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 2005-10-12

00:00:33Moosso, why don't commit? :)
00:03:28preglowmight as well
00:03:59preglowso, where should i stuff the main mul64 code? currently i just pasted it in the top of mpc_decoder.c, completely out of place
00:05:21linuxstbWhat about codeclib ?
00:05:46preglowyes, that does indeed sounds more decent
00:06:31preglowbut then we have duplicated code, but keeping mul64 in the plugin api sounds a bit... strange...
00:06:39preglowamiconn: what say you?
00:07:13amiconnPlugin api is definitely wrong
00:07:22preglowyup, agreed
00:07:25amiconnAre there other plugins using 64 bit math?
00:07:32preglownone that i know of
00:07:47preglowand i've tried my best to eliminate 64 bit muls in the codecs
00:09:35 Quit Moos ("Glory to Rockbox")
00:09:46preglowonly codec visible place to keep it is in codec.h and codec.c
00:09:58preglowthere is no codec.c....
00:10:09preglowwell, codec.h is the only header codecs include from rockbox itself
00:11:58amiconnIn fact *all* codecs except a52 and alac use 64 bit multiplies...
00:12:15amiconn$ grep -r -l 'muldi3' −−include *.map .
00:12:20amiconnin the H1x0 build dir
00:13:08amiconn6 codecs and the core itself
00:13:48preglowbut only libmusepack in performance critical spots
00:14:11preglowall codecs use long long maths for position calculation and so forth
00:14:30amiconnHow about placing it in codeclib.c ?
00:14:41preglowyup, but i'll have to declare it in codec.h
00:15:07preglowkeeping the code in codeclib.c and declaration in codec.h is a bit nasty
00:15:26 Part LinusN
00:15:30amiconnWhy not declaring it in codeclib.h ?
00:15:37preglowbecause codecs do not include codeclib.h
00:15:47preglowthey only include codec.h
00:15:56amiconnHow is that codec used, then?
00:16:45preglowwe could of course just start including codeclib.h as well
00:16:49preglownothing wrong in that, i suppose
00:16:59preglowfor those codecs that need it
00:17:08amiconnIn fact some functions are declared twice, in codec.h and codeclib.h
00:17:20amiconnSomehow that does look messy to me...
00:17:29preglowhow can that be? most codec plugins include both
00:17:34preglowi'd think gcc would complain
00:17:58amiconnDouble declaration doesn't hurt if both declarations are identical
00:18:14preglowi don't get the header system for codecs right now, i'd love it if all codec support stuff was in one set of source files
00:18:52preglownamely codec.h
00:18:56amiconnThe header system does indeed look messy
00:19:19preglowthey used to include tons of include files, yesterday i trimmed many of them down to just codec.h and lib/codeclib.h
00:19:28preglowstill, they don't include the one file they should: codecs.h
00:19:44preglowwhich define the main codec api
00:20:29linuxstbI think there are two things - functions needed to interface with Rockbox (the codec API) and helper functions needed by the codecs (codeclib).
00:20:52linuxstbBut I agree it's a mess at the moment.
00:20:55preglowand the last is currently split into both codec.h and codeclib.h, it seems
00:21:10amiconn...and xxx2wav.[ch]
00:21:21preglowyes, those are still there as well
00:21:25preglowthis needs to be rectified
00:21:40linuxstbMy suggestion would be to keep codeclib.[ch] and remove everything else.
00:22:06preglowyes, i think that sounds ok
00:22:32amiconnpreglow: The gcc generic arithmetic rightshift by n for long long looks like it could use some optimisation as well :)
00:23:42preglowamiconn: memcpy! :)
00:24:03amiconnDoes musepack copy a lot?
00:24:14preglowbtw, if you've got #define malloc codec_malloc, will that define trigger on the later usage of codec_malloc?
00:24:18preglowamiconn: nah
00:24:22preglowamiconn: not that i know of
00:24:49amiconnno, that define works the other way round
00:25:40preglowputting everything in codeclib.[c|h] should work just fine
00:26:45preglowanyone oppose that plan?
00:26:59preglowperhaps just move it from lib/ to codecs/ while we're at it?
00:28:05linuxstbI was just about to suggest the same thing - move it to codecs/
00:28:52amiconnHmm. That would mix this source with the codecs itself - and we need it as a library
00:29:18preglowwell, the name in itself should be more than enough to ensure people see it's a library
00:29:27preglowor do you means it's a build system issue?
00:34:35linuxstbWe could keep it in the apps/codecs/lib directory, but then add that directory to the INCLUDES - so we can #include <codeclib.h> instead of "../lib/codeclib.h"
00:35:21preglowsounds sane enough
00:35:48preglowoops, i broke mpa.c yesterday, think i need to fix that
00:37:17preglowlinuxstb: btw, i can't get a52 audio to reach full scale on the peak meter, you think it's possible liba52 is configured to not give full scale audio as a52.c is now?
00:38:03preglowlinuxstb: i didn't really understand how the level parameter to a52_frame works for fixed point audio
00:38:54linuxstbI think that's normal for AC3 encoding - it's generally at a low level.
00:40:20preglowlinuxstb: aphex twin - come to daddy decodes full scale in foobar2k here
00:41:49preglowlinuxstb: is vbr common for ac3 files? if not, i might try implementing seeking
00:43:11linuxstbpreglow: I asked the same question yesterday (about VBR). No-one answered, but I don't think I've ever seen a VBR file. So seeking should be trivial.
00:45:00linuxstbMaybe we need to disable dynamic range compression - it's the "-r" option in a52dec
00:45:11preglowlinuxstb: seems a52.c produces slightly lower output than normal
00:45:23preglowhmm, yes, i thought about that
00:45:31preglowi can't imagine why anyone would want that in a portable player
00:46:34Paul_The_NerdCan anyone think of a reason why I'd need to put an extra blank line in a .wps? Like, I put two lines, on separate lines, and they get displayed as one unless I separate them by a completely blank line.
00:47:00preglowlinuxstb: for cbr files, it should just be a case of fseeking, then walking forwards in the stream and probing for sync with a52_syncinfo()
00:47:17linuxstbLooking at a52dec.c (in the original distribution), the patch would seem to be to add the line "a52_dynrng (state, NULL, NULL);" after the call to a52_frame() in the main decoding loop
00:47:40preglowi'll give it a try
00:48:44linuxstbI think the code already takes care of syncing if there are frame errors. So we can probably just seek and set the decoder going again.
00:49:15linuxstbWe should be able to seek to the correct location anyway in the case of a CBR file.
00:50:00 Quit _FireFly_ ("Leaving")
00:50:07preglowlinuxstb: yes, looks like it does
00:52:12preglowlinuxstb: hah, that did the trick, alright
00:52:23linuxstbThe a52_dynrng() function?
00:52:26preglowlinuxstb: yes
00:53:07preglowi even made it clip
00:54:43 Join jeddy3 [0] (
00:55:06preglowlinuxstb: i need to know the frame sizes of ac3 files to do seeking, thebn
00:55:22preglowand btw, should i commit the dynrng thing?
00:56:04linuxstbLet me test it with my files - give me a minute.
00:57:54 Quit ashridah ("Leaving")
01:00:52 Join lostlogic [0] (
01:05:23linuxstbpreglow: I would say go ahead and commit the dynrng change.
01:05:38preglowwill do
01:08:15preglowbut ok, anyone lined up for doing the codeclib change, or should i have a go when i've fixed mpa.c?
01:09:32 Quit paugh (Remote closed the connection)
01:11:04 Quit lostlogic_ (Read error: 110 (Connection timed out))
01:23:00preglowlinuxstb: and btw, the mad_timer stuff in mpa.c, any reason for it? i don't think we'll ever need it
01:26:04amiconnlibmusepack uses floating point even in fixed point mode...
01:26:21linuxstbpreglow: It was there originally because I copied the code from a libmad demo program. So no real reason for it to be there.
01:27:24preglowamiconn: pretty magnificent, huh
01:28:05preglownot only floating point math, it even uses double precision
01:28:14amiconn...and some denglisch function names in requant.c
01:28:23***Saving seen data "./dancer.seen"
01:28:54preglowit's quite pretty
01:29:03linuxstbTime for bed. Goodnight all.
01:29:11preglowlinuxstb: nite
01:29:44amiconnmy favourite: mpc_decoder_initialisiere_quantisierungstabellen()
01:30:43preglowhahaha, yes
01:36:06amiconnmpc_decoder_scale_output() looks incredibly inefficient to me. Do you know how often that is used?
01:37:38amiconnIt goes through a 128-iteration loop doing 2 double precision calculations per round, plus 2 calls to find_shift() which is rather unefficient by itself
01:39:50preglowamiconn: it's just an init routine, afaik
01:40:00preglowany other use would be disastrous
01:47:48 Quit Nilisco (Remote closed the connection)
01:48:26 Join Nilisco [0] (
01:54:26amiconnpreglow: I have something for you:"> It's untested yet...
01:56:28preglowamiconn: you have any idea if it actually is faster or not? i can't really time it very well
01:57:05amiconnIt should be nearly 2 times as faster for 0 < shift < 31, even more for shift >= 32 and shift == 0
01:57:13amiconn*as fast
01:58:03preglowwill try it
02:00:13amiconn(compared to gcc's ashrdi3() )
02:01:07preglowamiconn: seems to work just ifne
02:01:09preglowfine, yes
02:01:49amiconnI get the impression that quite some of the gcc library routines are compiled C routines with low optimisation level
02:03:05amiconnIf you wanna compare:"> starting at 0x32f0a908
02:03:20amiconnAll that move'ing ...
02:04:05preglowyou know, that wouldn't surprise me
02:04:16preglowit seems musepack got even faster now, yes
02:05:41amiconnHmm. Musepack uses the 64bit mul always in conjunction with the 64 bit shift, either variable or by 14
02:05:56amiconnWe could make that 2 combined functions
02:06:08preglowplease don't invest too much time in this
02:06:13preglowi hope it's temporary, after all
02:06:15amiconnThis would save a number of stack accesses
02:08:25amiconnBtw, I did some calculations concerning memcpy. DRAM is so slow that it might even pay off to shift the whole line in-registers for odd alignments order to get burst writes addition to the (always possible) burst reads
02:09:24amiconnBut this shifting would need 16 (!) versions of the main loop...
02:18:59 Quit JoeBorn ("")
02:23:52 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
02:35:12preglowamiconn: but yes, i highly recommend you telling the gcc people about these routines
02:52:04 Quit solexx (Read error: 113 (No route to host))
02:52:47preglowbut bed, later all
02:53:40 Join BirdFish [0] (n=bradbox8@
02:53:57BirdFishDid anyone ever hear from the IAudio port?
02:54:41 Nick gromit` is now known as gromit_nique_le_ (n=gromit`
02:54:52 Nick gromit_nique_le_ is now known as gromit (n=gromit`
02:55:20 Nick gromit is now known as gromit` (n=gromit`
03:00:18markunBirdFish: There hasn't been info from the only developer in a long time I think.
03:00:36BirdFishBummer :(
03:01:00BirdFishToo bad I really don't know about firmware, otherwise I would jump onboard.
03:10:46 Quit Vladoman (Read error: 110 (Connection timed out))
03:11:56 Join Vladoman [0] (
03:18:23 Quit Mxm`Pas`Bien (Read error: 104 (Connection reset by peer))
03:22:24 Join Maxime [0] (
03:28:25***Saving seen data "./dancer.seen"
03:29:59 Join Vlad0man [0] (
03:41:28 Quit Vladoman (Read error: 110 (Connection timed out))
04:03:57 Join Vladoman [0] (
04:05:25 Join QT [0] (i=as@madwifi/users/area51)
04:08:34 Quit Vlad0man (Read error: 110 (Connection timed out))
04:18:19 Quit QT_ (Read error: 110 (Connection timed out))
05:13:36 Quit tvelocity ("Leaving")
05:28:30***Saving seen data "./dancer.seen"
06:07:01 Quit Maxime ()
06:14:22 Quit _coder` (Remote closed the connection)
06:21:41 Join coder` [0] (i=shani@Ariel.Atlantica.US)
06:42:24 Join Lynx0 [0] (
06:43:44 Quit Lynx_ (Read error: 104 (Connection reset by peer))
06:43:45 Nick Lynx0 is now known as Lynx_ (
06:55:45 Join amiconn_ [0] (
07:14:15 Quit amiconn (Read error: 110 (Connection timed out))
07:14:16 Nick amiconn_ is now known as amiconn (
07:18:26 Join webguest39 [0] (
07:20:28webguest39hi all
07:21:34webguest39I have a rockboxed recorder10 with a 20 gb drive in it. I WAS very happy with it until a coffee cup shattered the lcd.
07:22:03 Quit webguest39 (Client Quit)
07:22:08 Join webguest39 [0] (
07:22:25 Quit webguest39 (Client Quit)
07:22:30 Join webguest39 [0] (
07:22:54webguest39whats my best course of action?
07:23:11webguest39I can buy a new lcd for about $35 shipped
07:28:22 Quit webguest39 ("CGI:IRC (EOF)")
07:28:31***Saving seen data "./dancer.seen"
07:32:33Bgermorning :)
07:33:04Bgerthis fscking M$... wtf they are using cp866 for cyrillic in SMB shares
08:06:19 Join ender` [0] (i=ychat@
08:22:04*Bger is impressed by amarok's capabilities
08:49:49 Join Zagor [0] (
09:00:46 Join B4gder [0] (
09:11:04zewhoa, i just accidently found and noticed:
09:11:25B4gderdoes it rock the box?
09:11:50zeactually its not bad
09:13:36zei look in the dir its in for other z* artists
09:13:40zeand there's a Zagor/ in there
09:13:55B4gderit is the same Zagor
09:14:10B4gderwe are from the C64 camp
09:14:22B4gderI wrote the music editor/routine
09:14:32B4gderLinusN and Zagor composed
09:14:46B4gderthe music part that is
09:14:53B4gderwe did demos as Horizon
09:15:51zei don't see any linus or horizon in here
09:16:19zeah yep thats there
09:16:20B4gderwas linus' handle
09:16:49zethis rockbox track was pretty long
09:26:22BgerD$85 :))
09:26:41B4gderI was young!
09:26:54ryan_jB4gder: how old were you guys then?
09:27:09Bgerwhere does your old and current nicks come from ? ;)
09:27:10B4gderI turned 15 1985
09:27:22B4gderDS are my initials, Daniel Stenberg
09:27:30Bgeroh, yeah
09:27:35B4gderand I got the C64 '85
09:27:54B4gderand $ was just much cooler than S
09:28:00Bgeri know your name very well :P
09:28:21B4gderbut the name was not possible to pronounce and I never liked it
09:28:32***No seen item changed, no save performed.
09:28:38B4gderso I just invented this new one after a few years
09:28:51B4gderwhich was supposed to be the animal at first
09:28:58B4gderbut I got it wrong
09:29:04B4gderand have been misspelled since then
09:30:03Bgeri don't think the "to badger" describes u
09:30:24zeoh yeah its even in these tracks...
09:30:25zeModule Name: Arpeggio Beat
09:30:25zeAuthor : Bjrn Stenberg (Zagor)
09:30:25DBUGEnqueued KICK ze
09:30:25zeCopyright : 1989 Horizon
09:30:46Zagorrock on! :-)
09:30:50zeexcept it doesn't say Bjrn :/ i dunno whats wrong with my crap for that stuff heh
09:31:08BgerBjrn = Zagor
09:31:16B4gderah, I think he's better named Bjrn anyway
09:31:27zeright it has the umlaut'd or whatever o in there though
09:31:47zeexcept with my font/charset/whatever its like a divide symbol heh
09:32:10BgerZagor cyrillic "c" here :)
09:32:25B4gderyes, charsets and IRC are not standardized
09:32:58zewell its in a term being displayed by sidplayer
09:33:17 Join solexx [0] (
09:33:24zein irc it shows up as 3 chars heh
09:33:38zei think
09:33:45Zagoroff-topic: does anyone know how to make mozilla thunderbird use proper quoting?
09:34:21B4gderproper as in how?
09:34:38Zagorwith > prefix, not silly html:ized junk
09:34:46zeit uses >
09:34:51Zagornot for me :-(
09:34:52B4gderI believe it does right when you disable html
09:34:52zeit just renders it differently in its display
09:35:05zeZagor: you look at the source?
09:35:10ZagorB4gder: how do I do that?
09:35:42B4gderI'm not really sure, but I know you can set specific users in the address book to have certain settings
09:35:50B4gderlike my wife has me set to no-HTML
09:36:01B4gderso I get plain text mails with > in the quotes
09:36:04B4gderfrom her
09:36:26B4gderthere must be a "global" setting for that
09:37:12zeedit->preferences->composition->send and html options->configure text format behaviour
09:37:16zei guess
09:37:45zeexcept thats about when sending messages in html format to start with actually...
09:37:56zeand then it says use the address boook to specify preferred text format for recipients
09:38:57zeedit->account setting->composition & addressing has a checkbox for "compose messages in html format"
09:39:15Zagorahh, there it is. thanks!
09:39:57zemine still renders >'s as a blue line even when they're really >'s though (and its all plain-text) hehe
09:42:18 Join ryanj42 [0] (
09:42:40 Quit ryanj42 (Client Quit)
09:43:12 Join ryan_j_ [0] (
09:44:01BgerB4gder search for problesm in the
09:44:40 Quit ryan_j (Read error: 110 (Connection timed out))
09:46:01B4gderdyslexic we are ;-)
09:48:54Zagorze: you should take a look/listen at my vic20 track:
09:49:30 Join LinusN [0] (
09:50:19zehehe oldschool madness
09:50:34zeZagor: i'll check them out tomorrow, going to bed now
09:50:51zemaking glass beads and getting dichroic glass tomorrow at school... woo
09:53:13Bgernight, ze
09:58:02 Join ryan_j [0] (
10:02:19 Quit ryan_j_ (Read error: 110 (Connection timed out))
10:06:24 Join ryan_j_ [0] (
10:06:32 Quit ryan_j (Read error: 104 (Connection reset by peer))
10:39:15 Quit phaedrus961 ("Leaving")
10:54:31 Join ashridah [0] (
11:00:16 Join hardeep [0] (i=hardeeps@SDF.LONESTAR.ORG)
11:12:47 Join guillaumh [0] (
11:16:57 Join Paul_The_Nerd [0] (
11:28:33***Saving seen data "./dancer.seen"
11:30:24 Quit guillaumh (Remote closed the connection)
11:48:58 Quit Slasheri (
11:49:41NJoinSlasheri [0] (
11:57:30 Quit hardeep ("My damn controlling terminal disappeared!")
12:03:55Bger:P my favourite quit message :)
12:06:29XavierGrmy favourite is the one that says: Time to say mooo!!
12:06:29XavierGrMooo! :)
12:06:38XavierGrWhich one has it? Maybe Bagder?
12:06:39 Join Febs [0] (
12:06:43Bgeryes, B4gder
12:07:11XavierGrExcelleny choice B4gder!
12:18:02CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
12:18:02*B4gder bows
13:12:47preglownow, all rockbox needs is sid support, and we can start bundling those files!
13:19:28Bgeryep :))
13:28:35***Saving seen data "./dancer.seen"
13:40:42 Quit Bger ("brb, reboot")
13:46:11 Quit Paul_The_Nerd (Read error: 110 (Connection timed out))
13:52:54dwihnoZagor: I heard about you selling the bostadsbytare on the news. Way
13:53:14 Join Bger [0] (n=Bager@
13:53:41Zagordwihno: yeah. so now I'm a working man again. :-)
13:56:22preglowtough luck, i figured you'd retire and live the rest of your life on the money from the sale ;)
13:57:37dwihnoZagor: The question is, if it's a good or bad thing :)
13:57:39 Quit Febs (Read error: 110 (Connection timed out))
14:00:14Zagorit's a little of both, actually.
14:02:05dwihnoBad thing: limited freedom
14:02:15dwihnoGood thing: retirement fund fixed ;)
14:02:51 Join muesli- [0] (
14:02:53Zagor:-) that, and having co-workers again. working from home got a bit boring after a while, actually.
14:04:05Bgerhehe muesli-
14:04:22muesli-hi Bger ;)
14:05:02dwihnoZagor: yeah, co-workers is the best thing about working :) (well, second to salary day then)
14:05:36 Join actionshrimp [0] (
14:05:50muesli-actionshrimp *g+ nice nick :D
14:07:03muesli-guess forest inspired you ;)
14:07:34actionshrimpforrest gump?
14:07:36actionshrimpor what
14:07:38actionshrimpi dont get it
14:07:44actionshrimpno, unrelated :p
14:07:54muesli-get a shrimp cutter :D
14:08:25actionshrimpget a MUSELI.....spoon!
14:08:55muesli-yeah :D
14:09:20 Join Moos [0] (
14:10:33muesli-or we could have some moos ;)
14:11:00actionshrimpmoos lee?
14:11:05Moos:D Hello guys
14:11:06actionshrimpthe famous explorer
14:27:41muesli-TiMiD sick ;)
14:31:28muesli-TiMiD should rbx have an option acting as a last words recorder? ;)
14:34:51TiMiDI hope no one with rockbox gets involved into a crash :(
14:34:58TiMiD(at least :D)
14:36:42TiMiDsome words can sounds really funny taken away from their context, butin fact they are not ~_~
14:39:12Bgermuesli- it's very probable that it won't be usable after the crash ...
14:41:14muesli-this box will make useable as a black box (at least for h3xx's)
14:42:09Bgermuesli- u forget about the hdd in nearly all players
14:42:15Bgerthat rb supports
14:42:56TiMiDlook at the "truly bizarre" ^^
14:43:13muesli-B4gder thats why it must be directly written into flash rom
14:46:10TiMiDthey aree crazy :(
14:48:34muesli-Rhinoceros Beetles
14:48:35muesli-Rhinoceros beetles are popular pets to some children in Japan. Beetles can be purchased in this road side vending machine outside Morioka. The male beetles are 300 yen while the female ones are only 100 yen. The two curious kids here are part of the Shillingsberg clan.
14:51:16muesli-i've recently read an article about japanese sex life..very interesting..because they dont have any (at least not when you're marriaged...)
14:51:35TiMiDI don't want to live near this mashine (imagine all the insects escapes XD)
14:52:12muesli-married btw
14:52:23TiMiDyes ... sad for them :)
14:52:33Bgerlol flowers .. rice ... :)
14:52:37muesli-i would rather call it bizzare
14:52:54Bgerinsurances ...
14:53:50muesli-they have whorehouses with puppets...
14:54:16TiMiDthey don't need their wifes, they have "Porn Mashine" ^^
14:54:51Bgerand a condome machine between the porn and drink ones :)
14:56:35TiMiDha so they only use their wifes when they are totally waisted ?
14:59:05muesli-they've got once in month sex..and even this is more a duty than fun...
14:59:38muesli-guys go to whorehouses instead..
15:06:44Bgerwow the panties has been offered since 1993 ... (at least )
15:07:45muesli-i know a guy who sold such stuff over the internet
15:09:05muesli-he used raw liver to pretend as a used panty
15:09:15muesli-i am not kidding!
15:09:41muesli-and earned heaps of money with this stuff
15:18:54muesli-Bger are you a dev?
15:25:18 Quit ashridah ("Leaving")
15:27:15Bgerheh ... i can't tell i am myself
15:28:36***Saving seen data "./dancer.seen"
15:30:30muesli-you are Bger ;)
15:37:29 Quit Moos (Read error: 104 (Connection reset by peer))
15:44:43 Join Moos [0] (
15:54:43 Quit Moos (Read error: 104 (Connection reset by peer))
15:55:31muesli-TiMiD how's your german?
16:01:15muesli-After underwear is now girl-spit of Japan of newest Sex articles Tokyo (dpa) - whether the sales of used Schluepfer or the Posieren in school uniform - Japanese young girls on the search for the fast money for expensive luxury article find in of Tokyo relevant amusing quarters for all payment-willing customers. Now they constituted a new lucrative Sex article: Saliva. Up to converted 90 euro are ready some male customers, for a small bottle schoolgir
16:01:21 Quit lostlogic (Read error: 110 (Connection timed out))
16:01:50Zagormuesli-: i'm not sure that's on topic for this channel... :-)
16:01:52 Join lostlogic [0] (
16:02:01muesli-Zagor ;)
16:02:21muesli-we are working for a japanese panties plugin ;)
16:03:21 Join Moos [0] (
16:09:59Zagoropinion poll: what do you guys think about web sites that change the mouse pointer icon? I'd like to somehow indicate to users which links open a new window and which don't. one idea is to make the mouse pointer a crosshair for new-window links.
16:10:20crwlimho new-window links shouldn't exist at all, ever
16:10:33B4gderZagor: I prefer a little icon on the site of the link indicating something like that
16:10:38BgerZagor it depends ...
16:10:45preglowi agree with bagder
16:11:03muesli-Zagor do you do that with js?
16:11:04Bgerbut i'm sure that not everyone will be pleasured of changing his/her cursor
16:11:12Zagorcrwl: i used to think so too, but it's actually useful. in my case, it's for opening a map side-by-side with the page you're looking at (which describes what is on the map)
16:11:23Zagormuesli-: no, css
16:11:46muesli-mk..would be nice ;)
16:12:15ZagorB4gder: any idea what the icon would look like?
16:12:26B4gdernope :-)
16:12:33B4gderI'm icon illeterate
16:12:47B4gderand I can't type nor spell
16:13:28muesli-what about a butt?
16:13:49Zagornot very friendly :-)
16:14:12muesli-depends on your site ;)
16:16:33 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
16:21:07LinusNZagor: a tiny window
16:23:18Zagorthe icon idea has a problem. i already have small icons next to the links, which describe some aspects of what is behind the link...
16:23:29Zagorsuch as "photos available" etc
16:24:56Zagornot to mention, of course, the difficulty of actually drawing a intutive icons at about 10x10 pixels that conveys "this link will open in a new window"
16:27:20Zagori guess what I'm asking is: would changing the cursor be better or worse than doing nothing?
16:30:03B4gder"Major security fixes" release of openssl
16:30:22B4gderaffecting openssl-based eservers
16:33:46 Quit thegeek (Read error: 104 (Connection reset by peer))
16:33:58 Join thegeek [0] (
16:47:14 Quit B4gder ("time to say moo")
16:51:48Moos...s :-)
17:07:16Bgerdoes your nick have any connection with Bagder's quit message ?
17:07:50MoosI doubt
17:17:40 Join solexx_ [0] (
17:17:51 Quit Zagor ("Client exiting")
17:23:54 Quit muesli- (Read error: 110 (Connection timed out))
17:28:05preglowamiconn: i think the grayscale lib seems really sluggish at 45mhz
17:28:39***Saving seen data "./dancer.seen"
17:29:34 Quit solexx (Read error: 110 (Connection timed out))
18:17:15 Join muesli- [0] (
18:22:51 Join _FireFly_ [0] (
18:25:56 Part LinusN
18:26:06preglowamiconn: ghah
18:26:08 Join LinusN [0] (
18:26:12 Part LinusN
18:26:17preglowsorry about that, i just managed to commit something with no commit message
18:26:21preglowany way to fix that?
19:18:58 Join Paul_The_Nerd [0] (
19:24:00 Quit muesli- (Read error: 110 (Connection timed out))
19:27:52 Join DrMoos [0] (
19:28:40***Saving seen data "./dancer.seen"
19:28:50 Quit Moos (Read error: 104 (Connection reset by peer))
19:31:55 Quit BirdFish ("( :: NoNameScript 3.81 :: )")
19:37:20 Join djjsin [0] (n=djjsin@
19:37:35djjsindoes anyone have a good helvetica font for rockbox?
19:43:41 Part djjsin
19:47:19 Join muesli- [0] (
19:47:48 Join elinenbe [0] (
19:50:26amiconnpreglow: cvs admin -m <revision>:<message> <file>
19:51:25 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
19:52:08amiconnThe change will be immediately visible in viewcvs, but the front page will catch the change at the next real commit/rebuild
19:52:32preglowok, great
19:55:15preglowseems to have worked ok
19:55:21preglowthink i'll start on the codeclib change
19:55:52preglowand if i get that working, i'll commit the musepack stuff that uses your code
19:58:07amiconnI think I should try to make the two specialised functions for musepack for a bit more speedup
19:58:34amiconnI'd need test files for that. Where can I find the encoder?
19:59:56_FireFly_amiconn: for windows or linux??
20:00:24amiconnwindows (preferably command line)
20:00:53 Quit muesli- (Read error: 104 (Connection reset by peer))
20:01:03 Join muesli- [0] (
20:01:31_FireFly_first entry of a search result :)
20:01:31 Quit elinenbe (" Want to be different? HydraIRC -> <-")
20:03:27preglowamiconn: rasher's got test files
20:04:25amiconnDitch the capital flaw? ;)
20:06:31preglowamiconn: yes, i hate capitals in c source
20:15:26*preglow wants metadata handling in the codec plugins :/
20:16:00 Join djjsin [0] (n=djjsin@
20:16:27djjsini cant seem to find the convbdf executable for converting fonts for rockbox, all I can find is the source code
20:16:32djjsincan someone help me out?
20:17:08Bgerdjjsin windows ?
20:21:00preglowdjjsin: looks like rasher has some helvetica fonts
20:22:18 Quit linuxstb (Read error: 110 (Connection timed out))
20:22:49djjsinwhos that, I dont see him in the room?
20:22:55 Join linuxstb [0] (
20:23:02djjsinall i really need now is a windows executable for the conbdf
20:23:18_FireFly_djjsin: have you the sources ??
20:23:51preglowdjjsin:"> in both the 75 dpi and 100 dpi subfolders you'll find plenty of helvetica variations
20:24:00djjsinya i have the sources
20:24:01djjsinkick ass
20:24:03djjsinthats all i needed
20:31:16 Join webguest07 [0] (
20:32:51 Part djjsin
20:35:28Bgernite, guys
20:41:39 Join dpassen1 [0] (
20:59:16 Join webguest63 [0] (
20:59:28amiconnpreglow: Did you have a look at the newer versions of libmpcdec?
21:03:06 Join DangerousDan [0] (
21:09:50 Quit webguest63 ("CGI:IRC")
21:12:37 Join RotAtoR [0] (
21:19:56 Quit ender` (Read error: 104 (Connection reset by peer))
21:21:09 Quit webguest07 ("CGI:IRC")
21:23:31 Join ender` [0] (i=ychat@
21:28:45***Saving seen data "./dancer.seen"
21:31:18 Join Lear [0] (
21:35:01 Nick DrMoos is now known as Moos (
21:43:24preglowamiconn: yeah, nothing new there
21:43:41preglowamiconn: mostly bugfixes either me or t0mas have pushed through
21:48:05amiconnI think the two special functions will have even more advantages
21:48:40amiconn64 bits are only needed for the intermediate results, so the arguments and return value can be 32 bit
21:48:47 Quit paugh ("Leaving")
21:48:53amiconn-> less parameter passing
21:52:02preglowyes, that'd be neat
21:52:43preglowbut like i said yesterday, don't spend too much time on this unless you find it interesting, it should only be a temporary solution anyway ;)
21:52:51preglowwhere temporary might easily be more than six months, but hey...
21:54:34Learbtw, do you need to do anything special to get audio playback in linux?
21:54:48Learusing the simulator...
21:54:54preglowno idea
21:55:26Leargave it a quick try today, but it wouldn't even start playback before crashing... :/
21:56:01_FireFly_i can only play wav-files in the simulator in linux
22:03:30LearCould be codec size problem, but I tried to compensate for that...
22:04:14 Join vmx [0] (
22:04:42amiconnA propos codec size - I noticed that gcc includes floating point exception handling for mpc because of it using fp
22:05:02amiconnI wonder whether there is a way to tell gcc not to do that
22:07:04Learlinker script?
22:26:31LearOr maybe -fno-trapping-math?
22:26:40LearSorry, -fast-math. :)
22:28:44amiconnWee, 32->64 bit sign extension is really simple on coldfire :)
22:30:21 Quit _FireFly_ ("Leaving")
22:32:54amiconn..and the fact that the high part of the shift result can be dropped makes things faster too
22:48:17amiconnSomething is wrong with the playlist handling and .mpc
22:48:35amiconn.mpc files are skipped unless selected directly
22:48:51 Quit Lear (Read error: 110 (Connection timed out))
22:49:40amiconnI could play iriver-xtreme.mpc without skipping :)
22:49:59 Join Bagder [0] (
22:50:27dpassen1xtreme is -q6?
22:51:30preglowamiconn: weee
22:51:36dpassen1congrats, either way
22:51:41preglowamiconn: btw, i know of the playlist issue, haven't had time to look into it
22:53:41amiconnpreglow: Wanna try?">
22:55:06amiconnThe shifting gets incredibly simple with the implicit 64->32bit drop
23:00:32 Join matsl [0] (
23:00:47 Quit muesli- (Read error: 104 (Connection reset by peer))
23:22:42 Join JoeBorn [0] (
23:23:04 Join djjsin [0] (n=djjsin@
23:23:22djjsinis it possible to do a shuffle on all files on the harddrive, without creating a playlist
23:23:33djjsinso for example, you can choose a file you want to play, then after thathave it go random?
23:23:46djjsinlike you can with the original iriver firmware
23:24:30 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
23:24:32djjsincan you record yet with the iriver rockbox port
23:25:03amiconnpreglow: A bit faster again, replacing the two mulu.l (which only ever multiply by 0 or -1 now) with and.l + neg.l
23:26:19amiconn-> same old link now
23:26:32 Join ashridah [0] (
23:28:48***Saving seen data "./dancer.seen"
23:29:28amiconnBtw, mpc_multiply is unused, only mpc_multiply_ex is used by the decoder
23:33:31amiconnHmm, I can save a branch with no penalty for the zero shift case
23:34:13amiconnignore me :/
23:38:42preglowcan do
23:38:57 Join |Lupin| [0] (
23:39:15preglowwas watching a movie, i can try it now
23:39:23|Lupin|Hello, all. I was wondering how to record with the Iriver and Rockbox, please ?
23:39:57preglow|Lupin|: it's not completely supported yet, you need to use the debug menu
23:40:08 Part djjsin
23:40:21preglowand unless i remember incorrectly, you're blind, so that's not easily accessible
23:40:56preglowamiconn: looks really good
23:41:10|Lupin|preglow: Indeed...
23:41:26|Lupin|preglow: Which kind of skills is required to make recording work ?
23:41:26preglow|Lupin|: recording for iriver is going to be enabled properly in the not too distant future
23:41:29vmxmost probable that question gets asked quite often: any news about the iaudio port?
23:41:38preglow|Lupin|: so i really suggest you wait, unless you've got some urgent need
23:41:47Bagdervmx: nope, nothing
23:41:55preglowvmx: no, the guy who does the port just vanished, but says he's coming back soon
23:42:13vmxthanks for the info :)
23:42:43amiconnpreglow: Do you think this should be committed? Or better wait for you committing it together with your changes?
23:42:47|Lupin|Well, I can also try with the Iriver firmware, if someone explains me how to do, and if it is &ot too difficult.
23:43:21|Lupin|And btw: If there is something I can do I'll be glad to try doing it.
23:43:23preglowamiconn: well, i've got no pending changes, these supercede what i did with mul64
23:43:51preglow|Lupin|: can't help you there, i'm afraid, i haven't done any recording yet
23:44:55preglowamiconn: and yes, i definitely think this should be commited
23:45:17preglowamiconn: having realtime support until i can arse myself to do things properly (if ever) is most definitely a good thing
23:48:45preglowamiconn: btw, are you completely certain not using the emac is faster? i can't possibly imagine how it can be slower
23:49:00preglowamiconn: that is, using the emac to calculate the upper 32 bits, then an ordinary mul for the lower bits
23:49:37preglowonly problem is of course that the upper bit must be presumed useless, since it vanishes...
23:51:19amiconnThe problem with using the emac is that the result isn't readily available in a register, but needs to be moved from the accumulator
23:52:06amiconnPlus, you need more than 2 multiplications for that, and using mulu.l isn't possible for the lower part
23:52:25amiconn...since you need all result bits, with is only possible with mulu.w
23:52:47amiconnIf the emac had a 64bit accumulator...
23:58:29preglowthis exact technique for a 64 bit mul is used in asm_mcf5249 in MULT31_SHIFT15
23:58:44preglowemac for top bits, mulu.l for bottom bits

Previous day | Next day