Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2010-02-16

00:03:09 Join flydutch [0] (
00:03:23saratoga2stripwax: have you looked at optimizing BUTTERFLIES?
00:03:49saratoga2that seems like a good canidate for asm
00:10:17saratoga2stripwax: heres TRANSFORM_EQUAL on ARM [or rather the end off fft8 after your ASM code]
00:14:49saratoga2aside from all the sequential ldr/str its not too bad, though using multiple load store would speed that up a fair bit
00:15:06stripwax_saratoga2 - um, yeah. when I say I've optimised TRANSFORM macros, I mean everything in it...
00:15:43saratoga2can you pastebin it?
00:16:27stripwax_if you can wait a few secs I'll commit it :) But yep sure. Basically, BUTTERFLIES is only used in fft4, fft8, fft16, and pass() functions. I'll pastebin the asm for the new pass() fn.
00:18:16stripwax_just got to wait a sec (I'm stripping out the CPU_ARM - specific parts into its own proper header file)
00:18:54 Quit domonoky (Read error: Connection reset by peer)
00:20:30 Quit flydutch (Quit: /* empty */)
00:21:25*stripwax_ sighs - breaking your own code by refactoring it, and not having a last-good-working-version backed up :(
00:21:44JdGordon__been there.. done that :(
00:23:06 Nick stripwax_ is now known as stripwax (
00:24:41saratoga2in the current version of the TRANSFORM macros used by fft16, 68 out of 168 gcc generated instructions are single load/stores
00:24:59saratoga2for all 4 macro calls
00:25:20stripwaxsounds about right.
00:25:34saratoga2i guess thats misleading since all the mul macros are actually ASM
00:25:44saratoga2so its more like 68 out about 120 lines of gcc generated code
00:26:44stripwaxwith my optimised version, there's a small amount of spill over (so a bit of stack usage for I think one or two variables), but everything else is multiple load/store. If I actually wrote pass() in asm by hand I think it would fit in the available registers without any spillover
00:27:16saratoga2how do you coax gcc into using load/store multiple?
00:30:54stripwaxby writing it in inline asm :)
00:31:14stripwaxmy implementation started off simpler - but now its about 95% my asm and 5% c glue
00:32:58 Quit evilnick (Read error: Connection reset by peer)
00:33:23 Join evilnick [0] (
00:33:50*stripwax groans - now it's not even compiling
00:34:04stripwaxtrust me - it worked fine before trying to move code around :) ah, it's working now
00:38:05 Quit Buschel ()
00:40:55stripwaxsaratoga2 - . that's the whole fft-ffmpeg.c file (sorry)
00:43:24stripwaxargh, some bug has crept in and the audio's glitchy now. no, no, no, no, no ....
00:44:54*stripwax checks that it's just in the asm somewhere and that sdl still works ok
00:48:08 Quit bmbl (Quit: Bye!)
00:54:42***Saving seen data "./dancer.seen"
00:58:20 Quit bertrik (Quit: De groeten)
00:58:39mtDoes anyone have a problem if I add 2 fields - for the number of channels and extradata_size - in struct mp3entry ?
01:04:52 Join robin0800 [0] (
01:08:14 Quit Rob2222 (Read error: Connection reset by peer)
01:10:46 Join checker [0] (
01:11:17 Join Rob2222 [0] (
01:13:11checkerive written a guide for themes/wps'/sbs' based on the "CustomWPS" guide on the wiki, and id like to upload it to the wiki but it seems i need a writing permission
01:13:50stripwaxsaratoga - ok, panic over. it's actually a 'feature' that I hadn't yet tested, that doesn't actually work. but I'll commit without that thing and I think we're good.. the asm you've got doesn't "technically" work but it's hardly different from the working version
01:15:22CIA-8New commit by stripwax (r24678): Get rid of USE_PASS_BIG (you can always restore from an older version ..?). Use ARM-optimised TRANSFORM, BUTTERFLIES macros. Move ARM-specific code ...
01:15:37stripwaxsaratoga - enjoy
01:16:15stripwaxsaratoga - did you get a chance to move revtab to iram?
01:16:30JdGordon__checker: wiki name?
01:18:10JdGordon__you might want to link it from
01:19:44checkerim not too familiar with wiki...
01:23:11 Quit efyx_ (Remote host closed the connection)
01:27:25 Quit robin0800 (Remote host closed the connection)
01:27:42 Join robin0800 [0] (
01:28:58 Quit checker (Quit: CGI:IRC)
01:29:43 Join m3dlg [0] (~m3dlg@
01:35:26mtstripwax: Great effort !
01:35:31CIA-8New commit by stripwax (r24679): Fix the TRANSFORM macro in arm asm (the one that is only used by fft16 and was commented out in previous revision as didn't work properly). Proper ...
01:36:34 Quit m3dlg (Ping timeout: 276 seconds)
01:36:51 Join m3dlg [0] (
01:40:00 Quit m3dlg (Remote host closed the connection)
01:40:24 Join m3dlg [0] (
01:40:33stripwaxmt - glad you like :)
01:42:33 Join karashata [0] (
01:43:37 Quit robin0800 (Remote host closed the connection)
01:54:07 Quit m3dlg (Ping timeout: 276 seconds)
01:55:17*stripwax sleeps
01:55:32 Quit stripwax (Quit:
01:56:37 Part froggyman
01:56:57 Join webguest32 [0] (
01:57:32 Join webguest56 [0] (
01:57:32 Quit webguest32 (Client Quit)
01:58:16 Quit webguest56 (Client Quit)
02:01:55 Quit toffe82 (Quit: ChatZilla 0.9.86 [Firefox 3.6/20100115144158])
02:04:38 Quit JdGordon__ (Quit: Miranda IM! Smaller, Faster, Easier.
02:06:33 Join adam12345 [0] (
02:11:12adam12345Hi quick question about using the rockbox utility on WinXP with a sansa Fuze.
02:11:37 Join JdGordon1 [0] (
02:13:21adam12345The rockbox utiltiy in WinXP autodetected my sansa Fuze v1, but is asking me for a mount point. Would that be the root o the drive?
02:13:47adam12345such as G:\ ?
02:14:12adam12345thank you! Jd
02:24:13CIA-8New commit by mt (r24680): Add two extra fields to struct mp3entry, in preparation for making ATRAC3 more generic. (i.e, to avoid having to pass it RMContext even when the codec ...
02:25:39JdGordon1mt: I tihnk you need to bump the codec and plugin apis numbers
02:28:40mtJdGordon1 : Thanks for pointing that out.
02:29:11 Join makefile [0] (~god@
02:30:10 Quit linuxstb (Remote host closed the connection)
02:32:56mtJdGordon1: My changes doesn't directly affect the codec api though .. (struct mp3entry is a member of the codec api), so do you think I should still bump the version ?
02:33:15 Join Kitr88 [0] (
02:33:33JdGordon1any codecs or plugins accessing that struct will be pointing to the wrogn place
02:33:45JdGordon1for any members under the ones you added
02:33:55JdGordon1so they need to be recompiled
02:34:56 Quit Kitar|st (Ping timeout: 256 seconds)
02:35:53saratoga226.78MHz for 192k WMA!
02:36:03saratoga2299% realtime!
02:36:56saratoga2stripwax: pass now takes 5.69MHz on WMA
02:37:27saratoga2i can squeeze better performance out of the WMA core codec as well
02:37:39saratoga225MHz seems pretty easy, i wonder what the lower bound really is
02:38:20rasherDude, stop working on wma already, you're making MS look good
02:38:47saratoga2i'm pretty sure i'm faster then their own decoders, so they're not looking too good to me :)
02:39:22notByanhow fast are the FLAC decoders?
02:39:31saratoga2which player?
02:39:39CIA-8New commit by mt (r24681): Bump APIs' versions.
02:40:16notByanwhich player are those numbers for?
02:40:17saratoga2but the answer is "really fast" pretty much for any player
02:40:21mtsaratoga2: \o/ \o/ :)
02:40:34 Quit JdGordon1 (Ping timeout: 245 seconds)
02:42:26saratoga2vorbis 192k takes 27.6MHz
02:42:51saratoga2i want to email the tremorlo guy back with something like "we win"
02:43:53 Quit anewuser (Quit: =ooo ϢINTER ϾHIP 5iVE is OOON!!)
02:44:51 Join anewuser [0] (anewuser@unaffiliated/anewuser)
02:45:01mtsaratoga2: haha that would be kind of .. nice !
02:46:27 Quit freddyb (Remote host closed the connection)
02:46:38 Quit GeekShadow (Read error: Connection reset by peer)
02:47:24CIA-8New commit by mt (r24682): "Detach" ATRAC3 from RM. Since ATRAC3 isn't really specific to RM, it must not be obligatory to initialize the decoder through RMContext.
02:48:53mtsaratoga2: Any idea whether ATRAC3 in wav could be joint-stereo or just stereo ?
02:54:46***Saving seen data "./dancer.seen"
02:55:51CIA-8New commit by mt (r24683): Fix a comment in OMA metadata parser. No functional changes.
02:56:42saratoga2mt: AFAIK the limitation is just the bitrate, below 128k is joint stereo, above is dual stereo
02:56:53saratoga2i guess to save CPU power
02:57:02saratoga2seems like a really dumb design choice though
02:57:28saratoga2well heres the bitrev table assuming I did everything right:
02:57:30mthmm .. the problem is I can't find a flag in wav that would tell the coding mode
02:58:06mtMaybe I'll have to determine that according to the bitrate then
02:58:48mtwow the revtable is so tiny and pretty !
03:00:30CIA-8New commit by mt (r24684): Fix yellow.
03:01:54 Join cjcopi [0] (
03:05:32mtsaratoga2: : You were right, joint-stereo is used for 66 and 94 kbps
03:06:55 Nick fxb__ is now known as fxb (
03:07:06saratoga2mt: the bitstream really doesn't have a stereo mode flag?
03:07:56 Quit Adnyxo (Quit: Leaving)
03:08:01mtCan't find anything like that in the spec
03:08:34mtBut determining it according to the bitrate won't be a problem (luckily!)
03:10:24 Quit LambdaCalculus37 (Quit: Fwump)
03:16:06 Quit MethoS- (Remote host closed the connection)
03:19:37 Join Strife89 [0] (
03:20:05 Quit adam12345 (Quit: CGI:IRC (EOF))
03:20:28CIA-8New commit by saratoga (r24685): Remove runtime computation of the bitreverse table in favor of a constant table. ff_fft_init is still required however because it initializes the fft ...
03:21:15saratoga2i guess to get rid of the init functions we want to drop the FFTContext function pointers entirely
03:23:31 Join Schmogel [0] (
03:30:37 Quit HBK (Read error: Connection reset by peer)
03:31:52 Join HBK [0] (
03:33:14mtsaratoga2: Was there any speedup ?
03:33:36 Join Hillshum [0] (
03:33:52 Join grndslm [0] (
03:35:08saratoga2mt: i didn't test but there shouldn't be
03:35:26saratoga2i'm still having trouble getting it to run without the init function, must be missing a pointer somewhere
03:37:16 Quit komputes (Remote host closed the connection)
03:39:33saratoga2do we want rockbox or ffmpeg headers on the new fft?
03:43:49saratoga2somehow the fft gets screwed up if i kill the call to fft_init even though i've removed the revtab and number of bits stuff
03:51:05 Nick fxb is now known as fxb__ (
03:51:09CIA-8New commit by saratoga (r24686): Remove the need for ff_fft_init. FFT is now stateless. MDCT still requires an init function.
03:53:22CIA-8New commit by mc2739 (r24687): updated German translation ...
03:54:15JdGordondid someone break playback in the sim?
03:57:12mtJdGordon: playback is working fine for me
03:59:35saratoga2mt, stripwax: I've updated the mdct to be stateless, but the individual codecs need updating and I'm too tired to figure out how tremor/libatrac decides its block sizes
03:59:51saratoga2feel free to update and/or commit once its working
04:00:34mtsaratoga2: Nice work :)
04:00:47JdGordonmt: with svn? its no worky for me :(
04:01:56mtJdGordon: Yes with svn
04:02:09mtsaratoga2: IIRC, atrac3 just uses a size 512 mdct
04:02:46mtJdGordon: Did you try whipping it ?
04:04:50mtnvm I really need to sleep
04:04:54saratoga2mt: I said i'm tired ;)
04:05:13 Quit Hillshum (Ping timeout: 264 seconds)
04:05:52mtsaratoga2: I was just stating the mdct size .. didn't imply anything really. :)
04:06:36 Quit DerPapst (Quit: Leaving.)
04:07:53CIA-8New commit by saratoga (r24688): Remove a call to av_log2 which would result in an inlined function.
04:08:22 Quit saratoga2 (Quit: CGI:IRC (EOF))
04:09:15 Join Hillshum [0] (
04:12:23 Join Barahir [0] (
04:14:59 Quit TheSeven (Disconnected by services)
04:15:12 Join The_Seven [0] (~theseven@rockbox/developer/TheSeven)
04:15:17 Quit makefile (Read error: Connection reset by peer)
04:15:23 Nick The_Seven is now known as TheSeven (~theseven@rockbox/developer/TheSeven)
04:16:29 Quit Barahir_ (Ping timeout: 276 seconds)
04:18:48 Join AlexP_ [0] (~ap@rockbox/staff/AlexP)
04:19:08 Quit AlexP (Remote host closed the connection)
04:19:08 Quit kadoban (Write error: Broken pipe)
04:19:16 Join kadoban [0] (
04:32:26 Join woodensoul [0] (
04:33:03woodensoulHello all. Congrats on release 3.5! I'm loving it as always.
04:34:32woodensoulhas the idea of a favorite album designation in the database been suggested?
04:34:42CIA-8New commit by mt (r24689): Initial support for ATRAC3 streams in wav containers. ...
04:39:56 Quit woodensoul (Quit: CGI:IRC (Ping timeout))
04:40:31 Join woodensoul [0] (
04:41:30 Quit Strife89 (Quit: Bed.)
04:44:29 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
04:45:31 Quit woodensoul (Quit: CGI:IRC (Ping timeout))
04:46:52 Quit ecio (Ping timeout: 258 seconds)
04:54:49***Saving seen data "./dancer.seen"
04:55:08 Quit Sajber^ (Read error: Connection reset by peer)
05:00:38 Quit Schmogel (Quit: Miranda IM! Smaller, Faster, Easier.
05:14:40 Join yeahx [0] (
05:24:13 Part yeahx
05:39:10 Quit Horscht (Quit: Verlassend)
05:44:56 Quit panni_ (Read error: Connection reset by peer)
05:48:46 Join ecio [0] (
06:01:32JdGordonanyone around?
06:02:05 Join phanboy4 [0] (
06:02:26JdGordoncan you test a patch quickly?
06:02:39JdGordonyou want me to build for you?
06:02:55mc2739not necessary
06:03:39mc2739where is the pactch?
06:04:09JdGordonload up a theme or two and make sure the backdrops displays correctly please :)
06:04:43JdGordonthat makes skins reuse backdrops if the sbs and wps use the same bmp
06:07:38JdGordonthe sims seem to be broken on my pc!
06:13:20 Quit rphillips (Read error: Operation timed out)
06:17:20 Join kramer3d_ [0] (~kramer@unaffiliated/kramer3d)
06:17:44 Quit kramer3d (Ping timeout: 258 seconds)
06:18:10 Quit JdGordon (Quit: Leaving.)
06:20:48 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
06:22:12 Join rphillips [0] (
06:22:55JdGordonmc2739: ah, sim is working here again.. how you going?
06:23:04mc2739JdGordon: patch looks good - backdrop changes appear to work properly - tested on e200v2 with and
06:23:15JdGordonta :)
06:24:32CIA-8New commit by jdgordon (r24690): Reuse the backdrop buffers if 2 skins use the same backdrop (on the same screen of course)
06:27:09 Quit grndslm (Quit: Leaving)
06:30:08 Join freddyb [0] (
06:30:16JdGordonOH FFS!
06:31:24CIA-8New commit by jdgordon (r24691): fix checkwps
06:31:36JdGordonone day I'll remember out that one :p
06:37:40 Quit Hillshum (Quit: Leaving)
06:54:53***Saving seen data "./dancer.seen"
06:56:02 Join jeffp [0] (
06:56:52jeffphey guys, i was just wondering if anyone's working on the sansa fuze v2..
07:13:22 Join evilnick_ [0] (
07:17:09 Quit evilnick (Ping timeout: 256 seconds)
07:17:38*pixelma does not really like the German translation update committed this night
07:20:43 Quit freddyb (Quit: Konversation terminated!)
07:21:40pixelmaI also thought that for languages where there are native speakers among the committers, we agreed on letting those review updates
07:22:15 Quit advcomp2019 (Read error: Connection reset by peer)
07:23:27pixelmaand could someone explain to me the first hook in the diff, I mean the sense of it in the original? Aren't all the targets that are listed there behind the "recording" feature already included with "recording"?
07:23:43 Nick kramer3d_ is now known as kramer3d (~kramer@unaffiliated/kramer3d)
07:26:03pixelmaespecially seeing that the rest of the phrase only uses "recording"...
07:29:35 Quit liar (Remote host closed the connection)
07:30:45pixelmaoh, it's in a deprecated string anyways....
07:30:59jeffphey guys, i was just wondering if anyone's working on the sansa fuze v2..
07:32:42 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
07:33:10FlynDicejeffp: will give you the latest progress
07:33:23 Quit anewuser (Quit: =ooo ϢINTER ϾHIP 5iVE is OOON!!)
07:33:55jeffpthanks flyndice- i've looked there already, i was hoping for some more hopeful news here :)
07:40:40FlynDicejeffp: I don't think anyone is actively working on it at the moment, give it some time. Even though noone is actively working on it I would guess a lot of the work on the clipv2 and clip+ should transfer over to the fuzev2.
07:41:35jeffpi hope so! i want to play with rockbox :)
08:00:23 Quit karashata (Quit: The fluffy dragon has left entirely!)
08:07:17 Join Zagor [0] (
08:07:18 Quit Zagor (Changing host)
08:07:18 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
08:09:37JdGordondoes anyone tihnk we need an explicit "next preset id/freq/name" token for the fms?
08:15:33LloreanOnly if we also include a Prev one.
08:16:09LloreanIt might be nice, though I'd not say "need" at all
08:16:38 Join ender` [0] (
08:19:52 Quit kramer3d (Ping timeout: 260 seconds)
08:23:16jeffpjd, you the same guy who used to dev for miranda-im
08:26:25 Quit AndyI (Ping timeout: 248 seconds)
08:30:00JdGordonLlorean: well, we dont have a prev id3 track
08:30:07JdGordonjeffp: a looong time ago :)
08:30:48JdGordonby id3 track i mean in the wps
08:30:56jeffpah cool good to see you again.
08:31:03LloreanJdGordon: True, but browsing playlists is less bidirectional than presets
08:31:11JdGordonthe reason I ask is because the playlist viewer makes the tags pretty redundant
08:32:56 Join AndyI [0] (~pasha_int@
08:36:02amiconnmt: Those two extra fields should be made #if CONFIG_CODEC == SWCODEC, imo
08:37:29 Join flydutch [0] (
08:38:45 Part jeffp
08:47:51pixelmaLlorean: if I remember correctly, browsing the preset list from the radio is as easy/complicated as browsing the playlist (just not "moving" presets etc.) if you ignore that *some* targets have a quick access button to the playlist now
08:51:23 Quit feisar_ (Quit: leaving)
08:51:26Lloreanpixelma: This is true, but I just think that a user is more likely to go "back" in presets than back in playlists, so bidirectional tags make more sense there (if you're going to do a 'next' that is, I'm not sure next is even necessary)
08:51:57 Join feisar [0] (
08:52:25 Nick feisar is now known as Guest38414 (
08:54:56***Saving seen data "./dancer.seen"
09:05:04 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
09:07:29 Quit ps-auxw (Ping timeout: 248 seconds)
09:12:56 Join petur [0] (~petur@rockbox/developer/petur)
09:13:53 Quit koniiiik (Ping timeout: 248 seconds)
09:13:59 Join koniiiik [0] (
09:15:44 Join Tomis2 [0] (~Tomis@
09:15:53 Quit Tomis (Read error: Operation timed out)
09:15:53 Nick Tomis2 is now known as Tomis (~Tomis@
09:23:01 Join ps-auxw [0] (
09:23:41 Join TopyMobile__ [0] (
09:25:35 Quit CaptainKewl (Ping timeout: 245 seconds)
09:27:29 Quit TopyMobile_ (Ping timeout: 240 seconds)
09:30:40 Join dfkt [0] (dfkt@unaffiliated/dfkt)
09:30:44 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
09:37:44 Join phanboy_iv [0] (
09:37:49 Quit phanboy_iv (Client Quit)
09:37:52 Quit phanboy4 (Quit: Leaving)
09:46:40 Quit TheUser (Ping timeout: 260 seconds)
09:51:22 Quit kramer3d (Quit: Leaving)
09:51:28 Join stoffel [0] (
09:53:30 Join grndslm [0] (
09:54:18 Quit S_a_i_n_t (Quit: There are 10 types of people, those who understand binary, and those who don't.)
09:54:46 Join S_a_i_n_t [0] (S_a_i_n_t@
09:55:53 Nick AlexP_ is now known as AlexP (~ap@rockbox/staff/AlexP)
09:57:27 Quit grndslm (Client Quit)
10:03:54 Join grndslm [0] (
10:07:45 Quit grndslm (Client Quit)
10:10:49 Quit evilnick_ (Read error: Connection reset by peer)
10:11:12 Join evilnick_ [0] (
10:22:34 Join DerPapst [0] (
10:23:22*gevaerts agrees with pixelma. Translations should be reviewed by the usual translator first
10:26:06 Join LinusN [0] (~linus@rockbox/developer/LinusN)
10:28:41 Quit Topy44 (Ping timeout: 256 seconds)
10:44:58 Quit S_a_i_n_t (Quit: There are 10 types of people, those who understand binary, and those who don't.)
10:52:57 Join robin0800 [0] (
10:54:57***Saving seen data "./dancer.seen"
11:18:53 Join Buschel [0] (
11:19:12Buschelmt: just found some atrac3 joint stereo files ->
11:19:20 Quit ecio (Ping timeout: 245 seconds)
11:20:00mtamiconn: Yes, I'll look into that.
11:20:22mtBuschel: Yes, but we couldn't play those files before because they're in wav
11:20:44Buschelmt: I just tried 2 files (sample.ATRAC3.66kbps.44100Hz.Stereo.wav and mc_sich_at3_066.wav). Both of them definately use JS, but the first file shows heavy artifacts when playing, the second file seems to sound fine.
11:20:54 Quit Buschel (Client Quit)
11:30:37 Join webguest000 [0] (
11:31:52 Quit webguest000 (Client Quit)
11:35:24CIA-8New commit by mt (r24692): Add the Rockbox GNU header to some files in libatrac.
11:38:03 Join b0hoon [0] (
11:38:13 Quit b0hoon (Client Quit)
11:38:48 Join b0hoon2 [0] (
11:40:18b0hoon2Hi, why in the rtc_e8564.c the address of it is set to 0x51 and from where it was taken?
11:43:55b0hoon2shouldn't it be 0xa2? It is a clone of the philips pcf8563 used in the Packard Bell.
11:45:06b0hoon2besides it brakes the rule of read/write addressing
11:50:01b0hoon2is anybody here now? :P
11:51:11mtb0hoon2: Probably not one that could answer you, or maybe there's, but they are afk. :)
11:51:16gevaertsb0hoon2: yes, but probably not the people who know rtc_e8564.c :)
11:52:17b0hoon2ok, thanks :) from who i can find out something more about it then?
11:52:22gevaertssvn blame says that you want markun for this
11:53:17gevaertshm, maybe it's even older
11:53:33mtamiconn: question ; when I added those two fields, I added them in a similar sort to those added for vorbis, ac3, and xing. Shouldn't they be moved under #ifdef SWCODEC too ?
11:54:49b0hoon2gevaerts: alright then, thanks! i should look at svn, so simple :)
11:55:31gevaertsb0hoon2: the 0x51 was added in r10799
12:00:31 Quit mt (Read error: Connection reset by peer)
12:00:50 Quit rvvs89 (Ping timeout: 258 seconds)
12:01:04 Join mt [0] (~mtee@rockbox/developer/mt)
12:05:01 Join watto [0] (~watto@
12:05:26 Join Sajber^ [0] (
12:11:00mtDoes anyone see any point in why libfaad should stay, if we managed to have a fixed point aac decoder from libffmpeg ?
12:12:56 Join ecio [0] (
12:13:51 Join teru [0] (
12:16:50mtOkay, I'm changing that in gsoc wiki.
12:17:26 Join shai_ [0] (
12:17:31 Part LinusN
12:18:11 Quit shai (Read error: Connection reset by peer)
12:18:28 Join MethoS- [0] (~clemens@
12:30:02AlexPpixelma: I agree about the translations. mc2739 probably just didn't know - did you tell him?
12:31:44b0hoon2gevaerts: committed by barrywardell, patch by Laurent Baum. The funniest thing is it works on the pp-i2c bridge and the PB vibe while it shoudn't (with this address). I'll try to contact with Barry Wardell about it.
12:41:58 Join darkham [0] (
12:45:14 Quit b0hoon2 (Quit: CGI:IRC 0.5.9 (2006/06/06))
12:45:45 Join Schmogel [0] (
12:54:11 Join kugel [0] (~kugel@rockbox/developer/kugel)
12:54:59***Saving seen data "./dancer.seen"
12:59:41CIA-8New commit by teru (r24693): * surround member only used by charcells with #ifdef HAVE_LCD_CHARCELLS in struct scrollinfo in scroll_engine.h. ...
13:00:23 Quit stoffel (Remote host closed the connection)
13:02:49 Join stoffel [0] (
13:04:16 Quit darkham (Quit: Sto andando via)
13:05:01 Quit robin0800 (Remote host closed the connection)
13:05:27 Join krabador [0] (
13:14:52 Join perfectdrug_ [0] (
13:18:30 Quit perfectdrug (Ping timeout: 245 seconds)
13:40:16 Quit Sajber^ (Read error: Connection reset by peer)
13:41:51pixelmaAlexP: I assumed he would be sleeping by this time, planned to do so when I see him next active
13:46:39 Quit DerPapst (Quit: Leaving.)
13:47:13 Join DerPapst [0] (
14:18:05 Join mt_ [0] (~mtee@
14:19:47 Quit mt (Ping timeout: 265 seconds)
14:21:30 Nick mt_ is now known as mt (~mtee@
14:21:43 Quit mt (Changing host)
14:21:43 Join mt [0] (~mtee@rockbox/developer/mt)
14:29:00 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
14:29:15 Join Adnyxo [0] (
14:36:21 Join Strife89|PalmTX [0] (~upirc@
14:38:32CIA-8New commit by teru (r24694): scrollbar.c: Simplify code a bit.
14:43:10 Quit Strife89|PalmTX (Quit: The one thing I hate about the Palm OS: no multitasking. :/)
14:44:34 Join TheSphinX^ [0] (
14:46:06 Quit krabador (Quit: Sto andando via)
14:55:03***Saving seen data "./dancer.seen"
14:55:31Zagordoes anyone remember what caused the ondio sp to keel over?
14:55:51Zagor(in the build table. it says flash full.)
14:56:04Zagorand apparently it's been like that for quite a while
14:56:22gevaertsZagor: the last drop was multifont
14:58:41 Quit ecio (Ping timeout: 276 seconds)
14:59:57 Join fwoggyman [0] (
15:01:49pixelmawell rombox for it
15:04:18CIA-8New commit by zagor (r24695): Updated SVN log
15:06:03 Join Topy44 [0] (
15:10:58 Join freddyb [0] (
15:11:27 Quit Schmogel (Ping timeout: 272 seconds)
15:18:55 Quit Topy44 (Read error: Connection reset by peer)
15:21:26 Join MJ [0] (~MJ@
15:21:33 Join Topy44 [0] (
15:21:53 Quit Zagor (Quit: Clint excited)
15:22:31MJI am trying to configure a TTS engine in the RB Utility, on Mac
15:22:42MJthere is, however, no documentation on the subject that I can find
15:23:13MJnor does there seem to be a OS X option to use the OS built-in TTS engine.
15:23:18MJany hints?
15:24:40MJ(OS X comes with a TTS command line utility 'say' that can output to .aiff files if needed)
15:26:58 Join Zagor [0] (
15:26:58 Quit Zagor (Changing host)
15:26:58 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
15:28:52 Join robin0800 [0] (
15:38:08 Join ecio [0] (
15:47:56 Join evilnick_B [0] (~0c140464@rockbox/staff/evilnick)
15:51:44ranmachanMy Sansa C240v2 all soldered up for testing: ranma/S6000731.JPG"> (trying to find the JTAG pinout)
15:53:56Zagorooh, pretty!
15:54:13 Join krabador [0] (
15:54:21 Quit fwoggyman (Quit: tata)
15:54:43 Join TheSeven|Mobile [0] (~theseven@rockbox/developer/TheSeven)
15:58:08pixelmaMJ: I believe work on making use of the inbuilt MacOS TTS for RbUtil has just been started and especially slightly older OS versions have problems. Maybe you can talk to the developer who experimented with this later (bluebrother if I recall correctly)
16:00:17gevaertsranmachan: good luck!
16:07:06ranmachangevaerts: Thanks. I'll have to figure out a good way to test the pins now :)
16:10:03 Join jgarvey [0] (
16:15:00 Quit teru (Quit: Quit)
16:19:09AlexPZagor: I'd ass a the and have "All commits since the last release"
16:19:12AlexPdear god
16:19:17AlexPI'd *add* a the
16:30:08 Join Schmogel [0] (
16:43:41 Nick fxb__ is now known as fxb (
16:48:15 Join Qumat [0] (
16:48:52 Nick shai_ is now known as shai (
16:49:13 Join toffe82 [0] (~chatzilla@
16:53:37Qumathi there
16:53:54 Nick Qumat is now known as wojtek1991 (
16:55:04***Saving seen data "./dancer.seen"
16:58:45wojtek1991how to get the content as is there:
17:00:46 Quit wojtek1991 (Quit: MegaIRC v4.05
17:05:03 Join pamaury [0] (
17:08:04 Quit FlynDice (Remote host closed the connection)
17:08:31 Join FlynDice [0] (
17:11:54 Join wojtek111 [0] (
17:13:09wojtek111how to make content as it is on:
17:13:27gevaertsWhat do you mean?
17:13:41 Join fyrestorm [0] (
17:14:13dionoeawojtek111: is what you are looking for ?
17:14:18wojtek111to show sunversions
17:14:46wojtek111not major changes but eg RSS
17:18:16gevaertsAre you asking if there's an rss feed for the svn log?
17:19:24wojtek111but to subversion
17:19:28wojtek111on main page
17:19:50gevaertswe don't provide one I think
17:21:02wojtek111nut on page I write there is
17:21:15wojtek111nut -> but
17:25:36wojtek111I know how
17:25:41wojtek111I found out
17:25:47wojtek111on this pae itis
17:26:11wojtek111here we go
17:27:13wojtek111how aobut GSoC?
17:28:11krazykitwhat about it?
17:28:12 Quit TheSeven|Mobile (Ping timeout: 260 seconds)
17:29:11 Join Sajber^ [0] (
17:29:28wojtek111deos rockbox join in it
17:30:35 Quit mt (Ping timeout: 245 seconds)
17:31:03gevaertsWe plan to apply
17:31:26wojtek111what about students?
17:32:17Zagorwojtek111: suggest projects here:
17:33:14 Join mt [0] (~mtee@rockbox/developer/mt)
17:33:40 Join Strife89 [0] (~michael@
17:34:41 Quit B4gder (Ping timeout: 276 seconds)
17:39:16 Join mt_ [0] (~mtee@
17:39:20 Quit mt (Ping timeout: 245 seconds)
17:44:33 Quit wojtek111 (Quit: Leaving)
17:45:10 Quit mt_ (Ping timeout: 245 seconds)
17:45:19 Join phanboy4 [0] (
17:45:34 Join anewuser [0] (anewuser@unaffiliated/anewuser)
17:46:16 Quit MJ (Quit: homeward bound)
17:50:05 Join mt [0] (~mtee@rockbox/developer/mt)
17:52:03 Join Farthen [0] (
18:03:51 Join liar [0] (
18:09:21 Quit mt (Ping timeout: 240 seconds)
18:11:25 Join mt [0] (~mtee@rockbox/developer/mt)
18:12:09 Join komputes [0] (~komputes@ubuntu/member/komputes)
18:13:14 Quit flydutch (Ping timeout: 265 seconds)
18:20:09 Join flydutch [0] (
18:20:42 Quit Strife89 (Ping timeout: 256 seconds)
18:26:21 Quit mt (Ping timeout: 240 seconds)
18:27:56 Join mt [0] (~mtee@rockbox/developer/mt)
18:28:09JdGordonAlexP: kugel: pixelma: I've updated the fms patch to make the preset tags more sensible, if you could test some time. peakmeter updating still doesnt work on hwcodec, and no menu item, but everything else should be working
18:28:41gevaertsfreddyb, JdGordon: is FS #11008 ready for commit?
18:28:44JdGordonnegative offsets in the playlist viewer shuold work if you use them with the preset info tags
18:28:59JdGordongevaerts: unless you tihnk otherwise
18:29:21gevaertsI want it in as soon as possiblt
18:31:33gevaertsJdGordon: do you want to commit it, or shall I?
18:31:41JdGordonif you have time, go for it
18:31:46JdGordonI'm heading to work in a few min
18:34:41 Join bertrik [0] (
18:34:51CIA-8New commit by gevaerts (r24696): Turn the simulator background into a clickable area. i.e. you click on the play button to press play. It also maps mouse scrollwheel events to rockbox ...
18:35:49pamauryomg oO tagcache uses fancy types in some areas
18:36:11JdGordongevaerts: and can you look at 11019?
18:37:11AlexPJdGordon: Cool, I'll play in a hour or so
18:37:20*gevaerts isn't sure if he agrees with that one
18:37:48 Quit petur (Quit: *plop*)
18:38:41AlexPWhat is the point of FS #11019?
18:38:47AlexPWhy would you want to do that?
18:39:23freddybgevaerts: Fine with me!
18:42:19freddybAlexP: The reason for #11019 is so that you can have a background and use the −−root option at the same time. I built all the sims when I was testing my other patch and the only way to get the background was to cd to the simulator directory.
18:45:24 Quit kugel (Ping timeout: 264 seconds)
18:46:59 Quit Farthen (Ping timeout: 252 seconds)
18:48:21 Join Farthen [0] (
18:50:21 Join kugel [0] (~kugel@rockbox/developer/kugel)
18:55:08***Saving seen data "./dancer.seen"
18:57:47 Join sudoman [0] (
19:05:22*gevaerts starts on the experimental pebble code
19:05:58gevaertsmaybe I'll get somewhere :)
19:06:10AlexPfinger's crossed :)
19:08:41kugelgevaerts: the virtualisation thing?
19:09:04kugel*you* work on it?
19:09:07gevaertsWe can argue about the details later, I just want a running prototype
19:09:21*kugel didn't see that coming but can't argue against :)
19:11:51 Quit fyrestorm (Read error: Connection reset by peer)
19:13:41 Join fyrestorm [0] (
19:19:13 Join JdGordon_ [0] (~Miranda@rockbox/developer/JdGordon)
19:30:22CIA-8New commit by bluebrother (r24697): Remove column from tree widget −− this is set up by the caller.
19:30:26CIA-8New commit by bluebrother (r24698): Rename InstallFrm to InstallWindowFrm. ...
19:32:23 Quit Farthen (Disconnected by services)
19:32:45 Join Farthen_ [0] (
19:32:47 Nick Farthen_ is now known as Farthen (
19:32:54 Join Strife89 [0] (~michael@
19:33:07 Join efyx_ [0] (
19:34:38pamauryWhat is the difference between Actual size and Binsize in rockbox-info.txt ? Does the linker does some sort of section alignment ?
19:35:42 Join webguest35 [0] (
19:36:32 Join TheUser [0] (
19:36:38webguest35Hi, Whats the best way to get what button is pressed?
19:37:12 Join Bagder [0] (~daniel@rockbox/developer/bagder)
19:42:35JdGordon_webguest35: in what context?
19:43:14kugelpamaury: usually the actual binsize it's 8 byte less because we add a small header to it
19:43:15webguest35In a plugin
19:43:50kugelbut it can be different if the file format (e.g. mi4 on several PP devices) impose restrictions for the layout
19:45:28freddybIs there a right way to get the path of the rockboxui executable?
19:45:30gevaertsJdGordon_: woukd the parse_* function in skin_parser.c be the right place to actually load things from disk?
19:46:00JdGordon_gevaerts: no, probably not, but it depends...
19:46:23gevaertsJdGordon_: it's probably good enough for proof-of-concept code I guess
19:46:25JdGordon_webguest35: rb->button_get() for the raw buttton presses or rb->action_get()
19:46:32JdGordon_whatcha working on?
19:46:48gevaertspebble support :)
19:47:03JdGordon_oh cool
19:47:21webguest35thankyou JdGordon
19:47:55JdGordon_just be careful where you load the data into. if its the skin buffer then you HAVE to use the alloc() stuff and not the get_buffer() stuff or it will explode
19:48:36gevaertsso I have to find out in advance how much I need?
19:49:38JdGordon_crap, I was going to suggest loading it into the plugin buffer as temp space, but that is used for the skin text buffer
19:49:55*gevaerts grumbles a bit but accepts the facts
19:50:12gevaertsI can probably get the file size
19:50:18JdGordon_you might be able to use get_buffer as long as you move the data once its fully loaded
19:50:51gevaertsI guess I'll just see how big my first pebble is, and start with fixed sizes
19:51:11gevaertsI don't want to spend too much time on details until I know that the idea actually works
19:51:46JdGordon_the more time you spend, the more likelyhood of it eventually working
19:51:57JdGordon_+there is
19:52:12 Join moos [0] (moos@rockbox/staff/moos)
19:52:36gevaertswell yes, but I'm not even sure yet that this idea of loadable functions as such will work on all CPUs
19:53:09webguest35@JdGordan when I call rb->action_get() do I store it in an integer or string variable?
19:53:30 Quit bmbl (Ping timeout: 245 seconds)
19:55:35JdGordon_in whatever you like, this is , go nuts, you can store it in a float for all it cares
19:55:46JdGordon_(of course it wont work correctly if you do)
19:56:38kugeldamn still no icon set exception
19:57:29 Part Llorean
19:58:06JdGordon_whats the setting name?
19:58:12JdGordon_arnt' you a theme admin?
19:58:48 Quit shai (Quit: Leaving)
19:59:11 Join AlexP_ [0] (~ap@rockbox/staff/AlexP)
19:59:16 Join username [0] (~kosmaks@
19:59:21 Quit AlexP (Ping timeout: 240 seconds)
19:59:55usernamehello everyone
19:59:56usernamehello everyone
19:59:57kugelwtf man
20:00:05kugelthat cfg parsing is still broken
20:00:15 Join Tomis2 [0] (~Tomis@
20:00:18kugel"The file /.rockbox/themes/-.colours from the setting 'filetype colours' doesnt exist."
20:00:39 Quit freddyb (Quit: Konversation terminated!)
20:01:31 Quit username (Remote host closed the connection)
20:01:34 Join kosmaks [0] (~kosmaks@
20:01:40kugelhm, maybe dumping theme settings is wrong on that one
20:01:41 Quit kosmaks (Client Quit)
20:02:02JdGordon_using - for the inbuilt values may not be such a good idea there
20:02:28 Join username [0] (~kosmaks@
20:02:28 Quit webguest35 (Quit: CGI:IRC (EOF))
20:02:33usernamewell well well
20:02:41usernamei have few questions
20:02:47 Quit Tomis (Ping timeout: 272 seconds)
20:02:47 Nick Tomis2 is now known as Tomis (~Tomis@
20:03:00usernamefirst one
20:03:10usernamewhats new about Santa View?
20:03:25usernamei mean sansa
20:03:28usernameu know
20:03:54 Quit phanboy4 (Read error: Connection reset by peer)
20:04:39usernameand another one
20:05:09usernamecan somebody help me
20:05:15usernamei ve got it
20:05:18AlexP_Only if you ask a question
20:05:28AlexP_And please don't use the enter key as punctuation
20:06:04usernameда иди ты нахуй
20:06:11AlexP_And English please
20:06:12usernameshit =(
20:06:36*gevaerts is totally confused by the skin code
20:06:42usernameive got this thing and i need help
20:06:53AlexP_You really need to ask an actual question
20:06:55gevaertsJdGordon_: I think I understand the parsing bit, but how do I actually draw something?
20:07:43kugelJdGordon_: it should be "filetype colors: -" I guess
20:08:01gevaertsI mean, it must be possible. I have seen things on the wps!
20:08:55JdGordon_in get_line() or get_token_value()
20:09:07JdGordon_the first probably makes more sense
20:09:24JdGordon_remember, this is all from when it really was text lines only
20:11:06 Part username
20:11:43 Join GeekShado_ [0] (
20:12:28JdGordon_we need some way to differenciate between "text/value" tokens and umm... "action" tokens
20:13:39 Quit AlexP_ (Quit: No Ping reply in 180 seconds.)
20:14:08 Join AlexP [0] (~ap@rockbox/staff/AlexP)
20:14:20 Quit GeekShadow (Ping timeout: 245 seconds)
20:15:57sudomani did some benchmarks on my 3G ipod:
20:16:53 Quit Farthen (Ping timeout: 252 seconds)
20:16:59sudomando you think that the fact that my ipod lost power around 3.72V is due to the hardware or the battery?
20:17:36JdGordon_thats about where its expected to die
20:17:37 Nick YPSY is now known as Ypsy (
20:18:27 Quit amiconn (Disconnected by services)
20:18:29 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
20:18:53 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
20:19:55Zagor114 hours is an awesome time prediction
20:20:10 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
20:20:10 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
20:20:10 Quit pixelma (Disconnected by services)
20:20:27 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
20:24:29 Quit kugel (Disconnected by services)
20:24:37 Join kugel [0] (
20:24:49 Quit kugel (Changing host)
20:24:49 Join kugel [0] (~kugel@rockbox/developer/kugel)
20:26:41 Join `Assassin [0] (
20:29:32kugelwho can make theme site changes live?
20:31:05kugelrasher: ?
20:31:07 Nick fxb is now known as fxb__ (
20:32:08rasherkugel: I think me, domonoky and mcuelenaere (and scorche)
20:32:55kugelif this looks sane, I'd like to commit and need someone to make it live
20:34:27rasherAre you sure of the operator precedence? I'd prefer another set of braces to make sure
20:34:40 Quit anewuser (Quit: =ooo ϢINTER ϾHIP 5iVE is OOON!!)
20:34:42sudomani have a small patch for the "gray" theme for the ipod 3g but i've searched and have no way of contacting the creator.
20:35:13kugelrasher: yes, unless php is different to c in that regard. anyway another brace is fine with me
20:35:45rasherI would be very surprised if it is, but just to make things clearer.
20:35:56 Join Stephen__ [0] (
20:37:03 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
20:38:52CIA-8New commit by kugel (r24699): Accept "-" as a valid value for filenames settings (it's generally used to explicitely not load a file for those settings)
20:39:07kugelhighly untested :(
20:39:17JdGordon_don't say that untill after its live
20:40:03kugelrasher: would you be so kind?
20:40:33rasherkugel: I'll blame you if it breaks
20:40:49rasherLive now
20:41:06 Quit flydutch (Quit: /* empty */)
20:42:13JdGordon_you shoulda tested it :)
20:42:33pamauryIs a ~2KB ram use reduction (including 1800B of stack) worth a 300 bytes binsize increase and a hack in dircache.c ? :)
20:42:40kugelJdGordon_: how?
20:42:53JdGordon_that's totally irrelevant :D
20:43:07kugelI don't know how to update the image domonoky gave me (also, the machine it's on isn't on)
20:43:32JdGordon_pamaury: probably not, unless you can actually shrink the stack also and not just the usage
20:43:44JdGordon_depends on the hack though
20:43:52pamauryJdGordon_: I mean shrink stack
20:47:45 Quit Zagor (Quit: Clint excited)
20:53:16 Quit GeekShado_ (Ping timeout: 246 seconds)
20:53:40 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
20:53:47 Join Zagor [0] (
20:53:48 Quit Zagor (Changing host)
20:53:48 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
20:55:12***Saving seen data "./dancer.seen"
20:59:20TheSevenpamaury: which stack? dircache's or the main thread's?
20:59:43TheSevenand what kind of hack?
21:00:04JdGordon_does dircache really needs its own thread/stack?
21:00:13*gevaerts wishes that the linker actually linked his function into the binary
21:00:45TheSevenhow did you get rid of more than half of it's stack usage? dropping a sector buffer somewhere?
21:01:00TheSeventhat might even fix those nano2g dircache stkov's then...
21:01:09 Quit Zagor (Quit: Clint excited)
21:02:18pamauryThe current code uses a fat_dir structure to go through the directories. This fat_dir contains a 3*SECTOR_SIZE buffer which is huge. Previously it was allocated in memory but I moved it to the stack. Now the point is that there are already MAX_OPEN_DIRS of these buffers in dir_uncached and there really cost much ram so I just decided to use one of them during the generation by 'strealing' it. It's a hack because it uses two hi
21:02:18pamaurydden api functions and uses an open dir slot but I think it's not a problem during the dircache generation
21:02:45TheSevenahhhhhh that explains why we need 0x1800 bytes of additional dircache stack on nn2g
21:02:53pamauryTheSeven: what are those stkov's ? In which thread ?
21:03:09pamauryreally ? Where is it declared ?
21:03:13TheSevenremember that nano2g's sector buffers are 2K, not 512
21:03:40 Join panni_ [0] (
21:03:46pamauryThat explain things. I moved the fat_dir to stack because it fitted the stack but my computations use a 512 sector size
21:04:02pamauryMy mistake then, although the previous code was horrible also :)
21:04:03 Join AlexP_ [0] (~ap@rockbox/staff/AlexP)
21:04:13kugelmaybe the stack size should simply computed using the sector size
21:04:37TheSevenso that would free 6K of ramsize on nano2g...
21:04:57pamauryWhat I propose is to get rid of this buffer and use one from dir_uncached just for the time of the generation
21:05:17 Quit AlexP (Ping timeout: 276 seconds)
21:05:48pamauryThere already are several 'steal' hack (like tagcache stealing dircache) in RB but this hack is particular because it fails the generation if dir_uncached in unable to provide a slot.
21:05:52 Quit stoffel (Remote host closed the connection)
21:06:09pamauryBut that's quite unlikely to happen
21:06:31kugelis it save or just unlikely?
21:06:50pamaury'save' ?
21:07:19pamauryIt would fail iif you already have the maximum number of opened directories opened at the time of the generation
21:07:34pamauryAnd it that case, dircache would 'just' be disabled
21:08:45pamauryThat's seems highly unlikely to happen to me, that's why I propose it
21:09:09kugelyou never know what users/other code do
21:09:39kugelI think 2k (it's almost nothing afterall) isn't enough if it's more likely to fail
21:10:10gevaertsDoes look like it might run from wherever I load it?
21:10:31pamauryJust think about where a dircache generation happen: on startup, after tagcache generation and after usb plug-out I think
21:12:26 Quit Schmogel (Read error: Connection reset by peer)
21:12:34 Nick AlexP_ is now known as AlexP (~ap@rockbox/staff/AlexP)
21:12:38pamauryIn a normal situation, the core doesn't open the maximum number of directories, no ? There are buggy plugins that might do so but I don't think the core does it
21:12:40TheSevengevaerts: it should
21:13:48 Join webguest41 [0] (
21:15:05pamauryOk, I have to leave for a few moments, if someone else wants to give his opinion on that...
21:15:07 Join m3dlg [0] (~m3dlg@
21:15:57TheSevenwhat the heck is this doing? function whatever(int throw, int all, int of, int us, int away, int ptr) { (*(ptr + 68))(10, 10, 20, 20); return 0; }
21:16:27gevaertsit calls lcd_fillrect() :)
21:16:36 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
21:16:37JdGordon_obviously :p
21:16:41 Part watto
21:17:12 Join Zagor [0] (
21:17:12 Quit Zagor (Changing host)
21:17:12 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
21:17:22TheSeveni'm still a little confused by the args it gets
21:17:50gevaertsthe other args are not used yet
21:18:10*bluebrother spots that there are *still* problems with the rbutil osx dmg :(
21:18:22gevaertsah, data abort :)
21:18:37gevaertsthat's the way to go!
21:18:45 Quit webguest41 (Client Quit)
21:19:00bluebrotherdomonoky: how do you feel about requiring at least Qt 4.5 in the future? This would allow to get rid of BrowseDirtree, at least if I understood the documentation correctly
21:19:24domonokysure, no problem with that.
21:19:26AlexPJdGordon_: I don't see a patch with your latest version
21:19:31kugelgevaerts: are you fixing up sp correctly?
21:19:38AlexPer, latest comment
21:19:54JdGordon_oh bloody hell :p
21:19:56gevaertskugel: I'm hoping the compiler did that
21:20:13kugelah that was generated code? I thought you wrote that asm
21:20:19AlexPJdGordon_: I could make up some test results if you like :)
21:20:22gevaertsE8000088 doesn't look like a likely pointer
21:20:46 Quit AndyI ()
21:22:20bluebrotherdomonoky: do you think a web translation tool would be a good idea? I have a feeling that rbutil translations are in a bad shape also because translating kinda requires you installing yet another tool (linguist)?
21:22:43bluebrotherI've scrateched something up the last two days. Currently pretty rough, but it at least does something:
21:23:03JdGordon_AlexP: ok, attached
21:23:14AlexPcool, I'll try now
21:24:04domonokysounds like a good idea. are you aware of ? maybe we could integrate rbutil translation into that somehow ?
21:24:22 Join karashata [0] (
21:24:45bluebrotheryeah, I'm aware of translate.r.o −− that was something that got me the idea ;-)
21:25:32bluebrotherhowever, I found that rbutil translations are kinda different. Not sure if it makes sense to use the same kind of "presentation".
21:25:39 Quit mt (Remote host closed the connection)
21:25:45bluebrotherthe overview however does make sense.
21:26:15bluebrotherbut my current version only gets some files from a fixed location, nothing that updates itself right now. That needs changing too.
21:26:35domonokybluebrother: that only needs a cron job :-)
21:27:10bluebrothersure, if you have a machine you can do this on :)
21:27:42*bluebrother remembers that he even could do this on his router and push it to the ftp folder ... :)
21:27:48domonokyask rasher, he already hosts translate.r.o and themes.r.o :-)
21:28:17rasherdomonoky: Nope - scorche does
21:28:30 Quit m3dlg (Ping timeout: 265 seconds)
21:28:35domonokyoh... then ask scorche :-)
21:28:36bluebrotheranyway, there's heaps of work left for a nice translation script. This is just a first "does it make sense and is it useable" test ;-)
21:28:49moosbluebrother, domonoky: if you do so, you can count on me for the french part...
21:29:42TheSevengevaerts: yes, that looks more like an instruction... but a rather weird one
21:30:14TheSevenSTMED R0, {R3, R7} IIRC
21:30:21*JdGordon_ isnt really sure pebbles is the way to go for this
21:30:47kugelgah, I broke my image
21:32:30gevaertsMy pebble-loader works
21:33:04gevaertsAll it needs now is loads of skin code to pass it the correct viewport and things like that
21:33:29gevaertsand testing on non-arm
21:33:32pamaurygevaerts: what does you pebble-loader does ?
21:33:38JdGordon_loads pebbles!
21:33:40gevaertsit loads a pebble!
21:33:58gevaertsI now have a white square somewhere on my wps
21:34:37pamauryhum, either my english is not good enough or you are crazy or I'm stupid ;)
21:34:56gevaertsa pebble is a small rock :)
21:35:03JdGordon_gevaerts: can you change it so the pebble is given the skin text for the line? I dont like the idea of forcing a viewport for these.. especially when this could be useful for regular value/text tokens
21:35:27pamaurygevaerts: then I don't understand what you mean, are you programming a game ?
21:35:50 Join minus [0] (
21:35:56gevaertspamaury: no, the idea is to have some sort of plugin-like things that are loaded into the skin buffer for fancy wps effects
21:36:20pamauryso you are loading code ?
21:36:59pamaurydoes it has a special interface to read/write to screen ?
21:36:59domonokyhm, should themes on the themepage require to contain standard files like the icon set or the colours file, if they use it in their theme.cfg ?
21:37:53JdGordon_there is no standard colours file, but the icons comes in shuold it shouldnt be in the theme zip
21:38:34gevaertsJdGordon_: yes. The idea I'm working with now has an init/exit flag, the wps line, a viewport (that we can scrap), a pointer to some data area (which it should probably allocate itself), and *rb
21:38:48gevaertspamaury: my current code uses the standard plugin api
21:38:54domonokyso same handling for icons as for fonts ?
21:39:06*gevaerts re-works that a bit
21:39:19JdGordon_it also needs to verify the plugin api number
21:39:20kugeldomonoky: yea the tango icons need an exception
21:39:51kugeldomonoky: can you look at my last commit? it seems it didn't work
21:39:53gevaertsyes, but that's a job for the loader
21:40:15kugelif not, I managed to svn up the image so I might manage it myself
21:40:26 Join matsl [0] (
21:41:20domonokykugel: i think your last commit to the themepage needs some more () to work
21:41:25gevaertsI haven't done that yet because (a) it can't be exactly the same as for rocks (those have a header that has a pointer to the entry point, which assumes a fixed address), and (b) I'm not entirely sure yet if we should use the same api as normal plugins (although not doing that costs a lot of RAM)
21:41:29kugelcan I get the settings db file?
21:41:42kugeldomonoky: I thought so too
21:42:26domonokysettings db file ?
21:42:28JdGordon_gevaerts: I would think using the plugin api makes sense
21:42:53gevaertsJdGordon_: yes. It does have some unneeded stuff, but it doesn't make much sense to make a stripped down copy I think
21:42:53 Quit matsl (Client Quit)
21:43:22kugeldomonoky: well, whatever db file the allowed settings are stored in
21:43:22gevaertshm, which category should I use on flyspray? :)
21:43:25JdGordon_do things like globals work in pebbels?
21:43:28kugelor is that in the theme.db?
21:44:04domonokyall data is in the theme.db file.
21:44:12gevaertsI don't know
21:44:27kugelso I need to insert all allowed settings myself now?
21:44:34gevaertsThey can't have their own globals, but I haven't tried normal globals yet
21:44:46domonokyjup. :-)
21:45:24kugeloh dear :(
21:45:46domonokyor you could extend the automated db creation in db.class.php, and make it also fill in data if needed.
21:46:07 Join froggyman [0] (
21:46:29JdGordon_gevaerts: what about functions? or is the whole pebble contained in a single function?
21:46:37gevaertsit's a single function
21:46:53gevaertsMultiple functions would need relocation again
21:47:05TheSevenBL is relative...
21:47:13gevaertswell, you can try
21:47:17kugeldomonoky: adding settings doesn't seem to work in the image
21:47:28TheSevenso at least functions from the same C file should work, don't know what the linker will do to functions from other files
21:47:36kugelI already added myself as admin and added the sansa fuze target so it's not all broken
21:47:43domonokykugel: works for me.
21:47:54kugelmaybe because the table doesn't exist yet
21:48:02*gevaerts is busy uploading the patch to flyspray so people can experiment
21:48:08kugelhow does the table look like?
21:48:23domonokyit should auto create the table.
21:48:39domonokyif you updated the neccessary files in the vm, ofcourse.
21:49:20domonokyall tables are now described in the db.class.php file, and autocreated / columns added if needed.
21:49:40kugelright, the table is there
21:50:02domonokyso what doesnt work ?
21:50:14domonokyany error message ?
21:50:16kugelclicking on add setting leads to the same page
21:50:31kugeli.e. admin overview
21:50:33domonokyupdate index php ?
21:50:48domonokyadmin.php in this case.
21:51:17gevaertsFS #11027
21:51:21minuswhen installing rockbox bootloader on my sansa fuze, does it matter if i got the newest or an older version of the original firmware?
21:52:56gevaertsminus: not as far as rockbox is concerned. The one you pick is the one that's used if you boot to the original firmware though
21:54:35kugeldomonoky: I've no idea what to do there
21:55:14*kugel copes from viewvc
21:55:27domonokykugel: i mean is admin php in sync with svn ?
21:55:59domonokyjust ftp over the current admin.php
21:56:23 Join Tomis2 [0] (~Tomis@
21:56:37kugelthe page is showing now, but it doesn't insert
21:57:58 Join tomers [0] (
21:58:30AlexPJdGordon_: I think I might have killed it - first I got it couldn't open the backdrop or the icon files (rc: -2) and then I got "*PANIC* Stack overflow... viewportmanager"
21:58:50AlexPI just added a couple of very big fonts to the mix when this happened
21:58:56 Quit Tomis (Ping timeout: 252 seconds)
21:58:56 Nick Tomis2 is now known as Tomis (~Tomis@
21:59:01CIA-8New commit by tomers (r24700): Brickmania: Fix FS #10959 - Ball disappears
21:59:06AlexPbefore adding the big fonts it didn't do this
21:59:10JdGordon_i tihnk all that is unrelated
21:59:34JdGordon_the backdrop not loading sounds like buffer full
21:59:58AlexPSkin buffer usage: 547416/784384
22:00:01 Quit Strife89 (Ping timeout: 258 seconds)
22:00:09kugelwee it does insert but it doesnt show
22:00:19AlexPRestarting the sim it is displaying the backdrop etc now
22:01:06AlexPmultifont for the fmscreen is particularly ace
22:01:34gevaertsJdGordon_: no globals except the ones exported by the plugin API
22:01:42 Quit froggyman (Ping timeout: 256 seconds)
22:01:55 Join petur [0] (~peter@rockbox/developer/petur)
22:02:02JdGordon_and any globals the pebble would need could be in the extradata on the ksin buffer?
22:02:08JdGordon_its just functions thats really missing then
22:02:53domonokyis UI viewport a allowed theme setting ?
22:03:00kugelah, I begin to understand how this works :9
22:03:18minuswhat is it with using your rockboxed player as a HID?
22:03:21kugeldomonoky: no, it's a sound setting! :)
22:03:23JdGordon_domonoky: it better be :)
22:03:32gevaertsfunctions *might* work, but I think we'd be depending a bit too much on the toolchain then
22:05:39sudomani'm having difficulty with tools/
22:05:47 Join froggyman [0] (
22:06:03sudomani get this error:
22:07:52sudomani'm on ppc os x 10.5.8
22:08:15CIA-8New commit by Domonoky (r24701): correct config file checking. Allow rockbox standart icons to be missing. correctly allow - as a filetype value.
22:10:51 Join GeekShado_ [0] (
22:11:42*bluebrother hands domonoky a standard :D
22:11:55*domonoky likes flags :-)
22:11:59 Quit tomers (Quit: ChatZilla 0.9.86 [Firefox 3.5.7/20100106054534])
22:12:12kugeldomonoky: thanks
22:12:12 Join GeekShad__ [0] (
22:12:27AlexPJdGordon_: What does preset name map to in the playlist viewer - filename? And to get the "previous" preset do I set the playlist viewer start to -1?
22:13:10JdGordon_use the 3 preset tags for the playlister viewer, %Tn is name i think, and yes -1 will start with the previous one
22:13:31AlexPAh, OK
22:13:49*pixelma also prefers standards over stand arts ;\\
22:13:59bluebrotherstanding arts? ;-)
22:14:20 Quit GeekShadow (Ping timeout: 240 seconds)
22:14:59bluebrotherpixelma: btw, I've got OSX TTS running. There are potential issues with codepage translation, but that's nothing we can do much about except notifying the user about it. Hopefully I'll get it into a committable state the next couple of days
22:15:52 Quit GeekShado_ (Ping timeout: 256 seconds)
22:15:53pixelmaif you need testers on 10.4 ...
22:16:50 Quit danopia (Ping timeout: 245 seconds)
22:16:55pixelmabut someone else asked about it here (and had some problems), he doesn't seem to be around anymore though
22:20:21bluebrotherI've spotted it in the logs. Though there is currently no support for "say", and basically it's just a command line application around the TTS API
22:20:54kugeldid you spot fs#11023 too?
22:20:58AlexPJdGordon_: Do the %al/c/r tags not work with the playlist viewer?
22:21:13bluebrotherwhich, as the API itself, writes aiff files while everything else in the speech generation code wants wav
22:21:16 Join martian67| [0] (~martian67@about/linux/regular/martian67)
22:21:33bluebrotherkugel: yes, I've already commented on it :)
22:21:44kugeldomonoky: now I need to fix rockbox to actually generate '-'
22:21:45bluebrotherthis hdiutil thing is really giving me headaches :(
22:22:04 Quit martian67| (Client Quit)
22:22:08domonokykugel: :-)
22:22:08 Quit martian67 (Quit: Leaving)
22:22:26 Join stripwax [0] (
22:22:37AlexPWhat I'm trying to achieve is previous preset on the left, next preset on the right (of the screen)
22:22:51domonokybluebrother: i used the static, universal lib usb on earlier builds, that seemed to have been fine.
22:26:00JdGordon_AlexP: use %Tf (?) for that
22:26:12JdGordon_preset frequency in the playlist viewer
22:26:46AlexPI've got the preset names in the playlist viewer just fine - I just can't right align it
22:27:55JdGordon_course you can :) fix the viewport
22:28:09JdGordon_you may need to use more than on %Vp
22:28:10AlexPWhat I'd like is %Vp|1|%ar%s%Tn|| but the %ar is ignored
22:28:23AlexPNo, I want the info right aligned in the viewport
22:28:51JdGordon_the penny drops :p
22:29:02AlexPhehe :)
22:29:06JdGordon_i cant remmember if arlc are supported, I dont tinhk they are
22:29:14AlexPThey just seem to be ignored
22:29:20JdGordon_i thought arlc was frequency values :p
22:29:37 Join danopia [0] (~danopia@fullcirclemagazine/developer/danopia)
22:29:37AlexPaha :)
22:29:48bluebrotherdomonoky: the interesting thing is that the static linking options don't work on my Mac setup. Do you remember which version of gcc you did use?
22:30:34 Join Strife89 [0] (
22:31:18CIA-8New commit by mcuelenaere (r24702): Only build Lua overlay on low-mem targets
22:31:28 Join Biont [0] (
22:31:34domonokyi dont know which gcc i used. i just removed the libusb.dylib to force it to use the .a file :-)
22:33:10bluebrotherhmm. Ideally it should work with a dylib in the bundle. Why on earth does it keep using the wrong onw? :/
22:34:12 Quit killan (Quit: ( :: NoNameScript 4.22 :: ))
22:34:42CIA-8New commit by bluebrother (r24703): Use QFileDialog::getExistingDirectory() for path selection. ...
22:35:11AlexPJdGordon_: The fms backdrop isn't overwriting the menu backdrop properly, but otherwise it is all hunky dory
22:36:05pamauryTheSeven: is your dircache stkov systematic ? I can commit a fix (kugel seems against the hack I propose) to have stack size rely on SECTOR_SIZE
22:36:41pamauryOr I can put the buffer in memory instead, so it doesn't use stack (I can then reduce stack size without risk)
22:38:03CIA-8New commit by bluebrother (r24704): Start with mountpoint as folder to talk selection if old value is invalid.
22:38:34 Join martian67 [0] (~martian67@about/linux/regular/martian67)
22:41:28AlexPFM screen + multifont (including messed up background)
22:42:57Stephen__good work AlexP
22:43:15AlexPI did bugger all, it is just a demo of JdGordon_'s work :)
22:44:20gevaertsAlexP: if you apply FS #11027, you can have an extra filled square on there for free!
22:44:31BiontCould someone maybe post a link to a recent D2 windows simulator (please)? I tried compiling one myself and now my filesystem is broken and the image wont even boot. I would like to thest the new multifont feature, but I pretty much gave up on getting my own simulator now :(
22:44:40TheSevenpamaury: placing it in memory won't really make a difference regarding space...
22:44:56pamauryNo but at least it will fix the stkov
22:45:04TheSevenso do whatever is easiest...
22:45:05AlexPgevaerts: tempting... :)
22:45:50pamauryOk, I will put it in memory and reduce stack size, normally it will be safe after that
22:46:09gevaertsJdGordon_: would it be possible to get the enclosing viewport when parsing a token?
22:47:21AlexPJdGordon_: Incidently there are %?C type things all through that .fms and it doesn't seem to mind - not sure what the issue was the other day
22:48:38AlexPgevaerts: Did you start this purely so that as the originator you got to pick the name pebbles? :)
22:49:16 Join Strife1989 [0] (
22:49:25evilnick_BAlexP: Do you really have to ask that question? :)
22:49:25 Quit Strife89 (Read error: Connection reset by peer)
22:49:38gevaertsAlexP: what do you think? :)
22:49:58AlexPgood work :)
22:51:33kugelAlexP, JdGordon_: it is possible that %C makes rockbox crash if MAX_ALBUMARTS isn't increased
22:51:53kugelit's currently 2, so if you have 3 it may simply overflow the albumart slot array
22:51:56 Join Buschel [0] (
22:51:56 Nick Ypsy is now known as YPSY (
22:52:22AlexPI only have two I think - one on the WPS and one on the FMS, although the FMS doesn't do anything currently
22:54:11kugelhm, no it shouldn't be possible actually
22:55:16***Saving seen data "./dancer.seen"
22:55:43CIA-8New commit by mcuelenaere (r24705): Fix parallel building
22:58:20kugelgevaerts: that actually works? :)
22:58:40 Quit ecio (Ping timeout: 258 seconds)
23:01:20gevaertskugel: the pebbles thing? Yes, at least for simple code
23:01:32kugelso, port doom now :)
23:03:31kugeldoom-as-pebble for some shooter action in the wps :)
23:04:25 Quit n17ikh (Remote host closed the connection)
23:04:37*gevaerts leaves that to kugel :)
23:05:08stripwaxBiont - wait, you built a simulator and your "image" won't even "boot" ? What does that mean? Did you "make" and "make install" ?
23:05:54gevaertsgah, the usual memcpy/rb->memcpy issue
23:05:55bertrikis there a website where I can find out more about this pebble thing?
23:06:01Stephen__i take it he measn his vmware image wont boot ?
23:06:09gevaertsbertrik: FS #11027
23:06:21 Join saratoga [0] (
23:06:31saratogastripwax: did you see my pastebin last night?
23:07:09 Join m3dlg [0] (
23:07:17 Join n17ikh [0] (
23:07:25stripwaxsaratoga - yep, and saw that you committed it too :)
23:07:38stripwaxI'm wondering how you treated the Inverse flag? (Assumed it was 'True', I assume?)
23:07:55Buschelhmm, is shutting off @3700 mV normal for 3G iPods? and 4 hours runtime is also normal?
23:07:58stripwaxI've just realised that codeclib will only have an ifft, and not an fft, this way. No biggie, I think
23:08:00saratogastripwax: not all of it
23:08:06stripwaxBuschel - 4 hours is not normal.
23:08:30saratogai posted a pastebin that switches wma over to having no init, but i wasn't sure how tremor works so i didn't change that
23:08:43Biontstripwax: The vmware image of debian itself is broken :) I have no idea what wnet wrong. I'm deleting and reinstalling everything atm
23:08:44saratogaright now theres still an init function but it doesn't do anything beyond set the number of bits in the transform
23:08:47stripwaxsaratoga - yeah, I've actually just finished making all those changes for all the codecs; just about to test+commit
23:08:51stripwaxso - no more init
23:08:53saratogaoh ok
23:08:57saratogago for it
23:09:04saratogawma and atrac are easy but tremor was confusing
23:09:11saratogaac3 already has no init
23:09:26stripwaxI had to remove the init for atrac but yeah, it looks easy (Always 2^9 right?)
23:09:36stripwaxI have absolutely no way of testing atrac3, though
23:09:41saratogajust make the transform take nbits
23:09:45stripwaxdid that :)
23:09:55saratogawell 9 not 2^9 right?
23:10:57stripwaxwell, yeah ...
23:11:22stripwaxok, wma and vorbis seem to work. I'll assume that my change to remove atrac3 init didn't break it. committing ...
23:11:36stripwaxactually that was just a test on sdl, I'll check target also builds..
23:11:51pamauryTheSeven (and others too): What do you consider the acceptable max depth of directory recursion for dircache before a stkov happens ?
23:11:54saratogasounds good
23:12:09stripwaxBut you agree that codeclib now has only ifft and no longer has fft ?
23:12:11saratogaAFAIK I am concerned you can merge whenever you like
23:12:18saratogaopps that wasn't right
23:12:31saratogawe can fix that later if we need it
23:12:36stripwaxas far as I know, I am concerned also :)
23:12:56*domonoky sees new codec patches from Yoshihisa Uchida, can some codec guy look at them ? (new codecs are always good :-) )
23:13:16TheSevenpamaury: how much space do you need per depth level?
23:13:16stripwaxsaratoga - I'll do a coldfire target comparison vs trunk before merging?
23:13:25stripwaxfor wma+vorbis at least
23:13:37kugeldomonoky: he should be able to commit himself by now..
23:13:42Buschelanyone else has runtime infos on 3G iPods? the bench by sudoman shows 4h runtime and the bench ends before the typical knee in the voltage curve
23:14:09saratogastripwax: so to add the inverse transform back we'd need another bitrev table
23:14:14stripwaxsaratoga - would you mind doing the merge, by the way? I'm a little ... petrified ... by svn merge
23:14:14 Quit saratoga (Quit: CGI:IRC)
23:14:20 Join saratoga [0] (
23:14:21stripwaxsaratoga - (oops, where did you go?)
23:14:24stripwaxsaratoga - yeah
23:14:26pamauryTheSeven: on my sansa200 (it should be the same for all arm more or less), 30o per level plus an incompressible constant of approximately 800o. With DEFAULT(0x400)+0x200 that's 20 levels for example
23:14:44saratogastripwax: i figured you should do the honors since most of the work was yours, but i can if you like
23:15:04stripwaxsaratoga - so I'm not sure if it's just a bitrev table by the way. (I'd have thought the output would need to flip the Im sign at some point..)
23:15:24stripwaxsaratoga - oh, yeah - before merging we should actually compare the *output* (quality and similarity) to trunk.....
23:15:45stripwaxI have a nasty feeling that something in mdctexp might be flipping the sign of the entire output. (gut feeling, not actually observed)
23:15:50stripwaxjust from the math
23:16:02 Join checker [0] (
23:16:07saratogaisn't the sign arbitrary anyway for codecs?
23:16:14saratogatheres no absolute phase defined
23:16:25saratogabut i suppose fixing that would be good for neatness sake
23:16:33saratogaif it is a problem
23:16:44bertrikI'm sure you can find an audiophile who disagrees about that :)
23:16:56CIA-8New commit by stripwax (r24706): Removed mdct init. No more init. Yay!
23:17:10saratogaright now the only place the inverse flag is used is the bitrev computation function
23:17:16pamauryHum, I just saw mcuelenaere opened a task for an address error on mips the dircache I committed two days ago, did he said anything about that ?
23:17:16TheSeventhen i'd go for 32 levels or something
23:17:18saratogaso i don't think theres a sign flip involved
23:17:25saratogaanyway, have to run
23:17:25stripwaxsaratoga - worth checking. sign *isn't* arbitrary, if, say, you crossfade one track with itself under two different codecs, and happen to find that it cancels out!!!
23:17:29 Quit karashata (Quit: The fluffy dragon has left entirely!)
23:17:31TheSevennobody will ever exceed that, and it still isn't using too much space, just like 700 bytes
23:17:39TheSevenwhat is it currently for 512-byte targets?
23:17:48stripwaxsaratoga - yeah that's what I thought too, but wasn't completely 'convinced' ..
23:18:03saratogastripwax: it won't cancel out since they won't be phase matched
23:18:11pamauryTheSeven: It's independent of sector size because I put the sector size dependent part in memory.
23:18:24saratogaaudio streams are low coherence, so to cancel out you need to have identical phase delays
23:18:29saratogathe front and back of a track won't have that
23:18:35pamauryFor 32levels that's 0x700 bytes if I compute correctly
23:18:38stripwaxsaratoga - ah, no, I don't mean that.
23:18:41saratogaso you get an incoherent sum, rather then a coherent sum
23:18:52saratogaand an incoherent sum is not sensitive to the sign
23:19:06saratogaits the difference between combining flash lights and laser beams
23:19:07kugelcan someone quickly explain why this mdct init was so bad?
23:19:07stripwaxI mean if I mix the start of one track, with the start of the same track on a different codec that is incorrectly flipping the sign of the output, you get silence rather than twice the amplitde..
23:19:18saratogano you don't
23:19:27saratogathey'll mix incoherently
23:19:28stripwaxso (-A) + (A ) != 0 ?
23:19:37TheSevenpamaury: i mean how much recursion does it support currently, with the buffer still on the stack?
23:19:40saratogasqrt[2] the amplitude
23:19:45saratogatry it!
23:19:46stripwaxsqrt(0) = 0
23:20:05pamauryTheSeven: the previous code limited the depth to 16 also so if I put it to 32, I will increase RAM usage (but that's will be negligeable with the 30kb of the previous commit)
23:20:07saratogayou'll only get 2x the amplitude or zero amplitude if you have equal time delay
23:20:14saratogaso you can't do that with the beginning and end of a track
23:20:20saratogaonly if you tried to cross fade to beginnings
23:20:22stripwaxright, *not* talking about beginning and end
23:20:32stripwaxtalking about the start of the track plus the start of the same track on a different buggy codec
23:20:49stripwaxIf you think it's not a valid concern and not worth testing, that's fine too ...
23:21:06saratogahow would that happen?
23:21:10stripwaxif there was a bug
23:21:11kugelJdGordon_: did your backdrop commit also make same wps+fms backdrop work?
23:21:12stripwaxjust sayin'
23:21:14saratogai mean you'd need to cross fade sample per sample identically
23:21:18 Quit bluebrother (Disconnected by services)
23:21:18 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
23:21:29gevaertsok, pebbles as currently implemented can't handle rodata
23:21:41saratogaanyway got to go
23:21:49stripwaxsaratoga - I see your point. ok, see you
23:21:58TheSevenpamaury: keep it at 16 then
23:22:05pamauryTheSeven: I didn't made the computation but with the buffer on stack, on a 512-sector size, that's would be approximately 20
23:22:10TheSeventhis means that nobody seems to be using more depth anyways
23:22:40kugeldid someone change the sysfont? the keyboard input screen looks different now
23:23:05pamauryOk, I'll put it around 16/20
23:24:07 Join karashata [0] (
23:24:30 Quit petur (Remote host closed the connection)
23:24:53JdGordon_kugel: it should have
23:25:29 Quit n17ikh (Remote host closed the connection)
23:25:49kugelhm, I think I haven't updated to that revision yet, enabling a fms backdrop made other icons disappear (or show a inbuilt version)
23:25:56JdGordon_AlexP: grr, I thought I fixed that already
23:26:17JdGordon_gevaerts: the current viewport is curr_viewport->vp iirc
23:26:20JdGordon_in the parser
23:26:57JdGordon_kugel: I up the AA count I think
23:27:09 Join n17ikh [0] (
23:27:23stripwaxsaratoga- oh yeah, other thing that hasn't been tested is 8192-size blocks on vorbis. I just don't seem to have files that used them ....
23:27:57*stripwax would like to meet someone who has vorbis files that use 4096-size blocks or 8192-size blocks .. or at least meet the vorbis files ..
23:28:05 Quit karashata (Client Quit)
23:28:14gevaertsJdGordon_: ok, so we could technically go for a scheme where you define a viewport with only a pebble in it, and pass the viewport to the pebble
23:29:37 Join karashata [0] (
23:29:44gevaertsWe first have to get the things to actually work properly of course
23:29:46saratogastripwax: [back for one minute] someone told me that vorbis q-1 uses 4096 blocks
23:29:51saratogawant me to make you one?
23:30:10stripwaxmight have been me who said that :) yeah, if you could, would be great!
23:30:30stripwaxI think 8192 blocks are used if sample rate > 64kHz (if you could knock up one of those two that would be fantastic)
23:30:42stripwaxOh - and add to the rockbox download test files, of course.... :)
23:30:50-->"hello" received from d4red3vil (
23:30:57CIA-8New commit by pamaury (r24707): Fix several misused of 'struct dirent' instead of 'struct dircache_entry' in tagcache.c.
23:31:58 Nick fxb__ is now known as fxb (
23:32:12JdGordon_gevaerts: and decide if pebbles really are the way to go
23:32:25*JdGordon_ thinks boulders should be better
23:32:37*gevaerts throws obelisks at JdGordon_
23:33:00JdGordon_back in 10
23:33:04kugel%Tn1 is the next preset?
23:33:55 Quit karashata (Client Quit)
23:36:32 Join shai [0] (
23:37:12 Join karashata [0] (
23:38:11 Quit domonoky (Read error: Connection reset by peer)
23:38:13kugelAlexP: yea, %ac etc don't seem to work
23:41:56kugelin the playlist viewer
23:43:29kugelpretty useless this way, imo
23:47:10JdGordon_kugel: no, I got rid of the numbers on the %T tags
23:47:16JdGordon_you have to use the playlist viewer
23:47:44kugelthat one sucks though without align tags
23:48:16JdGordon_well, the freq is always 00.00 so just create a viewport of that size and you're set
23:48:35kugelthe name?
23:48:47JdGordon_goes in a seperate one
23:48:57kugelit doesn't have fixed length though
23:49:13CIA-8New commit by pamaury (r24708): -Cosmetic change in a comparison ...
23:49:18JdGordon_ is the way to use playlist viewer... 18 seperate ones!
23:49:38JdGordon_adding support for the alignment tags will bloat the code
23:49:59kugelit's too limited this way
23:50:08pamauryTheSeven: this commit should fix the nano2g stkov, if you have time, can you try it and report it to FS ?
23:50:47 Quit karashata (Quit: The fluffy dragon has left completely!)
23:50:51JdGordon_no its not
23:51:02kugelI can'T right align the next station name
23:51:16JdGordon_yes you can
23:51:25JdGordon_put the two viewports next to eachother
23:51:52 Nick fxb is now known as fxb__ (
23:51:56kugelhow does it solve right aligning?
23:52:11TheSevenpamaury: booting my build machine right now
23:52:17kugelI don't want the station name in a viewport that's on the right, I want the name to be right aligned
23:52:45kugelif I wanted center align it would be even worse
23:52:59JdGordon_1) why would you want that, and 2) thats not what I read your complaint to meaning
23:53:02 Join karashata [0] (
23:53:02gevaertsgevaerts/pebble.png"> \☺/
23:53:13JdGordon_a clock?
23:53:17JdGordon_a real live updating clock?
23:53:25gevaertsnot viewport-aligned yet
23:53:43kugelgevaerts: nice :)
23:53:56CIA-8New commit by Buschel (r24709): Add iPod 3G and iPod Color/Photo (FS #9072) current consumption and battery discharge curves to allow reasonable runtime estimation.
23:53:56Biontwow, how does that work?
23:54:06gevaertsBiont: magic :)
23:54:20*gevaerts is unfortunately serious
23:54:43gevaertsI patch up the GOT table while loading, based on a trial and error method that happens to work for me
23:54:47kugelJdGordon: you mean "why would you want to center or right align the station name"?
23:55:03gevaertsThe .pebble file is 376 bytes
23:55:07TheSevendang, why is somebody committing something while i'm trying to get my working copy up to date? :-P
23:55:15JdGordon_center maybe, right though seems odd to me
23:55:38TheSevenGOT table?
23:55:55kugelI can't do either, which means I see little sense in using it
23:56:02*JdGordon_ refuses to give up his "adds the most amount of pretty crap the skin stuff" crown
23:56:10gevaertsTheSeven: Global Object Table. It's basically simple relocation
23:56:17 Quit m3dlg (Read error: Connection reset by peer)
23:56:28TheSevenwhy can't you just use position-independent code?
23:56:33stripwaxso the pebble thing and the visualisations thing are kindof the same? I haven't actually really been following that particular discussion
23:56:35TheSevenis gcc too stupid for that?
23:56:46gevaertsTheSeven: basically, yes
23:56:49kugelI think the playlist viewer is way to limited to fully replace next-tags
23:56:49 Join m3dlg [0] (
23:57:28gevaertsstripwax: pebbles are my attempt at visualisation plugins. They're still pretty much in experimental proof of concept stage
23:57:37kugelgevaerts: is your configure change needed or does it result from copy&pasting plugin code?
23:57:50CIA-8New commit by theseven (r24710): Hopefully fix Nano2G rolo again. We have enough IRAM anyways.
23:58:22gevaertskugel: I could use the same define to control both plugins and pebbles. I'm not sure what's best
23:58:33BiontHm...I just don't get vmware to work. I think I'll settle for waiting for new simulators to appear on the extras site :(
23:58:50kugelgevaerts: pebbles surely could be enabled unconditionally?
23:58:57sudomanBuschel: do you think i should try to get a new battery?

Previous day | Next day