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 2009-03-07

00:01:02*amiconn oh-noes @ mono bitmap drawing performance on anything chunky
00:01:38 Join fml [0] (n=4fd3d3e0@gateway/web/cgi-irc/
00:01:42amiconnI.e. both 16bit and greylib
00:02:32fmlErr... My commit has caused bootloader build failures. How should I proceed? Revert? Or quickly fix?
00:03:34gevaertsfml: fix
00:03:41amiconn"NCHARS and max code mismatch "?
00:04:45pixelmabluebrother: Bagder said he's off for a few days (holiday)
00:04:50domonokyBagder: rbutil 1.1 binary for mac:"> please copy to the download server
00:04:53 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
00:05:30domonoky(or anyone else with access to the download server) :-)
00:05:41amiconnfml: Looks like you introduced a bug wrt font range limiting. The bootloaders must only get pure ascii.
00:06:07JdGordon|domonoky: yeah, you should stop using and use instead.... i dont knnow when ill change it but it could be soon
00:06:11 Join miepchen^schlaf [0] (
00:06:44domonokydomonoky/rbutilqt-v1.1.dmg"> :-)
00:07:28saratogaI just added two more GSOC ideas
00:09:41fmlamiconn: could you elaborate? Do you mean that bootloaders should only consider chars < 128?
00:09:46linuxstbsaratoga: "optimizing for ARM playback" ? What about our other targets? And where is the WMA pro codec - ffmpeg?
00:10:09saratogalinuxstb: yes its in ffmpeg
00:10:11amiconnfml: Yes. That's specific by a command line switch to convbdf
00:10:37saratogagenerally, coldfire has been fast enough that no arch specific optimization has been needed, at least in my experience
00:10:49saratogaplus its unlikely we'll find a student with a CF device
00:11:23fmlamiconn: ok, I'll check
00:11:29amiconnfml: convbdf has two switches for this, -s and -l . See getopt()
00:11:35saratogalinuxstb: well ffmpeg-soc branch, it hasn't been merged into main yet
00:11:43saratogabut its functional and very nearly complete
00:13:24fmlamiconn: Ok, I see the offending lines. I've introduced them recently. It's weird that it only fails now.
00:14:22amiconnWell, the check is new...
00:15:00amiconnstarting at line 467
00:16:01 Quit reacocard ("dammit screen give me my unicode already")
00:17:23 Join reacocard [0] (
00:17:30 Quit bluebrother ("leaving")
00:19:56fmlamiconn: I introduced it as a sanity check but it was too much of a good. I removed the check for now and will re introduce it (or something similar) in a later version.
00:20:17*amiconn thinks that convbdf now spits far too many warnings
00:20:44*bertrik too
00:21:08fmlI plan to introduce more command line options to allow to build fonts with modified ascent/descent (see mailing list and FS #9931)
00:21:27fmlOk, I'll introduce an option to be verbose or quiet
00:21:36 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
00:21:36 Quit pixelma (Nick collision from services.)
00:21:36 Quit amiconn (Nick collision from services.)
00:21:36 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
00:21:48 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
00:21:53fmlShould the last warning be always printed (how many glyphs were clipped)?
00:21:54 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
00:22:13fmlOr also only when in verbose mode?
00:23:49amiconnOnly in verbose mode imo
00:24:04 Quit fyrestorm ("ChatZilla 0.9.84 [Firefox 3.0.6/2009011913]")
00:26:35fmlamiconn: but then ther will be no indication of clipping at all
00:26:43amiconnThe warnings don't make sense during a normal build, because you get to see the warnings, but not the font(s) they're referring to
00:26:44 Quit goffa (Read error: 104 (Connection reset by peer))
00:27:11amiconnI don't care about this clipping, in fact. If the font is designed to clip, then let it...
00:27:56amiconnIt probably makes sense to display the warnings in verbose mode so that you can analyze a font and fix it if you want, but during normal builds it's just clutter
00:28:20fmlamiconn: OK
00:28:26amiconnI get >4 pages worth of warnings on a 'make 7zip' now
00:29:13 Quit planetbeing_ ()
00:32:49fmlamiconn: now you have something to read before you go to sleep! :-) Ok, fixed.
00:37:28 Quit flydutch ("/* empty */")
00:39:02***Saving seen data "./dancer.seen"
00:39:12 Quit kushalone ("Leaving. I cannot promise to be back but most likely will.")
00:42:53 Join yhuang [0] (n=yhuang@unaffiliated/yhuang)
00:44:44 Join plutonian [0] (
00:44:57plutonianI have a repeating freeze on my rockbox player
00:45:50plutonianwhen I load a certain song with a theme that has album artwork, the song plays for 9 seconds, I see the album artwork, and then at the exact same time I get "Data Abort at 000098BC (0)"
00:45:57plutonianand I have to reboot my iPod
00:46:16plutonianthe crash doesn't occur when I'm using a non-album artwork theme
00:46:52 Quit fml ("CGI:IRC")
00:48:18plutonianand if I start playing the song with a non-album artwork theme
00:48:26plutonianI can switch to an album artwork theme
00:48:32plutonianand then play from the middle of it
00:48:47plutonianbut if I try to play the first 9 seconds with album artwork enabled, it always crashes
00:49:19plutonianis this caused by the bmp scaling code?
00:52:54rasheramiconn: Any idea why archos sims don't get stripped by default whereas the rest do?
00:54:18 Join tyfoo [0] (
00:57:38pixelmaplutonian: could you tell the revision number of the build you're running?
00:58:52plutonianit's the one under official build right now
00:59:11plutonianI'm in disk mode right now, so I can't get the exact number
01:00:42pixelmasounds a bit like the bug to me that got fixed in r20081 which was after 3.1 ( Please try with a current build
01:00:59plutonianwill do
01:01:31 Quit jhMikeS (Nick collision from services.)
01:01:37 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
01:01:42plutonianthat bug doesn't sound like mine
01:01:47plutonianmy bug only happens with certain files
01:02:10 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
01:03:28pixelmathis was happening too for me (I could reproduce it reliably with one album), see my comment there
01:05:22pixelmait did not crash with all files as the original poster reported
01:06:21plutonianWell it'll take me a while to figure out what I need to back up from my .rockbox directory before I upgrade
01:06:29plutonianbut I'll try the nightly
01:07:05BigBambiplutonian: Current
01:07:14saratogayou don't have to back up anything
01:08:07BigBambiplutonian: The archive (daily) build is just the current built at the same time every day. There is no reason to use it over a current build, except in the very rare occasions when the current is known bad
01:08:19plutoniandon't I have to delete my current .rockbox in order to have a clean install?
01:08:39BigBambiJust overwrite it
01:09:01plutonianwhat if there are files in the old version that aren't in the new?
01:09:08plutonianor that shouldn't be, rather
01:09:19BigBambidoesn't matter
01:09:19saratogadoesn't matter
01:09:29saratogalook we're not trying to trick you here
01:09:37BigBambiIf the new doesn't have them, it won't try to use them as it doesn't expect to have them
01:09:56BigBambiplutonian: Honestly, we know whereof we speak
01:10:06plutonianI don't doubt it
01:12:03plutonianI'm just suspicious that my .rockbox contains some cached files that might be causing errors
01:18:03 Quit JdGordon| (" ajax IRC Client")
01:19:00BigBambiplutonian: You can delete everything if you want - if you want to save your settings first, then you can. Also, if you do stuff like runtime gathering with the database then you will need to export that too.
01:20:06Unhelpfulamiconn: yes, only the cache build uses any of the scalers, but we might be able to shave some off of that ~15min cache build if we can speed it up, and the multiplies that can be 16x16->32 are the ones performed most often
01:20:45amiconnI've just committed my tweaks
01:20:55 Quit tyfoo ("Carpe diem")
01:21:26Unhelpfulalso, if you can write up an asm 16x32->32, i think there are some places that could use that inline instead of calling the 32x32->32 routine... they're not as large a portion of the load as the ones in the line scaler, though.
01:21:42amiconnTurned out that the destination and the operands need to have the right size in addition to the casting in order to make gcc use the instruction instead of the subroutine
01:22:30saratogadoes get_buffer advance the buffer position or do i have to call advance_buffer for the amount of bytes read in?
01:23:09amiconnThe only changes in the generated code for arm and cf are now asr->lsr, verified by diffing the disassemblies. Took me a while to get fade_color() right to not change more...
01:24:33Unhelpfulmaking the scaling for cache dynamic is on my to-do list as well, though. it's probably never going to be possible on sh1, though, no matter how much the scaler can be optimized.
01:24:45 Join webguest [0] (n=63f3f065@gateway/web/cgi-irc/
01:24:53amiconnWHat do you mean?
01:24:58webguestHello everyone
01:25:21webguestWill music quiz feature be available anytime soon?
01:25:50Unhelpfuli mean i want to get rid of the "cache building" wait at the beginning. it can be done, with a few core scaler tweaks, and a very small static buffer for the scaler to use.
01:25:56saratogahmm my sample rate ends up being about 16 million
01:26:05 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
01:26:55Unhelpfulthe scaled images will still be cached to disk when they're done, but the startup will be reduced to making sure that the cache items are for the proper albums, and of the proper size.
01:28:17bertrikwebguest, unlikely
01:29:16BigBambiwebguest: If someone that wants it (like you) writes it
01:30:34 Quit domonoky (Read error: 104 (Connection reset by peer))
01:31:38plutonianare .voice files target specific?
01:35:13linuxstbsaratoga: get_buffer() doesn't change the buffer position - you call advance_buffer() when you're finished with that buffer. i.e. call get_buffer(), then process the data, then call advance_buffer with how many bytes you've consumed.
01:35:26plutonianand will the Sansa e200 .voice file on the wiki work for my ipod 5g?
01:35:47saratogalinuxstb: is the buffer invalidated when I call advance_buffer, or do I have to call get_buffer again to do that?
01:36:02 Join CaptainKewl [0] (
01:36:41linuxstbI'm not sure what you mean. What do you mean by invalidated?
01:37:06saratogawell sooner or later that buffer gets overwritten with subsequent data from the audio buffer right?
01:37:10linuxstbplutonian: Yes to your first question, and therefore no to the second.
01:37:12amiconnUnhelpful: For comparison, my H340 needs 4 minutes, and my H180 needs 3 minutes for preparing the slides for the same set of albums.
01:37:31amiconnSo the 15 minutes aren't that bad given the cpu speed difference...
01:37:31 Quit webguest ("CGI:IRC (EOF)")
01:38:01 Join webguest [0] (n=63f3f065@gateway/web/cgi-irc/
01:38:08saratogaoh advance_buffer actually returns a new pointer?
01:38:10linuxstbsaratoga: Yes - only after the codec has called advance_buffer.
01:38:24webguestAbout album arts, can I place all my album arts in one folder? Will rockbox find it?
01:38:26saratogaok so I don't call get_buffer again, just advance buffer?
01:38:40amiconnH340 achieves ~24fps while scrolling around, H180 does 32fps
01:38:42Unhelpfulno, but 15 min is probably not fast enough for cache-build-on-demand
01:38:59UnhelpfulH3xx is color, H1xx grey, right?
01:39:05linuxstbget_buffer() simply returns a pointer to the current "buffer position". advance_buffer will increase "buffer position" and then the memory you've advanced over is free to be overwritten by the buffering code.
01:39:51plutonianwhere do I get voice files for the current build?
01:39:59Unhelpfulthere are about 235 covers on my beast, it takes 75s to build cache
01:40:27plutonianthe newest voice file for the ipod 5g is for "Between 11/18/07 and 8/06/07"
01:40:53amiconnYeah, right. I don't have that many albums, but I do have many single tracks with an album tag without a cover (in 'various tracks' folders)
01:41:03Unhelpfulthe source bitmaps are what, screen-size?
01:41:24saratogalinuxstb: I guess what I'm unclear about is the order of events to step through a stream, do I call get_buffer once and then just call advance_buffer when I need data?
01:41:30webguestis there a program that can download album art that does not require 3.5 Net frame thingy
01:41:38saratogaor do I need to keep calling advance_buffer then get_buffer pairs
01:41:56Unhelpfulwow, you're scaling from that large on the sh1? my estimate of its performance has improved ;)
01:42:25*amiconn was too lazy to make a separate set of BMP files so far
01:42:52*amiconn will try to fix the start screen oddness on mono targets
01:42:53 Join fdinel [0] (
01:43:03Unhelpfuli've got the same problem, i don't really use my e200 aside from testing, so the covers and music on it are just copied from the beast, and therefore a good bit larger than the screen.
01:43:15plutoniansweet, the new build seems to fix my problem
01:43:24 Quit midijunkie ("?(???~•~)?")
01:43:31linuxstbsaratoga: The main decoding loop consists of a call to get_buffer() to get a pointer to the next frame/packet/whatever (you pass it the maximum possible size for a frame, unless you know it), then you call the codec decoding function to process it, then when you're finished, call advance_buffer() with the actual number of bytes you've consumed.
01:43:37Unhelpfulmight it be worthwhile to logf the time the cache build takes, and play around with optimizing the scaler a bit?
01:43:48 Quit yhuang ("Leaving")
01:43:55plutonianwhat is the meaning of the volume tag in wps syntax?
01:43:57linuxstbget_buffer() is guaranteed to return up to GUARD_BUFFER_SIZE (32KB) of contiguous data.
01:44:06plutonianwhat is meant by "0db", I mean
01:45:13 Quit webguest ("CGI:IRC (EOF)")
01:45:23saratogalinuxstb: thanks, now I understand
01:45:34plutoniandoes it mean that the music is being played at its recorded volume? and is there anything special about 0-6 that's different from -56 to 0 in terms of what is being done to scale the volume?
01:46:03rasherplutonian: Anything above 0 might clip
01:47:19 Join fool_ [0] (n=fool@unaffiliated/fool)
01:47:27 Quit toffe82 ("ChatZilla 0.9.84 [Firefox 3.0.6/2009011913]")
01:47:39 Quit n1s ("Lämnar")
01:48:22fool_does anyone know if "the sharper image" audio player work with rockbox ?
01:48:50linuxstbIt doesn't.
01:49:20fool_lol ok
01:49:50fool_do you know what's the manufacturer of it ?
01:49:55linuxstbEvery audio player is different - Rockbox needs to be ported to each one individually, which is a lot of work.
01:50:28saratogaopen it up and see who made it
01:51:19fool_i was hoping somebody'd came across it before me saratoga
01:51:30fool_looks like time to find the toolbox
01:51:38krazykitgoogle would help there ;)
01:51:43saratogawell, you're welcome to look at the wiki, but i doubt anyone here knows
01:52:20fool_krazykit: i did some googling
01:52:57plutonianwill rockbox ever support suspend/resume on ipods?
01:53:10fool_the screws are so small i'm afraid i'm gonna break it
01:53:55 Quit midgey ()
01:54:43saratogaplutonian: I think most people weren't that in favor of it, but if someone did the work well enough it probably wouldn't be rejected
01:54:52saratogaand we'd like to have it on the 1/2G since its needed on those
01:56:15fool_all there is is a huge samsung chip
01:56:42saratogagoogle it
01:56:49bertriktake pictures and write down the numbers on the chips
01:56:50plutonianI like the new LCD screen off
01:56:59plutonianhopefully my battery life will improve
01:57:44linuxstbfool_: That sounds like the flash memory. There are probably other chips hidden somewhere...
01:59:32LloreanSeeking isn't expected to be 100% accurate on long MP3s, right?
01:59:34fool_yep linuxstb
01:59:43fool_SAMSUNG 716 k9laGO8UOM , flash reader
01:59:57saratogawell then search harder until you find hte CPU
02:00:13 Join planetbeing_ [0] (n=planetbe@
02:00:14fool_does that sound like it ?
02:01:11bertrikactions semiconductor apparently
02:04:53saratogaah one of the S1mp3 generic players then
02:06:39fool_ there are a ton of firmware here
02:07:20 Quit bertrik ("Leaving")
02:07:31fool_nothing on the wiki
02:07:40fool_looks like i'm out of luck then
02:11:11 Quit n17ikh ()
02:18:20 Join n17ikh [0] (n=n17ikh@
02:24:20fool_i put it all back together and realize the onoff slider is stuck
02:24:40fool_i knew something's gonna fsck up
02:25:59 Join Q751 [0] (
02:26:46 Part Q751
02:28:18 Quit ender` (" Today's BOFH excuse: broadcast packets on wrong frequency")
02:32:11 Join JdGordon_ [0] (
02:32:53saratogaamiconn: ping
02:33:23fool_saratoga: what did you do with your s1mp3 player ?
02:33:34fool_the current interface sucks bad :/
02:33:46saratogaI don't have an s1mp3
02:34:02 Quit matsl (Remote closed the connection)
02:35:00LloreanSeeking isn't expected to be 100% accurate on long MP3s, right? Is it expected at least to always end up before the target time, I hope?
02:35:05 Join nuonguy [0] (
02:35:49saratogaamiconn: nevermind, found a bug in my makefile
02:39:06***Saving seen data "./dancer.seen"
02:40:02 Quit plutonian ("Lost terminal")
02:43:57saratogawhere is int32_t defined in rockbox?
02:45:09 Quit JdGordon (Read error: 110 (Connection timed out))
02:45:14saratogaah inttypes.h
02:50:42 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
02:53:09 Quit JdGordon_ (Read error: 145 (Connection timed out))
02:54:55 Quit saratoga ("CGI:IRC (EOF)")
03:07:41 Quit Nico_P (Remote closed the connection)
03:11:36 Quit HellDragon (Client Quit)
03:29:02 Join FlynDice [0] (
03:32:39 Join HellDragon [0] (
03:34:27 Quit Thundercloud (Remote closed the connection)
03:38:49 Quit BHSPitLappy (Read error: 110 (Connection timed out))
03:43:09 Part fool_ ("brb")
03:52:54 Join kevin__ [0] (
03:54:09 Join DC1 [0] (n=dc1@
03:55:23 Quit Aurix_Lexico (Read error: 110 (Connection timed out))
04:01:47 Quit efyx (Remote closed the connection)
04:02:12 Join z35 [0] (
04:06:04 Join blkhawk- [0] (
04:11:01 Join Strife89 [0] (n=nds@
04:22:30 Quit blkhawk (Read error: 110 (Connection timed out))
04:23:02 Nick blkhawk- is now known as blkhawk (
04:24:05 Join AndyIL [0] (i=AndyI@
04:26:23Unhelpfulamiconn: i assume i could do *basically* the same thing as MULUQ for signed values as well?
04:27:47 Join Barahir_ [0] (
04:28:30 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
04:28:41Unhelpfulthe linear line scaler could potentially use a MULQ for the same sort of optimization i described for the area line scaler. i doubt you'll often be scaling covers up on sh1, though, and the linear scaler is loads faster because of the less frequent repetition of per-input-pixel math
04:31:09 Join yhuang [0] (n=yhuang@unaffiliated/yhuang)
04:35:20 Quit AndyI (Read error: 110 (Connection timed out))
04:39:03 Quit Lss (Read error: 104 (Connection reset by peer))
04:39:08***Saving seen data "./dancer.seen"
04:39:48 Quit Barahir (Read error: 113 (No route to host))
04:47:16 Quit mirak (Remote closed the connection)
04:50:50 Quit miepchen^schlaf (Read error: 101 (Network is unreachable))
04:57:21 Quit planetbeing_ ()
05:01:18 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
05:07:43Unhelpfulamiconn: when you get a chance, if you don't mind testing them, i have a couple of patches. splashes the number of ticks spent on the cache build after it completes, you'll need to use whatever the ok/select button is on the target to clear it, and uses hardware multiplies in the row scalers. the limits this imposes are that the maximum output width be <2**16, and that the scale factor w
05:07:43Unhelpfulhen scaling down can be no more than 258x...
05:16:23 Quit Strife89 ("Bye Froggy! :-*")
05:26:21 Join kevin___ [0] (n=chatzill@
05:28:38 Join sarixe [0] (
05:37:02 Join planetbeing_ [0] (n=planetbe@
05:37:27 Join rodpod [0] (
05:44:49 Quit kevin__ (Read error: 110 (Connection timed out))
05:54:01 Quit Llorean (Read error: 131 (Connection reset by peer))
05:54:20 Join Llorean [0] (
05:59:59 Quit xSlack ("Ex-Chat")
06:03:10 Quit jhMikeS (Nick collision from services.)
06:03:16 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
06:05:21 Join rocko [0] (
06:05:56 Nick BUMBACL0T is now known as BUMBACLOT (n=ORF@unaffiliated/bumbacl0t)
06:08:12 Quit DaCapn (Read error: 60 (Operation timed out))
06:08:24 Join DaCapn [0] (
06:19:33 Quit BlakeJohnson86 (Remote closed the connection)
06:21:56 Quit DC1 ("$4e75")
06:23:39 Join BlakeJohnson86 [0] (
06:36:55 Quit CaptainKewl (Read error: 110 (Connection timed out))
06:39:12***Saving seen data "./dancer.seen"
06:57:35 Quit XavierGr ()
07:04:46 Quit perrikwp (" ajax IRC Client")
07:05:02 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
07:10:15 Quit rodpod (Remote closed the connection)
07:14:14 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
07:25:23 Join Alex00088 [0] (
07:27:34Alex00088I just installed a new motherboard without re-installing XP (used sysprep). SVN won't let me update. Here's the error: ==== SVN: Can't open file 'utils/MTP/beastpatcher/.svn/lock': Permission denied === Any ideas?
07:27:55Alex00088btw there is no "lock" file in that dir
07:34:38Unhelpfulyou'd probably best look for a more general windows or svn help channel for that. you could try deleting and doing a fresh svn checkout, first, though.
07:36:04Alex00088That would take hours. I'm gonna try copying my Rockbox folder from Ubuntu (a few months old). Thanks.
07:36:21Unhelpfula new svn checkout would take *hours*?
07:38:33Alex00088oops - Cygwin took hours.
07:42:08 Quit planetbeing_ ()
07:47:57 Quit BHSPitLappy (Remote closed the connection)
08:06:35 Join midijunkie [0] (
08:09:17 Join planetbeing_ [0] (n=planetbe@
08:14:43Alex00088I just copied my Rockbox folder to a FAT drive, then copied it back. SVN is updating.
08:16:28 Join einhirn_ [0] (
08:17:42 Quit Alex00088 ("ChatZilla 0.9.84 [Firefox 3.0.6/2009011913]")
08:36:30 Quit einhirn_ (Read error: 110 (Connection timed out))
08:37:35 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
08:39:13***Saving seen data "./dancer.seen"
08:50:57amiconnUnhelpful: I was thinking a bit about the MULUQ (and MULSQ) macros. Perhaps it would be better to use a little inline asm instead of relying on heavy casting to prod gcc into using the single instruction?
08:53:13 Join Rob2223 [0] (
08:53:22 Join midgey [0] (n=tjross@
08:53:41amiconnIt is mainly an SH1 specific optimisation. Coldfire also has a 16x16->32 multiply intsruction, in addition to 32x32->32, but they are equally fast.
08:53:56 Quit n17ikh (Read error: 104 (Connection reset by peer))
08:54:18 Join n17ikh [0] (n=n17ikh@
08:54:23amiconnIn fact the datasheet says that the 32x32 instruction does early termination, while the 16x16 one doesn't
08:56:44amiconnThe disadvantage of using inline asm is that it means the two instructions (lutliplication and fetching the result from the accumulator) will always ibe put together, causing a two-cyle pipeline stall. But in the cases I checked yesterday, gcc does the same anyway.
09:06:13 Quit kevin___ ("ChatZilla 0.9.84 [Firefox 3.0.7/2009021910]")
09:11:25 Quit Rob2222 (Read error: 110 (Connection timed out))
09:12:17 Quit midijunkie (Read error: 104 (Connection reset by peer))
09:13:12 Join n1s [0] (n=n1s@rockbox/developer/n1s)
09:15:34 Quit planetbeing_ ()
09:16:22 Join flydutch [0] (
09:19:38Unhelpfulif the casts work, i'd stick with them. where we *might* find some inline asm useful is an 16x32->32 unsigned multiply for the vertical scalers - that should be possible with two hardware multiplies on sh1, along with shift-add or using mac for one of the multiplies, right?
09:20:38amiconnThe problem with the casts is that it also depends on the operand and destination variables whether they work
09:21:06Unhelpfulbasically, the V scaler is doing the same operation the H scaler is, but on each 32-bit output value in a row from the H scaler, instead of on 8-bit values from the loader
09:21:31Unhelpfulah... so perhaps i'd better look more closely at how it handled the use of that macro in the scaler :)
09:22:14amiconnI discovered this while tracking down why the casts didn't work in my first attempt in pf. Both the operands and the output variable must be 32 bit (i.e. int or unsigned int - signedness doesn't matter) in order to trick gcc into using the instruction
09:22:55Unhelpfulso if the operands are 16-bit or 8-bit?
09:23:02amiconnIf either of the three is 16 bit or smaller, gcc will for some reason still use the subroutine. This even applies if an intermediate variable for the destination can be optimized away.
09:24:09amiconnThis is the cause for my changes to fade_color() operand and return value types
09:25:08amiconnSo using inline asm would make things more reliable.
09:25:27amiconnThis technique is not only used in pf, but also in the greylib and some other places
09:28:53amiconnAnd yes, a 32x16 multiply can be done with two hardware multiplies plus a bit of shifting and adding
09:29:25amiconnThat'll save a bit of time versus 32x32->32 (which needs 3 hardware multiplies + more shifting and adding)
09:30:57Unhelpfulalso, since it's not in core, inlining it for speed might be acceptable. for what it's worth, it appears the same macro works fine in the scaler - the values are uint32_t, the only one i'm not sure about are the bitmap struct's dimension fields, but the asm seems pretty clear
09:32:52 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
09:33:26 Quit z35 ("Leaving")
09:35:58amiconnRegarding your patches - limiting the output size should be fine, but I'm not sure about limiting the downscaling factor
09:38:42 Nick Barahir_ is now known as Barahir (
09:39:16Unhelpfulthat should give you sizes that approach the limits of the algorithm used, anyway, for reasonable target sizes.
09:40:03Unhelpfuleven if you're scaling down to 16x16, that lets you have input of 4128, which is large enough to possibly overflow the pixel value accumulator
09:44:07 Join Darksair [0] (n=user@
09:46:49 Quit _lifeless (Remote closed the connection)
09:47:05 Join _lifeless [0] (n=lifeless@
09:55:57 Join plutonian [0] (
09:56:14plutonianwhat controls how often rockbox refreshes wps tags?
09:56:42plutonianI've got a wps that puts the %pc tag in two places, one in the center viewport and one on the side
09:56:49plutonianthe one in the center refreshes, the one on the side doesn't
09:58:37 Join tvelocity [0] (
09:59:37pixelmais one of them in a line that has the %s tag? Or are the viewports perhaps overlapping?
10:01:14plutonianone of them is in a scrolling line, yes
10:01:20plutoniandoes that make it so it can't refresh?
10:03:08pixelmayes, scrolling lines have a problem with dynamic content that should refresh regularly
10:03:43plutonianI like AAFreak theme, but it has a problem with files that run over an hour
10:03:45 Join bertrik [0] (
10:03:56plutonianthe progress times don't fit
10:06:41 Join Acksaw [0] (
10:15:01plutonianis the fact that my ipod no longer automatically goes into disk mode when I plug it in under rockbox a bug or a feature?
10:16:54linuxstbWhich ipod?
10:31:49bertrikFlynDice, have you been able to test the new patch for the e200v2 rtc?
10:36:32linuxstbplutonian: Then yes, that seems like a bug. Does it never go into disk mode, or just sometimes?
10:39:17***Saving seen data "./dancer.seen"
10:42:30 Quit homielowe (Read error: 110 (Connection timed out))
10:42:53pixelmaif plugging in under Rockbox one should have Rockbox USB now on an 5g Ipod, no?
10:44:08pixelmaplug in USB I mean
10:48:03 Quit midgey ()
10:59:22pixelmaplutonian: since you upgraded from 3.1 to a current build yesterday - and if you still get a USB data connection even if it's not Apple's disk mode - I vote for "a feature"
11:00:40Unhelpfulpixelma: so your recommendation for the pictureflow album view would be direct use of core custom actions? it seems that PLA provides some of the boilerplate that will be needed for that sort of thing, but i don't really see how to make PLA itself useful in a way that's not a huge mess...
11:01:31Unhelpfulit would, except for the most simple cases, end up being each plugin having its very own PLA context, or else the stuff that happens now with combining contexts and some actions just having their mappings disappear
11:03:51 Join __lifeless [0] (n=lifeless@
11:06:04 Quit _lifeless (Remote closed the connection)
11:07:39pixelmacore custom actions? In my opinion PLA is not useful in its current state at all, I could only imagine 2..3 (maybe only one) very basic contexts in a lib like that. If you need more it's better if the plugin itself defines its own buttons or actions as that does prevents hidden "connections" and does not stuff things away. amiconn had an idea how to use the action "system" so you don't have to take care of pre conditions etc., I guess he can expla
11:11:20 Join _lifeless [0] (n=lifeless@
11:12:05Unhelpfulthe core get_custom_action function that PLA calls is what i'm thinking of using... the track list in PF can just use get_action with the list/browser context :)
11:12:56Unhelpfulbut the album list does something nothing in core does, scrolling horizontally. people using scrollwheel targets will expect the wheel to scroll the album list horizontally as well as the track list vertically.
11:13:32 Quit __lifeless (Remote closed the connection)
11:15:41Unhelpfuli can also see PLA *possibly* working, since combined contexts have a defined precedence, if we actually make the contexts even more simple, but sometimes define actions in them more than once. we might define several buttons as menu or select, for example, and then place a horizontal scrolling context before the one with menu and select, and they end up using whichever of their several definitions the scrolling context doesn't steal.
11:18:27pixelmain general?
11:21:14amiconnUnhelpful: Combining contexts is bound to break on some targets, as shown in the past
11:21:29UnhelpfulPLA working in general? i think a system of even simpler, but overdefined contexts could make it work in more places than it does now... but there will always be things that can't work.
11:21:45pixelmaI'd expect it would not work on all targets and you can't actually find out until you try
11:22:05amiconnThis is due to the fact that the contexts hide which buttons are actually used, and hence one context's actions can shadow actions of the other context(s), which are hence unreachable
11:22:06 Join ender` [0] (
11:22:11Unhelpfulso, best to define my own mappings, and call get_custom_action, then, if i'm not using something that fits to a core context?
11:23:09 Part pondlife
11:23:09bertrikis there an overview somewhere of rockbox targets with line-in?
11:23:16amiconnPLA can be useful for a few contexts. Those contexts must never be combined though. Plugins which have too specific needs to use one such context should use their own context(s)
11:23:27pixelmaman, even the first plugin (bubbles) ever converted to PLA had problems and got ugly exceptions
11:24:23bertrikoh, the BuyersGuide wiki shows info like that
11:24:40amiconnUnhelpful: The list context would probably work for the tracklist. Currently the tracklist behaviour is somewhat confusing
11:25:43Unhelpfulamiconn: you mean the fact that select returns you to the album list? we need to use back for that, for consistency, and because some day select needs to let you play tracks :)
11:26:43amiconnAdding playback start shouldn't be difficult in fact
11:27:23Unhelpfulusing the core's browser context means that there will be a useful back and select key, and vertical scrolling, already defined, right?
11:27:25amiconnIt might require exporting one or two more functions from the core, which can then simply be called with the track name
11:28:07amiconnOf course pf needs to release the audio buffer before starting playback in case it uses this buffer.
11:30:05Unhelpfulright, and then it would need to exit without trying to draw any more covers, since its allocation for them will be trashed
11:31:35Unhelpfulok, so i guess the question is, should i be using PLA with my own custom context, or calling get_custom_action directly? PLA already implements the needed get_context_map, although i think that perhaps a simpler one could be written if there is only ever one context.
11:34:45 Quit arohtar (Client Quit)
11:35:00 Join faemir [0] (
11:38:51 Join Lss [0] (
11:39:07 Join MethoS- [0] (
11:43:22 Join {phoenix} [0] (
11:47:36 Join tyfoo [0] (
11:55:24 Quit Darksair ("Use the Force, Luke!")
12:00:55 Join mrkiko [0] (
12:05:06 Join miepchen^schlaf [0] (
12:05:22 Quit rocko ("Leaving")
12:18:21 Quit Acksaw (Read error: 104 (Connection reset by peer))
12:24:37 Join robin0800 [0] (
12:27:26 Quit nuonguy ("This computer has gone to sleep")
12:31:10 Join ibseco [0] (
12:32:38 Join Lear [0] (i=chatzill@rockbox/developer/lear)
12:38:43robin0800soap think your revision number is perhaps the problem one and not the solution in
12:39:19***Saving seen data "./dancer.seen"
12:42:26 Join moos [0] (i=Mustapha@rockbox/staff/moos)
12:54:26 Quit ibseco (Read error: 110 (Connection timed out))
12:56:01 Quit plutonian ("Lost terminal")
12:57:20 Quit ender` (Read error: 60 (Operation timed out))
13:04:56 Join pyro_mania1 [0] (
13:05:42 Quit {phoenix} (Read error: 104 (Connection reset by peer))
13:13:21 Join ibseco [0] (
13:22:44pixelmahrrm... pictureflow.tex actually should now be "enabled" for the bitmap Archos manual and the button table needs to be filled out. I just don't like to do that with the kind of broken controls on Ondio :\
13:26:04 Quit robin0800 (Remote closed the connection)
13:30:28 Join robin0800 [0] (
13:31:09 Join MethoS-- [0] (
13:31:12 Quit Lss ()
13:32:05pixelmain the lang files - all current "user:" lines should now be "user: core", right?
13:32:39rasherYes. Which reminds me that I should have a look at what my translator does
13:38:40 Quit MethoS- (Read error: 104 (Connection reset by peer))
13:41:02 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
13:43:51amiconnUnhelpful: Btw, there seems to be a completely unused function in pf (static inline though, so doesn't end up in the binary): is_empty_rect()
13:47:11 Quit domonoky ("Leaving.")
13:48:40 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
13:51:02 Quit timc (Read error: 104 (Connection reset by peer))
13:53:39 Quit BigBambi (Read error: 104 (Connection reset by peer))
14:02:44 Join Lss [0] (
14:02:47 Quit bmbl ("Woah!")
14:02:51 Join timc [0] (n=aoeu@
14:03:37 Join rocko [0] (
14:08:49bertrikamiconn, lots of plugins have non-static functions that could be made static
14:09:26bertrikthey're easy to find by compiling with -Wmissing-prototypes
14:09:49amiconnWell, it saves space and it's a bit faster (saving a call if the function is only used once and hence inlined)
14:10:14bertrikI did a cleanup of the codecs some time ago, I wouldn't mind doing the same for plugins
14:10:15amiconnAll single file plugins can have all their functions static'ed, except plugin_start
14:10:41amiconnMulti-file plugins can't, of course
14:11:00 Join kugel [0] (n=kugel@rockbox/developer/kugel)
14:11:10amiconnbleh. Odd warning....
14:11:43bertrikhehe :P
14:12:18 Quit Lss (Read error: 54 (Connection reset by peer))
14:13:25 Quit domonoky (Read error: 104 (Connection reset by peer))
14:14:16 Join Lss [0] (
14:14:16*amiconn ponders reversal, 'cause somehow this static'ing messed up the tracklist display at least on recorder
14:14:27amiconnIf you go to the tracklist display, pf freezes....
14:17:27bertrikmaybe it's a different bug, revealed by making things static
14:17:37 Join BigBambi [0] (
14:17:52amiconnEeek. On H1x0 pf now crashes right at the start
14:18:33amiconn(panic: stkov main)
14:18:41 Join Sedgewick [0] (
14:19:46 Quit MethoS-- (Remote closed the connection)
14:22:32 Nick Lear is now known as Lear_ (i=chatzill@rockbox/developer/lear)
14:26:07FlynDicebertrik: I'll test your rtc patch this morning when I can
14:27:32 Quit BigBambi ("No Ping reply in 30 seconds.")
14:32:16 Quit ibseco (Read error: 145 (Connection timed out))
14:35:34 Join Aurix_Lexico [0] (
14:36:41 Join BigBambi [0] (n=alex@
14:36:56 Quit FlynDice (Remote closed the connection)
14:39:22***Saving seen data "./dancer.seen"
14:50:02 Quit robin0800 (Remote closed the connection)
14:58:03 Join rvvs89_ [0] (
14:59:04Unhelpfulamiconn: i think i cleared out the use of that before the renderer rewrite. it used to draw the center slide first, and use the outermost edge of each slide on screen to clip the next slide. the trouble was, that keeps you from seeing the top of the slide when it would project above the one in front of it.
14:59:21 Quit rvvs89 (Nick collision from services.)
14:59:24 Nick rvvs89_ is now known as rvvs89 (n=ivo@pdpc/supporter/base/rvvs89)
14:59:59Unhelpfulfix was to draw from the outside in and overdraw, and it killed all the uses of those rectangle structures
15:02:31 Quit Xerion (Connection reset by peer)
15:06:02 Nick Lear_ is now known as Lear (i=chatzill@rockbox/developer/lear)
15:06:19 Join rvvs89_ [0] (
15:13:44 Quit BigBambi (Remote closed the connection)
15:21:28 Join BigBambi [0] (
15:22:42 Quit rvvs89 (Nick collision from services.)
15:22:45 Nick rvvs89_ is now known as rvvs89 (n=ivo@pdpc/supporter/base/rvvs89)
15:35:45 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
15:43:03 Quit avis (Remote closed the connection)
15:53:12 Join avis [0] (n=ident@pdpc/supporter/student/avis)
15:56:23 Join CaptainKewl [0] (
15:57:03 Join efyx [0] (
15:57:40 Quit efyx (Remote closed the connection)
15:58:04 Join efyx [0] (
16:16:20 Join Wonko__ [0] (
16:21:23 Quit BUMBACLOT (Connection timed out)
16:24:04 Quit faemir ("Leaving")
16:25:46 Join BUMBACL0T [0] (n=ORF@unaffiliated/bumbacl0t)
16:35:09 Join ibseco [0] (
16:35:28 Quit Lear ("ChatZilla 0.9.84 [Firefox 3.1b4pre/20090305033203]")
16:39:23***Saving seen data "./dancer.seen"
16:39:58 Join ender` [0] (
16:46:05 Quit kugel (Read error: 110 (Connection timed out))
16:47:36 Join kugel [0] (n=kugel@rockbox/developer/kugel)
16:52:15n1samiconn: i wonder, is it a good idea to have a function called "main" in pf?
16:54:51 Join obo_ [0] (
16:55:43 Join Llorean1 [0] (
16:57:20 Quit kugel (
16:57:20 Quit Wonko__ (
16:57:20 Quit Aurix_Lexico (
16:57:20 Quit miepchen^schlaf (
16:57:20 Quit Rob2223 (
16:57:20 Quit yhuang (
16:57:20 Quit blkhawk (
16:57:20 Quit amiconn (
16:57:20 Quit pixelma (
16:57:20 Quit dionoea (
16:57:20 Quit maraz (
16:57:20 Quit rphillips (
16:57:20 Quit Ridayah (
16:57:20 Quit shadearg (
16:57:20 Quit tchan (
16:57:20 Quit Tuplanolla (
16:57:20 Quit Barahir (
16:57:20 Quit agaffney (
16:57:20 Quit daurnimator (
16:57:20 Quit alexbobp (
16:57:20 Quit tmzt (
16:57:20 Quit lightbulbjim (
16:57:20 Quit advcomp2019 (
16:57:20 Quit ChanServ (
16:57:20 Quit Seed (
16:57:20 Quit Galois (
16:57:20 Quit scorche|sh (
16:57:20 Quit BUMBACL0T (
16:57:20 Quit avis (
16:57:20 Quit AndyIL (
16:57:20 Quit kadoban (
16:57:20 Quit DataGhost (
16:57:20 Quit Unhelpful (
16:57:20 Quit Bagder (
16:57:20 Quit ze (
16:57:20 Quit ender` (
16:57:20 Quit rvvs89 (
16:57:20 Quit BlakeJohnson86 (
16:57:20 Quit shodanX (
16:57:20 Quit daurn (
16:57:20 Quit Dhraakellian (
16:57:20 Quit planetbeing (
16:57:20 Quit FOAD (
16:57:20 Quit flydutch (
16:57:20 Quit perrikwp (
16:57:20 Quit reacocard (
16:57:20 Quit SirFunk (
16:57:20 Quit einhirn (
16:57:20 Quit _Auron_ (
16:57:20 Quit intrados_ (
16:57:20 Quit courtc (
16:57:20 Quit freqmod_gq (
16:57:20 Quit Scoth (
16:57:20 Quit Tristan (
16:57:20 Quit pabs (
16:57:20 Quit Llorean (
16:57:20 Quit obo (
16:57:21 Quit The-Compiler (
16:57:21 Quit CaptainKewl (
16:57:21 Quit _lifeless (
16:57:21 Quit tvelocity (
16:57:21 Quit chrippa (
16:57:21 Quit killan_ (
16:57:21 Quit soap (
16:57:21 Quit jamiem (
16:57:21 Quit at0m (
16:57:21 Quit crwl (
16:57:21 Quit avacore (
16:57:21 Quit Dieterbe (
16:57:21 Quit jeffronius (
16:58:02NJoinChanServ [0] (ChanServ@services.)
16:58:02 Join Dauron [0] (
16:58:02NJoinkugel [0] (n=kugel@rockbox/developer/kugel)
16:58:02NJoinender` [0] (
16:58:02NJoinBUMBACL0T [0] (n=ORF@unaffiliated/bumbacl0t)
16:58:02NJoinWonko__ [0] (
16:58:02NJoinCaptainKewl [0] (
16:58:02NJoinavis [0] (n=ident@pdpc/supporter/student/avis)
16:58:02 Join rvvs89 [0] (n=ivo@pdpc/supporter/base/rvvs89)
16:58:02NJoinAurix_Lexico [0] (
16:58:02NJoinmiepchen^schlaf [0] (
16:58:02NJoin_lifeless [0] (n=lifeless@
16:58:02NJointvelocity [0] (
16:58:02NJoinflydutch [0] (
16:58:02NJoinRob2223 [0] (
16:58:02NJoinperrikwp [0] (i=4aa794a0@gateway/web/ajax/
16:58:02NJoinBlakeJohnson86 [0] (
16:58:02 Join Llorean [0] (n=DarkkOne@rockbox/administrator/Llorean)
16:58:02NJoinyhuang [0] (n=yhuang@unaffiliated/yhuang)
16:58:02NJoinBarahir [0] (
16:58:02NJoinAndyIL [0] (i=AndyI@
16:58:02NJoinblkhawk [0] (
16:58:02NJoinamiconn [50] (n=jens@rockbox/developer/amiconn)
16:58:02NJoinpixelma [50] (n=pixelma@rockbox/staff/pixelma)
16:58:02NJoinreacocard [0] (
16:58:02NJoinSirFunk [0] (
16:58:02NJoindionoea [0] (n=dionoea@videolan/developer/dionoea)
16:58:02NJoinrphillips [0] (n=rphillip@unaffiliated/rphillips)
16:58:02NJoinTuplanolla [0] (n=jani@unaffiliated/tuplanolla)
16:58:02NJoinmaraz [0] (
16:58:02NJoinshadearg [0] (
16:58:02 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
16:58:02NJoinRidayah [0] (
16:58:02NJoinshodanX [0] (
16:58:02NJoinScoth [0] (
16:58:02NJoinSeed [0] (
16:58:02NJoineinhirn [0] (
16:58:02NJoinagaffney [0] (n=agaffney@gentoo/developer/agaffney)
16:58:02NJoindaurnimator [0] (
16:58:02NJoindaurn [0] (n=daurnima@unaffiliated/daurnimator)
16:58:02NJoinkadoban [0] (
16:58:02NJoin_Auron_ [0] (
16:58:02NJoinintrados_ [0] (
16:58:02NJoinDhraakellian [0] (
16:58:02Mode"#rockbox +o ChanServ " by
16:58:02NJoinchrippa [0] (
16:58:02NJoincourtc [0] (n=court@unaffiliated/courtc)
16:58:02NJoinkillan_ [0] (
16:58:02NJoinfreqmod_gq [0] (
16:58:02 Join The-Compiler [0] (n=compiler@unaffiliated/the-compiler)
16:58:02NJoinDataGhost [0] (i=dataghos@unaffiliated/dataghost)
16:58:02NJointmzt [0] (n=tmzt@
16:58:02NJoinjamiem [0] (n=jm|
16:58:02 Join soap [50] (n=soap@rockbox/staff/soap)
16:58:02NJoinTristan [0] (i=tristan@
16:58:02NJoinalexbobp [0] (
16:58:02NJoinFOAD [0] (
16:58:02NJoinlightbulbjim [0] (
16:58:02NJoinadvcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
16:58:02NJoinjeffronius [0] (n=jomegata@
16:58:02NJoinDieterbe [0] (
16:58:02NJoinplanetbeing [0] (
16:58:02NJoinscorche|sh [50] (n=scorche@rockbox/administrator/scorche)
16:58:02NJoinGalois [0] (
16:58:02NJoinze [0] (i=ze@
16:58:02NJoincrwl [0] (
16:58:02NJoinUnhelpful [0] (n=Militant@rockbox/developer/Unhelpful)
16:58:02NJoinBagder [241] (n=daniel@rockbox/developer/bagder)
16:58:02NJoinpabs [0] (
16:58:02NJoinavacore [0] (
16:58:02NJoinat0m [0] (
16:58:07 Quit tarbo (SendQ exceeded)
16:59:09 Join tarbo [0] (n=me@unaffiliated/tarbo)
17:10:19 Nick rocko is now known as rockanta (
17:12:35 Quit _Auron_ (Connection timed out)
17:12:35 Nick Dauron is now known as _Auron_ (
17:15:08 Quit Llorean (Connection timed out)
17:16:46 Quit flydutch ("/* empty */")
17:16:50 Join gartral1 [0] (
17:20:58 Join toffe82 [0] (
17:26:06 Join webguest68 [0] (n=4c75e088@gateway/web/cgi-irc/
17:26:14 Quit Aurix_Lexico (Remote closed the connection)
17:29:29 Join Xerion [0] (
17:30:40 Join petur [50] (n=petur@rockbox/developer/petur)
17:30:40 Quit webguest68 (Client Quit)
17:30:52 Join Horschti [0] (n=Horscht@xbmc/user/horscht)
17:35:31 Quit ibseco (Read error: 110 (Connection timed out))
17:47:20 Quit Wonko__ (Read error: 104 (Connection reset by peer))
17:47:21 Join _Wonko [0] (
17:48:25 Quit Horscht (Read error: 110 (Connection timed out))
18:00:16 Join {phoenix} [0] (
18:00:45 Quit yhuang (Read error: 110 (Connection timed out))
18:12:13 Quit rvvs89 (Read error: 110 (Connection timed out))
18:17:23 Join SUSaiyan` [0] (
18:17:33 Quit BUMBACL0T ()
18:17:54 Join BUMBACL0T [0] (n=ORF@unaffiliated/bumbacl0t)
18:23:51 Quit rockanta ("Leaving")
18:28:13 Join midijunkie [0] (
18:29:25 Join kushalone [0] (n=kushal@
18:32:06 Join Aurix_Lexico [0] (
18:36:23 Quit SUSaiyan (Read error: 113 (No route to host))
18:39:27***Saving seen data "./dancer.seen"
18:46:08 Quit FOAD ("I'll be back")
18:46:37 Join FOAD [0] (
18:54:53 Quit jhMikeS (Read error: 104 (Connection reset by peer))
18:55:00 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
18:56:01 Quit Seed ("cu, Andre")
18:57:24 Join Landon [0] (n=Landon@unaffiliated/landon)
18:57:39 Join Thundercloud [0] (
19:12:31 Quit gartral1 (Read error: 54 (Connection reset by peer))
19:17:37 Join Seed [0] (
19:22:48 Join Strife89 [0] (n=nds@
19:26:33 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
19:26:53 Quit pondlife (Client Quit)
19:26:55 Join cellofellow [0] (
19:27:12cellofellowwhere can I find the reverse bootloader for 3.1?
19:28:17 Join gartral1 [0] (
19:28:28 Part gartral1
19:29:58 Join nuonguy [0] (
19:30:37rashercellofellow: That's not provided by Rockbox.
19:30:40 Join taylor_ [0] (
19:30:53cellofellowit's a hack, though, but the only one I can find is for 2.x.
19:31:14cellofellow(but, all or rb is a hack anyway :)
19:32:09taylor_This may sound crazy - and Im probably wrong. But the 5g ipod can run games from itunes store. Why cant we look at the code from the 5g(since its decrypted) and find out how to encrypt our own binaries?
19:32:56gevaertscellofellow: you didn't say which player you have, but I'm pretty sure it didn't have a 2.x release
19:33:36Llorean1cellofellow: The reverse bootloader is provided by someone else. This channel is for official Rockbox stuff (things we provide). We don't keep track of other peoples' personal projects, and they're considered off-topic for the channel
19:34:13n1staylor_: i'm not sure what you mean, why would we want to encrypt our own itunes games?
19:34:23 Part cellofellow ("Leaving")
19:34:29 Nick Llorean1 is now known as Llorean (
19:34:29taylor_to run code
19:34:52Lloreantaylor_: We already run code.
19:34:55LloreanWhat do you think Rockbox is?
19:35:08taylor_i Mean on the NEWER ipods
19:35:09n1staylor_: we can already run code on the 5g amd there is nothing that says this encryption would be the same as for newer ipods
19:35:25taylor_3g/4g nano
19:35:33taylor_Its most likely the same
19:35:47taylor_if the same game (for example vortex)
19:36:02Lloreantaylor_: Why would it be the same if, by that point, they knew we had access to the 5G OF _and_ were already taking active measures to make the hardware more secure?
19:36:16taylor_can run on 5g and 6g classic for instance, the code
19:36:20taylor_might be the same
19:36:36taylor_I dont know - it just sounds weird
19:37:37taylor_When you got to the apple download center it says: "games for 3g/4g nano classic, OR fifth generation ipod)
19:38:08taylor_that would suggest then (if you are right) they would have separate apps for different ipod generations
19:38:50LloreanOr that iTunes just does something different when syncing the apps.
19:39:09avisif i were to sell my sansa e270 with rockbox on it, since i installed rockbox on a linux machine logged in with my username, would any other username on another machine, be able to manage rockbox, or the files already located on the device ?
19:39:31Lloreanavis: It's FAT32, it doesn't have that sort of permission system.
19:40:05taylor_Interesting to find out though...
19:40:09taylor_Ill look into this
19:47:35 Join jaykay [0] (
19:47:43 Part pyro_mania1
19:51:47 Join ibseo [0] (
19:52:56 Quit Seed ("cu, Andre")
19:55:33 Join Seed [0] (
19:57:46 Quit kugel (Nick collision from services.)
19:57:51 Join kugel [0] (n=kugel@rockbox/developer/kugel)
19:58:49 Join toffe82_ [0] (n=chatzill@
20:11:02avisis the rockbox daily build the same as svn ? i updated my rockbox to it, and its not auto mounting for me, was told it had usb features in the svn build, thought the daily build might be it. this is for the e270
20:11:13BigBambiavis: The daily build is just a svn build built at a certain time each day. Get a current build to be bang up-to-date
20:11:46BigBambiBut both a current build and today's daily should have USB
20:12:27avishmm, well i'm guessing i dont need to be bang up to date, just to have my device auto mount. not sure whats wrong then. its got MTP enabled on the OF side, it stopped mounting right after the update to daily build
20:13:37BigBambiThe OF should not affect Rockbox USB
20:13:46BigBambiWhat version does Rockbox report?
20:14:08kadobanavis: rockbox doesn't have MTP, so if you're trying to do something from the OS side in terms of MTP, that won't work
20:14:42Lloreanavis: Do you mean the original firmware's not automounting?
20:14:58avisnope. the rockbox firmware doesn't mount
20:15:05avislet me sudo fdisk -l
20:15:34avissudo fdisk -l doesn't see it either. it charges though
20:16:04 Quit jaykay (Read error: 110 (Connection timed out))
20:17:10 Quit toffe82 (Read error: 110 (Connection timed out))
20:17:15avisgot a mount point through the official firmware. that a relief
20:18:56Lloreanavis: What SVN revision are you running, I don't want to try to guess where the typo was in the number you gave before.
20:19:16avisi think i just may have to continue using that usb test build i had. it was somewhat recent
20:19:44avisi'm using daily build r20226
20:20:01LloreanWhy didn't you download the current build instead?
20:20:27avisperhaps i should have. these build things confuse me
20:20:59avisi'll try the current build
20:21:53LloreanIs it the text in RBUtil that's confusing, or what?
20:22:09avisno i'm not using rbutil
20:22:39LloreanI don't think we even link to the daily build from the front page
20:24:03LloreanHow'd you end up there?
20:24:44avisi had followed it from a svn link
20:24:57LloreanWhat do you mean "a svn link"?
20:24:59avisyep the current built works great for auto mounting .
20:25:37LloreanDoes our website link to the daily builds somewhere other than the current build page?
20:26:23avislet me see
20:27:42avisthere is a daily snapshots link from the "using svn" page
20:27:55avisthat took me to the daily build page
20:27:59 Join jaykay [0] (
20:29:54LloreanAh, from a development page.
20:30:04avisregardless, the current build works great. thank you
20:30:43avisits sometimes difficult for me to communicate because of my medication. :) sorry about that
20:31:16Unhelpfuln1s: that main is probably a relic of the PC version, i would think. it could always be renamed or made static, but i don't really see what harm it could cause...
20:32:08n1sUnhelpful: ok, i was just curious, it seems a bit strange to me but if it's no problem
20:32:30Unhelpfulthere's a *lot* that's strange inside PF :)
20:33:42Lloreanavis: I'm removing that from the page anyway. It has nothing to do with using SVN (in fact, using SVN is something you do because you don't want to download the daily tarball)
20:35:59 Quit gromit`` (Read error: 110 (Connection timed out))
20:37:00 Join yhuang [0] (n=yhuang@unaffiliated/yhuang)
20:37:21 Join gromit` [0] (
20:38:03 Quit petur (Remote closed the connection)
20:39:28 Quit Strife89 (Read error: 60 (Operation timed out))
20:39:30***Saving seen data "./dancer.seen"
20:41:25 Join petur [50] (n=petur@rockbox/developer/petur)
20:42:32 Quit ibseo ("quit")
20:56:30 Quit midijunkie ("?(???~•~)?")
20:57:36 Quit Thundercloud (Remote closed the connection)
21:03:22 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
21:19:17 Quit kugel (Remote closed the connection)
21:23:07 Quit tyfoo (Read error: 104 (Connection reset by peer))
21:30:02 Join fdinel [0] (
21:40:19 Join midgey [0] (n=tjross@
21:40:34 Quit jhMikeS (Nick collision from services.)
21:40:40 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
21:41:50 Quit pixelma (Nick collision from services.)
21:41:50 Quit amiconn (Nick collision from services.)
21:41:51 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
21:41:51 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
21:42:05 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
21:42:09 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
22:00:36 Quit krazykit ("brb wireless shens")
22:04:06 Join krazykit [0] (
22:07:16 Join Chris_Black [0] (
22:07:59 Quit Sedgewick (Read error: 104 (Connection reset by peer))
22:08:43 Quit Chris_Black (Read error: 104 (Connection reset by peer))
22:09:02 Join pyro_maniac [0] (
22:09:51 Join kugel [0] (n=kugel@rockbox/developer/kugel)
22:11:59 Quit pixelma (Read error: 104 (Connection reset by peer))
22:11:59 Quit amiconn (Read error: 104 (Connection reset by peer))
22:12:01 Join pixelma [50] (n=pixelma@rockbox/staff/pixelma)
22:12:02 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
22:12:51 Join Rabbit^^ [0] (n=Rabbit^^
22:14:00Rabbit^^I am having to return a now DEAD View to Sandisk. When I receive my replacement, what are the "firmware" files I need to copy to the root directory?
22:15:17LloreanRabbit^^: The view isn't supported by Rockbox yet.
22:16:22 Quit kugel ("ChatZilla 0.9.84-rdmsoft [XULRunner]")
22:17:07Rabbit^^Some people have gotten it to go on View according to posts in Sansa forums at Sandisk. I, unfortunately, do not know of a way to chat with Sandisk, except via e-mail or phone.
22:19:23LloreanThere's development work going on. It's not ready to be used yet. If you want to get involved with development, you can read the thread in the New Ports forum and see how far it's come, and the instructions for doing what can be done so far.
22:19:49 Part Rabbit^^
22:20:52 Quit kushalone ("This computer has gone to sleep. Macs are like babies. They need a lot of sleep. :)")
22:28:12 Join kushalone [0] (n=kushal@
22:30:23 Join toffe82 [0] (
22:39:32***Saving seen data "./dancer.seen"
22:46:24 Quit gromit` (Read error: 145 (Connection timed out))
22:48:24 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
22:48:40 Quit toffe82_ (Read error: 110 (Connection timed out))
22:49:14 Quit jaykay ("ChatZilla 0.9.84 [Firefox 3.0.6/2009011913]")
22:53:54 Join toffe82_ [0] (
22:54:57 Join gromit` [0] (
23:01:31 Quit toffe82 (Read error: 145 (Connection timed out))
23:02:28 Join DC1 [0] (
23:09:07amiconnn1s: It isn't. Having main() in a shared object/dll sometimes causes weird effects, e.g. on cygwin
23:09:35amiconnBut the more puzzling part is why static'ing the functions causes pf to crash/hang
23:10:32saratogaI see Newegg has Fuzes quite cheap, at least in north america
23:16:01 Join krazykit` [0] (
23:25:09 Part toffe82_
23:36:43 Quit krazykit (Read error: 110 (Connection timed out))
23:36:54 Quit Landon (Remote closed the connection)
23:39:59 Quit yhuang ("Leaving")
23:45:34 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
23:47:00 Quit fdinel (Read error: 104 (Connection reset by peer))
23:50:45 Quit petur ("Zzzz")
23:54:12 Quit {phoenix} (Remote closed the connection)
23:58:12 Join Wraith| [0] (

Previous day | Next day