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 2005-10-27

00:00:23TiMiDand what was it about ? (protecting is quite big :p )
00:00:32XavierGrI can right Greek if the other has a Greek font.
00:00:38 Join muesli-- [0] (
00:00:56TiMiDso if you talkgreek on a grek forum it's ok
00:01:00muesli--it said something that words that can be translated must be translated
00:01:05TiMiDif the othercan understandgreek
00:01:19TiMiDabout that :)
00:01:21muesli--ie walkman (whatever it is in french)
00:01:22*preglow sets up the preglow whisky fund
00:01:51TiMiDwalkman -> balladeur (ugly word, not used)
00:02:13TiMiDthese words are produced by the academy but no one uses them
00:02:21TiMiDexcepted official documents
00:02:37XavierGrTiMiD: Normal greek language is preffered because it is more readable (especially in long texts)
00:03:12 Quit actionshrimp ("a bird in the bush is worth two in your house")
00:03:13muesli--balladeur..i remember that word in this context :D
00:03:27TiMiDwell balladeur is quite used
00:03:36XavierGrI thought that this is a card name in card games
00:03:37TiMiDyou can use both of them
00:03:45muesli--sounds strange ;)
00:03:55XavierGre.x a balladeur can be any card you want on Poker
00:04:06XavierGrso that you can make various combinations
00:04:09TiMiDin french ?
00:04:15Moosballadeur: hum french
00:04:22TiMiDyou mean jocker
00:04:26XavierGrballadeur is often designated as the Joker.
00:04:44XavierGryes but not only Joker
00:04:52TiMiDI never used balaldeur to play cards
00:04:58TiMiDit sounds odd
00:05:07muesli--i rather play with my balls ;)))
00:05:07XavierGrthough the balladeur rule is often avoided to make the game more interesting.
00:05:09tvelocityXavierGr, you use mirc?
00:05:38XavierGrno, right now I use Trillian, why?
00:06:16tvelocitydoes trillian support UTF-8?
00:06:34TiMiDuse a true irc client
00:06:47tvelocityirssi uber alles
00:06:56TiMiDyep irssi is the best
00:07:04XavierGrwell the text that TiMiD wrote wasn't readable I had to copy it to a notepad 2 and then convert it.
00:07:16TiMiD(hasn't been updated for ages)
00:07:42XavierGrWell I use Trillian because I have both MSN (which I dislike) and IRC together on 1 app.
00:07:43tvelocitycαn Υου reαδ thίς?
00:08:05tvelocityengreek :P
00:08:20TiMiDit's like that
00:08:28XavierGrtvelocity: No I can see only garbage on this.
00:08:42markunMoos is using Xchat for windows. Also does unicode.
00:08:44tvelocitythat's because trillian doesnt support unicode
00:08:51tvelocityor at least it's disabled by default
00:09:05TiMiDuse Xchat
00:09:07XavierGrbad Trillian
00:09:14TiMiDit runs under win
00:09:19tvelocityyou can use gaim too
00:09:21XavierGrWell I dont care I didnt paid for it so... :p
00:09:26 Join muesli|trillian [0] (
00:09:45tvelocitygaim support every protocol under the sun and is free as in beer and as in speech
00:09:47TiMiDmuesli|trillian: にはは
00:09:54TiMiDcan you read ?
00:10:05muesli|trilliancant read it in mirc and trillian
00:10:14muesli|trillianneiterh nor
00:10:17TiMiDI don't think gaim provides advanceds irc support
00:10:26TiMiDcαn Υου reαδ thίς
00:10:33tvelocityi doubt trillian provides that either :P
00:10:40TiMiD(the sentence of tvelocity)
00:11:06TiMiDxan others see what I wrote ?
00:11:34TiMiDok :p
00:11:41TiMiDwhat's your local encoding ?
00:12:14TiMiDoh :)
00:12:14tvelocityυηίcοδε rυΙεΖ :P
00:12:23TiMiDI love it ;)
00:12:39tvelocityit sucks tho that mirc doesnt support it
00:12:39markunSoon you can have all the unicode you want on rockbox :)
00:12:43muesli|trillianbut this could be my possibility to join #rockbox wheni am at the university..those blokes blocked the port apparently
00:12:49TiMiDfrench accents and japanese kanjis in the same file is very convenient :p
00:12:57markuntvelocity: xchat also is a nice client
00:13:03tvelocityxchat pwnz
00:13:04 Part muesli|trillian
00:13:06TiMiDmuesli|trillian: use irssi
00:13:16TiMiDif you have an access to ssh
00:13:18muesli--under linux?
00:13:21pregloweverything about this new flac decoder is better
00:13:26tvelocitybut most people use mirc, so i'm stuck with iso-8859-7 in greek chat networks :(
00:13:26preglowmy asm optimisation even ends up faster
00:13:51TiMiDtvelocity: you have a script that can convert encodings
00:14:03TiMiDyou specify it the inout encoding
00:14:20TiMiDand the per channel encoding and it handles the rest itself
00:14:52TiMiDI think
00:14:58tvelocitywhere can i find it?
00:15:06XavierGrpreglow someone said to move codecs to iram? (iirc) What would be the benefits from that move, and what is still missing from the multi-codec support
00:15:11TiMiDI use it since on french channels it's iso-8859-15
00:15:17TiMiDor iso-8859-1
00:16:15TiMiDI send you
00:16:28muesli--oki m8s...gotta get up in 6h...cya
00:16:42TiMiDgood night :)
00:18:25 Quit muesli- (Read error: 113 (No route to host))
00:18:32 Quit Moos ("Glory to Rockbox")
00:19:00 Join Moos [0] (
00:19:44TiMiDre Moos :)
00:20:07MoosTiMiD: Re
00:20:12linuxstb_preglow: Good news about your optimisation. Does it work with the "wide" version of the loop?
00:21:17preglowlinuxstb_: hell no
00:21:40preglowXavierGr: we already use iram
00:22:06preglowlinuxstb_: the wide version will require some more work
00:23:07preglowwe might do it the way we do for musepack, but then we have to count on never stumbling upon flac files that need all the 64 bits of the wsum variable
00:23:19preglowif we can't do it that way, it'll probably always be pretty slow
00:23:54linuxstb_It's not that important anyway. Just a nice thing to have. I don' think that loop will be used for 16-bit files.
00:24:29preglowi most certainly shouldn't
00:24:54 Quit Moos ("Glory to Rockbox")
00:25:05linuxstb_XavierGr: I think stability is the only thing really missing from the multi-codec support. Plus recording if you count that as part of the same thing.
00:25:05 Join Moos [0] (
00:26:07 Nick linuxstb_ is now known as linuxstb (
00:26:29preglowstability is pretty much missing over the entire iriver part of rockbox, heh
00:26:41linuxstbTime for a feature freeze? :)
00:27:09preglowlet's wait until slasheri is free to do bugfixing :P
00:27:14preglowi'd hate to fix bugs in playback.c
00:27:24 Join DrMoos [0] (
00:27:57linuxstbI think that's the problem - only one person seems to know the playback code. I know I should make more of an effort to understand how it works and help improve it.
00:28:35preglowi just can't be bothered yet, there's still much work i'd like to do for codecs and dsp
00:29:26preglowanyone know what the official comment code is in 68k gas?
00:29:40preglowi've always used /* */, but david bryant seems to use a |
00:29:45preglowand that's far preferable
00:32:11preglowamiconn: excellent
00:32:49 Quit DrMoos ("Glory to Rockbox")
00:33:04TiMiD <<kickass haircut ^^
00:33:59amiconnpreglow: gas allows both C style comment (for all architectures) and 'classic' comments for that platform
00:34:15amiconnI prefer C style because it's consistent across architectures
00:34:42amiconnClasic comments are | for 68k, ! for SH, @ for arm ...
00:34:53preglowi'm not exactly awestruck by our libmads performance on these files...
00:35:08 Quit muesli-- (Read error: 110 (Connection timed out))
00:37:43preglowamiconn: i've had a couple of "nice" bugs resulting from me messing up the end delimiter in classic comments, so i tend to go for the other types when writing asm
00:40:55 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
00:41:13XavierGrahhh I am off to bed.
00:41:17XavierGrGood night all!
00:42:13MoosHave a good night Xavier
00:42:43amiconnpreglow: In what way did you mess up the end delimiters? Afaik, classic comments are line comment (equivalent to C++ style // )
00:43:45amiconnAnyway, if you use an editor with syntax highlighting, messed up comments should never happen...
00:45:00preglowamiconn: i do use vim with syntax highlighting, and still it happened
00:45:13preglowamiconn: i meant c style comments, not classic comments
01:02:52preglownow, let's see if this works
01:13:47preglowof course, there had be a strange error
01:15:47***Saving seen data "./dancer.seen"
01:19:14preglow/home/thomj/rockbox-devel/apps/codecs/libffmpegFLAC/coldfire.S:228: undefined reference to `.default'
01:19:25preglowit's in plain sight, fifty lines above it
01:19:30preglowother branches to .default work just fine
01:21:01 Quit XavierGr (Read error: 110 (Connection timed out))
01:22:33preglowit was a comment delimiter error _AGAIN_
01:22:38preglowi'm bloody switching
01:24:23Mooshave a good night all
01:24:41 Quit Moos ("Glory to Rockbox")
01:29:21preglowit plays fine with opts a good twenty seconds
01:29:23preglowthen hangs...
01:32:40preglowkind of hard to say what the performance improvement is as well, since it never boosts anyway
01:34:42 Join linuxstb_ [0] (
01:36:51linuxstb_preglow: That's strange. Maybe it's a particular pred order that causes the problem.
01:37:10preglowi thought flac always used pred order 8
01:37:20preglowunless you specified another maximum
01:37:28linuxstb_No, I think it varies frame by frame
01:37:29preglowlets hope it's not so, i just found a bug in pred order 5
01:38:20preglowthat was it
01:38:43preglowbut like i said
01:38:52 Quit cYmen__ ("zZz")
01:38:53preglowimpossible to decide whether it helps much or little
01:39:06preglowit never boosts no matter what i do anyway
01:40:18preglowignore that, i was running the straight c version, doh!
01:40:28linuxstb_All my files boost around 10%.
01:40:46linuxstb_Some people even reported 40%.
01:41:13preglowk, if it works now, i'll send you a flac.codec you can try with
01:41:21pregloweven the q8 files don't boost here
01:42:39preglowhmm, doesn't crash anymore at least
01:43:37preglowbeware of sudden burst of noise, i haven't tried with many files yet
01:44:55 Quit linuxstb_ ("CGI:IRC")
01:45:12preglowwhat the hell, i've got a file here that just quits playing after twenty seconds or so
01:45:28linuxstbWith or without your new function?
01:45:36preglowi'll try without now
01:46:29linuxstbI'm just double-checking the boost ratio on one of my files. Then I'll test with your new version.
01:46:48preglowwith, it seems
01:47:07preglowit's still got bugs, i hear
01:52:15linuxstbIt's working fine for me. Boost ratio is now down to zero
01:52:39linuxstb(Apart from buffer filling and a boost at track transitions)
01:53:05linuxstbSo it's definitely a significant improvement.
01:53:06preglowi've found a couple of bugs in the order > 8 routine
01:53:58linuxstbThe album I'm playing sounds file - so I expect it doesn't have any > 8 frames.
01:55:27preglowi'm pretty sure the unrolled loops are fine, yes
01:55:38preglowouch, ouch, ouch
01:55:42preglowit's got major bugs
01:55:51preglowbut it doesn't quit anymore, so that's where the bu was
01:57:02preglowi'll try to fix it unless i cough myself to death first
01:59:36preglowdoes flac use the fixed predictors at all unless you force it?
02:00:13preglowthere, now i think i've got the last bugs ironed out
02:02:01 Join pinkutank [0] (n=ddd@
02:02:34linuxstbYes, I think FLAC can (and does) use any predictors. You can use "flac -a" to analyse a file if you want to find out what's in your test file.
02:02:57 Join bagawk [0] (n=lee@unaffiliated/bagawk)
02:03:11linuxstbSomeone on the forum just gave me a FLAC file which completely hangs Rockbox.
02:03:43linuxstblibFLAC decodes it fine, and there doesn't seem anything obviously strange with it.
02:04:12linuxstbstandard 4608
02:04:18linuxstbmplayer plays it fine as well.
02:04:31linuxstbI'll try my test program.
02:05:06pinkutankpreglow, mind if I ask about the status on the parametric eq
02:05:27preglowpinkutank: no, not at all, i'm about 50% done
02:05:57 Join webguest88 [0] (
02:05:59pinkutankglad to hear it, I would like to help, but I hav eno programmnig skills
02:06:11linuxstbMy test decoder seems to get into an infinite loop. So looks like it's something I've done wrong. Time to debug.
02:06:33pinkutankonly thing I can think I can do is beta testing and creating pngs for bars and knobs :)
02:06:45preglowpinkutank: don't expect a fancy gui at this point
02:06:52preglowpinkutank: at least i sure as hell wont bother making one
02:06:59pinkutankdont care about gui
02:07:04preglowbut the eq itself will be fully parametric
02:07:10pinkutankI just need an eq working thats all
02:07:16preglowadjustable Q, amplification and center frequencies
02:07:35pinkutankpeople will want ui tho, I'll try and help with that if someone decides to do it
02:07:37webguest88people, does anyone here knows how did the "Recent CVS activity" script was set?
02:07:51pinkutankwhat is the approx eta?
02:07:56webguest88the big table on the index page of the project ...
02:08:04preglowpinkutank: a bit hard to say, perhaps a week?
02:08:08preglowdepends on a lot of things
02:08:15pinkutankwoot, didnt expect that
02:08:42pinkutankI just heard you said you would be able to do it from someone on the forum thatas all
02:08:49preglowfirst of all i need to get the filtering to work properly, then i need to modify all the codecs to use 28 bit audio, then i need to integrate the eq into rockbox and make a settings menu
02:09:16pinkutankwhy 28?
02:09:31preglowso we've got a few spare bits left
02:09:41pinkutank:) ok
02:09:48webguest88I've spend a lot of time searching for google and haven't found anything that seem like a Recent CVS activity script .... :(
02:10:06pinkutankcustom script, I think
02:10:12 Quit AliasCoffee ("Leaving")
02:10:24webguest88... anyone knows where can i get it ?
02:11:14preglowlinuxstb: same url now for what i consider the final routine at this point
02:11:23pinkutankyou can't and even if you did, itd be hell hard to integrate
02:11:28preglowno major changes, just a few bugfixes to the order > 8 routine
02:11:29pinkutankbut ask the site developers
02:12:08webguest88okay ... i'm emailing the maillist them ....
02:12:17pinkutankbtw, does anyone know if therell be an improvisation to dac engine to further sound output quality?
02:12:20webguest88thanks! :)
02:12:23 Part webguest88
02:13:18preglowpinkutank: what kind of improvement would that be?
02:13:37preglowpinkutank: pretty much only thing we can do now is enable 20 bit audio instead of sixteen
02:13:53preglowand i have no idea how practical that would be
02:14:40pinkutankI'm not sure if there is a gap between the original firmware and rb
02:14:51pinkutankbecause I Dont have accurate enough testing phones
02:15:24pinkutankreferring to debates, some say rb has " worse " sq, but thats subjective, so I decided to ask you guys
02:15:28preglowi don't have accurate enough ears
02:15:50preglowpinkutank: i've read about these claims, and i've decided not to care about them until i see some abx tests done
02:16:00pinkutankthats reasonable
02:16:17preglowthere are a lot of factors that could play in here
02:16:19pinkutankI might to the abx, cousing has a proper setup
02:16:37preglowbut i'd like to see a proper test done, comparing rockbox and iriver on mp3s, oggs and wavs
02:16:48pinkutankoggs, rb is better
02:16:55pinkutankat least more accurate
02:17:00pinkutanksomeone had tests up
02:17:09preglowvorbis for rockbox should be very accurate
02:17:10pinkutankmp3s, I have no idea
02:17:13preglowbut the same applies to mp3s
02:17:24preglowi believe iriver too uses sixteen bit output
02:17:27preglowbut i'm not sure
02:17:34preglowand some people can possibly hear the difference there
02:17:54pinkutankI2ll try it tomorrow
02:18:00pinkutankbut I forget
02:18:05pinkutankthats my problem
02:18:22preglowdoesn't matter much for me, it's you who want this verified, heh
02:18:22pinkutankas soon as I turn it off to turn the other firmware on, I forget :)
02:18:35pinkutankwell ok, I'll try abx
02:18:38pinkutankthanks for the time
02:18:45preglowrockbox' sound quality is more than good enough for me
02:18:53preglowno problem
02:23:00pinkutankbtw, how many bands on the eq?
02:23:29 Join ]RowaN[ [0] (
02:23:37]RowaN[bah 100 views and no comments
02:23:43linuxstbpreglow: Your latest flac.codec seems fine - no boosting any more. :)
02:25:56preglowpinkutank: as many as you want, i was thinking of five
02:26:11preglowpinkutank: but some day someone can perhaps make an eq where you can decide yourself
02:26:59linuxstbI'm just listening to a 48KHz FLAC, and it's boosting a little. I guess that's the DSP working as well.
02:28:11preglowyes, the resampler does require a bit
02:28:31preglowbut still
02:28:34preglowa pat on the back
02:28:42preglowthis is leaps and bounds better than libflac in all ways
02:28:50preglowit's smaller, prettier, faster
02:29:29linuxstbYep. I'm surprised no-one found it earlier.
02:29:49preglowbut ok, should i commit the optimisation, then?
02:30:06preglowperhaps wait with enabling it until it's seen further testing?
02:30:13 Join ashridah [0] (
02:30:17preglowthat is, the main flac codec
02:30:51linuxstbThat's probably a good idea.
02:30:57TiMiDhehe Currently sending my resume to iriver japan ^^
02:31:12preglowi'll commit the code, though
02:31:34preglowbut i'll enable it first when you've added seeking and no more bugs are being reported
02:32:09 Quit bagawk (Remote closed the connection)
02:43:29pinkutankanyone knowledgeable of art education in japan
02:45:26TiMiDuh ?
02:46:02pinkutankermm since you mentioned japan
02:46:14TiMiDI'm in computer engineering
02:46:19pinkutankI though just maybe someone here might know a thing or two
02:46:24TiMiDnot in art :)
02:46:35TiMiDwhat do you want to do ?
02:46:38pinkutankwell, can i ask some baic college questions
02:46:50pinkutankphotography and illstration I think
02:47:06pinkutankdo you have to study in japanes or are the major schools in english
02:47:29TiMiDI don't know (I'm french and I live in france)
02:47:36preglowdepends, i think
02:47:57preglowi know some exchange students, and they said many lectures were in english
02:48:15TiMiDbut some students from my school spent a semester in a japanese university and the teaching was done in japanses
02:48:33TiMiDit depends on the school yes :)
02:48:42pinkutankeheh timid, J'ai pense que tu etudee en Japain (here goes my sucky french, dont beat me)
02:49:15pinkutankI looked at some databases
02:49:20TiMiDt'inquiete pas je vais pas te taper pour ca (mon anglais doit etre encore pire XD)
02:49:21 Quit ]RowaN[ (Read error: 110 (Connection timed out))
02:49:28pinkutankbut there werent enough info
02:49:43TiMiDyou are aiming for an art school in japan ?
02:49:52pinkutankqqc pire?
02:50:01pinkutankI am aiming at an art school
02:50:08pinkutankbut not sure where :)
02:50:50TiMiDanyway do you speak japanese ?
02:51:08pinkutankno, gonna start a course starting form jan or feb
02:51:15TiMiDhmm ok
02:51:26pinkutankall I know is basic stuff currently
02:51:41pinkutanklike this is my friend, hello , how are you and yadda yadda
02:51:42 Join sinizzl [0] (
02:51:56pinkutankand only kanji I memorized is -ka
02:52:00 Part sinizzl ("Leaving")
02:52:12TiMiDif the school dispenses courses in english it won't be a matter
02:52:20TiMiDyou mean kana
02:52:22pinkutankIm not much of a career seeker
02:52:36pinkutankno, - ka as in the "?"
02:52:44TiMiDit's a kana
02:53:06pinkutankso uits of hiragans are kana?
02:53:08TiMiDsyllabic alphabet
02:53:17pinkutankI know hiragana
02:53:23pinkutanki mean
02:53:27pinkutankwhat is hiragana
02:53:33pinkutankbut I didnt know ka was
02:53:36preglowone of the kana alphabets
02:53:44TiMiDkana is the generic term to call both katakana and hiragana
02:53:55pinkutankI thought it was katakana
02:53:57pinkutankbut then
02:54:03pinkutankits too round to be katakana
02:54:12TiMiDka is very similar in hiragana and katakana ...
02:54:27pinkutanklemme look up a table
02:55:43TiMiDany way
02:55:49TiMiDif it can help you
02:56:07TiMiDart school is bijyutsugakkou
02:56:52pinkutankisnt it without the j
02:56:58pinkutanki mean
02:57:22linuxstbThis FLAC file is troubling me: - it causes Rockbox to freeze.
02:57:26TiMiDdepends of the conversion method
02:57:49TiMiDit would be better to write it in kanas :)
02:58:01TiMiD(romajis are onliy a substitute)
02:58:16pinkutankwell, as I said
02:58:19pinkutankprimary aim is
02:58:21pinkutanknot a career
02:58:24preglowlinuxstb: debugging didn't yield anything?
02:58:24TiMiDplenty of art shcools
02:58:33pinkutankliving in a hut somewhere in a mountainc liff
02:58:51TiMiDthen you don't need to go to japan ;)
02:58:55pinkutanknot extremely remote, nor tresspassed by thousands
02:59:05pinkutankI like the weather and the setting
02:59:09linuxstbIt works fine on my PC - in mplayer (which uses ffmpeg), in libFLAC and in my test program using the Rockbox version of the ffmpeg decoder.
02:59:12pinkutanknot talking about tokyo
02:59:13linuxstbI'm about to try the sim.
02:59:21pinkutankI like cities too
02:59:30pinkutankbut I don't want to live in one forever
02:59:56preglowlinuxstb: is test program run on target?
03:00:06TiMiDso why an art school ?
03:00:36pinkutankermm because only thing I can stand is arts, I need to accumulate some wealth before going off
03:00:42linuxstbpreglow: No, it's the "main.c" program in the libffmpegFLAC directory.
03:00:54pinkutankbecause I want books, and obviously the house is not free
03:01:05pinkutankI've been photographing for 8 yuears now
03:01:08TiMiDand why japan ?
03:01:34pinkutankwell why not? I like the culture and the language
03:01:43pinkutankand its very similar to my culture and language
03:02:04pinkutankthe sentence structure is the same as in my language
03:02:07pinkutankwhich is Turksih
03:02:25TiMiDwith particules
03:02:53TiMiDenclytics particules I mean
03:03:32TiMiDthen all you nead is vocabulary :)
03:03:32pinkutankits a little more complicated in Turkish
03:03:41pinkutanklike we instead of -no
03:03:42linuxstbMmm. It plays perfectly in the Sim...
03:03:51pinkutankwe have 8 or so variables
03:03:55TiMiDwhat is more complicated ?
03:04:03TiMiDvariables ?
03:04:08preglowlinuxstb: wonderful...
03:04:12pinkutankhmm example
03:04:13TiMiDwhat's this?
03:04:20preglowlinuxstb: time for some heavy logfing?
03:04:26linuxstbtime for bed.
03:04:29pinkutankthink of -no
03:04:36TiMiDthe japanese no ?
03:04:59pinkutankwe have different nos, depending on the person doing the action
03:04:59preglowforgot about the time again
03:05:04linuxstbHow will logf help if it crashes? Do they survive a reset?
03:05:32preglowdoes it reset?
03:05:39preglowi thought you said rockbox hanged
03:05:46TiMiDjapanese particules are quite messy too ...
03:05:56linuxstbYes, it hangs. So I can't access the logs (I don't have a remote).
03:06:03pinkutanktoru taru like?
03:06:08TiMiDbasic usage is easy but at some point it becomes complicated
03:06:31TiMiDthe distinction between wa and ga for example
03:06:31preglowhow i'd like a bdm...
03:06:56TiMiDor wo and ni with transitionals verbs
03:07:44pinkutankwell, I'll learn I guess, I don't get bored reading it myself from other sources
03:08:13TiMiDI hope you can go there
03:08:19TiMiDit's a little hard
03:08:43pinkutankhard = going there?
03:08:48TiMiD(at least to find an internship in a sector as spread as computer enginering ...)
03:09:09pinkutankI may decide to work in the usa for a couple of years and then move to japan
03:09:16pinkutankwhich may be a smarter move
03:09:33TiMiDI'm thinking about the same thing
03:09:59TiMiDover ther just be sure to take japanese lessons
03:10:02linuxstbOK. Goodnight all.
03:10:09TiMiDgood night linuxstb
03:11:45TiMiD(no I won't bother you with another letter ;) )
03:12:20pinkutankI also started ninjustu
03:12:29pinkutankbut had to stop due to unknow knee injury
03:12:37pinkutankI'll probably oick that up too
03:12:55pinkutankwish japnese education was cheap in usa as it was here
03:13:09pinkutankwe have a women of japan foudnation, dont ask me why
03:13:15pinkutankand they teach japanese
03:13:23pinkutankfor ridiculuosly cheap prices
03:13:44pinkutankand if I take it in us, theyll go on aobut how it is reversely structured
03:13:47TiMiDsponsored by governement ?
03:13:56TiMiD'night preglow
03:14:01pinkutanknot sure timid, might be
03:14:11pinkutankthey train people to be tour guides and etc
03:14:52pinkutankthey do something up to 250 kanjis in writing, they focus on speaking
03:15:50***Saving seen data "./dancer.seen"
03:16:06TiMiDso it's not totally in one way ;)
03:16:19TiMiD250 kanjis is good
03:16:27TiMiDto survive
03:16:36pinkutankbut I cant start right know
03:16:45pinkutankbecause I ahve unv applicaitons and IB stuff at hand
03:16:54pinkutankjanuary to august
03:17:00pinkutankthats 7 months
03:17:15pinkutankIt would help as an initiation
03:17:38pinkutankhow long does it take to be able to read stuff written in kanji
03:17:44pinkutank, say mangas for example :D
03:19:01TiMiDI'm not able to read mangas
03:19:28TiMiDthe problem is not in my opinion to be know the kanji
03:19:40TiMiDit's to know the vocabulary behind
03:20:04TiMiDI personnaly think that to associate a kanji to a word helps you to remember it :)
03:20:32TiMiDI think with 500-600 kanjis you should be starting to understand 80% of what you read
03:21:11pinkutankI would think mangas would be hard because they have non-daily kanji
03:21:18pinkutankbut thats just a guess
03:21:55pinkutankwith my current meorizing abilities , 100 kanji would be a surprise, hope I'll develop some way
03:22:33pinkutankyou said it was hard to differ wa and ga, so I went and studied a little
03:22:41pinkutankbut I dont know japanese yet
03:22:50pinkutankso no its not very useful as of yet :)
03:23:30pinkutankwhat is ni?
03:23:34pinkutankis it to?
03:23:46pinkutankas in,tokyo-ni ikimashita,
03:23:59pinkutankit seems like the wa for places
03:24:09pinkutankbtw ,its 4:30 am here
03:24:23TiMiDit's 3:30 here (I was gone to eat a little)
03:24:30pinkutankgonna wake up for school in 2 hours, writing a tok essay atm
03:24:40TiMiDwhen I said it was difficult, I meant grammatically
03:24:47TiMiDthe japanese pronunciation is very clear
03:29:05pinkutankit is very similar to turkish again
03:29:28pinkutankIf c is read like sea, it is always read like sea
03:33:58pinkutankI have to squeeze 500 more words, turns out this was esay, gotta go drink some water
04:00:39pinkutankwell gnight all
04:00:51pinkutankkudos to rb
04:01:02 Quit pinkutank ()
04:07:12 Join Vladoman [0] (
04:24:48 Quit Vlad0man (Read error: 110 (Connection timed out))
04:56:58 Quit tvelocity ("Leaving")
05:15:51***Saving seen data "./dancer.seen"
05:34:50 Quit ]Maggie[ (Read error: 104 (Connection reset by peer))
05:36:01 Join ]Maggie[ [0] (
06:26:11 Join Seed [0] (
06:30:30 Quit lostlogic (
06:30:30 Quit preglow (
06:30:30 Quit Hadaka (
06:30:30 Quit Slasheri (
06:30:30 Quit markun (
06:30:30 Quit ashridah (
06:30:30 Quit Rick (
06:30:30 Quit Ismo (
06:30:30 Quit mbr (
06:30:30 Quit Nilisco (
06:30:30 Quit Seed (
06:30:30 Quit Vladoman (
06:30:30 Quit pabs (
06:30:30 Quit CoCoLUS (
06:30:30 Quit goa (
06:30:30 Quit TiMiD (
06:30:30 Quit Bagder (
06:30:30 Quit linuxstb (
06:30:30 Quit SeeSchloss (
06:30:30 Quit Zagor (
06:30:30 Quit ryan_j (
06:30:30 Quit solexx (
06:30:30 Quit merbanan (
06:30:30 Quit pill (
06:30:30 Quit yosemite (
06:30:30 Quit HCl (
06:30:30 Quit crash_ (
06:30:30 Quit ]Maggie[ (
06:30:30 Quit dwihno (
06:30:30 Quit amiconn (
06:30:30 Quit wireddd (
06:30:30 Quit ghode|afk (
06:30:30 Quit ze (
06:30:30 Quit phaedrus961 (
06:30:30 Quit Strath (
06:31:08NJoinSeed [0] (
06:31:08NJoin]Maggie[ [0] (
06:31:08NJoinVladoman [0] (
06:31:08NJoinashridah [0] (
06:31:08NJoindwihno [0] (n=dw@
06:31:08NJoinBagder [0] (
06:31:08NJoinlinuxstb [0] (
06:31:08NJoinSeeSchloss [0] (
06:31:08NJoinghode|afk [0] (
06:31:08NJoinZagor [0] (n=bjst@pdpc/supporter/sustaining/Zagor)
06:31:08NJoinryan_j [0] (
06:31:08NJoinmarkun [0] (
06:31:08NJoinamiconn [0] (
06:31:08NJoinsolexx [0] (
06:31:08NJoinlostlogic [0] (
06:31:08NJoinpabs [0] (
06:31:08 Join Rick [0] (i=rick@unaffiliated/Rick)
06:31:08NJoinCoCoLUS [0] (
06:31:08NJoinze [0] (
06:31:08NJoingoa [0] (
06:31:08NJoinSlasheri [0] (
06:31:08NJoinwireddd [0] (
06:31:08NJoinIsmo [0] (
06:31:08NJoinphaedrus961 [0] (
06:31:08NJoinmerbanan [0] (
06:31:08NJoinStrath [0] (
06:31:08NJoinmbr [0] (
06:31:08NJoinTiMiD [0] (n=TiMiD[
06:31:08NJoinHadaka [0] (
06:31:08NJoinpreglow [0] (
06:31:08NJoinNilisco [0] (
06:31:08NJoincrash_ [0] (
06:31:08NJoinyosemite [0] (
06:31:08NJoinHCl [0] (
06:31:08NJoinpill [0] (
06:31:09***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
06:31:39CtcpVersion from freenode-connect!freenode@freenode/bot/connect
06:31:40***Server message 477: 'logbot #RockBox :[freenode-info] if you need to send private messages, please register:'
06:44:41 Quit Seedy (Connection timed out)
06:45:18 Quit lostlogic (
06:45:18 Quit preglow (
06:45:18 Quit Hadaka (
06:45:18 Quit Slasheri (
06:45:18 Quit markun (
06:45:18 Quit Nilisco (
06:45:18 Quit mbr (
06:45:18 Quit ashridah (
06:45:18 Quit Ismo (
06:45:18 Quit Rick (
06:45:18 Quit TiMiD (
06:45:18 Quit CoCoLUS (
06:45:18 Quit Vladoman (
06:45:18 Quit Seed (
06:45:18 Quit pabs (
06:45:18 Quit goa (
06:45:18 Quit pill (
06:45:18 Quit ryan_j (
06:45:18 Quit linuxstb (
06:45:18 Quit solexx (
06:45:18 Quit yosemite (
06:45:18 Quit HCl (
06:45:18 Quit merbanan (
06:45:18 Quit SeeSchloss (
06:45:18 Quit crash_ (
06:45:18 Quit Zagor (
06:45:18 Quit Bagder (
06:45:18 Quit amiconn (
06:45:18 Quit dwihno (
06:45:18 Quit ]Maggie[ (
06:45:18 Quit wireddd (
06:45:18 Quit phaedrus961 (
06:45:18 Quit ghode|afk (
06:45:18 Quit Strath (
06:45:18 Quit ze (
06:45:55NJoinSeed [0] (
06:45:55NJoin]Maggie[ [0] (
06:45:55NJoinVladoman [0] (
06:45:55NJoinashridah [0] (
06:45:55NJoindwihno [0] (n=dw@
06:45:55NJoinBagder [0] (
06:45:55NJoinlinuxstb [0] (
06:45:55NJoinSeeSchloss [0] (
06:45:55NJoinghode|afk [0] (
06:45:55NJoinZagor [0] (n=bjst@pdpc/supporter/sustaining/Zagor)
06:45:55NJoinryan_j [0] (
06:45:55NJoinmarkun [0] (
06:45:55NJoinamiconn [0] (
06:45:55NJoinsolexx [0] (
06:45:55NJoinlostlogic [0] (
06:45:55NJoinpabs [0] (
06:45:55NJoinRick [0] (i=rick@unaffiliated/Rick)
06:45:55NJoinCoCoLUS [0] (
06:45:55NJoinze [0] (
06:45:55NJoingoa [0] (
06:45:55NJoinSlasheri [0] (
06:45:55NJoinwireddd [0] (
06:45:55NJoinIsmo [0] (
06:45:55NJoinphaedrus961 [0] (
06:45:55NJoinmerbanan [0] (
06:45:55NJoinStrath [0] (
06:45:55NJoinmbr [0] (
06:45:55NJoinTiMiD [0] (n=TiMiD[
06:45:55NJoinHadaka [0] (
06:45:55NJoinpreglow [0] (
06:45:55NJoinNilisco [0] (
06:45:55NJoincrash_ [0] (
06:45:55NJoinyosemite [0] (
06:45:55NJoinHCl [0] (
06:45:55NJoinpill [0] (
06:45:57***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
06:46:11CtcpVersion from freenode-connect!freenode@freenode/bot/connect
07:02:10 Join amiconn_ [0] (
07:15:55***Saving seen data "./dancer.seen"
07:20:33 Quit amiconn (Read error: 110 (Connection timed out))
07:20:34 Nick amiconn_ is now known as amiconn (
08:21:58 Join novimon [0] (
08:24:58 Join ender` [0] (i=ychat@
08:50:53 Quit ashridah (Read error: 110 (Connection timed out))
08:51:20 Join ashridah [0] (
08:57:28 Join einhirn [0] (
09:07:05 Join solexx_ [0] (
09:14:56 Join Lynx_ [0] (
09:15:59***Saving seen data "./dancer.seen"
09:23:20 Quit solexx (Read error: 110 (Connection timed out))
10:03:48 Join DangerousDan [0] (
10:51:18amiconnBagder, Zagor: r u there?
10:53:03Zagori'm here
10:59:56amiconnThere was a new level of wiki spam - huge link list in an invisible <div>
11:00:36amiconnI removed it, and I think we really need a more controllable way of registration
11:00:46 Join cYmen [0] (
11:01:39amiconnI think you should delete the user in question
11:02:23Zagoror who was it?
11:03:03Zagorah, SinerAram
11:03:29Zagorhe's deleted already
11:05:57 Join cYmen_ [0] (
11:06:49amiconnMaxSel it was
11:09:00amiconnyou can see the badness here
11:09:26Zagorok. I looked at the IriverPort page
11:11:07 Join cYmen__ [0] (
11:15:01*preglow strokes the new flac decoder
11:15:43*amiconn spots an active preglow
11:16:02***Saving seen data "./dancer.seen"
11:16:21 Join cYmen___ [0] (
11:16:21***Alert Mode level 1
11:16:21DBUGEnqueued KICK cYmen
11:16:21DBUGEnqueued KICK cYmen_
11:16:21***Alert Mode level 2
11:16:21DBUGEnqueued KICK cYmen__
11:16:21DBUGEnqueued KICK cYmen___
11:16:21***Alert Mode level 3
11:17:36 Join linuxstb_ [0] (
11:18:59Zagorunfortunately there doesn't seem to be any registration approval solution implemented for twiki. so we have to make one ourselves, meaning patching the distro code and thus making future upgrades more painful :-(
11:19:51 Quit cYmen (Connection timed out)
11:21:31 Join cYmen [0] (
11:21:31***Alert Mode level 4
11:21:31***Alert Mode level 5
11:21:31DBUGEnqueued KICK cYmen
11:21:31***Alert Mode level 6
11:21:31***Alert Mode level 7
11:21:31***Alert Mode level 8
11:22:40 Join LinusN [0] (
11:23:13 Part LinusN
11:23:39 Quit cYmen_ (Connection timed out)
11:27:01ashridahideally, you don't really want to have to okay each user before they can modify stuff, wouldn't it be better to just put them on a temporary status, and flag any changes they make for moderation up until they get accepted as a full member by someone?
11:27:08ashridahbe less work in weeding out spammers
11:28:43 Quit cYmen__ (Connection timed out)
11:29:23 Join Moos [0] (
11:30:34preglowwould also be harder to implement
11:31:32***Alert Mode OFF
11:31:46 Quit linuxstb (Read error: 113 (No route to host))
11:33:53 Quit cYmen___ (Connection timed out)
11:44:08 Nick linuxstb_ is now known as linuxstb (
11:46:00amiconnpreglow: I have some remarks concerning libffmpegFLAC/coldfire.S
11:47:01amiconn(1) The .align 4 for the jumptable is a waste. Afaik, .align n aligns by 2^n, so it should probably be .align 2
11:47:52amiconn(2) The jumptable isn't optimal, as it involves data access which isn't cached on 5249. The code is in DRAM...
11:48:17amiconnI have two suggestion how to solve (2)
11:48:57amiconn(a) put lpc_decode_emac() in IRAM
11:49:25amiconnor (b) convert the jumptable so that it doesn't involve a data access
11:49:27 Join djit14 [0] (
11:50:01amiconnThis is possible w/o increasing the # of cpu cycles
11:51:42 Join guillaumh [0] (
11:52:32 Join XavierGr [0] (
11:53:24preglowamiconn: i probably will put it in iram
11:53:49preglowthe flac codec has a small amount of iram left, and there's no more data to use
11:55:00preglowamiconn: it aligns by 2^n? this is new to me
11:58:27 Part djit14 ("Leaving")
11:59:26 Quit XavierGr ("Trillian (")
11:59:31 Join webguest82 [0] (
11:59:47linuxstbFor m68k, .align 4 aligns to 4 bytes, for i386, .align 4 aligns to 2^4 bytes...
11:59:50webguest82Do you do all those memories? :)
12:00:09linuxstb"This inconsistency is due to the different behaviors of the various native assemblers for these systems which GAS must emulate. "
12:00:38amiconnMrf, okay :/
12:00:45webguest82Changed translation Program.
12:01:13webguest82Now, can you understand my speech?
12:01:38 Join DrMoos [0] (
12:01:42amiconnSo for m68k this is the alignment in byte, while for SH1 it is the number of low-order zero bits (like on x86)
12:02:22webguest82No one does answer. T.T
12:02:30preglowwebguest82: yes
12:02:35linuxstbI wonder what the ARM situation is.
12:02:50 Quit Moos (Read error: 104 (Connection reset by peer))
12:03:02webguest82preglow: Changed translation program, how about?
12:03:18markunwebguest82: a lot better :)
12:03:42webguest82markun: Thank you. :)
12:03:43amiconnlinuxstb: We could use .balign and .p2align for consistent behaviour, but these are gas specific
12:03:49 Nick DrMoos is now known as Moos (
12:04:00preglowlike we're ever going to use anything but gas anyway
12:04:49preglowconsistent behaviour accross platforms is ok by me anyway
12:08:20pregloweverything else in assembler is platform specific
12:08:20webguest82'Muesli-' did not come yet?
12:08:20amiconnpreglow: With a no-data jumptable, the .align wouldn't be necessary at all
12:08:20linuxstbI agree - balign and p2align seem sensible.
12:08:20preglowamiconn: how would a no-data jumptable look like?
12:08:20webguest82There are much that wonder.
12:08:20amiconnpreglow: You now have:
12:08:20amiconnlea.l .jumptable, %a4
12:08:20DBUGEnqueued KICK amiconn
12:08:20amiconnmove.l (%a4, %d2.l*4), %a4
12:08:20amiconnjmp (%a4)
12:08:20***Alert Mode level 1
12:08:20amiconnand the jumptable consists of .longs
12:08:20***Alert Mode level 2
12:08:20amiconnYou could replace that with:
12:08:20***Alert Mode level 3
12:08:20amiconnlea.l .jumptable, %a4
12:08:20***Alert Mode level 4
12:08:20amiconnjmp (%a4,%d2.l*4)
12:08:20***Alert Mode level 5
12:08:20amiconnThen the jumptable would consist of bra.w insns
12:08:20***Alert Mode level 6
12:08:20amiconnTotal cycle count would be the same
12:08:21***Alert Mode level 7
12:08:21amiconnOld: 1 (lea) + 3 (move.l) + 3 (jmp)
12:08:35amiconnNew: 1 (lea) + 4 (jmp) + 2 (bra)
12:09:17preglowmight shave off a few cycles on the access if it's cached, of course
12:09:48amiconnYou could shave off 2 more cycles for the order8 case - by simply leaving out the last jumptable entry
12:10:44 Quit Moos (Read error: 104 (Connection reset by peer))
12:10:48 Join DrMoos [0] (
12:10:49preglowyes, indeed
12:10:53preglowand the order8 case is very common
12:11:34webguest82Is no there here H300 firmware programmer?
12:12:04preglowwebguest82: no
12:12:47preglowwebguest82: there is only one h300 programmer at the moment, and he is busy
12:13:00webguest82Where does live, preglow?
12:13:12preglowwebguest82: norway
12:13:29webguest82What time is there now?
12:13:55preglowwebguest82: 12:14 right now
12:14:19preglowamiconn: i can shave off a couple more bytes by doing PC relative jmp instead of using a lea
12:14:59preglowwebguest82: pm
12:15:11preglowwebguest82: as in the morning, or whatever
12:15:19preglowi never use pm/am ;)
12:16:45webguest82I am sorry, but I do not understand well. ^^;
12:17:39preglowamiconn: if i could only think of a clever way of doing it without hard coding the address difference :)
12:17:51webguest82Still, be so good becoming communication.
12:18:22***Alert Mode OFF
12:19:57webguest82Is not it so?
12:23:44 Quit webguest82 ("CGI:IRC (EOF)")
12:24:38 Join webguest82 [0] (
12:29:39preglowhah, i don't believe i need any address difference
12:29:41CtcpIgnored 6 channel CTCP requests in 51 minutes and 59 seconds at the last flood
12:29:41*preglow tries out
12:30:56*preglow resets his iriver
12:32:38webguest82Do you have an interested person here in South Korea?
12:33:08webguest82Or in iriver.
12:34:03preglowi don't know
12:35:55webguest82Programmers here?
12:36:28 Join muesli|tarn [0] (n=muesli_t@
12:36:47muesli|tarni could connect to freenode from here :D
12:37:07preglowthis jump table wont insult ME again!
12:38:03preglowlinuxstb: did you have a flac test program that wrote to disk on target?
12:39:17webguest82Although I want to know in detail, I am sad to be not so.
12:39:40 Join muesli- [0] (n=muesli-@
12:39:46 Part muesli-
12:39:52webguest82hi muesli-
12:40:17webguest82I waited you.
12:40:44muesli|tarnhere i am
12:41:06webguest82Are you H300 firmware programmer?
12:41:40muesli|tarnneither own a h300 nor can
12:41:40muesli|tarncode a single line :-/
12:42:45webguest82By the way, seem to know much about H300.
12:43:44linuxstbpreglow: No. Only a standalone test program.
12:44:05webguest82muesli-: Is not it so?
12:44:59 Join muesli- [0] (n=muesli-@
12:45:12muesli|tarnno...unfortunately my iq is quite coding skills
12:45:21 Part muesli-
12:45:48muesli|tarnsorry 4bothering...have to check my settings
12:46:27webguest82Do I mistake?
12:47:48linuxstbpreglow: But I may have to write one to try and trace the bug with the freezing file.
12:49:14 Join muesli- [0] (n=muesli-@
12:49:24 Part muesli-
12:49:45 Nick muesli|tarn is now known as muesli_ (n=muesli_t@
12:51:35preglowamiconn: will functions be aligned automatically by gas?
12:52:15webguest82muesli: Had not you spoken that reclamation is available from H300 to 25fps before?
12:53:17preglowit's clear that muesli hasn't done any advanced assembler optimising before :P
12:53:27muesli_preglow ;p
12:53:53muesli_never optimized anything in any coding language ^^
12:53:54webguest82Did I misguided?
12:54:18webguest82:) I am sorry .
12:54:25muesli_no worries m8
12:54:33muesli_ask preglow instead ;)
12:55:16webguest82What is 'm8'?
12:55:32webguest82I do not know well that do.
12:55:41muesli_no worries m8
12:56:18webguest82Are there much that preglow knows?
12:57:08preglowwell, i know that 25 fps video on h300 will be almost impossible
12:57:34preglowat least with divx/xvid
12:57:35webguest82Did not you speak that is possible at last time?
12:57:44preglowno, i did not
12:58:39webguest82I misguided, do not English.
12:58:59webguest82Then, how is it possible?
12:59:39amiconnpreglow: No, gas doesn't auto-align functions
13:00:42preglowwebguest82: like i said, i think it is not possible
13:00:48preglowwebguest82: i don't think the h300 has enough cpu power
13:01:19preglowamiconn: commited new version now, check it out and see if something can be improved still :)
13:01:54webguest82Then, is not the part improved?
13:03:35amiconnpreglow: This jumptable mechanism is clever :)
13:09:11webguest82Is H100's firmware working continuously?
13:11:55webguest82preglow: What is thing which is working now in priority?
13:12:14webguest82In Rockbox
13:14:03preglowwebguest82: no priorities, people just work on what they feel like
13:14:45webguest82Are you doing work that each wants?
13:14:46webguest82Are you doing work that each wants?
13:15:14preglowamiconn: what was the previous problem with const IDATA_ATTR again?
13:15:33webguest82What work is amiconn doing?
13:15:34preglowamiconn: i thought you couldn't use IDATA_ATTR on const data, but there are some arrays in libmad that do just that
13:15:47preglowamiconn does lots of different work
13:16:02preglowhe's got his hands in more or less all parts of rockbox
13:16:03***Saving seen data "./dancer.seen"
13:16:28webguest82aha~ ^^
13:16:44webguest82Very busy.
13:16:55amiconnpreglow: The problem with IDATA_ATTR and sconst was that you would get a section type conflict - but this happened only if you used it for const and non-const data in the same file
13:17:56preglowamiconn: ahh, that explains it, well, i'll update all occurences to ICONST anyway
13:18:59preglowliba52 has a lot of data that might as well be const, but isn't
13:19:37webguest82Who is operator of rockbox?
13:20:15amiconnpreglow: Maybe linuxstb removed the consts because of such conflicts?
13:20:39webguest82Because site structure of rockbox is complex, beginner such as me does not know well.
13:21:01 Join muesli- [0] (n=muesli-@
13:22:16 Part muesli-
13:23:14preglowoh well
13:23:24preglowi've fixed a lot of the IDATA_ATTRS now
13:23:28preglowwill test to see if i broke something
13:25:48webguest82Next, again see.
13:26:04webguest82preglow: I am appreciative today.
13:26:10preglowwebguest82: good to hear
13:26:41 Quit webguest82 ("CGI:IRC (EOF)")
13:30:26linuxstbamiconn: No, I don't remember doing that.
13:32:18preglownot all people use const
13:46:58linuxstbpreglow: Have we decided that all codecs should now be outputting 28-bit samples - in the same way as the new FLAC decoder does it?
13:47:25linuxstbIf so, I can quickly change ALAC. Not sure about the others though.
13:47:47preglowlinuxstb: well, that's my view, at least
13:47:54preglowi don't know if anyone disagrees
13:48:10preglowall codecs can be easily fixed, perhaps excepting wav
13:48:25preglowwavpack too stores all it's data in a 32 bit array as it is
13:49:02linuxstbSo the plan is to remove the SET_SAMPLE_DEPTH configuration option?
13:49:22linuxstb(after all the codecs are changed to set it to 28)
13:50:18preglowit will be redundant, yes
13:50:31preglowi plan to remove SET_SAMPLE_DEPTH, clipping limits and DSP_ENABLE
13:50:58linuxstbMy only concern is that we have unneccessary shifting in the case of the DSP doing nothing with 16-bit data.
13:51:07linuxstbBut that already happens.
13:51:36preglowwell, how do you suggest we avoid that?
13:52:00preglowwe could of course have a separate output loop in the codec plugin, and have the codec plugin query the dsp layer whether any processing is to be done
13:54:38 Quit Vladoman (Read error: 104 (Connection reset by peer))
13:55:04linuxstbI'm not sure what the best solution is - making the codecs always output 28-bit samples is the cleanest, but possibly less efficient.
13:57:44linuxstbBut if you forced me to decide, I think I would vote for the 28-bit solution. I think we'll probably end up always using the DSP for something.
13:58:26preglowbut btw
13:58:33preglowwe'll still need the SAMPLE_DEPTH
13:58:38preglowfor deciding where to apply dithering
13:58:51preglowwe can't just pretend all the 28 bits will always contain valid samples
13:59:18linuxstbOK. But now it will mean what it says.
14:00:28linuxstbMaybe it should be a parameter to the pcmbuf_insert function, along with samplerate and number of channels.
14:01:09preglowand btw, 28 bits dont mean the upper bits can't contain data
14:01:19preglowbut that data will be clipped later on if it's not properly scaled down first
14:23:03 Quit muesli_ ("ich will Khe!!!")
14:25:37linuxstbAnyone know how Rockbox/Coldfire reacts to an unaligned memory access? Would it give an error message, or just freeze?
14:27:10preglownothing special
14:27:18preglowan unaligned access i just slower
14:29:45linuxstbI'm just trying to think of reasons the FLAC codec could freeze on the target, but work in the sim.
14:30:26 Quit DangerousDan (Read error: 104 (Connection reset by peer))
14:31:38linuxstbI suppose a buffer overflow could be a reason.
14:33:01preglowoh yes
14:34:54linuxstbBut it only seems to happen with dbpoweramp encoded files... I've asked what encoder parameters it uses, but haven't got an answer yet.
14:35:20linuxstbThis time it isn't an ID3 problem - there are no unusual tags in the file.
14:36:16 Join DangerousDan [0] (
14:38:05 Join ender1 [0] (i=ychat@
14:38:23preglowit's closed source?
14:40:09 Join novimon_ [0] (
14:40:10linuxstbI'm not sure.
14:40:24 Quit Lynx_ (Nick collision from services.)
14:40:58preglowweird that it only crashes on target
14:41:00 Quit ashridah (
14:41:00 Quit ender` (
14:41:00 Quit novimon (
14:41:00 Quit lostlogic (
14:41:26preglowi could do a quick remote debug session for you if you get me the file, though
14:42:41linuxstbThat would help - thanks. I posted the link last night.
14:44:05NJoinashridah [0] (
14:44:05NJoinender` [0] (i=ychat@
14:44:05NJoinlostlogic [0] (
14:45:35preglowlinuxstb: 404
14:48:22linuxstbTry now.
14:50:58preglowlinuxstb: doesn't freeze here
14:51:43linuxstbI'm an idiot. I've just realised that.
14:52:13preglowgiven five tries it even plays!
14:52:16preglowno problems at all
14:52:24preglowbut there's clearly something funky going on
14:53:32linuxstbI broke my version of the FLAC decoder last night - at the same time I decided to test that file.
14:54:19preglowok, but someone clearly said it crashes?
14:54:21preglowor is this your file?
14:54:47linuxstbI think he must be doing something wrong.
14:54:53preglowit _is_ broken somehow, most times when i try to play it, the wps shows briefly, i get a little sound, then boom
14:54:58preglowand time to try again
14:55:39linuxstbYes - the first time it played it worked fine. The second and third times I am getting the same as you.
14:58:09linuxstbYes, a "codec failure" is popping up very briefly for the second play.
15:00:11 Quit ender` (Connection timed out)
15:00:15preglowwhy so briefly?
15:00:22preglowi didn't see it until you tols me
15:02:22 Join Lynx_ [0] (
15:05:08linuxstbThis is getting strange. I've now reverted back to the CVS version of the decoder, and it freezes on that track and plays my other files.
15:05:32preglowi've got cvs plus my emac routine
15:05:33preglownothing more
15:05:48linuxstbSo I'm thinking it's an uninitialised variable being accessed which is causing strange things.
15:07:05linuxstbWhich gcc for coldfire are you using?
15:10:14linuxstbI'm using 3.4.4, and I think the server does as well
15:11:07preglowthe server?
15:11:13preglowno, that uses 3.4.1 :)
15:11:21linuxstbAh. OK then.
15:11:53preglowdon't ask why
15:11:55preglowlaziness, i suppose
15:12:04linuxstbI've just added a memset to initialise the FLACContext structure, and that stops the crash. But on broken.flac I get an immediate codec error instead of a crash now.
15:13:53preglowwell, memseting the flaccontext is certainly a good thing
15:14:24linuxstbYes. I'll commit it.
15:16:05***Saving seen data "./dancer.seen"
15:17:40preglowlots of potential codec error places in flac.c
15:17:57preglowi wish we had several return codes here as well
15:18:11 Join LinusN [0] (
15:19:05preglowhi, linus
15:22:48LinusNhey ho
15:22:53 Quit dwihno (Read error: 110 (Connection timed out))
15:28:28LinusNTiMiD: u there?
15:31:34 Join fogcat [0] (
15:42:20markunLinusN: comments on his patch?
15:46:30preglowanything major that needs chaning?
15:47:15 Quit ashridah ("Leaving")
15:53:04 Quit fogcat ("CGI:IRC")
15:55:37LinusNpreglow: not at all
15:57:11preglowso it's time for official remote support soon?
16:08:20 Join Febs [0] (
16:20:51solexx_which patch is it? the complete one (wps, menu, dir browser)?
16:26:07 Quit Febs ("CGI:IRC (EOF)")
16:26:43 Join Febs [0] (
16:28:33LinusNsolexx_: afaik, there isn't a complete one from TiMiD
16:41:01 Part LinusN
16:57:08solexx_hm. i remember having used an inofficial version which had everything i can think of (wps, browser and menus)
16:57:32solexx_don't remember who built them but he's a regular in here
16:58:16amiconnProbably Firefly's version. That one won't be committed to cvs (q&d apprach with loads of code duplication)
16:59:18solexx_that's ok with me
16:59:45solexx_i promised ti give away my remote anyway. a friend of mine got his remote stolen and i don't use it anyway.
17:00:16solexx_...and as we all know, quality pays off in the long run :)
17:00:42amiconnTiMiD's approach is probably the way to go, but it's not yet complete
17:04:46 Join bagawk [0] (n=lee@unaffiliated/bagawk)
17:05:33solexx_does it always display the same thing as the main display?
17:08:08*solexx_ has to go home (iow: afk)
17:12:01*preglow looks at the mailing list patch
17:12:29preglowwill aligning function to 16 help much?
17:12:52preglowahh, right, for the code cache load
17:15:53preglowlinuxstb: what do you know about the block size in flac? is it always an even number?
17:16:07***Saving seen data "./dancer.seen"
17:30:52linuxstbNo - the block size is the number of samples. So if the file has an odd number of samples, the last block will be odd.
17:31:13linuxstbBut almost every block you see (apart from the last one) will be 4608.
17:31:20preglowi wonder what this rounding error he's speaking about is
17:34:20 Quit wireddd (Remote closed the connection)
17:35:15linuxstbHmmm. The person who reported the FLAC problem now says that all his files play OK - after the memset patch.
17:35:57linuxstbpreglow: Are you still getting unreliable playback on broken.flac?
17:35:58preglowperhaps i'll try as well
17:36:05preglowhaven't built it again
17:36:20linuxstbI'm going to download the official build and try it.
17:38:55 Quit bagawk ("Leaving")
17:39:01preglowlinuxstb: played on the first try
17:39:13preglownot the second
17:39:23preglowor third
17:39:32preglowor fourth...
17:39:34preglowbut fifth!!!
17:41:31preglowand is suddenly get a compiler warning
17:41:38preglowflac.c: In function `codec_start':
17:41:38preglowflac.c:116: warning: 'samplesdone' might be used uninitialized in this function
17:41:50pregloware additional warning flags turned on when you're compiling a debug build?
17:45:25linuxstbI find that it doesn't work on the first try - I get a codec error flashing up and am returned to the file browser. But if I select the file again immediately, then it will play. If I wait a second or two, then it doesn't play.
17:45:32 Join ep0ch [0] (n=ep0ch@
17:46:23linuxstbpreglow: That warning is due to the LOGF at the end of flac.c
17:46:25ep0chwhere's this flac file you two are obsessed with atm?
17:47:12linuxstbThe only known difference with other files is that it was encoded by dbpoweramp
17:49:59ep0chand dbpoweramp uses libflac?
17:50:11linuxstbI think so.
17:50:19linuxstbI don't know any other FLAC encoders.
17:50:30*amiconn wonders what this colon-separated path thingy is about
17:50:47ep0chfoobar says that file is libFLAC 1.1.2 20050205
17:50:47linuxstbNeeded for a new bootloader? :)
17:51:12linuxstbYes - now you mention it, I remember seeing that in the file header.
17:52:58ep0chheh first attempt at playing it has locked my player up
17:54:33linuxstbAre you using the very latest build? I committed a change a couple of hours ago that fixed that part of the problem.
17:54:47ep0chyeah built 3 mins ago
17:55:42linuxstbCan you check you have the very latest apps/codecs/flac.c version?
17:56:46ep0ch$Id: flac.c,v 1.10 2005/10/27 13:18:05 dave Exp $
17:58:58ep0chi'm stupid...
17:59:09linuxstbGlad to hear it :)
17:59:57novimon_the new flac decoder doesnt work for me either
18:00:21ep0chis working for me
18:00:29ep0ch20 seconds in, first time
18:00:50ep0chwhen does it crash?
18:01:11novimon_right after when I open a .flac
18:01:13preglowit doesn't crash anymore
18:01:15preglowat least not for me
18:01:17novimon_needs to reset
18:01:27ep0chnovimon_: rebuild rockbox
18:01:28novimon_and the hd is on
18:01:39novimon_i dont know how to :<
18:01:53ep0chdownload latest build
18:02:03novimon_i did
18:02:08ep0chwell its working now
18:02:15novimon_but ill do it again
18:02:20ep0chdaily or bleeding edge?
18:04:00linuxstbSomeone has just PM'ed me on the Forum with a link to another FLAC file that causes problems. I'm going to test it now.
18:04:04ep0chget bleeding edge
18:06:11novimon_thank god for the usb-boot mode
18:06:26novimon_I had resume on so my iriver crashed right after booting :)
18:07:11novimon_woot, works now
18:07:13novimon_thanks mate
18:07:34novimon_cant wait for the seeking to be implemented :)
18:07:46 Nick Lynx_ is now known as Lynx_awy (
18:08:42linuxstbI have no problems with the second test file.
18:08:52amiconnEven w/o bootloader USB it should be possible to recover from a startup-resume crash. The rockbox core itself has early USB detection (which doesn't do much on iriver because of bootloader USB)
18:27:02SlasheriI think it should be made possible to recover that kind of crash without usb at all
18:27:19 Nick ender1 is now known as ender` (i=ychat@
18:27:39SlasheriFor example some early button press detection or automatic crash indicator fla
18:29:06 Nick DrMoos is now known as Moos (
18:29:24linuxstbMmm. I don't understand. My libffmpegFLAC/Makefile still refers to libalac everywhere (that's where I copied it from). So I don't understand why it is working.
18:29:48preglownot bad
18:29:59preglowSlasheri: any news on the glitching track change bug?
18:30:06linuxstb"lowlight" just pointed this out on the Forum.
18:30:17Slasheripreglow: hmm, is that bug still present? :)
18:30:36preglowperhaps :P
18:30:42preglowi didn't know you tried to fix it
18:30:54preglowi haven't used my unit for actually listening to music for a while
18:31:06Slasheriyes i tried.. but you might be right, i haven' tried skipping tracks very fast
18:31:19Slasheriand in that case the bug might still appear
18:31:35Slasherithen we would just need to stop audio completely
18:31:46 Join matsl [0] (
18:32:36preglowwell, that's what i'm used to anyway
18:32:47preglowbut i'll try it out and see
18:45:53 Quit ep0ch ("Trillian (")
18:55:02 Join _aLF [0] (
18:55:59 Quit Moos (Read error: 104 (Connection reset by peer))
18:56:01 Join DrMoos [0] (
18:56:11 Nick DrMoos is now known as Moos (
18:59:20solexx_has anybody investigated which codec has the least overall power consumption
18:59:50linuxstbWAV ?
18:59:57linuxstbOr maybe not....
19:00:40solexx_i guess it depends on the proportion of hd and cpu power consumption
19:00:50solexx_which i don't know
19:05:41preglowsolexx_: why does it matter?
19:06:00solexx_i was just curious
19:06:17preglowyou would most certainly have to do a couple of very long tests
19:07:12solexx_the question arose mainly because i am considering to only put flac files on my iriver
19:07:34solexx_and i asked myself whether the heavier disk usage would drain the battery more
19:07:42preglowit will
19:07:57linuxstbsolexx_: It sounds like you're the person to do some tests :)
19:08:11solexx_but i guess it doesn't matter much because i seldomly have my player powered on for more than one or two hours
19:08:54linuxstbAt least by choosing FLAC, you can always change your mind in the future without loss in quality.
19:09:19solexx_yes, that is why i have ripped almost my whole cd collection again
19:09:29solexx_(and then i bought a new hard disk)
19:09:52 Quit ghode|afk ()
19:09:56linuxstbWavpack is probably slightly better more efficient than FLAC - but I'm not sure by how much now.
19:10:47solexx_is the runtime (in the debug menu) ever reset to zero without manual intervention?
19:12:54preglowi don't think wavpack has much on flac any more in terms of efficiency
19:14:43 Join Sandking [0] (
19:16:10***Saving seen data "./dancer.seen"
19:17:33 Quit matsl ("Leaving")
19:18:13 Join matsl [0] (
19:18:23 Join ghode|afk [0] (
19:23:14amiconnsolexx_: Yes, the runtime in debug is reset to zero when you charge the unit while powered on. Otherwise rockbox doesn't know that the battery was recharged...
19:24:14 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
19:25:12solexx_thanks. does rockbox need that to estimate remaining power?
19:25:44amiconnNo. It's just a statistics feature
19:25:46Sandkingi just thought that syncing rockobox with itunes would be quite cool - anybody thought about it?
19:27:56linuxstbMost people come to Rockbox as a way to be free of a specific PC music manager...
19:28:35linuxstbThe main problem would be that Rockbox supports a lot more file formats than iTunes, and doesn't support AAC (at the moment).
19:29:18linuxstbBut it may not even be possible - Rockbox has no control over the USB connection, so wouldn't be able to emulate an iPod, if that was needed.
19:30:31Sandkingi was just curious - thanks
19:30:49Sandkingi use itunes so it would great for me but i get it
19:31:39linuxstbMaybe here:
19:32:06Sandkingi got a msg "not a rockbox runtimedatabase" when i start - how to rebuild database? i can't find a guide about it on but i remember it was there
19:35:22 Join _FireFly_ [0] (
19:35:53 Quit linuxstb (Read error: 104 (Connection reset by peer))
19:36:38 Join linuxstb [0] (
19:37:20preglowwho the hell would want to emulate an ipod anyway?
19:37:28preglowi hate itunes
19:37:57Sandkingi would like to
19:38:04amiconnSandking: Afaik, a rebuild tool doesn't exist. Either your db became corrupt, or is a very early version with a different format from what is currently used
19:38:25Sandkingok, so i need only info about building a new one
19:38:34Sandkingi can't find it :/
19:38:49amiconnYou can simply delete rockbox.rundb and reboot the box
19:39:09amiconnAn empty db should then be recreated
19:39:29linuxstbSorry - runtimedatabase....
19:39:35Sandkingok, thx
19:40:19Sandkingi just wanted to say you guys that rockbox starts about 12 seconds while iriver 24... :D thanks for the soft
19:40:31Sandking[tell you - sorry for bad english]
19:43:57Sandkingamiconn - i deleted it and it works. brilliant :D
19:44:17Sandkingok, now it takes 8 seconds to start :D
19:46:39preglowwe rock
19:47:46*amiconn wants archos rockbox' boot time on iriver
19:48:03*preglow doesn't care that much
19:48:07preglowit's fast enough for me already
19:50:15*amiconn wants instant-on
19:51:00Moosoh yeah, I love the boot-time of my Ondio quasi instant-on
19:51:13_FireFly_amiconn: then you must have a standby function in rb :)
19:51:26_FireFly_but this wouldn't be could for battery-life
19:51:59amiconnPerhaps that's even possible. The sdram controller can switch the ram into low-power self-refresh mode and back
19:53:45linuxstbThat's how the ipod works - it has no power-off, just a sleep mode.
19:53:53linuxstb(with Apple's firmware)
19:54:10amiconnRockbox could have that as an option...
19:54:47_FireFly_after long travles without using it you would have to recharge it before using ;)
19:55:04linuxstbWould we be able to keep track of the time during the sleep?
19:55:41amiconnWe could even have a seletable shut-down threshold, like "go to standby, but power off when battery drops below 30%"
19:56:00amiconnlinuxstb: yes
19:56:07_FireFly_is there a automatik shutdown when battery gets low
19:56:40amiconnIt's almost the same as my wake-up timer idea, except the dram contents is preserved
19:56:52linuxstbSo we could emulate an RTC - as long as we never power down completely?
19:57:08_FireFly_i think so
19:57:17amiconnLet the standby code run in IRAM, don't touch the DRAM, and reduce the timer interrupt to 1 per second
19:58:20linuxstbThat sounds nice. I miss having a time display in Rockbox.
19:58:24amiconnI have to do some measurements what needs less power: running at 11 MHz with the pll powered down, or running at the lowest possible clock (which needs the pll running)
19:59:29amiconnThe power measurements are also necessary for estimating the possible standby time
20:02:38SlasheriWhat about frequency changes, those will cause significant drift to the clock?
20:04:48amiconnNot really. The most significant drift will result from the fact that the oscillator, although being an xtal, isn't fine tuned, because it wasn't intended to be used as a clock
20:04:59amiconnI'd expect an error of a few seconds per day
20:05:41linuxstbWe would also need to solve the problem of keeping the RTC during a firmware upgrade. I'm guessing that's possible though.
20:06:32_FireFly_maybe a timestamp could be saved on disk or in the config-block
20:06:41solexx_from a user's pov it wouldn't hurt having to manually reset it after an upgrade
20:06:51linuxstbI wouldn't want to though :)
20:07:03solexx_understandable :)
20:10:03amiconnSlasheri: You can do a test which I already did, but doing it on several units might be helpful:
20:11:02amiconn(1) Start playing some music. Either make a long long playlist, or enable repeat all. Choose a format which will cause the CPU to be boosted every now and then
20:11:40amiconn(2) Take a precise stop watch, load the stopwatch plugin, and start both the stopwatch plugin and the external stopwatch
20:12:22amiconn(3) Let it run for several hours. Then stop both the stopwatch plugin and the external stopwatch. Note the difference
20:13:04amiconnThe stopwatch plugin uses the timer tick, so it's as precise as an rtc emulation would be
20:14:08Mooswhy don't use it for a random playing files mode?
20:18:39Slasheriamiconn: hmm, that sounds good. I did a similar test earlier (but then the pll code was different and it needed to re-lock when changing operating frequency) and there was severe drift
20:19:07amiconnThe pll always needs to needs to re-lock on freq changes
20:19:25SlasheriAh, hmm. But anyway, i will re-test soon
20:19:32amiconnHowever, the drift you observed was probably caused by a different problem
20:20:58Moosamiconn: nevermind :-)
20:21:03amiconnThe timer cycle count for the tick was off by one, both on archos and on iriver. The effect was bigger on iriver because the absolute count was lower
20:21:36amiconnThe error caused by this was ~1 second per hour
20:21:38Slasherioh, that could be the reason. I think the drift was systematic
20:26:33amiconnMoos: I didn't understand what you mean, hence the ???
20:28:31MoosJust I remenbered one discussion few days ago about the fact that Rockbox in the iriver at least don't have one "randomly play song" ready, cause we have'nt RTC for this
20:29:10Moosbut it's nothing :)
20:29:45MoosIf we want to play son random we need to play the root playlist for this
20:30:02amiconnRockbox has very good randomness without an rtc.
20:30:18Mooswith archos?
20:30:44amiconnOn all platforms
20:31:04Moosneed to use the re-shuffle playlist option
20:32:29Mooswe can't play song random with the random option, it doesn't work well
20:32:35Moossorry for my english
20:33:22MoosI just remenbered one discusion about this, let me check google
20:33:35amiconnThe necessity for root.m3u has 2 other reasons: (1) The whole rockbox playback concept is playlist based. (2) Rockbox doesn't automatically know all files on the disk because it doesn't scan the whole disk on startup
20:34:01amiconnThe random option doesn't work well? It does here...
20:34:02Moosyes indeed
20:34:34Mooslet me see
20:35:09amiconnIn fact I wouldn't want 'random' playback, 'shuffle' is better. It doesn't repeat songs before the playlist ends (and starts over if you have repeat enabled)
20:35:37Moosah you understand what I mean
20:36:17amiconnOf course you still need a good random number generator for good shuffle
20:36:59Moosoh yes this is what I wanted to say
20:37:00amiconnThink of random playback as drawing a random card from a deck (=all tracks), playing it, then put it back to the deck
20:37:10Moosit's depand in RTC,no?
20:37:18amiconnShuffle is different in that it doesn't put back the card
20:38:27amiconnMoos: I wonder who told the fairy tale that you need an rtc for good randomness...
20:38:38preglowa skilled programmer, no doubt
20:38:48Moossorry mistake of me
20:39:00amiconnpreglow: ;)
20:39:57_FireFly_you only need function/var which returns/give a value which changes
20:40:15_FireFly_to initiate the random-generator
20:40:30preglowyes, and we happen to have that
20:40:38_FireFly_yepp the cpu-ticks
20:40:53amiconnThere are even several options to obtain a true random seed
20:40:57MoosI'm stupid sorry for this guys
20:41:18preglowno worries, moos
20:41:23_FireFly_Moos: not stupid only miss informed
20:41:25preglowyou're not the only one believing you need an rtc for random
20:41:34amiconnWe use the ticks, but there are more possibilities: last disk spinup time, current battery voltage, time between two button presses etc
20:48:19 Join muesli- [0] (
20:51:07_FireFly_na und ?? ;)
20:52:00muesli-had a panic warning today *Stkov*
20:52:16muesli-happened when i was browsing by remote
20:52:40muesli-i pressed the menu button but nothin happened
20:52:54muesli-and then i got that msg on units screen
20:53:54 Join miner49er [0] (
20:54:22muesli-and sometimes the selection bar disappears :o
20:55:09_FireFly_Stkov = stackoverflow
20:55:50muesli-did i press too fast in a too short time?
20:56:10_FireFly_i don't know i have only search for Stkov in the sources ;)
20:56:23muesli-mk ;)
20:56:56_FireFly_mk ??
20:57:11muesli-mmh okay ;)
20:57:27muesli-so isset ;)
20:57:31_FireFly_emkay ;)
20:58:16muesli-did you read TiMiD 'S plugin thread?
20:58:23_FireFly_a liitle bit
21:04:37 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
21:12:34miner49eri'm getting lots of warning: comparison between signed and unsigned when building GCC. Any ideas? Oh, hi there by the way!
21:15:01 Quit ender` (Read error: 113 (No route to host))
21:15:09paughminer49er: new warnings from gcc 4*something. probably safe to ignore unless you have a specific problem tha arises from it.
21:15:36 Join ender` [0] (i=ychat@
21:16:11***Saving seen data "./dancer.seen"
21:17:09miner49erwell, it hasn't been building until I noticed that I had to do something special for sh1. I know, I should've read the instructions. But I just did what it said and it's still giving me those warnings - maybe it will compile this time anyway? We'll see...
21:23:32 Join ep0ch [0] (n=ep0ch@
21:24:06_FireFly_linuxstb: i can't compile the latest cvs version
21:26:23 Join Sandking_ [0] (
21:27:01_FireFly_why does the vorbis-codec be linked against libffmpegFLAC ??
21:28:22ep0chwhat's a good free text editor in windows besides textpad?
21:28:30_FireFly_metapad :)
21:28:47ep0chgvim eh :)
21:29:14_FireFly_he whants only a texteditor not an IDE-like editor ;)
21:29:15preglow_FireFly_: works just dandy here
21:29:19ep0chdoes metapad let me have multiple files open?
21:29:39preglowTiMiD: yo, linus was looking for you
21:29:45ep0chwell, a very lightweight ide would be nice
21:29:53miner49eris a P3 533 a bit too slow to build rockbox - GCC takes forever to build...
21:29:55TiMiDI saw :)
21:30:11miner49erwhat's the best pretty-printer type editor I can use for debian?
21:30:12TiMiDbut he is not here atm :(
21:30:21preglowand he probably wont be until tomorrow
21:31:32ep0chhow hard do you think it will be for me to add the playlist counters to the status bar? bearing in mind i have little C knowledge
21:31:42_FireFly_ah it works i had to redo a ../tools/configure
21:32:11preglowahh, yes, you'll need that
21:34:17miner49erep0ch: I use Crimson editor in windows it's very nice. I also think the Bloodshed C IDE and compiler is very cool...but i'm trying to get into linux at the moment!
21:34:48_FireFly_devcpp is a ide which uses the cygwin/mingw gcc
21:35:10 Join linuxstb_ [0] (
21:35:20 Quit linuxstb (Read error: 110 (Connection timed out))
21:35:44ep0chok I'll look at Crimson too
21:38:08 Join LinusN [0] (
21:38:40TiMiDhi LinusN
21:38:53TiMiDyou were looking for me ?
21:38:56LinusNi worked a little on your patch today
21:39:08LinusNi moved the gui_xxx files to apps/gui/
21:39:18LinusNand removed the gui_ prefix
21:39:54TiMiDyou removed the gui_ prefix from the function names or just from the files ?
21:39:58 Quit Sandking (Read error: 110 (Connection timed out))
21:39:59LinusNfiles only
21:40:03TiMiDok :)
21:40:08TiMiDthat's perfect
21:40:17TiMiD(I wanted to do that btw)
21:40:32LinusNthat was the huge Makefile commit i did earlier today
21:40:48TiMiDI look
21:41:30 Quit SeeSchloss (Remote closed the connection)
21:41:32_FireFly_the "colon separated path"
21:41:37LinusNi was browsing through some of your code and i saw a thing that i wanted to ask about
21:43:06LinusNin tree.c, rockbox_browse(), you removed the "reload_dir = true"
21:43:47TiMiDI look
21:44:33LinusNi just wanted to make sure it was on purpose
21:45:10_FireFly_it is possible to find out if a command to the remote was successfully send ??
21:45:11amiconnLinusN: I found something concerning our rng. More specifically, how we can improve the seeding:
21:45:51TiMiDLinusN: hmm
21:45:54amiconnWe can leave out the array init, the standard 32 bit init is also improved
21:46:06TiMiDI don't remember to have done such a thing
21:46:12_FireFly_i mean that the remote has received the command
21:46:16LinusNamiconn: interesting
21:46:25TiMiDI must have deleted it accidentally :)
21:46:31LinusN_FireFly_: no
21:46:56TiMiDanyway don't think it was necessary here
21:47:08amiconnLinusN: I can try to adapt that to rockbox.
21:47:10LinusNTiMiD: it seems to have been redundant, yes
21:47:15LinusNamiconn: good
21:47:25amiconnHmm, it seems I'm trying to avoid working on memcpy :/
21:47:28miner49erAAAHRRGG!! GCC hasn't built, can anyone help me?
21:47:36LinusNminer49er: huh?
21:47:48TiMiDit's necessary yes :)
21:47:55_FireFly_it is mainly for a cosmetic problem
21:48:08TiMiDreload_dir= true
21:48:11TiMiDok no need
21:48:25LinusNTiMiD: also, on line 837, there is a commented call to gui_synchronized_statusbars_draw()
21:48:32miner49eri've been trying to build GCC, i've followed the instructions but it fails eventually (after about half-an-hour) with unrecognised option -big
21:48:32TiMiDthat why I removed it I think (I don't remember clearly since it has been a month)
21:48:49_FireFly_because the wps-tag for the remote-hold retuns that the hold is on when the remote-plug is only half plugged in
21:48:50miner49eralso getting warning: signed compared to unsigned
21:49:13LinusNminer49er: which gcc version, which target and which host operating system?
21:50:11miner49erokay, GCC 3.3.6, target SH1 (i think, I have an FMR), OS is GNU/linux Debian Sarge distro
21:50:11LinusN_FireFly_: iirc, that's a problem with the wps tag, where it doesn't properly handle the remote detection
21:50:39_FireFly_i have a patch running which solves the problem for the wps
21:50:42miner49er...I really don't want to use XP (which I have been doing happily for ages btw)
21:50:48TiMiDLinusN: it's because there was a call to the statusbar_draw
21:50:59TiMiDmaybe it should be uncommented
21:51:04_FireFly_if the remote-plug is complete plugged in but not if the plug is only half plugged
21:51:14TiMiDbut I don't see any reason why the status bar would be updated here
21:51:45_FireFly_so how does the iriver fw handle this because the "problem" is only in rb
21:51:48LinusNminer49er: did you do the newlib workaround?
21:52:17miner49eryes I did. Is there anyway of checking if I did it correcly? I am a linux newbie btw
21:52:26LinusNTiMiD: then the call should be removed, not commented
21:53:43LinusNminer49er: cd gcc-3.3.6
21:54:07TiMiDLinusN: the only reason for the call to be here is when the user leaves his player in filetree for a long time (the battry value may change)
21:54:19TiMiDbut I don't know if it's very useful
21:54:20LinusNminer49er: ls -ld newlib
21:54:38LinusNTiMiD: remember that the archos has a clock
21:54:38miner49er'kay, in the directory
21:55:05TiMiDoops :)
21:55:16TiMiDso it must be here ^^
21:55:40TiMiDa good thing I leaved it commented (I was not sure but I forgot to ask)
21:56:25miner49erI get: lrwxrwxrwx 1 root root 24 2005-10-27 20:06 newlib -> ../newlib-1.11.0/newlib
21:56:48LinusNTiMiD: what's left to do now, so we can get rid of the old gui implementation?
21:56:55LinusNminer49er: correct
21:57:13miner49erthat's good then :-)
21:57:30miner49er...I guess
21:57:36LinusNminer49er: did you install binutils and set the PATH correctly before configuring and building gcc?
21:57:37TiMiDLinusN: I planed to do in order : menus, playlist
21:57:40TiMiDand then wps
21:57:59TiMiDwe will gain some size with menu and playlist
21:58:01LinusNTiMiD: problem is that the archos code is *very* close to the code size limit
21:58:14TiMiDand then a little overload with wps I think
21:58:14miner49erAh, It won't be remembered from the last time I did it will it? That's probably it...
21:58:39LinusNminer49er: echo $PATH
21:58:50_FireFly_in the mean time my patch is ready when TiMiDs gets in cvs to provide the rest for those you whant a nearly full working remote
21:59:00TiMiDLinusN: That's why I do code that could reduce size first :)
21:59:41TiMiDsince your patch is here _FireFly_ there is no need to hurry we can do it in the most convenient way
22:03:00LinusNTiMiD: gui_statusbar.c line 108
22:03:23LinusNthe comment isn't correct anymore
22:03:56TiMiDthe comment must be removed :)
22:04:19LinusNbesides, how does it know if the info has changed or not?
22:04:38LinusNah, i see now
22:04:45ep0chis it normal for rockboxui.exe to take up 100% cpu all the time?
22:04:51TiMiDLinusN: it's a dirty way
22:04:56TiMiDit was like that
22:05:03TiMiDI kept it like it was
22:05:07amiconnep0ch: If you are building with sound support, then yes
22:05:14ep0chah. ty
22:05:37LinusNTiMiD: good
22:05:39amiconnIt's somewhat annoying, on a laptop
22:07:04ep0chyeah bye bye battery
22:07:50amiconnThe annoying thing is the fan noise
22:08:12ep0chwell atleast it will keep you warm when its cold ;)
22:08:37_FireFly_nearly 100% cpu-usage != nearly 100% cpu-usage ;)
22:09:26_FireFly_if the cpu-usage where really near 100% then the system wouldn't be usable anymore
22:10:49_FireFly_a simple endless loop(which really doesn't need much cpu-power) let the cpu-usage indicate a nearly 100% cpu-usage
22:11:05 Join XavierGr [0] (
22:12:09miner49erLinusN: sorry, I had a phonecall. I don't have anything in the PATH, except the regular stuff - i've added the sh1/bin path, now reconfigureing and make and make install
22:13:12LinusNminer49er: may the source be with you
22:13:12_FireFly_i had made a simple file which i "source" to add the path to the PATH-shell-var
22:15:16TiMiDafk for 10min
22:15:27miner49erLinusN: Just have to wait another half-hour now (damn slow PIII 533 CPU)
22:15:36ep0chwhat directory does the simulator use to simulate the players own harddisk filesystem?
22:15:45_FireFly_achos in build-dir
22:15:54_FireFly_is the root
22:16:22ep0chhaha i am using ihp so i didnt think to look in there :)
22:16:44_FireFly_but make had an output which indicates this ;)
22:16:52_FireFly_after compiling
22:17:19ep0chhmm i did make && make install so probably did see
22:17:21amiconnLinusN: Hmm, the modified mersenne twister is distributed under a bsd-style license, is that a problem?
22:17:40_FireFly_no make install is needed ;)
22:18:22amiconnHmm? 'make install' is part of the simulator build process.
22:18:45ep0chheh i've compile the simulator without sound and still CPU is 100%
22:18:46_FireFly_but i haven*t used it
22:19:01miner49erLinusN: Still getting the warning: signed and unsigned type in conditional expression
22:19:29amiconnminer49er: Your native gcc is version 4.0.x ?
22:19:49_FireFly_maybe the make install creates the .rockbox dir in archos and his content
22:20:17_FireFly_ah ok
22:20:26amiconnIt's a 'make zip' followed by unzipping the contents into archos/
22:21:09 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
22:21:44miner49eramiconn: I'm not sure what my native GCC version is, how can I know?
22:21:51amiconngcc -v
22:21:53_FireFly_gcc −−version
22:22:18_FireFly_hmm crossfeed doesn't compile with the new cvs (ACC not defined)
22:22:34miner49erit says, after a load of wierd stuff, GCC version 3.3.5
22:22:58miner49erI should be using GCC 3.3.5 not 3.3.6?
22:23:03amiconnStill, I remember that compiling gcc with gcc throws some warnings
22:23:16_FireFly_ups this is only for the sim
22:23:19 Join miner [0] (
22:23:50miner49ergoing into decent IRC client now (exciting!)
22:23:56 Quit miner49er ("CGI:IRC")
22:24:16minerHah, this is (probably) better than the browser thingy!
22:24:23amiconnminer49er: For the target, you should use the versions stated in the wiki. For the sim you will use your native gcc, whatever version that is
22:24:39amiconngcc 3.3.x and above should all work
22:25:06minerwell, that sounds promising I guess.
22:25:42minerI know this isn't the place, but linux/debian really does rock!
22:25:49amiconnYou can build the cross-gcc with a different version of native gcc.
22:26:36minerThis is pretty mind-blowing stuff - building a compiler with a lower version of the same compiler.
22:27:00_FireFly_that is the only possibility
22:27:11minerI have to admit, it's pretty much over my head - I just wanna continue with rockbox coding but got fed up with XP
22:27:16_FireFly_to compile a new one with the one before the new one
22:27:53minerFireFly: yeah, of course, you couldn't start from scratch each time - that in itself is mindblowing for me
22:32:52XavierGrmuesli finally succeeded into makinf 're' a valid vocabulary word for us!
22:33:18muesli-guess not...
22:33:38muesli-dont hold your breath..its done when its done ^^
22:34:11XavierGrwell I use it sometimes and TiMiD just did, I encountered it again by other people here too... :p
22:34:59XavierGrand I fancy the 'high' remark too....
22:35:28*LinusN is reading licenses
22:35:52muesli-but be careful... "high" and "re" are not open source :D
22:35:52XavierGrdriving licences? ;p
22:37:09LinusNamiconn: i don't see much of a problem just changing the init without relicensing
22:37:22LinusNhow much different is it?
22:38:11 Part LinusN
22:38:57muesli-gtg is not bad either!
22:39:24XavierGrgtg is normal like brb.
22:39:43muesli-never seen it in 8years of irc
22:40:05XavierGr8 years of irc!!! Wow that a lot...
22:40:23muesli-yeah..complete waste of time ;)
22:40:53_FireFly_to much free time ;)
22:41:44muesli-yeah..something like this ;)
22:42:35XavierGrmuesli-: Then fill a bit of this free time by learning C and give the devs a hand! ;)
22:43:16muesli-well...i had to pass 3 tests (including one oral one) to finish damn pascal ;)
22:43:35muesli-sorting numbers and such rubbish
22:44:09XavierGrthe you can just expand your knowledge. Anyway the logic is the same, the syntax changes a little bit.
22:44:41XavierGroh I just noticed, 5 consecutive lines with';)' in the end!
22:45:13_FireFly_;) *g*
22:45:30muesli-that was a problem writing code always ended with ; but id like to write ;) ;)
22:45:51_FireFly_you could try with a define :)
22:46:32_FireFly_but this wouldn*t work
22:47:13 Quit Febs ("CGI:IRC")
22:50:13 Part ep0ch
22:51:59_FireFly_night all
22:52:27 Quit _FireFly_ ("Leaving")
22:53:40muesli-but i am lost with coding...i am not very logical
22:58:37muesli-gtg..good n8
23:01:38 Quit muesli- (Read error: 104 (Connection reset by peer))
23:07:56 Join LinusN [0] (
23:08:49amiconnLinusN: It's not only the init that is different. I'd prefer to take the new version and cut it down instead of inserting the new init into the old version
23:09:10amiconnIn fact I already did that locally - it ends up slightly more compact than the old one
23:09:40amiconnIt's definitely random - the question is whether it is better than with the old init.
23:10:36preglowwhat are you talking about?
23:10:38preglowrandom gen?
23:10:56amiconnYes. An improved init for the mersenne twister
23:11:12preglowyou sure you actually need something as random as a mersenne twister?
23:11:40amiconnRockbox already uses the mersenne twister
23:12:16 Quit matsl ("Leaving")
23:12:54LinusNthe license is gpl compatible
23:14:13preglowi know it already does, i'm just wondering if you need something so big and random as a mersenne twister
23:14:18TiMiDwhat is GPL compatible ?
23:14:39LinusNTiMiD: the license for the improved mersenne twister code
23:15:21LinusNpreglow: no, i don't think we *need* it
23:15:35TiMiDok the random generator
23:15:47amiconnMoos (or anyone else): If you wanna test the improved rng, here is a fresh build:">
23:15:53LinusNbut it sure is nice to have a decent rng
23:16:13***Saving seen data "./dancer.seen"
23:16:14amiconnOther targets on request
23:17:31LinusNamiconn: we need to add the license to the binary distribution
23:18:00Moosamiconn: sure let me just few minutes I'm eating
23:18:12amiconnThe license statement is part of the .c file
23:18:22amiconnAh, to the binary...
23:19:36LinusNno, not to the binary
23:19:51LinusNjust in the zip file
23:20:24LinusNif we had to add it to the binary, the license wouldn't have been compatible
23:22:49amiconnThis would be the first license to be added to the distribution...
23:24:08LinusNin fact, i think the GPL license should be in there as well
23:24:22BagderI agree
23:24:31Bagderthe binary doesn't say that anywhere
23:24:35LinusNhi Bagder
23:25:07*LinusN hands the licensing issues to mr License himself
23:26:31LinusNTiMiD: if i commit your patch, will you continue working with the menus and stuff?
23:28:51Mooshe is at phone
23:29:06Moosamiconn: what build want you I test?
23:29:50TiMiDLinusN: yes
23:30:05TiMiDLinusN: sry, I'm on the telephone
23:30:39Moosthe moment you waited so long :D
23:31:38amiconnLinusN: Our rand() isn't a standard rand(). Standard rand is supposed to return a number between 0 and 2^15-1
23:31:48 Quit ender` (Read error: 104 (Connection reset by peer))
23:31:52amiconnOur implementation is an lrand() except on 16 bit platforms
23:32:19Bagder"The rand() function returns a pseudo-random integer between 0 and RAND_MAX."
23:32:26minerI think i'm gonna cry. I'm still getting the error when building GCC unrecognised option -big . Please someone help me!!!!
23:32:36preglowstandard rand has no set limit
23:32:39preglowjust RAND_MAX
23:32:56amiconnHmm, I didn't find that anywhere...
23:33:10preglowit's in all the c docs i've ever seen
23:34:03LinusNTiMiD: i think your patch looks good, except for one thing
23:34:26LinusNand that is the profoundly long function names
23:34:59minerLinuxN: I have /home/mat/sh1/bin in my PATH, i have a link to newlib in my gcc-3.3.6 directory. Is there anything else I need to do? Is there anything else I coudl try, a different version of gcc perhaps (I have no idea)
23:35:13*preglow kicks his filtering routine
23:35:18LinusNminer: still no luck?
23:35:27XavierGryeah I agree. Very long names when I checked the code.
23:35:39LinusN gui_synclist_get_selected_item_position(struct gui_synclist * lists)
23:35:48TiMiDLinusN: suggest me shorter and coherent names :/
23:35:49Moosouch :-(
23:36:18LinusNcould very well be synclist_get_selected_pos()
23:36:22amiconnBagder: Indeed, a man -S 3 rand says the same
23:36:25Bagderthere should be a huge comment describing the function, the name does not have to explain it all
23:36:26minerLinusN: no lick still, no change which is odd, as the newlib link is there and it wasn't before so surely something should be different?
23:36:28LinusNor even synclist_get_selpos()
23:36:59Bagderamiconn: yes for me too, but that site is plenty good for getting info from the standards
23:37:00TiMiDI prefer gui_ prefix
23:37:07LinusNTiMiD: why?
23:37:07TiMiD(still at tel :p)
23:38:18minerLinusN: I will start again, from scratch. Is it safer to use exactly the same version of gcc?
23:38:27Moosamiconn: codec error message, no mpa...
23:38:32LinusNsame as the native gcc?
23:39:05Moosamiconn: no sound at all
23:39:24minersame as the one that came with Debian (mine is gcc 3.3.5)
23:40:22Moosamiconn: I don't understand it return me codec failure, but is it in the .rockbox dir
23:41:34amiconnHmm... It works perfectly here
23:41:47Mooslet me retry
23:43:04amiconnThe code is identical to the official builds except the rng
23:43:31Mooserrm, I missed something sorry :) my fault
23:44:59LinusNminer: which binutils version are you building?
23:45:10TiMiDLinusN: ok remove them ...
23:45:39Moosamiconn: oh yes that works just fine
23:46:12 Quit cYmen ("zZz")
23:46:28Moosjust tested in my iriver I don't have my Ondio beside me :)
23:47:43LinusNTiMiD: care to share your thoughts about why you prefer the prefix?
23:47:53LinusNjust curious
23:48:02TiMiDtel :/
23:48:44Moostelephone is a french word TiMiD :)
23:48:46preglownothing more than a general namespace thing?
23:48:54TiMiDphone :)
23:49:13preglowtelephone is a perfectly ordinary english word
23:49:44Moosyes, I never hearded it in english :)
23:54:19TiMiDfinished tel :)
23:55:05Mooscongratulations XD
23:55:11TiMiDI prefer the prefix because when you read the code you see that the fn call is related to display
23:55:48TiMiDthen you have no need to read furthermore the parameters or to try to understand what the fn does if you are not interrested in that part of the code :)
23:55:55LinusNthat should somehow be evident from the rest of the code imho
23:56:11TiMiDwhen you want to find something fast
23:56:18TiMiDyou don't read all the code
23:56:20Bagderlong names are hard and awkward to read
23:56:27Bagderand they make code look funny
23:56:35Bagdersince it must be within 80 cols
23:56:35preglowwell, i don't believe the gui_ part of the name is what matters most here, heh
23:57:19TiMiDoh :) I avoid to make fn name more than 80chr :p
23:57:37amiconnBagder: What do you think about the license problem? Should we use the new rng init code and include the license in the .zip along with the gpl, or better not use the code
23:57:51Bagderhaven't checked the license
23:58:10LinusNBagder: a bsd style license without the announce clause
23:58:13TiMiDif you don't like the names change them
23:58:22Bagderoh, that's just fine to proceed with
23:58:28amiconnThe funny thing is that imho licensing of such a small piece of code is ambiguous, neither the original nor the new code is used verbatim
23:58:28TiMiDbut I would like at least to keep some consistancy in the naming
23:58:56amiconn...they have much in common themselves, but then they are licensed differently

Previous day | Next day