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 2005-10-01

00:01:08 Quit AliasAle ("Leaving")
00:16:15 Quit ender` (Read error: 110 (Connection timed out))
00:20:04 Join [ReDeViL]|[BlEaD [0] (
00:27:06 Quit [ReDeViL]|[BlEaD (Remote closed the connection)
00:43:45 Quit webguest19 ("CGI:IRC (EOF)")
00:44:49 Nick jborn_ is now known as JoeBorn (
00:46:14 Join XavierGr [0] (
00:47:07XavierGrI need some cvs help please!
00:47:26rasherWhat's up?
00:48:37XavierGrI have a file (radio.c) which is older than current version. I have 1.76 and current is 1.77
00:48:50linuxstbJust do "cvs update"
00:48:59XavierGri use cvs and I made an update. "cvs update"
00:49:15XavierGrit tells me that radio.c is merged.
00:49:37XavierGrbut when I open the file the header is still 1.76 and the changes have not been made.
00:49:50XavierGr(Linus added 2 lines of code which I cant find)
00:51:09linuxstbIf "cvs update" tells you that the files have been merged, then they should have been. Are you 200% sure you're opening the right file?
00:51:09XavierGrany idea what going on? I could make the changes manually but...
00:51:41linuxstbDid you have the file open in your editor when doing the "cvs update" ?
00:52:16linuxstb"yes" to which question?
00:52:22XavierGrto yours
00:52:28linuxstbI asked two :)
00:53:16XavierGrI just tried with the file closed and the same happens.
00:53:27XavierGroh didnt notice yes for the 2 then/
00:54:15linuxstbI don't know what to suggest - you typed the correct CVS commands.
00:54:34XavierGrwhats the difference between cvs update and cvs up -dP?
00:55:28linuxstb-d downloads new directories, and -P deletes (prunes) empty directories.
00:55:50linuxstb(I think)
00:58:01XavierGrOkay now another question: The fm preset patch uses the /.rockbox/presets/ directory. But by default there is no such directory. Do I have to change a file to add that folder, or make the program to check for the folder then create?
00:59:24linuxstbI think you should use the tools/ script to create it.
00:59:36linuxstbI don't know whether it is necessary to double-check in your program as well.
01:00:10linuxstb(see the mkdir lines already in
01:00:52XavierGrHmm ok
01:01:53linuxstbBut thinking about it, I'm not sure if a zip file can contain an empty directory though.
01:02:25XavierGrI tried firefly's remote patch. When I plug in the remote and then scroll the presets there is a lag when I release the button (scrolling continues a little more) without the remote response is good.
01:02:46XavierGrIs that normal, does the remote code (all those ifdefs) make the code slower?
01:03:02rasherthe ifdefs are gone by the time the code gets compiled
01:03:29XavierGrso what you think about the leag with the remote?
01:03:42linuxstbBut Rockbox does more work (probably double the work) when updating two screens at the same time. I haven't tried the patch though.
01:05:18linuxstbOr did I misunderstand? Is anything being displayed on the remote?
01:05:29XavierGrstrange it will not lag when scrolling in filetree. Also the remote code will not render the fm screen. Only the buttons are active.
01:05:38XavierGrmaybe there is a conflict with my patch.
01:06:17XavierGrBecause I remember that the patch supports fm radio too.
01:07:09XavierGrAny news from Slasheri and his dircache patch?
01:08:19rasherhe put out a "final test" this evening
01:08:32XavierGrand? :x
01:14:11***Saving seen data "./dancer.seen"
01:15:58XavierGris there a special function to create folders? I should check the context menu. If someone remembers shoot.
01:16:43rashermkdir I'd assume
01:17:33linuxstbYes: mkdir("dirname",0) - 0 is the mode, which I think Rockbox ignores. It's there for POSIX compliance.
01:18:36linuxstbThere's an example in onplay.c
01:19:12preglowanyone heard anymore from the faad guy?
01:20:34linuxstbfuzzie: Are you around?
01:20:53fuzzieOh. rasher got it running on a h120, but had logic problems. I haven't had time [or been too lazy] to work on it much this week, I guess it'll be tomorrow.
01:21:11preglowlaziness is a real problem
01:21:20preglowwhat do you mean by logic problems, btw?
01:21:22fuzzie(logic problems meaning 'it failed an error check on the second packet')
01:21:37preglowwhat compiler did you end up using?
01:21:43fuzzie4.0.2 without optimisations
01:21:55preglow4.0.2 fixed all ICEs?
01:22:11rasherOn -O0 at least
01:22:19XavierGrwhats the policy when a line passes 80 columns? I mean okay I will hit enter and change line but is there any need into aligning the line in a specific way?
01:22:23rasherO2 still gave errors
01:22:25fuzzieIt might work with optimisations, I don't know if it was tested with optimisations after we worked out the problem was overflowing the stack..
01:22:30preglow-O0 doesn't exactly cut it, though
01:22:37rasherfuzzie: Maybe not
01:22:41preglowfuzzie: oh, so there's still that
01:22:45linuxstbYou should try undoing the changes I made to fix the ICEs. Maybe I added some bugs when doing that.
01:22:48preglowfuzzie: do you know how much stack it requires now?
01:22:55fuzzielinuxstb: your diff looks fine
01:23:08fuzzieI made some changes to cut down the stack usage, just haven't tried them on the hardware.
01:23:32rasherfuzzie: shoot
01:23:39preglowyou guys rocks, i would never have bothered battling ICEs and too much stack usage to get it working
01:23:42fuzzieIt's stupid stuff, allocating a bunch of 2kb buffers on the stack.
01:27:37 Quit Moos ("Glory to Rockbox")
01:34:32rasherAt least -O1 works
01:34:35*rasher tries O2
01:34:50rasherwell, "works" - it works as much as O0 does
01:34:53XavierGrwhat are you working?
01:34:57rasherwhich is to say, not a whole damn lot
01:36:14rasherXavierGr: AAC support. Or, I'm just testing.
01:36:45rasherokay, -O2 compiles and runs as well
01:36:47rasherthat's good news
01:36:59XavierGrthe iriver target is trully going to be what we call "Multi-codec Jukebox"!!
01:37:37linuxstb:) Just WMA left if fuzzie succeeds.
01:37:47preglowand good luck to whoever starts that
01:37:51preglowbest wishes, etc
01:37:54linuxstbPlus Realaudio....
01:38:07preglowno, no, you're going too far!
01:38:09linuxstbTime for some serious reverse-engineering.
01:38:16linuxstbOh, and ATRAC....
01:38:17XavierGrIs fuzzie going to add WMA support or you say this as a joke?
01:38:29preglowATRAC isn't _that_ bad, at least i know how it works
01:38:39preglowwhich is more than i can say about all the different incarnations of realaudio
01:38:43linuxstbXavierGr: No - I meant that if fuzzie succeeds with AAC, then WMA is the only major codec missing.
01:39:05linuxstbWell, ATRAC is one incarnation of realaudio - so you know one.
01:39:14fuzzieI assume someone is working on speex. :)
01:39:24XavierGrI wish Midi could be done.
01:39:27preglowhmm, yeah, but i don't think some realaudio codecs look much like aatrac
01:39:27fuzzieAnd quite a lot of the modern realaudio streams contain AAC, that's two.
01:39:39fuzzieit's the old ones which are a pain
01:39:46preglowi don't think atrac scales good to low bitrates, like some RAs do
01:40:06fuzzieis it possible to build the toolchain with C++ enabled?
01:40:14preglowbut try to avoid it offers live concerts in 352kbps realaudio ATRAC if you want some test streams....
01:40:26preglowthe older rockbox boys aren't very fond of c++, and it's best to avoid it
01:40:30fuzziejust wondering about other file formats I'd like support for
01:40:54preglowfuzzie: if you get it working, you'd need to avoid tons of the c++ bells and whisltes like exception handling and rtti
01:40:59preglowfuzzie: apart from that, it might work well
01:41:16fuzziewell, my thinking was it'd be much easier to rip that out of existing code than to rewrite it
01:41:23XavierGris there a function to check if there is a directory or should I leave it to create every time the radio starts?
01:41:28preglowand i'm all for it
01:41:35XavierGrsorry for the storm of questions.
01:42:18preglowmost codecs i've seen use very basic c++
01:42:21preglowso might work good
01:42:23linuxstbXavierGr: "opendir" I think.
01:42:33preglowjust avoid the standard libraries
01:42:53preglowif so they can be called
01:43:02preglowmostly iostream, i guess
01:44:24linuxstbDoes anyone know if it's possible to use "dd" to write directly to a disk partition in Cygwin?
01:44:27fuzziemy old wish for c++ was for modplug, but it seems DUMB is just as accurate, if a little slower
01:44:30preglowthe lack of operator "new" might also be something of an obstacle
01:45:02preglowfuzzie: well, modplug isn't very accurate, last time i checked
01:45:12fuzzieit's a lot more accurate than mikmod
01:45:14preglowmy standards might be just a little high, though
01:45:17fuzziewhich I assumed was the only alternative
01:45:24preglowi'm used to playing music with the trackers themselves
01:45:41rasherAh, faad now running with the normal stacksize
01:45:46preglowrasher: !!!
01:45:49fuzzieunfortunately, porting them isn't something I think is going to happen :-)
01:45:58fuzzierasher: oh, good, so it was those
01:46:04rasherpreglow: say that to fuzzie
01:46:11rasherI just tested the changes
01:46:15preglowfuzzie: pleeeease, all you need to have fasttracker 2 running is a pascal compiler and the source code!
01:47:01preglowi haven't touched pascal in over ten years, god be praised
01:47:19*preglow suddenly realizes he's old
01:47:23fuzzieI still do maintenance on some oold Delphi stuff.
01:47:43fuzzieThankfully any pre-Delphi pascal I was responsible for has long since died.
01:47:53preglowlast pascal i used was turbo pascal 8
01:47:56preglowehh, 7
01:48:06preglowthe first Real programming language i learnt
01:48:09*linuxstb too
01:48:37preglowonly thing i touched before that was batch files, basic and the very occasional assembly
01:49:18*rasher proddled around some outdated COMAL80 in school
01:49:27rasherlord knows what it was doing there
01:49:33*preglow pats rasher
01:49:49rasherIt was pretty much BASIC, I think
01:50:05preglowi've never heard of it before
01:50:16rasherAh, it was created in Denmark, figures.
01:50:35fuzziethe DUMB guy seems to have a 'if this doesn't produce correct output, tell me so bi can fix the bug' thing on his page
01:50:39fuzziewhich sounds hopeful.
01:50:42*preglow misses denmark and it's beer selection
01:51:45*rasher directs preglow at
01:51:59preglowrasher: i know, i just can't stand correcting myself all the time
01:52:40*linuxstb remembers that "its" is a word like "his"
01:52:41preglowi should know, half of my time is spent correcting norwegians when they use apostrophes everywhere
01:52:45rasherI don't really care, I just suddenly remembered that group
01:55:37preglowdon't know if i care much either, but with 90% of the norwegian population being technically illiterate, i tend to be annoyed
01:55:50preglowand that _is_ a real newsgroupt? :P
01:56:00rasherIt is
01:57:04rasherwith recent posts, no less
01:57:15preglowi should subscribe
01:57:57preglowfuzzie: what exactly did you do to get rid of the excessive stack usage
01:58:27fuzziemade some 20kb structures static
01:58:50preglowany idea if it can still decode realtime?
01:59:25fuzziei don't think there's any chance of it decoding realtime without optimisations, but we haven't gotten anything working past the second frame yet
02:00:38fuzzieit doesn't seem to use much of those structures for 'normal' decoding
02:00:56fuzzieI might end up having to split them, some of them on the stack (and hence in IRAM, presumably), and others not
02:01:04preglowone would think aac was designed with embedded implementations in mind
02:01:16fuzziefaad was definitely not, alas
02:01:19rashernot this implementation...
02:01:29preglowi wonder what the chances of codecs getting more than 48kb iram is :P
02:01:53preglowthat's another problem
02:02:11rasherYeah, that should be... interesting
02:02:13preglow1. i haven't got one 2. how to handle one device having more iram than another
02:05:54fuzzieit looks like they have got this running on /some/ embedded device
02:06:04preglowoh well
02:06:05fuzziebecause it has a 'circumvent memory alignment errors on ARM' comment
02:06:22preglowit didn't much look like libmad was going to run on h1x0 when we ported it either...
02:07:23linuxstbAny idea how a 75MHz ARM will compare to the 120MHz Coldfile?
02:07:48linuxstbThe iPod does have two of them of course...
02:07:56preglowi expect the coldfire to be somewhat stronger, given proper emac usage
02:08:42linuxstbBut I think it will be fine with the codecs running on the COP, and the DSP and everything else on the main CPU.
02:08:45preglowi'd expect the arm to be somewhat faster per clock, but then again, i'm as far from an expert as they come
02:10:14preglowhaving dsp on the second processor is a pretty good idea
02:10:46preglowat least porting for an arm is easier
02:10:58preglowthe accumulator unit is fully 64 bit and not hidden from gcc's view
02:11:41preglowlinuxstb: the portalplayer propaganda data actually gives me the impressions they've got some separate mac unit going, you know anything about that?=
02:12:05linuxstbNope, no idea. I haven't got that far yet.
02:14:02linuxstbSpeaking of the iPod, I still haven't got used to the fact that there is no STOP button on the iPod - all you can do is pause, or skip to the end of the current playlist.
02:14:26linuxstbUnless I just haven't found it.
02:14:39preglowno stop button i know of
02:15:03preglowbut then again, on rockbox there's not THAT much of a difference either
02:15:08preglowone takes you out of the wps, the other doesn't
02:15:17preglowit resumes anyway
02:15:52linuxstbApple's approach does make a little sense - why do you need stop/resume as well as pause/unpause ?
02:16:18preglowcurrently, stop flushed the audio buffer on rockbox, i _think_
02:16:22preglowdoesn't make much sense
02:16:26linuxstbBut it's still odd not to have a stop button.
02:16:34preglowand if that's fixed, there's not much of a difference anymore
02:23:02 Join linuxstb_ [0] (
02:28:11preglowsomeone brave should just try the helix aac decoder for fun :P
02:28:19preglowthey claim it's good on embedded cpus
02:28:38rasherYou'd just be tainted
02:28:47preglowi feel dirty for mentioning it
02:29:07fuzzieit seems like such a wasted effort
02:29:09fuzziegiven the license
02:29:15preglowyes, indeed
02:29:32fuzziei would have definitely tried it first, otherwise
02:30:03linuxstb_Are the AAC specs public?
02:30:18fuzziei believe so
02:30:28preglowi sincerely doubt it
02:30:28linuxstb_Then write your own decoder... :)
02:30:36preglowif the mpeg specs are any indication
02:30:46preglowpublic, but you've gotta pay
02:30:59fuzzie"A copy of the MPEG-4 Audio standard can be purchased from the ISO online store (search for "14496-3")."
02:31:02fuzzie*rolls eyes*
02:31:15*preglow rests his case
02:31:18linuxstb_But I'm sure there's an iso14496-3.pdf file in google somewhere....
02:31:57fuzzieyes, i'm sure it wouldn't be a big problem to obtain
02:32:41preglowno, but you'd be hard pressed to make it seem a legitimate open development process...
02:33:17linuxstb_But how much does the standard cost?
02:33:19fuzziejust throw it out onto the web as GPL anonymously and cover your tracks well. :-)
02:33:46preglowlinuxstb: it's not a prohivitely high cost, afaik
02:34:06linuxstb_391 Swiss Francs
02:34:38rasherThat's not bad
02:34:49rasherWell, it's horrible. But it's not that much.
02:34:50preglowand for those of us who have no idea what a swiss franc is worth? :)
02:35:14 Quit matsl (Remote closed the connection)
02:35:18linuxstb_About 250 EUR
02:35:20rasherWell, I'm guessing it's a bit over half a euro
02:35:42preglowquite a sum for a bunch of dead tree
02:37:39preglowsomeone, make a wav writer!
02:38:52 Quit linuxstb (Read error: 110 (Connection timed out))
02:39:01 Quit dpassen1 ()
02:39:11linuxstb_Maybe we need to try and think of an alternative way to measure codec speed.
02:40:02*fuzzie was wondering how codec speed would be measured
02:40:05preglowproblem is it wouldnt be very hard
02:40:27preglowit's just that i can't be bothered to learn how all that playback code works
02:40:36preglowi don't have much time to code as it is
02:43:11fuzziethe AAC spec claims that 10 MIPS is sufficient for decoding, interestingly
02:43:18preglowfuzzie: in the olden days we had xxx2wav plugins that stressed the codecs as fast they could go, then measured how fast they ran compared to realtime
02:43:49preglowfuzzie: yes, and freescale has got a decoder that can deocde mp3s at about 20 mhz, we just can't be bothered to write an mp3 decoder in pure asm
02:44:17fuzziejust nice to know it's *possible* :)
02:44:21preglowhaha, sure
02:44:40preglowalmost anything is possible in 140mhz if you're persistent enough
02:47:02rasherIf you hadn't figured it out, 14496 pdf is an incredibly effective term to google for
02:47:13rasher(and if you're interested, of course)
02:50:52XavierGrwhat MIPS stands for? Is it a CPU power metric system?
02:51:00preglowmillion instructions per second
02:51:35preglowmilion, even
02:51:46preglowno, million :/
02:52:04preglowsuffice it to say 10 mips isn't much for us
02:52:26linuxstb_Does it mention the RAM requirements?
02:53:00fuzzieabout 9kb 'words'
02:53:11fuzzieso, presumably 36kb
02:54:10linuxstb_So the faad authors didn't work very hard then.
02:54:31preglowthey usually don't
02:54:40preglowlibmad is a pretty well-behaved codec in this regard
02:55:00fuzziewell, i assume the target platform for faad is a machine capable of running Nero :)
02:55:22linuxstb_With a nice big About box...
02:55:32fuzzieohh yes. :)
02:57:51 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
03:00:12fuzzielinuxstb_: is there some reason for the '#warning BIG EENDIAN' in config.h?
03:00:46linuxstb_Just a test I think - you can delete it.
03:01:05fuzzieokay, just wanting to know if you knew anything about it :)
03:01:16linuxstb_That sounds like my kind of thing.
03:03:43XavierGrDo you think that when someone loads a preset file it would be good to change the frequency to the first preset?
03:05:49rasherhrm, I'd say leave tuning changes to the user
03:06:00rasherhe may already have tuned into something
03:08:09XavierGryeah I agree.
03:14:15***Saving seen data "./dancer.seen"
03:16:22 Quit JoeBorn ("")
03:25:44 Quit preglow ("leaving")
03:33:38 Quit RotAtoR ()
03:49:36XavierGrOk the patch for fm preset saving/loading is finally done. Now only testing remains.
03:57:11 Quit rasher ("leaving")
04:05:18 Join QT_ [0] (i=as@madwifi/users/area51)
04:15:50 Join Peter200lx [0] (
04:16:43Peter200lxHow do I use the new ID3 Database tool on Rockbox 2.5?
04:17:08Peter200lxI've looked but havn't seen any documentation on how it's used
04:17:14 Quit QT (Read error: 110 (Connection timed out))
04:18:18 Quit XavierGr ()
04:18:20 Quit tvelocity ("Leaving")
04:21:34Peter200lxfinnaly found it
04:21:46 Quit Peter200lx ("CGI:IRC")
04:28:55 Join bagawk_ [0] (i=1000@unaffiliated/bagawk)
05:14:16***Saving seen data "./dancer.seen"
05:27:09 Quit bagawk_ (Read error: 110 (Connection timed out))
05:29:46 Join ashridah [0] (
06:20:59 Join pengo [0] (n=xtofu@
06:50:02 Quit ze ("new graphics card and kernel")
06:59:46 Join amiconn_ [0] (
07:05:35 Join ze [0] (
07:14:21***Saving seen data "./dancer.seen"
07:17:55 Quit amiconn (Read error: 110 (Connection timed out))
07:17:55 Nick amiconn_ is now known as amiconn (
07:20:54 Join banan_ [0] (
07:21:12 Join thegeek_ [0] (
07:21:29 Quit thegeek (Read error: 104 (Connection reset by peer))
07:38:16 Quit merbanan (Read error: 110 (Connection timed out))
07:58:06 Join random_man [0] (
07:58:33random_manhello i need some help on patching a file
08:08:57random_man this is my error
08:38:51Slasherirandom_man: at first you have to be in the rockbox directory where the source code is
08:39:02Slasheriif p0 doesn't do anything, try p1
08:41:28 Join Colddy [0] (
08:41:53random_manthought i had to be in where the patch was ?
08:42:49Slasherino, that doesn't matter where the patch is
08:43:12Slasheribut you have to be in the target directory where you want to apply the patch
08:43:31random_manok well this is in the guide
08:43:31random_man Back in the DevKit window, go to the directory containing the patch
08:44:21random_man Now you are in the sim directory. Create the make file as usual using rbconf but select (S)imulator build instead of (N)normal build, and then (w)in32
08:44:22random_manok now im trying to make simulator and i am this far. how do i do this part ?
08:46:00random_manhow do i type that part in
08:46:00 Join Vld0man [0] (
08:47:03Slasherirandom_man: so you want now to apply the patch?
08:47:26Slashericd .. (you don't want to be in the sim directory but the rockbox root directory where everything (firmware/, apps/) is)
08:47:53Slasherithen you can type it there (cat something.patch | patch -p0 (or -p1 if -p0 doesn't work at all))
08:48:27random_manso i want to be where the source code is right?
08:49:20random_manwhich should be in rockbox-bleeding folder right?
08:49:48Slasherihmm, i recommend using cvs to check out the latest source code
08:50:42Slashericvs login
08:50:52Slasheri(press just enter if it asks for a password)
08:50:59Slashericvs -z3 co rockbox
08:51:24Slasheriand when the command finishes downloading the files, you will se a new directory called rockbox there
08:51:43Slasherioh.. run cvs -z3 co rockbox-devel instead of the previous command i told you
08:52:24random_manok doing that now
08:54:41Slasheriamiconn: the newest code causes iriver build to fail when developer and logf options are selected with configure: gray_core.c:569: error: can't find a register in class `ADDR_REGS' while reloading `asm'
08:54:58random_manrockbox folder will be in C:\Rockbox\home\guest right?
08:55:09 Quit ghode|afk (Read error: 104 (Connection reset by peer))
08:55:16Slasheriit will be in the same folder where you typed the cvs command
08:55:21Slasherijust cd rockbox is enough after that
08:55:31Slasheripwd will show what folder it is
09:01:43 Quit Vladoman (Read error: 110 (Connection timed out))
09:14:22***Saving seen data "./dancer.seen"
09:27:45amiconnSlasheri: Hrmph.
09:28:54 Nick Vld0man is now known as Vladoman (
09:49:19amiconnSlasheri: This routine is really short on registers, and the developer builds don't use -fomit-frame-pointer
09:49:50amiconn...meaning that the compiler needs an addtional address register besides the stack pointer
09:50:06amiconn...and the routine doesn't leave such additional register
09:50:12Slasheriah, interesting
09:50:45amiconnIf you add -fomit-frame-pointer to the GCCOPTS lone in the Makefile, it should work
09:51:18amiconnI found a register for possible re-use, so it should be fixable.
09:51:31amiconnCan't do that atm because I have more changes pending
09:51:39amiconn(screendump for grayscale)
09:53:30 Quit Lynx_ (Read error: 104 (Connection reset by peer))
09:53:30 Quit elinenbe (Read error: 104 (Connection reset by peer))
09:53:30 Quit Maxime (Read error: 104 (Connection reset by peer))
09:53:34 Join Lynx_ [0] (
09:53:44 Join Mxm`Pas`Bien [0] (
09:53:44Slasheribut i will add that option, thanks :) hmm, is there any reason for iriver targets not to always omit frame pointers?
09:53:45 Join BirdFish[alt] [0] (n=bradbox8@
09:53:45amiconnUsing frame pointers might make debugging easier. No reason other than that, methinks
09:53:51Slasheriyep.. but iriver doesn't have a gdb stuff or anything like that
09:53:58 Join elinenbe [0] (
09:54:12amiconnSlasheri: Yes, nobody bothered to implement one
09:55:27amiconn...but iiuc it is possible to debug with bdm
09:55:27Slasherihmm, ok
09:55:40amiconnFor logf() builds it's definitely unnecessary to have frame pointers
09:56:02amiconnI had to disable that too when doing some logf() work on archos
09:56:23amiconnThe frame pointers increased the code size beyond the 200KB barrier...
10:09:56 Quit BirdFish (Read error: 113 (No route to host))
10:24:31 Join ender` [0] (i=ychat@
10:54:07 Join ghode|afk [0] (
10:54:59amiconnSlasheri: I fixed the problem in my local copy
11:14:26***Saving seen data "./dancer.seen"
11:25:39 Part Vladoman ("Verlassend")
11:38:48 Join rasher [0] (
11:46:34 Join Lost-ash [0] (
11:49:12 Quit ashridah (Read error: 110 (Connection timed out))
11:51:25 Join ep0ch [0] (
11:53:19ep0chI've written a WPS patch that displays the peak meter upper/lower values, it's not very elegant (i'm not a C programmer), so if someone could take a look at it and provide me with feedback it would be much appreciated.
11:53:25ep0chhere is the patch:">
11:57:15ep0choh and my wps file:">
12:05:30 Quit Ismo_ (Read error: 110 (Connection timed out))
12:17:49ghode|afk - has anyone been following this blog?
12:18:50rasherIt seems odd
12:18:53ep0chheh yeah, i posted a comment saying come and join us :)
12:18:54rasherthe whole thing
12:19:23Bagderhe obviously already knows about us as he's basig his work on our bootloader ;-)
12:19:44ep0chwhy do you think it's odd?
12:20:11rasherIt sounds like he's not using much of the information available from Rockbox
12:20:36Bagderyes, it seems he likes to start from scratch while still using code from us
12:20:37rasher <−− ghey
12:20:49BagderI find that very odd too
12:20:58pengoit's odd because someone trying to be that hardcore should have gotten the ihp120's silver edition rather than the black
12:20:59rasheralso, that url was a mis-send.
12:21:33*Bagder has a black h140 ;-)
12:22:47*amiconn just committed screendump for the grayscale lib
12:24:02 Join Vladoman [0] (
12:25:58 Join StrathAFK [0] (
12:27:54Bagdernice amiconn !
12:28:19amiconnNow, who wants to be the first in documenting the grayscale plugins with screenshots? ;)
12:28:50amiconnThe screendump now hooks into the core, so grayscale screendump works the same way as ordinary screendump
12:29:08amiconn...just the BMPs are a little larger
12:32:36rasherBagder: had any time to look at the bleeding edge builds being.. "lacking"?
12:35:05ep0chre blogspot: well it's just a hobby for him on how to learn, so it's not really that odd, using rockbox would be cheating ;)
12:35:30Bagderno, not yet
12:35:40Bagderrasher: you know any specific files that are missing?
12:36:15rasherBagder: at least all of apps/codecs/libmad/* EXCEPT for the all-caps README, COPYING and two others
12:36:22amiconnBagder: The build time estimation is off, as always...
12:36:40Bagder"as always" ?
12:37:11amiconnWhenever I commit something, it takes longer than the estimation says
12:38:07Slasherithat's true.. maybe the estimation timer should save the previous compilation and and use that for the estimation
12:38:17Bagderit does
12:38:21Slasherihmm, weird
12:38:23Bagderor rather, it saves the 4 latest
12:39:02Slasherimaybe there is a bug in the algorithm
12:39:11rasherMaybe it's just a clock-issue?
12:39:27Bagderits not my #1 issue
12:39:29rasherMake it say "in x minutes" instead?
12:39:49amiconnEeek, I broke the player builds
12:40:04Bagderrasher: then it would have to repeatedly update it until done
12:40:30rasheroh, of course
12:40:36rasherdidn't think of that
12:41:10 Quit Strath (Read error: 110 (Connection timed out))
12:41:27ep0chif it want comments about a patch is it best i post to the mailing list?
12:43:08Bagderthat, or in the patch tracker
12:43:27ep0chi dont think its ready for the patch tracker :)
12:43:54rasherit'll get ready in the patch tracker
12:43:59ep0choh ok
12:44:14rasheryou can update the patch there
12:44:28rasherit doesn't have to be 100% first-shot
12:44:37ep0chcan other people update it?
12:44:48rasheronly admins
12:44:54ep0chthat'll do
12:48:12BagderI think the reason for the low estimate is the seconds being chopped off
12:48:32Bagderand not rounded to nearest minute
12:49:32BagderI added seconds now to see if it this is it
12:49:32amiconnHmm. This time the estimate was ~13 minutes, but building took 20 minutes
12:49:36 Join Paul_The_Nerd [0] (
12:49:53Bagderwell, there's of course varying load on the server
12:49:55amiconnBagder: What's the current server time?
12:50:11BagderSat Oct 1 12:50:06 CEST 2005
12:50:31rasherlooks accurate to me.
12:50:34amiconn2 minutes late...
12:50:54amiconnAh, ignore me
12:51:31Bagderexpected build time is currently 842 seconds
12:52:39Bagderbut the missing files in the source package is very strange
12:52:48Bagderthe release script works fine when run stand-alone
12:52:51amiconnMy laptop clock was off by 2 minutes for whatever reason. It's configured to synchronise from the PTB, obviously it didn't work for a few times...
12:52:51rasherThe dailies are correct..
12:53:11rasherat least, the one yesterday was
12:53:22Bagderyes, must be something specific about this particular dir
12:53:38Bagderor something
12:59:46rasher fifteen seconds? I think not.
13:00:11BagderI think the recent uploaded movie proves that pretty good
13:00:23rasherIt's more like a second
13:00:31ep0chnah 3-4
13:00:40*Bagder runs off
13:00:46rasherguess I'll try and time it
13:02:35 Join muesli- [0] (
13:03:08rasher4 seconds, it seems.
13:03:16rasherfor the largest dir I have.
13:06:58 Join webguest95 [0] (
13:10:27ep0chi think its more down to disk spinup and seek time than dir size
13:11:25ep0chwhich is why i'd like a not so memory hungy dyamic cache
13:12:50ep0chjust reserve about 8k to cache where you've been and you get best of both worlds
13:13:00muesli-amiconn will there be grayscale support for the remote as well?
13:13:33rasheroh wait, don't mind me. I don't know.
13:13:45 Quit random_man ("CGI:IRC (Ping timeout)")
13:13:56rasherep0ch: but then you'd still have to cache quite often
13:14:30***Saving seen data "./dancer.seen"
13:14:38rasherI don't see how it'd work
13:14:45Slasheriep0ch: that would need the disk constantly spinning while browsing -> it will be battery hungry then
13:14:46rasherif I go browsing artists
13:15:00amiconnmuesli-: It should be technically possible, but I don't think it would make sense
13:15:09rasherI'd constantly hit uncached dirs
13:15:09ep0chyes but from a user experience point of view it will be quicker
13:15:27rasherBut pointless
13:15:31ep0chand regarding the battery, the disk will have to spin up anyway as the user is most likely going to select a new tune
13:15:32muesli-amiconn we would love it ;)
13:15:42amiconnWhat for?
13:15:54rashersince having everything cached isn't really a problem, memory-wise
13:15:59rasherep0ch: not really
13:16:06amiconnThe remote lcd is so tiny, especially compared to the main lcd
13:16:11muesli-its just eye candy
13:16:23rasherep0ch: building a playlist while listening to one song wouldn't require any disk spinup
13:16:39ep0chwhat about if you need to support 200 gig hard disks in the future, you dont wanna cache the whole lot
13:16:45ep0chrasher: true
13:17:34rasherThen maybe a limited cache will be implemented, but I'm guessing RAM would be increased as well in that case, I see no reason to do it for the irivers
13:18:13amiconnrasher: In fact a lazy cache in the browser can coexist with a full cache like Slasheri's
13:18:25Slasheri200 GB -> could be something like 2 MB cache -> in theory no problem with iriver
13:18:34Slasheriamiconn: yes
13:18:40amiconnThen you'd get the benefit of the lazy cache without needing additional ram, and if you want you can enable the full cache
13:19:24amiconnSlasheri: I think the cache ram depends on the number of files, not on the disk size
13:19:26 Join XavierGr [0] (
13:19:37Slasheriin my opinion the iriver ram is so large that reducing it a few megabytes shouldn't affect much (if at all) battery life
13:19:57amiconnI think I would get into trouble as soon as someone gets 'dumb' up & running
13:20:10amiconn>20000 mods...
13:20:11rasherOr sid.. :-O
13:20:12XavierGrare you discussing about the dircache?
13:20:35Slasheriamiconn: of course, but i just made some estimeta at the 200 GB disk would have 10x much files 20 GB disk have.. But true, i think so large disk would get filled with bigger (> 1 GB files)..
13:21:31SlasheriXavierGr: at the moment, yes :)
13:22:15XavierGrGreat! How you patch is going?
13:22:37SlasheriXavierGr: should be ready for commit. I will just give linus a try before that
13:22:39ep0chheh sid collection is currently at over 30,000
13:23:19Slasheriwe could have some compression for the cache in future so having a large sid collection may work also
13:23:37*XavierGr goes to read the logs
13:23:49rasherSlasheri: are you caching full paths?
13:23:55rasheror "these files are in this dir" things
13:24:01Slasherirasher: no, only the last filename part
13:24:06ep0chand what gets cached? all files or just supported?
13:24:14Slasheriit's a dynamic linked list
13:24:24Slasheriall files
13:24:49ep0chso i could put the sid collection on when i have time and see how big the cache is :s
13:25:07ep0chi think my ihp will die again if i do though
13:25:16Paul_The_NerdDid you eventually decide just to have it build every boot?
13:25:58SlasheriPaul_The_Nerd: yes, but the building is fast and transparent (it does not affect the bootup time)
13:26:17XavierGrHow fast?
13:26:25SlasheriXavierGr: just try out =)
13:26:26ghode|afkis there a build with dirache ver 2 floating around?
13:26:39Slasheriit depends how much there are files on the disk
13:26:44ep0chfor me about 4 seconds with a full 20 gb
13:26:50ep0chnot done a file count
13:27:21ep0chmake that a full 18.5 gb ;)
13:27:23Paul_The_NerdI honestly wouldn't have cared whether it was transparent anyway, as long as it was faster than iRiver, and whatever behaviour it had was consistent.
13:31:17 Quit ender` (Read error: 104 (Connection reset by peer))
13:31:19 Part ep0ch
13:34:22 Join DangerousDan [0] (
13:37:21ghode|afk - nice vid from MR
13:37:49XavierGrSlasheri: Where is your latest patch for testing?
13:37:55muesli-how big in size?
13:38:23XavierGrand anyone knows a site to quickly upload a small file?
13:38:44 Nick Lost-ash is now known as ashridah (
13:39:24t0masor if it's small enough... attach it to the rockbox wiki?
13:40:17XavierGryou send it is for sending not hosting
13:41:21amiconnghode|afk: Ooold bootloader
13:41:54XavierGrCtrl-A select all and save to a txt file.
13:42:01ghode|afknot my video :P#
13:42:06XavierGrPatch for radio preset loading/saving
13:42:15XavierGrPlease test for feedback
13:42:48XavierGr(Push Ctrl-A in the bottom window to catch only the code)
13:43:31XavierGroops forget what i said there is a download button in the upper left corner which let you save it.
13:49:01 Join linuxstb__ [0] (
13:51:16XavierGryes I searched and found it already.
13:51:23Slasheriah, ok :)
13:51:27*XavierGr awaits for commitment ASAP
13:51:35Slasherithat should be soon
13:52:01XavierGrIf you have free time test my patch too.
13:54:10 Quit rasher (Remote closed the connection)
13:55:12SlasheriXavierGr: ok, i will try soon :)
13:56:55XavierGrdid you removed the option to choose between dircache?
13:57:42XavierGrno forget it it seems that the patch didn't work for me I will try again.
13:58:05 Quit linuxstb_ (Read error: 110 (Connection timed out))
14:05:06XavierGrSlasheri about the warning, when you choose to enable dircache. How much time did you typed to remain on screen?
14:05:42XavierGrBecause I think it is too much to have it so lon into screen.
14:06:07XavierGrBetter make it HZ or if you find it too little HZ*2
14:07:04XavierGrMy god you have set it to 4 seconds!! I think 2 is most suitable.
14:07:28 Join _FireFly_ [0] (
14:08:11SlasheriXavierGr: Hmm, i will change it to 2 :)
14:08:36 Join preglow [0] (
14:10:48XavierGrHi preglow, hi _FireFly_
14:12:53 Nick _FireFly_ is now known as firefly (
14:13:01 Join muesli-_ [0] (
14:13:36 Nick firefly is now known as _FireFly_ (
14:15:05 Part _FireFly_ ("Leaving")
14:15:13 Join _FireFly_ [0] (
14:18:13XavierGrSlasheri: I remember changing all the warning in jpeg viewer to HZ/2 from HZ*2! It was so annoying to wait 2 seconds only to see that the file you opened is incompatible.
14:24:20 Quit muesli- (Read error: 110 (Connection timed out))
14:25:07 Nick linuxstb__ is now known as linuxstb (
14:31:33solexx_Would it (theoretically) be possible to use grayscale graphics in the wps?
14:32:18linuxstbsolexx_: I don't see why not - the LCD is running in greyscale mode (4 shades).
14:33:11Paul_The_NerdOoh... grayscale text could look really nice.
14:33:45linuxstbIf that was added, it would be nice to start thinking about colour support (for the H300).
14:33:49Paul_The_NerdRather... I think it would.
14:34:12Paul_The_NerdY'know... Track,title, album in black. Playlist name, and position, and other "less important information" in the darker gray.
14:34:35Paul_The_NerdAt least, that'd be *my* wps.
14:34:51 Join ender` [0] (i=ychat@
14:35:07solexx_how many shades are there (without dithering or sth like that)?
14:35:24Paul_The_NerdThere's black, white, and two grays natively if I recall.
14:35:58solexx_that's great. but you can't use 2bit bitmaps now, correct?
14:36:09Paul_The_NerdI don't believe so.
14:36:22*Paul_The_Nerd wouldn't know for sure.
14:37:34linuxstbLooking at the WPS code , it seems to use the lcd_mono_bitmap function to draw images. So the answer would appear to be no.
14:38:39solexx_what a waste!
14:38:46Paul_The_NerdMy big gripe with the WPS is actually a really minor irritation.
14:39:20Paul_The_NerdDrives me nuts that when the text scrolls, it doesn't clip where the string initially begins. So you end up with the first character wandering to the left of where you put it.
14:39:54solexx_never tried that
14:40:13solexx_my wps is a little bit broken atm anyway
14:40:19Paul_The_NerdI'm using a modification of the JamesGao wps from... well... somewhere
14:41:00Paul_The_NerdAnyway, it has a nice CD icon and then "album" and a music note and then "track" and a head and then "artist" (one line for each tag) and the text wanders *over* the icons. Heh.
14:41:10Paul_The_NerdThe icons being part of one large background .bmp
14:42:44solexx_cool, rasher has a backup of the wps gallery with my old wps!
14:42:45_FireFly_currently a line scrolls fully afaik
14:42:52_FireFly_in wps
14:42:56Paul_The_NerdYeah, that's what I figured.
14:43:16Paul_The_NerdIt'd be nice if elements scrolled within their own bounds, but that'd require you to be able to set a width.
14:43:17 Join arkascha [0] (
14:43:33 Quit XavierGr ()
14:43:42Paul_The_NerdI mean, what do you do if you put artist and album on a line, and the one on the left needs to scroll but the one on the right wouldn't?
14:44:00Paul_The_NerdYou'd have to be able to set an arbitrary number of characters as a length of space reserved for each thing.
14:44:07Paul_The_NerdIt'd get complex.
14:44:28Paul_The_NerdAnd as I said "minor irritation"
14:53:01solexx_did anybody ever have strange characters at the end of line in the wps?
14:56:20solexx_how they show up depends on the font
14:56:39solexx_with helvetica, i get these nice squares
14:56:40Paul_The_NerdI have not
14:56:54Paul_The_NerdThough I did have an issue where I had to insert additional endlines.
14:56:59Paul_The_NerdIt seemed to ignore at least one of them
15:02:28 Join actionshrimp [0] (
15:07:43solexx_does the wps have to use dos or unix newlines?
15:09:33linuxstbI think either are accepted.
15:13:26amiconnPaul_The_Nerd: Setting an element width in characters would indeed be complex, and not helpful at all, as we support proportional fonts
15:13:43amiconnWe need pixel positioning, and scrolling boxes instead of scrolling lines
15:14:13Paul_The_NerdThat's true.
15:14:32***Saving seen data "./dancer.seen"
15:14:34Paul_The_NerdBlasted proportional fonts, always making things complicated.
15:15:00amiconnNah, I'd rather live without scrolling boxes than without proportional fonts
15:15:49 Quit muesli-_ (Read error: 110 (Connection timed out))
15:15:55Paul_The_NerdIn something where the only real means of positional formatting are endlines and spaces, proportional width actually seems kinda out of place to me at least.
15:16:13amiconnI have some ideas regarding scrolling boxes (and font colours/ styles), but it's not exactly top prio for me
15:16:27amiconn(but certainly higher than remote support :) )
15:17:05Paul_The_NerdI think my biggest wish list for WPS would be definable fonts.
15:17:55linuxstbamiconn: What are your thoughts about colour support generally? Are you thinking that we should have a "low-depth" (e.g. 4bpp or 8bpp) colour mode for general Rockbox use, as well as a full-colour mode?
15:17:56amiconnDifferent fonts will definitely complicate things, but pixel-wise positioning will make things easier
15:18:08amiconnlinuxstb: No
15:18:24*linuxstb is planning to work on the iPod LCD driver this weekend
15:18:35Paul_The_NerdI wouldn't even want multiple fonts per WPS, just the ability to set the font within it. Heh.
15:18:43 Nick StrathAFK is now known as Strath (
15:18:54linuxstbamiconn: Sorry, I thought you suggested that a few days ago. I must have misunderstood.
15:18:58amiconnPalette modes tend to complicate things, so I'd rather use full-colour all the time
15:19:08amiconnlinuxstb: That was Linus iirc
15:19:21linuxstbOK, sorry for the confusion.
15:19:30amiconnFull colour means 16 bit for ipod, correct?
15:19:38linuxstbYes - RGB565
15:19:59linuxstbData is written in 32-bit words for two adjacent (in x-direction) pixels.
15:20:10amiconnThat's good
15:20:36linuxstbI've no idea how the 18-bit H300 LCD will operate. That gives me a headache thinking about it.
15:20:41amiconnFramebuffer layout will be simple: 2 bytes/pixel, natural order
15:20:55amiconn(ipod, and iAudio M5 - 16bit)
15:21:20amiconnFor H3x0, we could either limit ourselves to 16 bit, then it becomes the same as ipod/iAudio
15:21:32amiconnFor 18bit we'll need 3 bytes per pixel
15:21:46amiconnSomewhat odd, but not too bad
15:23:17solexx_hm. these characters I see are really strange
15:23:22amiconnSome things in lcd.h and other places are already somewhat prepared for colour support
15:23:30Paul_The_NerdIs there a noticeable difference in 16 and 18 bit?
15:23:42solexx_the viewer plugin doesn't display them, but when i enable the wps i can see them for a short time
15:23:50solexx_and they end up in the wps
15:24:09solexx_i tried to use a hex editor to spot them, but i can't find them
15:24:33Paul_The_NerdWhat kinds of lines are they on?
15:24:37Paul_The_NerdLike, data taken from tags?
15:24:46solexx_an image declaration
15:24:58solexx_the first one on the first line
15:25:06solexx_I already tried to retype it
15:25:15pengo16 bit = 65536 colours, 18 bit = 262144
15:25:38solexx_it looks like "%x|a|gfx/jrschulz.bmp|1|1|" and then two boxes behind it
15:26:18Paul_The_NerdPengo: I'm aware of the mathematical difference. I rather meant that the loss in color depth, would it be particularly noticeable?
15:26:23solexx_another font show something which looks like a "V" and a "T"
15:26:30solexx_in one character
15:26:46Paul_The_NerdSo, wait, it doesn't show the image, but instead shows the actual declaration line?
15:26:49pengoPaul_The_Nerd, for some things, i'd say so
15:26:50 Quit arkascha (Read error: 104 (Connection reset by peer))
15:26:57Paul_The_NerdSpecifically image viewing?
15:27:09pengoPaul_The_Nerd, only for image viewing i'd imagine
15:27:23solexx_the image is displayed as it should
15:27:43solexx_I only see the declaration line and the boxes when selecting this wps
15:27:50solexx_(it is displayed for a second or two)
15:27:58Paul_The_NerdAnd that's when you see the odd characters?
15:28:08solexx_yep. and then in the real wps
15:28:15solexx_first line, last two characters
15:28:23Paul_The_NerdWhat else is on the first line?
15:28:32amiconnsolexx_: VT as one character sounds like a vertical tabulator
15:28:37amiconn(control char)
15:29:06solexx_amiconn: thanks for the hint~
15:30:07amiconnlinuxstb: You've got an ipod photo, correct?
15:30:14 Join arkascha [0] (
15:31:46linuxstbI think we only need four iPod targets (plus one more if the Nano gets support) to cover all the iPod variations. Less than the 6 Archos targets :)
15:31:59 Quit arkascha (Remote closed the connection)
15:32:32solexx_my hex table tells me vertical tab is 0x0b
15:32:43solexx_but it's not in the wps
15:33:01pengostupid question.. what's wps?
15:34:37Paul_The_Nerdsolexx, do you have the ability to post the WPS and related images somewhere?
15:35:46solexx_yep, wait aminute
15:38:18solexx_the strange thing is, I copy-pasted this wps from rasher's versin of the old WpsGallery
15:39:38solexx_maybe the jrschulz.bmp is broken, since i had its %xl line at the top
15:39:46solexx_then deleted it and retyped it at the end
15:39:59solexx_but the character(s) are still at the end of this line
15:41:37Paul_The_NerdWhat font should I be using here?
15:42:11solexx_Erm, I am using some font from rasher's collection
15:42:38solexx_it's helvR12-L1
15:43:01solexx_should i out it online, too?
15:43:30solexx_and i was wrong: the characters are not at the line with jrschulz.bmp anymore
15:43:43solexx_they are in the new first line (with battery.bmp)
15:44:50Paul_The_NerdI found it at rasher's site
15:45:06Paul_The_NerdOkay, I'm seeing two dotted boxes after the battery icon and %... is that it?
15:46:21solexx_I have the battery icon, then battery percentage followed by a "%" and then the boxes
15:46:29solexx_so yes
15:46:58Paul_The_NerdIt... almost seems like a font issue.
15:47:13Paul_The_NerdI've tried it with some of the original Rockbox fonts, and it displays fine.
15:47:35solexx_not really. rockbox-default shows spaces at the end
15:48:57solexx_the same with nimbus and chicago
15:48:58Paul_The_NerdYou mean whitespace after the battery%?
15:49:41solexx_the text should be right-aligned
15:50:11solexx_so some characters are still there, it's only the font determining how they are displayed
15:50:15 Join XavierGr [0] (
15:50:26HCli need to update my rockbox badly..
15:51:27HClcause i'm horridly out of date :p
15:52:28Paul_The_NerdLemme poke around a minute
15:53:06XavierGrHCl: Which build you have?
15:58:46 Quit _FireFly_ ("Leaving")
15:59:24Paul_The_NerdThere's something odd about the WPS parser.
15:59:38Paul_The_NerdEven if I comment out every line, it doesn't show a blank screen.
16:00:10solexx_what does it show?
16:00:25Paul_The_NerdIt shows the battery icon and the volume icon.
16:00:28Paul_The_Nerdtwo images
16:01:10XavierGrFrom the status bar or the wps?
16:01:16Paul_The_NerdHCL, I think you're the one who mentioned that the Java and the .pl MakeDBs had somewhat different outputs... which would you suggest using?
16:01:23Paul_The_NerdXavierGr: From the WPS
16:01:34 Quit pengo ("Leaving")
16:02:40 Nick solexx_ is now known as solexx (
16:05:41HClPaul_The_Nerd: java, definately
16:05:59HClthe perl version has glitches that causes tags to be read wrongly
16:06:08HClending up in yPi entries
16:06:08Paul_The_NerdI had some problems with the database, but I think it may have been my build.
16:06:13HCland stuff like that
16:06:30Paul_The_NerdI'm checking again now, but it generated some interesting stuff... duplicates of folders, completely ignoring tagged MP3s, etc.
16:06:33HClXavierGr: i dunno, but i know my bootloader still boots the original firmware with the remote
16:06:42Paul_The_NerdLike... for one folder, with 1303 files in it... it found 1.
16:06:50HClperl or java?
16:06:58Paul_The_NerdI had problems with both of them
16:07:09Paul_The_NerdWhich is what lead me to believe something in my build was screwing it up
16:07:14Paul_The_NerdSo I'm trying a clean build right now.
16:07:16HClwell they both have glitches
16:07:26Paul_The_NerdThis wasn't "glitches"
16:07:32Paul_The_NerdThis was "an unusable mangling."
16:07:33HClbut i believe the java one has a much better tag reading backend
16:07:47Paul_The_NerdThe perl one gave me under any category a list of single character entries, and underneath those, nothing.
16:08:13HClthe java one only has trouble getting the offset of the music data
16:08:13Paul_The_NerdUn momento. The scan finished.
16:08:22HClfor the checksum
16:10:37Paul_The_NerdI wonder how the files are tagged.
16:10:57Paul_The_NerdIt seems to have Ape and ID3 data on it, looking back at one that's definitely missing from the Artists list.
16:11:18Paul_The_NerdSo the data may be split between them, and I assume it just does ID3 on Mp3 so far?
16:11:38 Quit ashridah ("Leaving")
16:13:15Paul_The_NerdHCl: I think I can definitely say the TagDB does not get along with my player on a clean build.
16:13:36Paul_The_NerdIt froze up on me while browsing. Then, on selection of a song, it got stuck HD reading. =/
16:14:34HClwith the java version?
16:14:38HClthe latest one too?
16:14:56Paul_The_NerdWell... I downloaded the generator yesterday
16:15:14Paul_The_NerdAnd the very latest "bleeding edge"
16:15:53HClthe songdb.jar on my ftp should be good
16:16:22HClif that one doesnt work properly your files must be pretty messed o-o
16:16:55 Join muesli- [0] (
16:17:40Paul_The_NerdOkay, lemme re-run
16:17:42linuxstbHCl: What file formats does SongDB.jar understand?
16:17:57HClmp3 and ogg
16:18:03 Join _FireFly_ [0] (
16:18:08linuxstbWhat does it do if it finds other types?
16:18:08HClmaybe even flac, but it doesn't know the extension flac
16:18:13HCland maybe also asf
16:19:10HCllinuxstb: it simply ignores them till someone adds them to the list of extensions
16:19:18HClit only checks files with extensions that it knows
16:19:23Paul_The_NerdIs there a good, say, tag management tool out there?
16:19:26HClit can be easily adapted
16:19:37HCli have one, but i dont remember the name :/
16:19:40Paul_The_NerdSomething that'd either readily let me merge data in tags, or perhaps just convert all my MP3s to one tag format?
16:20:00HClif you can wait till tonight
16:20:05HClthen i'll be able to check
16:20:17Paul_The_NerdThough I didn't tag any of these files myself, except the added replaygain data.
16:20:28Paul_The_NerdAnd OCRemix is supposed to require all the data be in ID3...
16:20:59NiliscoFoobar has some great tagging tools.
16:22:10Paul_The_NerdSame thing happened HCl
16:22:53 Join Maxime [0] (
16:22:54Paul_The_NerdWait a second
16:23:26Paul_The_NerdIt clearly hasn't wholly frozen.
16:23:42solexxtagtool (linux only, afaik) is pretty good
16:24:09*Paul_The_Nerd is sadly running that MS OS.
16:24:23Paul_The_NerdI think it just may be taking an exceptionally large amount of time generating the playlist.
16:24:32Paul_The_NerdEither that or it *really* hates files with ape tags. Period.
16:24:44Paul_The_Nerd'cuz it'll play anything but songs in my OCRemix folder
16:24:57Paul_The_NerdAnd those all have ID3v2.3 and ApeV2s on 'em.
16:27:31solexxtagtool (linux only, afaik) is pretty good
16:28:45Paul_The_NerdHCl: Okay, my mistake. It does definitely play the file.
16:29:00Paul_The_NerdIt takes 164 second to generate the playlist though.
16:30:08Paul_The_NerdMeanwhile if I browse to it using the normal filetree and play it... less than a second.
16:30:46 Join _DangerousDan [0] (
16:31:11HCli got a problem
16:31:19HClwho can tell me what ata error -71 is :/
16:33:47Paul_The_NerdLooking at an old log, Amiconn got that when his Jukebox didn't have enough battery to get to the HD in USB mode.
16:34:49 Quit Maxime (Read error: 104 (Connection reset by peer))
16:35:35 Join Maxime [0] (
16:35:37 Nick banan_ is now known as merbanan (
16:36:30solexxPaul_The_Nerd: did you investige the wps issue any further?
16:36:41Paul_The_NerdAs far as I can.
16:36:47Paul_The_NerdWhich isn't that far
16:37:20Paul_The_NerdThe fact that it's reproduceable on my end, and that the WPS interpretation ignored comments makes me wonder about some things.
16:37:31 Quit Mxm`Pas`Bien (Read error: 110 (Connection timed out))
16:37:56Paul_The_NerdEspecially my own problem with having to add an extra endline at one point, it may be reading too many or too few characters sometimes.
16:38:02solexxit looks like it is quite buggy as a whole
16:38:11solexxhad that issue with newlines, too
16:38:18solexxbut I cannot remember it exactly
16:39:24solexxanyway, thank you for looking at this
16:39:34 Join Chamois [0] (
16:40:15 Quit muesli- (Read error: 104 (Connection reset by peer))
16:41:38Paul_The_NerdSorry I can't really help other than to confirm the behaviour. =/
16:44:19 Join muesli- [0] (
16:44:56HClwe should have a list of ata errors on the wiki and what they mean :/
16:45:31preglowlinuxstb: know if the ipodlinux guys have done any testing on the nano?
16:45:59HClhey preglow
16:46:57 Quit Maxime (Read error: 104 (Connection reset by peer))
16:47:03 Quit muesli- (Read error: 104 (Connection reset by peer))
16:47:43 Quit DangerousDan (Read error: 110 (Connection timed out))
16:54:07linuxstbpreglow: Yes, I think leachbj owns one and is working on it.
16:54:33linuxstbBut the ipodlinux guys don't seem to talk much about what they are doing.
16:57:23preglowi'd like a flash based player, and the nano might fit my purposes well :)
16:59:07HCli guess
16:59:18HCli dislike ata panics :X
16:59:25 Join Maxime [0] (
16:59:33preglowwhy are you getting them+
16:59:47HCli dunno
16:59:55HCli got an -71 just an hour ago
17:00:03HClmy friend is running fsck.vfat on it
17:09:19markunHCl: I just tried your tagdb tool for the first time. It runs fine with kaffe (open source java)
17:09:51 Join bagawk [0] (i=1000@unaffiliated/bagawk)
17:10:03HCl :)
17:10:20markunAlthough I got a lot of false duplicates.
17:12:03 Join muesli- [0] (
17:12:27markunIt's only with filesnames with non-ascii chars I think.
17:14:36***Saving seen data "./dancer.seen"
17:16:42HClfalse duplicates?
17:17:02HClthats pretty odd, that'd mean the hash algorithm is failing to be accurate
17:18:35HClmaybe it hash fails on all of those files?
17:18:50muesli-how do i use comments in the *.wps file?
17:18:55HClthe bad side of not running in debug mode :p
17:19:01muesli-#bla bla (like this?)
17:19:12HCli wouldn't use comments since you have a limited file size
17:19:30 Quit Maxime (Read error: 104 (Connection reset by peer))
17:19:36 Join Maxime [0] (
17:19:49muesli-hcl. yeah i know...but without them i go crazy
17:20:21 Join novimon [0] (
17:20:22 Join webguest71 [0] (
17:21:04novimonman, I just came to thank you for the great piece of software you've done
17:21:17novimonit's awsome
17:21:46markunnovimon: Do you have an iriver?
17:21:56HClmuesli-: :p
17:22:04 Join tucoz [0] (
17:22:27novimonplus it has the usb-bootmode (or similar) which just saved me from sending my iriver to service
17:22:35novimonmarkun, oh jes
17:22:39tucozHi, do you know if gcc-java is included in the gcc 4.02 tar.gz file?
17:22:53HClwhy would you need/want that?
17:23:09tucozthe 30 MB file
17:23:31tucozHCl: talking to me?
17:23:53tucozok, well, I heard that gcc 4 has better support for java
17:24:05HCli hope so
17:24:10tucozI just wanted to try out a thing or two
17:24:12HCl3.3 failed to compile the songdb.jar
17:24:13Slasheritucoz: just use ibm or sun jdk
17:24:31novimonbtw what does the "battery" in the setting mean? and how come can you add your battery? does it actually work?
17:24:43tucozSlasheri: I want to compile a standalone application
17:24:47tucoz*try to
17:25:01Slasherinovimon: yes, it specifies the battery capacity that is used to calculate the estimated runtime
17:25:02markuntucoz: kaffe also works.
17:25:11tucozok, cool
17:25:35tucozthen we might be able to distribute binaries for the songdb
17:25:35markuntucoz: But doesn't help produce an application
17:25:57tucozOk, hmm. But gcc-java is able to produce binaries?
17:26:17tucozat least, that's what I thought when I tried last time some months ago
17:26:45novimonSlasheri, but you cant "upgrade" your existing battery? :P
17:26:59HClyou can
17:27:08novimonis it safe?
17:27:10HCli did a battery upgrade
17:27:15novimonyeah yeah
17:27:24HClto 2200mah
17:27:40muesli-where do you got it?
17:27:47HClpaid 13 euro
17:28:01muesli-could you give me the link?
17:28:09HClit was ages ago
17:28:15HClits somewhere in the rockbox logs
17:28:15Slasherii also upgraded to 1900 mAh ionity
17:28:34HClits pretty much an ipod 1st gen battery replacement
17:28:35muesli-Slasheri second this
17:28:38HCland you swap the wires
17:28:43HClgonna go
17:29:26muesli-but wasnt ipods battery not known for poor quality??
17:29:35muesli-they had to replace them imho
17:29:37novimonis there an "recorder" and "play mode" in the iriver rockbox?
17:29:52tucozmuesli: the shipped batteries. Not the replacements (I think)
17:30:14 Join einhirn [0] (
17:30:16Slasherinovimon: the true recording menu will be enabled soon
17:30:47markunmuesli-: HCl and I have the same battery. Works fine.
17:31:03 Quit webguest71 ("CGI:IRC (EOF)")
17:31:08muesli-coz...13e are a gift
17:31:24muesli-i paid 40e for the ionity
17:32:01novimonSlasheri, is there better manual for irvier rockbox than the official 2.5 manual?
17:32:02markunJust search for "ipod 2200" on ebay. We bought them from Hong Kong.
17:32:27Slasherinovimon: please take a look at the wiki manual, it should be more up-to-date
17:32:40novimonSlasheri, ok will do that
17:35:19Paul_The_NerdHow much additional battery life did you get upgrading to a 2200 (like, percentage wise?)
17:36:32tucozaccording to some place I read, the standard battery will run for 16h. With a 1900mah you will be able to have the player running for 29h under the same circumstances.
17:37:05tucozBut this was at dapshop or whatever it is called
17:37:10Paul_The_NerdWow... there's a lot of 3 ipod 2200 batteries up for US $15 (buy it now).
17:37:18muesli-juergen @dap-store tested it
17:37:27Paul_The_NerdBut... they say "no returns" like 500 times with warning that "we cannot test the batteries or insure they work"
17:37:33Paul_The_NerdThat raises confidence.
17:38:09novimonwhats the default iriver battery in mah?
17:38:40muesli-whats the dimension of pixels from the remote?
17:38:55_FireFly_128x64 afaik
17:39:13muesli-ah..there you are _FireFly_
17:40:15 Quit Chamois (" HydraIRC -> <- The future of IRC")
17:40:37Paul_The_NerdCan you tell if a button was pressed on the remote vs. the player?
17:41:58Paul_The_NerdThe play button on the remote. Is it different from the play button on the box itself? (I'm not talking about how they're handled now, just... can they be interpreted differently if someone wanted to?)
17:42:46_FireFly_yes as BUTTON_RC_ON for the play-button on the remote and BUTTON_ON for the main device
17:43:00_FireFly_so are the defines in button.h
17:44:26Paul_The_NerdSo, when the remote's finalized it'd be possible for say... one to retain control of the music/wps on the remote, even while, say, playing a game in a plugin?
17:45:05_FireFly_that should no problem but currently not with my patchset :)
17:45:16Paul_The_NerdYeah, I'm somewhat familiar with how yours works.
17:45:23Paul_The_NerdI just wasn't familiar with the button handling at all.
17:45:50 Quit Maxime (Read error: 104 (Connection reset by peer))
17:46:04 Join Mxm`Pas`Bien [0] (
17:54:11novimonwhats the difference between 'mix' and 'crossfade'?
17:58:49muesli-title.bmp|0|5| is that a line for a solid pic?
17:59:12_FireFly_solid pic ??
17:59:22muesli-not conditional just solid
17:59:42_FireFly_the wps tag is missing ;)
18:00:24 Quit bagawk (Read error: 110 (Connection timed out))
18:00:26muesli-%x|title.bmp|0|5| ?
18:00:30_FireFly_it should looks like this %x|a|title.bmp|0|5|
18:00:51_FireFly_every image in the wps has an identifier
18:00:57_FireFly_must have
18:01:09muesli-but then i have to call that images?
18:01:53 Join bagawk [0] (i=1000@unaffiliated/bagawk)
18:02:17_FireFly_look at the wiki-page ;)
18:02:38_FireFly_tha %xl images are for conditional images
18:02:51_FireFly_the %xl tag
18:03:15muesli-aah ;)
18:03:29muesli-those tags drive me crazy ;)
18:05:16Paul_The_Nerdnovimon: One fades the ending song while bringing up the new song, and the other just brings up the new song, from what I understand.
18:07:21novimonPaul_The_Nerd, oki
18:09:51 Quit tucoz ("CGI:IRC 0.5.4 (2004/01/29)")
18:16:35 Quit bagawk ("Leaving")
18:19:35 Join hshah [0] (
18:21:57 Quit hshah (Client Quit)
18:29:20 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
18:42:04 Quit muesli- (Read error: 110 (Connection timed out))
18:43:03 Part linuxstb ("Leaving")
18:49:42 Join linuxstb [0] (
19:03:11 Join dpassen1 [0] (
19:09:36linuxstbCan anyone explain what the following ARM assembler commands do?
19:10:08linuxstb ldrcc r2, [r4], #4
19:10:09linuxstb strcc r2, [r5], #4
19:12:39linuxstbIt's OK, I'm understanding them now.
19:14:01Mxm`Pas`Biennever understood asm :x
19:14:04 Nick Mxm`Pas`Bien is now known as Maxime` (
19:14:40***Saving seen data "./dancer.seen"
19:24:14 Quit crash__ ("leaving")
19:46:20 Join wrecker [0] (
19:47:29wreckerSomeone's posted on Misticriver that Splitedit plugin is not available on Iriver, despite it indicating that it is in the rockbox Wiki. Looks like he's right. Does anyone know anyhting about this? Thanks..
19:48:23 Quit einhirn (Read error: 110 (Connection timed out))
19:52:25 Join crash_ [0] (
19:53:44 Join linuxstb_ [0] (
19:54:00linuxstb_You can see from the apps/plugins/SOURCES file (the last few lines) that splitedit and metronome don't work on players with software codecs.
19:54:34linuxstb_I'm sure it will eventually - but any documentation that says it works at the moment should be corrected.
20:05:16SlasheriHmm, the metronome should work, with very minimal fixes
20:05:29Slasherijust trying it
20:08:11 Join Chamois [0] (
20:08:44wreckerok. thanks. I'll update the wiki when I get chance.
20:08:53 Quit wrecker ("CGI:IRC")
20:09:10 Quit linuxstb (Read error: 110 (Connection timed out))
20:09:50linuxstb_I've just updated the Plugin pages for Metronome and Splitedit to remove the iriver reference.
20:10:17linuxstb_Slasheri: If you get Metronome working on the iRiver, then don't forget to update the Wiki :)
20:10:24Slasherilinuxstb_: hehe :)
20:10:34 Nick linuxstb_ is now known as linuxstb (
20:12:55linuxstbSlasheri: Does that mean you are back working on Rockbox?
20:13:22Slasherilinuxstb: basically yes, but i have limited time to work with it :)
20:13:50*linuxstb wonders why the Metronome and Splitedit plugin pages contain key mappings for iRiver
20:16:38XavierGrhow the military is going Slasheri?
20:16:54 Join Moos [0] (
20:17:44MoosHello folks !
20:17:54XavierGrHi Moos
20:18:11MoosHow are you there?
20:18:31SlasheriXavierGr: well, the place i work is pretty good but i have to do long days there. [TJ 326|17.4%] :) (tj=remaining days of the service)
20:18:51Moos:-D Hi Slasheri
20:18:56Slasherihi Moos :)
20:19:21MoosXavier: How the FM preset things goes?
20:19:51Moosnot read the logs yet :)
20:21:01MoosSlasheri: Linus didn't see your dircache code yet?
20:21:39SlasheriMoos: he will try it when he has some time. Maybe soon
20:22:01Mooslike everytime :D
20:22:05Moosbusy man
20:22:13Slasherithen i can commit it if it doesn't have any critical issues
20:22:37Moosany bugs reported?
20:24:38Moostime to read the logs for me :)
20:28:24XavierGrMoos: I am proud to say that the fm presetloading/saving is pathc is ready and waits commitment.
20:28:57XavierGrSpeaking of it did anyone tested it? I need feedback and bug tracing (though I doubt it will have any)
20:29:01Moosvery fine, waiting in Linus too :)?
20:29:21 Join hshah [0] (
20:30:53Moosfor your jpeg browser too ;) ?
20:31:02Moosok thanks
20:31:09XavierGrhere is the diff if anyone is interested.
20:31:18XavierGrpush download and save it.
20:31:31XavierGryes the jpeg viewer is waiting too.
20:31:43XavierGrmany fronts I could say.
20:31:50XavierGrSlasheri's dircache
20:31:55 Quit solexx (Read error: 104 (Connection reset by peer))
20:31:57XavierGrjpeg viewer and fm presets
20:32:08XavierGrand now TiMiD is working for proper remote support
20:32:20Mooslots of very cool things :-)
20:32:41XavierGrthen recording remains, what else?
20:32:59MoosSlasheri? :P
20:33:08XavierGrmaybe rockboy optimizations but that is DAP irrelevant
20:33:19Moosyes indeed
20:33:26 Join linuxstb_ [0] (
20:33:53XavierGrlinuxstb: What's with the multiple join quit?
20:34:16SlasheriMoos: yes, i will commit some basic stuff to enable the recording screen and peakmeter there soon
20:34:35Moosthanks ;)
20:36:48 Join solexx [0] (
20:38:46XavierGrAlso the peak meter for the radio is still pending.
20:39:06XavierGrIf only I knew hot to fix that, but I don't have a clue. Any ideas
20:41:50linuxstb_XavierGr: Sorry, unreliable connection.
20:42:47XavierGrlinuxstb: No I am not bothered just asking...
20:43:08XavierGrIs there a connection between radio peakmeter and wps peakmeter?
20:43:50 Quit hshah ("Leaving")
20:44:36linuxstb_I've just looked quickly at your radio patch, and I think you will need to add some #ifdef CONFIG_TUNER around some of your changes - so they don't get compiled for targets without FM.
20:46:46 Join matsl [0] (
20:47:54 Quit linuxstb (Read error: 110 (Connection timed out))
20:50:09Slasherilinuxstb_: yes there is, but especially the recording peak meter would be the same as radio peak meter
20:50:17SlasheriXavierGr i meant
20:52:17XavierGroh yes you are right
20:52:30XavierGryou mean changes outside fmradio.c
20:56:51 Join linuxstb__ [0] (
20:58:17XavierGrso put #ifdef CONFIG_TUNER in every segment of code outside radio.c (which compile only on FM targets) that has to do with the radio right?
20:59:21linuxstb__Yes, I think so. Other radio changes are like that - e.g. the FM settings already in settings.c
21:00:54 Quit _DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
21:03:22XavierGrhmm when reentering the FM screen you can still hear a faint pop.
21:03:33 Join sangre [0] (i=sangre@
21:03:52sangrehey, is there any way to speed on scrolling through long song lists?
21:04:20XavierGrAlso when exiting the FM screen the reception gets slightly more bad. And I didn't changed that part of the code. I will test with a clean build.
21:05:06XavierGrpress play while scrolling
21:05:14XavierGrand hold it
21:08:48XavierGrSlasheri: I think I know why the radio peak meter isnt working. Peak meter on wps (on iriver) is working from the data that the pcm buffer has to offer. But the radio will not use the pcm buffer.
21:10:17 Join guillaume [0] (
21:12:28markunlinuxstb__: Have you been working on the root menu?
21:12:41linuxstb__Not really. I've been busy on the iPod.
21:12:49 Quit linuxstb_ (Read error: 110 (Connection timed out))
21:13:04 Nick linuxstb__ is now known as linuxstb (
21:13:18sangreahh savage XavierGr
21:13:21sangreim delighted :)
21:14:10markunlinuxstb: Maybe its better to wait for the 'widget' code?
21:14:41***Saving seen data "./dancer.seen"
21:14:42XavierGrsangre: Excuse me, savage :p?
21:15:07XavierGroh it worked out?
21:15:13XavierGrforget it now I got it
21:15:44fuzzielinuxstb: being busy on the iPod sounds promising?
21:20:29guillaumeis an audioscrobbler/ support planned in rockbox?
21:20:43linuxstbfuzzie: Yes, things are progressing well. Slowly, but well.
21:22:44markunguillaume: It will only work for units with a real time clock I think.
21:24:22guillaumehmm i haven't thought about this problem
21:24:38guillaumeindeed there must me some kind of dating to the listens
21:24:48guillaumes/must/has to be
21:27:12sangre<XavierGr> sangre: Excuse me, savage :p? // heh..yeah great/deadly
21:29:36Slasherize: that's true and that's why it has to use the recording peak meter
21:29:55zeSlasheri: spooky...
21:30:08zeSlasheri: i just /back and right away you accidently direct something to me
21:30:10zeSlasheri: :p
21:32:44Slasherihehe, sorry :D i meant trying x<tab> but pressed z :)
21:33:35XavierGrDo you think it is difficult to implement?
21:33:39zei know... just strange timing for it...
21:33:46XavierGrLot of code and low level.
21:37:09SlasheriXavierGr: it's already implemented for recording (on my local build) so it shouldn't be hard
21:38:05XavierGron your local build you mean last daily build?
21:38:33XavierGrWell as far as I know radio.c mustn't be changed for the peak meter to work.
21:38:47XavierGrmaybe peakmeter.c wants modifications.
21:40:33XavierGranyway this stuff is out of my league
21:44:56 Join Maxime [0] (
21:56:12 Quit Chamois (Read error: 104 (Connection reset by peer))
21:56:51 Join Mxm`Pas`Bien [0] (
21:57:48 Quit Maxime` (Read error: 110 (Connection timed out))
22:04:06 Quit XavierGr ()
22:06:25 Quit _FireFly_ ("Leaving")
22:15:28 Quit Maxime (Read error: 110 (Connection timed out))
22:19:35 Quit matsl (Read error: 110 (Connection timed out))
22:22:23 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
22:36:34 Join RotAtoR [0] (
22:38:37 Join matsl [0] (
22:42:56 Join linuxstb_ [0] (
22:57:04 Join linuxstb__ [0] (
22:59:02 Quit linuxstb (Read error: 110 (Connection timed out))
23:12:06 Quit linuxstb_ (Read error: 110 (Connection timed out))
23:14:44***Saving seen data "./dancer.seen"
23:20:04 Quit paugh ("Leaving")
23:26:44 Quit guillaume (Read error: 104 (Connection reset by peer))
23:33:54 Nick linuxstb__ is now known as linuxstb (
23:34:05amiconnHmm. Where's the irc log of 2005-09-30 gone?
23:41:10 Join MipsIrv [0] (
23:54:30 Join XavierGr [0] (
23:59:42 Quit matsl (Read error: 60 (Operation timed out))

Previous day | Next day