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 2011-12-03

00:00:24kugelyay, succesfully rolled my own build for the r0
00:02:28lebelliumkugel: nice :)
00:04:00bluebrotherfunman: true. Once we have gitweb :)
00:04:08*bluebrother curses chinachippatcher
00:04:22bluebrotherwhy can't it use return codes?
00:07:19bluebrotherthis userdata and function pointer concept is really weird :/
00:09:55kugelwill fight the sdl dependency tomorrow
00:24:36bluebrotherhmm, we have onda support in Rockbox Utiltiy but no manual?
00:39:42 Quit pamaury (Remote host closed the connection)
00:44:50 Quit lebellium (Quit: ChatZilla 0.9.87 [Firefox 9.0/20111122192043])
00:46:48 Quit domonoky (Read error: Connection reset by peer)
00:52:02 Join Keripo [0] (
00:52:14 Quit bertrik (Ping timeout: 248 seconds)
00:58:47 Quit bluefoxx (Ping timeout: 245 seconds)
01:08:11***Saving seen data "./dancer.seen"
01:10:45 Join ReimuHak_ [0] (~reimu@
01:16:22 Quit ender` (Quit: Democracy: the theory that the common people know what they want and deserve to get it good and hard. -- H. L. Mencken)
01:17:31 Quit ReimuHak_ (Quit: Leaving...)
02:25:28 Quit Lyude (Remote host closed the connection)
02:27:42 Quit Thra11 (Ping timeout: 248 seconds)
02:44:27 Join bluefoxx [0] (
02:46:53 Quit kugel (Remote host closed the connection)
02:46:59 Join kugel [0] (~kugel@
02:46:59 Quit kugel (Changing host)
02:46:59 Join kugel [0] (~kugel@rockbox/developer/kugel)
02:48:43 Quit tjb0607 (Quit: No Ping reply in 180 seconds.)
02:48:59 Join tjb0607 [0] (
03:08:14***Saving seen data "./dancer.seen"
03:40:49 Quit Guest_13323 ()
03:53:07 Quit evilnick (Ping timeout: 255 seconds)
03:53:39 Join evilnick [0] (
03:53:39 Quit evilnick (Changing host)
03:53:39 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
04:00:03 Quit robin0800 (Ping timeout: 260 seconds)
04:07:46 Quit hilbert (Quit: Textual IRC Client:
04:10:46 Quit amiconn (Disconnected by services)
04:10:47 Quit pixelma (Disconnected by services)
04:10:47 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:10:49 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:10:51 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:11:09 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:13:36 Join mshathlonxp [0] (msh@
04:17:15 Quit [7] (Disconnected by services)
04:17:26 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:33:58 Join dreamlayers [0] (~bgjenero@rockbox/developer/dreamlayers)
04:40:46 Quit bluefoxx (Quit: bluefoxx)
04:50:44 Quit Keripo (Quit: Leaving.)
04:57:11 Quit dreamlayers (Remote host closed the connection)
05:07:21 Quit pixelma (Disconnected by services)
05:07:23 Quit amiconn (Disconnected by services)
05:07:23 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
05:07:23 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
05:07:25 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
05:07:45 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
05:08:17***Saving seen data "./dancer.seen"
05:48:56 Join Rob2223 [0] (
05:53:04 Quit Rob2222 (Ping timeout: 244 seconds)
06:04:19 Nick icarusfactor is now known as factor (~factor@
07:08:22***Saving seen data "./dancer.seen"
07:32:42 Join hskf [0] (
07:51:08 Join y4n [0] (y4n@unaffiliated/y4ndexx)
08:29:23 Quit MethoS- (Remote host closed the connection)
08:41:14 Join Keripo [0] (
09:08:26***Saving seen data "./dancer.seen"
09:10:21bluebrotherI'm wondering if we should remove PPC support from Rockbox Utility, since OS X 10.7 is now out for a while.
09:10:40bluebrotherlooking at chinachippatcher makes me believe that it won't work as expected on PPC.
09:16:27 Quit hskf (Quit: Colloquy for iPhone -
09:22:57kugelfunman: your commit message says CROSS_COMPILE but the code says CROSS_PREFIX
09:30:13 Join nosa [0] (
09:30:39 Quit nosa-j (Ping timeout: 260 seconds)
09:30:39 Nick nosa is now known as nosa-j (
09:49:11 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
10:02:49 Join n1s [0] (
10:02:49 Quit n1s (Changing host)
10:02:49 Join n1s [0] (~n1s@rockbox/developer/n1s)
10:03:30 Join lorenzo92 [0] (
10:04:00lorenzo92kugel: some news? :)
10:06:13funmankugel: oops
10:06:53funmankugel: i hope you don't want me to use git svn propedit svn:log ......
10:09:10bertrikfunman, which git mirror are you using?
10:18:13kugelfunman: you could fix the makefile :P (/me prefers CROSS_COMPILE)
10:19:23kugellorenzo92: hacked a bit in the path handling and configure
10:20:19lorenzo92and for alsa still nothing?
10:20:34kugelhey, I just woke up and wanted something light :)
10:20:43lorenzo92haha ok :D
10:21:12funmanbertrik: no git mirror, i'm using git-svn only
10:21:30funmankugel: hm no problem if you want to change it
10:21:39funmani think linux use CROSS_COMPILE
10:21:41bertrikso instead of git pull, you do git svn rebase?
10:22:25*kugel does both :)
10:22:37funmanbertrik: i didnt change clip zip status wrt plugins
10:23:10kugelbertrik: you need svn rebase too if you're using the git mirror
10:26:18bertrikI did git svn rebase only just before git svn dcommit
10:30:12bertrikfunman, any plans to work on other clip zip stuff?
10:30:34funmanyeah i want to make a C lcd_blit_yuv
10:31:08funmansuch a small screen with such a fast CPU : should be good enough, AND give us a reference C implementation
10:31:39funman-> bedtime
10:32:25bertrikRDS basically works, but it is a bit stuck on some weirdness with the RDS station name and radio text not getting updated properly somewhere in the skin engine
10:32:38kugelfunman: SDL has a reference C implementation
10:32:54bertrikthe SSP connection is relatively slow
10:33:16kugelit's used on android too
10:33:35bertrikfor every byte sent over SSP, the CPU can do 320 cycles
10:34:25 Quit evilnick (Ping timeout: 252 seconds)
10:34:55 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
10:34:56bertrikand 1 pixel is two bytes
10:39:16CIA-10New commit by bluebrother (r31114): Split up chinachippatcher. ...
10:41:19CIA-10r31114 build result: 2 errors, 0 warnings (bluebrother committed)
10:41:47CIA-10New commit by bluebrother (r31115): Extend return codes for chinachip_patch(). ...
10:41:51CIA-10New commit by bluebrother (r31116): Consolidate libraries list in project file a bit. ...
10:43:40CIA-10r31115 build result: 2 errors, 0 warnings (bluebrother committed)
10:44:19 Quit Bagder (Ping timeout: 252 seconds)
10:44:23CIA-10New commit by bluebrother (r31117): Rockbox Utility: build chinachippatcher as library. ...
10:45:10CIA-10r31116 build result: 2 errors, 0 warnings (bluebrother committed)
10:46:42CIA-10r31117 build result: 2 errors, 0 warnings (bluebrother committed)
10:47:22lorenzo92kugel: was thinking...I used 16 bpp for the screen, but R0's screen is actually capable of 24 bit. Should we improve that?
10:52:09kugellorenzo92: definitely!
10:52:25lorenzo92to add to the todo list so :)
10:52:29kugelunfortunately rockbox can't do 24bit :)
10:52:35lorenzo92uhm ok
10:52:39lorenzo92no matter then
10:56:30 Join Bagder [241] (~daniel@rockbox/developer/bagder)
10:58:10kugellorenzo92: got rid of libuisimulator :)
10:58:38 Join lebellium [0] (
10:59:23 Join AlexP_mob [0] (~AlexP@rockbox/staff/AlexP)
11:08:29***Saving seen data "./dancer.seen"
11:13:58kugellorenzo92: working on native linux tick timer now (more leightweight than the sdl one)
11:14:27lorenzo92yeah indeed. I had all these ideas in mind...;)
11:14:45kugelit's not that tricky
11:14:59kugelI implemented that for android, it's mostly copy&paste from android instead of sdl :P
11:19:50 Join TheLemonMan [0] (
11:29:29lorenzo92ah the latest svn compiling fine? because lovasoa told me he had a strange error....but I didn't test it yet
11:42:19kugellorenzo92: compiling without PLATFORM_SDL now
11:43:42 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
11:44:37 Quit AlexP_mob (Quit: Bye)
11:44:43 Quit bertrik (Ping timeout: 255 seconds)
11:56:11 Quit desowin_ (Read error: Connection reset by peer)
12:09:26kugellorenzo92: any way to speed up boot?
12:11:01lorenzo92eh I guess not...need to load the kernel
12:11:27lorenzo92or what do you mean?
12:14:41kugellorenzo92: the time during the first samsung logo is annoyingly log
12:14:57lorenzo92I know....
12:15:12lorenzo92ah I hope you use the safe mode to test executables
12:15:21kugelno :)
12:15:28kugelhow to use that?
12:15:39lorenzo92eh do it! It's faster....just insert usb cable and hold "Back" key
12:15:53lorenzo92this is one of the feature of my mod :)
12:17:21kugelgonna see what I can do with alsa now
12:20:32 Join bertrik [0] (~bertrik@
12:20:37 Quit bertrik (Changing host)
12:20:37 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
12:21:26lorenzo92you will notice also the "internal" hidden drive...
12:22:17 Nick bertrik is now known as bertrik_train (~bertrik@rockbox/developer/bertrik)
12:22:59kugellorenzo92: do you happen to know which alsa version that is?
12:23:06kugelon the device that is
12:23:17lorenzo92yes wait I need to find the script logs :)
12:23:27kugel0.9 by chance?
12:24:19lorenzo92[ 0.960000] Advanced Linux Sound Architecture Driver Version 1.0.15 (Tue Nov 20 19:16:42 2007 UTC).
12:25:16lorenzo92ah I needed to static link the library (alsa) to sdl, otherwise I had always a segmentation fault
12:25:23lorenzo92maybe we need to do the same with RB
12:26:03kugelI hope not :(
12:26:44 Quit funman (Read error: Operation timed out)
12:28:09 Join funman [0] (~fun@rockbox/developer/funman)
12:31:11 Quit funman (Read error: Connection reset by peer)
12:31:59 Join liar [0] (
12:32:44bertrik_trainI think I know now what is going on with the git mirror. It's still a proper git repo, it's just not getting mirrored anymore beyond 31106
12:33:11 Join funman [0] (~fun@rockbox/developer/funman)
12:36:53kugellorenzo92: works, no segfault
12:37:00lorenzo92yeah yeah!
12:37:05lorenzo92really sound is working?
12:37:07lorenzo92great :D
12:37:27kugellittle hiss when starting playback but that's also in the simulator
12:37:36lorenzo92okay then I will pass to lebellium to test if random glitches are gone...
12:37:50kugelI suspect header version missmatch when you linked dynamically
12:38:26 Join bluefoxx [0] (
12:38:26kugelotoh it would perhaps be better to link statically. unless you need to recompile alsa?
12:38:40lorenzo92nono it's not needed
12:38:48lorenzo92just link the cramfs's lib
12:39:54kugelnot sure what to do with the libs now
12:40:07kugeleither put them into the toolchain, or into the rockbox tree
12:40:15kugelwhat would people prefer?
12:40:46lorenzo92the easies way :)
12:41:05lorenzo92so probably rockbox tree
12:46:23lorenzo92when you have a patch, let me know that I want to test :p
12:46:35kugellorenzo92: if we're distributing it we would need to include its source as well
12:46:57 Join hilbert [0] (
12:47:01TheLemonManwhat board is the player based off ?
12:47:14lorenzo92thelemonman: imx37
12:47:48lorenzo92I guess the linux stuff is more or less the same on q2!
12:48:34TheLemonManyessir, but i went for the hard way
12:49:17TheLemonManthe bootloader is kind enough to load whatever you upload and run it
12:49:29 Quit qnm (Ping timeout: 244 seconds)
12:49:54TheLemonManand lots of the needed infos (namely lcd and keys) can be found in the bootloader with little effort
12:50:48lorenzo92eh with R0 the story is a bit different I guess....apart from a special cable easy to build, it needs a special soft with samsung specific files
12:51:07 Join qnm [0] (
12:51:11TheLemonMandoes it enters the recovery mode by pressing the hold button ?
12:52:14lorenzo92it needs a 300 kohm resistor between pin5 and gnd
12:53:12kugellorenzo92: special cable?
12:53:41TheLemonManmight be because of the different wiring
12:54:10kugelthis ypr0 from ebay came with a kodak micro usb cable
12:56:35lebelliumkugel: the recovery tool requires a special USB cable. Only me outside Samsung owns an original one and some others like Lorenzo built their own one with a 300kohm resistor
12:56:58TheLemonManhow hard has been to port RaaA to the R0 ?
12:59:14kugellebellium: oh!
13:01:54lorenzo92TheLemonMan: not too much but I needed to figure out something like ioctl calls for afe.ko module
13:02:23TheLemonManohwell, you can just ask samsung for the kernel source as i did
13:02:32TheLemonManthey're pretty kind
13:02:54 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
13:03:11lorenzo92yes we have it! but not the source for every module :(
13:05:43lorenzo92and moreover recompiling the kernel by us produces lots of strange runtime things (crashes etc)
13:07:13lorenzo92but at least it's not necessary to do this for rockbox
13:07:41TheLemonManmmh, where did you put the hook for running the custom application ?
13:08:08TheLemonMani wanted to include it in the init.rc but i was afraid of breaking permissions
13:08:30***Saving seen data "./dancer.seen"
13:09:52 Quit Keripo (Quit: Leaving.)
13:10:04lorenzo92uhm I exploided "profile" file, here I call a script that call every mod/DRK/safe mode etc
13:10:11lorenzo92and also rockbox
13:10:38lorenzo92of course is all rooted
13:14:23lorenzo92i'll back in 20 mins ;)
13:14:29TheLemonMandoes your rootfs have symlinks in it ?
13:21:50 Join dfkt [0] (dfkt@unaffiliated/dfkt)
13:22:01 Join MethoS- [0] (~clemens@
13:22:49 Quit evilnick (Ping timeout: 260 seconds)
13:33:54 Join stripwax [0] (
13:35:02kugelare rockbox pcm samples interleaved or not?
13:36:44*kugel always forgets this
13:39:12 Quit stripwax (Ping timeout: 245 seconds)
13:39:55 Join wodz [0] (
13:40:27wodzkugel: RLRLRL....
13:43:39kugelso interleaved. thanks
13:46:47 Join Horscht [0] (
13:46:48 Quit Horscht (Changing host)
13:46:48 Join Horscht [0] (~Horscht@xbmc/user/horscht)
13:50:14lorenzo92I'm back
13:50:33lorenzo92yes rootfs has lots of symlink
13:50:43lorenzo92mainly for busybox
13:50:46wodzkugel: You inspired me how to solve backtrace() from panicf(). I'll have to verify this on hardware but calling swi at the end of panicf() should do the trick
13:57:02 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
14:06:10 Quit bertrik_train (Remote host closed the connection)
14:06:37kugelpwodz: the swi isn't going to work on raaa
14:06:59wodzyes, thats the downside of this idea
14:07:03kugelpdid you try Inliner assembly to get pc and so?
14:07:18wodzit is done like this currently
14:07:40Tornecan you not just trigger a data abort?
14:08:34wodzany 100% proof asm snippet? I am not that good at arm asm
14:08:42kugelpwodz: the last patch does builtin_return_adress no?
14:09:13wodzkugelp: can't remember
14:09:32kugelpthat would be lr, not sp
14:10:20wodzyes but I had a feeling that we are interested in skiping backtrace() function itself
14:10:48kugelpskipping how?
14:11:24wodzif you pass LR you start unwinding the stack *AFTER* backtrace() itself
14:11:48Torneyes, but the stack will be wrong
14:12:11Torneyou need the pc to match sp..
14:12:23Tornereturn address is never useful
14:12:52kugelppassing the current pc and lr with inline asm should be simple
14:13:18Tornei dont see why you would need to, though..
14:13:41wodzpassing current pc, sp is easy but it produce empty backtrace
14:13:43Tornecant it just use the current pc and sp on entry to the unwinder?
14:14:12Torneyou only need to pass something in if it's not the current context
14:14:29Tornei.e. if you are in an exception
14:14:41kugelpyes the patch does that
14:15:19kugelpwe want it to work also from panicf
14:15:36Tornewhy does that not work in pancf?
14:15:41wodzah yes, thats why I diverged from original because I was originally interested in backtrace from exception
14:15:43Torneit should work from anywhere
14:15:47Torneexvept exceptions
14:16:01Tornewhere you need to explixitly get the normal mode lr/sp
14:16:20kugelpthe patch works from exception, but not panicf :-)
14:16:44Tornei havent read it
14:17:22Tornei am talking about using the *current* stack/pc when you get int the unwinding code
14:17:25Tornenot panicf's
14:17:40Tornesuch that the top thing on the stack is the unwinder itself
14:17:52Tornei.e. do not pass anything in at all
14:18:54Tornehm, alternatively
14:18:56wodzhmm you mean if unwinder gets SP=0 PC=0 it should get the current one?
14:19:18Torneanyway, thwres a different optipn
14:19:38Tornejust use a naked inline asm shim ariund panicf
14:20:20Torneasm(mov r0, lr); asm(mov r1, sp); asm(b real_panicf)
14:20:34Torneto pass the location in as arguments :)
14:20:52Torneneeds to be a naked function. or just written in real asm
14:21:30kugelpyou surely mean around backtrace, not panicf?
14:21:35Tornepanicf doesnt return, which is probably why the unwinder is giving you an empty stack
14:21:59Tornethe unwinder goes forwards to work out what to do
14:22:18Tornefunctions that dont return will terminate it, no?
14:22:22kugelppanicf still wants to print the panic message
14:22:32Torneyes, and?
14:22:39Tornewhat's the problem
14:22:50Torneyou are juat capturing where you were first
14:23:01Torneprint what you like
14:23:16Tornethen pass the params on to thw backtrace generator
14:23:24wodzI'll play with this a bit during next week
14:23:25kugelpr0 has the parameter for panicf, don't overwrite it :-)
14:23:47Tornewell use r1 and r2 then :)
14:23:56Tornethe numbers dont matter
14:24:09kugelppanicf also calls other functions
14:24:11Torneyoubadd two more parameters passed in registers, is all
14:24:21Tornewhy does that matter?
14:24:26lorenzo92kugelp: a little question. Rockbox has a log stream right? where is outputted? why isn't created any log file in R0 (not a register lol)?
14:24:45Tornethe logs go to a buffer in ram
14:24:47kugelpno it doesn't
14:24:50Torneand they are disabled hybdefault
14:25:29kugelpor you enable logf built in which case the logs are in the debug menu
14:25:46lxs-makot0win, my sansa crashed again
14:25:48lorenzo92ah! ok thanks! Now I remeber once I enabled it :)
14:26:00Tornekugelp: you are saving sp and pc as of the place that *called* panicf
14:26:12Tornenothing thatvhsppens after that matters
14:26:31Tornethe naked function means the compiler wont emit a preamble that changes te stack
14:27:04Tornewo it will unwind starting at the caller of panicf
14:27:06kugelpI worry about the regs you save sp and pc to
14:27:22Tornewhat about them?
14:27:43Tornethey are jyst function parameters
14:28:19Tornevoid panicf(char* msg)
14:28:26kugelpah you mean to change the signature of that reak_panicf?
14:28:37Torneadd two new parametera
14:28:43Tornepopulate them from naked asm
14:28:45kugelpokay that would work
14:28:48Tornethen tailcall
14:29:55Tornethe unwinder really shpuld work from anywhere though :)
14:30:14Tornei would expect it to be just fine to call it and keep executing
14:30:51 Join Misan [0] (
14:30:53 Quit bluebrother (Disconnected by services)
14:30:53Torneand have a stack produced that starts with unwinder, then the rhing that called it
14:30:54 Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother)
14:30:56Tornebut noreturn functions will probably break it.
14:30:58Torneso pancf may need something special
14:31:37Torneanyway nearly arrived at my family xmas dinner :)
14:31:50Tornebad typingwas due to cellphone ;)
14:33:02 Quit fs-bluebot (Ping timeout: 248 seconds)
14:34:10 Join fs-bluebot [0] (
14:34:19 Quit lorenzo92 (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111115183813])
14:53:16 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
14:55:36 Join Xerion [0] (
14:58:46CIA-10New commit by bluebrother (r31118): Don't show the full path when building librbspeex.a
15:00:21CIA-10r31118 build result: 2 errors, 0 warnings (bluebrother committed)
15:01:51 Quit Horscht (Ping timeout: 244 seconds)
15:02:11 Join Horscht [0] (~Horscht@xbmc/user/horscht)
15:08:32***Saving seen data "./dancer.seen"
15:10:49 Join jlbiasini [0] (
15:12:20 Join stripwax [0] (
15:14:33 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
15:14:33 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
15:17:50 Quit stripwax (Ping timeout: 252 seconds)
15:46:37 Join lovasoa [0] (~olojkine@2a01:e35:8a2e:8080:e2b9:a5ff:fe5b:ca7b)
15:51:04TheLemonManhrm im trying to use the rockbox toolchain to cross compile an utility but looks like ld its missing crt0
15:54:29n1swe provide our own crt0 so maybe you need to link to it
15:54:37 Join lorenzo92 [0] (
15:58:22 Quit efyx (Ping timeout: 245 seconds)
15:59:46lorenzo92kugel: Now I have some time to test your patch if you want :)
16:01:51TheLemonMann1s, how do i remove the rockbox cross compiler ?
16:02:17TheLemonMani mean, is everything installed in /usr/local ?
16:02:56n1sshould be
16:03:19n1sin fact i'm certain it is :)
16:03:25lovasoakugel: I'd really like to test the patch too :)
16:03:46CIA-10New commit by desowin (r31119): Sansa Connect: Fix GIO initialization and screen wakeup.
16:05:22CIA-10r31119 build result: 2 errors, 1 warnings (desowin committed)
16:07:31lorenzo92lovasoa: it seems that kugel has sucefully compiled last svn without problems, maybe has he fixed something?
16:12:46 Join efyx [0] (
16:16:42 Join T44 [0] (
16:20:09 Quit Topy (Ping timeout: 260 seconds)
16:21:13CIA-10New commit by desowin (r31120): Sansa Connect: Fix yellow
16:22:54kugellorenzo92: the one with alsa?
16:23:00CIA-10r31120 build result: 2 errors, 0 warnings (desowin committed)
16:23:10lorenzo92yes and maybe other fixes you did ;)
16:25:29kugelgive me a moment
16:30:06 Part jlbiasini
16:30:11 Join jlbiasini [0] (
16:30:15lorenzo92I'm not in hurry ;)
16:33:51 Part jlbiasini
16:34:41CIA-10New commit by pamaury (r31121): imx233: add control for more clocks, add debug clock debug screen
16:36:22CIA-10r31121 build result: 2 errors, 0 warnings (pamaury committed)
16:42:31pamaurywhat is the difference between CPUFREQ_DEFAULT and CPUFREQ_NORMAL ? Most target pick the same, gigabeat-s seems to imply that normal<default<max and system.c seems to imply that default<normal<max !
16:51:00 Join lebellium_ [0] (
16:52:56 Quit lebellium (Ping timeout: 244 seconds)
16:53:10 Nick lebellium_ is now known as lebellium (
17:00:45 Quit Llorean (Quit: Leaving.)
17:02:40kugellorenzo92: uploaded to the tracker
17:04:40lorenzo92uhm okay I've seen you changed something for the wasn't necessary because in the new firmware I did something else but fine anyway ;)
17:05:05lorenzo92now I try it
17:05:59kugelI dont have the new firmware :)
17:06:17kugel.rockbox is nicer anyway IMO
17:07:30lorenzo92yes yes indeed
17:07:43 Quit niekie (Read error: Connection reset by peer)
17:08:34***Saving seen data "./dancer.seen"
17:10:49lorenzo92I'm using svn :(
17:14:24 Quit wodz (Ping timeout: 260 seconds)
17:21:38kugellorenzo92: patch -p1
17:21:48lorenzo92thanks :)
17:24:38 Join Thra11 [0] (~thrall@
17:25:45lorenzo92I get an erro
17:25:57lorenzo92multiple definition of `fat_size'
17:26:06lorenzo92undefined reference to `timer_create'
17:26:18lorenzo92undefined reference to `timer_settime'
17:26:40lorenzo92it is after LD rockbox...
17:29:44 Join jlbiasini [0] (
17:29:57kugellorenzo92: add -lrt to the LDFLAGS
17:30:03kugelI wonder why it's not in the patch
17:30:09kugelfor fat_size I don't know
17:33:41lorenzo92home/lorenzo/YP-R0_DEV/rockbox/rockbox/uisimulator/common/io.c:535: first defined here
17:33:53lorenzo92the first error is solved, this fat is still there
17:34:12lorenzo92seems that there is still around the simulator
17:35:04kugeltry make clean
17:35:11kugelio.c shouldnt be compiled
17:36:34kugeland run configure again
17:47:16 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox)
17:50:28CIA-10New commit by thomasjfox (r31122): Backport r30935 - Fix format string
17:55:06CIA-10New commit by thomasjfox (r31123): Backport r30964 - Fix copy'n'paste bug in skin engine ...
17:56:49 Quit TheLemonMan (Quit: Screw you, im going home)
18:02:01thomasjfoxAnything wrong with the automated build?
18:02:33thomasjfoxOr is it not enabled for the 3.10 branch?
18:02:41gevaertsIt only builds trunk
18:03:31thomasjfoxOk. Thanks
18:03:57thomasjfoxThen let's commit something to trunk - I need my medicine
18:06:47CIA-10New commit by thomasjfox (r31124): Remove redundant assigment to "dummy" variable ...
18:07:30CIA-10r31124 build result: 404 errors, 0 warnings (thomasjfox committed)
18:08:32kugeljhMikeS: ping
18:09:34gevaertsthomasjfox: I think we usually do that sort of thing with "(void) dummy;" insead of "dummy = dummy;"
18:09:46thomasjfoxYeah, I just looked it up ;)
18:10:19thomasjfoxI clearly remembered that we marked unused variables
18:11:59CIA-10New commit by thomasjfox (r31125): Fix red: Mark variable as unused the rockbox way
18:13:42CIA-10r31125 build result: 2 errors, 0 warnings (thomasjfox committed)
18:18:14CIA-10New commit by thomasjfox (r31126): Backport r31124 - Remove redundant assigment to "dummy" variable ...
18:18:34gevaertsthomasjfox: is backporting that sort of thing *really* needed?
18:19:03thomasjfoxgevaerts: As far as I can tell, the code is quite central
18:19:14gevaertsYes, but it's also not a bug
18:19:32thomasjfoxThat is true. Do we only backport bugfixes?
18:19:42thomasjfoxOh sorry
18:19:55thomasjfoxThen it's not needed
18:20:09gevaertsNo point in reverting it again though
18:20:16lorenzo92kugel: okay testing rockbox on device ;)
18:20:25lorenzo92noticed new icons :D
18:20:37thomasjfoxWell, at least the other two backports are bugfixes
18:22:50 Quit lorenzo92 (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111115183813])
18:24:30 Quit JdGordon (Ping timeout: 248 seconds)
18:29:25 Quit thomasjfox (Remote host closed the connection)
18:31:20jhMikeSkugel: sir?
18:32:04 Join JdGordon [0] (
18:32:05 Quit JdGordon (Changing host)
18:32:05 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
18:32:13lovasoakugel: With -Wall, I get a lot of Warning SAMSUNG_YPR0_PAD is not defined. Is it normal?
18:32:45 Join stripwax [0] (
18:34:17funmankugel: any idea how to share the sdl yuv code without copying it in lcd-clipzip.c ?
18:39:21funmanperhaps move that lcd_blit_yuv somewhere so we can include it in SOURCES
18:40:02funmanalso i have no idea why we don't use yuv dithering for all models
18:43:30kugeljhMikeS: does pcm_get_more_callback() always give 1024k?
18:43:57kugelI have trouble, for some strange reason, to implement other sizes in my alsa backend
18:45:15kugelas snd_pcm_writei() outright rejects smaller sizes I try to collect samples in a spare buffer which gives very broken and noisy sound
18:45:35jhMikeSkugel: it will with the mixer running
18:45:44jhMikeSdirect pcm is another story
18:46:37kugelI'm also a bit confused about pcm_play_dma_started_callback(). why not just call it directly after get_more()?
18:46:51 Quit stripwax (Ping timeout: 244 seconds)
18:47:30jhMikeSkugel: It's sort of is, after the transfer to the audio device has begun though
18:48:36kugelit seems I need more than 1 pcm_play_get_more_callback() calls between feeding the data to the hw, so I'm just calling it after
18:49:16jhMikeSit should only be called once per call to pcm_get_more_callback, if pcm_get_more_callback returns 0, don't call it
18:50:22kugelwhy make it so complicated to call it after feeding the hw?
18:50:39jhMikeSit doesn't seem to be normally
18:51:10kugelI'm looking at pcm-sdl.c and pcm-android.c
18:52:21jhMikeSnot the nicest things in the world. normally it expects to be called in parallel with DMA, not with the hardware dangerously low on data
18:52:22kugelit seems I need more than 1 pcm_play_get_more_callback() calls between feeding the data to the hw, so I'm just calling it after
18:53:00jhMikeSif you call the callback, call the started callback sometime afterwords
18:53:31kugeljust so the hw buffers can be filled earlier?
18:54:29jhMikeSthey must be or mixing will cause dropout. there isn't enough time to do mixing when a hw fifo has only has a few samples in it
18:55:00kugelpcm-android is unecessarily complicated then. the hw buffers won't be filled before the callback returns (the callback only fills a temp buffer)
18:55:32kugelthanks for the explaination
18:56:10Lalufureading through the tagcache code, it's not a good idea to spend too much time between tagcache_search() and tagcache_search_finish(), right, as tagcache_search creates a lock on the tagcache?
18:56:13kugeloh uh, I'm wrong. it that way at one point
19:00:58 Quit ukleinek (Ping timeout: 244 seconds)
19:01:27 Join wodz [0] (
19:01:37jhMikeSkugel: isn't it possible to set a watermark on android yourself? maybe I misunderstood what the one object is for.
19:02:37wodzpamaury: when I was porting rb to HD200/HD300 I was told that CPUFREQ_DEFAULT is with PLL off, CPUFREQ_NORMAL is well normal (with PLL on), CPUFREQ_MAX is when boosted
19:05:29kugeljhMikeS: yes it is, and it's set rather conservative
19:08:38***Saving seen data "./dancer.seen"
19:09:05kugeljhMikeS: what would be a good watermark for alsa? what does sdl use? is 2k too high or too low?
19:11:04jhMikeSI'm not sure but it might have been 4k or so
19:12:46jhMikeSsdl sometimes had bad drivers that ask for 48K (it's rather horrid when it does)
19:13:43jhMikeSthat even varies with how the library is made
19:16:53 Quit MethoS- (Read error: Connection reset by peer)
19:24:18 Join stripwax [0] (
19:29:34 Quit stripwax (Ping timeout: 248 seconds)
19:35:11 Join MethoS- [0] (~clemens@
19:42:05 Join Steff8000 [0] (
19:44:53 Quit Steff8000 (Client Quit)
19:48:03 Join Pedrital [0] (
19:48:13PedritalHey people !
19:49:28 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
19:50:12PedritalI read through the FAQ and the iPod FAQ but I have not been able to know how a rockbox iPod behaves on a "dock" , like this :
19:51:44gevaertsPedrital: it depends on which ipod you mean. Rockbox has some support for docks on ipods up to the 1st generation nano and the video, but not for the 2nd gen nano or the classic
19:52:06PedritalToo bad...
19:52:28gevaerts has some details
19:54:13PedritalWell I guess i'm ok with the normal "RCA to Jack" option, but docks are cool.
19:54:26Pedritalthanks for the link !
20:00:04PedritalIt seems to be quite random, depending on the hardware you're plugging it in.
20:03:46PedritalOh, and does Rockbox support FLAC ?
20:08:30gevaertsof course
20:26:19 Quit mgue (Ping timeout: 260 seconds)
20:27:35 Join mgue [0] (
21:02:17 Join ender` [0] (
21:08:42***Saving seen data "./dancer.seen"
21:19:00 Join swilde [0] (~wilde@2001:470:9c7c:abc:948f:fcff:fef6:e221)
21:24:01 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
21:35:02 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
21:37:13 Quit hilbert (Quit: Computer has gone to sleep.)
21:38:58 Join hilbert [0] (
21:39:22 Quit wodz (Quit: Leaving)
21:39:31 Part lovasoa
21:42:33funmanhow about target/generic or target/common to host common functions?
21:43:08funmanor perhaps this could go in firmware/drivers/lcd-16bit*
21:43:27funmanwith e.g. #ifndef LCD_OPTIMIZED_YUV_BLIT
21:50:53 Join lorenzo92 [0] (
21:51:00 Join notben [0] (
21:51:14 Nick notben is now known as iamben (
21:54:23 Quit lorenzo92 (Remote host closed the connection)
21:56:15 Join lorenzo92 [0] (
21:58:37funmanperhaps with weak symbols there would be no need for #ifdef HELL
22:02:01kugellcd-16bit Sounds good
22:02:42 Part iamben
22:06:35 Join Keripo [0] (
22:11:30 Quit y4n (Quit: only amiga makes it possible)
22:17:41 Part jlbiasini
22:24:32funmanusing attribute((weak)) the function is persent in lcd-16bit.o
22:24:39funmanbut optimized as 'bx lr' in rockbox.elf
22:25:09 Join Stummi [0] (~Stummi@rockbox/developer/Stummi)
22:28:03kugelfunman: weak breaks with cygwin (old gcc)
22:29:58funmancan't we upgrade gcc?
22:32:02funmanalso who needs cygwin when there's mingw
22:32:13kugelwho knows
22:32:27*Torne alllmost explains why mingw is cygwin, but can't be bothered to do that one again :0
22:32:49kugeli think recent cygwin has newer gcc
22:32:56funmanif mingw is cygwin, then cygwin has recent enough gcc
22:35:17funmanweird the function is here in the .elf but not referenced by the global symbol
22:35:40funmanaaah sorry
22:35:50funmani didn't remove the empty lcd_blit_yuv()
22:36:23funmanthat proves overloading works correcrly
22:38:39 Quit lorenzo92 (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111115183813])
22:39:00funman# Your branch is ahead of 'origin/master' by 16 commits.
22:39:03funmanrockbox-old is not updateD?
22:39:36CIA-10New commit by funman (r31127): lcd_blit_yuv: move from sdl driver to lcd16bit-common ...
22:39:41CIA-10New commit by funman (r31128): Clip Zip: fix yuv playback
22:40:15dfktdoes that mean video playback on the zip is possible? :)
22:41:14CIA-10r31127 build result: 7 errors, 0 warnings (funman committed)
22:41:32 Quit Keripo (Quit: Leaving.)
22:42:58CIA-10r31128 build result: 7 errors, 0 warnings (funman committed)
22:44:53funmani don't get the recorder error
22:45:19funmancode was under -#ifdef HAVE_LCD_COLOR
22:46:57scorchedfkt: so did you make these images?
22:47:07dfktscorche, yes
22:47:22 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
22:47:45scorchedfkt: ok - just making sure there are no potential license issues ;)
22:48:06dfktnah, they were based on the official, freely available press shots
22:48:36CIA-10New commit by funman (r31129): lcd_blit_yuv: move to lcd-16bit-common.c
22:50:20*funman spots code duplication in drivers/lcd*
22:50:25CIA-10r31129 build result: 2 errors, 0 warnings (funman committed)
22:50:37funmanah recorder error was not my fault ...
22:53:45funmanmrobe +432 bytes
23:07:11 Quit Misan (Remote host closed the connection)
23:08:02 Join jlbiasini [0] (
23:08:43***Saving seen data "./dancer.seen"
23:10:08 Join b0hoon [0] (
23:13:58 Quit liar (Read error: Connection timed out)
23:15:25funmanhum i didnt spot that every target got red
23:15:36 Join saratoga [0] (9803ec71@gateway/web/freenode/ip.
23:15:42funmanlooks like the function is not removed at all, just not referenced by lcd_blit_yuv
23:19:19CIA-10New commit by funman (r31130): lcd-16bit: move common code to lcd-16bit-common.c
23:20:56CIA-10r31130 build result: 2 errors, 0 warnings (funman committed)
23:22:41funmanand of course sometimes only one version of the duplicated function got optimized
23:29:14 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
23:32:32 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
23:34:42 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
23:36:04 Join ungali [0] (
23:39:10CIA-10New commit by funman (r31131): lcd-16bit: move lcd_bitmap and lcd_bitmap_transparent to common code
23:40:56CIA-10r31131 build result: 2 errors, 0 warnings (funman committed)
23:41:36 Join Keripo [0] (
23:51:05 Quit Stummi (Quit: Bye!)
23:52:28scorcheas a note to anyone else who has admin on the themesite - it takes its player images from which is in www on svn - to "add" the image to the themesite, scroll down to edit/add targets on the admin page, edit the target you wish to change, then put the image name for the picture field
23:52:47 Join lovasoa [0] (~olojkine@2a01:e35:8a2e:8080:e2b9:a5ff:fe5b:ca7b)
23:57:34 Part ungali

Previous day | Next day