deinit
14:08:57preglowso we purely need to find the static limits that would work, for some codecs this is easy (aac), for some it is hard (vorbis)
14:10:26 Join LambdaCalculus37 [0] (i=44a04303@gateway/web/ajax/
14:10:44 Join Schmogel [0] (
14:11:43B4gderI committed tools/ now
14:13:54 Nick GodEater is now known as Goodie_Godeater (i=c2cbc962@gateway/web/ajax/
14:14:18 Nick Goodie_Godeater is now known as GodEater (i=c2cbc962@gateway/web/ajax/
14:15:20linuxstbIf anyone is interested - results for an ipod build (minus the encoders) -
14:16:21LambdaCalculus37We should have a wiki page for that information.
14:17:08linuxstbLambdaCalculus37: It would be better to generate it from the builds, rather than put it statically in the wiki. But see
14:17:40LambdaCalculus37linuxstb: Are your results for iPods in general, or any specific iPod?
14:18:03linuxstbIt's the ipod color build. I don't know if the codecs make use of the extra iram on nano/video...
14:18:05*GodEater guesses for the photo
14:18:24*LambdaCalculus37 will get some results later on
14:18:31preglowi was about to ask about that, should we create them for all builds, or just find which targets share the same?
14:18:39preglowi guess the last would require maintainance, but be preferable
14:19:18linuxstbpreglow: Is there a reason speex uses very little IRAM? it looks like lots of code could go there (if it helps...)
14:19:52linuxstbAlso, it seems like the entire wavpack codec (code and bss) could fit in IRAM...
14:20:33B4gderin general iram usage seem somewhat low
14:20:56preglowlinuxstb: speex just doesn't need much working ram, and yeah, speex is a work in progress
14:21:12linuxstbpreglow: I was thinking of icode
14:21:22preglowi know, just saying that's the reason it uses little iram
14:21:42preglowstuffing code in iram is always good, and can come up with some candidates for stuffing
14:21:43linuxstbOK - I was just wondering if you had tested and found it to not be useful.
14:21:46preglowbut need to go soon
14:22:01 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
14:22:09 Quit amiconn (Nick collision from services.)
14:22:15 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
14:22:21preglowspeex too will probably fit nicely (at least the parts that matter) in iram
14:23:55*linuxstb wonders what causes wma to use so much bss
14:24:20LambdaCalculus37linuxstb: But SPC uses even more BSS.
14:24:56LambdaCalculus37338784 for SPC versus 149056 for WMA... those are the highest two in terms of BSS usage.
14:24:57linuxstbIsn't that an emulator format though? Meaning it's probably the RAM of the target device?
14:25:38LambdaCalculus37linuxstb: Yes, it's an emulation of the SPC700 chip in the Super NES.
14:41:22 Join tvelocity [0] (
15:02:24***Saving seen data "./dancer.seen"
15:06:32 Join pierre- [0] (n=pierre@
15:10:45 Quit Schmogel (Read error: 104 (Connection reset by peer))
15:11:08 Join Schmogel [0] (
15:12:00B4gderdoes that mean anything?
15:12:38 Join Yskyflyer [0] (
15:13:10 Quit Yskyflyer (Client Quit)
15:15:15 Join toffe82 [0] (
15:15:16 Join evilnick_ [0] (i=0c140464@gateway/web/ajax/
15:22:02 Join lol3izer [0] (n=lol3izer@
15:24:54 Join Kopfgeldjaeger [0] (
15:45:55 Join robin0800 [0] (
15:53:33B4gderthe d2 manual isn't working yet is it?
15:53:36 Join Nibbl [0] (
15:54:27*B4gder read the dev ML post
15:54:56LambdaCalculus37B4gder: I don't think there's any reference or any work done on a D2 manual.
15:55:39*B4gder tries it
15:55:53B4gderFile `platform/cowond2.tex' not found.
16:04:24 Join hogeh- [0] (
16:08:00bluebrotherthe d2 isn't released yet too, right?
16:08:24*bluebrother checks
16:08:46B4gderit's not "supported" at all, no
16:09:19linuxstbThe m3 is missing a manual too...
16:09:45bluebrotheryep. IIRC there was still this issue with the m3 not having a main remote
16:10:08*bluebrother spots that this "don't show metadata lines without info" has a read bin delta :(
16:10:11 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
16:10:22linuxstbSo less is more...
16:11:10bluebrotherme never was annoyed by those empty lines
16:12:16bluebrotherthus I never saw a point in removing those
16:12:55linuxstbIsn't it also useful to know what you're missing?
16:13:31bluebrotherwell, my files are usually well tagged, but for some fields I have no use. Still it's useful to see those fields IMO
16:13:45pixelmawhere have you been yesterday? ;)
16:14:10*bluebrother scratches head
16:15:24bluebrotherI have no too strong feelings about showing or hiding empty fields (though it smells kinda like dynamic menus, and we don't want those ...) but if it's a red bin delta I kinda dislike hiding them
16:15:28*linuxstb reads the logs...
16:15:44linuxstbpixelma: I was there, but missed it. JdGordon kept the discussion brief...
16:17:11linuxstbI almost never use that screen anyway,, so don't really mind. But IMO the desired behaviour is debatable, so should have been debated...
16:17:31pixelmawell, I didn't have too strong arguments and time to discuss (was shortly before lunch break)
16:18:15 Quit vitja_ (Remote closed the connection)
16:20:12*bluebrother agrees with linuxstb
16:20:30 Quit blahrus (Read error: 104 (Connection reset by peer))
16:23:08 Quit Seed ("cu, Andre")
16:26:11 Join blahrus [0] (n=blahrus@
16:28:05 Quit EspeonEefi ("さよなら")
16:29:17XavierGrbluebrother: what's your preffered "fixed" length about 9455? Because as I see it, it would need a big value to please all habits.
16:29:36 Quit pvbcharon ()
16:30:15XavierGrand as I comment, 32 bytes is not something that large-screen targets miss.
16:34:20bluebrotherXavierGr: well, to have a "good" spacing, especially if you use something different than spaces, 5 would be enough IMO −− something like the "+++" winamp uses is the usual thing I could think about
16:34:55bluebrotherand for spaces, well ... if you have 5 user characters and add a space in front and after them you can get to up to 7 spaces. Shouldn't that be enough?
16:36:04XavierGrbluebrother: but thats why this became an issue. Some need a larger padding string.
16:36:39XavierGrI am ok for a fixed value as long it is big to please all, but then small screens will just waste memory
16:36:54bluebrotherone could argue about 7 user defineable characters. But everything else is quite special-case IMO. And we can't care for every special case like people wanting the complete line as padding
16:37:40bluebrotherIMO the number of characters doesn't make much of a difference for the screen size −− on small screens you usually will use a much smaller font, so the number of characters isn't too different
16:38:33bluebrotherpeople repeatedly complained about rockfont on larger screens, so anyone using small fonts on a large screen can be considered a special case. IMHO :)
16:38:58XavierGrof course that's why I am still arguing about it, array[30] from array[LCD_WIDTH / 4] is not more complex but still you get something more for only some bytes
16:39:35XavierGror save something more for smaller screens
16:39:56 Quit Nibbl (Read error: 113 (No route to host))
16:39:58bluebrotherwell, I don't see a point why larger screens should require more *characters*.
16:40:06XavierGrbecause if it is done as a setting to only increase the space from 3 - 7 then it is a waste
16:40:24 Join PaulJam [0] (
16:40:44bluebrotherwell, with that setting you can include other characters, so this definitely makes things more flexible
16:41:20bluebrotherwith that arguing you could consider the whole thing a waste on players with small targets if LCD_WIDTH / 4 comes around as something like 7 ...
16:42:03bluebrotherhmm, while thinking about it ... how making it part of the wps? I.e. define the scrolling line including the padding?
16:42:31bluebrotherthat way one could even use different padding strings depending on tag / subline / whatever
16:42:35XavierGrmy idea first, was to achieve something like iriver scrolling which shows an empty screen at the end of scrolling, using a large pad and a big step size I get this. At the expense of 30-40 bytes
16:42:57 Join Nibbl [0] (
16:43:12XavierGrbluebrother: it is working on wps too, or do you mean as a tag?
16:43:19bluebrotherI mean as tag.
16:43:35bluebrothersomething like %scroll(%artist +++ )
16:43:45bluebrothersorry for the pseudo-wps-code ...
16:44:02XavierGrbut what about the file browser? or you mean haveing both?
16:44:17bluebrothernever thought of lists −− is it really that of an issue there?
16:44:28XavierGrbut having both could be confusing
16:44:32bluebrotherI might use such a feature in the wps but definitely not in lists
16:44:36linuxstbDoes FS #9455 currently affect all scrolling - i.e. in list (file) browser and WPS?
16:45:56linuxstbAnd the binsize increase is still almost 1KB? (I see 848 bytes quoted for an earlier version of the patch on h300)
16:46:02*bluebrother points out that he usually uses settings sane enough to rarely require scrolling −− slow lcds make scrolling a pain
16:46:16XavierGrlinuxstb: nope it should be rather small now
16:46:19XavierGrI should check
16:46:29*XavierGr checks
16:46:37lol3izerwhy are you guys so obsessed with binsize?
16:47:06bluebrotherlol3izer: because more RAM for buffering means less disc spinups. Less disc spinups mean longer battery runtime
16:47:27lol3izerbut a few K
16:47:29lol3izeris nothing
16:47:30bluebrotherand if we didn't care about it the binary would explode rather quickly.
16:47:43LambdaCalculus37lol3izer: Every single K counts.
16:47:43bluebrotherI've seen this at work ...
16:47:53 Quit Zagor ("Client exiting")
16:48:12bluebrotherevery byte counts ;-)
16:48:19linuxstblol3izer: Various reasons. Avoid inefficient code, avoid overly-complicated code, avoid feature bloat
16:48:43bluebrothernow consider some of the old targets (archos) and never targets have not much RAM, 1MB or even less
16:49:19bluebrotherthat's not much given the fact that we need codecs, filesytem access, a customizable UI, ...
16:49:37LambdaCalculus37The Archos targets have only 2MB of RAM; some like the iFP-7xx, Clip, m200 and others have even less.
16:49:39bluebrotherand we do want to use RAM for buffering ...
16:50:07bluebrotherah, right, the archos were 2MB. Still not much
16:50:39PaulJamHi, it looks like test_codec is broken in current SVN. (probably related to the changes in r18834)
16:50:54LambdaCalculus37PaulJam: Which target did you try it on?
16:51:10PaulJamit doesn't compile
16:52:01PaulJamtest_codec.c:430: error: structure has no member named `get_codec_memory'
16:52:16 Join {phoenix} [0] (
16:52:50XavierGrlinuxstb: it dropped to 700 for H300. I hoped for something more
16:53:26*bluebrother starts to kinda like the idea of padding in the wps tags
16:54:35XavierGrstrnage... since you were the father of the previous idea
16:54:37pixelmawhy? You can just add your "padding string" as plain text in the WPS
16:55:04XavierGrhmm indeed
16:55:24bluebrotherXavierGr: why? I just expanded that idea ;-)
16:56:13XavierGrhehe. But indeed in wps you can alter the scrolling line with any text string you like
16:56:27linuxstbPaulJam: I expect you just need to rename it to codec_get_buffer()
16:56:34XavierGrthe line in general though, not the "scrolling" per se
16:56:36bluebrotheris this already possible? Not really sure about that
16:57:03linuxstbPaulJam: I'll test and commit, thanks for reporting.
16:57:06XavierGrbluebrother: yeah you can insert any text you like before and after, though it will appear even if the line isn't scrolling
16:57:29PaulJamlinuxstb: ok, thanks. i wasn't sure, because the commit message sounded as if there were functional changes too.
16:57:35bluebrotherwell, then this isn't what I've thought.
16:59:53XavierGrit is quite similar though
17:02:26***Saving seen data "./dancer.seen"
17:04:39 Part hogeh-
17:06:05 Join Seed [0] (
17:10:45linuxstbPaulJam: A quick fix is possible, but I think test_codec needs more changes to accurately emulate the new way codecs use memory. So I'll leave it for now...
17:11:01linuxstbPaulJam: But if you need to use it, the search/replace I suggested should work.
17:12:39PaulJamlinuxstb: i don't need test_codec. i just had it enabled in my build, so i noticed it.
17:19:07 Join Arathis [0] (
17:19:57 Join bughunter2 [0] (n=Jelle@
17:21:01 Quit bughunter2 (Client Quit)
17:21:47 Join bughunter2 [0] (n=Jelle@
17:24:33 Join n1s [0] (n=nils@rockbox/developer/n1s)
17:24:50 Part johwil
17:25:47 Quit sarixe (Remote closed the connection)
17:26:21 Join sarixe [0] (
17:28:55 Join kachna [0] (
17:42:18 Join Hillshum [0] (
17:42:43 Quit BigBambi (" ajax IRC Client")
17:52:46 Quit evilnick_ (" ajax IRC Client")
18:00:18 Quit Rob2223 (Read error: 104 (Connection reset by peer))
18:07:12 Join Rob2222 [0] (
18:09:12 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:09:43 Quit Rob2222 (Read error: 104 (Connection reset by peer))
18:10:22 Quit domonoky (Client Quit)
18:10:50 Quit petur ("work->home")
18:11:44 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:12:47 Join nuonguy [0] (
18:12:51 Part B4gder
18:14:45 Quit robin0800 (Remote closed the connection)
18:15:25 Join funman [0] (n=fun@
18:18:49 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:20:57funmanisn't this test run automatically on each commit (on the builders) ?
18:21:46 Join Rob2222 [0] (
18:23:43 Join Tetracomm [0] (n=nicholas@
18:24:12 Quit rvvs89 (Read error: 60 (Operation timed out))
18:25:23 Join lol3izer_ [0] (n=lol3izer@
18:25:35 Quit Arathis ("Bye, bye")
18:25:45 Join rvvs89 [0] (
18:26:05pixelmatest_codec is a plugin which is not part of the downloadable builds (hence not tested if it still builds)
18:29:36funmanah I thought it was a kind of test suite runnable on the build host
18:30:28 Join miepchen^schlaf [0] (
18:30:51domonokytest_codec is a plugin to test codecs (especially decoding speed) on target..
18:31:42 Join Bagder_ [0] (
18:32:07 Quit Rob2222 (Read error: 104 (Connection reset by peer))
18:32:40 Nick Bagder_ is now known as Bagder (
18:32:46funmanexcuse me I never used rockbox ;)
18:33:08LambdaCalculus37funman: You're excused. :)
18:36:38 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
18:41:35 Join stoffel_ [0] (
18:42:55 Quit lol3izer_ ()
18:43:32 Quit lol3izer (Connection timed out)
18:46:14funmanis the m200v2 supposed to embed SDRAM ?
18:46:31funmanthe memory setup routines are 100% identical between Clip & M200
18:47:18 Quit miepchen^schlaf ()
18:47:32funmanand read timings from uninitialized stack apparently :'(
18:48:37 Join miepchen^schlaf [0] (
18:50:13J-23What are the video pins on Sansa e200 USB connector?
18:51:44linuxstbJ-23: I think there's a wiki page describing the Sansa connector
18:52:45 Join Rob2222 [0] (
18:52:56linuxstbfunman: I've no idea. What about the e200v2? I couldn't see any RAM chips on the photos linked from the wiki.
18:53:11funmanand the e200 routine is 99.9% identical (one setting has another value)
18:53:32vitjaI've mounted iaudio7!!
18:53:45funmanvitja: great \o/
18:54:21 Quit Rob2222 (Read error: 104 (Connection reset by peer))
18:54:29vitjayeah but a bit hacky wo interrupts for tx
18:55:24 Quit Schmogel ("Miranda IM! Smaller, Faster, Easier.")
18:59:21 Join mf0102 [0] (
19:01:47funmanI found still the same code in Clipv2 OF
19:01:51linuxstbvitja: Congratulations
19:02:15gevaertswell done!
19:02:24vitjastill isn't done ;)
19:02:27***Saving seen data "./dancer.seen"
19:02:56vitjaI'll commit workaround soon, so you can test it
19:05:39 Quit Exorcist- (Read error: 110 (Connection timed out))
19:07:44 Join Rob2222 [0] (
19:08:13 Join karashata [0] (n=kimi@
19:11:02 Join culture [0] (
19:11:46 Join bertrik [0] (
19:13:23 Quit Rob2222 (Read error: 104 (Connection reset by peer))
19:13:31 Quit Seed ("cu, Andre")
19:13:57 Join Seed [0] (
19:14:55 Join hannesd [0] (
19:15:04 Join kharo1 [0] (
19:16:05 Quit kharo (Read error: 110 (Connection timed out))
19:21:22 Join domonoky1 [0] (
19:21:26 Join Rob2222 [0] (
19:22:17 Quit Rob2222 (Read error: 104 (Connection reset by peer))
19:22:57 Join Rob2222 [0] (
19:23:16 Quit Seed ("cu, Andre")
19:24:02 Join Thundercloud [0] (
19:24:09 Quit karashata (Read error: 113 (No route to host))
19:24:17 Join karashata [0] (n=kimi@
19:30:17 Quit Rob2222 (Read error: 104 (Connection reset by peer))
19:30:28 Quit karashata (Read error: 113 (No route to host))
19:30:54 Join karashata [0] (n=kimi@
19:33:20 Quit sarixe ("Ex-Chat")
19:37:36 Join Rob2222 [0] (
19:40:16 Quit domonoky (Read error: 110 (Connection timed out))
19:41:24amiconnRegarding the test plugins - maybe we should have a 'developer build' option in configure, which will enable all the test stuff?
19:42:07amiconn(perhaps a sub-option for advanced builds)
19:42:22gevaertsgood idea
19:42:39amiconnThen the build system could produce such builds, and we would see whether the test stuff still compiles
19:42:54LambdaCalculus37amiconn: Good call.
19:43:09amiconnIt would mean that the master would need to strip the test stuff from the .zip files before putting them into the download folder
19:43:48amiconnHmm, or perhaps just build them, but don't include them in the .zip? Saves some upload time (build server -> master) as well
19:45:07amiconnThe exclusion could be part of the standard 'make zip', 'make 7zip' etc. There could be a special 'make devzip' which includes them
19:46:24*LambdaCalculus37 opts to have a developer option added to configure, alongside normal, debug, et. al.
19:46:56 Quit Rob2222 (Read error: 104 (Connection reset by peer))
19:47:52gevaertsThey could get a separate category. That would make it trivial to exclude them
19:47:59 Quit stoffel_ (Read error: 113 (No route to host))
19:48:30amiconnWell, there are 3 cases. (1) A user who wants to build his own rockbox (perhaps with a few patches). The test_* plugins should neither be built nor zipped. (2) A dev who wants the test_* plugins. In this case they should be built and zipped. (3) The build system. Here the test_* plugins should be built but _not_ ziiped
19:49:46amiconnMy suggestion is to make building of the test_* plugins a configure option, and zipping them a separate make target
19:50:34linuxstbWhat's the point in not building them? They're all tiny compared to the rest of the plugins needed to be built.
19:51:00 Join stoffel_ [0] (
19:51:02amiconnBuild time
19:51:11*linuxstb thinks the build system is complicated enough already...
19:52:17gevaertsHow big would the build time difference be in practice?
19:52:49 Join Rob2222 [0] (
19:53:44amiconnI didn't measure. I usually do build all test plugins, even on cygwin
19:55:27amiconnActually I don't. I only build 6 of them
19:55:28 Join nanok [0] (n=nanok@
19:55:38*gevaerts now has a dump of the DAX nand flash
19:55:52linuxstbgevaerts: Cool... The whole 1GB?
19:55:56 Join tessarakt [0] (
19:56:08gevaertsyes. As correct as the current tcc nand driver allows
19:56:17*nanok salutes respectfully
19:56:19linuxstbHow long did that take?
19:56:34gevaerts303.998 seconds
19:57:00 Join Siku [0] (
19:57:03*domonoky1 loves exact statements..
19:57:49gevaertsSpeaking of exact, this is actually exactly 10^30 bytes :)
19:58:07gevaertsvitja: usb seems to work on the dax
19:58:38 Quit stoffel_ ("Reconnecting")
19:58:48 Join stoffel_ [0] (
19:59:13linuxstbgevaerts: That will make a nice recovery tool, assuming you can do the reverse...
19:59:54 Join peerlessdeepak [0] (n=peerless@
19:59:54bluebrothergevaerts: so it's 1GB ;-)
20:00:15gevaertslinuxstb: indeed. I
20:00:18vitjagevaerts: problem was TX fails on big(>512) packets, RX will not work now for packets >512
20:00:31gevaertsI'm not trying to writr yet though
20:00:38 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
20:00:59vitjahope that can help with TNFL
20:01:07gevaertsvitja: >512 byte transfers I guess. Packets are always 512-bytes max
20:01:46vitjayes, but some times they are fragmented
20:02:13vitjae.g. when host asks for 2 blocks usb_drv_send is called with len 1024
20:02:14linuxstbvitja: shotofadds was here recently, sounding confident about breakthroughs with the NAND. Hopefully this will help him too.
20:02:56gevaertsYes. usb_drv_send is for an entire transfer
20:03:14vitjaso entrie transfer sometimes fails
20:03:15 Quit Thundercloud (Read error: 104 (Connection reset by peer))
20:04:16 Join jgarvey [0] (
20:04:34 Quit peerlessdeepak (Remote closed the connection)
20:05:36vitjalinuxstb: nice
20:07:01denes_btw how did somebody find out how the telechips_nand works? reverse engineering?
20:07:35linuxstbdenes_: Yes - I think work started almost a year ago...
20:07:47linuxstbBut only one person has really been working on it afaik
20:08:15*gevaerts tries to find a FAT header in the dump
20:08:18denes_i see. the meizu m3 does something funky too (different from the tcc7XX)
20:11:10 Nick fxb is now known as fxb__ (
20:11:15 Quit stoffel_ ("leaving")
20:12:04 Quit Hillshum ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
20:12:57 Join Hillshum [0] (
20:14:56denes_so on the ipod nano there is some kinf of an ata-nand bridge?
20:15:12 Quit Rob2222 (Success)
20:15:16LambdaCalculus37denes_: Which generation of nano?
20:15:36denes_LambdaCalculus37: only one generation is supported, that generation
20:15:37linuxstbdenes_: Yes - for the Nano Rockbox supports.
20:16:04 Quit GodEater_ (Remote closed the connection)
20:16:05LambdaCalculus37denes): Ahh, you mean first. I thought you were asking about other generations.
20:16:08denes_linuxstb: interesting.
20:16:13*LambdaCalculus37 slaps his tab key
20:16:33 Join Schmogel [0] (
20:16:56 Quit bmbl ("Woah!")
20:17:04denes_LambdaCalculus37: yes, sorry I realize I was a bit vague
20:18:36 Quit homielowe ()
20:19:00linuxstbLooking at the 5g resource files with readelf, it just says "Machine: <unknown>: 0x5f"
20:20:40bertrikdoes anyone know how the oled brightness on the clip is varied?
20:21:01 Join anselm [0] (
20:22:33funmanlinuxstb: what about using OSX's readelf ?
20:22:58funmanbertrik: the command should be defined and used in the init routine
20:23:00 Join Rob2222 [0] (
20:23:32bertrikI do have a clue and I'm about to try it out now
20:23:34Hillshumwho admins flyspray?
20:24:06BigBambiThe Swedes
20:24:16BigBambiShout at Bagder :)
20:24:20n1sSo about the crypto on newer ipods (like nano 2g) i find it interesting that there seems to be no reports of people trying to brute force it... IMHO it seems it could be worthwhile to try brute forcing the RC4 crypto as:
20:24:26*Bagder covers ears
20:24:28funmanhow many people are 'the Swedes' ?
20:25:30BigBambifunman: Three
20:25:35funmanBagder: shouldn't rockbox website / flyspray specify unicode character set ?
20:25:40BigBambiBagder, LinusN and Zagor
20:25:50domonoky1our 3 swedes are: Bagder, Zagor and LinusN, hooray to them.. :-)
20:25:52Bagderfunman: yes, "should" being the keyword there
20:26:13Bagderwe are more dev swedes than us though
20:26:15BigBambidomonoky1: There are only three Swedes as far as I am concerned :)
20:26:28funmanis the website static? in svn?
20:26:41Bagdermost is static and in svn, yes
20:26:56BigBambi(apologies to the other Swedish people here ) :)
20:27:07*funman is going to flame webmaster@rockbox.o
20:27:23n1s1) Apple used it on ipod flash images before, 2) (AFAIK) noone else that uses the same SoC has encrypted firmware which suggests that apple did it themselves 3) analysis supports the theory that it is a stream cipher 4) if the key is short it should be doable (32 bits used on earlier flash images) reports of up to 48 bti keys brute forced in less than two weeks on distributed systems
20:27:41Hillshumso flyspray seems to request unicode, rb home doesn't
20:28:07Bagderfunman: feel free, that address gets lot of crap ;-)
20:28:07funmann1s: as long as you know the stream cipher algorithm
20:28:08bertrikany idea where the key is hidden?
20:28:40funmanor maybe bruteforce all algorithms you know until you can grep 'Apple' ;)
20:29:18Bagderbertrik: if it's an asymetric algorithm the generating key isn't to be found, only the decrypt one
20:29:41Bagderwell, until you brute-force that too
20:30:17n1sfunman: i would suggest starting with rc4 at least and also there should be no problem to find unique strings to search for...
20:30:26 Quit Rob2222 (Read error: 60 (Operation timed out))
20:30:40bertrikasymmetric sounds a bit unrealistic (and perhaps quite slow?). I assume they want to protect the firmware from being decrypted, not from being encrypted.
20:31:08*Hillshum seconds bertrik
20:31:10n1sbertrik: if we can't encrypt we probably can't run any code so...
20:31:39*n1s Votes for Mr Someone to try brute forcing the thing :)
20:31:56 Quit J-23 (Remote closed the connection)
20:32:23n1sOr if someone sends me a new ipod and a _really_fast computer i promise i'll try it :)
20:33:35funmanBagder: unless you want to replace the code by yours
20:33:45 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
20:34:02*funman gives some quarks to n1s
20:34:31*shotofadds 's ears are burning
20:34:39n1shmm quarks tastes funny...
20:34:48*Hillshum has access to a 3g nano...
20:36:42funmanhm I'm looking at the Clipv2 again
20:37:04funmanIt setups stacks past 0x50000 (more than 320kB)
20:37:20Hillshumhow different from v1?
20:38:16funmanthe v1 uses stack (and everything) below 0x50000
20:38:29funmanbut I just read that you can remap external memory (up to 4MB) at address 0
20:38:56funmanthat means that when booting the external memory is mapped at 0 (not the Embedded 1T-RAM of 320kB)
20:39:13shotofaddsgevaerts: about the tcc nand driver, I've a better understanding of how the various 0x13,0x15 and 0x17 block types fit together now (I'll add this info to the wiki at some point), but unfortunately the results from my "super improved" driver weren't quite as good as expected.
20:39:23shotofaddsI'll produce a patch for you to try later, though
20:40:13gevaertsOk. My current dump doesn't make much sense
20:40:24bertrikok, it seems clip oled brightness is not controlled by the dcdc15 voltage as it is on the c200v1 and e200v1
20:41:30funmanbertrik: brightness != contrast ?
20:41:54domonoky1funman: would this mean the external memory is already mapped when we branch to rockbox ?
20:41:55 Join n17ikh|Lappy [0] (
20:42:14funmandomonoky1: Yes, but I'm talking about Clipv2 here, not v1
20:42:26domonoky1ah, different clips.. :-)
20:47:38gevaertslinuxstb: I'd send you the compressed dump, but after dd-ing /dev/zero to a file via the OF and making a new dump, bzip2 compresses it to 7300 bytes so I'm pretty sure that this isn't very useful
20:47:51 Quit kharo1 (Read error: 60 (Operation timed out))
20:48:31 Join kharo [0] (
20:50:26 Quit pierre- (Read error: 110 (Connection timed out))
20:50:34 Join fml [0] (n=4fd3cefa@gateway/web/cgi-irc/
20:51:07fmlXavierGr: ping. Want to talk about FS #9455?
20:53:18 Join Rob2222 [0] (
20:53:48 Join Exorcist- [0] (
20:54:46amiconnlinuxstb: Afaics it's neither mips, nor arm, sh or m68k
20:56:09n1samiconn: is the flashing chapter in the manual in a reasonable state now? (i.e. can FS #5953 be closed?)
20:56:50linuxstbgevaerts: That sounds odd - the nand flash should contain the firmware.
20:57:16gevaertslinuxstb: I blame the FTL in svn
20:57:27gevaertsLet's see what happens with shotofadds' new version
20:58:00 Join homielowe [0] (
20:58:19 Join DerDome [0] (
20:58:49fmlbluebrother: you here? Have you seen my last comment in FS #9455? Do you agree with all points?
20:58:57linuxstbgevaerts: Ah, I thought you were dumping the raw NAND pages?
20:59:05funmanbertrik: IntBootSel (xpc[0]) selects ROM or external memory on the AS3525
20:59:13shotofaddsgevaerts, linuxstb: if your dump is from the rockbox driver it won't contain the firmware. that's in a separate, non exposed, area
20:59:17 Join MarcGuay [0] (
20:59:20amiconnn1s: Yes. The wiki pages still need to be updated. I'll remove the basic flashing instructions as they are covered by the manual now, and collect technical background and faq-like stuff there
20:59:36shotofaddsdump that raw NAND would be useful, though
20:59:47n1samiconn: great, I'll close that bug then
21:01:27bertrikfunman, what do you mean to say with that? Maybe if clip v2 references "external memory" that may actually be extra RAM inside the chip, it would also allow us to boot from that RAM?
21:01:51gevaertsshotofadds: sure, how?
21:02:23funmanbertrik: no, see page 188 of the AS3525 datasheet, this setting determines what is mapped at address 0
21:02:28***Saving seen data "./dancer.seen"
21:03:21shotofaddsgevaerts: I would guess some combination of nand_read_raw() and black magic ;)
21:03:35 Quit Rob2222 (Read error: 104 (Connection reset by peer))
21:04:16MarcGuayTargetStatus could use some love from people in-the-know on the Meizu, m:robe500, GoGear HDD, and Elio. Also wondering if a "Boots Rockbox" column is needed...
21:04:28 Join Rob2222 [0] (
21:05:02BigBambiMarcGuay: Only if you define what boots Rockbox means
21:06:00bluebrotherfml: well, I don't see a reason why someone would want to stick the first letter to the last, so why no automatic space?
21:06:14funmanI have a problem I can't solve
21:06:27 Join Rob2223 [0] (
21:07:08funmanAS3525 ships with a ARM922T CPU, which is in the armv4 family. the BLX instruction appeared in armv5T family. The Clipv2 uses the BLX instruction.
21:07:24linuxstbMarcGuay: The phrase "currently built by the Rockbox build system" isn't true for all of those targets. See - maybe that should be a column as well.
21:07:47linuxstbMarcGuay: Plus the DAX can dual-boot, and the hardware can't charge, so the charging column should be "N/A"
21:07:58linuxstbSimilarly for the m200v1
21:08:23bluebrotherdoes gcc allow creating arrays with a variable as size when putting them on the stack?
21:08:25bertriknooooo, sandisk must have sensed we were about to port rockbox to a clip so they quickly started working to make it obsolete!
21:08:42linuxstbbluebrother: alloca?
21:09:26bluebrotherlinuxstb: I'm referring to 2. of fml's last comment in FS #9455
21:09:42funmanlinuxstb: no, char mystring[i];
21:10:06linuxstbfunman: afaik, you can't do that.
21:10:17funmanjust try
21:10:22linuxstbWhich is why I suggested alloca. But I may very well be wrong...
21:10:33linuxstbWhen/where is i defined?
21:11:27 Join dany_21a_ [0] (
21:12:00 Join AndyIL [0] (i=AndyI@
21:13:09fmlbluebrother: because it makes more things possible at less cost (the code becomes simpler)
21:13:55gevaertsn1s: FS #9380 was masked on arm by the crash bug I guess
21:14:08n1sgevaerts: very likely
21:15:16*linuxstb should probably read a book about the other things invented by C99....
21:16:09MarcGuaylinuxstb: Does the m200 run off of AA batteries? Or did you mean that it can dual-boot?
21:16:18funmanlinuxstb: I think there is a (shorter than the standard) document which lists those
21:16:34bluebrotheramiconn: thanks.
21:16:51funmanMarcGuay: can you recharge the AA batteries with USB ?
21:17:21bluebrotherMarcGuay: it uses AAA batteries
21:17:34MarcGuayfunman: Not as far as I know. bluebrother: Thanks.
21:17:37funmanlet's make a deal on A*
21:17:45gevaertsshotofadds: you could actually make a special mode that exposes the translated device as LUN 0 and the raw nand blocks as LUN 1
21:18:32*bluebrother is not sure if he likes this C99 feature in gcc
21:18:55linuxstbMarcGuay: Both (well, AAA). I'm pretty sure the m200 can't charge them.
21:19:29LambdaCalculus37The m200 can't charge the batteries.
21:19:31MarcGuayThings get confusing with the m200.. The v1 (NAND) can dual-boot but the v? (HARP) can't?
21:20:50linuxstbMarcGuay: dual-boot really (IMO) refers to being able to load the OF. They should all be able to do that, as the bootloader is patched into the OF image, and then that modified image is flashed using the normal firmware upgrade facility in the OF
21:20:51LambdaCalculus37MarcGuay: The v3 (HARP)? I'm not even sure if it can; I haven't tried.
21:21:23linuxstbMarcGuay: Being able to load Rockbox is obviously dependent on a NAND driver - and you have a separate column for that.
21:21:55 Quit Exorcist- ()
21:23:12MarcGuayAnd we can't refer to the first three types as v1,v2,v3 without confusing them... Okay, I see what you're saying...
21:23:27 Quit AndyI (Read error: 110 (Connection timed out))
21:25:15 Join Strife89 [0] (n=michael@
21:25:30LambdaCalculus37Speaking of the bootloader patched into the OF image, is it possible to do this already with the m200/c100 bootloaders? Or is there something still missing for that to happen?
21:26:54 Quit markun ("leaving")
21:27:03 Quit Rob2222 (Connection timed out)
21:27:10 Quit nplus (Remote closed the connection)
21:28:16 Join m0f0x [0] (
21:31:50linuxstbLambdaCalculus37: Yes, I think it's possible. You need to #define TCCBOOT to build a dual-boot bootloader, and then patch an OF with tools/mktccboot. But I'm not 100% sure if crt0.S has the correct dual-boot button check for the m200...
21:33:46funmanIn Clipv1 OF external memory is initialized just after NAND, and just before LCD
21:35:06funmanI have figured part of the settings, but I have trouble understanding how the remaining ones are read from the stack ..
21:35:46shotofaddsMarcGuay: the v3 (HARP) doesn't really have a port of its own right now. it really needs a new configure target, and someone to work on the SD interface
21:36:22shotofaddsgevaerts: yes, something like that might work, and shouldn't take too long to implement. shame I'm in the middle of something else right now...
21:37:33 Join midkay__ [0] (
21:38:09 Quit gevaerts (Read error: 113 (No route to host))
21:38:37LambdaCalculus37shotofadds: HARP is similar to the e200/c200 way of accessing the flash, isn't it?
21:38:58 Quit shodanX (Read error: 60 (Operation timed out))
21:39:03 Join shodanX [0] (
21:39:37 Quit fml ("CGI:IRC")
21:39:39shotofaddsLambdaCalculus37: yes, it's that SD/NAND bridge. preglow's SD driver should work just fine, if and when he gets it working
21:40:07*LambdaCalculus37 looks in preglow's direction
21:40:15 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
21:40:22 Quit karashata ("I go, only to return again some time...")
21:41:09n1sfunman: regarding your blx question, the AMS3530 uses a ARMv5TEJ core, maybe they switched to that?
21:41:51shotofaddsLambdaCalculus37: "someone" will still have to port that SD driver to tcc77x
21:42:02shotofaddsmy "should work just fine" was probably a bit hasty ;)
21:42:05funmancould be, the OF still mentions "AS3525_2_0.cr_5_0_develop."
21:42:20linuxstbgevaerts: Does this mean that logf-over-usb-serial is close to reality on the telechips targets?
21:42:28funmancould be the AS3525 devkit with a different -mcpu option for gcc
21:42:47funmanbut then AS3525 an 3530 have to be VERY similar
21:42:58 Quit midkay_ (Read error: 60 (Operation timed out))
21:43:11gevaertslinuxstb: I think so, yes. Let me try :)
21:43:56linuxstbDo we know the exact markings on the AMS chips inside each device?
21:45:31funmanno, on the Clip the chip reads 'SanDisk' not AMS
21:45:32bertrikI haven't seen any scans or photos of v2 clips
21:45:54 Join markun [50] (n=markun@rockbox/developer/markun)
21:47:08 Quit Hillshum (Read error: 104 (Connection reset by peer))
21:49:26 Quit bughunter2 ("bye")
21:50:03 Join bughunter2 [0] (n=Jelle@
21:51:48 Quit shotofadds ("Leaving")
21:56:57 Quit kharo (Read error: 110 (Connection timed out))
21:57:25 Quit funman ("leaving")
21:58:59 Quit LambdaCalculus37 (" ajax IRC Client")
21:59:23 Join kharo [0] (
22:00:38 Join shotofadds [0] (
22:01:52 Quit MarcGuay ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
22:02:36 Quit massiveH ("Leaving")
22:04:08 Join petur [50] (n=petur@rockbox/developer/petur)
22:04:21 Join funman [0] (n=fun@
22:04:42funmanjust wanted to tell you I can write at 0x30000000 which is the Clip SDRAM (going to verify its size)
22:07:48funmanmy incomplete code was correct, just forgot to use system_init() in the bootloader ><
22:08:37 Join Acky [0] (
22:11:00funmanhmm the memory is quite slow in my opinion, may need tweaking
22:11:36linuxstbNice work. Is the cache enabled?
22:11:44funmanthe what ? ><
22:11:57 Quit Acksaw (Read error: 60 (Operation timed out))
22:12:48XavierGrwhich is faster? memcpy or strncpy?
22:13:10funmanmemcpy obviously (no \0 check)
22:14:18XavierGrand if I make sure that the string has \0, what is more suitable strcpy or memcpy on the whole array?
22:14:41funmanmemcpy() with a nice comment why you don't use str*cpy ;)
22:15:33XavierGrI used strncpy with the 3rd parameter strlen(string) + 1 in order to avoid writing the whole array
22:15:38XavierGrbut I was told it is not so good
22:15:55linuxstbIf there is no trailing 0, what will strlen return?
22:16:06XavierGrthen yeah you will have a problem
22:16:15Unhelpfullinuxstb: don't do that. ;)
22:16:24Bagderand if you do a strlen, you can just as well use memcpy afterwards
22:16:36funmanhmm according to my test there is more than 8MB of RAM in the Clip (didn't found the upper limit yet)
22:17:00bertrikoh nice, but did you check for aliases?
22:17:11gevaertsAre you sure it's not just wrapping around?
22:17:22linuxstbfunman: Any idea what the extra 10MB of clipv2 firmware contains?
22:17:35XavierGrBagder: imagine a char array of 10 characters, but the string in question might be less e.g 5. Shouldn't be more efficient to just call strcpy?
22:17:35funmanI use the alias for MPMC Bank 4
22:17:56funmanlinuxstb: code, the firmware block is > 320kB and so it has to be in external RAM (which is mapped at 0)
22:17:57BagderXavierGr: yes
22:17:59XavierGrBagder: instead of memcpy or strncpy on the whole array?
22:18:13funmanbut that doesn't explain the large overhead with no new features advertised
22:18:18funmanmaybe it's a debug build ;)
22:18:45XavierGrI see, so strcpy would be better if I make sure that the \0 always exists
22:19:51 Join massiveH [0] (
22:20:01linuxstbIf the array is only 10 bytes, don't worry about speed - just write something that's safe and clear (e.g. strncpy)
22:20:32XavierGrlinuxstb: it can be as much as 60 butes actually (but lower too on other targets)
22:21:34XavierGrlinuxstb: can you please check #9455 in scroll_engine.c and display_menu.c, I was told by fml that strncpy with strlen is bad
22:22:28XavierGror does he mean that if you call strncpy the check for \0 is done internally so the third argument needs only to be sizeof the arrya?
22:22:33XavierGrarray too
22:23:24*linuxstb isn't sure he wants to get involved with 9455
22:23:42n1sXavierGr: using the array size is safer since strlen will keep counting until \0 so if the string wouldn't be terminated it would stop in some random place but if you use the array size it will at least stop there
22:24:03XavierGrn1s: yup now it is more clear thanks
22:25:25*domonoky1 thinks you should use strncpy(dest,source,sizeof(dest)); dest[sizeof(dest)-1] = '\0' to be always safe .. :-)
22:26:10Bagderif 'dest' is reasonably small I'd do it similar to that too
22:26:20*linuxstb was about to suggest what domonoky1 suggested, but would write it "= 0", not ='\0'
22:26:38domonoky1(if source is longer then dest, dest isnt null-terminated)
22:27:04XavierGrdomonoky1: in my case it can't be
22:27:27funmanweird I used a magic to check memory wrapping and I'm already past 4MB
22:27:45domonoky1XavierGr: its always better to be safe, than sorry. you dont know what future modifications will do..
22:27:49 Quit {phoenix} (Remote closed the connection)
22:27:50funmanmaybe if the timings are wrong the data isn't correct
22:28:39 Quit kharo (Read error: 110 (Connection timed out))
22:29:47 Join kharo [0] (
22:29:47bertrikfunman, you could just write each SDRAM location with it's own address, then check all locations again
22:30:22 Quit perrikwp (" ajax IRC Client")
22:30:25 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
22:30:58XavierGrdomonoky1: sure, it is just that source points to an array which is the same length as destination (and they should be like that in any modification), but I understand your point
22:31:06funmanbertrik: thanks
22:32:58*bluebrother was thinking to simply memset(dest, 0, sizeof(dest); memcpy(dest, src, sizeof(dest) - 1);
22:37:32Bagderthat's slower though
22:37:49XavierGrbut that would write the whole array two times
22:38:28Bagderremember that strncpy() zero-pads as well
22:41:41bluebrothertrue. But what happens if src is shorter than dest and dest has an old value in it?
22:41:56bluebrotherargh, overlooked that it's a local array. Ok, scrap that
22:42:15funmanif I read continuously a word in memory, its value stays valid
22:42:33funmanif I loop over a bigger part, the values disappear :/
22:42:40 Quit anselm ("Ex-Chat")
22:42:54bertrikbus capacitance maybe?
22:43:17funmanI don't know what you are talking about
22:43:34funmanI'll try to understand what I'm doing, but tomorrow ;)
22:44:02Bagdersounds like bad memory refresh or similar
22:44:53 Quit funman ("leaving")
22:46:26bertrikif you read very quickly after writing, the data value can be remembered by the capacitance of the data wires, so the data bus itself can act as a kind of memory (for just 1 word of data...)
22:47:23domonoky1bertrik: but if it the bus, it should only work for 1 word. even 2 would destroy it..
22:47:41Bagderbut if the value stays "if I read continuously" it is something else
22:48:00Bagderto me it sounds exactly like bad memory refresh
22:48:03bertrikok, agreed
22:51:10 Join sarixe [0] (
22:56:30 Quit m0f0x ()
22:58:43 Quit domonoky1 (Read error: 104 (Connection reset by peer))
23:01:44 Quit massiveH ("Leaving")
23:02:31***Saving seen data "./dancer.seen"
23:04:54 Part dany_21a_
23:05:27 Quit mf0102 (Remote closed the connection)
23:05:33 Join dany_21a_ [0] (
23:06:45denes_shotofadds: so you wrote the telechips nand ftl stuff right? you "simply" reverse engineered the OF? how long did it take?
23:10:37shotofaddsdenes_: no, the tcc nand driver was initially written by examining raw nand blocks and trying to piece them together into contiguous data. Since then I've been trying to understand the OF code for almost a year now. Hope that doesn't put you off ;)
23:10:58 Part dany_21a_
23:11:35denes_shotofadds: ;) well... I looked at the meizu raw nand blocks, and couldn't yet make any sense of it.
23:12:37denes_and which device uses firmware/drivers/ata_flash? because that's a similar idea
23:12:43shotofaddsI hope the meizu has a larger screen than the d2. maybe you can display a whole sector at a time :)
23:13:11gevaertsdenes_: that's iriver ifp7xx
23:13:16denes_shotofadds: yeah right :) I only cared about the additional few (64) bytes in the flash page
23:13:42gevaertsshotofadds: it should be the same usb driver as the tcc ones
23:15:05shotofaddsdenes_: check if any of the 'spare' bytes are the same in all sectors of a physical block. if so it's possibly a logical block number.
23:15:20shotofaddsyou can also try searching the physical blocks for known data
23:16:04 Quit bluebrother ("leaving")
23:16:09shotofaddsthe fun starts when you find multiple physical blocks with the same logical block number ;)
23:16:45 Join lostnfound [0] (n=a81087b3@gateway/web/cgi-irc/
23:16:51 Join Slack [0] (
23:17:15denes_shotofadds: yes, I have found the first logical block. it's the 22016. page ;)
23:17:17lostnfoundI found an iPod today, and it had this operating system on it
23:17:38lostnfoundis there anyway it has information on the person that used it
23:17:45shotofaddsgevaerts: yes, usb should be a rather significant help too! any idea what I'd have to do to get logf-over-usb working (either on m200 or d2)?
23:18:05shotofaddsdenes_: but is it always in the same physical location? I suspect not.
23:18:15denes_shotofadds: probably not
23:18:18Bagderlostlogic: connect it to a computer and investigate the files on it
23:18:27lostnfoundThere are a few pictures
23:18:28Bagderlostnfound: even
23:18:34lostnfoundbut the OS is hard to get used to
23:18:59Bagderlostnfound: I think some of us may disagree with that
23:19:11lostnfoundI didn't want to plug it in to my computer and have it sync to the foreign iPod instead of mine
23:19:24gevaertsshotofadds: I'm now trying to get it to work on DAX
23:19:37Bagderlostnfound: what, it syncs without you asking it to?
23:20:19 Join webguest64 [0] (n=1824bd5b@gateway/web/cgi-irc/
23:20:36webguest64is anyone there?
23:20:51lostnfoundI'm not sure
23:21:01lostnfoundThat is why I haven't connected it to the computer
23:21:12shotofaddsgevaerts: excellent work! hopefully we can find a way to spit binary data through "logf" too
23:21:19lostnfoundfor fear of it syncing with my computer and giving me more work to do
23:21:24webguest64Do you think rockbox is going to work on the v2 versions of the sansa e200?
23:21:27gevaertsshotofadds: I
23:21:38gevaertsI'd really go for the mass-storage way for that
23:21:48webguest64I tried that
23:21:48webguest64yeah I know
23:21:51Bagderwebguest64: yes, we have great hopes of that it eventually will
23:22:04webguest64great that is just what I wanted to hear
23:22:19webguest64if you could estimate how long do you think it would take?
23:22:21lostnfoundits an interesting OS concept
23:22:29lostnfoundbut it won't work on my touch
23:22:35Bagderwebguest64: if we could estimate we would, but we can't
23:22:44webguest64ok thanks
23:22:57shotofaddsgevaerts: yes, that sounds sensible. a quick hack to allow that should take all of 30 seconds if mass storage works
23:23:10webguest64No mass storage doesn't work
23:23:15 Join BigBambi_ [0] (n=Alex@rockbox/staff/BigBambi)
23:23:23Bagderlostnfound: true, but it runs on some 28 other music players
23:23:24webguest64it detects it then when you try to install it thinks it isn't there
23:23:27 Quit BigBambi (Nick collision from services.)
23:23:34gevaertsshotofadds: I can post a patch that uses a block of RAM instead of the ata_* layer if you like
23:23:34 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
23:23:35 Nick BigBambi_ is now known as BigBambi (n=Alex@rockbox/staff/BigBambi)
23:23:46webguest64Auto-detect finds it but when you try to install it thinks it isn't there
23:23:53Bagderwebguest64: that discussion has nothing to do with your v2 question
23:24:07Bagderthere are 140 persons here and there are more than one conversation at once at times
23:24:17webguest64that makes sence
23:24:20webguest64I think
23:24:32BagderRockbox does not work on the sansa v2 at this point in time, so tryign to install it is pointless
23:24:33soapblubrother (for logs) - I can get you a H10 with "battery issues" for $15 + shipping.
23:24:33 Quit webguest64 (Client Quit)
23:24:41soapbluebrother even
23:24:44shotofaddsgevaerts: if mass storage "works" currently, I'll just tweak the ata_read_sectors to return raw data instead for now
23:24:51 Join Munkie [0] (
23:25:01lostnfoundGuess I'll sell this bitch
23:25:18gevaertsshotofadds: it works on tcc77x anyway
23:25:19 Join hd [0] (
23:25:49saratogalinuxstb: the high WMA DRAM ussage is almost all look up tables
23:25:56MunkieIf any of you are in need of a laugh, go to the apple forums and search for rockbox.
23:25:59saratogathe format needs a very large number of look up tables
23:26:04lostnfoundI haven't looked but is there a walkthrough on how to uninstall
23:26:04 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
23:26:13gevaertsshotofadds: I didn't get it to run last time I tried on the D2, but you probably know that one better
23:26:29gevaertslostnfound: easiest way is to let itunes do a restore
23:26:31saratogathe LSP feature needed for files less then 16kbps uses a 32KB lookup table for one function
23:26:43 Quit Schmogel (Read error: 104 (Connection reset by peer))
23:26:44shotofaddsgevaerts: I have my m200 here ready to try, but it'll probably have to wait til tomorrow. Time is a bit scarce at the moment :/
23:26:53saratogathis could be paired down, but I haven't due to a poor understanding of how some of the codec's features work
23:26:53lostnfoundAlright, as long as it wont sync automatically.
23:27:01 Quit HellDragon (Read error: 104 (Connection reset by peer))
23:27:21linuxstbsaratoga: But in bss? So they're calculated at init time?
23:27:42saratogalinuxstb: oh I didn't realize BSS was all runtime
23:27:45saratogaa lot are, but not all
23:28:16saratogatext is predefined tables and code then?
23:28:17linuxstbYes, "BSS" is zero-initialized data.
23:28:32linuxstbYes, text is code and const data
23:29:07saratogathen its probably the huffman tables, various trig tables, etc
23:29:25saratogadue to an odd detail about how the MDCT works in WMA, each trig table is roughly 2x as big as most other MDCT codecs
23:33:57 Quit lostnfound ("CGI:IRC (EOF)")
23:35:50 Quit fyre^OS (Remote closed the connection)
23:36:06 Join fyrestorm [0] (
23:42:44 Quit ender` (Read error: 104 (Connection reset by peer))
23:43:19 Join ender` [0] (
23:46:31linuxstbBagder: Can you remember enough about the V1 Sansas to think of what could be the issue here?
23:46:46shotofaddsgevaerts: could you post that UMS-from-RAM patch? I'll try it out when I get a chance tomorrow
23:46:50 Quit Siku ()
23:46:56gevaertsshotofadds: actually I'
23:47:03gevaertsshotofadds: actually I'm thinking of just committing it
23:47:12 Quit Munkie ("Leaving")
23:47:13shotofaddsor that :p
23:48:12Bagderlinuxstb: the only hint is his read-mode comment, which usually implies file system breakage
23:50:06 Quit lasser (Remote closed the connection)
23:51:38gevaertsshotofadds: committed. The size near the top is in sectors
23:52:58 Quit hd (Remote closed the connection)
23:53:25 Join HellDragon [0] (
23:55:52 Quit ender` (" A computer program will always do what you tell it to, and seldom what you want it to.")
23:58:08 Quit bughunter2 ("bye")

Previous day | Next day