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 2010-07-29

00:06:56***Saving seen data "./dancer.seen"
00:08:58 Quit mt (Ping timeout: 240 seconds)
00:09:23 Join mt [0] (~mtee@rockbox/developer/mt)
00:09:38 Quit Jerom (Read error: Connection reset by peer)
00:28:25 Quit krazykit (Quit: Connection reset by beer)
00:28:33 Join halmi_ [0] (
00:28:48 Join krazykit [0] (~kkit@
00:29:43 Quit bertrik (Ping timeout: 245 seconds)
00:31:35 Quit halmi (Ping timeout: 276 seconds)
00:33:18 Quit krazykit (Ping timeout: 240 seconds)
00:33:50 Quit bunnyboi (Quit: Don't follow me)
00:34:50 Quit halmi_ (Read error: Connection reset by peer)
00:37:14n1sBuchsel: That asm will probably output different samples than the compled c version since the as is accumulating the result of two multiplications into a 64 bit accumulator before constructing the final result while the compiled c code will construct individual 32 bit results for all multiplications and accumulate those
00:37:59n1sAlso, did you have an answer to my question about the MPC_MULTIPLY_EX macro in fs#11502 ?
00:40:01 Join kugel [0] (~kugel@rockbox/developer/kugel)
00:40:45kugelpreglow: ping
00:45:30kugelgevaerts: ping
00:52:45 Quit Topy44 (Read error: Connection reset by peer)
00:52:49 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
00:55:37 Join panni__ [0] (
00:55:55 Quit panni_ (Ping timeout: 264 seconds)
01:02:32 Join Topy44 [0] (
01:03:09 Join bzed_ [0] (
01:03:37 Nick bzed_ is now known as Guest94406 (
01:04:56 Quit bzed (Ping timeout: 276 seconds)
01:06:06 Join krazykit [0] (~kkit@
01:17:13 Quit JdGordon (Ping timeout: 245 seconds)
01:25:47CIA-97New commit by nls (r27607): libwmapro: coldfire asm for vector_fixmul_scalar, 1% speedup
01:27:28CIA-97r27607 build result: All green
01:32:28 Join Davide-NYC [0] (~chatzilla@
01:36:51Davide-NYCAnybody given any thought to a WebOS port?
01:37:17n1swhat does that run on?
01:37:55Davide-NYCPalm (now HP) phones.
01:38:07Davide-NYCC/C++ and SDL.
01:38:34Davide-NYCI put a post in the forum but it may have been posted in the wrong place.
01:42:07n1ssounds like a rather nice os to develop for
01:50:14 Join wodz [0] (
01:55:10 Quit n1s (Quit: Lämnar)
01:58:30krazykitDavide-NYC, you'd want to look into the rockbox-as-an-app work being done
02:00:14 Join JdGordon| [0] (
02:00:14 Quit JdGordon| (Changing host)
02:00:14 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
02:04:46wodzamiconn: I did DMA in lcd_update() and lcd_update_rect(). This gives 12-26% speedup on HD200. I wonder if using DMA has any side effects.
02:06:58***Saving seen data "./dancer.seen"
02:19:43 Quit chrisb (Ping timeout: 245 seconds)
02:23:23 Quit AlexP (Ping timeout: 246 seconds)
02:23:32 Quit MethoS- (Read error: Connection reset by peer)
02:23:38 Join AlexP [0] (~alex@rockbox/staff/AlexP)
02:23:40 Join MethoS- [0] (~clemens@
02:24:24 Quit juju2143 (Ping timeout: 246 seconds)
02:27:53 Join juju2143 [0] (juju2143@gateway/shell/
02:34:38 Quit wodz (Quit: Leaving)
02:40:43 Join CaptainKwel [0] (
02:42:38 Quit kugel (Ping timeout: 245 seconds)
02:48:20 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
02:51:27 Quit insp_ (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
02:56:17 Join milz [0] (
03:07:09 Join hebz0rl_ [0] (
03:09:18 Quit milz (Remote host closed the connection)
03:10:08 Join milz [0] (
03:10:45 Quit hebz0rl (Ping timeout: 276 seconds)
03:20:54 Quit komputes_ubuntu (Remote host closed the connection)
03:33:30 Join Hillshum [0] (4ba5edf2@gateway/web/freenode/ip.
03:40:48 Quit pixelma (Ping timeout: 260 seconds)
03:40:48 Quit amiconn (Ping timeout: 260 seconds)
03:42:29 Join bunnyboi [0] (
03:52:31 Join pixelma [0] (quassel@rockbox/staff/pixelma)
03:52:51 Join amiconn [0] (quassel@rockbox/developer/amiconn)
03:57:16 Quit MethoS- (Remote host closed the connection)
03:58:16 Quit t0rc (Remote host closed the connection)
04:06:28 Join Barahir [0] (
04:07:01***Saving seen data "./dancer.seen"
04:09:38 Quit Barahir_ (Ping timeout: 245 seconds)
04:19:13 Join guymann [0] (
04:28:08 Quit Hillshum (Quit: Page closed)
04:32:27 Quit pixelma (Disconnected by services)
04:32:28 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:32:43 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:32:48 Quit amiconn (Disconnected by services)
04:32:50 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:32:57 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:41:11 Join flan_suse [0] (~linux@unaffiliated/flansuse)
04:42:13flan_suseI cannot seem to install the voice files. I keep getting this error message: Encoding of /tmp/rbvoice/LANG_SET_BOOL_YES.wav failed
04:42:21flan_suseIt aborts after that.
04:44:49flan_suseI have the espeak package installed on my openSUSE 11.2 computer.
04:48:22 Join [Saint] [0] (S_a_i_n_t@
04:48:54 Quit S_a_i_n_t (Ping timeout: 276 seconds)
04:50:07flan_suseHm seems to work with the festival engine instead. Espeak just wouldn't work, period.
04:50:21flan_suseGoing to see how it sounds.
04:54:12flan_suseYup. Voice works when I downloaded, installed, configured, and used the festival TTS engine; as opposed to espeak.
04:57:38 Quit flan_suse (Quit: Leaving)
05:05:09 Quit Davide-NYC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
05:06:56 Quit mikroflops (Ping timeout: 265 seconds)
05:07:23 Join mikroflops [0] (
05:10:05 Quit jordan` (Read error: Operation timed out)
05:10:09 Join jordan` [0] (
05:19:05 Join robin0800 [0] (
05:22:48 Quit anewuser (Quit: Votinz Remix Phase concluded. The votes! The mad votingzes!!)
05:33:41 Quit robin0800 (Quit: Leaving)
05:54:06 Quit panni__ (Read error: Connection reset by peer)
05:54:31 Join Sidein [0] (
05:55:12 Quit bunnyboi (Ping timeout: 276 seconds)
06:01:07 Join bunnyboi [0] (
06:07:05***Saving seen data "./dancer.seen"
06:11:20 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
06:15:26 Quit [Saint] (Ping timeout: 265 seconds)
06:37:11 Quit hebz0rl_ (Quit: Ex-Chat)
06:39:10 Quit AndyI (Ping timeout: 248 seconds)
06:44:42 Join AndyI [0] (~pasha_int@
07:03:46 Quit CaptainKwel (Quit: Ex-Chat)
07:09:21 Join jstaley [0] (
07:14:54jstaleyHi everyone. I have an SDL app I'd like to move to Rockbox, but was a little confused where to start. My goal is to get it running on the similator first, but I'm a little confused where to start.
07:15:33jstaleyAnybody have some advice or can point me in the right direction?
07:27:18JdGordon|you need to replace anything sdl with rockbox's plugin api calls
07:28:04JdGordon|and rockbox doesnt provide a full libc so things like malloc() need to be removed also
07:36:30rvvs89Also, eliminate any floating point variables.
07:42:18jstaleyThanks for the replies. All of that should keep me busy for a while.
07:43:11 Quit sasquatch (Ping timeout: 248 seconds)
07:44:03bieberJdGordon|: You showed me a piece of an XML tag description a while back, has that progressed any further?
07:49:22 Join Buschel [0] (
07:55:53Buscheln1s: you've made a good point regarding the difference in output between the asm and the c version. maybe I'll test this via using the same operation in the asm to verify this. the bad thing thing is that the asm routine is slower than the compiler's output... maybe the compiler does some loop unrolling...
07:57:54 Join sasquatch [0] (
07:59:15Buscheln1s: regarding your question to mpc -> yes, MPC_MULTIPLY_EX is called with shifts from 0..32. For arm and coldfire target only 0..31 as there is a special asm routine for the 32-case. For arm and coldfire the only place where MPC_MUL_EX is called, is the decoder initialization. so, I see no reason further optimize here. thanks for taking a look into the routines!
08:07:08***Saving seen data "./dancer.seen"
08:10:32CIA-97New commit by bieber (r27608): ThemTheme Editor: Enabled sbs rendering underneath fms documents
08:12:13CIA-97r27608 build result: All green
08:15:07JdGordon|bieber: no because perl libxml isnt installed by default which makes it hard to put into the build
08:15:15bieberOh, okay
08:15:45bieberJust use plaintext, then?
08:16:09JdGordon|I've got a odd issue with the editor, when I start it the timer dialog opens and I cant move it around
08:16:14JdGordon|yeah, but I dont want to :)
08:17:35bieberThat's really weird. Is it happening every time you start it up?
08:18:36JdGordon|and I deleted the config file
08:19:24bieberAre you running the newest build? What platform are you on?
08:22:28bieberJdGordon|: Someone on the forum is telling me this is a rendering bug, but I'm not sure if it is or not
08:23:21bieberIn the widecabbie theme, there's a line that goes like this %?cs<othercraphere> that's apparently intended to render only if the player is currently on the menu screen, which would select the first conditional branch
08:24:14 Join mitk [0] (~mitk@
08:24:19bieberBut there aren't any other branches there, even empty ones, and I was under the impression that when there aren't enough branches to satisfy all the tag's options the renderer should select the last valid branch the user specified, which in this case means that that text will always show up unless they add a | after it
08:28:25JdGordon|bieber: ubuntu 10.04 ie86, svn from last night
08:28:56JdGordon|and yeah, that sounds correct
08:29:10bieberThe current behavior, or the way he wants it to work?
08:29:20JdGordon|umm, no, it will probably do different things because there is only one option
08:29:57JdGordon|check apps/gui/skin_engine/skin_display.c for evaluate_conditional() (i tinhk thats the ritght place)
08:30:00JdGordon|cant look now though
08:30:09bieberYou mean it should work the way they want it to?
08:30:11bieberTaking a look now
08:31:41JdGordon|I tihnk so yeah
08:31:58JdGordon|it depends on the order of the values
08:32:10JdGordon|when there is only one value the renderer pretends there are two
08:32:16 Quit tmzt (Read error: Operation timed out)
08:32:20 Join tmzt [0] (
08:32:35JdGordon|if the value == 1 then the true branch works, when it is >1 the false (nothing) is used
08:33:06bieberYep, I see what it did there
08:33:35 Join einhirn [0] (
08:34:17 Join ender` [0] (
08:39:23 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.7/20100713130626])
08:41:35CIA-97New commit by bieber (r27609): Theme Editor: Renderer now correctly handles special case with only one branch in a conditional
08:42:20bieberNow I get to figure out how to implement the whole auto-complete thing
08:42:38 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
08:43:08CIA-97r27609 build result: All green
08:51:59 Join Judas_PhD [0] (
08:53:07 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
08:56:30 Quit bunnyboi (Quit: There are two major products that come out of Berkeley: LSD and UNIX. We don't believe this to be a coincidence.)
09:05:26 Join Szpila [0] (~sszpila@
09:07:38 Quit Strife89 (Quit: Bed.)
09:17:16CIA-97New commit by bieber (r27610): Theme Editor: Project Exporter will not check for internal settings keys as files
09:18:09 Quit rvvs89 (Ping timeout: 240 seconds)
09:18:52CIA-97r27610 build result: All green
09:20:48 Join S_a_i_n_t [0] (S_a_i_n_t@
09:24:13 Quit bertrik (Quit: :tiuQ)
09:38:03CIA-97New commit by bieber (r27611): Theme Editor: Recent documents/projects that no longer exist are now cleared from the menu
09:39:42CIA-97r27611 build result: All green
09:40:14 Join bimbel [0] (~Miranda@unaffiliated/bmbl)
09:40:59 Quit bmbl (Ping timeout: 258 seconds)
09:44:21 Quit n17ikh (Remote host closed the connection)
09:51:18 Join insp_ [0] (~chatzilla@
09:52:00 Join n17ikh [0] (
09:53:00 Join Rob2222 [0] (
09:55:05 Join Jaykay [0] (
09:56:27 Quit Utchybann (Ping timeout: 265 seconds)
09:56:31 Quit Rob2223 (Ping timeout: 248 seconds)
09:57:01 Join Utchybann [0] (
10:07:11***Saving seen data "./dancer.seen"
10:08:24 Join rvvs89 [0] (
10:08:51 Quit rvvs89 (Changing host)
10:08:51 Join rvvs89 [0] (ivo@pdpc/supporter/base/rvvs89)
10:14:01 Nick bimbel is now known as bmbl (~Miranda@unaffiliated/bmbl)
10:22:02 Join swilde [0] (
10:28:20 Join insp [0] (
10:28:25 Quit insp (Client Quit)
10:31:53 Join DerPapst [0] (
10:36:34 Quit ajb (Quit: ERC Version 5.3 (IRC client for Emacs))
10:38:59 Join n1s [0] (
10:39:03 Quit n1s (Changing host)
10:39:03 Join n1s [0] (~n1s@rockbox/developer/n1s)
10:40:39 Join insp [0] (
10:40:39 Quit insp (Client Quit)
10:44:12 Nick insp_ is now known as insp (~chatzilla@
10:44:34 Nick insp is now known as insp_ (~chatzilla@
10:49:31 Quit jstaley (Ping timeout: 260 seconds)
10:51:56 Join Kitr88 [0] (
10:53:14 Quit Kitar|st (Ping timeout: 246 seconds)
11:01:48 Join kugel [0] (~kugel@rockbox/developer/kugel)
11:07:52 Quit powell14ski_ (Ping timeout: 276 seconds)
11:07:59 Quit BlakeJohnson86 (Ping timeout: 245 seconds)
11:07:59 Quit n17ikh (Ping timeout: 265 seconds)
11:08:58 Join BlakeJohnson86 [0] (
11:12:28 Join n17ikh [0] (
11:14:52 Part Szpila
11:18:31 Join mk429_ [0] (
11:20:54 Quit mk429 (Ping timeout: 245 seconds)
11:25:22 Quit JdGordon| (Quit: leaving)
11:27:35 Quit n17ikh ()
11:28:37 Join n17ikh [0] (
11:32:53 Join MethoS- [0] (~clemens@
11:33:03 Quit n17ikh (Client Quit)
11:37:51 Join dfkt [0] (dfkt@unaffiliated/dfkt)
11:53:54 Join Casainho [0] (
11:54:15kugelI saw the main menu!
11:54:43 Quit Sidein (Quit: leaving)
11:54:52kugelbut it crashes rather quickly, there must be some threading problem
12:02:07kugelah, my core_sleep() doesn't quite work
12:02:26 Join watto [0] (~watto@
12:04:24 Join petur [0] (~petur@rockbox/developer/petur)
12:07:15***Saving seen data "./dancer.seen"
12:12:48 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
12:13:14 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
12:18:58 Join Jerom [0] (~jerome@
12:22:09 Quit insp_ (Ping timeout: 245 seconds)
12:23:35 Quit Casainho (Quit: ChatZilla 0.9.86 [Firefox 3.6.6/20100628124739])
12:32:13 Nick Guest94406 is now known as bzed (
12:35:33 Join hebz0rl [0] (
12:43:28 Join insp_ [0] (~chatzilla@
12:49:11 Quit Judas_PhD (Quit: Aaarrrggghhh!!!)
12:53:01 Join DislocateHen [0] (
12:56:37 Quit Xerion (Ping timeout: 276 seconds)
13:01:32CIA-97New commit by nls (r27612): libmwapro: use codeclib asm macros for XNPROD31, speeds up decoding of a 55kbps wma pro file by 40% on h300.
13:03:21CIA-97r27612 build result: All green
13:16:31 Join T44 [0] (
13:18:19 Quit bmbl (Read error: Connection reset by peer)
13:20:05 Quit Topy44 (Ping timeout: 245 seconds)
13:24:05 Join bmbl [0] (
13:24:10 Quit bmbl (Changing host)
13:24:10 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
13:25:02kugeln1s: shouldn't codeclib_misc.h be enough?
13:27:20 Quit Jerom (Quit: Leaving.)
13:27:25JdGordonLadies and gents.... I think its just about time to commit my stupidly mammoth skin patch unless someone wants to complain?
13:28:13JdGordonand by complain I mean offer to test
13:32:49 Quit Kitr88 (Ping timeout: 260 seconds)
13:32:52JaykayJdGordon: gevaerts said in FS that players doesnt compile
13:33:09JdGordonthat has since been fixed
13:35:52 Join Kitar|st [0] (
13:43:36mitkJdGordon: Commit can be best call for testers :)
13:45:09JdGordonS_a_i_n_t: ping
13:47:26 Quit hebz0rl (Quit: Ex-Chat)
13:50:01S_a_i_n_tJdGordon: pong...kinda, eally busy.
13:50:23JdGordonhow many targets cabiie havnt been converted?
13:52:07JdGordon9 files?
13:52:14S_a_i_n_tall pretty bastard ones, I haven't had the time to get to them.
13:52:28JdGordonah bugger :p
13:52:47JdGordonI suppose I'll work on them tonight then
13:55:23JdGordonlooks like only 5 of them actually need to be converted
13:55:35JdGordonthe mono/2bit dont really need viewports for me
14:07:17***Saving seen data "./dancer.seen"
14:17:41 Quit soap (Ping timeout: 276 seconds)
14:18:42 Join bluebrother [0] (
14:18:42 Quit bluebrother (Changing host)
14:18:42 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
14:20:39 Quit petur (Quit: Leaving)
14:21:59 Quit bluebroth3r (Ping timeout: 258 seconds)
14:27:43 Join wodz [0] (
14:28:25JdGordonthey dont have the possible screwy AA viewports left behind maybe
14:29:25pixelmagreyscale ones have AA viewports except the M3/Iaudio remote IIRC
14:30:30pixelmaor maybe just AA currently
14:31:46 Join chrisb [0] (
14:32:29 Join soap [0] (~soap@rockbox/staff/soap)
14:34:11 Part DislocateHen
14:36:42kugelJdGordon: will viewport-less wps still work with that patch?
14:37:13JdGordonthe problem I found was pb outside a viewport (mini2g cabbie was doing that)
14:37:35JdGordonfingers crossed....
14:37:51CIA-97New commit by jdgordon (r27613): FS #11470 - new skin code, finally svn uses the new parser from the theme editor. This means that a skin that passes the editor WILL pass svn and ...
14:37:52 Quit soap (Ping timeout: 264 seconds)
14:38:18JdGordon44 files changed, 3161 insertions(+), 4382 deletions(-) :)
14:38:20JdGordononly a little one
14:39:39n1skugel: i would guess so but someone went throug the trouble of commenting out those includes from codeclib:misc.h i assumed there was a reason for it
14:40:02CIA-97r27613 build result: 54 errors, 35 warnings (jdgordon committed)
14:40:23kugelpb outside a viewport doesn't work anymore?
14:41:20 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
14:41:21JdGordonif it *really* pisses people off I can reeenable it
14:41:21JdGordonbut it isnt nice
14:41:28n1sJdGordon: congratulations! those are some mighty green deltas btw :)
14:41:40JdGordonoh hell yeah :D
14:41:49JdGordonprobably means i fucked something up :p
14:42:12pixelmawhat's with the progressbar in the built-in WPS then?
14:42:18JdGordonmaybe not :)
14:42:32JdGordonpixelma: that doesnt ave any viewports so it is fine
14:43:21pixelmaaha, so progressbars on default viewport only still works, just not when you start using more viewports
14:43:59JdGordonyes, only images and AA will draw in the default vp if others are set
14:45:40 Quit Kitar|st (Ping timeout: 240 seconds)
14:45:51pixelmaseems fair enough I think
14:46:21 Join LambdaCalculus37 [0] (~LambdaCal@rockbox/staff/LambdaCalculus37)
14:46:24*pixelma wonders what the database toll build failures have to do with the patch
14:46:33pixelmaor tool...
14:46:52JdGordongevaerts: can you fix that break, its root.make which is AFAIK black magic :D
14:47:12Topic"Rockbox 3.6 now released! | Please read before speaking: | Please direct offtopic/social chat to #rockbox-community | This channel is logged at" by ChanServ (ChanServ@services.)
14:49:01CIA-97New commit by jdgordon (r27614): fix touchscreen errors
14:49:10LambdaCalculus37JdGordon: Question... since the disktidy plugin rewrite, all it takes to add another file or directory to include by default is just to edit disktidy.config, right?
14:49:30JdGordonumm, if you say so :D
14:50:53LambdaCalculus37I'm going to add .Spotlight-* and Mac OS 9 Desktop files to the config file.
14:51:13CIA-97r27614 build result: 4 errors, 0 warnings (jdgordon committed)
14:52:01JdGordon2MB less ram usage on the mr500! I definetly messed something up...
14:52:06CIA-97New commit by nls (r27615): Make codeclib_misc.h include the asm headers to simplify inclusion magic.
14:52:59JdGordonis there a bloat-o-meter for ram usage?
14:53:02n1sJdGordon: some buffer depending on lcd size and depth perhaps?
14:53:26JdGordonmust be
14:53:56CIA-97r27615 build result: 4 errors, 0 warnings (nls committed)
14:54:39wodzWhy don't we use DMA module for ATA transfers on CF?
14:54:56n1swodz: don't we? i thought we did
14:55:10n1sof course it might differ between players
14:55:27wodzn1s: I mean 5249/5250 DMA module not ATA DMA mode
14:55:29n1samiconn: probably knows that stuff though
14:55:49 Join soap [0] (~soap@rockbox/staff/soap)
14:56:04chrisbnow that i've seen the software on an ipod classic, i can see that rockbox is so much better
14:57:38wodzcopy_read_sectors() is asm beast. It is complicated because it utilizes line transfers. With DMA module You have line transfers for free
14:57:39n1smt: are you fine with deleting the old tremor mdct from svn, saratoga mentioned it might be useful to keep it around when you debugged wma pro but that seems done now
14:58:30n1swodz: i'd guess that either something in the hardware prevents using dma or that the asm beast turned out to be faster
14:59:40wodzn1s: or exiting solution was optimized and optimized and becomes this monster :-)
15:00:31n1sif you really want to know, either ask amiconn or search the irclogs, the answer might be hidden there somewhere
15:01:07wodzasking amiconn is faster route probably
15:01:45 Quit antil33t (Read error: Connection reset by peer)
15:01:53 Join antil33t [0] (
15:02:00 Quit soap (Ping timeout: 260 seconds)
15:02:05n1swodz: i think he's away on some trip though
15:12:36 Join soap [0] (~soap@rockbox/staff/soap)
15:16:16JdGordonwhy the hell does metadata.o want tag_Table.h?
15:17:29 Quit soap (Ping timeout: 276 seconds)
15:19:33JdGordongevaerts: that red isnt Makefile issues... ignore my request :)
15:21:05 Quit krazykit (Ping timeout: 240 seconds)
15:23:35CIA-97New commit by jdgordon (r27616): Fix Database red, move the cuesheet->token code into the skin engine
15:24:42wodzAm I correct that if I use unit16_t to encode fp number with fractional part 14bits I'll get [-1,1] range of possible values?
15:25:02 Join robin0800 [0] (
15:25:11wodzor int16_t rather
15:25:33CIA-97r27616 build result: All green
15:25:41n1swodz: afaiu, yes, i'm no fixed point expert though'
15:25:58n1swodz: what are you working on?
15:26:15wodza game :-)
15:26:32 Join insp__ [0] (~chatzilla@
15:26:35 Quit insp_ (Ping timeout: 260 seconds)
15:26:36 Nick insp__ is now known as insp_ (~chatzilla@
15:26:45 Quit insp_ (Client Quit)
15:28:05 Join soap [0] (~soap@rockbox/staff/soap)
15:28:38 Join WalkGood [0] (~FU@unaffiliated/walkgood)
15:38:21kugeldlopen() fails :(
15:38:30JdGordongrumble has a b in it!
15:38:41JdGordonwhat cant it link?
15:38:50 Quit Jaykay (Ping timeout: 260 seconds)
15:38:58 Join funman [0] (~fun@rockbox/developer/funman)
15:38:59 Quit robin0800 (Quit: Leaving)
15:39:53 Join Judas_PhD [0] (
15:40:52kugelI'm so close to playing music, but it doesn't want to load the codec :(
15:42:06 Quit soap (Ping timeout: 258 seconds)
15:42:07JdGordonfile permissions?
15:42:15funmanapp permission?
15:42:32JdGordongot a WIP apk?
15:45:50JdGordoncare to share? :)
15:46:10kugelin a few seconds
15:46:32 Join soap [0] (~soap@rockbox/staff/soap)
15:46:54kugelJdGordon: you need a 320x480 device
15:47:07JdGordonoh poo :p
15:48:45kugel(or the emulator from the sdk)
15:49:45 Quit JdGordon (Read error: Connection reset by peer)
15:49:46 Join JdGordon1 [0] (
15:50:22 Join krazykit [0] (~kkit@
15:53:36 Quit bieber (Ping timeout: 240 seconds)
15:54:55kugelcodecs should go to /sdcard/rockbox/codecs
15:55:42JdGordon1dont apps need extra permisions to be able to read from /sdcard?
15:55:54LambdaCalculus37kugel: I have a 320x480 LCD on my Magic.
15:55:57JdGordon1or have you got other files there working fine?
15:56:30kugelJdGordon1: yes, I added that permission
15:56:36LambdaCalculus37Send me a WIP apk when you get one working.
15:56:43kugelwell, you need permissions to *write* to /sdcard, not for reading
15:56:45 Quit mitk (Quit: Leaving)
15:57:00kugelLambdaCalculus37: the link is right above :)
15:57:25JdGordon1does the ndk maybe not let you run code from /sdcard?
15:57:33JdGordon1which would explain why dlopen no worky
15:57:47kugel07-29 15:47:25.020: DEBUG/RockboxBootloader(32075): dlopen(/sdcard/temp_codec/_temp_codec0.dll): Cannot load library: load_segments[875]: 56 failed to map segment from '_temp_codec0.dll' @ 0x80b00000 (0x0001533c). p_vaddr=0x00000000 p_offset=0x00000000
15:57:53JdGordon1FS equivilant of no execute pages in ram
15:57:54kugeldoesn't sound like a permission thing
15:57:57LambdaCalculus37kugel: Downloading now.
15:59:35kugelLambdaCalculus37: but I can't promise that it works on the magic. it runs 1.5 right?
15:59:36 Join jgarvey [0] (
16:00:05 Quit soap (Ping timeout: 245 seconds)
16:00:49LambdaCalculus37kugel: Nope. CyanogenMod 5.0.8, so it's more or less Android 2.1 with some 2.2 bits.
16:00:56JdGordon1kugel: well you arnt the only one with dlopen problems
16:01:32kugelLambdaCalculus37: ah, I see
16:02:48kugelJdGordon1: :'(
16:03:12JdGordon1just noticed the date in the thread I was looking at was from 08
16:03:21JdGordon1try #android-dev
16:03:34kugelgood idea
16:06:08 Quit evilnick (Ping timeout: 248 seconds)
16:06:15kugelI wonder what p_vaddr and p_offset are
16:06:49 Join evilnick [0] (~evilnick@2001:0:4137:9e76:3cde:1636:b5a6:f655)
16:06:59LambdaCalculus37kugel: Installed. Let's see what it does.
16:07:21***Saving seen data "./dancer.seen"
16:08:09JdGordon1480x800 "just" needs the art and skins done right?
16:08:42pixelmaJdGordon1: did you enable some tags that weren't enabled on teh Player before? I'm really just asking because the delta looks weird compared to the others, binsize etc. isn't that tight on it :)
16:09:12JdGordon1yes and no
16:09:24kugelJdGordon1: and a rockbox logo
16:09:32JdGordon1the increase is probaly because the player didnt lose the code that the others did to make it worth it
16:09:40kugel320x480 has a rockbox logo at least, but no cabbie
16:10:01JdGordon1ok, cool
16:10:06LambdaCalculus37kugel: Codec failure whenever I play something.
16:10:19kugelall targets got a binsize increase. only the ram usage went down
16:10:21pixelmaJdGordon1: how do you mean that - it doesn't use the new parser or something else?
16:10:30kugelLambdaCalculus37: that's exactly my problem :(
16:11:15JdGordon1the new parser added about 8KB or so, the other targets all managed to drop about 7K thanks to it so came out close to even
16:11:26JdGordon1the player didnt lose that much dead code
16:11:43pixelmaok, but the RAM usable for other things got bigger I guess (e.g. for voice on small mem)
16:12:22JdGordon1the ram deltas are confusing. I sure wasnt expecting that at all
16:12:38JdGordon1but same deal, the new code isnt optimized for the players shitty display
16:12:52 Join Strife89 [0] (
16:13:36pixelmathe display might not look nice but its readability is great ;)
16:14:21kugelJdGordon1: did anything happen about the skin buffer?
16:14:54JdGordon1no, making it dynamic is for later
16:15:20kugelI suspected the ram delta change would have something to do with the skin buffer
16:15:27pixelmalet's see if and maybe which bugs creep up. Unfortunately I'm quite busy with RL this week, if I'm lucky I'll have some time tomorrow or Saturday to test
16:16:32JdGordon1anyone remember a massive ram increase recently?
16:16:46JdGordon1I have a feeling svn was using 2 skin buffers accidently
16:17:34pixelmamaybe the binsize graphing people know (if they also do RAM usage, not sure) ;)
16:18:03JdGordon1ah yeah, the libskin_parser has a static 400KB buffer which accidently got in a while ago :D
16:18:04kugelnow I'm mighty confused about lib/skin_engine vs apps/gui/skin_engine
16:18:58kugelthe 400KB doesn't explain it though
16:19:11 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
16:23:53 Quit evilnick (Read error: Connection reset by peer)
16:26:49n1scould someone running a sim on windows confirm FS #11225?
16:29:48 Quit funman (Quit: free(random());)
16:29:59 Quit Judas_PhD (Quit: Aaarrrggghhh!!!)
16:29:59 Join funman [0] (~fun@rockbox/developer/funman)
16:34:30 Quit JdGordon1 (Quit: Leaving.)
16:36:45 Quit CIA-97 (Ping timeout: 245 seconds)
16:42:14 Quit n1s (Quit: Lämnar)
16:45:01 Quit wodz (Quit: Leaving)
16:45:42 Join fdinel [0] (
16:46:32 Join captainewkll [0] (2669ecc2@gateway/web/freenode/ip.
16:49:18 Join panni_ [0] (
16:58:47 Join CIA-7 [0] (~CIA@
17:01:25 Join soap [0] (~soap@rockbox/staff/soap)
17:07:54 Quit soap (Ping timeout: 276 seconds)
17:13:24 Join Jaykay [0] (
17:18:43 Join Marco3 [0] (
17:20:22 Quit pamaury (Remote host closed the connection)
17:20:50Marco3hey, I tried using rockbox on my sansa clip+, but it didn't sort the files right. I tried both interpretation "as digits" and "as numbers", but neither one fixed the problem. it's always like "01", "10", "11", "12" instead of "01", "02", "03" and so on
17:21:25Marco3does anyone know how to fix that?
17:21:43amiconnwodz (logs): Did you actually try DMA line transfers (aka auto align) for ATA? I did on iriver, and it just didn't work
17:22:28Marco3I don't even have a clue what that is. I simply copied the files using windows explorer
17:23:06amiconnAll I got was garbage. DMA without auto align did work, but was significantly slower than CPU line transfers
17:23:54amiconnFurthermore DMA is impossible for unaligned transfers, so you need CPU transfers for that anyway
17:24:56amiconnBtw, X5 OF uses DMA - *without* auto align
17:27:26 Quit antil33t (Read error: Connection reset by peer)
17:27:33 Join antil33t [0] (
17:29:35funmanmirsal: audiobuffer transfers should be aligned
17:29:56funmanamiconn: ^
17:29:56 Quit Marco3 (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
17:30:13funmanfor the others, TheSeven is working on a patch (see flyspray)
17:30:57amiconnYou probably mean swcodec audio buffers
17:31:14amiconni doubt the hwcodec engine does alignment
17:31:27funmandunno hwcodec
17:32:12amiconnBesides, even if the buffers are usually aligned, the ata driver still has to support unaligned transfers
17:32:21funmaneasy to do
17:32:45funmanand, only until TheSeven patch got committed
17:34:13TheSevenbtw did anybody check if my patch works on multivolume in the meantime?
17:34:26 Quit mt (Ping timeout: 258 seconds)
17:34:31 Quit scorche (Read error: Connection reset by peer)
17:34:32amiconnIt will always have to support unaligned. You never know what plugins may want to do
17:34:43 Join scorche [0] (~scorche@rockbox/administrator/scorche)
17:34:46funmanamiconn: no it will not
17:35:00funmanTheSeven: not yet
17:35:03TheSevenit wraps those into aligned transfers
17:35:13TheSeven(using a buffer allocated from the pool)
17:35:52TheSevenIIRC this is configurable using the STORAGE_NEEDS_ALIGN define though
17:35:58amiconnThat's definitely slower than doing unaligned directly, at least on coldfire and SH1
17:36:16TheSeventhat's why that define exists
17:36:41TheSevenon targets that would need to wrap it into aligned buffers at driver level anyway, the cache layer can take that over
17:38:25 Quit CIA-7 (Ping timeout: 248 seconds)
17:38:39amiconnWhich target(s) would need that?
17:38:48TheSevennano2g, for example
17:39:37amiconnOk, so nand + software ftl, possibly
17:39:50TheSeventhe ftl doesn't care, but dma does
17:40:04TheSevenand we don't know if non-dma nand accesses are even possible on the 2g
17:40:24amiconnCF and SH1 ata prefer aligned, but handling unaligned direvctly is faster
17:41:29amiconnSH1 mmc (Ondio) doesn,t care since it does byte transfers anyway (to/ from SPI )
17:41:45*TheSeven wonders if he should drop the write cache part from it, as it seems to have a negative performance impact
17:42:17TheSevenit's still not clear why those few cache lookups seem to take that much time
17:42:58TheSevenamiconn: the problem on nano2g is cache consistency, not dma transfer alignment
17:43:00*amiconn should probably try this patch soonish
17:43:41funmanTheSeven: there's no restriction on alignement for DMA ?
17:44:11amiconnAh, this problem doesn't exist oncoldfire (no data cache) and SH1 (no cache at all)
17:44:25TheSevenwe don't know the details as we don't have proper documentation for that soc, but i haven't had trouble with dma transfers, at least if they were word aligned, when the caches are off
17:45:52funmanwith ARM PrimeCell PL081 transfers need alignement of burst size
17:46:22TheSevenif the caches are on, it might happen that something different in the cache line gets written while dma is writing to the memory behind it, which will either be forgotten when flushing the cache after the DMA transfer to get rid of old cache lines in the dma buffer, or it might even write back old data from the cacheline being written to, overwriting the DMA data
17:53:48 Join insp_ [0] (
17:59:40 Join komputes [0] (~komputes@ubuntu/member/komputes)
17:59:58TorneTheSeven: you don't know the details? isn't what you just said the details?
18:00:19Tornecacheline interference is the same problem everywhere, it's not hardware dependen
18:00:29Tornewhat don't you know? :)
18:01:04TheSevendma alignment requirements and such things
18:01:28Tornewell, what the DMA controller actually needs is independent of what you have to do to prevent cacheline interference
18:01:37Torneif the DMA controller objects to the alignment it will probably let you know :)
18:01:51TheSeveni have the impression that misaligned (at least if it's word aligned) dma will work, there might be a performance hit though
18:02:05TorneProbably, DMA engines are pretty smart these days
18:02:19Tornebut it's not safe to do it unless you know exactly what's it's sharing a cacheline with
18:03:28Tornethe only thing it's reasonable to share the cacheline with is another part of the same buffer that you are filling/draining in the same operation (even if that's not the same actual dma transfer)
18:04:14 Join mt [0] (~mtee@
18:04:22funmanon PL081 misalignement is not detected, it just ignores the last bits when starting the transfer
18:04:39Tornefunman: That's a pretty reasonable way for it to let you know
18:04:45TorneYou'll get totally wrong data and you'll explode
18:04:46funmanright :)
18:05:02Tornecacheline interference is hard to spot, is what i meant
18:05:08Tornesince it may work fine 99% of the time
18:05:42Torneon PP it just doesn't do DMA if the request isn't cacheline aligned
18:05:56Tornei haven't yet gotten around to making sure as many requests as possible are :)
18:06:17Tornebut if you can't do non-DMA transfers that's a pain. you only need to do the transfer via a temp buffer for the first and last sector, though
18:06:40Tornethe sectors in the middle are safe to DMA because you know they only share a cacheline with data you already transferred (or will later transfer) in the same operation
18:07:23***Saving seen data "./dancer.seen"
18:08:17TheSevenheh, if i knew what sectors are belonging to the same operation...
18:08:34TheSeventhey usually end up as a bunch of single sector transfers at the nand layer
18:08:59Tornewell, that's tough
18:09:11Tornei mean, when transfers are >2 sectors long
18:09:20Tornewhich they should be for contiguous files, no?
18:09:23TheSeventhe ftl will probably take those apart
18:09:47TheSevenat least in some cases
18:09:48Tornethat sounds like a shitty ftl
18:09:58Tornefiles written contiguously should be contiguous up to erase-block size
18:10:09Torneish, give or take alignment
18:10:13TheSevenif you're writing a whole erase block (1MB) at once, yes
18:10:33Tornedoes it not always do that?
18:10:40Tornethat would be a very terrible ftl design
18:10:52TheSeveni've never seen a >64K write coming through the storage api
18:10:59TheSevenso these will always be split up
18:11:18Torneyes, but that doesn't matter
18:11:23Torneconsecutive writes are fine
18:11:29TheSevenif you end up lucky and the block wasn't touched before, they will hopefully be written in the right order to a scattered page block, which will be detected once it's full
18:12:39Tornefat.c only does 256 sectors at a time
18:12:48Tornewhich is our fault
18:13:01Torneit should really be made to do more, but i haven't gotten around to that yet either
18:13:14Torne(now that LBA48 is compiled on all ATA targets it should be able to use that to do 1MB transfers)
18:13:48TheSevendo we need to?
18:14:06TheSeveni doubt it will make much of a difference for reads
18:14:17TheSevenand 99% of the writes issued by rockbox will be smallish anyway
18:14:29Tornebuffering also limits to that much so that'd need poking as well
18:14:37TheSeventhe only thing doing huge writes is USB, and that's currently capped at 24K per write
18:15:25Torneit might be worthwhile for reads on rotating media
18:15:28Torne*shrug* i've not tested it yet
18:15:34TheSevenoh, right
18:15:40Torneit hsouldn't be particularly difficult, anyway, so it's just a matter of getting around to it ;)
18:15:43TheSevenwell, the drive's cache will probably take care of that
18:16:07Tornesometimes :)
18:17:15 Join CIA-6 [0] (~CIA@
18:17:21 Join r0b- [0] (
18:20:03 Quit rob (Ping timeout: 258 seconds)
18:30:10funmanshould i commit the mkamsboot patch and tag the releases, even if the binaries aren't on yet?
18:31:57kugelTheSeven: is that thread safe?
18:32:13kugelthe "wrapping into aligned buffer" thing I meant
18:32:57kugelI had once a similar idea, but using some sort of alloca() so the buffers remain on the stack
18:33:48TheSevenhm, having them on the stack isn't really a good idea
18:34:44kugelwhy not?
18:35:15funmanoverflow if the sectors are too big (2k on nano2g)
18:36:24kugelmake the stack bigger if that's the problem
18:42:42TheSevenit's a huge source of unpredictable misaligned buffers, and reserving 4-6K on *every* thread's stack just for the fat code isn't really efficient
19:01:18 Quit jfc^3 (Ping timeout: 276 seconds)
19:06:35 Join jfc [0] (
19:13:07 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
19:15:08 Join hebz0rl [0] (
19:22:26 Join Jerom [0] (~jerome@
19:23:34 Quit funman (Remote host closed the connection)
19:23:50 Join funman [0] (~fun@rockbox/developer/funman)
19:24:03 Nick evilnick_B is now known as evilnick_withasp (0c140464@rockbox/staff/evilnick)
19:26:14 Nick evilnick_withasp is now known as evilspoon (0c140464@rockbox/staff/evilnick)
19:27:41 Join stoffel [0] (
19:30:02marazwas MWDMA implemented before r13990-ish?
19:31:38maraza MWDMA-only build of r27545 runs almost without a hitch on my nano at 80MHz, i'm kind of wondering if that's the key
19:37:14 Join evilnick [0] (~evilnick@2001:0:4137:9e76:3cde:1636:b5a6:f655)
19:38:02 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
19:40:39 Quit Strife89 (Ping timeout: 240 seconds)
19:44:57 Nick evilspoon is now known as evilnick_B (0c140464@rockbox/staff/evilnick)
19:45:55 Join Strife89 [0] (
19:48:47 Quit evilnick (Ping timeout: 252 seconds)
19:53:52 Quit Jaykay (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
19:54:13 Join Jaykay [0] (
19:54:56kugel07-29 19:54:22.717: DEBUG/RockboxBootloader(6871): dlopen(): Success
19:55:03kugelfunman: thanks for that hint
19:55:40 Quit bug2000 (Ping timeout: 260 seconds)
19:55:57kugelthe reason is indeed that /sdcard is mounted with noexec
19:56:24funmandunno why we need a temp_codec.dll anyway
19:58:14kugelmaybe dlopen on windows desperately wants a .dll file, dunno
20:06:06 Quit komputes (Ping timeout: 276 seconds)
20:07:24***Saving seen data "./dancer.seen"
20:07:25 Join Xerion [0] (~xerion@
20:08:16 Quit Strife89 (Quit: Escorting someone to a doctor's appt.)
20:08:22 Join bug2000 [0] (~bug@unaffiliated/bug2000)
20:08:43bertrikIs there a plan for the next release of rockbox?
20:08:50 Quit Xerion (Client Quit)
20:09:05bertrikend of august / beginning of september for example?
20:09:07 Join Xerion [0] (~xerion@
20:09:39funmandid we move to 4 months schedule?
20:09:54 Quit Jaykay (Ping timeout: 276 seconds)
20:10:34bertriknot that I know
20:11:19funmanthere was 4 months between last 2 releases
20:14:38 Quit milz (Remote host closed the connection)
20:21:17 Join milz [0] (
20:25:53funmanbertrik: perhaps we should ask jdgordon if we need an early release (because of skins changes)
20:26:11 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
20:28:19bertrikAnother 4 weeks before release seems like a good time to shake out any new issues
20:28:36pixelmaI would think it's better to wait a while after the big change today
20:32:55 Join Jaykay [0] (
20:35:15 Quit DerPapst (Quit: Leaving.)
20:36:57 Join komputes [0] (~komputes@ubuntu/member/komputes)
20:40:00 Quit LambdaCalculus37 (Remote host closed the connection)
20:57:34 Join evilnick [0] (~evilnick@2001:0:4137:9e76:3cde:1636:b5a6:f655)
20:58:24 Join bunnyboi [0] (
20:59:29 Part watto
21:00:20 Quit hebz0rl (Quit: Ex-Chat)
21:04:15 Join DerPapst [0] (
21:07:27 Quit evilnick (Ping timeout: 252 seconds)
21:07:35 Quit mt (Read error: Connection reset by peer)
21:08:54 Quit insp_ (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
21:09:58 Quit Jerom (Quit: Leaving.)
21:18:56 Quit stoffel (Ping timeout: 260 seconds)
21:23:33bertrikfunman, did the fix you made to mkamsboot for fuze v2 also apply to newer clip+-es (svn r 27441)?
21:24:10 Join ender [0] (
21:24:14 Join Jennifur [0] (
21:24:33bertrikAlso, do you think there are/will be clipv2 players that will need this and have the sd variant thing?
21:24:50funmanthere won't be
21:25:05funman1/ this version number is present in none of the clipv2 OF
21:25:12funman2/ clipv2 (and v1) is abandoned by Sandisk
21:25:52funmanif there's a new OF anyway, if it follows the same scheme than fuzev2/clip+ it will be recognized by mkamsboot
21:26:36funmanthere's no checking of fuzev2/clip+, we preserve the first block (except the first 8*4 bytes for vectors) of OF unconditionally
21:26:42 Join Jerom [0] (~heidi@
21:27:10 Quit ender` (Ping timeout: 265 seconds)
21:27:15 Quit bunnyboi (Ping timeout: 276 seconds)
21:27:46 Nick ender is now known as ender` (
21:29:22funmanfuzev2 still freeze without possibility for powering off, still related to plugging usb
21:34:10 Join Kitar|st [0] (
21:46:54 Quit captainewkll (Quit: Page closed)
21:47:58 Quit funman (Quit: Reconnecting)
21:48:06 Join funman [0] (~fun@rockbox/developer/funman)
21:49:07 Quit milz (Remote host closed the connection)
21:51:38 Quit bertrik (Read error: No route to host)
21:52:26 Join Kitr88 [0] (
21:52:44 Join milz [0] (
21:52:52 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
21:53:41 Quit Kitar|st (Ping timeout: 240 seconds)
21:54:36 Quit rasher (Quit: Lost terminal)
21:55:02 Join rasher [0] (
21:55:02 Quit rasher (Changing host)
21:55:02 Join rasher [0] (~rasher@rockbox/developer/rasher)
22:02:14 Join evilnick [0] (~evilnick@2001:0:4137:9e76:3cde:1636:b5a6:f655)
22:05:49 Join Buschel [0] (
22:07:28***Saving seen data "./dancer.seen"
22:08:41 Join bunnyboi [0] (
22:10:38 Quit Jennifur (Ping timeout: 240 seconds)
22:11:03Buscheln1s: I'm impressed. with r27612 you pushed wmapro performance a lot! +16% (-5 MHz !) on PP5022.
22:18:04 Join Strife89 [0] (
22:20:17 Join joxx [0] (~jox@
22:45:11 Join B4gder [0] (~daniel@rockbox/developer/bagder)
22:47:58 Quit guymann (Ping timeout: 246 seconds)
22:48:44 Quit Strife89 (Quit: Going home.)
22:49:49 Quit milz (Remote host closed the connection)
22:50:18 Join guymann [0] (
22:55:37 Join milz [0] (
22:59:56 Quit milz (Remote host closed the connection)
23:03:33 Join milz [0] (
23:10:24 Quit milz (Read error: Connection reset by peer)
23:14:08 Join stripwax [0] (
23:14:34 Join milz [0] (
23:15:53 Join bieber [0] (
23:21:30 Quit bertrik (Quit: :tiuQ)
23:21:51 Quit Curulan (Quit: Curulan)
23:22:17 Join Zarggg [0] (
23:22:44 Quit milz (Remote host closed the connection)
23:26:37 Quit guymann (Ping timeout: 260 seconds)
23:29:47 Quit evilnick_B (Quit: Page closed)
23:31:15 Join wodz [0] (
23:33:44wodzamiconn: I didn't try DMA transfers for ATA simply because I only have MPIO as CF target where bytes in ata word are swapped.
23:33:50 Join milz [0] (
23:34:00 Join guymann [0] (~charlie@
23:34:42preglowkugel: pong? :>
23:37:18kugelpreglow: no idea why I pinged you actually :)
23:38:39preglowi'll go back to my tea then :D
23:38:52preglowhow's the android work goin?
23:39:19 Quit jgarvey (Quit: Leaving)
23:40:01 Quit ender` (Quit: What if there were no hypothetical questions?)
23:40:25kugelpreglow: quite good
23:40:40kugelI just got "hello codec" from mpa.codec, but it doesn't decode :(
23:41:10kugelwell, the progressbar and elapsed time doesn't move, I guess it's doing something
23:41:24funmantried PCM already? (easier to debug)
23:41:46preglowgot myself an android phone now, so this stuff is very exciting
23:42:53preglowthe stock music player is just as good as they usually are on mobiles
23:42:56preglowabsolutely shit
23:43:25kugelLadies and Gentlemen :)
23:45:15funman... yes? :)
23:46:12kugelmusic :)
23:46:31funmanmail rockbox-dev@ or it didn't happen!
23:48:44amiconnTorne: LBA48 is *not* compiled on all ATA targets
23:49:59amiconnBut imo the ata transfer limit should be handled in the ata driver, not fat.c, since it's ata specific (and not even fixed)
23:50:39 Quit Jaykay (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
23:51:19preglowkugel: status report plox
23:51:58kugelpreglow: I can play music here :)
23:52:56amiconnFor plain ata it's 256 sectors, for lba48 it's 64K sectors, and e.g. for Ondio mmc it's unlimited
23:53:30funmani don't think SD is limited either
23:53:50preglowkugel: waht, one second you go "codecs are sad :/" then suddenly "chillin' with some music"? :P
23:54:25kugelyea, sort of
23:54:47kugelI just enabled my pcm driver code, then it suddenly started making sound!
23:55:20amiconnMMC over SPI knows 2 sector transfer modes, predefined and unlimited
23:55:32preglowkugel: voodoo
23:56:00preglowi need to sit down and read how android works with native code, i thought shit like this was supposed to be mad cumbersome
23:56:24amiconnOn Ondio we're using unlimited mode, since it's actually easier
23:56:37 Quit dfkt (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
23:56:42funmanpreglow: NDK
23:56:47B4gderkugel: congratulations!
23:56:56kugelpreglow: I think it usually is, but with rockbox you have basically zero dependencies (we don't even need a c library)
23:57:00*funman spots B4gder
23:57:10kugelB4gder: thanks :)
23:57:17*B4gder does his gevaerts impression to hide
23:57:23 Quit guymann (Ping timeout: 265 seconds)
23:57:29preglowkugel: seriously sweet stuff, good work!
23:57:33funmanB4gder: no time to copy some files on the download server?
23:57:41 Join soap [0] (~soap@rockbox/staff/soap)
23:57:56 Join FlynDice [0] (
23:58:06funmanplease cp -r funman/sandisk-sansa/">
23:58:16funmanand cd sandisk-sansa; ln -sf mkamsboot-1.4 mkamsboot
23:58:51 Join guymann [0] (
23:58:54B4gderit'd be nicer if you ship things like that in a tarball or zip or something

Previous day | Next day