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 2010-03-26

00:01:08xiainxokay, how long would the translations stuff take?
00:01:13xiainxProbably not too long
00:01:59CIA-5New commit by funman (r25336): Make storage alignement use cache alignement macros ...
00:03:02domonokyxiainx: maybe you can split the rest of the summer into different improvements for existing services. But make sure to pick specific tasks :-)
00:03:30xiainxyeah, so what exactly has to be fixed with the translations?
00:03:44CIA-5New commit by funman (r25337): Accept expressions in CACHE_OVERLAP() macro
00:05:39domonokyxiainx: for the core translations i am not really sure what needs fixing, maybe you should try it a bit, and see if you find something missing ? :-)
00:06:13xiainxOkay, so some testing and bug fixing
00:06:28domonokyfor example i think there is no way to start a new translation for languages not already existing in rockbox
00:07:09domonokyand ofcourse the rbutil transaltions, bluebrother should be able to tell you more about that :-)
00:07:18 Join funman [0] (~fun@rockbox/developer/funman)
00:07:27RoronoaZoroI would like to discuss about what to do this summer as a project i have 2 ideas in mind 1. Audio codec optimisation or 2. Improved video playback
00:07:42RoronoaZoroi am not able to decide which one to do..
00:08:06domonokyRoronoaZoro: nice, do you have expirience in this area ?
00:08:38ender`amiconn: if we're going exotic: :)
00:09:13RoronoaZoroI know C and Assembly on x86 and DSP... not much in Audio/Video Codec Theory.
00:09:38funmanapps/plugins/mpegplayer.c:862 : i'm not sure why an uncached address is used there
00:10:00funmancould it be that the buffer is shared between CPU & COP ?
00:10:07bluebrotherwell, the basic code for rbutil translation is dead simple. It's less than 200 LOC right now, and the most time I've spent was to understand how the simplexml in php works
00:10:38RoronoaZoroi had talked to saratoga about it he said i could read about MDCT and FFT and i have done that part but i am still not able to decide on one
00:11:42xiainxOkay, and does it need to be improved in some way to add something?
00:11:45xiainxOr is it fine as is?
00:12:03amiconnender`: Pfft, x86 is boring ;)
00:12:07funman00:11 -logbot( I have never seen 'jhmikes'
00:12:12funmanwho else could answer?
00:12:24domonokyRoronoaZoro: so codec optimisations is diving really deep into the codecs and optimising specifc parts (like fft) with better code/ asm code. The Video project is a bit defferent.
00:13:08bluebrotherIt does work. It's ugly. It misses a status overview. It does not produce diff files (like does).
00:13:17domonokyits more the devloping of a multi-codec able mpegplayer, and ofcourse porting work (optimising too, but probably not as much as for codecs)
00:13:46amiconnfunman: Line 862 reads /* Draw the movie duration */ here
00:13:51xiainxOkay, so there are a couple of improvements to make
00:14:20funmanamiconn: ah it's apps/plugins/mpegplayer/disk_buf.c (there's no plugins/mpegplayer.c !)
00:14:53domonokyRoronoaZoro: do you have any rockboxable mp3player ?
00:15:16RoronoaZoroI had discussion on the irc yesterday about video playback and i am basically starting to understand what is to be done.
00:15:33bluebrothersure. It's a basic version I've put together in just a couple of hours.
00:15:39xiainxbluebrother & domonoky: I have to run right now, but thanks for all your help, I will be back to talk some more about this later! :)
00:16:11domonokyRoronoaZoro: so you will have to buy one for this project :-)
00:16:34RoronoaZoroi was planning to have one by summers
00:17:00funmanamiconn: with lack of comment (in the code) i was about to use #if NUM_CORES > 1, which would enable it only for PP, like it was before I defined PROC_NEEDS_CACHEALIGN for nano2g
00:17:33 Quit pamaury (Quit: Page closed)
00:19:02RoronoaZorodomonoky: you said Video project is bit different can you tell me more as to what can i keep as my goals for summer
00:19:53RoronoaZoroif i decide on it
00:20:04domonokyi mean its different, because the video project is not only optimisations, part of it is also to create a way to have multiple codecs in mpegplayer
00:20:25 Quit xiainx (Ping timeout: 276 seconds)
00:21:14domonokyand porting work of a new video codec. like extracting it from a different project, adapting it to rockboxs interface, maybe make it fixed point, etc
00:21:22amiconnfunman: PROC_NEEDS_CACHEALIGN is only defined for PP anyway
00:21:35funmannot since r25336
00:22:31funmancache alignement is useful for DMA
00:23:40amiconnBut it's not required, or is it?
00:24:36funmanhm not strictly
00:25:23amiconnThat means NEEDS_STORAGE_ALIGN is a bit misleading
00:25:52 Quit GeekShadow (Ping timeout: 265 seconds)
00:26:06RoronoaZoroso for the Video project, what is expected to be known from me .... so that once i know that much i can understand and decide the goals
00:27:05funmanaccording to ata-pp5020.c:ata_dma_setup() reads need to be aligned on 16 bytes, writes on 4 bytes.
00:27:37funmanPerhaps it's nothing to do with cache lines size?
00:27:53 Part toffe82
00:28:04funman /* Require cacheline alignment for reads to prevent interference. */
00:28:04funman /* Require cacheline alignment for reads to prevent interference. */
00:28:33RoronoaZorolike i know about the basic video decoding stuff but have not read much on the codec specific decoding
00:29:02linuxstbRoronoaZoro: I don't think you need to know much in particular. But you'll need to be competent in C, and be motivated to learn about the details of the codecs you want to work with.
00:29:43saratogayou should also have done enough research in advance to propose a reasonable project and time line
00:29:59saratogathat project is particularly open ended since most of us aren't sure which codecs are reasonable to implement
00:30:07funmanr25103 added storage alignement for nano2g but there's no rationale
00:30:49linuxstbRoronoaZoro: What do you find more interesting? Video or audio? (you said you weren't sure which to do).
00:32:31CIA-5New commit by funman (r25338): fix red
00:32:45RoronoaZoroi liked the Video Decoding stuff when i read it...
00:34:08linuxstbRoronoaZoro: But the best way to impress us in your application is to start work on it now - i.e. look at the existing mpegplayer code, try and understand it, and look for potential video-related source code to port to Rockbox.
00:35:18 Quit domonoky (Read error: Connection reset by peer)
00:35:59saratogalooking online there are many vendors selling ARM optimized codecs with performance figures provided, and papers on porting decoders (xvid's, etc) to ARM cores
00:36:15saratogai think some research should clear up whats reasonable
00:36:23saratogathis might also be interesting:
00:39:24 Quit S_a_i_n_t (Ping timeout: 265 seconds)
00:40:17RoronoaZoroi have already started to see different decoders and have about 2 papers and 3 sites
00:41:02RoronoaZoroi have been doing this since yesterday
00:47:08 Quit liar (Ping timeout: 258 seconds)
00:48:58saratogaits probably worth emailing the people who have done arm optimized open source decoders in those papers and ask them if they provide the source
00:49:03saratogaor maybe check their websites first
00:49:16saratogaxvid is GPL IIRC, so they'll probably be willing to give it to you
00:50:08 Quit ender` (Quit: I like work: it fascinates me. I can sit and look at it for hours. -- Jerome K. Jerome)
00:51:44 Join JdGordon [0] (~7bf38c1f@gateway/web/freenode/x-efsqsrikgsjjtruk)
00:54:52 Quit RadicalR (Ping timeout: 265 seconds)
00:55:32RoronoaZoroso i can start on it today itself...
00:56:18linuxstbsaratoga: That paper said they developed the decoder "using Xvid as a reference" - so it sounds like they rewrote it. But yes, it's of course worth asking.
00:58:51 Nick fxb is now known as fxb__ (
00:59:07RoronoaZorore implementation might have been done because it must have been better than porting it .... each possibility can be considered while writing a decoder
00:59:17Tornefunman: looks good, except I would prefer not to call it STORAGE_NEEDS_ALIGN - it's not actually *needed*, it's just desirable for performance. the PROC_NEEDS_CACHEALIGN one is actually mandatory :0
01:02:22funmanthat's what amiconn also suggested, now i'm not sure how to separate PROC_NEEDS_CACHEALIGN & STORAGE_WANTS_ALIGN
01:06:19funmanoh right I know, just make them use CACHEALIGN_BITS regardless PROC_NEEDS_CACHEALIGN
01:09:31 Quit DataGhost (Ping timeout: 240 seconds)
01:11:09 Join S_a_i_n_t [0] (S_a_i_n_t@
01:11:53CIA-5New commit by funman (r25339): Use STORAGE_WANTS_ALIGN to make clear it's not a strict necessity ...
01:11:53 Join xiainx [0] (
01:13:43 Quit komputes (Remote host closed the connection)
01:16:09 Nick animAgus is now known as _silentAssassin (
01:17:29 Nick fidencio[AWAY] is now known as fidencio (
01:18:27funmanI wonder about the green delta for nano2g, there was no red in r25336-7-8
01:18:30***Saving seen data "./dancer.seen"
01:19:20 Quit funman (Quit: free(random());)
01:22:50 Join S_a_i_n_t_ [0] (S_a_i_n_t@
01:24:26 Quit rasher (Ping timeout: 248 seconds)
01:24:29 Join rasher [0] (
01:24:29 Quit rasher (Changing host)
01:24:29 Join rasher [0] (~rasher@rockbox/developer/rasher)
01:24:51 Quit Adubb (Read error: Connection reset by peer)
01:24:58 Quit Bagder (Ping timeout: 248 seconds)
01:25:19 Quit S_a_i_n_t (Ping timeout: 265 seconds)
01:25:30 Join Bagder [0] (~daniel@rockbox/developer/bagder)
01:29:05 Nick fidencio is now known as fidencio[AWAY] (
01:30:00 Join naag [0] (~harish@
01:34:16 Quit naag (Ping timeout: 258 seconds)
01:43:47 Join CaptainKewl [0] (
01:50:38 Quit perfectdrug (Read error: Connection reset by peer)
02:07:52 Join Hillshum [0] (
02:15:07 Join CGL [0] (~CGL@
02:18:22 Quit aholic (Read error: Connection reset by peer)
02:21:12 Quit planetbeing (Quit: planetbeing)
02:22:32 Join planetbeing [0] (
02:24:52 Quit RoronoaZoro (Ping timeout: 258 seconds)
02:30:37 Join RadicalR [0] (
02:35:52 Join RoronoaZoro [0] (~vijayss@
02:45:19 Quit kadoban_ (Remote host closed the connection)
02:45:43 Join kadoban_ [0] (
02:52:32 Join DV [0] (~DV@
02:54:18 Join komputes [0] (~komputes@ubuntu/member/komputes)
03:15:18 Quit kadoban_ (Remote host closed the connection)
03:15:45 Join kadoban_ [0] (
03:18:32***Saving seen data "./dancer.seen"
03:20:15 Quit Darkknight512 (Remote host closed the connection)
03:22:15 Join Lss__ [0] (
03:23:03 Quit Hillshum (Ping timeout: 260 seconds)
03:25:11 Join Hillshum [0] (
03:25:41 Quit Lss (Ping timeout: 268 seconds)
03:25:53 Quit kadoban_ (Read error: Connection reset by peer)
03:26:25 Join kadoban_ [0] (
03:35:26 Join Strife89 [0] (~michael@
03:35:38 Part RoronoaZoro ("Leaving")
03:37:57 Nick S_a_i_n_t_ is now known as S_a_i_n_t (S_a_i_n_t@
03:56:28 Quit kadoban_ (Remote host closed the connection)
03:56:55 Join kadoban_ [0] (
04:13:06 Join sejanis [0] (
04:14:15sejanisanyone tried to format a iriver h10 and install just rockbox?
04:15:38 Join Barahir [0] (
04:18:06 Quit sejanis (Quit: CGI:IRC (EOF))
04:19:12 Quit Barahir_ (Ping timeout: 264 seconds)
04:21:42 Quit TheSeven (Disconnected by services)
04:21:55 Join The_Seven [0] (~theseven@rockbox/developer/TheSeven)
04:22:05 Nick The_Seven is now known as TheSeven (~theseven@rockbox/developer/TheSeven)
04:25:22 Quit SpyBot (Ping timeout: 258 seconds)
04:25:49 Join SpyBot [0] (
04:28:31 Quit komputes (Ping timeout: 260 seconds)
04:31:33 Join JdGordon| [0] (~jonno@
04:43:58 Join Rob2222 [0] (
04:44:20 Join Adubb [0] (~aldubuc@
04:44:23 Quit xiainx (Quit: Good Bye)
04:47:57 Quit Rob2223 (Ping timeout: 265 seconds)
04:50:32flybackwhich h10 model
04:50:41flyback5/6gig or the bigger ones
04:53:53 Quit CaptainKewl (Quit: ( :: NoNameScript 4.22 :: ))
04:59:24 Quit kadoban_ (Remote host closed the connection)
04:59:51 Join kadoban_ [0] (
05:04:26 Quit JdGordon| (Quit: Bye)
05:10:27 Quit Strife89 (Read error: Connection reset by peer)
05:10:56 Join Strife89 [0] (~michael@
05:18:35***Saving seen data "./dancer.seen"
05:22:44 Quit n17ikh (Ping timeout: 265 seconds)
05:28:17 Join n17ikh [0] (
05:29:41 Join planetbeing_ [0] (
05:31:01 Quit planetbeing (Ping timeout: 246 seconds)
05:31:01 Nick planetbeing_ is now known as planetbeing (
05:32:46 Join funman [0] (~fun@rockbox/developer/funman)
05:34:09funmanlinux can see my fuze over usb! (sometimes.. i must miss some initialization) : usb 2-9: device descriptor read/64, error -62 (ETIME)
05:34:37RadicalRThat's nice, I guess.
05:35:16funmanno, v1. rockbox doesn't run on v2
05:35:32RadicalRAh, sorry. I thought you were talking about the OF.
05:35:38RadicalRHence, my question.
05:35:58FlynDiceuh oh, funman's got insomnia again, lookout!!
05:36:04RadicalRCongrats on cracking the clip though. And with that, I'll bow out of this channel.
05:36:11RadicalRSince this is for developers.
05:36:51funmanFlynDice: eh, lack of write support on Clip+ cause me nightmares, please save me !! :)
05:36:57 Quit S_a_i_n_t (Ping timeout: 260 seconds)
05:37:48FlynDicedoes the phrase "gerbil on a wheel" translate to french? ;-)
05:37:54 Join S_a_i_n_t [0] (S_a_i_n_t@
05:38:58funmanyeah, keep running ^^
05:41:33 Quit Horscht (Quit: Verlassend)
05:43:45 Quit panni_ (Quit: ( :: NoNameScript 3.81 :: ))
05:44:46 Quit Adubb (Read error: Connection reset by peer)
05:45:07 Join Adubb [0] (~aldubuc@
05:46:38 Quit bluebrother (Disconnected by services)
05:46:39 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
05:47:07 Quit arbingordon (Ping timeout: 246 seconds)
05:48:50rasherCould someone have a look at to help them out?
05:49:01rasher(getting Rockbox support in udev/hal/whatever)
05:49:40funmanis audio playback stopped on usb connection ? (not when only charging at least)
05:56:13 Quit Strife89 (Read error: Connection reset by peer)
05:59:19ranmafunman: BTW I have seen some codec issues by now. Sometimes it stops playing and the codec thread hogs cpu (100% boost)
06:01:11funmanother threads still work ? (button/backlight)
06:02:02funmando you have high bitrate files ? (flac/ape)
06:02:12 Quit Zarggg_ (Read error: Connection reset by peer)
06:02:26S_a_i_n_tTorne: If you happen to see this...just wondering how your iPod is going so far with the ass-half-of-iram-zero-ing shutdown.
06:02:28ranmamp3 only, may be 320kbps though (didn't check)
06:03:03funmaniirc high bitrates files made playback problems more frequent
06:04:02ranmaWhen I changed UI language during codec hang, the UI hung too until I connected with the debugger, saw that it was waiting for the codec to unload and manually set the variable it was waiting on from 1 to 0 using gdb over jtag.
06:06:23 Join Strife89 [0] (~michael@
06:13:23ranmafunman: Another issue I'd like to debug first are display / keypad glitches.
06:13:55ranmaMy gut feeling is that it's an interaction between both display update and keypad read using DBOP. At first glance the code seems fine though...
06:14:27funmanhm i thought we had cleared that once for all :(
06:14:45funmanwhat's the problem?
06:17:20 Join phanboy4 [0] (
06:18:30ranmafunman: In debug menu show hw info keypad barely responds at all. Sometimes I have to press a key dozens of times before it gets recognized.
06:19:30ranmaAnd I sometimes get short display glitches which looks like either a pixel didn't get sent (whole or part of the screen moves about one pixel left)
06:19:31funmandbop_debug() is only read in I/O ports, not in hw info
06:19:51ranmadbop_debug() only reads the variable anyway :)
06:20:00funmanoh :)
06:20:43ranmaOr like a partial refresh doesn't work and garbage data gets sent ot the lcd
06:22:06funmanbertrik and kugel are your men for dbop
06:23:23funmanmpegplayer would be good to stress the lcd but it doesn't run on 2MB
06:24:09ranmaLike here in the upper part of the screen: ranma/S6001018.AVI">
06:25:57funmanis it really *one* pixel?
06:26:41funmanhm there were some talks about statusbar
06:26:59funmantry disabling it
06:29:01ranmaWell, not sure if it's exactly one pixel, but sometimes the whole screen seems to shift a tiny bit to the left for one update.
06:29:06funmanranma: isn't the freedesktop problem that rockbox uses the mtp id of the c200 OF ?
06:29:10funmanrasher: ^
06:29:36ranmaYeah, looks like it should use the rockbox id as said on
06:29:43 Quit togetic (Quit: WeeChat 0.3.0)
06:29:50 Join planetbeing_ [0] (
06:30:09funmanranma: e200v2/fuze have a delay before reading a button in dbop-as3525.c
06:31:07ranmaYep, I've seen that, but haven't got around to trying it with the delay yet. But today I'm going to do a bit of debugging on that.
06:31:25ranmaBut, first I'm taking a shower :) (Ugh, 14:30 already...)
06:32:23 Quit planetbeing (Ping timeout: 264 seconds)
06:32:24 Nick planetbeing_ is now known as planetbeing (
06:32:39 Join togetic [0] (~togetic@unaffiliated/ibuffy)
06:35:21funmanhm can't tell what happens with USB : sometimes I get an interrupt for both input & output control endpoints, sometimes I get something in dmesg (device descriptor read timeout), sometimes nothing at all. I don't think I forgot to setup a register so I assume it's a timing problem
06:36:39ranmaDo you have a patch I can test?
06:37:24funmansure, let me try one last thing
06:37:25 Nick ranma is now known as ranma|brb (
06:39:16 Quit CGL (Quit: Saliendo)
06:41:07funmanranma|brb: (apply the sansafuze.h diff to sansac200v2.h)
06:42:36rasherfunman: that was fixed. The more general freedesktop problem still exists
06:47:22 Join naag [0] (~harish@
06:58:04 Part naag
07:07:35 Nick ranma|brb is now known as ranma (
07:12:57 Join DerPapst [0] (
07:18:37***Saving seen data "./dancer.seen"
07:21:08funmanranma: /"clear soft disconnect" : should be USB_DEV_CTRL not USB_DEV_CFG
07:24:24ranmaOk, I was just building it. BTW I think turning on UVDD is unecessary as that should be only needed if you want to connect usb peripherals to the sansa...
07:25:34funmanoh right, power comes from the host
07:25:48funmani'm mostly doing the same thing than OF / linux
07:26:21funmanhoping to understand in the process
07:26:26ranmaHmm, I get a reboot loop and sometimes data abort.
07:26:48funmanhow much audiobuffer is left?
07:29:58ranmaIs that recorded somewhere during build? Firmware just reboots directly after the rockbox logo, so...
07:30:36funmanyou can make the diff between audiobufend & audiobuf in
07:30:48funmannot sure if usb would steal from audiobuffer though
07:31:52ranma604396 Bytes
07:32:20funmanOo that's a lot
07:33:44funmanhm I get more than 1MB on normal clipv1 build ..
07:34:19funmanah but debug menu shows ~500kB for buffering thread
07:35:27ranmaHmm, with while(1) in usb_init it still doesn't work. I think it may be beacuse of the PLL thing in system-init that was in the patch.
07:36:21funmantry CGU_COUNTB = 0xff; before ?
07:36:30 Join Zagor [0] (
07:36:30 Quit Zagor (Changing host)
07:36:30 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
07:37:05topikamiconn: RE fuze refreshing: i don't think making the system folder read-only still works. i'll try again though
07:50:33ranmafunman: Hmm, the problem is unrelated to your usb patch, I must have accidentally broken something in my tree here.
07:53:01 Nick ranma is now known as ranma|away (
08:11:05CIA-5New commit by funman (r25340): Clip+: fix battery voltage reading when charging : use ADC_RTCSUP like e200v1
08:21:18 Join DV_ [0] (~DV@
08:22:08 Quit DV (Read error: Connection reset by peer)
08:30:37 Quit DV_ (Ping timeout: 276 seconds)
08:31:37 Join ender` [0] (
08:35:42 Quit Bagder (Remote host closed the connection)
08:36:43 Join Bagder [0] (~daniel@rockbox/developer/bagder)
08:36:55funmanhmm voltage alternate between 75 & 81% in the battery debug screen
08:38:43 Quit DerPapst (Quit: Leaving.)
08:42:19 Quit JdGordon (Changing host)
08:42:19 Join JdGordon [0] (~7bf38c1f@rockbox/developer/JdGordon)
08:49:47 Join mitk [0] (~mitk@
08:50:18 Join DerPapst [0] (
08:52:12 Quit DerPapst (Client Quit)
08:57:08 Quit JdGordon (Quit: Page closed)
08:57:33 Join Luca_S [0] (
09:00:41 Join petur2 [0] (~petur@rockbox/developer/petur)
09:17:06 Join naag [0] (~harish@
09:18:40***Saving seen data "./dancer.seen"
09:19:35 Join flydutch [0] (
09:21:46 Join DerPapst [0] (
09:27:21 Quit phanboy4 (Read error: Connection reset by peer)
09:29:21 Join stripwax [0] (
09:30:28 Part naag
09:31:54 Join naag [0] (~harish@
09:32:03 Part naag
09:32:05 Join liar [0] (
09:33:06 Join naag [0] (~harish@
09:34:26 Quit stripwax (Quit:
09:37:49 Join Professional [0] (~69@unaffiliated/shani)
09:53:58CIA-5New commit by bluebrother (r25341): Fix uic warning.
10:09:19 Quit Professional (Ping timeout: 264 seconds)
10:10:15 Part naag
10:11:50 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
10:13:23 Join swilde [0] (
10:18:24 Join Professional [0] (~69@unaffiliated/shani)
10:20:07 Nick petur2 is now known as petur (~petur@rockbox/developer/petur)
10:37:59 Join TheSeven|Mobile [0] (~theseven@rockbox/developer/TheSeven)
10:38:32 Join B4gder [0] (~daniel@rockbox/developer/bagder)
10:41:16 Quit blairb (Quit: Leaving)
10:45:40 Quit Professional (Ping timeout: 246 seconds)
10:46:00 Join Professional [0] (~69@unaffiliated/shani)
10:48:57 Quit DerPapst (Ping timeout: 265 seconds)
10:50:45 Quit Strife89 (Quit: Going home soon)
10:54:22 Quit Professional (Remote host closed the connection)
10:58:40 Join LinusN [0] (~linus@rockbox/developer/LinusN)
10:59:35 Join wodz [0] (
11:03:05 Join kugel [0] (~kugel@rockbox/developer/kugel)
11:06:53 Join Kitr88 [0] (
11:08:19TorneS_a_i_n_t: works fine so far :)
11:10:09 Quit Kitar|st (Ping timeout: 268 seconds)
11:18:42***Saving seen data "./dancer.seen"
11:28:02 Nick ranma|away is now known as ranma (
11:28:13 Quit RadicalR (Quit: Nettalk6 -
11:39:46funmanRTCSUP looks fine when not charging though
11:40:22 Quit planetbeing (Quit: planetbeing)
11:54:48 Join _Dhiraj [0] (~7371398d@gateway/web/freenode/x-fjeczxzeckeflucc)
11:56:20 Nick fxb__ is now known as fxb (
11:56:28 Quit TheSeven|Mobile (Ping timeout: 260 seconds)
11:57:46wodzehh TBLCF was ok when debugging bootloader. Trying to debug rockbox itself shows how slow it is :-/
12:04:27 Quit funman (Quit: free(random());)
12:17:35 Join jgarvey [0] (
12:25:30 Quit kugel (Ping timeout: 264 seconds)
12:41:17CIA-5New commit by funman (r25342): Fuzev2: fix build, use suited bitmap tool
12:41:34 Quit liar (Quit: brb)
12:43:32 Join xiainx [0] (
12:45:19 Join mt_ [0] (~mtee@
12:46:19 Join liar [0] (
12:46:43 Quit mt (Ping timeout: 258 seconds)
12:46:44 Nick mt_ is now known as mt (~mtee@
12:47:55 Join Zarggg [0] (
12:55:58 Join m3dlg [0] (
12:58:51 Nick fxb is now known as fxb__ (
12:59:07 Quit yawny_ (Remote host closed the connection)
12:59:08 Quit yawny (Remote host closed the connection)
13:05:30 Join panni_ [0] (
13:06:11 Join kugel [0] (~kugel@rockbox/developer/kugel)
13:06:20 Join JohannesSM64 [0] (
13:18:45***Saving seen data "./dancer.seen"
13:22:03 Quit xiainx (Ping timeout: 276 seconds)
13:22:53 Join perfectdrug [0] (
13:25:01 Join anewuser [0] (anewuser@unaffiliated/anewuser)
13:29:28 Join salty-horse [0] (~ori@
13:35:06_DhirajHi! I wanted to work upon the Text to Speech as part of GSOC
13:35:12_DhirajHow should I proceed?
13:36:06_DhirajPlanning o use eSpeak right now
13:39:33gevaertsDo you have a player that runs rockbox?
13:41:28pixelma wasn't there something about licensing with espeak (and Rockbox)?
13:42:15_DhirajI have vlc
13:42:37 Quit anewuser ()
13:42:57_DhirajIs there a licensing issue with espeak?
13:43:08TorneIt is GPLv3, we are GPLv2
13:43:28Torneso including espeak in the rockbox core would change our license to GPLv3
13:44:06_Dhirajthen how about flite?
13:44:11linuxstbOr we use an older version of espeak, before they changed the license.
13:44:45_DhirajBut then the older version wouldn't be supported later?
13:45:13pixelma_Dhiraj: I think gevaerts meant a digital audio player that can run Rockbox, not which program you use on your PC ;)
13:45:14linuxstbWhat do you mean by "supported" ?
13:45:39_DhirajI mean the community wouldn't look to its development, bugs etc.
13:46:01_DhirajSo would have to be completely continued by our communiy I guess
13:46:56Tornehow big a problem that is depends on a lot of things
13:49:02Torneso.. don't rule it out, but something different might be worth considering too
13:50:02_Dhirajpixelma: I was just going through some page which stated vlc with reference to rockbox
13:50:39_Dhirajpixelma: Which audio players are supported?
13:51:50_DhirajI haven't really used Rockbox but have used espeak in a different project
13:54:34pixelmathere's a list on the our project's homepage... ;) ;)
13:57:47 Quit xavieran (Ping timeout: 260 seconds)
13:59:01 Join robin0800 [0] (
14:02:55_Dhirajpixelma: I don't have a player
14:03:03_DhirajHow can I move forward/
14:08:09pamaurygevaerts: [resume yesterday's beginning of conversation] 1) are you working on usb host ? 2) do you still think it might a good idea to switch to blocking/nonblocking version of usb_drv{send,recv} ?
14:12:06pixelma_Dhiraj: maybe you could have a look at some simulators, I could *imagine* that especially the ones of software codec devices (anything but Archos devices) already give a good impression, especially if it comes to the current voice system and playback. I can't tell from a code point of view though
14:13:20_Dhirajpixelma: Ok. Thanks!
14:18:17 Join RoronoaZoro [0] (~vijayss@
14:18:42gevaertspamaury: (a) no. I once started on ohci, but I didn't get further than a set of register definitions (firmware/export/ohci.h). (b) cleaning up the send and recv functions is a good idea I think. The current asymmetry isn't too nice
14:18:51 Part RoronoaZoro
14:22:48 Join xiainx [0] (xiainx@wpa106019.Wireless.McGill.CA)
14:25:54 Quit Luca_S (Quit: CGI:IRC)
14:26:13 Join RoronoaZoro [0] (~vijayss@
14:27:43 Quit xiainx (Ping timeout: 276 seconds)
14:34:25 Join Schmogel [0] (
14:36:30 Join webguest28 [0] (
14:38:59 Part RoronoaZoro ("Leaving")
14:39:28 Quit kugel (Ping timeout: 265 seconds)
14:42:06S_a_i_n_tTorne: Awesome ;D I have high hopes for this...
14:52:51 Join evilnick_B [0] (~0c140464@rockbox/staff/evilnick)
14:53:39_Dhiraj\quit Bye!
14:54:10 Quit _Dhiraj (Quit: bye)
14:55:04 Quit webguest28 (Quit: CGI:IRC (Ping timeout))
14:59:46salty-horseshortcut files don't have an icon
15:00:23S_a_i_n_tsalty-horse: edit your .icons file.
15:01:31salty-horseyou're saying it's in rockbox but not in my installations?
15:01:54S_a_i_n_tno, I'm saying "edit your .icons file"
15:01:58salty-horseI don't see it in filetypes.c
15:02:08S_a_i_n_tIts in the "icons" folder in .rockbox
15:02:26S_a_i_n_ton the DAP, not in the source
15:02:37salty-horseI have such a file?
15:02:38 Nick fidencio[AWAY] is now known as fidencio (
15:02:47S_a_i_n_tyou should do...
15:03:52S_a_i_n_tmake it include "link:X" where X is a number from 0 to 31 (for the main iconset) or 0 to ? (for the viewers iconset you're using.
15:04:46S_a_i_n_tif you're using an icon from the main iconset, it needs to be "link:*X"
15:04:59S_a_i_n_tthe * tells it to use the main iconset.
15:05:22salty-horseI don't have .rockbox/.icons at all. I have a dir called .rockbox/icons/
15:05:51S_a_i_n_t*inside* the "icons" dir?
15:06:43S_a_i_n_ttango_small_viewers.icons to be specific.
15:07:35salty-horsethe only dot files in .rockbox are: .glyphcache, .playlist_control
15:08:12S_a_i_n_tGAH! .icons is the extension! look in .rockbox/icons/tango_small_viewers.icons
15:08:36 Join Luca_S [0] (
15:09:25S_a_i_n_tOnce you open the file, you'll see the syntax is pretty easy and you'll know what you need to do to add more extensions to it.
15:09:28salty-horseoh! but I only have icons/tango_small_viewers.bmp
15:09:29salty-horseno *.icons file
15:09:59S_a_i_n_tThat's weird...its in the source.
15:10:16S_a_i_n_tMaybe it doesn't build with it for some reason.
15:10:17salty-horseS_a_i_n_t, I'd rather like it added to the default cabbie theme so everyone would have it
15:10:35salty-horseI may be using an old version (a month old, or so)
15:10:49S_a_i_n_tthat shouldn't make a difference.
15:11:14S_a_i_n_tJust copy the tango_small_viewers.icons file to the "icons" dir on your DAP
15:11:27S_a_i_n_tthen edit it appropriately.
15:12:18salty-horsebut I don't have such a file anywhere. there's only a sample doc/sample.icons
15:13:08salty-horseS_a_i_n_t, do *you* have icons for shortcut (.link) files with the default cabbie theme?
15:13:29S_a_i_n_tI use my own iconset
15:14:55salty-horsethen I'll file a bug against cabbie
15:15:25S_a_i_n_tThere is already a bug regarding viewer icons.
15:15:31S_a_i_n_tI lodged it myself ;)
15:15:51 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
15:16:21salty-horselink please?
15:17:28S_a_i_n_tFS #10981
15:17:36wodzregarding mpio hd200 in order to incorporate dualboot rockbox bootloader into flash I have to hack original firmware and do firmware update. The process is simple - a) put rockbox bootloader at specified offset into original firmware file b) take first 8 bytes from rockbox bootloader and copy it to the very begining of the firmware file. The process itself is very similar to this used in h100/300 but without any encryption/decryption/checksum generation.
15:17:36wodz So should I write little separate utility do patch of or should I hack mkboot ?
15:18:49***Saving seen data "./dancer.seen"
15:19:39S_a_i_n_tsalty-horse: As you can see from FS #10981, there is problems with a *lot* more than just .link file icons ;)
15:20:43S_a_i_n_t*the ugly grey background in the screendump is to make the weirdness of the 'strange black squares instead of icons' stand out better.
15:21:54salty-horsebut what's that got to do with "Cabbie is missing a .link icon"? :)
15:22:10Tornebecause cabbie doesn't have explicitly specified viewer icons
15:22:20S_a_i_n_tall themes use that file.
15:22:37 Quit wodz (Quit: Leaving)
15:27:53salty-horseso my issue is covered by that bug? I just want to know if I need to file a separate one or not :)
15:28:23S_a_i_n_tsalty-horse: FS #10981 covers your bug, yes.
15:29:10S_a_i_n_tNo way of telling when it'll be looked at however.
15:31:51salty-horsethen I'll add a comment, so that issue won't get lost
15:32:11S_a_i_n_tIt won't get lost...its on the tracker.
15:32:28S_a_i_n_tThat's what the tracker is for ;)
15:32:47salty-horseI meant my mini-issue, which you say may not be related to the bug you filed
15:33:15S_a_i_n_tI didn;t say it "may not be related"...its directly related to it.
15:33:40S_a_i_n_tDisplay of icons in general is slightly broken.
15:35:00LloreanDo any themes even use the file types colors functionality? Or does that even still exist?
15:36:33S_a_i_n_tI'm not sure of themes are allowed to include colours for filetypes anymore..I think it was decided that should be up to the user to set.
15:37:24LloreanIt's something that has to be set by editing a text file
15:37:53LloreanIt doesn't really seem like a user setting.
15:38:10S_a_i_n_tIt can be set on-device
15:38:20S_a_i_n_tQuite easily in fact.
15:39:31S_a_i_n_tI forget exactly how (I don;t use filetype colours) but its definitely in the manual.
15:39:43salty-horseS_a_i_n_t, sorry. I misread your "No way of telling when it'll be looked at however." as "No way of telling until it'll be looked at"
15:39:51domonokyfiletype colors is a allowed theme setting.
15:39:58pixelmaI don't think that's easily settable through the menu
15:40:45S_a_i_n_tThere's something in the manual about it, I think the text editor understands the syntax of the .colours file..and does some magic
15:40:51S_a_i_n_t...something like that.
15:41:13pixelmaexcept loading a colours file (or however it's called). I can't remember seeing a theme that uses it
15:41:32LloreanThe manually explicitly says ".colours file belonging to a theme .cfg"
15:41:42LloreanSo the manual at least suggests they're to be treated as part of a theme
15:42:07pixelmabut IMO that could be something a theme can include - in contrast to e.g. the language setting
15:42:13LloreanIt's also only mentioned in one line in the text editor section of the manual
15:42:16 Join kugel [0] (~kugel@rockbox/developer/kugel)
15:42:27LloreanBasically, nobody probably even knows this feature exists.
15:42:47S_a_i_n_tI think very few people use it.
15:42:52LloreanBut considering colours need to be compatible with the backdrop image used, they should always be either reset to default (use foreground color) or explicitly set by themes
15:42:54LloreanJust like everything else
15:43:01S_a_i_n_tprobably for that reason, or maybe not.
15:43:31 Join toffe82 [0] (~chatzilla@
15:44:43S_a_i_n_tI think there is quite a few features rockbox has that people would be surprised existed to be honest...
15:45:02 Quit mitk (Quit: Leaving)
15:45:07S_a_i_n_tWhen *most* people read the manual, its probably only to find out one specific piece of info.
15:45:21S_a_i_n_tAs opposed to actually "reading the manual"
15:45:51LloreanWell, this is a core feature that's only briefly mentioned in a plugin description
15:45:58LloreanAnd not demonstrated in the default theme
15:46:18S_a_i_n_tI knew I saw it *somewhere* though ;)
15:46:39*Llorean thinks it might be nice if, for example, we slightly used filetype colors by making music/supported files one shade of gray brighter than other files in the default theme.
15:46:52linuxstbpixelma: Speaking of the manual, are you able to produce a Clip+ image for it? Does an SVG already exist somewhere (I noticed the different Clip+ image on the download pages).
15:47:03S_a_i_n_tI do seem to remember a conversation along the lines of "it should be up to the user to set filetype colours...*if* they want to"
15:47:31S_a_i_n_tBut you're right, they should be more *aware* of it as a feature.
15:47:32pixelmalinuxstb: I believe Zagor committed a Clip+ small png from an svg in the patch tracker
15:47:55LloreanS_a_i_n_t: By that very same logic, it should be up to them to set the foreground
15:48:00LloreanYet themes must set it, to prevent unreadability.
15:48:01S_a_i_n_ts/they/people - users/
15:48:27S_a_i_n_tI think themers just set colours that work to begin with.
15:48:35*linuxstb thinks things would be much simpler if we didn't give users all this choice about independently setting different theme-related elements...
15:48:37S_a_i_n_tWell, that's what I do.
15:48:58 Quit kugel (Remote host closed the connection)
15:49:03Lloreanlinuxstb: Agreed, honestly
15:49:08 Join kugel [0] (~kugel@rockbox/developer/kugel)
15:49:41salty-horsecan anyone confirm this on other devices?
15:50:10S_a_i_n_tLike, dark text on a dark background won't work for any text...not just specific filetypes...So themers set a decent contrast between the two
15:50:32Lloreanlinuxstb: At the very least though, themes on the theme sight should basically "reset" to default all theme settings they don't explicitly set a value for of their own
15:50:40LloreanThat way themes are more or less always applied to a known state.
15:50:55LloreanFiletype colours, if not required by themes now, means there's an unknown state in the visual appearance that can affect usability
15:50:58pixelmathere's a theme sight? ;)
15:50:58S_a_i_n_tPersonally, none of my themes ever include a .colours file as I think thats up to the user to set to their preference, not mine.
15:51:07linuxstbWouldn't it just be easier to move all theme-related stuff out of the main .cfg file, and have a ".theme" file storing it all?
15:51:07LloreanI think it's still mostly okay if users set such things after the theme is loaded.
15:51:19Lloreanlinuxstb: It might help.
15:51:45linuxstbThat way, you don't have to worry about defaults - loading a .theme will always reset everything to defaults.
15:52:36LloreanS_a_i_n_t: But here's the thing. If Theme A sets filetype colors, and Theme B just assumes they're default. Theme A sets them dark, Theme B *is* dark, so dark colors are hard/impossible to see... what happens if you load Theme A then Theme B
15:53:00pixelmathere is a "save theme" which will be a cfg though in the end
15:53:07LloreanIf a user wants custom colors, they'll have a .colours file around defining them. So they can load their theme then load the .colours file, simple enough
15:53:59S_a_i_n_tWouldn;t it be easier to default those values while changing themes if its not set by the theme?
15:54:01linuxstbpixelma: I think treating themes like a normal .cfg is the complication. If we move away from that, I imagine things being much simpler. (perhaps less powerful, but that may be a good thing...)
15:54:52LloreanS_a_i_n_t: There's not really a system for "default if not set by a file" plus, since a theme is just a .cfg, any time *any* .cfg file loaded without a .colours set would then default it
15:55:06LloreanS_a_i_n_t: The easiest would be for theme authors to explicitly set colours, even if what they explicitly set is "no colours"
15:55:38linuxstbLlorean: And then we get a new theme feature, which older themes don't set to a default, and more complications...
15:55:39S_a_i_n_tAll my themes include "filetype colours: -"
15:57:51LloreanS_a_i_n_t: And that's exactly what I was saying they should do.
15:57:55LloreanYou set it.
15:58:16Lloreanlinuxstb: Yeah. I honestly don't know why we don't have a separate .theme file. I'm sure it's come up in the past.
15:58:18S_a_i_n_tYes, well...kinda. I set it to "not set"
15:58:38LloreanThere's no "kinda" about it.
15:58:39S_a_i_n_tLlorean: I believe that conversation has come up, yes.
15:58:40LloreanYou set it to a value
15:58:44LloreanThe value is "default"
15:59:03pixelmaI'd be opposed to e.g. losing the ability to set colours easily on the device, as they can look so much different on different devices and compared to the sim
15:59:26pixelmaI guess a plugin could help there though
15:59:36LloreanI'd say a theme editor plugin would not be amis.
15:59:45S_a_i_n_tIsn't a theme plugin in the works somewhere?
15:59:51linuxstbpixelma: I don't think we would have to lose the settings - they could just get saved to the current ".theme" file.
15:59:58LloreanJust the existing setting menus, plus perhaps a "preview" button that shows you a sample of the selected .wps + colors in list, so you could see if it's readable/usable
16:00:19Lloreanlinuxstb: Well, moving the settings away from the user might be beneficial too, you said
16:00:55pixelmalinuxstb: I thought that#s what your "less choice" boils down to
16:01:28*linuxstb isn't sure what he meant any more then... ;)
16:02:39 Quit JohannesSM64 (Quit: WeeChat 0.3.2-dev)
16:02:40pixelmaS_a_i_n_t: I haven't heard about someone working on a theme editor plugin
16:03:16S_a_i_n_tI'm fairly sure its been talked about, perhaps somewhere in GSoC discussion...
16:03:28S_a_i_n_tI'm sure I remember seeing *something* about it.
16:03:46linuxstbI think it's the fact that if theme settings are part of a .cfg, you can't clear all values to a default before loading it. So all I think I'm suggesting is to explictly store all theme settings in a .theme file, and ignore them if they appear in a .cfg The .cfg would just reference a .theme
16:04:28LloreanThat seems like a good starting point at least, since nothing is lost there.
16:04:35LloreanS_a_i_n_t: Maybe you're thinking of the theme editor PC app
16:04:37pixelmaso that would mean loading two files?
16:05:33S_a_i_n_tLlorean: Hmmmm, possibly, I'm sure I saw mention of an "on-device" editor. But I could be wrong.
16:05:35LloreanWell, you load the .cfg, then it loads the .wps, the .sbs, the .rwps, and the .colours
16:05:39linuxstbpixelma: Yes. Although we already load many files (main .cfg, .wps, .rwps etc etc)
16:05:40S_a_i_n_tIt would definitle be handy.
16:05:41LloreanAdding the .theme to the end of that list isn't a big deal
16:05:52 Nick fidencio is now known as fidencio[AWAY] (
16:12:21 Join hebz0rl [0] (
16:13:02 Quit toffe82 (Read error: Connection reset by peer)
16:13:45 Quit jgarvey (Quit: Leaving)
16:14:04 Join jgarvey [0] (
16:34:05 Quit leavittx (Read error: Operation timed out)
16:34:34 Join fyrestorm [0] (
16:35:56 Join leavittx [0] (~leavittx@
16:39:23 Quit hebz0rl (Quit: Ex-Chat)
16:41:56 Join hebz0rl [0] (
16:53:54 Quit petur (Quit: connection reset by beer!)
16:58:38 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
16:58:47perfectdruglinuxstb: the flyspray task for the clip+ is here everything needed for the manual should be there
16:59:41 Join phanboy4 [0] (
17:00:26 Quit bluebroth3r (Read error: Operation timed out)
17:02:05linuxstbperfectdrug: Thanks. So I just need to drop those files into the write place, and that's it?
17:02:30perfectdrugi hope so, you may have to check the naming though
17:02:32 Join komputes [0] (~komputes@ubuntu/member/komputes)
17:03:30linuxstbperfectdrug: OK, I'll try and do that this evening, if no-one gets there first.
17:04:58perfectdruglinuxstb: fine:) i did my best to provide all files necessary, hope everything is right
17:08:05 Quit kaniini (Remote host closed the connection)
17:11:07 Join perfectdrug1 [0] (
17:11:08 Quit perfectdrug (Read error: Connection reset by peer)
17:17:43 Quit FlynDice (Remote host closed the connection)
17:18:06 Quit pamaury (Quit: Quitte)
17:18:50***Saving seen data "./dancer.seen"
17:21:29 Join FlynDice [0] (
17:27:00 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
17:29:30 Quit robin0800 (Remote host closed the connection)
17:38:52 Quit ender` (Read error: Connection reset by peer)
17:39:26 Join ender` [0] (
17:40:32 Nick fidencio[AWAY] is now known as fidencio (
17:42:28 Quit B4gder (Quit: It is time to say moo)
17:42:53 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
17:43:58 Quit Guest3190 (Ping timeout: 258 seconds)
17:49:15 Quit Luca_S (Quit: CGI:IRC)
18:00:29CIA-5New commit by FlynDice (r25343): sd-as3525v2.c Organize construction of MCI_COMMAND so it is more apparent which bits are being set and why. ...
18:11:38CIA-5New commit by dave (r25344): Add Clip+ player image from FS #11048 by Marko Pahlke
18:12:16linuxstbperfectdrug1: Your images seemed fine - they're now committed, thanks.
18:19:35 Join mitk [0] (
18:20:08 Join CGL [0] (~CGL@
18:21:18perfectdrug1linuxstb: awesome:)
18:22:47linuxstbperfectdrug1: I assume you're already in the CREDITS ? (I didn't check)
18:23:10*linuxstb sees that perfectdrug1 is
18:23:11perfectdrug1yes I am thanks
18:23:52 Join feisar [0] (
18:24:20 Nick feisar is now known as Guest6244 (
18:26:27CIA-5New commit by dave (r25345): Update copyright year to 2010
18:26:55 Quit Hillshum (Ping timeout: 264 seconds)
18:29:29 Join pamaury [0] (~c2c7a50a@rockbox/developer/pamaury)
18:31:48 Join phanboy_iv [0] (
18:33:51 Quit phanboy4 (Ping timeout: 265 seconds)
18:35:42 Part LinusN
18:36:51 Quit GHF (Read error: Connection reset by peer)
18:37:37 Join GHF [0] (~meow@unaffiliated/ghf)
18:41:15 Quit phanboy_iv (Ping timeout: 245 seconds)
18:42:16 Quit _silentAssassin (Ping timeout: 264 seconds)
18:49:26 Join Edwerd [0] (
18:49:40Edwerdi have a quastion about clipV2
18:49:49Edwerdi cant update the database
18:49:59Edwerdbecuase it aint developed yet
18:50:09Edwerdthat it can write files to the system
18:50:13Edwerd"read only"
18:50:25Edwerdor i have a specific problem with my clipV2
18:51:47gevaertsAs far as I know rockbox doesn't have write support yet on the clip v2
18:52:05Edwerdbecause of that i cant update the database ?
18:52:15Edwerdor create or whatever ?
18:52:45Edwerdis there any way i can create the database through my pc ?
18:52:51Edwerdand upload it to the player ?
18:52:56kugel yes
18:53:38Edwerdhow ?
18:54:06Edwerdi will be very much obliged if you could help me with that :)
18:54:26 Quit jgarvey (Quit: Leaving)
18:56:36 Quit liar (Ping timeout: 258 seconds)
18:58:21Edwerdany one can instruct me on how to create a database file on my pc ?
19:00:58kugelEdwerd: you can use the database tool which you can build from svn
19:01:17kugelno time to guide you through it now though
19:02:57Edwerdcan you give me a link for the tool ?
19:03:40 Quit GHF (Ping timeout: 246 seconds)
19:04:41LloreanYou'll need to compile it, from the Rockbox source code.
19:04:45 Join GHF [0] (~meow@unaffiliated/ghf)
19:06:16 Quit kugel (Ping timeout: 264 seconds)
19:06:57Edwerdwhere do i put the data base files
19:07:42Edwerdin witch directory in .rockbox
19:08:52 Join adnyxo [0] (
19:12:17LloreanJust in .rockbox
19:14:34Edwerdhe wont turn on
19:14:40Edwerdwont even charge
19:14:46Edwerdwhat could have caused this ?
19:15:06 Join Davide-NYC [0] (
19:16:09LloreanWhat did you do?
19:16:12Edwerdit worked moments ago
19:16:25Edwerdi've installed rockbox
19:16:32Edwerdthe latest build for clipv2
19:16:37Edwerdand it worked just fine
19:16:46Edwerdi've used some software for sourceforge
19:16:52Edwerdto create the database files
19:17:06Edwerdand put em on the .rockbox directory
19:17:12Edwerdthen it updated the files
19:17:17Edwerdand it wont turn on
19:17:21Edwerdsince then
19:17:32Davide-NYCIs it off topic to talk about "Rockbox as an Application" in this channel?
19:17:42LloreanEdwerd: Please, don't use the enter key as punctuation. Try to write clear and complete statements
19:17:46LloreanDavide-NYC: It's Rockbox...
19:17:57FlynDiceEdwerd: hold the power button to off for 15 secs
19:18:12Davide-NYCHas there been any discussion about attempting to port some form of RB to Palm's WebOS?
19:18:31Davide-NYCIt is a linux based, SDL ready platform that is very open to developers.
19:18:54***Saving seen data "./dancer.seen"
19:19:03FlynDiceEdwerd: when you start it up again press the << button and you should get to OF
19:19:11Davide-NYCI have had the phone for almost 9 months and absolutely love it.
19:19:20Edwerdthe 15 seconds solution worked for me
19:19:26Edwerdand the database works aswell
19:19:27gevaertsDavide-NYC: most people here are of the opinion that the exact target is a relatively minor portion of the total work, so we don't care too much about it
19:19:30Edwerdthank you guys :)
19:19:46LloreanDavide-NYC: Really, the first step should almost certainly be porting it to a full PC app.
19:19:56LloreanOnce that's done, target-specific ports for a lot of things become somewhat easier.
19:20:36Davide-NYCgevaerts: Llorean: I see.
19:21:11Davide-NYCHas there been any work in this direction recently?
19:21:17 Join _silentAssassin [0] (
19:21:21evilnick_BDavide-NYC: It's a GSOC project
19:21:31gevaertsa project idea anyway :)
19:21:31Davide-NYCWaddap Nick.
19:21:32evilnick_Bs/project/project suggestion/
19:22:40Davide-NYCFWIW, the game community (gameloft etc) claims that porting to WebOS is particularly easy. It may be worth taking a quick look. How is the simulator not a PC app?
19:23:51LloreanIt's a PC app
19:23:54gevaertsDavide-NYC: it is a PC app, but it's full of hacks
19:23:54LloreanIt's not "Rockbox as an App"
19:25:00Davide-NYCThen it's just "The Rockbox App"
19:25:08gevaertsno. It's a hack :)
19:28:16 Join Xerion_ [0] (
19:28:28 Quit Xerion (Ping timeout: 264 seconds)
19:28:34 Nick Xerion_ is now known as Xerion (
19:28:56 Quit Davide-NYC (Quit: ChatZilla 0.9.86 [Firefox 3.6.2/20100316074819])
19:30:10 Quit swilde (Quit: ERC Version 5.3 (IRC client for Emacs))
19:30:29Edwerdlol i've created the database and says some error with a playlist writing or something
19:30:48Edwerdbut i've managed to creat the database manually
19:33:52 Quit flydutch (Quit: /* empty */)
19:45:14 Join kugel [0] (~kugel@rockbox/developer/kugel)
19:46:13 Join xiainx [0] (
19:46:36 Join planetbeing [0] (
19:46:46 Quit planetbeing (Client Quit)
19:49:56 Join planetbeing [0] (
19:55:22 Join S_a_i_n_t_ [0] (S_a_i_n_t@
19:55:32 Quit S_a_i_n_t (Ping timeout: 265 seconds)
20:01:43LloreanIs there some reason rbutil doesn't try to download OF images when they're necessary?
20:07:59domonokycopyright fears, and also the OF Urls might change.
20:08:25LloreanCould it at least provide the last known link?
20:08:40Llorean"please click this link, save the file in the same folder as rbutil, and click OK to continue" or something
20:08:50linuxstbWe could have the OF URLs on, with a script to test them nightly.
20:09:07 Quit kugel (Ping timeout: 246 seconds)
20:09:10LloreanRight now, for players that require an OF image people often need to know to look in the manual install section.
20:09:21LloreanWhich, by the way, is a URL that might change. ;)
20:11:30 Join kugel [0] (~kugel@rockbox/developer/kugel)
20:22:50domonokyrbutil shows a message with the neccessary urls (sometimes pointing to the wiki=
20:24:43 Join liar [0] (
20:26:11 Join shivkrish [0] (
20:29:50 Join jgarvey [0] (
20:32:20 Quit Schmogel (Quit: Miranda IM! Smaller, Faster, Easier.
20:33:10 Join BHSPitLappy [0] (~BHSPitLap@unaffiliated/bhspitmonkey)
20:34:11 Quit BHSPitLappy (Read error: Connection reset by peer)
20:34:29 Join BHSPitMini [0] (
20:35:19 Quit jgarvey (Quit: Leaving)
20:37:59 Join efyx [0] (
20:39:26 Join petur [0] (~petur@rockbox/developer/petur)
20:53:47 Quit BHSPitMini (Ping timeout: 276 seconds)
20:56:42 Join intrados [0] (
20:58:25 Quit moos (Ping timeout: 246 seconds)
20:58:31 Quit intrados (Client Quit)
21:03:03 Join Darkknight512 [0] (
21:12:25 Quit Xerion (Quit: )
21:13:30 Join moos [0] (moos@rockbox/staff/moos)
21:18:17 Quit hebz0rl (Quit: Ex-Chat)
21:18:57***Saving seen data "./dancer.seen"
21:18:58 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
21:18:58 Quit pixelma (Disconnected by services)
21:19:11 Quit amiconn (Disconnected by services)
21:19:14 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
21:19:17 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
21:19:36 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
21:23:26 Quit fyrestorm (Quit: Ur skills' fireproof like a wooden panel -- U got feds talking leet on your IRC channel!)
21:24:21 Quit Battousai (Read error: Operation timed out)
21:25:26 Join buz_ [0] (
21:25:37 Join Xerion [0] (
21:27:24buz_i am experimenting with rockbox on a clip+. works beautifully so far. only thing I cannot quite get to work is the database. its sitting there for 5min now saying it found 0...
21:28:55 Join Battousai [0] (~bryan@gentoo/developer/battousai)
21:29:21 Join Hans-Martin [0] (
21:30:35 Join froggyman [0] (~me@unaffiliated/froggyman)
21:31:00 Join planetbeing_ [0] (
21:32:13Hans-Martinhi folks, after my old PNA went belly-up (GPS won't work anymore) I would like to use it for the rest of its lifetime as a rockbox media player. I know there have been efforts in teh past to run rockbox as an app on WM5+6 (by saratoga), but there's nothing on the web site.
21:32:29Hans-MartinI could invest some time into this, but I'd prefer not to start from scratch.
21:33:01Hans-MartinI have some experience with cross-compiling for CE (aka WM) using the MinGW compiler
21:33:02 Quit planetbeing (Ping timeout: 240 seconds)
21:33:03 Nick planetbeing_ is now known as planetbeing (
21:33:46LloreanHans-Martin: The simulators already run as SDL apps. Assuming there's SDL for WinMo, adapting one of them to your purposes might be a starting point, though ideally there's a lot of things that would need to work differently for the ideal rockbox as an app
21:34:45kugelbuz_: there's no write support yet, hence you cannot create the database on target
21:34:59*kugel thinks rockbox should handle no-write-support better
21:39:05Hans-Martinhm, I think I take a look at it.
21:41:34 Join robin0800 [0] (
21:47:51 Quit Zagor (Quit: Clint excited)
21:48:26 Join Halborr [0] (
21:49:06 Quit xiainx (Ping timeout: 264 seconds)
21:50:49 Quit moos (Ping timeout: 245 seconds)
21:50:57 Join xiainx [0] (xiainx@wpa106060.Wireless.McGill.CA)
21:52:29 Join moos [0] (moos@rockbox/staff/moos)
21:57:22 Quit robin0800 (Remote host closed the connection)
21:57:37 Quit evilnick_B (Quit: Page closed)
22:05:38 Join Strife89 [0] (
22:08:34Halborrhey does anyone know what the notation (1<<4) means?
22:09:16gevaertswell, it's standard C...
22:10:39Halborrsorry if it's a stupid question
22:10:46domonokyHalborr: its a one shifted by four bits :-)
22:11:04domonoky*shifted left*
22:11:27Halborrthank you *to google!*
22:11:42gevaertsWell, this isn't really a C support channel
22:12:36Halborryeah, sorry. I was just looking at the rockbox code and wondered what that meant. I've got a long-term project of understanding at least part of the code in rockbox (in order to contribute later)
22:16:56gevaertsHalborr: I'd recommend either ##c, #rockbox-community, or a good reference manual. We really want to keep things on-topic in #rockbox
22:18:00HalborrWill do. I figured it was okay since there wasn't much (any?) traffic.
22:18:44LloreanHalborr: The channel's logged, so we discourage unrelated traffic independent of time so that there's less noise in the logs for people reading through them to see if they've missed anything important.
22:19:21 Nick fidencio is now known as fidencio[AWAY] (
22:25:34 Join webguest11 [0] (
22:27:55 Quit webguest11 (Client Quit)
22:28:12 Join webguest87 [0] (
22:29:20 Quit webguest87 (Client Quit)
22:29:31buz_kugel: ah that makes sense. maybe giving an error would be useful, though. also, failing a db, a feature to just play all songs in shuffle would be a big step for me (i mostly use the clip in shuffle mode all)
22:29:57kugelyou can d that, can't you?
22:30:21buz_not sure, have not figured out how to say the least :)
22:30:38LloreanIf you have a music folder, just "insert shuffled" it.
22:30:41LloreanIt's pretty easy.
22:31:03buz_ok, ill give that a shot
22:31:14FlynDicebuz_: We're paddling as fast as we can, please be patient...
22:31:30 Join Alexandru_Criste [0] (acr@
22:31:30buz_otherwise i like what i hear, maybe placebo but it seems to have better soundquality
22:33:19FlynDicebuz_: I believe you are correct though, you cannot shuffle between multiple directories as it stands now. I beleive write support will remedy this.
22:33:51buz_i may just drop all files into one directory then :)
22:34:03FlynDiceI hink that would work
22:34:09FlynDicethink even
22:34:29LloreanFlynDice: If you just have all your music organized like \Music\Artist\Album\Song.mp3 (note the top single folder) you can just insert the single folder.
22:35:02LloreanIf you have it all as individual folders in the root (which is quite cluttered anyway), yes you'll need to insert every folder individually, or use the "create playlist" option, and then shuffle that playlist (which is still quite quick)
22:35:21buz_so recursing down trees works? even better then
22:35:28Torneyes, recursing is on by default
22:35:36buz_thanks guys
22:36:09buz_the clip was in bad need of a better eq - any place to send a few beers to?
22:37:14saratogaideally the database and other features should be conditional on write support working
22:38:05Tornebuz_: you can donate to the project through the paypal button on
22:38:18LloreanHow many features really require write? Database and dircache and?
22:38:22buz_probably will
22:39:41LloreanDoes resume present any weird behaviour or does it just say "nothing to resume"?
22:39:46pixelmasaving settings
22:39:56 Quit pamaury (Quit: Page closed)
22:40:53LloreanI don't know if you could readily disable settings, though. ;)
22:41:05pixelmaI guess the weird behaviour there is when you stop playback. You'll get the "can't write playlist.control (or however it is called exactly)" if I remembercorrectly
22:41:48LloreanMaybe the message could be clarified to "unable to write resume data" or something?
22:41:52LloreanSomething more user understandable?
22:41:56FlynDiceLlorean: You probobly understand this better than me I can only go on what I observe. On my clip+ right now I cannot insert an entire directory and have the player play music. I can do that on my e200v2 and fuze. I am guessing write support will enable this.
22:42:27FlynDiceI can play from a directory in shuffle mode though
22:42:33Alexandru_Cristein the GSoC project list I read Rockbox as an Application and I would like to know more about the project
22:42:36LloreanThat's very strange. And no, I'd have no clue why. :)
22:42:44Alexandru_Cristecould anyone point me in any direction?
22:42:47LloreanAlexandru_Criste: What questions did you have?
22:43:17Alexandru_Cristewhat is the main goal of the project?
22:43:31linuxstbTo run Rockbox as an application...
22:43:33saratogaAlexandru_Criste: the goal is to get rockbox running as an application on another embedded operating system
22:44:05Alexandru_Cristeand this would be acomplished by porting the Rockbox code?
22:45:02*domonoky thinks the biggest work for this project, would be to change the sdl simulator to be more similar to real targets, and adapt it to a specific embedded os (android for example)
22:45:06LloreanAlexandru_Criste: Right now Rockbox makes a lot of assumptions based on the fact that it is also the operating system. As well, has provisions for its plugin architecture and memory use and file access that don't necessarily make the same sense when run as an application within another OS
22:45:26CIA-5New commit by alle (r25346): Remove the unneeded appendix declaration. It's already present in appendix.tex and does not make any sense here (=at the end) anyway.
22:45:38saratogadomonoky: imo getting the sim to compile in the target tree isn't really that hard, but getting it to compile and also run in a useful way is a LOT of work :)
22:45:54pixelmaFlynDice: I think inserting and other playlist manipulation updates the playlist control file which it can't without write support
22:46:06linuxstbdomonoky: I wouldn't describe it like that. I would describe it as separating Rockbox's application part more from its kernel, so that it can run on other kernels.
22:46:19Alexandru_Cristethe Rockbox code is written in C/C++
22:46:25LloreanNot C++
22:46:29domonokyC and Asm :-)
22:46:50saratogaAlexandru_Criste: right now rockbox is an operating system that only runs one program (which is also rockbox), eventually it would be nice to be able to run it on other operating systems as a program
22:46:56gevaertsAsm won't be an issue for this project though
22:46:58FlynDicepixelma: Yes, that's what I think is going on but I haven't tried to dig into the code...
22:47:37Alexandru_Cristefrom what I read the preffered platform is Android
22:47:41linuxstbdomonoky: i.e. creating a clearer API between apps/ and firmware/, and then implementing the firmware/ part as a wrapper for different target OSes.
22:47:55Alexandru_CristeI've been working on Android for some time
22:47:56domonokylinuxstb: sounds good :-)
22:48:03LloreanAlexandru_Criste: There's not too much of a "preferred" platform. It's just one of the most obvious/significant candidates.
22:48:14Alexandru_CristeI didn't had the chance working on NDK
22:48:18Alexandru_Cristebut I read about it
22:48:32saratogawe don't really have a prefered platform, but android is probably one of the easiest to work on
22:48:35gevaertsOne other problem with the simulator is that several bits of it are basically hacked together, like e.g. the threading support
22:49:04Alexandru_Cristeand I am not so sure that Android OS let's you use the hardware directly
22:49:10saratogaimo even things like threading aren't so important for this project, the goal is really to get something working so that people can use it and see how to improve it
22:49:37saratogaright now its impossible for another person to work on rockbox as an app without also being a core rockbox developer since we haven't yet really decided how rockbox as an app should work
22:50:06BagderAlexandru_Criste: android lets you do things in plain C using the ndk, if you want to
22:50:22Alexandru_Cristeyes,but there are limitations
22:50:30Bagdernot really, no
22:50:30Alexandru_Cristeyou have the player already implemented
22:50:55gevaertsAlexandru_Criste: you need (a) access to the screen, (b) the ability to output sound, and (c) the ability to read buttons and the touchscreen
22:51:06gevaertsI'm assuming that those are possible
22:51:10saratogaafaik all of those things can be done from c code, using sdl on android
22:51:14Alexandru_Cristeyou have all those
22:51:20Alexandru_Cristebut on a higher level
22:51:32Alexandru_CristeI tried to modify the MediaRecorder
22:51:33domonokygevaerts: + d) storage access :-)
22:51:39Bagderthe highest level may involve some java stuff on androd
22:51:43gevaertsWell yes, this is where "as an application" comes in
22:51:48Alexandru_Cristeto enable RTSP streaming
22:51:52linuxstbBut for Android, wouldn't a better approach be to write the higher-level code in Java, and just use Rockbox's playback engine and codecs? (hasn't someone even done that already?)
22:51:52Alexandru_Cristeand got to an dead end
22:51:54saratogawe don't stream anything
22:52:17saratogalinuxstb: ideally the sdl people already did that and we can start with their code and maybe go from there later . . .
22:52:22Alexandru_Cristeno no
22:52:23Alexandru_CristeI know
22:52:24BagderAlexandru_Criste: that's not a proof that it really is a dead end though...
22:52:26 Join stripwax [0] (
22:52:31Alexandru_CristeI was just talking about the posibilities
22:52:46domonokylinuxstb: i think rewriting the gui in java, is a too big project. java warppers for drawing etc should be enough.
22:52:48saratogai don't think this project involves adding any new features to rockbox playback
22:52:59saratogait should just get the current ones working
22:53:01gevaertslinuxstb: that might be useful later on, to be done after the rest actually works :)
22:53:08saratogafor GSOC the network stack might as well not exist in android
22:53:29BagderI agree, let's get the playing native music done first
22:53:30 Join kaniini [0] (
22:53:39Bagderor local perhaps
22:53:46saratogaonce you have a proof of concept port, i'm sure other people will take interest
22:53:52Alexandru_Cristeto use the sdk/ndk
22:53:57Alexandru_Cristeto create a player
22:54:11Alexandru_Cristeit's not very hard
22:54:32saratogai would phrase it more as using the sdk/ndk/sdl to get an existing player working on android
22:54:35Bagderit's not about "to create a player"
22:54:39saratogaone with very complicated requirements
22:54:43 Join S_a_i_n_t [0] (S_a_i_n_t@
22:54:57linuxstbBut I think even a project that gets Rockbox running as a real application on a Linux desktop (in a clean way that allows more ports) would be a successful summer.
22:55:13saratogai think once we had something usable on android others would take interest, like that guy who wrote the simple android player using our codecs
22:55:19 Quit komputes (Quit: I haven't slept for ten days, because that would be too long.)
22:55:20gevaertsThe hard bit shouldn't be the android/maemo/something specific bits, it's getting the rockbox code itself in a good shape to run as an application, i.e. with an OS running under it
22:55:23saratogai'm sure he'd be very intertested if a rockbox port existed
22:55:56 Quit CGL (Quit: Saliendo)
22:55:56saratogagevaerts: well i have an ancient patch for doing that now, although its a bit ugly
22:57:25 Quit S_a_i_n_t_ (Ping timeout: 268 seconds)
22:58:11linuxstbAlexandru_Criste: I think the important thing is that this shouldn't be thought of as an Android project - it's a project to work with the Rockbox code to separate the application part of it in a clean way, and then perhaps port it to Android as a secondary task.
22:59:31Alexandru_Cristehow can I access the repository?
23:00:32saratogalinuxstb: i was able to compile rockbox for linux/sdl under the target tree years ago
23:00:47saratogaimo that should be less then the primary objective or this project is not much use
23:00:49linuxstbsaratoga: Yes, but there's a lot more to it than that.
23:01:11gevaertssaratoga: there's more than compiling under linux/sdl in the generic part as far as I can see
23:01:15saratogayes but if the primary objective is something we've already shown to not be very useful, i think it will be a dull project
23:01:24saratogaimo without the port whatever changes are made will simply rot
23:02:08saratogawe can't just separate things without giving rise to a new port and expect others to do the rest, we must have a proof of concept application port
23:02:22saratogaotherwise this exercise will be for nothing like my attempt at rockbox as an app
23:02:33linuxstbI don't think it would rot at all - I for one would use a proper Rockbox app for my main desktop linux player, which is my proposal for a first step, ignoring any embedded OSes.
23:03:08 Quit shivkrish (Quit: CGI:IRC)
23:04:15saratogawhy don't you use the sim now?
23:04:37gevaertssaratoga: getting things to compile from proper locations is really just the beginning. You'd basically get the same code as the sims have now after that, and running the sim on n900 takes lots of CPU and stutters. That really shouldn't happen on a 500MHz armv7
23:05:10saratogasure but if you have a useful device, you can make incremental improvements until it doesn't
23:05:16 Join LambdaCalculus37 [0] (~rmenes@rockbox/staff/LambdaCalculus37)
23:05:22saratogapeople already run our sim on various linux phones
23:05:23linuxstbsaratoga: I do sometimes. But it's things like not having access to my full filesystem (unless I setup symlinks), having a small screen, etc.
23:05:36saratogalinuxstb: those things could quite easily be fixed
23:05:43saratogathe screen res is just a define in a config file
23:06:16linuxstbsaratoga: Yes. But if RaaA existed, they wouldn't need fixing, so I would use that - that's my point.
23:06:35domonokysaratoga: all those small things probably add up time fast..
23:06:59linuxstbAnd ports to actual targets is going to be relatively easy (and fun - meaning lots of people will want to do it) once the infrastructure is there.
23:07:07saratogayes of course, but i mean the changes right now are quite simple, imo the real problems with using rockbox as a media player on a PC are much more fundimental and unlikely to be addressed by this project
23:07:28*gevaerts still thinks that the very first goal (possibly even before the target tree work) should be to use a real cooperative threading library
23:07:30saratogaif using rockbox as a media player was as simple as changing the resolution someone would already be doing that right now
23:07:53saratogai really dont' think messing with the threading is a good idea for this project, at least not initially
23:08:04saratogathe moto porters already showed that what we have is tolerable if poor
23:08:12saratogawe have bigger problems then some inefficiency
23:08:17kugelgevaerts: if we use sdl for drawing, why not use its threads as well?
23:08:33gevaertskugel: have you looked at how the sim does threading?
23:08:45kugela third party threading library only makes sense if we drop the other sdl bits as well, doesn't it?
23:09:08saratogain the long term we should probably just use whatever libraries the OS API provides right?
23:09:19saratogai assume these embedded systems have some prefered threading tools
23:09:31gevaertsI don't think so, no
23:09:42saratogaso they just give you processes and nothing else?
23:09:44domonokywe probably wont easily get around the need of cooperative sheduling.
23:09:47kugeldo OSes provide cooperative threading?
23:10:05kugelrockbox heavily relies on the cooperative bit
23:10:20*domonoky would just write a custom sheduler like the rockbox one, and run it in one host thread.
23:10:23gevaertskernels don't provide cooperative threading because they don't need to. libc provides building blocks you can use
23:10:45TorneYou can do cooperative threading with multiple host threads
23:10:46saratogawell any OS that provides preemptive should also be able to provide cooperative provided its latency isn't too high and we don't need to directly touch hardware registers and interrupts
23:10:48 Quit mitk (Quit: Leaving)
23:10:48gevaertsBut there are libraries around that build on those libc mechanisms to give you something pretty close to what native rockbox does
23:10:49Torneand there are some advantages to doing so
23:11:08linuxstbsaratoga: What fundamental problems are you talking about? Are they different on a PC than something like an Android device?
23:11:22saratogawell on a PC our interface sucks for using a mouse IMO
23:11:26TorneYou just put a semaphore on each thread..
23:11:35gevaertsTorne: is that really efficient?
23:11:36Lloreansaratoga: Touchscreen WPS buttons help with that a lot
23:11:36saratogaand we can't dynamically resize windows which is also annoyinjg
23:11:37Torneand inside yield() you pick a thread to run, signal its semaphore, then wait on your own semaphore
23:11:39LloreanThough the file tree still sucks
23:11:41Tornegevaerts: it's not horrible
23:11:47 Join webguest77 [0] (
23:11:49saratogait helps but its still pretty bad imo
23:11:52Tornegevaerts: and it gives you the ability to "pop outside" the scheduler for a while
23:12:01Tornegevaerts: to do synchronous OS calls or whatever
23:12:10Torneor to pretend to be an interrupt handler, or so on
23:12:13LloreanHeck, just adding a right click to be the "context menu" button would be a vast improvement for sims if it's not.
23:12:17TorneThis is how the Symbian emulator works, for reference ;)
23:12:25 Quit _silentAssassin (Remote host closed the connection)
23:12:31Tornethough we also have preemption
23:12:32 Join EBG [0] (
23:12:38Torneimplemented using a further nasty trick ;)
23:12:41CIA-5New commit by alle (r25347): Correct quotation marks; simplify the viewport display example
23:12:41linuxstbsaratoga: That's the same as integrating with any host UI framework.
23:13:01gevaertsTorne: the rockbox sim does this as well (except for the useful bits :), and I've seen it elsewhere too
23:13:10Tornegevaerts: yeah. it's really not htaat bad an idea
23:13:10saratogaIMO rockbox is never going to be all that great on a mouse based OS as long as we have the current menu system
23:13:12 Quit webguest77 (Client Quit)
23:13:14Tornegevaerts: it works properly in debuggers :)
23:13:22kugelTorne: "we have preemption"?
23:13:28 Quit EBG (Client Quit)
23:13:33Tornekugel: sorry, different we. Symbian.
23:13:44saratogathe menus work great on button based devices, and can probably be made to work pretty well on touch devices, but they're just plain odd on a PC
23:13:49gevaertsFor a simulator it's definitely good enough, but for a real app I'd really go for the most efficient threading possible
23:14:33Tornegevaerts: well yes as an eventual goal
23:14:36Tornebut that's hard..
23:14:45Torneyou would need to start by surrounding almost the entire OS with a preemption mutex :)
23:14:58Tornethen break it off a bit at a time
23:15:16gevaertss/the most efficient threading/the most efficient cooperative threading/ :)
23:15:41gevaertsI'm not convinced that preemptive threading is a good idea on native targets, and I don't think having both is a good idea
23:16:02TorneWell, you can do that too, you just need the relevant asm magic to context switch
23:16:11Tornebut then you have a problem if you need to call blocking OS functions
23:16:20*domonoky mentions that rockbox sheduler on target is not only cooperative, but als has some custom dynamic prioritys.. dont know if that is important for the app part.
23:16:24Torneyou probably still need more than one thread.
23:16:36gevaertsI've very briefly looked at and I think it should work
23:17:26 Join _silentAssassin [0] (
23:18:11linuxstbsaratoga: I'm not talking about Rockbox being a great PC music player. I'm just saying that if the first step of RaaA is to make it run as a PC app, which can be used as an easy-to-debug platform to solve other issues, that's not a waste of time, or something that would rot if taken no further.
23:19:00***Saving seen data "./dancer.seen"
23:19:08 Part Hans-Martin
23:20:06Alexandru_Cristefrom what I understand Rockbox currently has some kernel functions
23:20:14saratogaits a complete os, so yeah
23:20:22Alexandru_Cristeso I suppose it's based on a linux kernel
23:20:52saratogawe're an os, not a linux application
23:21:03domonokyrockbox has its own rockbox kernel :-)
23:21:15saratogaif we ran on linux directly that would make this a LOT easier
23:21:38saratogaright now the only parts that run on unix or windows are the sim, and they're ugly
23:21:42gevaertsTorne: there's also makecontext()/swapcontext(), but those are apparently gone from recent POSIX
23:21:57 Part domonoky
23:22:07Alexandru_CristeI thought the application started from a modified linux kernel
23:22:14Alexandru_Cristeso you just wrote the kernel
23:22:23Alexandru_Cristeall the kernel
23:22:27Tornegevaerts: we could always use the multicore code
23:22:28Alexandru_Cristefrom nothing?
23:22:32TorneAlexandru_Criste: yes
23:22:48Tornegevaerts: to get it into two threads, so the UI is isolated from codec
23:22:57 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
23:22:58AlexPAlexandru_Criste: The whole thing is custom, nothing to do with linux
23:23:02linuxstbAlexandru_Criste: From the top of the Rockbox home page - "Rockbox is an open source firmware for mp3 players, written from scratch. "
23:23:14Alexandru_Cristehow dependent is the application code from it's kernel?
23:23:55kugelvery :)
23:23:56linuxstbAlexandru_Criste: At the moment it's very closely tied - that's the main task for RaaA.
23:25:20saratogawas going to suggest a note on our front page about this, but its already there: "Rockbox is an open source firmware for mp3 players, written from scratch. "
23:25:36*scorche has booked his trip over to europe
23:26:37Alexandru_Cristeso the main task is lowering the coupling between the application and the kernel
23:26:58Alexandru_Cristeand then porting the application under another kernel\
23:26:59Torneyup, though actually implementing the support to run the application on *some* OS is important too
23:27:12saratogai don't know about reducing coupling, but certainly porting
23:27:26Alexandru_Cristeand is it feasable under a 3 months term?
23:27:51 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
23:29:33linuxstbAlexandru_Criste: What's feasible is for you to put in your project proposal.
23:30:11 Join planetbeing_ [0] (
23:31:36 Join salty_horse [0] (
23:31:54 Join planetbeing__ [0] (
23:33:27 Join blairb [0] (
23:33:32 Quit planetbeing (Ping timeout: 258 seconds)
23:33:32 Nick planetbeing__ is now known as planetbeing (
23:34:06kugelgevaerts: it looks like pth uses make/swapcontext
23:34:37 Quit planetbeing_ (Ping timeout: 252 seconds)
23:35:08gevaertskugel: I think that despite what posix *now* says, it's still reasonable to use those if you really want cooperative threading
23:36:14 Quit salty-horse (Ping timeout: 276 seconds)
23:36:37saratogai still want the rockbox test driver project
23:36:45saratogai think that could have a huge impact on the quality of our code
23:38:11 Quit petur (Quit: Zzzzz)
23:38:47gevaertssaratoga: that's basically isolating the playback engine, right?
23:39:09saratogathat or else not isolating it but adding a scripting interface
23:39:26saratogabut isolating it is probably best
23:44:34 Quit S_a_i_n_t (Read error: Connection reset by peer)
23:45:32kugelwhat's the problem with tacking our thread library onto a single OS thread?
23:45:57 Join S_a_i_n_t [0] (S_a_i_n_t@
23:46:27kugelit's simply, but I guess it's only simple because we can do asm tricks we cannot do on a OS?
23:46:54gevaertsYou can do asm tricks on an OS, but those would basically reimplement make/swapcontext
23:48:37 Quit _silentAssassin (Remote host closed the connection)
23:49:40 Quit froggyman (Quit: ta ta)
23:49:47 Join S_a_i_n_t_ [0] (S_a_i_n_t@
23:49:52Tornekugel: our thread library only has it implemneted for arm/mips/sh/coldfire, is the main one ;)
23:50:13 Quit S_a_i_n_t (Ping timeout: 240 seconds)
23:52:08kugeldon't we rely on being able to disable interrupts in some places?
23:52:41kugel(which requires supervisor mode which AFAIK won't have as on app?)
23:52:52Torneyou don't really need to disable interrupts
23:52:56gevaertsoutside drivers?
23:53:09Tornesince you also aren'y handling them, as an app
23:57:25 Join RadicalR [0] (

Previous day | Next day