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 2007-11-01

00:01:10preglowhow i hate arm and it's complete lack of hardware accumulators
00:01:17 Join mr_pink [0] (
00:02:14 Quit linuxstb__ (Read error: 113 (No route to host))
00:06:20 Quit davina ("xchat on Ubuntu 7.04")
00:07:15 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
00:07:19*jhMikeS sings the hardware accumulator song
00:07:21 Quit bertrik ("bye")
00:09:59*preglow sings the hardware saturation ballad
00:10:21 Quit jmspeex ("さようなら")
00:14:15preglowone reg too little for all the accumulators
00:14:51preglowi'd give a bloody leg for SMULxy now
00:16:25mr_pinkLooking at the logs I noticed there was a mention last night about the gigbeat custom dock idea... Has anyone actually put much thought or work into this yet? And is there anyone keen that has the right knowledge to make it happen? Or is it more of a pipe dream?
00:17:06karashatamr_pink: there's a wiki page on it if you wanna look into it a bit more
00:19:01mr_pinkYeah I noticed the other day that someone had updated it so it got me thinking about it again
00:19:38***Saving seen data "./dancer.seen"
00:19:50karashataahh, kay
00:21:19 Join TradeJack [0] (
00:21:48preglowamiconn: hmmm, wouldn't multiplying two 16x16 numbers by each other with mul result in the correct answer, even if the top 16 bits of both aren't zero?
00:22:04preglowamiconn: correct answer with regards to the lower 16 bits of reach register, that is
00:23:37preglowamiconn: i mean two packed 1616 numbers, i'm a bit too tired now...
00:26:30 Quit Domonoky_ (Read error: 104 (Connection reset by peer))
00:30:42 Quit obo ("bye")
00:31:48 Join jmspeex [0] (
00:34:03 Quit Nico_P (Remote closed the connection)
00:44:46 Quit moos ("good night")
00:44:47 Join J3TC- [0] (
00:45:05 Quit jhMikeS (Nick collision from services.)
00:45:11 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
00:45:52 Join JdGordon [0] (
00:48:02hcsthe great pumpkin will destroy you for such an off topic remark
00:50:17amiconnYay! 101fps fullscreen on G5 :D
00:50:56rasheramiconn: Improvement from what?
00:50:59 Join midgey [0] (
00:51:16rasherThat's quite a leap
00:51:18rasherWell done
00:51:49amiconnNow 2 things need fixing before commit: yuv blitting, and the background update should be made dualcore safe
00:52:14hcsin other news, SPC codec is failing a lot in the current build
00:52:31hcs(for gigabeat and sansa e200)
00:53:29hcsI'll look into it in a bit, when I can find an outlet for my laptop...
00:53:29amiconnPartial updates profit even more: 390fps for quarter screen (was 59fps before)
00:53:37 Quit hcs ("Leaving.")
00:53:51preglowamiconn: what the hell
00:53:57preglowthat's quite insane
00:54:20amiconnThis is still with the simple bcm rom firmware
00:54:42amiconnJust some neat trickery to get around the 14 ms bcm wait
00:55:34amiconnNow someone make a more optimised yuv->rgb convesion :>
00:56:04amiconnTo beat the current one, it needs to convert in less than 30 cpu cycles per pixel
00:56:20amiconn(including clipping and all)
00:56:25preglowamiconn: anyway, mul r0, r1, r2 where r1 and r2 are two 16 bit packed values will give the same answer as the two low parts multiplied together separately, yes?
00:56:36amiconnNot sure...
00:56:46 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
00:56:49preglowand yes, i am really low on registers to even consider using this :P
00:56:52midgeyat 101 fps, doesnt that make the 5g one of fastest targets for lcd in terms of byte/s?
00:56:52amiconnThe low part will be the same
00:57:06 Quit advcomp2019 (Nick collision from services.)
00:57:08 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
00:57:14amiconnmidgey: Nope, the gigabeat and sansa are faster
00:57:50amiconn(because they use dma from ram)
00:58:07midgeywell, i saw the e200 on LcdFrameRate, but I didn't know the speed of the gigabeat
00:58:11amiconnG5 is now about 15MByte/s
00:58:55amiconnIn fact the transfer speed already was that high, just the bcm wait limited the speed of back-to-back updates
00:59:08midgeyvery nice
01:00:15 Join Mouser_X [0] (i=cf9bb003@gateway/web/cgi-irc/
01:00:44jhMikeSamiconn: How long will it spend in the tick? Just a few uS? If so, don't worry about blocking in a interrupt handler on one core while the other completes.
01:01:03 Quit advcomp2019 (Client Quit)
01:01:09amiconnYes, it's just a few us
01:01:28amiconnRight now I'm using a plain old tick task
01:01:39jhMikeSno worry then. the queues do that.
01:01:55preglowamiconn: using the tick task to transfer image data?
01:02:25amiconnUsing the tick task to initiate a delayed update if necessary
01:02:32amiconn <== this is the whole tick task
01:02:58 Quit ender` (" I like long walks, especially when they are taken by people who annoy me. -- Noel Coward")
01:04:05 Nick fxb is now known as fxb__ (
01:04:15 Quit mr_pink ()
01:04:54preglowthere's some real change this qmf_synth asm will be faster after all...
01:04:55 Quit Arathis ("Bye, bye")
01:05:31preglowman, oh man, gcc does a shit job at that one
01:08:15 Quit mf0102 ("Verlassend")
01:10:04jhMikeSamiconn: core-safe version:
01:10:47 Quit kugel ("ChatZilla [Firefox]")
01:10:48jhMikeSoops, missed that one return;
01:11:11 Nick fxb__ is now known as fxb (
01:11:58amiconnBtw, fullscreen elephants dream is now playing at ~13fps, widescreen at ~17fps
01:12:30amiconnVersions smaller than the screen are currently broken (due to lcd_yuv_blit needing adjustment)
01:12:35preglowhow much would that improve with a nice yuv blit?
01:12:57amiconnIt's already asm'ised - if you could come up with a better one...
01:13:59amiconnPure fullscreen yuv blit is currently running at 27.3fps
01:14:38preglowand you're out of ideas on how to optimize it?
01:14:41amiconnThat means each mpeg decoding and yuv blit need roughly half of the total time
01:15:37amiconnIt will even become a tiny bit slower for less than screen width
01:16:04 Quit midgey ()
01:16:08preglownow, find an arc assembler and make the bcm do the yuv conversion for you :P
01:16:32amiconnNot the conversion itself, but I need to either split the line pairs into 2 separate calls, or let the function deal with bcm address setup
01:17:42preglowyeah, i think the bcm is an arc chip
01:17:45preglowwhich means it uses mips instructions
01:17:45amiconnjhMikeS: I take it that all accesses to lcd_state would need to be corelocked?
01:19:11amiconnjhMikeS: The lcd_update does this: set lcd_update to something != LCD_NEED_UPDATE, then transfer data to the bcm (during that time the tick must not access the lcd.
01:19:15jhMikeSamiconn: I don't know...perhaps not if the interrupt will catch them at some point
01:19:27 Join RoC_MasterMind [0] (
01:19:46amiconnWhen finished, it checks whether the bcm is busy. If it is, it sets lcd_state to LCD_NEED_UPDATE, to let the tick do the update
01:19:54 Quit RoC_MasterMind (Client Quit)
01:20:12amiconnIf the bcm is ready, it kicks off the update itself, and leaves lcd_state set to LCD_IDLE
01:20:22jhMikeSthen yes if you can't risk one extra getting by
01:20:34amiconnOne extra would be a waste
01:21:05jhMikeSthe corelock there is just swp, so no big deal to lock it and change the var and unlock
01:21:25amiconnYeah, but I don't want to lock during the whole update
01:21:42amiconnI mean during the whole data transfer to the bcm, which can be lengthy
01:22:08 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
01:22:25amiconnAs long as lcd_update is never called on COP, I don't need locking at all
01:23:21jhMikeSlock, set the var to something != LCD_NEED_UPDATE, unlock. copy the data, and just do the same procedure when allowing updates again. and it is called on COP.
01:23:43amiconnis it?
01:24:11jhMikeSby mpegplayer but not at the same time
01:24:31amiconnHmm, mpegplayer runs video on cop?
01:24:42amiconnI thought it would run audio on cop, hmm...
01:24:45jhMikeSon pp targets
01:25:17jhMikeSmuch better to use COP for video since audio isn't so heavy and CPU must buffer and handle keys
01:25:42amiconnOn greyscale pp targets it would be better to put audio on cop...
01:25:58amiconn...or better, put the greyscale isr on cop
01:26:11*amiconn wonders whether that would work
01:28:41*amiconn will look into that while implementing the all-new greyscale lib
01:29:00jhMikeSI think the ISR should just switch to whatever core starts it
01:29:55jhMikeSit should work so long as it's not enabled for both at we've seen in other places
01:30:18amiconnIt's a TIMER2 isr
01:30:44amiconnCurrently running on CPU only, as the timer api only enables the interrupt for cpu
01:31:12amiconnI think the timer api should allow core selection on PP
01:31:34jhMikeSright so, that needs a small tweak to select the core starting it or select a core explicitly
01:31:37preglowahaha, hole smokes, just replacing the qmf_synth function with asm shrunk the binary with 1kb :P
01:31:44 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
01:32:07preglowgo gcc
01:32:08amiconnI am even thinking about alternation, in order to balance the load
01:32:17hcsdoh... how might I reboot a frozen sansa (e260)?
01:32:39amiconnAlthough that might cause problems with cache coherency
01:32:51pixelmahcs: hold power for about 15 sec.. core flush, other invalidates. the data should be aligned properly too then...or use uncached aliases and alignment
01:33:34amiconnUncached would be nasty...
01:34:16jhMikeSI'm not seeing any framrate penalty for uncached yuv writes at all...not on e200 anyway
01:34:38 Join Nilisco [0] (
01:34:50amiconnHmm, I guess it's just speculation right now
01:34:52NiliscoThis might be a bit off topic for this channel, but has anyone ever opened a ipod here (and have any tips? )
01:35:14jhMikeSgo ask in #rockbox-community?
01:35:17hcspixelma: ah, so that shuts it off, I thought nothing had happened, was expecting a reboot, thanks!
01:36:11NiliscoI'll take that as a no. Thanks anyway!
01:36:13 Part Nilisco
01:36:30pixelmawell yeah, you need to press the power button again ;)
01:36:38preglowjmspeex: qmf_synth() seems to have a tremendous impact on arm also, 250% realtime on uwb now
01:37:01amiconnup from?
01:37:50preglowhowever, there is still a bug, so it might still not be the right performance figure
01:38:11preglowand further optimizations in the innerloop is possible
01:38:23preglowamiconn: it really isn't a "wow" when you see the bloody mess gcc made of that function
01:38:52*amiconn is curious about the before-and-after values on PP5002
01:38:52jhMikeSpreglow: vorbis builds seektables at load time?
01:39:04preglowjhMikeS: vorbis builds _everything_ at load time, afaik...
01:39:21preglowbtw, i don't think vorbis uses a seek table
01:39:40preglowit just skips from ogg header to ogg header in search of where it's going, afaik
01:39:43preglownot 100% certain
01:39:44 Quit lee-qid_ ("aufwiederbyebientotsayonara")
01:39:49jhMikeShmmm...I think mpegplayer needs that if I want video scrubbing with the wheel. 100KB would hold 1s precision for the entire timestamp duration.
01:40:09preglowjhMikeS: so, that's what you're working on now? :P
01:40:28jhMikeSI said I would after dual core was in place.
01:40:35preglowdid you, know
01:40:41preglowgod, i should sleep
01:40:46preglowputting down asm is so hard...
01:40:54jhMikeSyes, look in the IRC logs
01:41:15preglowi believe you :)
01:41:31jhMikeSnow that all kernel objects are useable on DC it's much easier to do whatever
01:42:48amiconnjhMikeS: Any news regarding DC on PP5002?
01:43:06jhMikeSthis thing could support playlists too since it's not hard coded to end the plugin at video end...but do we need yet another full-blown playback engine? :)
01:43:53jhMikeSamiconn: haven't gotten there yet. I suspect that most of it could be thrown-in rather quickly.
01:44:04amiconnImho video is just an add-on for a dap, and we don't need to support video playlists
01:44:44amiconnOn a true video target this might be different, and video might find its way into the core. But then some brilliant wps ideas are needed ;)
01:44:49jhMikeSIt doesn't really feel like an addon on a gigabeat. Feels meant for such a thing which a good screen and fast processor.
01:45:25preglowand g5 ipods _are_ meant for video
01:45:31jhMikeSWPS video mode? YUV blitting can leave a margin for OSD.
01:47:03jhMikeSpreglow: but that needs the special broadcom support, right? so until that, it's sort of clunky eh?
01:48:15*jhMikeS wonders if gigabeat wasn't meant for video but they abandoned it to bring it to market faster.
01:49:21amiconnNow cube.rock on G5 is almost as fast as on H300 (h=1, both targets boosted), even though G5 has significantly more pixels...
01:49:35preglowgourad shading! :P
01:49:38jhMikeSvideo players are damn complex :\
01:49:58Mouser_XVideo on Gigabeat = nice.
01:50:05Mouser_XI'm really glad to have one of those.
01:50:33karashatavideo on H10 seems okay, aside from the really tiny screen...
01:52:31 Join hc1 [0] (n=agashlin@nat01-southtowers-ext.Rutgers.EDU)
01:52:39 Join formulaONE`pende [0] (
01:52:52 Quit hcs (Nick collision from services.)
01:52:56 Nick hc1 is now known as hcs (n=agashlin@nat01-southtowers-ext.Rutgers.EDU)
01:53:05 Part formulaONE`pende
01:54:00jhMikeSyeah, H10 looks pretty good
01:54:15*karashata nods
01:54:45karashataall the videos I've encoded and tested on it played back at full fps, no issues really
01:55:22jhMikeSI have some that don't quite make it but they're a little too big in frame size
01:55:34 Quit jhulst ("Konversation terminated!")
01:56:53 Quit spiorf (Remote closed the connection)
01:57:22*jhMikeS doesn't have anything he's specifically codec for a device...just what was laying around in his archives and various test cases
01:58:25karashatait takes me about 15 minutes to encode a ~23 minute video for playback on my H10
02:03:01hcsCan anyone think of any recent changes that may have broken the SPC codec?
02:03:18 Join alienbiker99 [0] (
02:03:31jhMikeSman, I botched the seektable estimate. That thing would be huge at 1s res over a full-range video.
02:04:02jhMikeShcs: it's broken? at last check it seemed to work ok.
02:04:12karashatahcs: SPC is working fine here...
02:04:43hcsjhMikeS: crashes the sim and breaks on my gigabeat and sansa e260, just built it an hour ago
02:04:47 Join donsdw [0] (
02:05:01 Join midgey [0] (
02:05:32jhMikeShcs: then it's quite recent. I did SVN up today and will verify on the build on gigabeat now.
02:06:26jhMikeShcs: what happens for you? It's running ok.
02:06:50hcsseems to not play the first few songs in a directory
02:07:00hcsand occasional codec failures
02:07:19preglowgod damn!
02:07:31jhMikeSI got those right after MoB but only after starting it during playback of something else.
02:07:32preglowthe amount of copy&paste errors i commit
02:08:13jhMikeShcs: it is only when playback switches directories itself? I can start it explicitly in a directory and it's ok.
02:09:07hcsjhMikeS: looks like it happens when starting explicitly in a directory, too build is from an SVN earlier today around 4:00 EDT
02:10:21hcsprobably something sensitive with what memory it's overwriting
02:11:59*jhMikeS suggests a settings reset and retry since his own build is >= r15384
02:12:41hcsI'm using the simulator to test now, with a clean slate.
02:14:30safetydanDo we have a standard response to bugs raised where the user is using a custom build?
02:15:24jhMikeSr15386 actually
02:15:24safetydan"Closing this as you are using an unsupported build. Please retest on an official build and create a new bug if the issue is reproducible there."?
02:16:05hcsjhMikeS: I don't understand what you're mentioning revs for
02:16:53jhMikeShcs: just to say "my build is basically latest SVN and not having the trouble you describe"
02:17:39jhMikeSthough I did just witness something interesting for a moment
02:18:30hcsI would like to not that the crashes seem to go away if i leave out the pcmbuf_insert call
02:19:11jhMikeSThe track info in the WPS was just a bunch of boxes for a few seconds while the track switched over. Probably clearing the metadata when the codec is unloaded but not when the track is actually done sounding.
02:19:37jhMikeSheh, just did it on the next switch too
02:19:42***Saving seen data "./dancer.seen"
02:19:50hcsjhMikeS: I have rev 15388
02:20:56hcsnothing of import seems to have changed in the meantime, though
02:21:27jhMikeSnope. I looked at the playback update but that seems to not be affecting me.
02:22:02jhMikeSbut then I don't seem to remember the boxes in WPS either...hmmm
02:25:22hcshum, why are the samples int32_t?
02:26:04jhMikeSwhich samples?
02:26:37hcsthe main buffer I render into, samples[]
02:27:06jhMikeSbecause it outputs 24-bit data?
02:27:15hcsah, quite
02:27:54hcsbeen a while since I've worked on this
02:29:07jhMikeSshould I throw the ARM asm in? Some real improvement but it won't get you echo on PP without COP use. (the parallel processing will have to wait a little)
02:29:41hcsIf you wish, I'd like to find out what's wrong just now first.
02:30:32 Part pixelma
02:30:48jhMikeSNothing's been altered in the track handling (nothing I did that I remember or intended anyway) :p.
02:31:16hcsI'll do some regression, see if I can narrow it down
02:31:22hcsI hadn't built a build for a while
02:35:18jhMikeShmm...last SPC commit is the building with its own -O level. last before that was coldfire assembly work.
02:36:10hcsI don't think there had been any SPC changes since I checked
02:36:17hcs*since I last built
02:36:43pregloww00t, 254%
02:37:04jhMikeShcs: it's been 3 months
02:37:22hcsjhMikeS: right, definitely been since then
02:38:12preglowjhMikeS: if i have two packed 16x16 registers and use a standard arm mul instruction, will the result be the lower two 16 bits multiplied by each other?
02:38:19preglowi just can't wrap my head around that now
02:39:31preglowmy head says yes, but the ipod says no...
02:40:23preglowbeh, the lower 16 bits will be right
02:41:16jhMikeSthe upper 16 will have the carries added of course
02:41:19preglowi've got a bad register squeeze going on
02:42:02jhMikeSmaybe that's why gcc mangled it so much
02:42:39preglowanyway, even having to fetch shorts from memory all the time, i'm way, way faster than gcc
02:43:03jhMikeScan SP be stored out of the way and restored?
02:43:27preglowsure, if you can write to your code section :P
02:43:59preglownot very flash friendly
02:44:01jhMikeSno, just stash it per core and use it for scratch unless you need to call some other routine
02:44:37preglowstash it at some constant offset, you men?
02:45:00jhMikeSno, at some static location
02:45:06preglowahh, i meant that, yes
02:45:14preglow0x10000000 would be nice :P
02:45:47jhMikeSthough that is a really desperate shortage then
02:46:03preglowanywho, can you think of any faster clipping for arm than: cmp r1, limit \n movgt r1, rlimit \n cmn r1, limit \n rsblt r1, rlimit, #0 ?
02:46:12preglowthat is symmetric clipping
02:46:35preglowfor assymmetric you'd just do mvnlt r1, rlimit, i guess
02:48:03preglowyes, as in you know of something faster?
02:48:04jhMikeSwell, I can think of it but I'll get bopped again just for doing so :)
02:48:20preglowfour cycles isn't too bad, really
02:48:58jhMikeSit turns it into asymmetric, unsigned clipping
02:49:13 Join psycho_maniac [0] (
02:49:55 Quit keanu (Read error: 104 (Connection reset by peer))
02:50:02preglowi can live with assymetric clipping in this particular spot
02:51:58 Join grndslm [0] (
02:52:17jhMikeSbasically like the arbitrary one in dsp.c. is the range variable or fixed?
02:52:44preglowsixteen bit clipping
02:53:25jhMikeS-32768 to 32767? use the one from DSP.c as compiled. it's four instructions
02:53:49 Join keanu [0] (n=keanu@unaffiliated/keanu)
02:53:54preglowwith how many cycles on clip not taken?
02:54:20jhMikeSI think it's four always
02:54:47preglowthen i'll just continue using this one, since this is also four always :)
02:55:42preglowi guess it's kind of hard to get under four, since that just about leaves you with one instruction and one branch, and unless you also want to always execute the clipping case, you'll need a branch
02:56:22jhMikeSwait and the result should be sign-extended to bit-31 or not?
02:56:39preglowyou know what, in this particular scenario, it would be for the better if it was not
02:56:59jhMikeSheh, you can have your pick
02:57:13preglowsince then i can or together the sixteen bit results and transfer them with stmia
02:57:25preglowif they're sign-extended, i can't or them together if they're negative
02:58:05 Quit donsdw ()
02:58:53preglowhah, if i change the rsblt r1, rlimit, #0 to addlt r1, rlimit, #1, i achieve just that
02:59:40preglow0x8001 is -32867, isn't it?
02:59:43preglowsleepy brain
02:59:49preglow-32767 ...
02:59:58jhMikeS <==
03:00:18jhMikeSI'm sure you can make whatever at the 4th instruction and after
03:00:20 Join Falco98 [0] (
03:01:47Falco98hey all
03:06:35keanuJdGordon, other than "rejected" what was the reason for closing FS5047?
03:08:55preglowseems like a case of settings bloat to me
03:09:12preglowbut i can't remember him mentioning it
03:10:12 Quit karashata ("Leaving.")
03:10:47 Join karashata [0] (
03:11:27Falco98I wanna see if anyone has seen anything like this: i'm having a pesky recurring issue (iriver h140); some time after installing a new build, the player starts "losing" its playlist on resume after an idle poweroff... and if i instead manually power-off, it loses its settings instead.
03:12:28Falco98i ran a scandisk a few weeks ago at someone's suggestion and found that like 2 of the system files were crosslinked, so i corrected this. but, it remains. I'm starting to think that rockbox might be doing this thru improper handling somehow
03:12:41Falco98it's pestersome as it's impossible to trace and hard to replicate
03:12:55safetydanFalco98: do you property unmount your player before unplugging the USB cable?
03:13:02Falco98i don't believe i'm experiencing disk corruption, as it's always those specific files
03:13:11Falco98generally, yes
03:13:20Falco98though sometimes the unmount fails and i have to unsafely unplug it anyway
03:13:30Falco98i do, at least, make sure the disk is inactive when i do
03:13:49safetydanthat's likely to lead to filesystem corruption though and disk activity isn't a reliable indicator depending on write caching and so forth
03:14:00Falco98"safely remove"ing doesn't do anything additional, as far as i've seen
03:14:13safetydanit ensures the write cache is flushed at least
03:14:19Falco98ah ok
03:14:38Falco98wouldn't i have write caching disabled in my OS?
03:14:44Falco98or would that not matter?
03:14:53safetydandepends on the os and how dumb it is about removable storage
03:15:06safetydanIt's "Optimise for quick disconnect" or something like that on XP
03:15:25Falco98whatever it is, i believe i have it enabled
03:15:53jhMikeSgo to device manager while the device is pluggin, select the disk driver for the device and check the policies tab
03:19:11Falco98yeah, that's how it was set
03:20:25Falco98hmm.. what's the command name again? i thought it was "scandisk" but that's not working now
03:21:11Falco98not that one.. just the scan / fix system errors
03:21:46 Quit Echelon (Remote closed the connection)
03:22:08Falco98oh, i guess it was chkdsk
03:22:13 Join Echelon [0] (
03:25:55 Nick fxb is now known as fxb__ (
03:28:35Falco98also... does anyone have any idea why tetrox was taken out of the standard build?
03:28:59jhMikeSit was renamed rockblox?
03:29:45Falco98oh? cool, i never saw that announced or anything.
03:29:45safetydanI think so yes
03:29:53safetydanit was a long time ago
03:29:59jhMikeSlegal bs from the company with the TM
03:30:36Falco98grumble :-P
03:30:52 Join Shaid` [0] (
03:31:13Falco98btw.. would it be safe to delete the directory called "system volume information" from my player? that seems like windows BS to me
03:31:32Falco98i also don't see why it has a "recycled" directory
03:32:03safetydanFalco98: you can delete them but they'll return as soon as you connect to Windows again
03:32:29Falco98ah, yeah.. the problem is that SVI has like 20 megs of crap in it atm
03:32:42jhMikeSyeah, just don't let recycled fill up (hit Shift when deleting files from Explorer)
03:33:19Falco98yeah, i never "recycle" anyway, especially if removing stuff from my player
03:35:36 Quit grndslm ("Leaving")
03:38:30hcsjhMikeS: son of a... one problem solved, I had copied the wrong codec binaries to my gigabeat
03:39:54Mouser_XThat could be part of it...
03:40:48Falco98would it be a good idea or useful at all to defrag my disk?
03:41:33Mouser_XIt's certainly not a bad idea.
03:41:49Mouser_XIn fact, it could probably help in disk spinup.
03:42:49Falco98oh, would a defrag really recognize the rockbox files and prioritize them?
03:42:52 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
03:45:19 Part keanu ("Leaving")
03:45:52karashataFalco98: no
03:45:59karashataI can tell you from experience
03:46:55 Quit Shaid (Read error: 110 (Connection timed out))
03:46:56 Nick Shaid` is now known as Shaid (
03:48:19Mouser_XFalco98: What I was refering to, is that it wouldn't have to spin the disk as much, when loading data, because it wouldn't be scattered all over the disk.
03:48:54karashataI have to agree with Mouser_X, it does help a bit
03:49:20Falco98i'll try it sometime
03:49:22karashatathough it may take a little while for it to complete, if my experience tells me anything
03:49:35Falco98really atm i'm less concerned about spinup than i am about this weird shutdown error i'm having
03:49:43Falco98it's kinda frustrating, as you might imagine
03:50:00karashataI also have to note...
03:50:10karashataI never get the SVI folder on my player
03:50:17 Quit alienbiker99 (Read error: 104 (Connection reset by peer))
03:50:33 Nick lostlogic is now known as foundlogic (n=lostlogi@rockbox/developer/lostlogic)
03:50:43Falco98i just went into windows and told it to disable system restore on that drive.. i dunno why i hadn't done that already :-P
03:50:49Falco98i still have the folder but it's now empty, at least
03:51:23 Nick foundlogic is now known as lostlogic (
03:53:53Falco98so when i try to safely remove my drive, it says "cannot safely remove right now."
03:54:03Falco98but it's absolutely motionless.
03:57:00jhMikeShcs: rotf xD
03:57:23hcsjhMikeS: still, doesn't explain the sansa
03:57:47jhMikeSwell, you did rule that out there too right?
03:57:59Falco98grr, systemrestore came back on when i plugged it back in...
03:58:56jhMikeSFalco98: I get that all the time on first attempt or when an Explorer window is on the drive. I'll pull it anyhow if I know I don't have it open in some program.
03:59:32jhMikeSIt delayed write is off. No worry. I never use safe removal.
03:59:46Falco98i tried a second time and for some reaason it worked then :-P
03:59:48hcsjhMikeS: yeah, sansa is still inexplicable. and the crashing in the simulator is odd as well
04:01:37 Quit Falco98 ("ChatZilla [Firefox]")
04:04:56 Join karashat1 [0] (
04:05:25 Quit psycho_maniac (" HydraIRC -> <- IRC with a difference")
04:11:47 Join saratoga [0] (i=9803c6dd@gateway/web/cgi-irc/
04:14:37 Quit karashata (Read error: 110 (Connection timed out))
04:19:46***Saving seen data "./dancer.seen"
04:21:20 Nick karashat1 is now known as karashata (
04:22:27saratoga(for the logs)
04:22:59 Quit midgey ()
04:23:26saratogacan we please give Toni SVN access
04:23:58jhMikeShe should have it already
04:24:26saratogaBuschel too then while i'm at it
04:28:02 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:28:17 Join miepchen^schlaf [0] (
04:30:36 Quit saratoga ("CGI:IRC (EOF)")
04:35:46 Quit MournBlade ("Her lips were wet with venom")
04:36:38 Quit DataGhost (Nick collision from services.)
04:36:44 Join DataGhost [0] (
04:52:50 Part TradeJack
04:57:35 Join karashat1 [0] (
05:00:19DEBUGReceived signal 15 (SIGTERM), terminating (snapshot: fplrun.c line 385)
05:00:19***Saving seen data "./dancer.seen"
09:13:06***Started Dancer V4.16
09:13:06***Connected to on port 6667
09:13:06***Logfile for #rockbox started
09:13:05CtcpVersion from freenode-connect!freenode@freenode/bot/connect
09:13:06***Server message 501: 'logbot :Unknown MODE flag'
09:13:06Mode"logbot :+i" by logbot
09:13:07***Server message 477: 'logbot #rockbox :[freenode-info] if you need to send private messages, please register:'
09:13:07 Join logbot [0] (i=bjst@gateway/web/cgi-irc/
09:13:07 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
09:13:07 Join Zagor [0] (n=bjst@rockbox/developer/Zagor)
09:13:07 Join Buschel [0] (
09:13:07 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
09:13:07 Join Rob2222 [0] (
09:13:07 Join hannesd [0] (
09:13:07 Join Genre9mp3 [0] (n=yngwiejo@rockbox/contributor/Genre9mp3)
09:13:07 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
09:13:07 Join homielowe [0] (
09:13:07 Join scorche [0] (i=Blah@rockbox/administrator/scorche)
09:13:07 Join keanu [0] (n=keanu@unaffiliated/keanu)
09:13:07 Join bb [0] (n=bb@unaffiliated/bb)
09:13:07 Join DataGhost [0] (
09:13:07 Join miepchen^schlaf [0] (
09:13:07 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
09:13:07 Join Shaid [0] (
09:13:07 Join Echelon [0] (
09:13:07 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
09:13:07 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
09:13:07 Join J3TC- [0] (
09:13:07 Join jmspeex [0] (
09:13:07 Join Soap [0] (n=Soap@rockbox/staff/soap)
09:13:07 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
09:13:07 Join Soap_ [0] (n=Soap@rockbox/staff/soap)
09:13:07 Join TotallyInfected [0] (
09:13:07 Join freqmod_nx [0] (
09:13:07 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
09:13:07 Join crashd [0] (
09:13:07 Join rasher [0] (n=rasher@rockbox/developer/rasher)
09:13:07 Join tedrock [0] (
09:13:07 Join BigBambi [0] (n=alex@rockbox/staff/BigBambi)
09:13:07 Join nanok [0] (n=nanok@
09:13:07 Join idnar [0] (n=mithrand@unaffiliated/idnar)
09:13:07 Join illissius- [0] (
09:13:07 Join in-jane__ [0] (
09:13:07 Join Bagder [0] (
09:13:07 Join Weiss [0] (
09:13:07 Join joshin [0] (n=joshin@unaffiliated/joshin)
09:13:07 Join pill [0] (
09:13:07 Join Isolinear [0] (
09:13:07 Join FOAD [0] (
09:13:07 Join DogBoy [0] (
09:13:07 Join billytwowilly [0] (
09:13:07 Join SirFunk [0] (
09:13:07 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
09:13:07 Join ATravelingGeek [0] (n=ATG@pdpc/supporter/student/ATravelingGeek)
09:13:07 Join atsea-39 [0] (i=atsea-@gateway/tor/x-40174253f69b664a)
09:13:07 Join Jon-Kha [0] (
09:13:07 Join Slasheri [0] (i=miipekk@rockbox/developer/Slasheri)
09:13:07 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
09:13:07 Join EnterUserName [0] (n=dave@pdpc/supporter/student/GeekZoid)
09:13:07 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
09:13:07 Join n17ikh|Lappy [0] (
09:13:07 Join thegeek [0] (
09:13:07 Join darkless [0] (
09:13:07 Join sbeh [0] (
09:13:07 Join lostlogic [0] (
09:13:07 Join TTThomas [0] (
09:13:07 Join JRoT [0] (
09:13:07 Join XavierGr [0] (
09:13:07 Join parafin|away [0] (
09:13:07 Join RaRe [0] (
09:13:07 Join markun [0] (n=markun@rockbox/developer/markun)
09:13:07 Join HellDragon [0] (n=Nocebo@unaffiliated/helldragon)
09:13:07 Join jumijoze [0] (n=jumijoze@unaffiliated/mike-zed)
09:13:07 Join sd__ [0] (n=sd@
09:13:07 Join Seed [0] (
09:13:07 Join kclaf [0] (i=kclaf@
09:13:07 Join rvvs89 [0] (n=rvvs89@pdpc/supporter/active/rvvs89)
09:13:07 Join daurnimator [0] (i=daurn@unaffiliated/daurnimator)
09:13:07 Join gromit` [0] (
09:13:07 Join andrewg867 [0] (
09:13:07 Join JETC- [0] (
09:13:07 Join crashmatrix [0] (
09:13:07 Join Guile`` [0] (n=Guile@
09:13:07 Join Xerion [0] (
09:13:07 Join midkay [0] (n=midkay@rockbox/developer/midkay)
09:13:07 Join nnakihfdn [0] (i=0@
09:13:07 Join Dave2 [0] (i=dave@freenode/staff/dave2)
09:13:07 Join bagawk [0] (n=lee@unaffiliated/bagawk)
09:13:07 Join [mbm] [0] (i=mbm@openwrt/developer/mbm)
09:13:07 Join lastebil [0] (
09:13:07 Join Hadaka [0] (
09:13:07 Join fxb__ [0] (
09:13:07 Join crwl [0] (
09:13:07 Join Kohlrabi [0] (
09:13:07 Join preglow [0] (n=thomj@rockbox/developer/preglow)
09:13:07 Join maddler [0] (
09:13:07 Join Toki_ [0] (
09:13:07 Join Presence [0] (
09:13:07 Join w0rd54 [0] (
09:13:07 Join Galois [0] (
09:13:07 Join marcosource [0] (
09:13:07 Join Zom [0] (n=zom@
09:13:07 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
09:13:07 Join Cronoh [0] (
09:13:07 Join ze [0] (
09:13:07 Join Rick [0] (i=rick@unaffiliated/rick)
09:13:07 Join jepler [0] (n=jepler@emc/developer/jepler)
09:13:07 Join Nimdae [0] (
09:13:07 Join gtkspert [0] (
09:13:07 Join courtc [0] (n=court@unaffiliated/courtc)
09:13:07 Join micols [0] (
09:13:07 Join krazykit [0] (
09:13:07 Join puzzles [0] (n=dan@xmms2/developer/puzzles)
09:13:07 Join GodEater [0] (n=bryan@rockbox/staff/GodEater)
09:13:07 Join GodEater_ [0] (n=bryan@rockbox/staff/GodEater)
09:13:07 Join TheDarkOne[lappy [0] (n=marc@
09:13:07 Join dionoea [0] (
09:13:07 Join guyzmo [0] (
09:13:07 Join feisar [0] (
09:13:07 Join Bjoern-Erik [0] (
09:13:07 Join sslashes [0] (n=sslashes@
09:13:07 Join solatis [0] (
09:13:07 Join sneakums [0] (
09:13:07 Join [omni] [0] (
09:13:07 Join Ave [0] (
09:13:07 Join delYsid [0] (n=user@debian/developer/mlang)
09:13:07 Join newbyx86 [0] (
09:13:07 Join maxkelley [0] (n=max@unaffiliated/themaximus)
09:13:07 Join J [0] (
09:13:07 Join tuplanolla [0] (
09:13:07 Join shodanX [0] (
09:13:07 Join pabs [0] (
09:13:07 Join blithe [0] (
09:13:07 Join jurrie [0] (
09:13:07 Join DiDjCodt [0] (
09:13:25linuxstb_That's a good sign...
09:13:30Zagornah, same disk failure as once before. I haven't had time to change that disk yet (a backup disk).
09:13:49Zagorso fsck failed and stopped to ask what to do
09:16:21Zagor...and fsck failed because I reboot that box once a month
09:21:57 Join Elbo [0] (
09:22:08 Join SpaceFuzz [0] (i=43aa2316@gateway/web/cgi-irc/
09:23:24Elbois the rockbox site down?
09:23:35JdGordonnot anymore
09:24:02Elbowow, I guess i should have tried one more time before asking
09:25:30 Join davina [0] (
09:26:41 Quit nanok (Read error: 110 (Connection timed out))
09:33:21 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
09:36:57SpaceFuzzHas anyone else been getting [build] Error 2? I'm compiling ipodmini firmware from rev 15390, and seeing "multiple definition of pcm_is_..." errors in pcm_playback.c (and pcm.c).
09:37:25linuxstb_Have you tried a "make clean" ?
09:38:28SpaceFuzzshort answer: no. trying now... ;-)
09:56:33 Join n1s [0] (
09:57:18 Join ender` [0] (
10:01:57n1sjhMikeS: did you figure out why the spc codec broke with Os on coldfire, it might be worth investigating as higher O levels usually give a nice speed up on coldfire
10:04:42 Quit Isolinear ()
10:06:30 Join LinusN [0] (i=linus@gateway/web/cgi-irc/
10:11:51SpaceFuzzOK. had to chmod to get make clean working. Now fresh config & make gives this:
10:12:43 Part LinusN
10:12:43SpaceFuzzsorry. what's pastbin?
10:12:57 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
10:13:03 Part LinusN
10:13:27 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
10:14:06n1sSpaceFuzz: put the text here and post the link instead of filling the channel
10:15:35linuxstb_SpaceFuzz: If you had to use chmod, then something is wrong with your setup. Are you a normal user at the moment, but did the previous build (and maybe the svn checkout) as root?
10:16:18JdGordonamiconn: you round?
10:16:45 Quit LinusN (Remote closed the connection)
10:16:50SpaceFuzzapologies. learning etiquitte.
10:16:58 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
10:17:55SpaceFuzzused the chmod to enable permissions as user, per
10:19:15pixelmaSpaceFuzz: did you also run configure again (sometimes that is also necessary)
10:19:23 Quit Elbo ()
10:19:58n1sSpaceFuzz: do you get any other errors earlier in the build?
10:22:10SpaceFuzznls: yes, I now see ealier ones. copy out of my vmware into Windows browser (ugh) and don't have browser/irc in vmware. 1 min.
10:22:37SpaceFuzzcant copy
10:23:23 Nick fxb__ is now known as fxb (
10:25:25n1shmm, the comment in system-pp502x.c line 76-80 seems to be out of date
10:26:12 Quit TotallyInfected ()
10:26:40n1sSpaceFuzz: have you altered your source or used any patches?
10:26:56 Quit LinusN (Remote closed the connection)
10:27:17 Join LinusN [0] (i=linus@gateway/web/cgi-irc/
10:28:40 Part LinusN
10:29:00 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
10:29:14 Part LinusN
10:29:21SpaceFuzznls: removed sleep function from ata.c. & at one point undef'd cpu_freq_scaling (?). but learned that cpu freqency scaling bug had been resolved, so I undid it.
10:29:36 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
10:29:55SpaceFuzzIf no one else has issues, I'll blow my repository and start afresh.
10:31:27 Join spiorf [0] (
10:32:45n1sSpaceFuzz: undefing cpu scaling is what is causing your probelm
10:34:01n1sbecause the #ifdefing in system-pp502x.c is not complete, the function scale_suspend_core() is #ifdefed out when freq scaling is disabled but some calls to it are not...
10:34:47n1salso if you have build problems you are supposed to try a clean source...
10:35:44SpaceFuzznls: thanks. thought my svn co would overwrite. guess not.
10:36:29 Join karashata [0] (
10:39:12 Quit DataGhost (Nick collision from services.)
10:39:18 Join DataGhost [0] (
10:39:30n1sSpaceFuzz: "svn revert- R ." in the source root will get you back to a clean state
10:39:43n1summ, "svn revert -R ."
10:40:33SpaceFuzznls: bing! thx.
10:49:06 Join MrRichard [0] (i=Tentacle@
10:49:54MrRichardSorry to be such a nuisance but i've been lookinng everywhere and can't work out what the size of the iriver H20s screen is in pixels. Anybody able to help?
10:50:39karashataMrRichard: do you mean the H120 or the 20GB H10?
10:51:06Shaidat the bottom of the page.
10:51:09Shaidgo look
10:51:38 Part linuxstb_ ("Leaving")
10:52:04*karashata blinks slightly
10:52:19karashataapparently they're both the same resolution
10:52:38karashata'cept one's 4 colour greyscale and the other's 16-bit colour
10:53:37MrRichardkarashata: the 20gb one
10:53:50karashataah, 160 by 128
10:54:03MrRichardI never realised it was so samll D:
10:54:06MrRichardThanks for that.
10:56:42 Part SpaceFuzz
10:56:52MrRichardAnyone have any idea why my database always ends up glitching after my 2nd sync?
10:57:05MrRichardon the H20 20gb
10:57:36 Join lee-qid [0] (
10:59:44MrRichardwell, to be specific, after my 2nd sync the database will no longer function. Instead, after selecting database it sits on searching...0 found. And from then on I cannot use the database
11:04:46preglowLinusN: have speex down to 50kb on coldfire now
11:05:16preglowstill 3.5kb in floating point code, i'd expect. plus ogg framing code we don't need, and the entire codec control file
11:05:21LinusNMrRichard: wow, sounds bad
11:05:38LinusNpreglow: good, so there is more to trim then
11:05:58ShaidI had that happen on my 5g ipod, MrRichard, I made it recreate the whole DB, seemed to fix it
11:06:02jmspeexpreglow: what's "the entire codec control file"?
11:06:20Shaidbut it's always possible you have something funny in your db that it doesn't like
11:06:30preglowjmspeex: our speex.c, the layer that handles the interface between rockbox playback and the speex api
11:06:35 Quit BigBambi (Read error: 113 (No route to host))
11:06:46MrRichardshaid: how do I force create a new database? ive just relied on reformating the whole device, which is quite a lengthy and iritating chore
11:06:58jmspeexah, OK
11:07:30karashataMrRichard: the easy way to clear the DB is to delete the database files from the .rockbox directory
11:07:49karashataafter that you can rebuild it from a clean slate
11:08:11ShaidMrRichard: And the Initialize Now option in the Database options menu recreates the DB
11:08:44MrRichardshaid+karashata: thanks a lot for the tips
11:08:59rasherMrRichard: I'm pretty sure you don't need to delete the database files. Just "Initialize Now" will start from scratch
11:09:27MrRichardI just reformatted, I'll play around a bit and try that out
11:09:37preglowjmspeex: btw, would you happen to know anything about libogg2?
11:09:50jmspeexpreglow: a bit, yes
11:10:00preglowjmspeex: does it do less mallocing and memory copying than libogg?
11:10:14jmspeexyes. libogg2 is zero-copy
11:11:06jmspeexWell, it's zero copy for Tremor −− which uses the libogg2 bit-packet. Speex uses its own bit-packer, so there a copy involved.
11:11:39jmspeexThat is, unless you want to change it to use the libogg2 bit-packer.
11:11:50preglowand that would be non-trivial, or?
11:12:11jmspeexpreglow: speaking of bit-packer, you may be able to save on the memory that the Speex one allocates.
11:12:55preglowwow, there's an alloc i didn't even see
11:12:56jmspeexI've never used libogg2, so I can't say how hard it would be. Possibly not that hard. All you'd need to do is reimplement bits.c using calls to the libogg2 bitpacker
11:13:08***Saving seen data "./dancer.seen"
11:14:04jmspeexpreglow: the best thing for you would be to not use speex_bits_init() at all and instead use speex_bits_init_buffer() or speex_bits_set_bit_buffer()
11:14:58preglowjmspeex: how do i know if i can save any memory on it, though? currently it's 2kb, it seems
11:15:09jmspeexpreglow: also note that the zero-copy aspect of libogg2 gives you very little if you consider how small Speex frames are
11:16:05jmspeexpreglow: You can skip the allocation altogether and just use speex_bits_init_buffer(). That will just make the SpeexBits struct use your existing buffer.
11:16:31preglowjmspeex: sure, done, but that still doesn't help me shave any bytes off it by itself
11:16:44jmspeexsorry, I meant speex_bits_set_bit_buffer() which is probably more useful for decoding
11:17:08jmspeexpreglow: If you don't call speex_bits_init(), then it doesn't allocate *any* memory.
11:17:30jmspeexyou just use whatever libogg gave you
11:17:37preglowahhh, right
11:19:43preglowand it's always safe to do that? no bit reservoir or nothing?
11:20:11jmspeexwhat do you mean safe?
11:20:31jmspeexyou've currently extracting the packet and copying it, right?
11:20:47jmspeexAll you need to do is call speex_bits_set_bit_buffer() instead of copying
11:21:04 Quit spiorf (Remote closed the connection)
11:22:16jmspeexIn theory, there is no upper bound for the size of an Ogg/Speex packet constructed by a idiotic asshole, but in practice anything larger than 1200 bytes can be assumed to come from an idiotic asshole.
11:22:35preglowi'm currently looking at that now, i didn't write the entire ogg parsing stuff
11:22:46jmspeexI have Ogg code
11:23:49preglowi hate all code monty writes, but that's primarily because of his spaced out white space usage
11:24:05preglowie. don't use any
11:26:13preglowwhy, oggframing.c seems to comply nicely with that rule
11:27:43preglowbtw, does the enchancer do any comb filtering?
11:31:12MrRichardanother question: when using custom themes that allow coverart to be displayed when playing, where do I save the cover art?
11:31:44jmspeexpreglow: the enhancer is pretty much *only* a comb filter. Why?
11:33:01 Quit J3TC- (Read error: 113 (No route to host))
11:33:06preglowjmspeex: listened to some low bitrate files yesterday and heard some heavy flanging
11:33:24preglowjmspeex: couldn't figure where that happened in encode, so figured it had to be the enhancer
11:33:32markunMrRichard: in the folder of the album I think
11:33:40jmspeexpreglow: can you send a short example (preferably the worst)?
11:33:49markunMrRichard: maybe you can find some more info in the patch tracker
11:33:55jmspeexpreglow: try −−no-enh and see what it does
11:33:57MrRichardpatch tracker?
11:37:13karashataif I recall correctly
11:37:17MrRichardthank you.
11:37:23preglowjmspeex: this has it:
11:37:33preglowjmspeex: but it seems to be present also with −−no-enh
11:38:06jmspeexthen it's not the enhancer
11:38:22jmspeexalso note that uwb q1 is totally useless
11:38:32preglowjust using it as a test file
11:38:34karashataMrRichard: or save yourself the trouble and just go here:
11:38:48karashata^ wiki page on Album Art status
11:39:07 Join desowin [0] (
11:39:16preglowand i wouldn't say it's useless, it's quite intelligible
11:39:25jmspeexpreglow: uwb q1 uses the 4 kbps mode for narrowband.
11:39:54jmspeexThat mode uses an open-loop pitch gain, which isn't very good.
11:40:26jmspeexWhen I say useless, I mean that you can do better for the same bit-rate if you reduce the bandwidth
11:40:38preglowperhaps i like sibilance :)
11:40:39jmspeexDoes -q2 sound better BTW?
11:40:51preglowanyway, the set bit buffer trick worked
11:41:06pregloweven shaved off a few percent cpu usage, for some reason
11:41:10 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
11:41:15preglowthat'd be no copy
11:41:26preglowi don't need to init the bit buffer when i do this, no?
11:41:34jmspeexwhat you save is probably from the reduced pressure on the cache
11:41:44preglowjmspeex: coldfire has no data cache
11:41:58preglowwhich is why using iram is so crucially important on it
11:41:59jmspeexpreglow: you need to make sure you *don't* call init() actually
11:43:20preglowjmspeex: but no, speex has no bit reservoir or anything? each frame is self-contained?
11:44:10 Quit atsea-39 (Read error: 104 (Connection reset by peer))
11:49:08 Join moos [0] (
11:55:09 Quit jhMikeS (Nick collision from services.)
11:55:15 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
11:57:45 Join atsea-39 [0] (i=atsea-@gateway/tor/x-182a0ab7cee28bc4)
11:58:09jmspeexpreglow: no reservoir or anything
12:02:38 Join Bagder_ [0] (
12:06:48jmspeexpreglow: BTW, you're aware that an Ogg packet can contain up to 10 Speex frames, right?
12:06:53 Quit n1s ()
12:10:04preglowjmspeex: not at all
12:10:27preglowjmspeex: but it seems they guy who coded speex.c was, so it's alright
12:18:33MrRichardwhat file extension should video files be for me to play them on my iriver?
12:20:04 Quit Bagder (Read error: 110 (Connection timed out))
12:20:50karashataMrRichard: they need to be MPEG-1 or MPEG-2 format, at a resolution of 160 by 128 or smaller
12:21:09MrRichardoh thank you
12:26:52 Join john [0] (
12:27:29MrRichardThis is starting to get confusing. I need to download a build for this theme to work but the link supplied it dead. Are there any alternatives?
12:29:02 Join mf0102 [0] (n=michi@
12:34:54BigBambiMrRichard: The size the video should be depends on which iriver you have
12:35:24BigBambiMrRichard: And for the build question, it should say what patches you need. You can patch your own build
12:35:41MrRichardI have a list of relevant patches. its just...
12:35:45MrRichardhow do I install them?
12:35:53BigBambiMrRichard: For the video look at the PluginMpegPlayer wiki page
12:36:06BigBambiMrRichard: You patch the source code then compile it
12:36:17BigBambiLook at the SimpleGuideToCompiling wiki page
12:36:37 Join class37 [0] (
12:36:37pixelmaPluginMpegplayer ;)
12:36:43class37good morning guls
12:36:58class37quick question.
12:37:06class37my ipod video screen is dead
12:37:23 Quit john ("Leaving.")
12:37:23BigBambipixelma: Thanks :
12:37:29BigBambi:) even
12:37:43class37is there a way to set up a playlist that will automatically start or something so I can have functionality until my budget allows for a new one?
12:38:29 Join john [0] (
12:38:43BigBambiclass37: Are you using itunes to transfer music?
12:38:55pixelmayou could set everything up on your computer, editing the config.cfg file and/or even use the voice UI...
12:39:24 Join midgey [0] (
12:39:25BigBambiclass37: Right, but how do you get audio files onto the ipod. Drag and drop or itunes
12:39:45class37drag and drop
12:40:14MrRichardIs drag and drop the preferred method?
12:40:24MrRichardAs I've been using windows media player to sync with my iriver.
12:40:39BigBambiclass37: Good - it is then easy to set up a playlist on the pc, then as pixelma says this would seem an ideal use of voice ui to navigate to that playlist and also to set autoresume to on
12:40:58BigBambiMrRichard: You can do it how you want, but personally I will only use drag and drop
12:41:10MrRichardthanks, I've been wondering about it.
12:42:59 Quit SirFunk (Remote closed the connection)
12:43:02BigBambiclass37: And as pixelma also says you can create config.cfg on your pc so you don't have to change settings within rockbox, so all you would have to do is locate the playlist
12:43:10 Quit class37 (Read error: 104 (Connection reset by peer))
12:43:51 Join class37 [0] (
12:43:51 Join J3TC- [0] (
12:43:52preglowwho has the real copyright for vorbis.c? i assume it's not zagor...
12:46:35MrRichardIf a theme says it requires a certain build, but i don't install the build but still put all the patches in. will it work?
12:47:05BigBambiMrRichard: Probably not
12:47:14*MrRichard sighs
12:47:24BigBambiThe WPS parser is quite strict now, so if it finds errors it displays the default WPS
12:47:24MrRichardI have teh patches but all the links to the build I need are dead
12:47:46BigBambiAnd an error can include unsupported tags which some WPS patches introduce
12:47:49class37when i set talk menu: on
12:47:58BigBambiMrRichard: Have you looked in the unsupported builds forum?
12:48:04class37i'm not hearing anything
12:48:11MrRichardI'll have a look, thanksBigBambi
12:48:19BigBambiMrRichard: You don't need the exact same build, just one that includes the patches you need
12:48:31BigBambiclass37: Do you have voice files installed?
12:48:42class37probably not
12:48:48class37got it
12:50:32BigBambiMrRichard: Sorry, I misread your original question - you don't need the exact same build, just one with the right patches, so building your own is fine
12:51:02MrRichardThanks again BigBambi
12:51:16MrRichardI found a build that's just missing 1 or 2 patches. Is there a way to include the ones it's missing?
12:51:17class37copying voice files to ipod
12:51:24class37should be that simple
12:51:43BigBambiclass37: I'm not too sure tbh, I have never used voice
12:51:58BigBambiBut I believe it tells you in the manual what you need to set up
12:52:04class37i'll let u know
12:52:11class37great idea though, thx
12:53:14*BigBambi deflects credit to pixelma
12:53:23MrRichardJust out of interest, would an iAudio build work on an iRiver?
12:53:30BigBambiMrRichard: no
12:53:36pondlifeclass37: Copy it to /.rockbox/langs/english.voice
12:53:45pondlife(or whichever language)
12:53:55MrRichardI noticed the H10 and a certain iAudio are compatible with the same themes. But I guess the build is different
12:54:19pondlifeThemes might be compatible if the screen dimensions are the same, builds won't bt.
12:54:19BigBambiThemes are essentially just text and pictures and rockbox knows how o deal with them
12:54:21 Quit class37 (Read error: 104 (Connection reset by peer))
12:54:33MrRichardI thought that may be the case
12:54:41 Quit midgey ()
12:54:43BigBambiBut the build talks to the hardware, so has to be very specific - different players = different hardware
12:54:58 Join class37 [0] (
12:55:05class37don't see docs?
12:55:18class37for talk menu
12:55:43BigBambiclass37: Which player qgqin?
12:55:56class37ipod video 30 gig
12:56:17*BigBambi curses azerty and qwerty keyboards having q and a swapped, and me having to use both
12:57:28MrRichardWhat a bummer. the only build I can find is missing quite a few patches
12:57:29 Join spiorf [0] (
12:57:39johnHi, I wanted to correct an error in the wiki on the cross compiler page. So I created an account, asking for write permissions here. Alternatively I can just tell the changes here and somebody with write permissions does the job. It is not much.
12:57:50class37 a q q
12:57:54BigBambiclass37: Section 7.9 of the manual and links therein
12:58:02BigBambijohn: Wiki name?
12:58:06 Join Domonoky [0] (
12:59:08BigBambijohn: done
12:59:17johnoh fine. thanks.
12:59:33MrRichardOh man. I have another silly question. Where do I extract themes to?
12:59:53MrRichardThis .zip hasnt got .rockbox as its root, just 'RedEye2' with everything inside that folder
12:59:54BigBambiThe zip should be set up with the correct directory structure so you just unzip to the root
12:59:59BigBambiAh, OK
13:00:07BigBambiwhat is inside that folder?
13:00:17MrRichardwps, themes, fonts, background
13:00:50BigBambiThose folders should be inside .rockbox
13:00:59class37you guys mentioned voice and one other way to handle it
13:01:05class37please restate the other way?
13:01:12MrRichardshould i rename redeye to .rockbox and then extract?
13:01:31BigBambiclass37: when rockbox saves changed from default settings it writes config.cfg inside .rockbox
13:01:40BigBambiMrRichard: that is one way
13:02:22BigBambiclass37: So if you edit that to change rockbox to autoresume on startup and create a playlist on your computer and save it somewhere easy like the root of the player
13:02:25 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
13:02:38 Join miepchen^schlaf [0] (
13:02:55 Quit Genre9mp3 ("I don't suffer from Rockbox psychosis. I enjoy every minute of it.")
13:03:15MrRichardI'd just like to thank BigBambi and everyone else for all the help so far D:
13:03:19MrRichardIts not easy being a newbie
13:03:26BigBambiclass37: The all you have to do once is from the menu structure in the manual know you have to press down twice to get to files (check how mny times it actually is, then select, the select again to load the playlist (if it is the only file n the root)
13:03:37 Join Thundercloud [0] (
13:04:01BigBambiMrRichard: No problem. By the way, if you haven't already done so, we have an excellent manual that you should take a look at
13:04:29MrRichardIve read the manuals.
13:04:32MrRichardlots of them D:
13:05:25MrRichardso far the theme is working flawlessly!
13:05:30BigBambiGood news
13:06:14MrRichardonly letdown so far is that the wps screen is just default.
13:06:15BigBambinot flawless? :)
13:06:30BigBambiOK, so that means there are errors in the WPS file itself
13:06:32MrRichardhaha yeah ;)
13:06:42MrRichardWell it was missing 2-3 patches
13:06:50BigBambiThat'll do it
13:07:02MrRichardNow I have another issue...
13:07:12MrRichardI've already installed a build, lets say I come across the build I need.
13:07:19MrRichardCan I safely just install that build on top?
13:07:20BigBambiThose patches probably introduce new tags, so as official rockbox doesn't know about them it refuses to load it
13:07:24karashataMrRichard: yes
13:07:29BigBambiYep, just overwrite
13:07:39BigBambiThat is also how you upgrade
13:07:40 Quit lee-qid (Connection timed out)
13:07:42MrRichardI have all the patches, now I just need to learn how to compile them =/
13:07:56BigBambiMrRichard: It isn't too bad as long as they are uptodate
13:08:34 Join lodesi [0] (
13:08:47MrRichardgood thing I slept in thismorning. Looks like a late night for me
13:09:08BigBambiKeep in mind that the only supported version of rockbox is the up to date current build (which normally changes several times per day). For patched builds etc., we can' really support as we don't know what the patches might have changed
13:09:37BigBambican' = can't in the previous sentance :)
13:10:44MrRichardohh okay
13:11:24MrRichardso my up-to-date current build may not accomodate the patches I need?
13:12:31karashataMrRichard: as long as the patches have been synced to a recent SVN commit they should work fine
13:12:58karashataif they're a few months old they might need a little work to make them fit with the current SVN
13:13:07MrRichardNow my only remaining issue is not knowing how to compile patches >,< the guide is confusing
13:13:09BigBambiYes, up-to-date current has no patches. You are free to use patches and unsupported builds, but if you have a problem we ask you test with an official build before reporting is all
13:13:10***Saving seen data "./dancer.seen"
13:13:35BigBambiMrRichard: Are you on windows?
13:13:44MrRichardCurrently, yes =/
13:14:02BigBambiFirst step is setting up a build environment
13:14:13BigBambiFor ease of use I would suggest VMWare
13:14:30MrRichardDownloading now
13:15:23MrRicharddo i need all the 7zip etc that the wiki mentions?
13:15:48BigBambiYou need something that can unzip .7z (7zip files)
13:16:16BigBambi7zip itself is free and open source, but if you have another tool to do it you prefer you are welcome to use it
13:16:24MrRichardMightaswell get 7zip
13:16:45BigBambiMrRichard: Possibly, I don't use winrar so don't know if it can do it or not (I'd guess yes)
13:17:24karashatayep, winRAR can open 7-Zip files
13:17:50karashatait's even enabled as the default app on install if you don't change the defaults on install
13:18:18MrRichardoh okay. excellent\
13:19:15MrRicharddidnt realise these files were over 100mb.
13:19:23MrRichardHopefully my net pulls through and gets them doen pretty quick
13:19:32BigBambiMrRichard: It is a whole operating system :)
13:19:57MrRichardI thought it may have been smaller
13:20:07MrRichardso it's basically cygwin, except set up differently?
13:20:17BigBambiIt is a linux install
13:20:21MrRichardoh o,o
13:20:24BigBambiThat runs on top of windows
13:20:29BigBambivia VMWare
13:20:34 Join darkapostrophe [0] (
13:20:46MrRichardAhh I get it.
13:20:51darkapostropheI noticed that there's a new download on the downloads page called Voice
13:20:55darkapostropheWhat is that?
13:21:01darkapostropheSpeech synthesis or something?
13:21:04BigBambidarkapostrophe: voice files
13:21:08MrRichardWell once this is out of the way atleast I can live happily if I can get my iriver running well.
13:21:25darkapostropheBigBambi: Again, what are they, and what are they used for?
13:21:27BigBambidarkapostrophe: Not synthesis, but prebuilt files to voice menus etc
13:21:37darkapostropheVoice files = ambiguous, leaves me guessing
13:21:40karashataMrRichard: you'll like it, I'm sure
13:21:44darkapostropheOh, I see
13:21:50Domonokydarkapostrophe: with this voice files, rockbox can speak its mens..
13:21:51BigBambidarkapostrophe: well, you could have searched the wiki or manual
13:21:58darkapostropheSo that if there's a directory called podcasts, it'll read it out loud?
13:22:03BigBambiFor instance VoiceHowTo
13:22:20BigBambidarkapostrophe: No, that's a directory
13:22:23darkapostropheBigBambi: The wiki is quite strange, it's not easy to search or find stuff in, as MediaWiki is
13:22:32BigBambidarkapostrophe: I managed it
13:22:43darkapostropheOkay, so if I have a file called Artist - Song.mp3, it'll read that, then?
13:22:55BigBambiIt is not synthesis
13:22:57markundarkapostrophe: those are all examples of speech synthesis
13:22:58BigBambiIt is prebuilt
13:23:00Domonokyno it will read the mens..
13:23:11darkapostropheSo the stuff under System, Settings etc...
13:23:13MrRichardI love you 'u's ;P
13:23:28darkapostropheMrRichard: Those are called umlauts
13:23:30Domonokyfor files and folder you would have to generate .talk files.. the rockbox can also speak those..
13:23:32BigBambiTo read filenames etc, you must record yourslef some talk clips
13:23:46BigBambiDomonoky: you in :)
13:23:56darkapostropheOkay... Not really a big deal for me, that was really it. I just wondered about that, thanks. :)
13:24:04*karashata sits quietly
13:24:05Domonokyrbutil can generate those talk files, if you whish.. :-)
13:24:16BigBambiDomonoky: That's cool
13:24:25BigBambiI haven't tried RBUtil in a while
13:26:14*Domonoky would like to have more feedback on rbutil.. :-)
13:26:48BigBambiI shall try to next time, but I either build my own or just unzip
13:27:00BigBambiAnd don't use voice etc
13:27:23karashataI've used it once, it works alright but I find it easier just to download the .zips and install manually
13:27:32karashatasince that's what I'm used to
13:30:35 Quit pill (Nick collision from services.)
13:30:45 Join pill [0] (
13:32:26 Nick darkapostrophe is now known as Dark_Apostrophe (
13:34:14 Quit pill (Nick collision from services.)
13:34:30 Join _pill [0] (
13:36:28MrRichardone of the download links from just fried my mozilla D:
13:36:50MrRichardDang. I had so many things i needed open.
13:37:06BigBambiI doubt it was the download
13:37:12BigBambiIt is just another website
13:37:24MrRichardIt was just the website that my computer didnt like
13:37:48BigBambiIt was more likely to be having lots of tabs open, but this is getting a little off topic
13:38:12MrRichardSorry x)
13:38:27MrRichardIn any case, I have the VMware player
13:38:33MrRichardIm just downloading the debian file now
13:38:39BigBambiOK, cool
13:46:07pondlifeDomonoky: Is there a new RBUtilQT build around?
13:46:50Domonokyno.. but we should make a new release, since i fixed some big bugs.. :-)
13:47:02pondlifeIf you could, I'll see if I can get talk files working
13:47:16pondlifeThe wiki is a bit confusing
13:47:33 Quit Thundercloud (Remote closed the connection)
13:47:37pondlifeYou have ticks under 1.0.2 for fixes not in 1.0.2.
13:48:05pondlifee.g. "Talkfile creation puts the file in the wrong place." will be in 1.0.3, right?
13:48:30pondlifeBut not in 1.0.2
13:48:36pondlife(at least, not here)
13:48:42 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
13:48:47Domonokythe page is a bit confusing... feel free to change..
13:49:15pondlifeI'm not sure what it should be. Should the 1.0.2 read 1.0.3 etc.?
13:49:31Domonokythe notes under a version, are bugs found in this version.. if the are fixed in svn, the get a tick.. .:-)
13:49:37pondlifeAh, ok
13:50:14pondlifeI'll let you do 1.0.3 first.
13:50:44Domonokyi could build a svn windows binary for you, if you want..
13:51:06pondlifeYes please, if you're not planning on doing 1.0.3 soon anyway
13:51:20pondlifeProbably better to do a 1.0.3 though
13:51:54pondlifeSee if the Sansa problem is resolved
13:52:17MrRicharddamn... the download jumped to 1 hour. And I've got work in the morning x)
13:52:22Domonokyyes, but i dont have qt correctly configured, to make a static version of rbutil...(without the need for mingw.dll) :-/
13:52:27 Join Cliffer [0] (
13:52:29pondlifeAh, ok
13:52:45pondlifeI'll get editing then.. and testing a windows build should it appear ;)
13:52:51Clifferhow do i detect if i have an ipod video 5.5 or 5?
13:52:52Domonokythe last releases are all made by bluebrother... . :-)
13:53:22tuplanollais there any way to change the low battery warning threshold in the bootloader?
13:53:54tuplanollabecause my sansa said 40% left and it won't start up anymore without charging
13:54:10linuxstbCliffer: If you run ipodpatcher, it will tell you the sector size - 512 bytes is 5g, 2048 bytes is 5.5g.
13:54:25 Quit jmspeex (Remote closed the connection)
13:55:42 Join Arathis [0] (
13:58:27 Join jmspeex [0] (
14:00:53MrRichardThis download is going to take a while. I'm going to take a nap. Thanks everyone so much for the help
14:03:56pondlifeHmm, how does the MajorChanges page use %E% to get that enhancement graphic?
14:04:15pondlifeI'm attempting to make RockboxUtilityQt do the same.
14:05:06 Join darkapostrophe [0] (
14:06:50JdGordonpondlife: arnt they just local images? rbutil should probably just have those 3 images built in also then
14:07:07pondlifeI need to attach the graph one I think
14:08:24pondlifeDomonoky: Does that look ok to you?
14:08:54 Nick parafin|away is now known as parafin (
14:10:15pondlifeClear enough?
14:10:23pondlifeI may have misunderstood!
14:12:11 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
14:18:14 Join [1]seb [0] (
14:18:54JdGordonamiconn: you round?
14:19:14 Nick Bagder_ is now known as Bagder (
14:19:20[1]sebhow long does it take to write 2gb to a microsd via e200?
14:19:52 Join Thundercloud [0] (
14:20:13JdGordonhow long is a piece of string?
14:20:20Domonokypondlife: looks nice..
14:21:46 Quit [1]seb (Client Quit)
14:22:21 Quit Dark_Apostrophe (Read error: 110 (Connection timed out))
14:22:21 Nick darkapostrophe is now known as Dark_Apostrophe (
14:34:11 Join Arathis2 [0] (
14:35:34 Join Rondom [0] (
14:38:19 Quit moos ("Glory to Rockbox")
14:40:45 Quit mf0102 ("Verlassend")
14:41:45JdGordon2 bugs fixed... whos going to challenge me in my half asleep state of awakedness..?!
14:44:31LinusNoh, you're at work too? ;-)
14:45:39JdGordonno.. i should be in bed
14:47:52 Quit Arathis (Read error: 110 (Connection timed out))
14:48:36LinusNi get that feeling all the time :-P
14:48:52pondlifeJdGordon: for number 3 ??
14:49:18pondlifeSeems to be a button action thing.
14:49:37JdGordonmy brother has my h300, but ill hae a quick look
14:49:53pondlifeYou are allowed to sleep, should you prefer :)
14:50:50JdGordonor maybe not... that looks like im definatly going to need the h300 to debug
14:51:19 Nick Arathis2 is now known as Arathis (
14:51:24 Quit class37 ()
14:56:14 Quit Toki_ (Read error: 110 (Connection timed out))
14:58:57JdGordonis fs#6129 fixed? I just checked and think mob may have got it
15:01:14LinusNpondlife: does it only happen with the nonlcd remote and not with an lcd remote?
15:02:00LinusNpondlife: and what exactly is "Toggling the remote"=
15:02:47 Join stacks [0] (
15:02:54JdGordonis there any reason why 7218 (coloured sliding puzzle) was never commited?
15:03:07stacksis there an autoplay
15:03:13LinusNJdGordon: my guess is lack of interest
15:03:20LinusNstacks: autoplay?
15:03:20stackslost use of my screen
15:03:37 Join scorche|w [0] (n=8dc5049d@rockbox/administrator/scorche)
15:03:59LinusNstacks: you can make it resume playback automatically, if that is what you mean
15:04:24stackst see to change directories
15:04:43stackswould like a way to start a playlist without having to navigate
15:05:16hcsyou could name it AAA_AARDVARK.m3u
15:05:21LinusNyou can generate a playlist with your pc, and then play it
15:05:50stacksok so I make a play list aaa_aaa.m3u
15:05:55LinusNbut you will have to navigate blindly to play it in the first place
15:05:57stackshow do I get it to play
15:06:14hcswell, the idea is that it's the first in the list, so you just hit select once
15:06:18LinusNafter that, you can resume it by pressing play after starting rockbox
15:06:32LinusNor edit the config file to set the WPS as the startup screen
15:06:56LinusNyou can even manually edit the .playlist_control file to make it play your playlist
15:07:12pixelmaif you set filenames to spell, you can find that .m3u with the voice UI
15:07:28LinusNpixelma: good idea
15:07:45pixelma(or prepar voice clips, but that's a bit more involved)
15:10:10JdGordonlatest aussie beer ad
15:12:55stacksis there a step by step for voice?
15:13:14***Saving seen data "./dancer.seen"
15:15:29 Join Spliffer [0] (
15:15:45stacksthanks guys
15:15:47 Quit stacks ()
15:19:48 Quit J3TC- (Read error: 110 (Connection timed out))
15:19:55JdGordonwould anyone object to adding ata_disk_is_usable() which would be true if the disk is still spinning? (or always true on flash targets) as apposed to ata_disk_is_active() which is true if the disk is actually being used) ?
15:20:33 Join J3TC- [0] (
15:21:19JdGordonactually ignore that
15:21:21*JdGordon going to sleep
15:21:24 Quit JdGordon ("Konversation terminated!")
15:25:33 Quit spiorf (Read error: 104 (Connection reset by peer))
15:25:45 Quit Cliffer (Read error: 110 (Connection timed out))
15:26:09 Join spiorf [0] (
15:27:49 Join The-Compiler [0] (
15:27:54 Quit pondlife ("Read error: 110 (Connection slimed out)")
15:28:06scorche|wquery scorche
15:28:55 Part pixelma
15:36:50 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
15:37:02Buschelhi, can anyone give write permission to wiki? i'd like to add the current mpegplayer framerates for ipod video.
15:39:09LinusNhang on
15:40:20LinusNBuschel: done
15:40:46Buschelworsk, thanks :)
15:41:19DomonokyLinusN: would it be possible to provide the features file alongside the voice files on the rockbox servers ? ( to make rbutil generate voice files :-) )
15:44:06LinusNDomonoky: interesting idea
15:45:17BuschelLinusN: finished
15:46:04LinusNDomonoky: i'll see what i can do about that. i'll have a chat with daniel
15:49:07Domonokypondlife: could you test this rbutil version : domonoky/"> ? i hope i have now removed the mingw.dll depency on my build system.. :-)
15:49:27 Quit hcs ("Leaving.")
15:50:33pondlifeDomonoky: Thanks, will play
15:51:35pondlifeDon't know if it's been suggested, but a button to hear the configured voice talk would be useful... I guess RBUtil audio output wouldn't be simple though.
15:52:05pondlifei.e. To check your configuration produces the expected voice before you make the files
15:52:47Domonokypondlife: would be nice.. and shouldnt be to hard, you just have to change the parameters to the tts engines..
15:53:28pondlifeWell you can easily produce the MP3 (or speex?) file, it's playing it back that would be new
15:53:47Domonokythe other (possibily easier) way would be to just generate a short wav, and play it via Qt s sound functions..
15:53:48pondlifeOr do all our supported TTS systems have an audio out option#
15:54:05 Quit J3TC- (Read error: 110 (Connection timed out))
15:54:09pondlifeOK, although it doesn't test the encoder config
15:54:17 Part LinusN
15:54:31 Join J3TC- [0] (
15:54:33Domonokyso playing a generated file would be the best..
15:55:49 Join J-23 [0] (
15:56:10pondlifeOne other improvement would be if the voice generation progress bar really showed progress - I suppose a prepass to count dirs/files would be too slow though.
15:56:22Domonokyhm, Qt can only play wav, on all plattforms..
15:56:33pondlifeOK, wav it is then!
15:57:53Domonokyif you have a good idea how to calculate the progress, without first reading all dirs and files... tell me :-)
15:58:44pondlifeNo, you'd have to preread
16:00:16linuxstbDomonoky: Do you mean it can only play .wav files from disk, or is there an interface to play uncompressed PCM? If the latter, we could incorporate some codecs...
16:01:28pondlifeDomonoky: p.s. Was that a debug (symbols) build of RBUtil? It was 7MB, as opposed to the previous 2.5MB...
16:02:04pondlifeOn another topic, what's the downside for H300 users having "Reduce Ticking" enabled?
16:02:13Domonokyi think it can only play wavfiles from disk..
16:02:20 Join toffe82 [0] (
16:02:29Domonokypondlife: ah forgot to ucl compress :-)
16:02:30 Quit Zagor ("Client exiting")
16:04:01 Join Crash91 [0] (n=evil91@
16:05:46 Quit Crash91 (Client Quit)
16:07:13 Join Crash91 [0] (n=evil91@
16:09:38 Join TQuid [0] (
16:10:10TQuidAhm, anyone know how to reboot when you are stuck on a data abort and the keys don't respond?
16:10:35scorche|wthat depends on which device you have
16:10:45TQuidSansa, that is.
16:10:51scorche|whold the power button
16:10:57tuplanollafor 15 sec
16:11:14scorche|wholding it till it resets will work ;)
16:11:16TQuidTaht's a bit of a "duh."
16:11:37TQuidJust not used to having to hold it that long. Now if I could just remember what the hell I was doing.
16:12:02 Join PaulJam [0] (
16:12:21TQuidSo is the infacility with USB just a Sansa thing or is that something not yet fully developed in rockbox generally?
16:12:36 Quit J3TC- (Read error: 110 (Connection timed out))
16:13:07TQuidI also find that Amarok can't quite see the device, with the OF, but I guess I'll bug the Amarok guys about that. I think it's getting in a fight with KDE somehow.
16:13:41TQuidDidn't help that Sansa lies like a rug on the casing−−said I had an e280r but not true! :)
16:14:00 Quit MrRichard (Read error: 110 (Connection timed out))
16:14:25scorche|wit isnt developed in rockbox for all PP devices
16:14:55scorche|w*yet* :)
16:15:02Splifferhow can i rebiuld the database?
16:15:43scorche|wthe manual should describe that
16:15:46TQuidI have heard it's close! Exciting.
16:16:56TQuidIt would be fun to be able to operate without using the OF at all.
16:18:30 Join Toki [0] (
16:19:14TQuidOh sweet.
16:19:19TQuidThat db thing is hyper awesome.
16:19:31 Quit Toki (Read error: 104 (Connection reset by peer))
16:19:39TQuidWhy the hell don't the vendors think of obvious cool stuff like this?!
16:20:57TQuidanyway, again, my humble thanks. Gotta get to work now.
16:21:03 Quit TQuid ("Leaving")
16:26:46 Join Toki [0] (
16:27:37 Quit scorche|w ("CGI:IRC (EOF)")
16:27:43 Quit The-Compiler ("Verlassend")
16:28:42 Join webguest31 [0] (i=5848e9b8@gateway/web/cgi-irc/
16:30:03 Quit in-jane__ (Read error: 104 (Connection reset by peer))
16:30:53 Join scorche|w [0] (n=8dc5049d@rockbox/administrator/scorche)
16:32:15 Join J3TC- [0] (
16:34:08 Quit webguest31 ("CGI:IRC (Ping timeout)")
16:34:18 Join in-jane [0] (i=jane@
16:35:11 Join jgarvey [0] (
16:35:12 Join Soap__ [0] (
16:37:27 Quit linuxstb ("Client Exiting")
16:38:26 Join webguest18 [0] (i=5848e9b8@gateway/web/cgi-irc/
16:40:16 Nick fxb is now known as fxb__ (
16:43:52 Quit webguest18 ("CGI:IRC (Ping timeout)")
16:43:56 Join MajorC [0] (
16:44:05 Quit Rondom (Read error: 110 (Connection timed out))
16:44:47 Join kubrick [0] (n=repulse@unaffiliated/funky)
16:44:50 Join Rondom [0] (
16:52:16 Quit Soap (Read error: 110 (Connection timed out))
16:52:50 Quit Crash91 ("Bye Bye!")
17:07:38Domonokypondlife: do you have any idea what the tts test button should say ?
17:11:00 Part J-23
17:11:03 Quit toffe82 (Remote closed the connection)
17:13:00 Join toffe82 [0] (
17:13:15***Saving seen data "./dancer.seen"
17:22:19Domonokypondlife: could you test again ? now with tts test button, in the talk file generation screen : domonoky/">
17:23:39 Join midgey [0] (
17:24:08 Join sdoyon [0] (
17:24:21pondlifeDomonoky: Yes, sorry -was away
17:25:21pondlifeClicking Test TTS doesn't do anything..
17:26:04Domonokyreally ? , for me it speaks something..
17:26:17pondlifeI'm using SAPI5.
17:26:23*Domonoky to..
17:26:58pondlifeWindows Control Panel is talking ok.
17:27:17 Join spaceinvader [0] (n=server@unaffiliated/spaceinvader)
17:28:03 Quit midgey (Client Quit)
17:28:28pondlifeAny diagnostics I could try?
17:29:23spaceinvaderim looking for a device to play ogg and flac, and i dont really care about colour/video
17:29:32spaceinvadercan anyone suggest anything?
17:29:41spaceinvaderat least a couple of gig storage preferable
17:30:20ATravelingGeekspaceinvader: I have a Gigabeat and a Sansa and am happy with both, and they both with rockbox should play ogg/flac
17:30:35Domonokypondlife: try domonoky/rbutilqt.exe">
17:31:08Domonokyspaceinvader: take a look at:
17:31:24pondlifeDomonoky: Nope, not talking.. :(
17:31:57Domonokystrange, so qt sound doesnt fail, but no output... :-/
17:32:11spaceinvadersansa looks nicer :P
17:32:37Domonokypondlife: are sure its not your mixer steeings ?
17:32:47pondlifeWell, default WAV is audible
17:33:02pondlifee.g. if I go into Control Panel and test speech there
17:34:07pondlifeI do have a spare stereo output from my audio card, but there's no reason it should use that.
17:34:19Domonokystrange, but why does it work for me ? could someone else here test it please ?
17:34:58 Join alexxx [0] (n=alexxx@
17:35:10alexxxmuie muie muie
17:35:54 Part alexxx
17:36:07spaceinvadersansa e200 looks ok
17:36:18spaceinvaderor perhaps a used h100
17:36:23 Nick fxb__ is now known as fxb (
17:37:06ATravelingGeekspaceinvader: I have the sansa e250 (2Gb)
17:37:15pondlifeDomonoky: Does the Qt audio interface have anyway to specify a particular output channel/device?
17:37:29pondlifeI have multiple audio cards
17:37:31ATravelingGeekbut I know they exist up to at least 8Gb internal storage w/ miniSD card slot
17:37:46krazykitit is NOT a minisd, it's a microsd slot
17:37:53tuplanollae280 ftw
17:37:56spaceinvaderhmm "No FM-tuner is available in Europe"
17:37:58krazykitthis is important, as they're different form factors
17:37:59 Quit Rondom (Read error: 110 (Connection timed out))
17:37:59ATravelingGeekkrazykit: er, thank you
17:38:08ATravelingGeekkrazykit: yes, my mistake
17:38:16Domonokypondlife: no, no settings for QSound, its very basic..
17:38:39pondlifeI have Windows restricted to the one output anyway
17:38:40krazykitATravelingGeek, just wanted to make it clear, so no one goes out and buys the wrong card :)
17:38:56ATravelingGeekkrazykit: good point
17:39:08rasherATravelingGeek: and those will only be usable in Rockbox (Sansa firmware doesn't support microsdhc which is anything >2GB)
17:39:12ATravelingGeekAlthough the last two microSD cards I've bought have come with a miniSD and SD adapter
17:39:33spaceinvaderwhere is the sandisk uk site? still has US prices..
17:39:52ATravelingGeekrasher: the 8GB was referring to internal Sansa storage
17:40:02rasherATravelingGeek: ah, like that.
17:40:46ATravelingGeekWhich will hopefully have usb support sometime soon, since the sansa firmware has weird issues with OS X tiger... haven't tried it with Leopard
17:43:28spaceinvaderATravelingGeek: but its fine with rockbox? :P
17:43:39Domonokypondlife: i just read in the Qt docs: if QSound fails on windows, it will fail silently, no error.. :-/
17:44:03pondlifeAny config (or registry) to enable Qt diags?
17:44:16ATravelingGeekspaceinvader: Totally. I love my super-top-secret-usb-rockbox build :p
17:44:46Domonokypondlife: i dont think there is any diagnostics.. :-/
17:45:31pondlifeDo you check isAvailable()?
17:45:55tuplanollathe lack of usb doesnt really matter
17:45:58Domonokyyes... but that always returns true on windows.. see qt docs.. :-/
17:45:58pondlifeAh - always true on Windows
17:46:34tuplanollabecause if you plug the player in when it's not turned off, it'll start the OF
17:46:54pondlifeDomonoky: Which OS are you using? I'm on Win2K here
17:47:04rashertuplanolla: really annoying to have to reboot all the time though
17:47:07rasherand slow..
17:47:14pondlifeGot XP around somewhere though - will test on that too
17:47:24Domonokypondlife: i am on winXp .. maybe i should try the other method and let the tts speak directly..
17:47:42pondlifeIf that's not too hard
17:47:46ATravelingGeektuplanolla: Well, it's a bit of a hassle to get the sansa firmware to work with osx
17:48:15tuplanollaum, osx, what if you put the player to msc mode?
17:49:20ATravelingGeekWell, that works, once. Then, inorder to get Tiger to recognize it, I have to switch USB modes from MSC to the other mode and then back to MSC for it to be recognized
17:49:29ATravelingGeekand I'd check with Leopard but I left my cable at home
17:49:31 Join illissius` [0] (
17:49:41tuplanollaok, that sounds a bit weird
17:49:50Domonokyhm, letting the tts speak directly is more work.. would have to rework the tts handling in rbutil a bit.. :-/
17:50:18ATravelingGeekI'll try it again when I get home
17:56:15 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
17:57:15 Join midgey [0] (
18:08:35 Quit Cronoh (Read error: 110 (Connection timed out))
18:11:57Spliffercan i copy itunes playlists to rockbox?
18:12:15Splifferdont say its in the manual :)
18:13:19 Quit illissius- (Read error: 110 (Connection timed out))
18:13:50Splifferi know where to store them, but can rockbox convert the itunes playlists?
18:14:10BigBambiAny people with forum modding powers around?
18:14:28BigBambiIf so could someone change this guys name, I don't think it is appropriate;u=13870
18:15:34sdoyonAre there any freely redistributable short audio clips that I could use? I'm looking for very short beep-like sound effects, with a little more personality than pcmbuf_beep...
18:16:06ATravelingGeeksdoyon: you might be able to use audacity to create beeps
18:16:19rashersdoyon: perhaps gnome or kde has some sound-bites?
18:16:42sdoyonrasher: Hmm.. right...
18:16:55 Join SirFunk [0] (
18:17:42sdoyonATravelingGeek: I suppose... :-)
18:24:42 Join Alonea [0] (
18:27:18lostlogicNico_P: <3
18:28:07lostlogicNico_P: I hope my commit message and comments in here lay plain the issue that I hacked around, the proper solution was non-obvious to me and I had to give up
18:28:44Nico_Plostlogic: yeah, I need to look at it... haven't had much time today
18:29:25Nico_PI'll be back in 5 minutes
18:29:28lostlogicNico_P: understandable −− seems you've been kinda busy lately, was part of why I decided to commit the hack, since it does solve the playback stopage w/o horrid side effects.
18:29:55 Nick fxb is now known as fxb__ (
18:32:41rasherDoes sdhc work in the sansa c200?
18:39:45 Quit J3TC- (Read error: 110 (Connection timed out))
18:40:10 Join J3TC- [0] (
18:44:40 Join J-23 [0] (
18:56:18 Quit keanu (Read error: 110 (Connection timed out))
18:56:34 Join keanu [0] (n=keanu@unaffiliated/keanu)
18:58:02 Join mf0102 [0] (n=michi@
19:00:10 Join Robin0800 [0] (
19:02:11 Join bertrik [0] (
19:02:16Robin0800Couple of things that appear to be broken now 1) Replay Gain and 2) Go to wps on startup is this due to the new buffering?
19:04:23lostlogicRobin0800: both of those things work fine for me, are you on the latest build? what player?
19:05:45Robin0800Video ipod
19:05:58lostlogicand latest build?
19:06:02lostlogic(I have an ipod video as well)
19:06:08Robin0800Build from two days ago
19:06:25Robin0800Could it be album art?
19:07:05lostlogicwait... you have album art patched onto a new buffering build!? I know nothing about that patch.
19:07:45 Nick fxb__ is now known as fxb (
19:11:05 Join adel [0] (n=zizo@
19:12:06 Part adel
19:13:16***Saving seen data "./dancer.seen"
19:14:33 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
19:16:10 Quit midgey ()
19:16:26 Join nanok [0] (n=nanok@
19:16:29 Join midgey [0] (
19:16:33 Quit sdoyon ("ircII EPIC4-2.4 -- Are we there yet?")
19:17:00 Join webguest67 [0] (i=53c37844@gateway/web/cgi-irc/
19:17:03pixelmarasher: microSDs don't work on c200 at all yet
19:19:33 Join ackbahr [0] (
19:20:22 Quit ackbahr (Remote closed the connection)
19:22:28 Quit webguest67 ("CGI:IRC (Ping timeout)")
19:24:21 Join iriver120 [0] (i=84d84220@gateway/web/cgi-irc/
19:24:37iriver120anyone know if the current build for the H120 has some bugs in it?
19:24:52iriver120mine cant play music right now after i flashed to this firmware
19:25:34iriver120I get the I04:IllInstr at 32EC0000 error
19:26:45 Quit PaulJam (".")
19:27:53 Join linuxstb [0] (
19:28:36bertriklostlogic: your latest commit seems to have stopped FS #8037
19:29:02bertrik(from occurring)
19:29:33 Join darkapostrophe [0] (
19:30:09 Quit Guile`` ("KVIrc 3.2.0 'Realia'")
19:31:42 Join Guile`` [0] (n=Guile@
19:33:03 Quit J-23 ()
19:35:06 Quit iriver120 ("CGI:IRC (EOF)")
19:35:25 Quit Alonea ("ChatZilla [Firefox]")
19:36:34 Join TotallyInfected [0] (
19:38:09 Quit TotallyInfected (Client Quit)
19:39:30 Join TotallyInfected [0] (
19:41:01 Quit keanu (Read error: 110 (Connection timed out))
19:41:44 Join keanu [0] (n=keanu@unaffiliated/keanu)
19:46:07 Quit Dark_Apostrophe (Read error: 110 (Connection timed out))
19:46:08 Nick darkapostrophe is now known as Dark_Apostrophe (
19:51:55*midgey notices that a significant amount of the log seems to be missing...
19:52:34linuxstbYes, was down for a few hours last night.
19:52:59midgeyah, nothing gets saved during that period?
19:53:07linuxstblogbot runs on that server IIUC
19:53:17*scorche|w nods
19:53:26midgeywell, then that makes perfect sense why it'd be gone
19:53:36nanokhi there
19:53:37scorche|wi can likely take from my logs and toss it over though
19:53:49nanokseem the mp3-ogg bugg is squashed
19:54:43midgeynanok: bertrik reported the same thing about 30 min ago, so thats two reports of the bug being gone
19:59:27nanokmidgey: yup. that sounds good
19:59:32markunthere is not 80GB ipod 5.5g, right?
19:59:45pixelmasure there is
20:00:04bertrikbut is the bug really fixed or did it just disappear?
20:00:10nanokdo we know how it happened, though? or it just got..swiped away somehow, with other changes? :)
20:00:17nanokbertrik: :)
20:00:30markunpixelma: ok, thanks
20:01:35bertrikthis bug triggered me to look into playback.c and I noticed some fishy things there
20:01:57linuxstbmarkun: 5g was 30 or 60GB, 5.5g was 30 or 80GB
20:02:31 Join obo [0] (n=obo@rockbox/developer/obo)
20:02:36Nico_Pnanok, bertrik: talking about FS #8037?
20:02:41markunlinuxstb: strange, I wonder why not 40 and 80GB
20:03:23pixelmamarkun: that's the one which took longer to support - amiconn made it work shortly after devcon with the help of LinusN's 80GB
20:03:40bertrik15390 made it disappear
20:04:08nanokNico_P: yes
20:04:19Nico_PI'm looking at the commit right now
20:04:31markunpixelma: yes, I remember now
20:06:48Nico_Plostlogic: still here?
20:09:22 Join petur [0] (
20:09:36BigBambiIs it just me there are no images loading in the forums for?
20:10:52 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
20:11:02 Quit Thundercloud (Remote closed the connection)
20:17:21 Join donutman25 [0] (n=chatzill@
20:21:53linuxstbSeems fine for me, but maybe they're cached.
20:22:43BigBambilinuxstb: Cheers for looking. For some reason a hard refresh did nothing, but closing and reopening the tab worked
20:23:41 Quit scorche|w ("CGI:IRC (EOF)")
20:27:41nanokNico_P: thanks for the cleanup. terribly sorry, i will be more careful with refresh on that page
20:27:45 Join low_light [0] (i=c730180b@gateway/web/cgi-irc/
20:27:54Nico_Pyou're welcome ;)
20:29:07 Join Crash91 [0] (n=evil91@
20:29:23 Quit MajorC ()
20:29:52 Quit Buschel (Remote closed the connection)
20:30:19 Quit Robin0800 (Read error: 104 (Connection reset by peer))
20:30:49low_lightlinuxstb: do you understand how the MMAP stuff for pp5020 works? I'm looking for an interpretation of MMAP2_LOGICAL = 0x36003f00; MMAP2_PHYSICAL = 0x37002f84
20:31:13Crash91rockbox accepts mpeg format right? is that mpeg 1 or 2
20:31:25Nico_PCrash91: both
20:31:33 Quit donutman25 ("ChatZilla [Firefox]")
20:32:01linuxstblow_light: No, I don't understand them. Have you checked MrH's documents?
20:32:31preglowthat's the finest source of documentation you'll find on the area, just check that
20:33:06low_lightI'll look
20:34:34nanoki have noticed something "strange". it seems older builds are a bit hungrier for power on the sansa than the newer ones
20:34:52Nico_Pnanok: older how?
20:35:08nanokespecially the pre MoB ones, it seems. is this normal, or am i just imagining it :)
20:35:56nanokNico_P: i understand there was a major change, regarding buffering i think, nickamed here mob
20:36:09BigBambinanok: Nico_P wrote it
20:36:14Nico_Pnanok: yes, I committed it a few days ago
20:36:18 Quit Crash91 ("Bye Bye!")
20:36:22nanoka build from 27/10 (just fter that change), iirc, seemed a bit more hungry
20:36:38BigBambinanok: Unless you have proper battery benches from before and after there is no way of knowing
20:36:39nanokone before that, that i tryed today, seemed even hungrier
20:36:51BigBambi'seemed' doesn't really mean anything
20:37:00Nico_PI'd be very interested in comparative battery benches
20:37:06Nico_Pespecially on HD targets
20:37:07nanokBigBambi: i do not, that is why i am asking. it is just the impression after "normal" usage
20:37:43BigBambinanok: I'm glad it seems better, but without tests it'll just be a feeling (albeit a nice one)
20:37:49nanokNico_P: my rockbox runs on a sansa (to put things in the right order ;) ), but i can do some benches, it would be interesting indeed
20:38:01 Join Crash91 [0] (n=evil91@
20:38:44Crash91mpegplayer give me a missing start packet error (or something like that) when i try to play a particular mpg
20:39:05Crash91should i try using mediacoder to reconvert it?
20:39:13BigBambiNico_P: I'm afraid I won't be doing them anytime soon even though I'd like to see the results, I upgraded the battery in my h100 and it lasts ~ 30 hrs, and the gigabeat I use daily :)
20:39:14linuxstbWhat format are your mpeg files? They need to be Program Streams, not avi
20:39:47Crash91linuxstb: no idea...its a quicktime trailer i got from my friend...mpg so i guessed it would work
20:40:19nanokBigBambi: that's exactly my problem: i am addicted to it, i use it daily now, all day. really hard to make benchmarks (i could if i had a spare battery, i guess, during the night)
20:40:34Nico_PBigBambi: don't worry about it ;)
20:40:41linuxstbCrash91: Have you read the PluginMpegplayer wiki page?
20:41:04Crash91linuxstb: one upon a time
20:41:24BigBambiCrash91: You sure it isn't q mov with the wrong extension?
20:41:30linuxstbDon't you think a re-read would be helpful? ;)
20:41:37Crash91could be...ill use mediacoder anyway
20:42:12 Join darkapostrophe [0] (
20:42:53Crash91:O its lagging for some reason
20:50:48 Join Lear [0] (i=chatzill@rockbox/developer/lear)
20:51:36 Quit Crash91 ("Bye Bye!")
20:52:41LearDomonoky: The Sansa fix for rbutil should be identical for the ipod. Don't have any ipod to test with though... I could make a blind commit, as the change is small.
20:54:47Domonokylear: would be good if someone could test it..before commiting :-)
20:58:24 Join FOAD_ [0] (
20:59:34 Quit Dark_Apostrophe (Read error: 110 (Connection timed out))
20:59:34 Nick darkapostrophe is now known as Dark_Apostrophe (
20:59:43linuxstbLear: Do I understand correctly that the problem is that rbutil is being built as unicode (meaning CreateFile is #defined to be CreateFileW), but that sansapatcher is using ANSI filenames?
21:00:56preglowthere's not much point in not using unicode on windows anymore
21:01:32*Domonoky thinks that rbutil switched to unicode, sometime ago, to complety support l10n
21:03:09linuxstbIMO, sansapatcher and ipodpatcher are so similar, I would be confident with a blind commit.
21:03:19 Join scorche|w [0] (n=8dc5049d@rockbox/administrator/scorche)
21:03:34Domonokyif linuxstb is confident, lear should commit it :-)
21:03:37lostlogicbertrik: that's interesting, wouldn't have 'spected it.
21:03:41lostlogicNico_P: am back now
21:04:12Nico_Plostlogic: cool. the issue was that buffering of the codec would be interrupted, wasn't it?
21:04:37lostlogicNico_P: no, of the file
21:04:39lostlogic(audio file)
21:05:06lostlogicat least that was the particular one that I was resolving
21:05:11lostlogicbut codec could also trigger it now that I think of it
21:05:22*bertrik looks into buffering.c
21:05:22lostlogicany time there was a request for data larger than file_chunk it could happen
21:05:30lostlogicwhich spc, nsf an codecs will do
21:05:56 Quit kubrick ("leaving")
21:06:09linuxstbHow big is file_chunk?
21:06:23lostlogic(on targets with > 1m ram iirc)
21:06:32Nico_Plostlogic: but the request doesn't say the required size
21:06:37linuxstbThat's odd - I thought codecs were written to only expect a maximum of 32KB.
21:06:45 Join kubrick [0] (n=repulse@unaffiliated/funky)
21:06:57lostlogiclinuxstb: nsf and spc violate that if so, which I'm pretty sure could lead to future bugs with wrapping the buffer
21:07:02linuxstb(i.e. GUARDBUF_SIZE)
21:07:05lostlogiclinuxstb: because guardbuf is only 32k :)
21:07:17Nico_Plostlogic: these formats can't wrap IIUC
21:07:31Nico_Pand they need the whole file to be present
21:07:52lostlogicNico_P: well currently nothing prevents them from wrapping afaics
21:08:17 Join mrkiko [0] (
21:08:23mrkikoHi all!
21:08:23Nico_Pthat's true. they should be bufopened with a special type
21:08:24 Quit nanok (Read error: 104 (Connection reset by peer))
21:08:32linuxstbHmm, nsf just requests the entire file - so it's not obeying the current API...
21:08:44lostlogiclinuxstb: spc does same according to hcs
21:09:12linuxstbHow big are nsf and spc files ?
21:09:29 Join nanok [0] (n=nanok@
21:09:32nanokit seems that bug that caused mp3's to stop in the middle of playback rubbed off on ogg's also
21:09:38nanokwith the latest build..
21:10:14nanokanybody noticing the same (i've only been playing with it for an hour or so, and it just happened for the first time)
21:10:23Nico_Plostlogic: what I don't get is what was preventing the buffering from doing more than a filechunk?
21:10:35hcslinuxstb: where is the API documented such that it says not to request the entire file?
21:10:53 Quit obo (Remote closed the connection)
21:11:13linuxstbhcs: I don't know, but a single call to request_buffer is only guaranteed to return a minimum of 32KB.
21:12:15lostlogicNico_P: I can only presume that yield_codec() immediately returns true
21:12:34Nico_Plostlogic: yeah but buffering should resume at some point
21:12:35lostlogicNico_P: my guess is that another event gets written to the buffer ~simultaneously to the buffer_handle
21:12:50lostlogicNico_P: yeah, I can't figure that out either −− that's what I'd hoped you would know
21:12:59lostlogicbecause it should just hit that file again in the buffering loop
21:13:05*lostlogic not know
21:13:19***Saving seen data "./dancer.seen"
21:13:26*Nico_P smells a race conditions
21:14:31linuxstbhcs: Imagine if your file was loaded at the end of the audiobuffer, and wrapped. The guard buffer (32KB) is there to provide a known minimum size that can be returned.
21:14:49mrkikoWhy FS #8003 disappeared from the web-site? Or it's just i CAN'T FIND IT?
21:15:15hcslinuxstb: understood, it's just hard to know any of this without reading through the thoroughly inscrutable playback code
21:15:27 Quit FOAD (Read error: 110 (Connection timed out))
21:15:28 Nick FOAD_ is now known as FOAD (
21:15:32mrkikosorry for the caps-lock
21:15:37Nico_Plostlogic: what do you think is sending the buffer handle request?
21:15:54lostlogichcs: don't talk bad about the playback code... among that inscrutability is sentience and it hears all.
21:16:07lostlogicNico_P: I _know_ that the buffer_handle request is sent in bufgetdata
21:16:14lostlogichcs did some detective work to determine that
21:16:41Nico_Pthat's not possible
21:16:53lostlogichcs: huh?
21:17:02Nico_Pbufgetdata returns ERR_DATA_NOT_READY if data isn't ready
21:17:07lostlogicer not hcs, Nico_P
21:17:23Nico_Pbufgetdata doesn't request more data, it lets the caller do it
21:17:24lostlogicer maybe it's codec_request_buffer_callback
21:17:29Nico_Pthat sounds better :)
21:17:40lostlogicok then it's codec_request_buffer_callback −− yeah, that's where hcs had his debugging :)
21:18:37Nico_Pthis one sends only on request for handle buffering, so if it gets aborted it could wait a bit long for the data
21:19:21lostlogicNico_P: yeah, which is why I just introduced that force flag, realistically though, the right thing might be that we need to fix the codecs to only request filechunk of data at a time
21:19:26lostlogicbecause then the bug just goes away on its own
21:19:32Learlinuxstb: Yes. Hence the compiler warning I saw (for Sansa only...) about a bad pointer.
21:20:05hcslostlogic: I recommend returning an error rather than locking up...
21:20:08Nico_Plostlogic: so the ogg codec asks for too much?
21:20:09 Quit SirFunk (Remote closed the connection)
21:20:09linuxstbLear: So ipodpatcher doesn't give a warning? Are the gcc flags the same?
21:21:00lostlogicNico_P: unsure, it shouldn't, forgot about that side of the buglet
21:21:08linuxstbhcs: I agree, these things need to be documented. I discovered it a couple of years ago when talking to Slasheri about my ALAC codec...
21:21:11Learlinuxstb: Didn't say that... :) It must've scrolled by when I was busy with other stuff.
21:21:12lostlogichcs: yeah, we should do limits checking and return an immediate error
21:21:15*Nico_P also doesn't get why the problem only showed on the sansa :/
21:21:27Learlinuxstb: Oh, sorry.
21:21:54LearI haven't buildt ipodpatcher.
21:22:19linuxstbI mean when you build rbutil - I'm assuming you only got the sansapatcher warning when doing an rbutil build?
21:22:37Nico_PI'll be back in about 1/4 of an hour (hopefully this time it'll be true too)
21:23:08LearYes, rbutil got the warning. I only saw it for sansa, but that was more luck than anything. I just happened to see it as it scrolled by.
21:23:40 Quit nicktastic ("Leaving")
21:24:06LearChecked now, and yes, same warning. (No surprise there.)
21:24:40 Quit amiconn (Nick collision from services.)
21:24:48 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
21:28:00amiconnlinuxstb: Thatnks. In fact my technique isn't difficult to understand, you'll see when I'll commit it
21:28:35amiconnWill still take some time because I need to split the yuv_blit helper function in order to make yuv_blit work properly again
21:28:54linuxstbIs lcd_yuv_blit running on the COP an issue?
21:29:25preglowjmspeex: are N and M in qmf_synth always going to be multiples of eight?
21:29:32amiconnI will now test my shiny new dualcore safe update tick, including the anti-bcm-lockup measure
21:32:43linuxstbNico_P: (When you're back) Have you thought about how to deal with files that require a contiguous area of the buffer?
21:33:00 Quit advcomp2019 ("Leaving")
21:33:50 Join kugel [0] (i=kugel@unaffiliated/kugel)
21:34:52 Join WarWolf [0] (
21:37:22linuxstbhcs: Is there a maximum size for an nsf file?
21:37:24 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
21:38:45lostlogiclinuxstb: there are several data types on the buffer that aren't allowed to wrap already
21:38:54lostlogicfor that matter only audio is allowed to wrap currently
21:39:06lostlogicso we'd just hafta give another type for "this is audio that must not wrap"
21:39:51linuxstbget_metadata() could determine that.
21:39:51 Join localhero [0] (n=dinzdale@
21:40:23lostlogiclinuxstb: sure, it's no problem from a playback side, just there isn't such a type on the buffering API yet
21:41:46lostlogiclinuxstb: it would probably actually work to just lie to the buffering API aobut the type of file for NSF/SPC
21:41:53linuxstbBTW, what can't wrap at the moment?
21:42:36lostlogiclinuxstb: as of now we only wrap audio data, everything else is forced
21:42:51lostlogicso I guess that means codecs and id3 are not wrapped
21:43:26linuxstbI would have thought codecs could wrap.
21:43:47lostlogiclinuxstb: they are currently requested as a unit, that could be changed, but it hasn't yet
21:44:12lostlogichmm, or are they done with a read instead of a get
21:44:37linuxstbThey _should_ be read IIUC.
21:44:52 Quit bertrik ("bye")
21:45:26lostlogicit's a bufgetdata right now
21:45:38lostlogicplayback.c 367
21:46:45lostlogicah, needs to be called fromt he codec_load_ram function instead
21:46:47linuxstbAre the semantics of bufgetdata and bufread defined anywhere?
21:46:57WarWolfso whats up with svn build now, get "*panic* stkov audio" on my x5 currently :(
21:46:58lostlogiclinuxstb: not that I've seen
21:47:14lostlogicWarWolf: works fine for me :-P
21:47:39WarWolfhad some time over so i made a new build, first in months :)
21:48:00linuxstbSo based on the old playback engine, bufgetdata() is guaranteed to return 32KB, unless the handle refers to something that is marked as "no wrap", in which case it will return the whole file?
21:48:01lostlogicWarWolf: you use patches?
21:48:10WarWolfdon't work so well, yesterday build archive works fine tho
21:48:19WarWolfyea i do use album-art
21:48:23linuxstbAnd bufread() will provide the whole file?
21:48:26WarWolfdon't work without it
21:48:42lostlogicbufread copies to a dest pointer and will provide as much as it has available up to size
21:49:05linuxstbAh, so it won't block until the whole file is read?
21:49:07lostlogicbufgetdata will not return unless it fails or can satisfy the whole request, which is the result of the current loops, because it gets into a never satisfy
21:49:31lostlogicstill though codec_load_ram should be able to use bufread to read out the codec
21:50:07linuxstbSo it's up to the caller to check the return value of bufread, and keep calling it until the whole file is read?
21:50:21lostlogicyeah, just like a read()
21:50:31*Nico_P is back
21:50:35linuxstbWouldn't it be simpler to make it block until all the data is available?
21:50:35WarWolfahhh seems current build from webpage don't work either so it's just not my build atleast :)
21:50:58lostlogicWarWolf: something specific to x5?
21:51:12WarWolfdunno i only got the x5 :)
21:57:48preglowamiconn: where did you find details on the arm9 pipeline?
21:58:44amiconnSame page as the arm7 stuff...
21:58:59preglowon the arm site?
21:59:23amiconnThis is arm920t, i.e. gigabeat F/X
21:59:43 Join Frazz [0] (
22:00:11jmspeexpreglow: yes
22:02:02preglowjmspeex: goodie, i just unroll some stuff here and there
22:03:41 Join Gnu47 [0] (i=Gnu47@
22:04:56 Join MrRichard [0] (i=Tentacle@
22:05:02MrRichardHello all.
22:05:15MrRichardI just installed VMware player and I'm trying to compile some patches
22:05:22MrRichardbut I have basolutely no idea what to do...
22:05:41scorche|wMrRichard: that is what the wiki is for :)
22:06:02 Join Domonoky_ [0] (
22:06:07linuxstbMrRichard: How far have you got? Have you compiled a normal Rockbox build (without patches)?
22:06:09MrRichardI've just been up all night reading the wiki and all the guides.
22:06:15MrRichardI've never compiled
22:06:33MrRichardIn any case I've gotta' leave for work so I'll be back to pester the kind souls of #rockbox later.
22:07:03MrRichardbut before I do leave, which file do I delete to delete my database files on the player?
22:07:20linuxstbAs scorche|w said, the wiki has lots of instructions for compiling...
22:07:45rasherMrRichard: .rockbox/database*
22:08:08MrRichardit has plenty of instructions, I just lack any knowledge in the area and find some of it hard to follow.
22:08:10MrRichardrasher: thanks
22:09:07nanokMrRichard: unless you are some kind of security guard or something, i am worried how you can work if you were up all night ..
22:09:22MrRichardLots of coffee
22:09:38WarWolfwhen does the "daily build" build?
22:09:40MrRichardHeck, I'm only a breakfast chef. Who cares what their eggs look like ;)
22:09:49rasherWarWolf: midnight utc, I believe
22:09:56WarWolfok, ty
22:10:05 Quit desowin ("use linux")
22:10:20nanokMrRichard: "just do like i do, sir, and in the morning you won;t even be able to see the damn eggs"
22:10:29linuxstbI thought it was closer to around 6am UTC...
22:10:31 Join barrywardell [0] (
22:11:20 Join isforinsects [0] (
22:11:21 Join webguest68 [0] (i=4a469ecd@gateway/web/cgi-irc/
22:11:28MrRichardnanok: well I'm off to mangle some poor soul's bacon. Thanks for the help so far :)
22:11:31webguest68where can i download the e200r bootloader
22:11:38*linuxstb checks the timestamps and now agrees with rasher ;)
22:11:48lostlogiclinuxstb: any interest in rewriting codec_load_ram to directly talk to the buffer so that codecs can be allowed to wrap in the new world order?
22:12:04webguest68never mind, i don't need it
22:12:11preglowjmspeex: gonna try porting that stereo code to fixed point now
22:12:22rasherlinuxstb: I'm not too sure now..
22:12:53linuxstbrasher: Yes, I was looking at a mirror...
22:13:11 Part low_light
22:13:16rasherdefinitely not midnight.. perhaps 02:00 utf
22:13:26Bagder6am CET iirc
22:14:01rasherBagder: the timestamps on haxx. seem to disagree
22:14:12rasherI assume that's where they show up first?
22:14:21linuxstbDepends what timezone they're in - haxx could be UTC
22:14:23Bagderthere's more to it
22:14:24 Quit webguest68 (Client Quit)
22:14:40Bagderthe builds are built on a different server and rsynced over to the download master
22:14:57Bagderand on the build server they are done starting 6am our time zone
22:15:02nanokrasher: wouldn't "crontab -l" be an easyer way to figure it out? :)
22:15:30lostlogicmy server pretends to be in PDT even though the physical box is probably in EDT yay timezones
22:15:30rasherSo now we need to know what MrRichard wanted to know.. when they are built, or when they are available
22:15:35linuxstbBagder: Maybe it's worth adding that info to the daily builds page?
22:16:07Bagderwell, I think publishing the svn rev number is more important
22:16:12Bagderthe exact time won't help a lot
22:16:13rasherLooks like they might be available an hour early some days (presumably because the build server finished faster?)
22:16:14WarWolfjust trying to track down whats the problem with last current build for my x5, not that many changes to source tho
22:16:47*amiconn wonders why rsync would mangle time stamps
22:17:00linuxstbBagder: I still think it's nice to know - you can't have too much information ;)
22:17:00Bagderrasher: the download server rsyncs them every 20 minutes during 5 hours every morning
22:17:04lostlogicamiconn: it's mroe that it doesn't mangle them −− they keep the origin server's stamp
22:17:08amiconnThat said, I never used rsync, but I'm used to the fact that copying files keeps correct timestamps
22:17:31linuxstbThe default with "cp" is to not preserve the timestamps...
22:17:34amiconnSo why are they different from 6:xx cet then?
22:17:50amiconnlinuxstb: Odd...
22:17:56Bagderthere's not
22:18:11lostlogicamiconn: build starts at 5 cet, runs on a server where it's 1300pst, copies back with a 1300 timestamp
22:18:19amiconn[22:14:05] <rasher> Bagder: the timestamps on haxx. seem to disagree
22:18:20BagderI mean, they say 06:xx on the download server locally as well
22:18:38Bagderyes, but he's wrong
22:18:39linuxstbamiconn: I think it's logical - the copy was created at that time, but I always use "-p" to preserve timestamps.
22:18:40rasheramiconn: not really. If you're copying a file, you're creating a new one - not moving the old one. It makes perfect sense to me.
22:19:11Bagderwell, the last builds seem to go at around 07:00 cet
22:19:17amiconnlostlogic: Well, I'm used to the ntfs way of copying files. Ntfs is timezone aware, so timestamps are still correct even if you copy files from a server in cet to server in pst, or wherever
22:19:29WarWolfok, seems like the last changes to settings_list.c makes my x5 crash when i play a mp3
22:19:49WarWolfreverted it to last version seems like everything works just fine now
22:20:00amiconnlinuxstb: I'm also used to having 3 timestamps: creation time, last modification time, and last access time
22:20:13amiconnCopying a file touches the 1st and the 3rd, but not the 2nd
22:20:34amiconnAnd the 2nd is what file browsers show by default...
22:20:45rasheramiconn: creation time is especially yucky, since many modifications are actually done by creating a new copy, then moving it back, so the filesystem has to do all sorts of guessing as to whether this is a new file or not
22:20:55 Join Isolinear [0] (
22:20:55Bagderand "last access" is a great performance killer ;-)
22:21:02lostlogicnoatime ftw
22:21:03amiconnIs it?
22:21:07Bagderit is
22:21:27Bagdermost *nix systems have it as well
22:21:29lostlogicamiconn: forces an actual disk write on every read, specially bad on journaled file systems
22:21:42amiconnThere are caches...
22:21:50rasherAnd especially bad on systems where the file is in cache, but the disk is spun down
22:21:50lostlogicamiconn: journals are not cached
22:22:38linuxstbWarWolf: So you've confirmed that r15389 is fine, but r15390 and later freeze?
22:22:59linuxstbWarWolf: Sorry, add one to those numbers...
22:23:08lostlogiclinuxstb: did that change introduce an additional level of indirection?
22:23:23amiconnSo far I didn't see any negative impact due to last access time handling... I don't even know if it's disableable in windows, and even if it were, I probably wouldn't
22:23:48WarWolfhold on need to check on revs
22:23:56nanokamiconn: if you have caching on your journal (it can happen), you can get in a lot of trouble
22:23:57linuxstblostlogic: Which change?
22:24:22nanokor so i've heard
22:24:28WarWolfim running 15142
22:24:59nanokamiconn: the performance penalty for atime is hardly visible on a workstation. it is there, but you cannot actually feel it, most of the time
22:25:14lostlogicamiconn: the one that warwolf thinks causes his breakge −− he mentioned an audio stkov
22:25:28 Quit Domonoky (Connection timed out)
22:25:57preglowwhy the yellow? gcc bugs?
22:26:02 Join webguest13 [0] (i=c3bd8ee1@gateway/web/cgi-irc/
22:26:04 Nick Soap__ is now known as Soap (
22:26:06linuxstbalways ;)
22:26:15 Quit webguest13 (Client Quit)
22:26:17lostlogicer that last line was to linuxstb
22:26:54linuxstblostlogic: I'm not sure what commit WarWolf is referring to any more...
22:27:24WarWolfnot commit, file reversion :)
22:27:31lostlogicamiconn: <−− shows at least some of the performance impact of atimes (and also shows how it hurts smoe filesystems but not others)
22:27:52lostlogicWarWolf: no such thing as a file revision :) only repository revisiosn, which are commits.
22:28:31 Join webguest80 [0] (i=0c9b3414@gateway/web/cgi-irc/
22:28:33 Part isforinsects
22:28:42scorche|wwell, there are file revisions in some SCMs...
22:29:30lostlogicscorche|w: not modern ones :-P
22:29:30WarWolfwell reverted last changes to settings_list.c
22:29:47linuxstbscorche|w: And importantly, not our one...
22:29:50 Join SirFunk [0] (
22:30:18WarWolfyea it's just svn, makes me confused
22:30:23WarWolfnot used to it
22:30:33amiconnAh, it can be disabled for ntfs as well... but if it's really just a few percent, I'd rather keep it enabled
22:30:35linuxstbWarWolf: If you're running r15142 (which is old), why do you think it's the settings_list.c change?
22:30:48*scorche|w = pedantic...
22:31:18WarWolfdownloaded the old settings_list.c and replaced old the one from last commit, and now it works fine
22:31:22*lostlogic pees on your dang tick
22:31:39*scorche|w frowns
22:31:45WarWolfthe settings_list.c was from the 15142 commit
22:31:46*scorche|w vows revenge
22:31:55 Quit webguest80 (Client Quit)
22:32:08linuxstbWarWolf: Have you tried an official build, or just your own builds?
22:32:24WarWolflast daily build don't work
22:32:30WarWolfyesterdays works just fine
22:32:35linuxstbThat commit to settings_list.c doesn't look as if it would affect playback...
22:32:49lostlogicneither that one nor the one at 15141-15142 do.
22:32:53linuxstbAnd the last daily build was made before that settigns_list change.
22:33:11n17ikh|Lappyhooray for rockbox
22:33:13WarWolfit could as it don't scroll any titles before u start to play an mp3
22:34:49*preglow continues with his barrage of speex commits
22:35:46jmspeexpreglow: are you just adding assembly or are you modifying existing files (except for the stuff that disables the encoder)?
22:36:03linuxstbWarWolf: Yesterday's daily build was r15381, and the previous one was r15370 - so the troublesome commit must be in that range. The settings_list.c change was 15391
22:36:18linuxstbI meant today's daily build was r15381...
22:36:27WarWolfyea ofc :)
22:37:12linuxstbpreglow: You don't want to "svn rm" the files you remove from SOURCES ?
22:38:58 Nick parafin is now known as parafin|away (
22:39:13WarWolfcurrent daily build is r15392 for me
22:39:39WarWolfso it must be between r15381 and r15392
22:39:51linuxstbNo, that' the current "current build"...
22:40:10linuxstb(or was, 30 minutes ago...)
22:40:24WarWolfstill is
22:40:39WarWolfahh sry that was the build that did not work
22:40:53WarWolfdid not read on the page :/
22:41:07WarWolfjust thought there was daily builds no current
22:43:17n17ikh|Lappythis is what I get for using a pre-patched branch
22:43:31n17ikh|LappyI have no idea if these behaviors are default for the rockbox on the e200 series or not :/
22:45:03preglowlinuxstb: no
22:45:20 Quit animeloe ("This computer has gone to sleep")
22:45:27preglowlinuxstb: i might want to make speex encoder
22:48:05jmspeexpreglow: if you have a decoder-only version of Speex linked in, it might be hard to add just the encoder.
22:48:43nanokn17ikh|Lappy: ?
22:49:54 Join safetydan [0] (n=safetyda@rockbox/developer/safetydan)
22:50:23n17ikh|LappyI'm using the cpchan branch on the e200, but I've never used rockbox (or the OF, even) on the e200 so I don't know if the behaviors are silly patches done by the maintainer of the branch or not
22:50:34n17ikh|Lappylike, the player becomes useless when you plug a usb cable in
22:50:43nanokn17ikh|Lappy: what behaviours?
22:50:59nanokn17ikh|Lappy: that one s normal :) usb support isn;t ready yet
22:51:16n17ikh|LappyI'm not talking about mass storage
22:51:21nanokn17ikh|Lappy: for file transfers, use the of,
22:51:23preglowjmspeex: the encoder won't be linked in
22:51:23n17ikh|LappyI'd like it to just charge
22:51:37n17ikh|Lappybut it displays a usb-cable icon and becomes useless
22:51:50preglowjmspeex: all our regular encoders and decoders are plugins, the speex decoder will be the sole exception, thanks to tackling voice ui
22:51:58nanokfor limited charging while playing, hold the middle button while plugging in (rockobox), but in my experience you will not get a full charge
22:52:03rashern17ikh|Lappy: reading the manual would help you here - ther'es some button you can hold to avoid that
22:52:08nanokfor a full charge, again, use the OF :(
22:52:18n17ikh|Lappythanks, nanok
22:52:41preglowcan someone please give me the name of a linux audio editor that actually works? don't mention audacity, that can't even bloody cut a stereo wav file without corrupting it
22:52:46 Quit keanu (Read error: 110 (Connection timed out))
22:52:54nanokn17ikh|Lappy: np. rasher is right too, allthough i found it the hard way, myself (good old trial and error)
22:52:58 Join DerPapst [0] (
22:53:30 Join keanu [0] (n=keanu@unaffiliated/keanu)
22:53:30preglowlinuxstb: kinda hard to find cut points with it, but if it did have a waveform view, i'd go for it :>
22:53:34nanokpreglow: try rosegarden, audacity is..something else :)
22:53:51nanoklinuxstb: :-D
22:53:53 Quit Nico_P (Remote closed the connection)
22:53:59preglowi thought rosegarden was a sequencer
22:54:05n17ikh|Lappystill, the behavior where it boots into the OF if you plug in a usb cable when it's off is sort of nice
22:54:12n17ikh|Lappybecause that way you can do the mass storage thing right away
22:54:27n17ikh|Lappywith my h10, you have to hold down the select button while booting for "emergency mode" or whatever
22:54:44nanokpreglow: i am far from being able to understand these terms fully, but i think, iirc, rosegarden should do it for you
22:54:51jmspeexpreglow: I mean that if you want to load the encoder plugin, you might actually end up with two copies of the common code/data
22:55:03*linuxstb uses dd to edit ac/3 files - lots of fun ;)
22:56:00amiconnn17ikh|Lappy: With the ums firmware, the H10 also just enters USB mode when powering on with usb plugged
22:56:20 Part Domonoky_
22:56:26preglowjhulst: yeah, we'll just have to live with that
22:56:35preglowjmspeex: yeah, we'll just have to live with that
22:56:50n17ikh|Lappyums firmware?
22:57:02preglowjmspeex: that really can't be helped anyway, we can't share data/code between core and a plugin
22:57:04amiconnThere are 2 firmwares for the H10, ums and mtp
22:57:15n17ikh|Lappylike, the actual bootloader
22:57:24amiconnThe mtp firmware only does "emergency" ums mode
22:57:32n17ikh|LappyI see
22:57:37n17ikh|Lappywhere can I get said firmwares?
22:57:38jmspeexpreglow: did you say you want to do a voicec ui using Speex?
22:57:40n17ikh|Lappythe actual rockbox site?
22:57:41amiconnOn the small H10 you can choose which one to install
22:57:49n17ikh|LappyI don't have the small h10
22:57:59n17ikh|LappyI've got the beefy-h10
22:58:09preglowjmspeex: aye
22:58:14amiconnThe H10 pure is only mtp officially, but can be hacked to use the standard H10 (small) firmware. YOu'
22:58:24n17ikh|LappyI see
22:58:24preglowjmspeex: basically just using it to decode all our voice clips used for menu navigation, etc
22:58:27amiconnll get an unusable radio menu, but that's it
22:58:38n17ikh|Lappywell, I might give my h10 to a friend since I've bought a sansa e260
22:58:42amiconnHmm, not sure about the big H10
22:59:02n17ikh|Lappyso I'll leave the original mtp firmware on there
22:59:14n17ikh|Lappysince he's not the type to use rockbox
22:59:20*nanok is unsure about giving any rockboxable iriver for a sansa..
22:59:45n17ikh|Lappywell, I'm still playing with the sansa, it just came today
22:59:47nanokallthough i do like the sansa a lot
22:59:50n17ikh|Lappyit's a lot smaller, that's for sure
22:59:55 Join TMM [0] (
23:00:00n17ikh|Lappythe H10 is a brick
23:00:15n17ikh|Lappytoo bad there's no rockbox for the 2g nano, I'd really be in business then
23:00:19 Quit WarWolf ("zzzz")
23:00:28amiconnWell, you can't really compare hdd and flash players wrt size
23:00:31 Part toffe82
23:00:31n17ikh|Lappybut, the nano I have is only 2 gigs, which is lame
23:00:42amiconnFor a flash player, the e200 series is quite bulky
23:00:46rasherThe h10 is a brick? What's the h120 then, or *gasp*, the archos?
23:00:55n17ikh|Lappythe archos? a turd >_>
23:01:01*amiconn would prefer the c200 when he had to choose a sansa
23:01:11 Quit jhulst (Remote closed the connection)
23:01:13nanokreal men like bricky players
23:01:20nanoksize denotes reliability
23:01:26nanoksize and weight
23:01:35n17ikh|Lappyor just that it's a previous-gen player
23:01:42nanokin the imortal words of boris :)
23:02:22n17ikh|Lappysee, if I'm gonna have a player as big as the h10 it needs to be at least 300gb >_>
23:02:31nanok"if it doesn;t work, at least you can allways beat yourself with it"
23:02:43n17ikh|Lappythis is true
23:02:51n17ikh|Lappyalso, I find the sandisk usb cable hilarious
23:03:00n17ikh|Lappywhy can't these people just use a standard mini-usb B jack?
23:03:14n17ikh|Lappythe dock connector thing is as big as the whole bottom of the player
23:03:16nanokn17ikh|Lappy: don;t even get me started :(
23:03:27n17ikh|Lappyeven the iriver's connector is smaller
23:03:36n17ikh|Lappyand it has that weird two-cable thing going on
23:03:38nanokn17ikh|Lappy: i think it would be nice to move this to -community though
23:03:45 Join Zagor_ [0] (
23:03:46n17ikh|Lappythere's a -community?
23:03:56nanokn17ikh|Lappy: see /topic
23:04:06n17ikh|Lappynew addition?
23:04:24 Quit scorche|w ("CGI:IRC (EOF)")
23:04:46preglowjmspeex: should all fixed point math in speex be signed?
23:06:01jmspeexpreglow: I believe so. Why?
23:06:17 Quit pondlife ("Read error: 110 (Connection slimed out)")
23:06:23preglowjmspeex: just wondering, seems some of the stereo stuff just needs unsigned ints
23:06:40preglowbut i'll stick to signed ints
23:07:12 Quit midkay ("Leaving")
23:07:49lostlogicpreglow: I think that your recent commits trimmed 11k off of the
23:07:58 Join scorche|w [0] (n=8dc5049d@rockbox/administrator/scorche)
23:08:11preglowlostlogic: trimmed around 25kb++ off speex.codec, so not unlikely
23:08:16jmspeexpreglow: yes, anything that is float should be converted to signed int, even if you know it's positive
23:08:20preglowdelta table for zip size, please! :P
23:08:26lostlogicpreglow: :)
23:08:58 Quit localhero ("later, gator")
23:10:23 Join Odd-rationale [0] (
23:10:34 Part Odd-rationale
23:11:17 Join darkapostrophe [0] (
23:12:02 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
23:12:44 Nick Zagor_ is now known as Zagor (
23:13:20***Saving seen data "./dancer.seen"
23:13:55nanokhm, that bug which makes mp3's stop while playing _is_ now definetly "working" with ogg also
23:14:14nanokis there a flyspray entry for that allready?
23:14:34 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
23:16:13 Join freqmod__nx [0] (
23:16:45lostlogicnanok: I don't think so, do you have a repeatable test case, or is this still a "it'll happen if I play long enough" thing?
23:17:00 Quit Dark_Apostrophe (Success)
23:17:01 Nick darkapostrophe is now known as Dark_Apostrophe (
23:17:31nanoklostlogic: it's still random and unpredictable :(
23:17:31 Quit jhMikeS (Read error: 104 (Connection reset by peer))
23:17:50nanokif i convince it to play the same song again it will even play it fine the second time
23:18:22 Quit Lear ("ChatZilla [Firefox]")
23:20:10*preglow misses floats :/
23:20:34lostlogicnanok: yeah, that makes sense unfortunately
23:21:19lostlogicnanok: I forget what is the status of the buffering thread when this bug strikes?
23:21:33lostlogicnanok: is usefl close to zero or high?
23:22:12preglowjmspeex: what would be a torture test for the stereo decoder?
23:22:27preglowjmspeex: just something panning wildly between channels?
23:23:38 Quit jgarvey ("Leaving")
23:24:50 Join donsdw [0] (
23:24:51 Quit donsdw (Remote closed the connection)
23:25:57 Quit jhulst (Read error: 104 (Connection reset by peer))
23:26:51 Join MajorC [0] (
23:27:11nanoklostlogic: hm, i shamefully have to admit i wouldn't know how to check that
23:27:46 Join donsdw [0] (
23:28:45lostlogicnanok: when it stops playing hit menu, system, debug, buffering thread
23:28:57lostlogicnanok: and just report the vaules of the four sliders in there
23:34:43 Join darkapostrophe [0] (
23:37:31 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
23:38:22jmspeexpreglow: There's no real "torbute" test for that. Just check on the "middle" case an on the extremes (completely left or right)
23:39:15 Quit Frazz ("Leaving")
23:40:32 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
23:41:30preglownot really friendly code for fixed point, this
23:42:02 Join Buschel [0] (
23:42:30 Quit mrkiko ("Lost terminal")
23:45:09jmspeexpreglow: you know that you can replace exp() with a 16-entry table right? (because the input is an int with 16 values possible)
23:46:07 Quit jmspeex ("さようなら")
23:46:12 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
23:47:42 Quit midgey ()
23:47:57amiconnlinuxstb: around?
23:48:08 Join midgey [0] (
23:48:34*amiconn has a patch that should receive some testing on an old G5 before commit
23:48:44amiconnBuschel: around?
23:49:02Buschelamiconn: yep, just read all the news in the log :o)
23:49:13Buschelyou've got it?
23:49:25amiconnYup, it works now
23:49:53Buscheleven for the blit-stuff?
23:49:54amiconnIncluding proper core locking, resolving occasional bcm stalls, and yuv blitting fixed
23:49:58 Quit davina (Remote closed the connection)
23:50:17Buschelwhat are the fps`s?">
23:51:12 Quit Dark_Apostrophe (Connection timed out)
23:51:12 Nick darkapostrophe is now known as Dark_Apostrophe (
23:51:34amiconn80MHz: lcd_update 101fps (full), 391.5 fps (1/4), yuv_blit: 27.3fps (full), 108fps (1/4)
23:51:47 Join webguest57 [0] (i=4a469ecd@gateway/web/cgi-irc/
23:52:11webguest57how do i install a patch on a bootloader?
23:53:02Buschelwow, thats exactly what was expected when removing the 14ms :)
23:53:21Zagorwebguest57: if you don't know, you don't want to do it
23:53:29amiconnYes, but my patch does never miss to make updates visible :)
23:54:18webguest57someone gave me a bootloader and a patch. i need to know how to use that patch on the bootloader
23:55:00amiconnAnd it fixes the occasional bcm stall (frozen gui)
23:55:03preglowbut yuv blit cannot be further optimized?
23:55:58linuxstbamiconn: Let me switch computers...
23:56:03 Quit linuxstb ("ChatZilla [Firefox]")
23:56:16webguest57this is my post and the problem i am having.
23:56:29 Join Thundercloud [0] (
23:56:42webguest57If you read this, you will know what i mean
23:56:51krazykitthat's not supported here
23:57:09krazykitit's an unofficial bootloader, and #rockbox only supports official builds and bootloaders
23:57:42webguest57okay, sorry for asking
23:57:46amiconnBuschel: That code still contains my test indicator for bcm stall: If it resolved a bcm stall, the backlight will light up
23:57:46*lostlogic tests amiconn's patch because it sounds cool
23:58:17amiconnJust 2 if()'s which will be removed before commit
23:58:37webguest57but you can't tell me how to patch a bootloader?
23:58:50amiconnCompiling it throws a warning because I was too lazy to #include "backlight.h" just for this test...

Previous day | Next day