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).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2005-11-29

00:00:03JdGordonuno momento
00:00:45markunlinuxstb: what type? unsigned char I think
00:00:58JdGordonhttp://users.monash.edu.au/~jdgor1/sudoku2.jpg
00:01:16linuxstbmarkun: Do you use the standard string.h functions with your utf-8 strings?
00:01:17preglowseriously, are builds going on for longer than half an hour now?
00:01:25preglowtime to cut out some targets? :/
00:01:27Bagderat this hour, yes
00:01:44preglowbut the estimate was correct
00:01:49preglowand that was based on commits done this morning
00:01:52linuxstbJdGordon: 404 error
00:02:01JdGordonhttp://users.monash.edu.au/~jdgor1/rb/sudoku2.jpg
00:02:04JdGordonsoz
00:03:10linuxstbYou _may_ have done something wrong there.
00:03:21linuxstbBut I don't really mind if you leave them at 16x16.
00:03:34linuxstbJust that the second set of numbers didn't align properly.
00:03:37JdGordoni did nothing but change the cell_height and cell_width macros
00:03:48JdGordonill fix them
00:03:59Bgerheh that shouldn't be enough ... (or i'm wrong)
00:04:02linuxstbYou'll obviously need new bitmaps as well.
00:07:19amiconnLinusN: Do you think it's better to have a mini crt0.S for the decompressor, or put the asm in the C file?
00:07:20BgerJdGordon: 16x16 as well
00:07:39JdGordon?
00:07:44amiconnOf course it shouldn't be called crt0.S
00:10:06BgerJdGordon: the current bitmaps aren't 16x16
00:12:51 Quit muesli_- (Read error: 110 (Connection timed out))
00:13:45preglowlinuxstb: btw, still don't get the file browser when you boot?
00:14:03 Join cool [0] (i=Benacool@modemcable035.68-202-24.mc.videotron.ca)
00:14:44 Quit Benacool (Read error: 110 (Connection timed out))
00:14:46 Nick cool is now known as Benacool (i=Benacool@modemcable035.68-202-24.mc.videotron.ca)
00:14:59LinusNamiconn: do whatever you feel is kiss
00:15:11linuxstbpreglow: I fixed that with my last lcd commit.
00:15:22preglowgoodies
00:15:59linuxstbSilly copy and paste errors
00:15:59 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
00:17:05 Join RotAtoR [0] (n=e@12-210-82-91.client.insightBB.com)
00:17:11 Join schnittn2005 [0] (i=schnittn@xdsl-81-173-149-219.netcologne.de)
00:17:15JdGordonk, fine done
00:17:19JdGordonlast time :p
00:20:30LinusNd'oh, iaudio x5 has no stop button
00:20:44linuxstbJoin the club. The ipod is the same.
00:20:58LinusNand headphone connector is on the side!
00:21:11linuxstbNot even the ipod is that crazy
00:21:13preglowyeah, that last is quite ingenious
00:21:28LinusNit's horrible
00:21:43mirakThe only flaw of the iriver H300 is that they are just a bit to big
00:21:46preglowat least they've made it clear they don't want you keeping it upright in a pocket
00:21:59mirakthey also have a crap firmware, butt there is rocbox
00:22:34LinusNi haven't run the x5 long enough to have an opinion about the firmware
00:22:49LinusNthe lcd isn't that good though, it's rasterized
00:23:15JdGordonlinuxstb: does the solver functoin fill the the game state board?
00:23:17mirakH300 firmware is not that bad, but it could be perfect. They don't develop anything on it
00:23:25mirakthey should have released the source code
00:23:46ashridahmirak: they'd probably have to strip out half of it to do so
00:23:56mirakashridah: you mean codecs ?
00:23:57linuxstbJdGordon: Yes - it fills currentboard
00:24:09JdGordonok good... so ill try adding a game maker also
00:24:22ashridahmirak: at least one of them, other things as well tho
00:24:44JdGordonhttp://users.monash.edu.au/~jdgor1/rb/sudoku2.jpg better?
00:24:44LinusNi like the x5 joystick, much better than the h100
00:24:48linuxstbJdGordon: That would be cool. But before you make any more changes, can I have your latest bitmaps, and I will then commit it to CVS.
00:25:04mirakthe positions of hardware manufacturers relative to software is a bit weird
00:25:36mirakH300 would be could with a click wheel like the ipod
00:25:51mirakI guess the wheel is patented
00:26:26LinusNgood, then other manufacturers won't get the same stupid idea :-)
00:26:34mirakimagine if a guy patented the wheel. We would have square wheels for our cars
00:26:36JdGordonbetter be... just added in for good :p
00:26:37preglowahahah
00:26:41*JdGordon is sick of these stupid numbers :D
00:26:52ashridahmirak: not really.
00:27:09mirakashridah: or the guy would have golden bollocks
00:27:14linuxstbJdGordon: I think I prefer the first set....
00:27:14ashridahmirak: you have to understand that it doesn't matter what you actually sell, you can't give away what you don't own.
00:27:26LinusNreboot, brb
00:27:28 Part LinusN
00:27:39ashridahand i'm willing to bet that iriver don't own the wma codec, SRS, their video codec, and a bunch of other stuff.
00:27:54JdGordonna these are better than 1st set
00:28:04mirakashridah: how is it given away if anyway it would be used on that same hardware anyway ?
00:28:05JdGordoncept the 9 and 6
00:28:32mirakashridah: but I understand
00:29:08mirakI am going to bed
00:29:09ashridahwhat the code runs on is irrelevent, it's who can see the code.
00:29:10mirakbye
00:30:30 Join DJDD_ [0] (n=DJDD@220-245-186-182.static.tpgi.com.au)
00:31:18JdGordonstupid irc
00:32:17JdGordonis there rand() on rb?
00:32:57 Join LinusN [0] (n=linus@labb.contactor.se)
00:33:07preglowyeah
00:33:22ashridahrb->rand() or something
00:33:37ashridah(assuming it's from a plugin
00:34:13JonSeniorLinusN: A quick an' dirty hack (ifdef...def...undef) implemented the H100 USB code on the H300 target. I notice that the USB detection works in the Read I/O ports menu.
00:34:22JonSeniorI don't recall if it did before however.
00:34:33JonSeniorWhat is the problem with the two power lines to the chip?
00:34:47JonSeniorAre we missing an IO line somewhere?
00:35:37 Quit iobound ("Leaving")
00:35:48LinusNthe usb detection works fine
00:36:35JonSenior...but the rest of the USB code doesn't?
00:36:50JonSeniorDo we need the rest of the driver for both targets?
00:37:02LinusNthe rest?
00:37:18LinusNthey should work identically
00:37:23*JonSenior is horribly confused!
00:37:49JonSeniorWhat is currently missing from the USB implimenation on the Hxxx platforms?
00:37:56preglowrockbox entry point always is 0?
00:38:08LinusNpowering the USB chip properly, i guess
00:38:31LinusNpreglow: ?
00:38:36preglowfile offset
00:38:59preglowhmm, forget it, i can just overwrite the entry point code when writing exception vectors anyway
00:39:43preglowipod rockbox will remap sdram space to 0x0, which is also where exception vectors are
00:40:41 Join Kyl3 [0] (i=Kyle@cpe-24-90-232-130.nyc.res.rr.com)
00:41:55 Part paugh ("Leaving")
00:42:31amiconnpreglow: Does ipodlinux remap sdram to 0 ?
00:42:58preglowamiconn: yes, you need to, default setup is flash at 0x0, and then you can't write to the reset vectors
00:42:58 Join matsl [0] (n=user@1-1-4-2a.mal.sth.bostream.se)
00:43:07preglowehh, exception vectors
00:43:16amiconnYes, except if you run from flash
00:43:27preglowyeah, but we'll never do that, i'm sure
00:43:45amiconnI would be icking to do that, for boot time...
00:43:52preglowit's not that big an issue with ipods
00:43:55amiconnThat's another con for iPod
00:43:58preglowthe sleep mode is very efficient
00:44:17amiconnPerhaps for the nano...
00:44:30 Quit mirak (Remote closed the connection)
00:44:38 Quit matsl (Client Quit)
00:44:38amiconnBoot times are *always* too long imho
00:44:40dwihnoLinusN: I received the reply from iriver today. They told me to return the unit to the point of purchase... I think it's just more than my unit. Something in the software.
00:44:55preglowamiconn: boot time is one of the things that don't annoy me too much
00:44:56LinusNaha
00:45:25dwihnoLinusN: don't you have an IFP? *wink wink*
00:45:37LinusNnope
00:46:08linuxstbamiconn: It takes about 7 seconds for the apple bootloader to load our bootloader. Then about 2-3 seconds before the file browser appears.
00:46:10dwihnoHm. I need to take my anti-senility pills.
00:46:31linuxstbBut I'm not sure if we'll ever know enough about the ipod to get rid of the apple bootloader - it initialises all the hardware for us.
00:46:38amiconnlinuxstb: At least two times too slow...
00:46:51linuxstbamiconn: I agree.
00:47:02linuxstbBut there are worse things in the world...
00:47:10*amiconn just successfully RoLo'ed compressed unicode rockbox on recorder v1 :-)
00:47:20LinusNcoooooooool
00:47:34dwihnoThat's really cool!
00:47:41preglowlinuxstb: i've got three seconds total boot time
00:47:43dwihnoWhat is cool?
00:47:50dwihnoRiiight, the anti-senility pills....
00:48:00linuxstbpreglow: No wonder you don't care about running from flash.
00:48:07markunamiconn: great job!
00:48:10amiconnNow I need someone to fix the Makefile for integration, and write makeimage.pl
00:48:22amiconnBagder?
00:48:52markunSo, what was the problem in the end?
00:49:03JonSeniorLinusN: Sorry to keep pestering... Is there a wiki page with details of which pins of the CY7C... USB controller are connected to GPIOs?
00:49:17amiconnAn ooooooolllllddd bug in RoLo, plus me not making sure where the stack pointer is
00:49:29JonSeniorLinusN: Or should I start tracing the hardware scans?
00:49:35amiconnSo the stack ended up *inside* the compressed data area...
00:49:47LinusNJonSenior: the thing is that i dont know the hardware details
00:49:58JonSeniorLinusN: Ahhh
00:50:31JonSeniorTime to start looking then... I'll let you know if I find anything! ;-)
00:50:38JdGordondo i need to worry about calling srand, or does rb start it off with a nice seed?
00:50:46JdGordonumm.. i mean a new seed each start
00:50:57preglowyou need to call srand
00:51:10amiconnlinuxstb: In fact preglow's iPod *is* running from flash ;)
00:51:18markunLinusN: I wonder how I can find out how tha ata bridge is connected to the gigabeat's cpu without desoldering it..
00:51:25JdGordongrr... whats a nice way to keep the seed sandom without having time(NULL)??
00:51:31preglowamiconn: pretty weird hardware, flash based AND 32 megs of ram
00:51:32LinusNmarkun: bga?
00:51:36markunyes
00:51:50preglowamiconn: but then again, the file storage flash might of course be fragmented and hard to use directly
00:51:53LinusNthen i think disassembling the stock firmware is the only option
00:51:57markunand no vias on the other side of the pcb
00:52:03LinusNif you don't want to slaughter it
00:52:05amiconnpreglow: Design reuse... cheaper than designing something new
00:52:14amiconnIt's the same with the Ondio, fortunately
00:52:18preglowamiconn: yep, figured that, but still
00:52:29*preglow huggles his 32kb extra iram
00:53:12RotAtoRJdGordon: something like "rb->srand(*rb->current_tick);"
00:53:59JdGordonhmm.. just saw thta... but that wont wok on whatever uses PP5020 cpu..
00:54:04JdGordonwhich is?
00:54:58 Part JonSenior
00:55:10JdGordonwhat does audio_get_file_pos return if no file is being played?
00:55:22 Join stripwax [0] (n=stripwax@i-83-67-214-206.freedom2surf.net)
00:55:24JdGordonmight use that for the seed.. doesnt seem to be player dependant
00:55:25stripwaxlo
00:55:28LinusNJdGordon: what won't work on the pp5020?
00:55:37JdGordonrb->current_tick
00:55:39 Quit Moos ("Glory to Rockbox")
00:55:42LinusNwhy?
00:55:45 Join Moos [0] (i=DrMoos@m79.net81-66-158.noos.fr)
00:55:50JdGordon#if CONFIG_CPU!=PP5020
00:55:50JdGordon long* current_tick;
00:55:50JdGordon#endif
00:55:59LinusNhow odd
00:56:07JdGordonshould thta not be there?
00:56:24amiconnThis is temporary
00:56:29LinusNwhy?
00:56:41amiconnAsk linuxstb or preglow
00:57:09linuxstbWe don't have interrupts yet, so we fake a current_tick using the rtc.
00:57:22linuxstbcurrent_tick is a macro - so we can't put it in the plugin struct.
00:57:31LinusNoh
00:57:35linuxstbAs amiconn said, it's a temporary thing.
00:57:56JdGordonso its safe to but it in the plugin? what player ses the pp5020?
00:58:18linuxstbJdGordon: Yes, it's safe to assume current_tick is always there.
00:58:20LinusNJdGordon: don't worry about the pp5020
00:58:23 Part stripwax
00:58:24JdGordonok
00:58:43LinusNthose lamers won't finish that port anyway ;-)
00:58:49JdGordonhaha
00:58:54preglow!
00:59:14linuxstb:)
00:59:16 Join KaL| [0] (n=gather@h60n2fls304o1119.telia.com)
01:00
01:00:03amiconnreboot, brb
01:00:06 Part amiconn
01:02:19Bgernite
01:03:01markunnite
01:03:27 Join DJDD__ [0] (n=DJDD@220-245-186-182.static.tpgi.com.au)
01:03:58JdGordon(rb->rand()%9)+1 will give a number between 1 and 9 inclusive right?
01:03:59linuxstbJdGordon: Sudoku patch committed.
01:04:08JdGordonwoops.. that 9 should be a 10.. right?
01:04:27 Quit schnittn2005 ()
01:04:33JdGordonawesome... the game generator is almost done.. i tinhk :d
01:04:36 Join amiconn [0] (n=jens@p54BD5E42.dip.t-dialin.net)
01:05:27LinusNJdGordon: no, 9 was correct
01:05:44JdGordonok, thanx
01:06:09***Saving seen data "./dancer.seen"
01:06:31linuxstbJdGordon: How are you generating games?
01:07:02JdGordonfuck... im asleep still
01:07:28JdGordoni was randomly adding numbers.. but obviously that wont work :p
01:08:01JdGordonim gonna add 1 number to each diagnol group then run the solver and it should make a full board...
01:08:05JdGordonthen remove numbers?
01:08:26 Quit Bger (Read error: 104 (Connection reset by peer))
01:08:39linuxstbJdGordon: I don't think writing a generator is a five minute job.
01:08:51JdGordonno, but using the solver should make it one
01:09:01JdGordonassuming it can work with an almost empty board...
01:09:32*JdGordon 's brain hasnt been working since exams finished 3 weeks ago
01:09:48*preglow tries his remap code...
01:10:19 Quit San (Read error: 110 (Connection timed out))
01:10:25preglownot a complete success...
01:11:12amiconnJdGordon: Just using the solver and then removing number doesn't ensure 'good' sudokus
01:11:25JdGordonsure it does
01:11:35amiconnAfaiu, a good one should only have exactly one solution
01:11:45JdGordonthe difficulty comes in how many cells u remove...
01:12:00amiconnHow many and/or which
01:12:02linuxstbBut doesn't it also matter _which_ cells you remove?
01:12:17markunamiconn: can I commit unicode after bagder has done his magic, or are there some things I should change?
01:12:20 Join muesli_- [0] (i=muesli_t@Bc0a6.b.pppool.de)
01:12:50 Quit Maxime (Read error: 104 (Connection reset by peer))
01:13:05JdGordonfine... im looking for a gpl generator i can nick
01:13:09amiconnmarkun: Hmm. I'd like to test a bit more, but I would need files with exotic filenames etc...
01:13:22linuxstbJdGordon: There was one linked from the wiki
01:13:50linuxstbhttp://www.laurasia.com.au/sudoku/
01:13:55markunamiconn: the viewer is still not finished, so maybe I should wait with it's commit.
01:14:02amiconn...and of course I need to know how it's supposed to look like
01:14:21linuxstbJdGordon: I had a quick look at that generator, and it seemed suitable for us.
01:14:28markuna lot of people have been using the unicode patch and I have not received complaints about major bugs
01:14:32JdGordonok
01:14:41amiconnBtw, your unicode patch with disabled arabic joining defines NO_ARABJOIN in two places
01:14:42linuxstbBut feel free to look around some more.
01:14:50markunAh, does it?
01:14:51amiconn...one of which is probably wrong
01:14:55 Join Maxime [0] (n=flemmard@fbx.flemmard.net)
01:14:56markunThat's not good
01:15:11amiconnYes, once in config-*.h (correct) and a second time in bidi.c
01:15:33markunthe one in bidi.c should be removed of course
01:15:55amiconnJdGordon: I hope this generator doesn't increase code size too much
01:16:19JdGordonhow big is too big for a rock?
01:16:34markunbut maybe the NO_ARABJOIN could be replace by ifdef HAVE_LCD_BITMAP, or is it useful for charcell as well?
01:17:00amiconnJdGordon: 32KB on archos
01:17:06linuxstbamiconn: It's only about 5KB on the Archos devices at the moment.
01:17:10amiconnyes
01:17:18JdGordonk
01:17:26amiconnmarkun: The other way round I presume?
01:17:47markuninstead of ifndef NO_ARABJOIN..
01:17:50amiconnarabjoin doesn't make sense on charcell lcd
01:17:52Benacoolis it me or the rockbox firmware only turn off the backlite of the H3x0 LCD? (the original firmware turn off the LCD not only the backlite)
01:18:03amiconnmarkun: ah, yes
01:18:17amiconnBenacool: You are right
01:18:27amiconnw-i-p, you know?
01:18:28amiconn;)
01:18:33*preglow wonders how long it'll be until he wears out his flash
01:19:12Benacoolamicon: since we see absolutely nothing when the backlite is off it will be beter to turn it off ;-)
01:19:37amiconn...work in progress...
01:19:42Benacoolok ;-)
01:19:47markunI'm sure there will be some bugs after the initial unicode commit, but it's getting harder to maintain outside the cvs so I would prefer it to be committed soon
01:20:03*amiconn prods Bagder
01:20:36 Quit bobwise ("CGI:IRC")
01:20:56markunmakefile wizard
01:21:21 Quit DJDD_ (Read error: 110 (Connection timed out))
01:23:18 Quit KaL| ()
01:26:54 Quit skorpyon_ (Read error: 110 (Connection timed out))
01:32:41preglowargh!
01:33:03 Quit Paul_The_Nerd ("Chatzilla 0.9.68.5.1 [Firefox 1.5/undefined]")
01:35:22 Quit Moos ("Glory to Rockbox")
01:37:04JdGordonFARK
01:37:38preglowlinuxstb: you think we'll ever need to support pp5002? :>
01:38:15 Quit muesli_- (Read error: 110 (Connection timed out))
01:46:52 Quit Cassandra (" HydraIRC -> http://www.hydrairc.com <- IRC has never been so good")
01:47:16preglowwoah, it loaded
01:48:08grant_?
01:49:15preglownothing, just me getting too accustomed to bugs
01:49:28grant_lol
01:52:12linuxstbpreglow: It would be nice to support the pp5002 (lots of people have them), but I don't think we should worry about it at the moment.
01:52:33linuxstbLet a pp5002 owner worry about it in the future.
01:53:17preglowi do so wish i could just fire an interrupt soon
01:53:29 Join DocBrown [0] (n=508ffea6@labb.contactor.se)
01:55:31preglowwell
01:55:36preglowat least the memory remapping works
01:55:43preglowi've written to the exception vector table
01:56:45DocBrownLinusN: Hi! I'm one of the unlucky H3x0 black screen guys. Have managed to boot Rockbox once by pressing arbitrary key combination
01:57:46DocBrownLinusN: Trying to reproduce this (which didn't work) I founf out that any PLAY-other key-PLAY (long) combination will start the iriver firmware
01:58:09DocBrownDon't know if that information is helpful
01:58:29 Join Cassandra [0] (i=Cassandr@elmyra.coraline.org)
01:58:48preglowhah
01:58:49preglowweird
01:59:31LinusNDocBrown: ehh, in sequence you mean?
02:00
02:00:00DocBrownLinusN: Yes
02:00:03 Quit Cassandra (Client Quit)
02:00:04 Join Cassandra [0] (i=Cassandr@elmyra.coraline.org)
02:00:26LinusNcool
02:01:05DocBrownLinusN: would be even cooler if I knew the combination for Rockbox :-)
02:01:41*Cassandra wonders why everyone is blaming Rockbox for their USB connection problems.
02:01:45LinusN14-35-73-12
02:01:47grant_PLAY
02:02:14LinusNCassandra: there is probably a bug in the rb usb handling
02:02:27linuxstbpreglow: Are you saying you can change the exception vectors, but your interrupts still don't work?
02:02:40grant_Has any work been begun for the usb support?
02:02:43preglowlinuxstb: yeah, i haven't managed to enable any of them
02:03:03CassandraWell, I've seen crashes on connecting while playing. But surely once it's plugged in, it's just a matter of the chip handling the USB.
02:03:51CassandraYou know, if I had an H3xx I'd be quite tempted to port ScummVM to Rockbox.
02:05:20Cassandra(Although I imagine the sound support would be kind of tricky.
02:05:38linuxstbThe C++ involved would be trickier....
02:05:57CassandraAh, it's C++ is it? Not much of a chance then.
02:06:37linuxstbYep. And from a quick look it looks like "real C++", not just C disguised as C++.
02:06:53CassandraCurse them and their object orientation.
02:07:05CassandraMonkey Island on Rockbox would be damn funky.
02:07:31linuxstbI'm guessing that's the only program of its type?
02:07:41CassandraYup.
02:10:25DocBrown14-35-73-12
02:10:32DocBrownLinusN: ?
02:12:40preglowLinusN: i suspect that was a joke
02:12:43preglowDocBrown: !
02:12:50preglowi need to consider sleep
02:13:08preglowhas anyone even tried to make a c++ plugin?
02:13:12DocBrownpreglow: Me too
02:13:36preglowshould be possible, probably depends on how much dynamic allocation it does
02:14:54preglowmonkey island has to be played on an old dos box with an adlib sound card in it anyway ;)
02:14:59linuxstbI think that's the only way scummvm could be ported.
02:15:18linuxstbBut it's the start of a slippery slope....
02:15:24preglowindeed
02:15:32preglowi don't care about scummvm on rockbox anyway
02:15:37CassandraI played it on a Sound Blaster. Ph34r my l33t technology.
02:15:47preglowCassandra: well, yeah, me too, but in adlib mode!
02:15:58preglowman, those games are leet
02:16:05Cassandra(A microchannel Sound Blaster Pro, if I remember correctly.)
02:16:19preglowdon't remember what mine was, only that it was a nightmare to get it working
02:16:21CassandraThe music was really great too.
02:16:24preglowindeed
02:16:36CassandraAll sound cards were a nightmare back in 1992.
02:16:49 Join dropandho [0] (n=dropandh@cpe-24-193-36-91.nyc.res.rr.com)
02:16:55preglowSET BLASTER...
02:17:38preglowahh, those were the days
02:17:45dropandhoslasheri- you around?
02:17:55dropandhoor anyone that worked on the tick-fix?
02:18:13CassandraOoh - only 5 more people need to sign up for us to get a campaigning digital rights organisation in the UK. (http://www.pledgebank.com/rights/)
02:18:47CassandraI'd buy Monkey Island, the album.
02:19:52Cassandra(Unfortunately the Interweb is full of shitty remixes.)
02:20:00DocBrownLinusN: Another information (maybe more useful): Having set the HOLD switch and pressing PLAY the Rockbox bootloader starts!
02:20:26DocBrownLinusN: But then switches off because HOLD switch is set :-)
02:20:58LinusNinteresting indeed
02:22:19DocBrownLinusN: Rockbox booted! Pressing play and directly setting the hold switch off
02:22:30preglowthis sounds pretty weird...
02:23:48 Join CheeseBurgerMan [0] (n=BurgerBo@63.150.80.142)
02:24:22 Join }{CBM}{ [0] (n=BurgerBo@63.150.80.142)
02:24:39 Quit }{CBM}{ (Remote closed the connection)
02:26:27DocBrownLinusN: Booting iriver via Rockbox bootloader also works with the same trick when pressing REC and PLAY
02:30:19LinusNDocBrown: so the hold switch is the problem?
02:30:56DocBrownLinusN: Looks like it... its completely reproducable
02:31:03CassandraGood grief. A Sam and Max computer game has risen from the dead again. http://www.gamecloud.com/omg.wtf?lolz=2459
02:32:21RotAtoRhaha, ".../omg.wtf?lolz=..."
02:32:51RotAtoR..but that's very good news, i loved sam & max hit the road
02:34:09preglowgreat game
02:34:45RotAtoRpreglow: any progress lately on the musepack updates/fixes? ;)
02:35:12DocBrownLinusN: Anything I can check for you in order to understand what happens?
02:35:38LinusNi'm thinking...
02:35:58preglowRotAtoR: none, i got the musepack people to do a couple of updates, and i'm waiting for them
02:36:29RotAtoRoooh, fun, maybe they'll release sv8 while they're at it
02:36:32RotAtoR:)
02:36:33preglowahaha
02:36:34preglow"yes"
02:36:52preglowno, just a couple of things i had them fix
02:37:05preglowto minimise the size of the deocder struct
02:37:30preglowas to the buzzing problem, i have no idea what's causing it
02:37:43preglowbut i haven't been able to debug it either, my libmusepack lies in shambles for now
02:37:54RotAtoRhmm, ok, thanks!
02:38:11LinusNDocBrown: will you stay here for a while longer?
02:39:05DocBrownLinusN: I could, if that helpful.
02:39:16DocBrownLinusN: that's
02:39:22lostlogicis skippiness during applause in OGG/Vorbis unique to H3x0, or has it also been reported on H1x0
02:39:24LinusNi'll prepare a test bootloader for you
02:39:38DocBrownLinusN: OK
02:39:38preglowlostlogic: ehh???
02:39:39JdGordonwhen usb support if done on the h300 will u be able to listen to music while using it as a portable hdd?
02:39:46LinusNlostlogic: most likely a h300 only bug
02:39:47preglowJdGordon: no
02:39:56JdGordon:( ok
02:40:00LinusNreboot, brb
02:40:03 Part LinusN
02:40:08lostlogicJdGordon: control of the Hard Drive must be handed over to the host system for UMS mode.
02:40:10preglowlostlogic: skippiness during applause?
02:40:33lostlogicpreglow: and other times. I'm using Q6 or Q7 oggs, and applause in live albums regularly triggers the problem.
02:40:37linuxstbJdGordon: You can listen to the radio though...
02:40:50preglowlostlogic: it's probably a high bitrate issue
02:41:01preglowclapping is pretty wide bandwidth
02:41:06lostlogicpreglow: yes, that was my figuring.
02:41:07 Join [-CBM-] [0] (n=BurgerBo@63.150.80.142)
02:42:10lostlogicpreglow: CPU or disk buffer limitation do you think? CPU probably?
02:42:19preglowlostlogic: cpu
02:42:22preglowlostlogic: but still a bit weird
02:42:38preglowlostlogic: if you could have a look at the audio debug thread screen while it happens, that would be helpful
02:42:54lostlogicpreglow: will do.
02:43:01 Join LinusN [0] (n=linus@labb.contactor.se)
02:43:09preglowjust note what state the top meter is in when it skips
02:43:12preglowi bet it's at the bottom
02:43:38preglowin which case it's pretty strange anyway
02:43:40JdGordonhow the heck do u get cards out of the pack n solataire?
02:45:32dropandhoany tick-fix experts in the house?
02:46:13lostlogicyeah, the PCM buffer fluctuates wildly during the skippy sections. the codec buffer stays fairly stable (fills up gradually toward the long anti-skip buffer I have set)
02:47:05preglowlostlogic: fluctuates how? it does hit bottom as it skips, yes?=
02:47:43lostlogicit never shows it at the bottom, but yes, just before a skip, it's on a downward trend, and is at the lowest point that I see displayed just as it skips.
02:48:10lostlogicI'm sure it's just because the H3x0 is still only at 90mhz not 124 or whatever the H1x0 is at.
02:48:37LinusNDocBrown: http://linus.haxx.se/H300.zip
02:49:02DocBrownLinusN: Okay
02:49:22preglowlostlogic: if so, then tremor performs lousily
02:49:26preglowand needs to be chastised
02:50:11ashridahit hasn't had any optimisation work in a while iirc.
02:50:15preglowbahhh
02:50:19ashridahonce it got realtime, it kinda got left
02:50:22lostlogicpreglow: IIRC tremor was hard to get to realtime, so someone should chastise it ;)
02:50:40ashridahbut yeah, improving it for battery's sake would be nice :)
02:50:43preglownot _that_ hard
02:50:47preglowthere's still more to be done for it
02:51:12 Join San [0] (n=Test@212.2.163.159)
02:51:53lostlogicwhat kind of optimizations would be doable do you expect?
02:52:12preglow1. there's more iram to be used, 2. faster imdct
02:52:24lostlogicplayijng oggs, I'm running 94% boost ratio
02:52:49preglowlostlogic: that's very bad
02:52:51Kyl3Sounds SO much better
02:53:04preglowKyl3: what does?
02:53:08lostlogicyes. I'll see if I can dig up some time to poke at it.
02:53:23Kyl3rocbox
02:53:36Kyl3rockbox* i dunno why, but it sounds better
02:53:41preglowhaha
02:53:48preglow'course it does
02:54:02linuxstbKyl3: Which codec do you use?
02:54:05lostlogicit does, iRiver firmware artificially bass boosts, makes it unlistenable with my subwoofer even _on_
02:54:18Kyl3now where are those crossfade settings?
02:54:26DocBrownLinusN: Success!!! Rockbox and iriver original via bootloader now boots as intended
02:54:48Kyl3I use MP3 192K and OGG 6/7
02:55:06ashridahpreglow: i'm pretty suve i've got some q10 oggs that have given rockbox issues even on the H140 when running without cpu caps
02:55:11preglowlinuxstb: oooh, what did you do?
02:55:17Kyl3LinusN, any idea on fixing the "dead" bug?
02:55:29preglowashridah: and it would be a small surprise if they did
02:55:53preglowtremor is hard to optiomise properly
02:55:54linuxstbpreglow: I guess you meant to oooh at LinusN
02:55:55DocBrownLinusN: Ooooops, now its gone...
02:55:55preglowit's got no limits
02:56:01preglowlinuxstb: i guess i did
02:56:27preglowargh, btw
02:56:37preglowi've unmasked all interrupts i can find
02:56:40LinusNDocBrown: no go?
02:56:42DocBrownLinusN: Works again... I'll investigate that...
02:56:54preglowand inited keyboard code, etc
02:57:00preglowstill no bloody interrupts
02:57:15 Quit CheeseBurgerMan (Read error: 110 (Connection timed out))
02:58:40Kyl3The new bootloader sort of fixed my "dead syndrome" problems but it still happenes occasionally
02:59:08Kyl3but it restarts back immedatly after i reset it
03:00
03:05:01preglowYES!
03:05:43Benacool?
03:05:51preglowgot timer interrupt going
03:06:14***Saving seen data "./dancer.seen"
03:06:38DocBrownLinusN: Preliminary result: Everything works fine as long as i do NOT touch the HOLD button now
03:06:57LinusNi don't get it
03:07:05LinusNwhy touch the hold button?
03:07:25DocBrownSorry, "move" the hold button
03:08:12DocBrownEven if i set the hold button to On and back to Off with the unit being completely OFF will cause the black screen error
03:08:21LinusNi see
03:08:36DocBrownSo it seems like an electrical problem?
03:09:14linuxstbpreglow: Hurray...
03:09:41LinusNDocBrown: i'm puzzled
03:10:56preglowlinuxstb: at least i'm not doing too much wrong
03:12:00DocBrownLinusN: Me too. The bootloader always starts if the Hold button is set. And then turns the unit off of course. However, the trick from half an hour ago still works
03:16:29 Quit ashridah ("gone.")
03:16:54preglowhahah
03:16:57preglowneato, neato, neato
03:17:26preglowseems interrupt ack works as well
03:17:45DocBrownLinusN: Sometimes it boots by only pressing play, sometimes it doesn't. And now i habe a new behaviour. It stopped at the "noise" pattern that appears first for some
03:18:06DocBrownmilliseconds when starting the bootloader
03:18:49linuxstbpreglow: If you try to get the current_tick working, then don't forget that the ATA driver (and maybe other things) rely on it.
03:20:07preglowlinuxstb: shouldn't be hard to have a current_tick working now
03:20:14preglowjust need to clean up this awful mess
03:21:29preglowtick counter is 100hz, yes?
03:21:37preglowor is it 1000 ?
03:21:38linuxstbYes.
03:21:40preglowokies
03:21:42linuxstb100Hz
03:21:51preglowthen i've actually got it running at the correct frequency already
03:26:12DocBrownLinusN: Really strange. Now it won't boot anymore by just pressing PLAY. But I swear it did for some times. The "HOLD button flipping" trick still works however
03:27:49LinusNi replaced the h300.zip with a new one, try it
03:29:17DocBrownLinusN: Okay. I Have just reflashed the old h300.zip and it works again. Really strange. Will try the new one now
03:32:42 Join skorpyon [0] (n=skorpyon@i216-58-27-81.cybersurf.com)
03:41:42DocBrownLinusN: Strange.... Flipping trick still works. Direct boot doesn't. Then I put my PSU because of the battery drain from flashing, spinning etc... Now it boots (very rare occasions), mostly black
03:42:06DocBrownscreen or flickering white screen and then off.....
03:42:36LinusNweird...
03:44:01preglowi'm out, gnight
03:44:30LinusNi have to sleep too
03:44:49LinusNDocBrown: could you document your findings in the wiki, so i won't forget?
03:45:22 Join jaydpb [0] (n=4432b355@labb.contactor.se)
03:45:24DocBrownwas about to say the same. Will investigate further tomorrow
03:45:59DocBrownNever worked with a wiki before. Think I need to register and stuff?
03:46:32LinusNyes, it's easy
03:47:11DocBrownJust booted without PSU... strange. There semms to be a time constant in it... electrical stuff i think
03:47:24DocBrownWill document that tomorrow, okay?
03:48:19DocBrownOn the iriver boot page?
03:48:26LinusNdo that
03:48:37LinusNnite all
03:48:42 Part LinusN
03:48:50DocBrownquit
03:48:54 Quit DocBrown ("CGI:IRC 0.5.7 (2005/06/19)")
03:50:18 Quit phaedrus961 ("Leaving")
04:00
04:01:16lostlogicman, I haven't worked in C in such a farking long time, let alone C + asm. ALL THE JAVA, IT ROTS MY BRAIN!
04:16:07 Join Membrillo [0] (n=cb59493a@labb.contactor.se)
04:17:17lostlogicdamn, preglow is one smart fucker.
04:17:59 Quit Membrillo (Client Quit)
04:19:44 Quit Benacool ()
04:27:12 Join tvelocity [0] (n=tony@ipa241.7.tellas.gr)
04:27:49 Quit DreamTactix291 (Read error: 110 (Connection timed out))
04:28:17 Quit [-CBM-] (Read error: 104 (Connection reset by peer))
04:29:08 Quit dropandho ()
04:31:05 Join CheeseBurgerMan [0] (n=BurgerBo@63.150.80.142)
04:31:26 Join crown292 [0] (i=clunk@CPE000c6e94cf09-CM001225d870de.cpe.net.cable.rogers.com)
04:33:22 Quit skorpyon (Read error: 110 (Connection timed out))
04:34:19lostlogicit's not making good use of the PCM buffer, it never fills to even half way.
04:36:38linuxstblostlogic: Is this vorbis on a h300?
04:36:51lostlogicyeah
04:36:58linuxstbWhat's the boost ratio?
04:37:14lostlogicnever mind, it's gotten higher, I was just using too torturous of music.
04:37:24lostlogic94-96%
04:37:29 Quit Kyl3 (Read error: 104 (Connection reset by peer))
04:37:30linuxstbIt probably just can't decode fast enough to fill it any more.
04:37:35lostlogicyeah, noted
04:38:36lostlogicdo you know how I would go about testing code changes for optimization aside from just watching boost ratio on the same track?
04:39:56linuxstbThat's the best (only) test we've got.
04:40:18 Join webguest69 [0] (n=45dda986@labb.contactor.se)
04:40:27lostlogickk
04:40:44linuxstbTwo things to be aware of - disk filling and backlight fading will also boost.
04:40:58lostlogicfigured
04:41:06lostlogic*sets backlight to always on*
04:41:28linuxstbOr set the fade to 0s
04:41:52lostlogicalways on will do
04:42:22lostlogichmm... at least one of the skips happens after the nearly gapless track transition, like it forgets to resume disk fill.
04:44:09 Quit webguest69 (Client Quit)
04:44:20lostlogicdoes it always try to buffer all the way up to the size of RAM (30m) for the codec buffer, or is that configurable somewhere?
04:47:15linuxstbI don't think it's configurable.
04:48:11linuxstbI'm not 100% sure what the buffering strategy is, but I think it first fills the buffer with about 2MB of data, and then when the low watermark (about 512KB) is reached, it fills up to the full buffer (about 29MB I think)
04:49:15lostlogicmakes sense
04:50:38lostlogicand it does that for each track, so at EOT, if the codec buffering thread can't get the CPU, there is a time where the buffer is deliberately allowed to run down in order not to waste energy buffering when the next track will just be skipped any way.
04:51:18lostlogicdave matthews band live albumbs are a wonderful ogg torture chamber it seems.
04:53:25linuxstbNo, it doesn't do it for every track - it fills the buffer with the upcoming tracks.
04:53:40 Join Jungti1234 [0] (n=jungti12@211.248.102.131)
04:53:41 Quit jaydpb ("CGI:IRC (EOF)")
04:53:59linuxstbBut because vorbis is very slow, Rockbox may struggle to read the data from disk as well.
04:54:11lostlogicok, that was my second theory, thanks!
04:54:40lostlogic4 tracks played, 99% boost, backlight always on.
04:55:22lostlogicThis is the first time I've dug into the rockbox code, I must say I'm very impressed with it.
04:57:00 Join Paul_The_Nerd [0] (n=paulthen@cpe-66-68-93-2.austin.res.rr.com)
04:59:23 Quit Jungti1234 ("Http://www.ZeroIRC.NET Zero IRC Ver 2.8")
05:00
05:00:19 Quit hardeep ("[BX] Silly faggot! mIRC is for kids!")
05:05:34grant_any thing new been commited today?
05:05:45BBublook at the page :)
05:05:57grant_linkage?
05:06:08BBubhttp://www.rockbox.org/ right on the main page
05:06:10BBubat the bottom
05:06:17***Saving seen data "./dancer.seen"
05:06:18grant_thanks :)
05:06:44Paul_The_NerdHehehe
05:06:50BBub"Enabled playback speed adjustment support for H1x0. Modified the resampler to do both channels in one pass." i guess that is the nicest feature
05:06:55BBubsomeone on the forums asked for it
05:07:45grant_Nice, nothing too major for the h3x0 but good progress
05:07:53grant_the team are doing an excellent job :)
05:08:05BBubyes, they are making progress every day now
05:08:25lostlogicHmm... if the codec buffer reaches some large size, it stops filling to save battery, drains down to threshold (512k) and starts filling again. so very smart. now if only ogg could be sped up enough to give it time to actually fill the buffer in reasonable time.
05:08:48BBublostlogic: on h3xx or h1xx?
05:08:56lostlogich3xx
05:09:01lostlogicbut it would effect both.
05:09:02Paul_The_Nerdlostlogic: The main problem is that the H3x0 isn't stable running at the full processor speed right now.
05:09:10BBubyou think so?
05:09:18BBubafaik the h3xx still runs at only 94 mhz
05:09:21grant_Yeah I'm so delighted with the speed of things. Rockbox is useable for mp3 playing right now and it's fun.
05:09:34Paul_The_NerdIt's useable for up to Q7 or Q8 oggs too, I understand.
05:09:44lostlogicPaul_The_Nerd: sure, but even so, it takes 100% boost to decode q7 ogg at 90mhz, then it takes 75% boost to do it at 120
05:09:47lostlogicthat's not a very good ratio.
05:10:10grant_I understand that the h3x0 version is pretty much just a port of the h1x0 rockbox and that's why it was so fast yet not adjusted correctly as of yet. Correct?
05:10:45BBubi think so, as the hardware is identical in many parts
05:10:50lostlogicgrant_: all of the various rockboxes are built from the same codebase
05:11:21lostlogicbut yes, H1 and h3 are very similar
05:11:26grant_Ah thanks. :)
05:11:27Paul_The_Nerdlostlogic: It's hard to say what "good" is. The ogg decoding had not so great battery life under iRiver too, so it's hard to say what the fastest it could be done *is*
05:11:45grant_Is there much that can be contributed without knowing coding and such? besides testing and donating?
05:12:10BBubnot really
05:12:19lostlogicgrant_: WPSs, otherwise, not that I know of... maybe when they need graphics for color bitmap WPSs...
05:12:19BBubyou could help out others on the forums though :)
05:12:56lostlogicPaul_The_Nerd: well I'm optimization hunting, I think I've found a couple of functions that belong in IRAM so far, trying to find code inefficiencies too, but smarter people than I have already been over it.
05:13:01grant_Thanks, oh btw. Are the bleeding edge releases equal to the cvs commits?
05:13:06lostlogicyes
05:13:20lostlogicbleeding edge is built before the commit is sent to the front page, or so someone said yesterday
05:13:44grant_ok cool, just wasn't sure if I needed to use CVS to get these latest commits.
05:13:45Paul_The_Nerdloslogic: Honestly, from my understanding, the Vorbis decoder hasn't been optimized much at all since it hit the 'realtime' mark, so you might be surprised as to what you can do.
05:14:11lostlogicstill, preglow is a smart guy, and he is who got it to realtime, so that means that all of the 'obvious' optimizations are done.
05:14:17lostlogicbut I'm working on it :)
05:14:32Paul_The_Nerdlostlogic: And that's all anyone can ask.
05:14:55BBubbut im very happy about mpc playing in realtime
05:14:59 Join Benacool [0] (i=Benacool@modemcable035.68-202-24.mc.videotron.ca)
05:15:04BBubthats what i really missed on the iriver
05:15:41Paul_The_NerdThere's not really any hardware out there that supports MPC is there?
05:16:13BBubi would bet rockbox on iriver is the only one so far ;)
05:17:03BBubi also never heared of other players supporting flac or wavpack
05:17:18BBubso you got an exclusive piece of h/w with rockbox :)
05:17:52Paul_The_NerdFLAC is actually supported elsewhere
05:17:56Paul_The_NerdI think on some by Cowon
05:18:03BBuboh, i didnt know
05:18:25Paul_The_NerdYeah
05:18:37Paul_The_NerdBut Rockbox has a pretty damn fast FLAC implementation
05:18:40Paul_The_NerdIt doesn't even boost.
05:18:50Paul_The_NerdAnd that's with replaygain and crossfeed enabled.
05:18:58 Join clunk- [0] (i=clunk@CPE000c6e94cf09-CM001225d870de.cpe.net.cable.rogers.com)
05:18:59BBubyep, they changed the codec a few weeks ago to a very fast one
05:19:04Paul_The_NerdSo I saw
05:20:22 Quit crown292 (Read error: 104 (Connection reset by peer))
05:24:08 Quit HCl (Remote closed the connection)
05:24:09 Join HCl [0] (i=hcl@titania.student.utwente.nl)
05:25:34lostlogicanyone know if -funroll-loops is included in -O?
05:28:49lostlogicsnot
05:29:27linuxstblostlogic: The easiest way to get more speed out of vorbis is to move more code and data (but mainly data) into iram.
05:29:40linuxstbYou have 48KB available and should be using it all.
05:30:47lostlogiclinuxstb: yeah, that was my original hunt, but I'm seeing a lot of places where loops of small constant size are used in oft-called code, and wondering about unrolling them.
05:33:44linuxstbI would be surprised if that gave any noticable improvement. I would start with iram, and then look at the rest.
05:34:06 Quit Benacool (Read error: 110 (Connection timed out))
05:34:10lostlogic*nod* how to track IRAM usage from adding a given function? will compile time or runtime fail?
05:34:38linuxstbLook at the build-dir/apps/codecs/vorbis.map file
05:34:43lostlogicthanks!
05:35:31linuxstbBut yes, if you try to use too much, the linker will give you an error.
05:37:14 Quit tvelocity ("Leaving")
05:53:36lostlogicby jove, I think it's getting better!
05:53:56Paul_The_NerdHehehe
05:54:29lostlogic96% boost on a song that was 99 with skips last time around... took 8k of IRAM to do it though :-\
05:55:31Paul_The_NerdTo quote linuxstb "You have 48kb available and should be using it all
05:56:01lostlogicyeah, only have 8k left, important to only use it for the _most_ often referenced symbols...
05:56:09Paul_The_NerdGotcha
06:00
06:04:40 Join DreamTactix291 [0] (n=DreamTac@adsl-149-140-175.bna.bellsouth.net)
06:04:53lostlogicDT: how's battery life on ogg on the H1x0?
06:07:53 Quit grant_ ("Leaving")
06:19:40DreamTactix291lostlogic: Vorbis battery life is at least as good as the iRiver firmware in my experience
06:19:46DreamTactix291so like 12-14 hours
06:20:26lostlogicwow, they must not have really optimized the hell out of their vorbis codec either.
06:20:47lostlogicBAH, how do I lose performance this time!?
06:21:22DreamTactix291i don't notice any real degradation
06:21:31DreamTactix291though to be honest i haven't done any serious battery test in a long time
06:21:36lostlogicnod
06:21:53DreamTactix291both mp3 and vorbis decode really well
06:23:19lostlogic*sigh* at 90mhz, it still can't decode the beginning of Dave Matthews - Live at Red Rocks (applause) smoothly
06:24:07 Quit HCl (Remote closed the connection)
06:26:03 Join HCl [0] (i=hcl@titania.student.utwente.nl)
06:29:00DreamTactix291applause = huge bitrate spikes
06:29:24lostlogicyeah, hehe, and sometime in the next few days, I'm going to get this thing to deal with it. :-P
06:29:25DreamTactix291probably in the 500-700kbps range
06:30:54lostlogicmoving the lookup tables that vorbis uses to compute floors didn't seem to impact performance nearly as much as I thought it would
06:31:22lostlogicmy initial move of several of the oft used functions ot IRAM had more impact... *sigh* I obviously have a lot to learn.
06:33:54 Nick CheeseBurgerMan is now known as CBM-away (n=BurgerBo@63.150.80.142)
06:34:04Paul_The_NerdEven if you do have a lot to learn, trying out different ideas and seeing how much they help *does* qualify as "learning"
06:34:51lostlogic;)
06:35:04lostlogicPaul_The_Nerd: you sure are the encouraging type
06:35:18Paul_The_Nerd:-P
06:35:26Paul_The_NerdWhat can I say, I have a talent. ;-)
06:43:15lostlogicok, last rendition for the night, residue and floor calculations should be almost entirely in IRAM now.
07:00
07:06:19***Saving seen data "./dancer.seen"
07:10:26 Join perplexity [0] (n=joust@83.110.209.107)
07:14:18 Quit RotAtoR ()
07:23:53 Join B4gder [0] (n=daniel@static-213-115-255-230.sme.bredbandsbolaget.se)
07:24:49 Join ModernExecutive [0] (n=bob@pool-141-152-178-224.norf.east.verizon.net)
07:25:31ModernExecutivehi yall
07:28:13Paul_The_NerdThat's y'all, durnit. :-P
07:28:46 Quit clunk- ("fractal2 mirc script (ver 1.0betar2) http://fractal2.net")
07:36:22 Join Jungti1234 [0] (n=jungti12@58.77.81.144)
07:53:11 Join Bger [0] (n=Bager@83.222.160.88)
07:53:16Bgermorning :)
07:53:21Paul_The_NerdMornin'
07:54:06ModernExecutivehi'
07:54:44 Quit DJDD__ (Read error: 110 (Connection timed out))
07:56:56 Join _FireFly_ [0] (n=FireFly@p54A46B29.dip.t-dialin.net)
07:57:41 Quit ModernExecutive ()
08:00
08:04:45 Quit lostlogic (Read error: 110 (Connection timed out))
08:20:39 Quit Jungti1234 ("bye")
08:38:23 Quit _FireFly_ ("Leaving")
08:38:31 Join lostlogic [0] (n=lostlogi@node-4024215a.mdw.onnet.us.uu.net)
08:46:20 Join Zagor [0] (n=bjst@194-237-150-170.customer.telia.com)
08:50:11 Join Febs [0] (n=4363c602@labb.contactor.se)
08:57:43 Quit perplexity (Excess Flood)
08:58:42 Join perplexity [0] (n=joust@83.110.209.107)
08:58:54perplexity<Naive question> Has anyone thought about what might prevent us from compiling rockbox with -pg (profiling) enabled and starting to look at some hotspots? </>
08:59:12 Join einhirn [0] (i=Miranda@bsod.rz.tu-clausthal.de)
08:59:58B4gderthe "hotspots" are in the codecs
09:00
09:00:03B4gderwhere else could they be?
09:00:48perplexityPrimarily I'm interested in instrumenting the tremor codec yes.. I'm just throwing ideas in the air at the moment..
09:02:13perplexityin any case.. compiling with -pg requires the removal of -fomit-frame-pointer which then causes us to run out of registers in some other code.. so I need to think about it a bit harder..
09:05:35 Join JdGordon41 [0] (n=Miranda@c211-28-93-42.smelb1.vic.optusnet.com.au)
09:06:22***Saving seen data "./dancer.seen"
09:09:15 Join amiconn_ [0] (n=jens@p54BD5557.dip.t-dialin.net)
09:13:15JdGordon41haha.. mp3 ripping + dvd encoding are killing my comp :D
09:14:29 Quit JdGordon (Nick collision from services.)
09:14:31 Nick JdGordon41 is now known as JdGordon (n=Miranda@c211-28-93-42.smelb1.vic.optusnet.com.au)
09:18:30 Quit amiconn (Nick collision from services.)
09:18:30 Nick amiconn_ is now known as amiconn (n=jens@p54BD5557.dip.t-dialin.net)
09:23:48 Quit DreamTactix291 (Read error: 110 (Connection timed out))
09:28:06 Join ashridah [0] (i=ashridah@220-253-123-200.VIC.netspace.net.au)
09:41:45 Join _Vladoman [0] (n=Vladoman@p54A7D3AF.dip.t-dialin.net)
09:56:43 Join Jungti1234 [0] (n=jungti12@58.77.81.144)
09:58:51 Quit Vlad0man (Read error: 110 (Connection timed out))
10:00
10:03:36 Join sublah [0] (n=jasonmd@dsl-203-113-239-74.SA.netspace.net.au)
10:10:43 Join baobab68 [0] (n=ca3421eb@labb.contactor.se)
10:10:50baobab68hi all
10:11:14 Join ender` [0] (i=ychat@84.52.165.220)
10:11:56markunmorning
10:13:01Bgermorning, markun, baobab68
10:14:22Jungti1234markun
10:15:10markunhi Jungti1234
10:15:17Jungti1234morning
10:15:39Jungti1234are you busy?
10:15:40Jungti1234i need you
10:16:20markunI emailed samsung to ask what is different bout their S3C2440AL-30 (CPU of the Gigabeat) compared to the S3C2440A of which I have the datasheet
10:16:42markunInstead of answering my question they send me an old version of the datasheet of the S3C2440A.. very strange
10:16:56 Part sublah
10:16:59Jungti1234hmm..
10:17:27Bgermarkun: there's probably not a big diff :)
10:17:52Paul_The_NerdIt's also possible that they misunderstood the question and thought you were requesting information about *that* one?
10:17:53markunmaybe not, I was just wondering what is the difference was because I couldn't find it on the website.
10:18:16ashridahmy guess is packaging of the chip
10:18:20markunThe datasheet they send me was revision 1. I have revision 10 from their website :)
10:18:40markunsent
10:18:44Bgerhehehe
10:18:59BgerWhy don't send them another e-mail ?
10:19:08markunMaybe I will
10:19:14JdGordonany1 know where to dl roms for rockboy?
10:19:15Bgerjust do it (tm)
10:19:28BgerJdGordon: torrents, p2p ... :)
10:19:40t0masWett? are you around?
10:19:45JdGordonwhat roms work? just game boy? or game boy advanced?
10:19:59BgerJdGordon: gameboy & gameboy color
10:20:14JdGordonok
10:20:44Jungti1234markun: Did you get gigabeat?
10:20:53markunI didn't check my mail yet
10:22:02markunno, didn't receive it yet
10:26:49baobab68for those who have used replaygain, anyone got recommendations for a db level? i was going to try 89db, seems to be the commonly selected value.
10:27:16baobab68fantastic that rockbox supports it, can't wait to hear how it works
10:29:33markunbaobab68: yes, 89 is now the default
10:30:48baobab68thanks markun, will try that value. interesting to see if it has any effect on CPU usage. My h320 is coping fine with my Q0.7 Ogg's but not sure what will happen with gaining...
10:31:02markunnot much will happen I think
10:31:02baobab68(yes I did say 0.7) :-)
10:31:18baobab68cool.
10:31:36ashridahvorbis supports fractional qualities?
10:32:06markunbaobab68: damn :)
10:32:10perplexityI would have assumed 0.7 would be a pretty low quality.. What sort of average bitrates are you getting with that?
10:32:22baobab68it does in media monkey. the slider is from -1 to 10, in increments of 0.1
10:33:10baobab68perplexity: 70-90kbps. Sounds fine to me, fine for rocking around the city.
10:34:01baobab68though dreamtactix and Paul_The_Nerd disagree :-)
10:34:32baobab68anyway, thank you for the info markun. why you say "damn"? wish you had transcoded like me? (i doubt it!)
10:34:33 Quit CoCoLUS (Read error: 104 (Connection reset by peer))
10:34:34 Join CoCoLUS [0] (n=coco@h081217139221.dyn.cm.kabsi.at)
10:34:39Paul_The_NerdNaah, I'm fine with lossy music at levels below the transperency point, especially if there's a lot of background noise or space is at a premium.
10:34:58Paul_The_NerdIt's the fact that they're transcoded from 64kbps WMAs that makes me cry.
10:35:01Paul_The_Nerd:-P
10:35:03perplexityOk.. I have my collection encoded at several values for testing.. i) -b 256, ii) -q 6, iii) -q 8.. I can't imagine going that low
10:36:06baobab68i can't stand 'burbling noises' but i've spotted very few in my sample 1500+ Ogg's. Other artifacts annoy me, but aotuv dll's have done a great job....
10:36:29baobab68my poor pc has been very busy since RB came out for the H300.
10:36:43perplexityMine also.. takes me 12 hours to re-encode my collection :)
10:37:00markunbaobab68: I said damn because I encoded everything at Q3 and I thought that was on the low side already :)
10:37:03baobab6812 *hours* - mine has been like 3 *days*
10:37:19perplexityI only have 111 cd's to encode
10:38:16baobab68well thanks for the replaygain info, everyone, got to run and catch up on some telly - got some arrested development and queer as folk to catch up on....
10:38:19crwl-b 256 and -q 8 is the same thing, btw
10:38:19Bgernice number
10:38:34baobab68just remember, "0.7" :-)
10:38:37 Part baobab68
10:39:06JdGordonis rockboy actually usable?? mortal combat was working ta like 1fps
10:39:06perplexitybrad@srv:/raid3$ find cds-ogg-Q6/ -name "*.ogg" | wc -l
10:39:06perplexity1468
10:40:34Paul_The_NerdRockboy is useable for things like turn based RPGs primarily. It's not at full speed yet, as it kinda got abandoned before being optimized fully and/or finished (there are some evident bugs)
10:40:56JdGordon:( damn
10:41:42markunA lot of people are trying out rockboy now that rockbox runs on the H3x0. Maybe one of them will pick it up.
10:41:44perplexitycrwl: Perhaps as an average, but Q8 will make bigger than 256k files on particularly difficult tracks.. it's an experiment in any case.. I have a script that re-encodes the whole collection in a batch anyway
10:42:17crwlperplexity, no, it should be exactly the same thing
10:42:35crwl-b xxx will just map to some -q value... or the other way around
10:43:01perplexityOh.. well then I should be able to diff the ogg encoded at -q8 and -b256 and they should be the same size exactly..
10:43:10crwl-b 64 is -q 0, -b 128 is -q 4 , -b 192 is -q 6, etc
10:43:21crwlit's the same VBR encoding mode nonetheless
10:44:04perplexityindeed.. the file sizes are very close
10:44:23perplexityin which case -Q8 is going to skip on the h3x0 at this point.. (as 256k does)
10:44:40crwlyou could always use FLAC ;)
10:44:46 Join webguest92 [0] (n=d5c88e91@labb.contactor.se)
10:45:05 Join LinusN [0] (n=linus@labb.contactor.se)
10:45:20webguest92hello all
10:45:27perplexityI already have all my albums as flac also, I'm just getting my head around metaflac to be able to set the tags and skip points so I can skip tracks.. (All my flacs are a single file per album)
10:45:31markunmorning LinusN and webguest92
10:45:37webguest92morning
10:45:41 Quit ghode|afk (Read error: 110 (Connection timed out))
10:46:19webguest92does anyone know if it's still unclear whether or not the ipod 5g will run linux/rockbox?
10:47:00markunwebguest92: when ipod linux figured out how to run on 5g, rockbox can also be ported
10:47:25webguest92i'm on the verge of exchanging my 4g for a 5g...
10:47:45B4gderdig up that disassembler and plunge!
10:48:15B4gderarm assembler is nice to read too ;-)
10:48:28webguest92so, there's nothing new about the 5g's then, i take it?
10:48:52B4gderwebguest92: I would assume #ipodlinux is a better place to ask for such details
10:49:07webguest92probably! thanks a lot.
10:49:09markunwebguest92: they placed a video decoder chip between the cpu and the lcd I think. That's the biggest problem.
10:49:22webguest92aha
10:50:57markunB4gder: now that amiconn's self extracting rockbox works, could you do some of your makefile/perl magic?
10:51:16B4gderyes, as soon as I get some info on what's necessary
10:51:37markunamiconn: are you there?
10:55:42linuxstbwebguest92: I saw a first version of an LCD driver for the 5G LCDs checked into the ipodlinux CVS yesterday. But there was a warning attached saying it was very slow.
10:57:48linuxstbwebguest92: If you have any desire to run ipodlinux or Rockbox in a useful way, I would avoid the 5G ipod.
11:00
11:02:34webguest92linuxstb: why?
11:02:58 Join Polo_o [0] (n=polo_o@82-69-160-166.dsl.in-addr.zen.co.uk)
11:03:22webguest92linuxstb: i mean, isn't it only a matter of time until the issue with the video decoder chip is figured out?
11:04:04linuxstbwebguest92: Hopefully - the ipodlinux people are making good progress. But you can never be 100% sure.
11:04:36 Join DreamTactix291 [0] (n=DreamTac@adsl-149-140-175.bna.bellsouth.net)
11:05:34B4gdera good opportunity to dive in and do some low level digging!
11:05:43webguest92linuxstb: hmm... that's a tough one... I can get a 5G for free, but then I have to give up my 4G...
11:05:52linuxstbI'm just saying that if you want to use ipodlinux or Rockbox, you are much better off sticking with your 4G. If you are happy with Apple's firmware, then upgrade to the 5G.
11:06:24***Saving seen data "./dancer.seen"
11:06:50linuxstbBut I agree with B4gder - best option is to get the 5G and help the cause...
11:06:53*LinusN spots a combined-bitmaps patch in the tracker
11:07:03LinusNlovely
11:07:13B4gderI bet amiconn jumps of joy ;-)
11:07:25LinusNi do to :_)
11:07:29B4gderhehe
11:07:42LinusNloading a wps takes quite a while nowadays
11:07:50B4gderyes
11:08:30*JdGordon is bored.. anythign not so complicated i can help with?
11:09:12B4gderI could use another cup of coffee
11:09:17B4gder:-)
11:09:27JdGordoncoming guvna
11:09:54linuxstbJdGordon: You could make some of the other plugins more h300 friendly.
11:10:10JdGordonwhich r in need?
11:10:15LinusNJdGordon: how about a color bmp loader?
11:10:16B4gderand add "background" image support to all screens! ;-)
11:10:26linuxstbJdGordon: I've no idea, probably most of them.
11:11:52B4gderand we need color support in the X11 sim
11:11:59LinusNa larger status bar on the bigger lcd:s
11:12:27LinusNcolor support in x11 sim, yeah
11:12:29JdGordonthe small status bar is good..
11:15:02webguest92linuxstb: i wish i could
11:16:13linuxstbJdGordon: OK, then a status bar with configurable size :)
11:16:56JdGordonye, maybe
11:18:53B4gderbtw, I'm considering making it a requirement to build the tools only from make in a build directory - to be able to do specific builds for specific targets. Would that be bad to people?
11:19:59B4gderbut now lunch!
11:23:14 Join linuxstb_ [0] (n=linuxstb@i-83-67-212-170.freedom2surf.net)
11:23:14 Quit linuxstb (Read error: 104 (Connection reset by peer))
11:23:40 Nick linuxstb_ is now known as linuxstb (n=linuxstb@i-83-67-212-170.freedom2surf.net)
11:32:33 Quit DreamTactix291 (Read error: 110 (Connection timed out))
11:36:20markunB4gder: There was a Gigabeat F60 for sale on ebay for about they price I payed for my F40. Too bad I couldn't reach you in time.
11:38:41linuxstbmarkun: Stop that - B4gder's buying an ipod :)
11:38:51markunno he isn't :)
11:38:58markunIs he?
11:39:14LinusNhe's buying an x5!
11:39:18LinusN:-)
11:39:40Paul_The_NerdJust had to throw in another one?
11:39:46markunCan't he just buy all three to keep us all happy?
11:41:21linuxstbmarkun: Still no sign of your gigabeat yet then?
11:41:26 Join Bgr [0] (n=Bager@83.222.160.88)
11:41:28markunno :(
11:41:35markunfucking snow..
11:41:35linuxstb(or the Linux source from Toshiba...)
11:42:02solexxfucking pl/sql!
11:42:11markunI received an outdated version of the S3C2440A datasheet from them today ;)
11:42:32linuxstbI read that. A strange response.
11:43:09*amiconn appears
11:43:14amiconnB4gder: u there?
11:43:14LinusNi love snow
11:43:25LinusNB4gder: is having lunch
11:43:31LinusNdamn completion
11:45:22perplexityGrr.. H3x0 build of X11 sim is bombing in the vorbis codec :(
11:45:35amiconnBackground support for all colour and greyscale targets shouldn't be hard
11:45:52amiconn...which (hopefully) very little slowdown
11:46:00amiconns/which/with/
11:46:16markunamiconn: B4gder wants to do the makefile and perl stuff, but you will have to tell him what exactly there is to do
11:46:18 Join Zak1392 [0] (n=zkeeping@CPE-139-168-223-188.sa.bigpond.net.au)
11:46:37Paul_The_Nerdamiconn: You mean a background image type thing?
11:46:41amiconnmarkun: Yes, 've read that
11:46:41LinusNnothing like the smell of unicode in the morning
11:46:54markun:)
11:46:56amiconnPaul_The_Nerd: yes
11:47:00markunWell, I can almost smell it now
11:47:08 Join muesli_- [0] (i=muesli_t@Bc1fa.b.pppool.de)
11:47:19muesli_-re
11:47:21LinusNit would be great to finally add it to cvs
11:47:30markunI was looking at freetype2, would freetype support be overkill?
11:47:41markuntrue type support..
11:47:41amiconnImho yes
11:47:43linuxstbI think so.
11:48:09linuxstbEven the ipodlinux people think freetype2 is overkill - they have used it in their new UI toolkit, but are going to remove it.
11:48:59markunok, then I'll not bother
11:49:24*amiconn actually used his Ondio FM today
11:49:35LinusNwow
11:49:36amiconn...and noticed that there is another not-quite-feature
11:49:42amiconn...for archos
11:50:05LinusNoh?
11:50:17amiconnThere's code to switch between preset and scan mode, but it's not 'wired' for archos
11:50:28LinusNmarkun: freetype2 in rockbox?
11:50:49markunLinusN: nah, forget what I said :)
11:50:56*LinusN mumbles something about the temperature in hell...
11:51:03amiconnI think of adding this
11:51:12LinusNamiconn: good
11:51:29amiconnI probably need to make it a menu option - too few buttons you know :/
11:51:37Zak1392hello all
11:51:42LinusNamiconn: a real pain
11:51:59amiconnFor the fm recorder, a button should be possible
11:52:03*LinusN looks for this "all" dude
11:52:24Zak1392hardy har har ;)
11:52:40amiconnThere's another feature - the pitch screen - which isn't available on Ondio
11:52:51Zak1392any progress for h300s linus?
11:52:54amiconnCould also be made a wps context menu option...
11:53:11Paul_The_NerdWhatever happened to that patch that let you load various FM presets files, so you could say load one for a city you're visiting?
11:53:37LinusNZak1392: not really, but we had an interesting session last night where we found out some details about the "no-boot-black-lcd" problem
11:53:41amiconnOh, and the Auto Scan feature has a button-related bug
11:53:55LinusNPaul_The_Nerd: coming...
11:54:00Zak1392what was the problem linus?
11:54:04LinusNamiconn: go fix
11:54:19LinusNZak1392: it seems to be a problem with sensing the ON buttons at boot
11:54:22amiconnIf I cancel the request to overwrite current presets with Off, it also drops me out of the radio screen
11:54:35LinusNoh
11:55:09Paul_The_NerdLinusN: Neat. I was just curious if it had maybe been rejected or something.
11:55:09Zak1392i love the music on rockbox. thanks for all your efforts ;)
11:55:35LinusNwe didn't make the music, you know ;-)
11:55:54*LinusN is sooooo funny today
11:56:02Zak1392unless you're a part of children of bodom ;)
11:56:16Paul_The_NerdYou do seem to be in a good humor today LinusN.
11:56:47amiconnLinusN: Could you check whether thsi also happens on the fmr?
11:57:10 Quit Bger (Read error: 110 (Connection timed out))
11:58:39Zak1392anybody here like children of bodom?
11:59:42Paul_The_NerdQuestion: Is it still going to be a matter of loading individual .fmr files? (IE, if it is, would it be welcome if I started converting the FmPresets wiki page into useable FMR files, with some sort of standard naming like Country.City.fmr (or in the case of USA Country.State.City.fmr)?
12:00
12:00:02 Join lamed [0] (n=d4b3395e@labb.contactor.se)
12:00:11lamedhello.
12:00:15markunAnyone willing to help Jungti1234 to compile rockbox in windows? He installed "Rockbox Development Kit" and downloaded the source with cvs, but now he can't find rbconf.
12:00:19lamedpreglow: around?
12:00:59lamedjungti1234, markum. rbconf is an alias for ../tools/configure\
12:01:10lamedsorry for last slash
12:01:52lamedjungti1234, just type at your install directory (say /rockbox/build) "../tools/configure"
12:02:18Jungti1234hmm..
12:02:53lamed(but i see no reason why shouldn't rbconf be properly configured)
12:03:36 Quit Zak1392 ()
12:03:51 Join ghode|afk [0] (n=garudin@host-84-9-104-109.bulldogdsl.com)
12:03:51lamedbagder - around?
12:04:20LinusNrbconf should be shot and buried
12:05:14lamedare you just saying that because of disagreaments with
12:05:16amiconnLinusN: I use some scripted shortcuts as well... not for configuring though
12:05:27lamedarrr... disagreements with bluechip?
12:05:29Jungti1234T.T
12:06:12LinusNno, but all those "convenient" scripts are unnecessary
12:06:12lamedlinusn, amiconn, do you know what's up with my patch?
12:06:31LinusNall they do is hide how things really work
12:06:45LinusNand those scripts are only part of the "devkit"
12:06:58amiconnWell, they make it easier to perform a task
12:07:18amiconnI agree that it's not a good idea to include them in the devkit
12:07:25LinusNexactly
12:07:41amiconn...but I made some scripts to save me the hassle to type longish commands over and over
12:07:44LinusNnothing wrong with scripting, but you should learn the basics first
12:08:35lamedlinusn: I must say that as a begginer you really have to get everything as easy as can be. you have no one to direct you to the smallest thing that stops you from progressing.
12:09:00LinusNamiconn: fm recorder can use OFF to cancel without problemsd
12:09:27lamedi'll again use the example of > file
12:09:51amiconnLinusN: So it doesn't drop you out of the FM screen? Interesting...
12:10:14lamed|<lamed> linusn, amiconn, do you know what's up with my patch?|
12:10:25LinusNno, i've been busy with other stuff
12:10:59lamed:( I shell wait some more then.
12:11:15LinusNlamed: btw, i'm obviously not good at catering for beginners
12:11:45amiconnGrmpf, I know why I usually avoid looking at radio.c
12:11:54LinusNamiconn: it's a mess
12:11:55amiconnThe button handling is a mess...
12:12:00LinusNlunch!
12:14:25 Join webguest55 [0] (n=d33a593d@labb.contactor.se)
12:14:44amiconnAh, it's because leaving the FM screen on Ondio is triggered on release...
12:14:57 Quit webguest55 (Client Quit)
12:15:10lamedlinusn: ouch! that hurts... as soon as i'll have some time I'll start a tip page for begginers... I was collecting tips from several irc questions that have been answered (!). I hope that people like yourself find themselvs answering questions over there. then same questions would not be asked twice.
12:15:17amiconn...without having a _PRE
12:15:26 Join gadasi [0] (n=d33a593d@labb.contactor.se)
12:16:16gadasiHi~
12:17:12amiconnLinusN: Can you switch between preset and scan mode on FMR (should be mapped to long ON) ?
12:18:03 Join Membrillo [0] (n=sam_kill@CPE-60-229-178-125.nsw.bigpond.net.au)
12:26:30Membrilloanyone know anything about editing WPS's?
12:28:06ashridahnah, it's all a black box ;)
12:28:44Membrillo.... *crickets buzz* .....
12:29:49Paul_The_NerdMembrillo: http://www.rockbox.org/twiki/bin/view/Main/CustomWPS
12:30:02Paul_The_NerdOr are you having specific problems?
12:30:10Membrillosweet. i just wanted to fix a few things with the iPodvol WPS
12:30:35Membrilloor isnt that possible
12:30:57 Quit linuxstb ("Leaving")
12:31:16Paul_The_NerdDepending on what you want to change, it's probably easy enough.
12:31:33Membrilloyeah fair enough
12:32:11Membrillojust add a battery % and volume % mainly. and stretch out the graphics on top. They are terribly sized
12:33:46Paul_The_NerdThat'll require a little bit of redrawing of the .bmp files, and a couple quick changes to the .wps file, but everything you need to know about the .wps itself are on the page I linked.
12:34:50markunYOOHOO!! I got the Gigabeat!!
12:35:08MembrilloPaul_The_Nerd: yep i got it sussed out
12:35:08Bgrhappy rb-gigabeating then ;)
12:35:20 Join muesli- [0] (i=muesli_t@Bc12c.b.pppool.de)
12:35:53markunThe screen is just so beautifull, you can read it with no problem without the backlight
12:36:04Bgrreally ?
12:37:52B4gderwow
12:39:42markunI F40 is thicker than the F20, but just as thick as my H120
12:40:58FebsMembrillo, if you adapt that WPS for the H300, you should post the WPS in the WPS Gallery so that others can benefit from it.
12:45:08Jungti1234 I made already suitable WPS to H300...
12:46:01 Join frederic [0] (n=chatzill@i577BB393.versanet.de)
12:46:50 Quit frederic (Client Quit)
12:47:13Membrillofebs: will do
12:48:29 Quit muesli_- (Read error: 110 (Connection timed out))
12:49:12 Quit elinenbe (" HydraIRC -> http://www.hydrairc.com <- The future of IRC")
12:54:14chopped_porkJungti1234: could you post it to the WPS Gallery?
12:54:49Membrillook mines finished
12:54:52Membrilloill just test it
12:54:57Jungti1234I didn't take Screenshot.
12:55:30chopped_porkJungti1234: rockbox on h3x0 doesn't support screenshots yet afaik
12:55:42Jungti1234I know
12:55:48Jungti1234Do you have H300?
12:56:18Membrilloi do
12:56:39chopped_porkno, my friend has one and none of the existing wps's looks good on it
12:56:49amiconnchopped_pork: rockbox on H3x0 _should_ support screenshot
12:56:49Jungti1234Can you test WPS that I make?
12:57:11Membrilloim just testing my own first
12:57:14Membrillothen ok
12:57:15amiconnThat is, it should do if USB _detection_ works
12:57:23Jungti1234amiconn: really?
12:57:35chopped_porkJungti1234: probably, that might take a day or two since i have to get in touch with him first ;)
12:58:05 Quit muesli- (Read error: 110 (Connection timed out))
12:58:14chopped_porkamiconn: right, rockbox on h3x0 doesn't make screenshots yet;)
12:58:26amiconnDid you try?
12:58:49chopped_porkhm, no tbh
12:59:11Jungti1234Membrillo: Can you test WPS that I make?
12:59:16Jungti1234now?
12:59:22chopped_porki read it somewhere and haven't seen a relevant cvs entry since then
12:59:58amiconnhttp://www.rockbox.org/twiki/bin/view/Main/ScreenDump
13:00
13:00:12ashridahamiconn: i seem to recall people trying and it not working.
13:00:16Jungti1234I know that
13:00:19amiconnIt might already work although USB mode itself doesn't work yet
13:00:37Jungti1234It didn't work.
13:01:00amiconnThe function is definitely there, so it's just not called yet
13:01:36amiconnAnyway, it's possible to take a screenshot on the H3x0 win32 sim
13:02:38 Join tucoz [0] (n=81b17b04@labb.contactor.se)
13:02:41Jungti1234I dont know whether I use it how.
13:03:05tucozHi, how do I change playback speed on the h1x0 series?
13:03:29 Join ep0ch| [0] (n=ep0ch|@84.12.67.172)
13:03:37tucozI can not find a setting for it in the menus
13:04:04amiconnIn wps, press & hold play, then use joystick up/down
13:04:17 Quit Paul_The_Nerd ("Chatzilla 0.9.68.5.1 [Firefox 1.5/undefined]")
13:04:34tucozamiconn: thanks!
13:05:18ep0ch|just been reading the logs, someone was mentioning profiling with regard to Vorbis decoder. Would profiling on x86 give a good indication as to what functions to optimise/place in iram for coldfire?
13:05:54 Join DreamTactix291 [0] (n=DreamTac@adsl-149-140-175.bna.bellsouth.net)
13:06:26***Saving seen data "./dancer.seen"
13:11:28tucozbye bye
13:11:30 Part tucoz
13:15:18perplexityI'm working on that right now ep0ch| :)
13:16:03perplexityI figure that at least profiling Tremor on x86 might help identify some of the low hanging fruit in any case.. biggest problem at the moment is getting it to work with the X11 sim.. but I'll get there..
13:17:43 Quit webguest92 ("CGI:IRC")
13:19:18Jungti1234It's H300 rockbox firmware install manual that make in Hangul. : http://cafe.naver.com/iriverh300/557
13:22:32Membrillohow do i put a % sign in my WPS so it places a % character on the screen and doesnt interpret it as syntax
13:23:00ashridah%% ?
13:25:28 Quit gadasi ("CGI:IRC (EOF)")
13:25:31 Join muesli_- [0] (i=muesli_t@Bbc62.b.pppool.de)
13:26:27Membrilloah
13:26:29Membrilloterrific
13:29:03ep0ch|perplexity: cool, hope you get some good results
13:29:53FebsFor new H300 Rockbox users, the QuickStart guide has been updated with H300 information (including button assignments): http://www.rockbox.org/twiki/bin/view/Main/QuickStart
13:30:30muesli_-Febs btw is it ok to start a german rbx support forum?
13:31:20muesli_-LinusN
13:31:23*perplexity bangs head on desk.. my first foray into gdb.. think I need a drink
13:31:23muesli_-amiconn
13:31:29muesli_-same q to you
13:34:01Membrillohmmm, for some reason my background on my WPS isnt showing
13:34:38Membrillowhat could it be
13:34:58Jungti1234hmm..
13:35:02Membrilloah
13:35:09Membrillodoes it have to be saved in monochrome format?
13:35:22chopped_porkMembrillo: yup
13:35:25Membrillothere you go
13:36:12Membrillook
13:36:20Membrilloim pretty sure my iPodvol WPS is done
13:36:46Jungti1234:)
13:37:05Bgrhttp://www.sunbelt-software.com/stu/quitcomplainingaboutyourjob.htm heh :)
13:37:46Jungti1234ow
13:38:10Jungti1234Terrible. :)
13:38:28Bgryep
13:38:55Jungti1234It's a Chinese.
13:39:46Membrilloit does indeed work
13:40:19Jungti1234ow
13:40:40Jungti1234Show it to me.
13:40:54Membrillonah, bed time now
13:41:02Membrillobut i will be back :)
13:41:05Membrilloi like to linger
13:41:24Jungti1234-_-;
13:41:43Jungti1234no..........
13:44:07 Join DocBrown [0] (n=80b0bc65@labb.contactor.se)
13:45:32Febsmuesli_ - Not my call on the German Rockbox support forum, although it sounds like a good idea to me.
13:46:02muesli_-we have a forum..but those mods are damn slow
13:46:11muesli_-useless
13:46:23muesli_-thus we'd like to set up a new forum
13:46:27muesli_-similar to mr
13:47:39Jungti1234I wonder.
13:47:39Jungti1234If I apply WPS that make, why does H300 stop operation?
13:47:50ep0ch|i think misticriver needs a bigger site logo
13:47:56Bgrwhat do you mean ?
13:48:03Bgr(to Jungti1234)
13:48:30Jungti1234As I use WPS that make, H300 stops.
13:49:20Bgrwhere "stops" mean ?
13:49:25Bgrexplain more
13:50:23Jungti1234system down
13:50:32Bgrpower off ?
13:50:36Jungti1234no
13:50:38Bgrthe unit powers of ?
13:50:39Bgror ?
13:50:40Jungti1234stop
13:50:48Bgrand returns in the file browser ?
13:50:59 Join DMJC [0] (n=DMJC@60-240-214-67.tpgi.com.au)
13:51:33Bgror it just sits in the wps screen
13:52:00Jungti1234Stop in wps browser.
13:52:22Bgrwps screen you mean
13:52:45 Quit Membrillo ()
13:52:49Bgri suggest u to put your wps code somewhere so someone could look at it ...
13:54:13Jungti1234May I show it to you?
13:54:34Bgri'm not the right person for wps ...
13:54:57Bgrand i don't have time right now to learn the wps codes ...
13:55:06lostlogicgah, can't simulate tremor to profile :(
13:55:11markunI thought the LCD of the Gigabeat was visible with backlight turned off.. but it was just dimmed.
13:55:20Bgr:(
13:55:32Bgrhm
13:55:42Bgrwhat's the battery life then ?
13:55:45perplexitylostlogic is it segfaulting ?
13:55:56markunwithout backlight 16 hours of mp3
13:56:01lostlogicperplexity: yes
13:56:04Bgrit doesn't stay dimmed for a long time time i guess ?
13:56:15Jungti1234qegnklqgnip
13:56:23Jungti1234What should I do?
13:56:38perplexitycomment out "errno=0" in _get_data in vorbisfile.c and it will play the first segment of the ogg in any case..
13:56:45BgrJungti1234: i said you: put it somewhere on the web and post a link here
13:56:51lostlogicJungti1234: post it on the misticriver forums or somewhere where people will look at it, ask for help
13:56:57Bgrand wait ...
13:56:58Jungti1234ok
13:57:17lostlogicperplexity: does it still segfault after?
13:57:30perplexitynope... just only plays about a second of sound then just sits there..
13:57:38lostlogicthat might work.
13:57:51perplexitywell.. I have just compiled with profiling enabled.. let's see what happens
13:58:22Jungti1234Membrillo, are you there?
13:58:42Jungti1234Did he go out?
13:58:54lostlogicI tried putting almost all of the residue and floor code in IRAM last night, it helped performance, but not as much as I'd hoped, so I want to profile it and see what exactly is worth moving to IRAM.
13:59:53perplexityI suspect it's going to require more than just moving code to iram.. I was thinking about some of the really critical code segments in asm perhaps.. but then that's why I'm profiling to find out :)
14:00
14:00:16lostlogicdepends on your goal
14:00:30lostlogicI'm just trying to get it to stop skipping on q7 ogg at 90mhz
14:00:32perplexitymy goal is to get -q10 oggs playing at 90Mhz :)
14:00:43lostlogicquite the goal :)
14:00:59lostlogic:-\ mine still segfaulted w/o that errno=0
14:01:00perplexityWell, I figure I might learn something along the way..
14:01:12perplexitydid you remove vorbis.codec and re-zip/unzip ?
14:01:23ep0ch|does iriver fw play -q10 oggs ok?
14:01:27perplexityrm -f apps/codecs/vorbis.codec && make && make zip && cd archos && unzip -o ../rockbox.zip && cd ..
14:01:41Bgrep0ch|: at least the specifications say so
14:01:43lostlogicI was a mo-ron and didn't run make install on that attempt
14:01:53perplexityApparently on the H1x0 it does..
14:02:21lostlogicyay, it's playing!
14:02:23ep0ch|iirc iriver fw uses 90 mhz for vorbis, so its probably do able
14:02:42perplexityI'm sure it is. it's just a matter of making the coded efficient enough..
14:02:45lostlogicep0ch|: it's too slow still, which is why perplexity and I and probably others are working on it
14:03:31perplexityI'd not go as far as saying "I" am working on it.. more like fumbling, bumbling, and crashing through the code like a wounded bull.. but you never know.. I might learn something..
14:03:39 Join muesli- [0] (i=muesli_t@Bc1e3.b.pppool.de)
14:03:50lostlogicperplexity: same.
14:04:52 Quit TCK (Read error: 110 (Connection timed out))
14:05:06amiconnIirc the iriver firmware doesn't play oggs >500 kbits/s
14:05:25amiconn(Not that I think such bitrates are sane...)
14:06:09Bgryes, this is near the loseless sizes
14:06:27amiconnExactly
14:06:43perplexityYeah, I'm thinking at that rate you might as well be using FLAC
14:07:06Bgrheh
14:07:06Bgr File Type MPEG 1/2/2.5 Layer 3, WMA, ASF, OGG
14:07:07BgrBit Rate MP3: 32Kbps ~ 320Kbps
14:07:10dwihnoWhat's with everybody disliking wavpack?
14:07:13ashridahhmm. some flacs i;'ve made have come out at 900kbps
14:07:15dwihnoWavPack to the people!!! :)
14:07:44amiconnashridah: I even have some >1000kbps. That's with -8 ...
14:07:45ashridahdwihno: it's not that i don't like it, i just haven't run into much that can play it.
14:07:52perplexityOnly that I already have over 100 full CD images in FLAC and I don't feel like converting them :)
14:07:57ashridahamiconn: whoa
14:07:58Bgr(taken from detailed specs for h100 @ iriver.com ... not very detailed obviously)
14:08:01dwihnoashridah: Same here... People just don't want to support it...
14:08:19dwihnoashridah: I'll probably end up converting all my wavpack to flac.
14:08:25perplexitylostlogic: are you able to play more than about 1 second of the song?
14:08:25ashridahdwihno: it has no mindshare, which is a pity, it sounds like it's decent.
14:09:01lostlogicperplexity: yeah, played about 2 minutes of the song
14:09:11lostlogicperplexity: but now I have to figure out multithreaded profiling
14:09:40lostlogicand I have to get it to show the symbol calls inside the vorbis codec
14:09:53perplexityheh.. I'm still suck with it stopping after 1 second.. I get some profile data but not enough to actually do anything :|
14:10:31lostlogic*nod* I get data, but it just shows the high-level calls into the codec, not the internals, missing something important here
14:10:53 Quit muesli_- (Read error: 104 (Connection reset by peer))
14:11:24dwihnoashridah: mindshare?
14:11:35perplexityYeah, sounds like thread selection.. well you are ahead of me in any case
14:13:19perplexityhttp://sam.zoy.org/writings/programming/gprof.html
14:13:49lostlogicyeah, seen it... working on it
14:13:59perplexityheh.. always one step ahead
14:14:21lostlogiccould barely sleep overnight, because I was planning to dive into profiling this morning ;)
14:14:28lostlogicbut you're the one who knew to disable that errno call.
14:14:33perplexityOh I'm not even close to that hard core..
14:15:42lostlogichardcore or stupid, hard to tell sometimes :-P
14:16:09B4gderI'm definately on the stupid side
14:16:17perplexitylol.. I know how you feel though.. I'm absolutely embarrased to say it took me nearly 2 hours to figure that errno segfault out.
14:18:48lostlogicI hadn't looked at rockbox code at all till last night, so it took a lot of patience from linuxtsb to get me pointed in the right direction with starting to optimize
14:19:18ashridahdwihno: random audiophiles haven't heard of it as much, it has no traction, no early adopter spread, etc.
14:19:27lostlogicI was able to make some headway with just moving things randomly to IRAM based on interpretation of frequency from code, but that isn't right, profiling is right.
14:20:03Jungti1234good night all
14:20:05Jungti1234bye
14:20:11Bgrnite, JdGordon
14:20:13perplexityYeah.. I only compiled my first version a couple of days ago.. and I'm pretty green with "C". Most of my stuff has all been directly in ASM, so this high level language stuff is taking some getting used to :)
14:20:14BgrJungti1234
14:20:20Jungti1234^^
14:20:28 Quit Jungti1234 ("bye")
14:20:40JdGordondont u hate crappy name comletion??
14:21:05lostlogichaha, other way around for me −− I used to hack about in the linux kernel, but currently I'm a Java coder by trade
14:21:06perplexityYou just don't hit tab enough...
14:21:09ep0ch|perplexity: oh just dive straight in and asm mdct.c :)
14:21:17BgrJdGordon: sorry :)
14:21:24dwihnoashridah: It also beats flac compression-wise
14:21:33perplexitygive me a little time to come up to speed with the coldfire instruction set ep0ch|, I only got the manuals yesterday
14:21:59ashridahdwihno: i didn't say it sucked.
14:22:14ashridahbut i'm just saying, beta didn't suck either, yet VHS still won
14:22:17dwihnoashridah: Sure you didn't, I just try spread some wavpack evangelism here :)
14:22:44lostlogicdangit, still not getting things that happen inside the codec
14:23:50preglowconverting wavpack to flac? now what kind of nonsense is that
14:23:59preglowlostlogic: getting what?
14:24:27lostlogicprofiling the simulator, I only see things that happen in the main rockboxui executable, not things that happen inside the codecs
14:25:32ashridahlostlogic: they run in separare threads
14:25:37perplexitySame here lostlogic... I only get vorbis tag reads and metadata info..
14:25:50ashridahseparate even
14:25:58lostlogicashridah: used: http://sam.zoy.org/writings/programming/gprof.html to try and capture all threads, may not be working.
14:26:10 Quit DocBrown ("CGI:IRC (Ping timeout)")
14:26:36perplexityI'll look at it again later.. I suspect forcing the ITIMER_PROF event from inside the codec thread directly will do what we want..
14:26:36JdGordongnite all
14:26:51 Quit JdGordon (Read error: 104 (Connection reset by peer))
14:26:53perplexityanyway I'd better leave before the parking inspector tows my car away.. ciao all
14:26:56 Quit perplexity ("*pop*")
14:28:21 Join Jungti1234 [0] (n=jungti12@58.77.81.144)
14:28:36Jungti1234hey guys
14:28:36Jungti1234Really, new original firmware of H300 release do?
14:28:52Bgrwhat???
14:29:11Bgrreally ?
14:29:16Jungti1234http://www.misticriver.net/showthread.php?t=32558
14:29:26Jungti1234I desire that that is true.
14:29:36ashridahiirc, gprof's threading support sucks balls, btw.
14:29:43lostlogicnoted.
14:31:05amiconnB4gder: r u there?
14:31:20B4gderyes, but only for a short while more
14:31:32amiconnRegarding the self-extractor...
14:31:58amiconnThe first thing necessary is to make the Makefile use the proper directories for building
14:32:11amiconnRight now it does everything in the source dir
14:32:53amiconnFor integration into the build system, I think the following things are necessary:
14:33:41amiconn(1) A script (build-image.pl ?) that builds the .ajz instead of a direct call to scramble, which does the following
14:33:56amiconn- call scramble and check return code
14:33:58Jungti1234Does no one know whether that is true?
14:34:15ep0ch|wait and see
14:34:15amiconn- when ok, the job is already done
14:34:45ep0ch|i cant see them adding any features though
14:34:47amiconn- if error: firmware too big, then:
14:35:02amiconn- call 'make' in firmware/decompressor
14:35:03BgrJungti1234: i don't think SoulEata has dreamed this ...
14:35:19preglowwho cares if it's true?
14:35:21Jungti1234hehe..
14:35:23amiconn- call scramble, this time using compressed.bin built in the last step
14:35:25preglowit's probably just a few small shitty features anyway
14:35:47Bgrpreglow u're probably right ;)
14:35:55preglowof course i'm right
14:35:58amiconn(2) For this to work, rockbox.ucl is a prerequisite for building ajbrec.ajz/archos.mod
14:35:59ep0ch|nah it'll be a bug fix, or something to do with lyric support
14:36:04Bgror maybe SDK for plugins :D
14:36:06amiconnThat's all...
14:36:12preglowi have wide experience with iriver not doing what the users want
14:36:32ep0ch|preglow: well they gave us gapless ;)
14:36:36Jungti1234Do you know here? : http://www.wikihouse.com/cipher/index.php?rockbox
14:36:49preglowep0ch|: yes indeed... the famous gap delete
14:36:51amiconnOh, and regarding the build directories, the Makefile should usually work in BUILDDIR/firmware/decompressor ...
14:36:55BgrToday, 01:35 AM
14:36:55Bgr...gapless... gapless... gapless... gapless... gapless... gapless... gapless... gapless... gapless... gapless... gapless... gapless...
14:36:55BgrPlease Father Christmas, make it gapless... I've been a good boy all year... well, mostly good for most of the year...
14:36:56DBUGEnqueued KICK Bgr
14:36:56Bgrhahaha
14:37:01Jungti1234hahaha
14:37:03preglowhahah
14:37:03amiconn...but take rockbox.ucl directly from BUILDDIR
14:37:04preglowyes
14:37:09preglowfat chance you'll get gapless
14:37:22preglowi can guarantee you: no gapless
14:38:03ep0ch|any idea if iriver are going to come out with a successor to the H series? or are they going down the video route?
14:38:28amiconnB4gder: That should work for all archos target, so *in theory* we could exceed the 400KB limit on fmr/Ondio as well, although I strongly hope that'll never happen
14:40:33B4gderok, gotta go now
14:40:35 Quit B4gder ("time to say moo")
14:42:24 Join zaZZ [0] (i=acupk@dyn36.her1.nas.panafonet.gr)
14:43:06Jungti1234bye
14:43:07 Quit Jungti1234 ("bye")
14:43:16zaZZhi. i have an iriver 140 and using rockbox. I also update daily builds. Do i need to update the bootloader or is updated by the daily builds?
14:43:41LinusNthe bootloader is separate
14:43:47ep0ch|zaZZ: you will need to reflash the firmware
14:44:27zaZZdo i need to get the new bootloader? i mean is it getting better like the daily builds?
14:44:28markunThe max volume of the gigabeat is insane
14:44:34markunBlows your head off
14:46:26 Join perplexity [0] (n=joust@de23559.alshamil.net.ae)
14:46:44 Join KN|stiff [0] (i=phhome@141.48.5.112)
14:47:30amiconnmarkun: Well, the max. iriver volume isn't really high
14:47:38amiconnArchos is significantly louder
14:47:48amiconnNot that I'd need it...
14:48:26ep0ch|zaZZ: what version bootloader do you currently have?
14:48:56zaZZv.5
14:52:14ep0ch|zaZZ: imho the new bootloader doesnt seem to add anything essential, but if you can bare going into iriver firmware to update there's no harm in it.
14:53:41zaZZok, i will now try to find how to reflash the firmware and go to bootloader v.6
14:53:42zaZZthx
14:54:20 Join DocBrown [0] (n=80b0bc65@labb.contactor.se)
14:55:25DocBrownLinusN: Hi! Just put the results from last nights 'bootloader session' to the wiki
14:56:07 Quit ashridah ("Leaving")
15:00
15:00:48 Quit zaZZ (" Scp Script 2004 ! with Self clearing virus System! Download it at www.scoopsite.com")
15:01:51 Quit lamed ("CGI:IRC")
15:02:46thegeek_ooh!, scp script!
15:06:29***Saving seen data "./dancer.seen"
15:07:10Bgrwhat ? :)
15:08:53 Join Kohlrabi [0] (n=Kohlrabi@dslb-082-083-134-154.pools.arcor-ip.net)
15:19:38LinusNDocBrown: great
15:22:32lostlogicwhere does the simulator's main method hide?
15:23:12LinusNx11 or win32?
15:23:17lostlogicx11
15:24:27LinusNuibasic.c iirc
15:24:49LinusNno, screenhack.c
15:25:12lostlogiccoo', ty
15:25:24perplexitygot those profiles yet lostlogic ?
15:25:31lostlogicperplexity: no :(
15:25:56lostlogicused moncontrol(0) to turn it off in screenhack.c and moncontrol(1) to turn it on in the audio_thread, and still getting J4CK
15:26:59perplexityBummer..
15:27:18perplexityLooks like gprof is a bit of a three legged dog for threaded stuff
15:27:24lostlogic2 legged.
15:28:33ep0ch|are you trying to profile rockbox or Tremor?
15:28:59perplexityTremor (well I was in any case)
15:29:11lostlogicTremor by way of rockbox
15:29:14 Join BBub_ [0] (i=belzebub@dslb-084-059-201-017.pools.arcor-ip.net)
15:29:28ep0ch|why not just profile Tremor without Rockbox
15:29:38amiconnLinusN: Btw, is there a reason why the 'Force mono' radio setting is handled specially?
15:29:43lostlogicdon't want to write a whole test application to do that?
15:29:57LinusNamiconn: specially?
15:29:59amiconnIt's a bit faster to operate this way, but it breaks consistency
15:30:43amiconnWell, this is the only boolean option that has its setting *in* the menu item, and calling the item toggles it
15:30:56 Quit einhirn (Read error: 104 (Connection reset by peer))
15:31:38LinusNah, it was an experiment really, to see if it was nice to have a direct toggle
15:31:43amiconnAlso, did you check whether switching FM modes works on fmr?
15:32:05LinusNscan/preset?
15:32:10amiconnyup
15:32:17 Quit DreamTactix291 (Read error: 110 (Connection timed out))
15:32:20amiconn(direct toggle) what do you think?
15:32:39LinusNmode change works fine
15:32:39amiconnI'm asking because the scan/preset item for Ondio would also be boolean
15:32:53LinusNi like direct toggles
15:33:06LinusNin fact, i'd like more settings to be like that
15:33:11amiconnHmm.
15:33:37LinusNbut it's hard to do platform independent
15:33:42amiconnFor more than 2..3 settings, the values should be right justified, imho
15:33:50LinusNabsolutely
15:34:05amiconn...and it will be difficult to implement this for multiple-choice options
15:34:11amiconnThe strings can be rather long
15:34:39amiconnSome numerical settings have quite long unit names as well
15:34:50LinusNyes, there are countless problems with that approach
15:35:15amiconnBoolean could use the checkbox widget
15:35:16LinusNi have a few ideas, but let's discuss them another time :-)
15:35:37amiconnI just need to decide how to handle the mode switch on ondio right now
15:35:45LinusNyes, but how is the checkbox voiced?
15:36:11LinusNdo the in-menu boolean thing, like the mono mode
15:36:17LinusNi like it
15:36:28lostlogicperplexity: new hunt, let's use google. http://ipodlinux.org/forums/viewtopic.php?t=150&sid=32cf5dec452d2038401b9ff7b5165288 has some info that may be useful. (/gprof)
15:36:39LinusNDocBrown: u there?
15:37:40amiconnLinusN: I think the voicing could stay as-is. Yes/No or On/Off
15:37:48LinusNme too
15:39:38 Join einhirn [0] (i=Miranda@bsod.rz.tu-clausthal.de)
15:40:16LinusNanyone here with an h300 that can't boot with the v2 bootloader?
15:40:51perplexitywell spotted lostlogic.. now I have been digging in all the generated asm files for Tremor only to find that some nice person has already optimised a great deal of the real workers..
15:41:28 Quit KN|stiff (Read error: 110 (Connection timed out))
15:42:44DocBrownLinusN: In principle, yes. Can look into brownser only occasionally
15:42:46preglowperplexity: yes, some friendly soul appeared out of nowhere and optimised it
15:43:07LinusNDocBrown: http://linus.haxx.se/H300.zip
15:43:20perplexityOh I'm sure they just "appeared out of nowhere" :) I just have not been around long enought to equate real names to irc nicks ;)
15:43:25DocBrownLinusN: Okay, will try...
15:43:44preglowperplexity: at least i haven't seen him for at least four months
15:43:49preglowhe never hung around here anyway
15:44:14perplexityI'm impressed by the code anyway..
15:44:22LinusNin fact, his optimization patch came as a surprise
15:45:01perplexityMan I wish we could run gprof on the actual target :\
15:45:30LinusNthere is an old profiling patch in the tracker
15:45:39perplexityI might have a look.. ta :)
15:46:00LinusNbut it needs some work, a lot of things have changed since then
15:46:25 Quit BBub (Read error: 110 (Connection timed out))
15:46:31perplexitylostlogic it looks like there could be a small gain to be had by copying the sincos_lookup* tables to fast ram.. if they are not there alreay
15:46:46perplexityIt would give me a leg up anyway.. Thanks for the pointer.
15:47:32perplexityIgnore me.. I see they already are..
15:47:38 Quit BBub_ (Read error: 110 (Connection timed out))
15:47:47LinusNperplexity: http://sourceforge.net/tracker/index.php?func=detail&atid=439120&group_id=44306&aid=1157222
15:47:57amiconnLinusN: Hmm, I want to reuse the language strings for the menu item, but they are 'Scan Mode' and 'Preset Mode'
15:48:10LinusNbummer
15:48:13amiconnPutting a 'Mode:' in front of them will probably look silly,
15:48:26amiconnbut not putting anything would be confusing...
15:49:41LinusNthen change it to Mode: Scan/Preset in the main screen
15:49:43perplexityThanks LinusN, that was how I intended to do it myself, but seeing as someone else has already done the donkey work.. I'll just stand on their shoulders :)
15:50:29 Quit Zagor ("Client exiting")
15:56:28preglowperplexity: but yeah, in addition to iram use, you can also recode the mdct
15:57:05perplexityby recode you mean convert more of it into better asm, or fundamentaly change the way the algorithm is implemented?
15:57:42perplexityI have been combing the final generated asm for the last hour or so, and to be frank.. it's not too shabby, especially when you take into account the hand optimised portions..
15:58:07preglowperplexity: it can be coded by doing an (i)fft (which are faster), and a coupleof sort step
15:58:10preglows
15:58:46perplexityIn which case this programmer, bereft of mathematical dsp talent is going to have to do some serious study :)
15:58:53preglowhehe
15:59:00preglowsome of the current codecs already do this
15:59:03preglowlike a52 and flac
15:59:06preglowyou could study them
15:59:18preglowbut yeah, ffts exist already in numerous fast implementations
15:59:24perplexityAnd I will.. FLAC is super fast..
15:59:25preglowand the sort steps aren't too hard
15:59:34preglowi meant aac
15:59:36preglowahh, sorry
15:59:41preglowflac is a time-domain decoder
15:59:59perplexityindeed.. I just defered to the greater experience :)
16:00
16:01:02perplexityI had naively assumed that as Tremor was written for these sort of apps, that the author might have chosen the most efficient algorithms off the bat..
16:01:29 Join mirak [0] (n=mirak@AAubervilliers-151-1-47-69.w83-114.abo.wanadoo.fr)
16:01:34miraksalut
16:01:38 Join DreamTactix291 [0] (n=DreamTac@adsl-149-140-175.bna.bellsouth.net)
16:02:01LinusNmirak: salut
16:02:05 Part miah
16:02:23preglowperplexity: no, he deliberately did not go into that issue at all
16:02:26mirakanyone heard of fixed point codecs ?
16:02:30mirakfor video
16:02:36preglowperplexity: i views tremor as a sort of baseline implementation, where everyone else does work on top of it
16:02:48preglowperplexity: s/i/he/
16:02:53LinusNi know next-to-nothing about video codecs
16:02:53preglowmirak: of course
16:03:10markunffmpeg is fixed point I think. Most are.
16:03:36 Join DJDD_ [0] (n=DJDD@220-245-186-182.static.tpgi.com.au)
16:03:45perplexityOk, thanks for the headsup preglow.. now it's off to find/hit the dsp books.. later all :)
16:04:03preglowperplexity: goodie, could need another dsp fellow around here, i'm wildly inefficient ;)
16:04:10preglowbut okays, i'm out, later
16:05:40DocBrownLinusN: Not a real improvement. Have the impression that the first one was closest... now it either stays black, flickers or boots iriver fw
16:05:42markunmirak: on 2nd thought, I'm not so sure
16:07:28 Join Kaggen [0] (n=kaggen@nl-67-116.netlogon.liu.se)
16:10:37 Join webguest47 [0] (n=c13354c1@labb.contactor.se)
16:11:21webguest47Hello here
16:11:32webguest47I'm wondering something
16:12:09LinusNDocBrown: it doesn't boot the iriver firmware anymoe?
16:12:20LinusNsorry, i misread
16:12:36webguest47When I go to the audio tread for see cpu boost, I noticed that witout song the cpu boost increased, sometimes more than with sound 15-20%
16:12:48webguest47is it one known thing?
16:13:04LinusNwebguest47: which platform?
16:13:18webguest47sorry, iriver h1xx platform
16:13:29LinusNdo you have backlight fade enabled?
16:13:30markunmaybe the backlight fading?
16:13:31webguest47boost cpu without song
16:13:57webguest47yes I have,
16:14:13webguest47but this backlight fadind cost cpu more than song?
16:14:13LinusNthe backlight fading boosts the cpu
16:14:21webguest47ok
16:14:28webguest47a lot, no?
16:14:57webguest4715-20%, when we know FLAC for exomple doesn't eat cpu around
16:15:34webguest47now you said me that I will never use this option again (pitty it was so fun)
16:16:30amiconnNo, backlight fading doesn't cost much CPU by itself, but it has to boost the CPU in order to avoid frequency changes during the fade
16:16:41amiconn...which would cause flicker
16:17:07webguest47but I tested to dosen't touch any key, and that still boost a bit :(
16:17:20webguest47amiconn: ah ok
16:17:40ep0ch|while were on the subject of backlight fading. Something that slightly annoys me is the way backlight fading is on by default. Shouldn't the default be to have no fading?
16:17:50amiconnIt only boost during the fade
16:17:53webguest47taht is true
16:18:01webguest47epOch
16:18:26webguest47amiconn: can the dircahe boost the cpu, if is'nt the backlight?
16:18:33LinusNep0ch|: it's sometimes good to have the eye candy on by default, for the first-impression effect
16:18:52webguest47rellay true too :)
16:18:53 Join Pieter [0] (i=Pieter@pieter.student.utwente.nl)
16:19:01Pieterhi!
16:19:03LinusNi believe the dircache boosts it during the background scan
16:19:05ep0ch|LinusN: true but i think as it boosts cpu it should be off
16:19:05webguest47Hello
16:19:16Pieterrockbox stopped compiling on the development kit
16:19:25Pietermentions something about uclpack...
16:19:31amiconnep0ch|: It only boosts during the fade
16:19:42CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
16:19:42*amiconn repeats himself
16:19:42ep0ch|i know
16:19:49webguest47LinusN, amiconn: if isn't dircahe or backlight I'm wondering what can cause this cpu cost?
16:19:53LinusNPieter: install cygwin for real, or wait for an updated devkit
16:21:54webguest47an other system question please guy
16:22:12amiconnThe dircache boosts the CPU during the background scan
16:22:31*amiconn has the dircache disabled
16:22:43Pieteri'll install cygwin then.. thought i'd notify you :)
16:23:24webguest47I make the spining down option to 3seconds, but sometimes that is much more than 3sec
16:23:26LinusNPieter: we've had lots of reports
16:23:47Pieterah. sorry then :)
16:23:51webguest47isn't normally force the HD to sleep in 3sec for my exemple?.
16:23:51LinusNPieter: but the devkit is not the officially supported way of developing
16:24:02 Quit gromit` (Read error: 104 (Connection reset by peer))
16:24:41webguest47LinusN, amiconn: any thought please
16:25:08amiconnIt spins down 3 seconds after the last activity once it is spinning
16:25:19LinusNi gotta go, cu later
16:25:20amiconnIf you keep pressing buttons, it keeps spinning
16:25:53webguest47but for exemple I do nothing and I can still hear HD activity
16:26:07webguest47and few seconds after it go sleeping
16:26:16webguest47(more than 3sec)
16:26:28webguest47and I didn't do nothing, just browser
16:27:00 Quit DocBrown ("CGI:IRC 0.5.7 (2005/06/19)")
16:27:01webguest47OMG 45minutes now for building :)
16:27:13webguest47CVS one I mean
16:27:40PieterlinusN: thanks.. i'll try cygwin again and will stop whining ;)
16:28:20amiconnLinusN: Teh iwindows nstaller didn't get built again
16:29:14webguest47oups Linus gone
16:30:03webguest47thanks folks for your answers
16:30:29webguest47have a good day to all
16:30:34 Quit webguest47 ("CGI:IRC")
16:34:56 Quit muesli- (Read error: 110 (Connection timed out))
16:37:23 Join gromit` [0] (n=gromit`@ras75-5-82-234-244-69.fbx.proxad.net)
16:37:29 Quit gromit` (Remote closed the connection)
16:37:32 Join gromit` [0] (n=gromit`@ras75-5-82-234-244-69.fbx.proxad.net)
16:38:34 Join webguest20 [0] (n=542b6bfa@labb.contactor.se)
16:41:01 Quit webguest20 (Client Quit)
16:42:47 Quit Pieter ()
16:51:12 Join muesli_- [0] (i=muesli_t@Bbc81.b.pppool.de)
16:59:52 Quit DreamTactix291 (Read error: 110 (Connection timed out))
17:00
17:00:54 Join muesli- [0] (i=muesli_t@Bbc8e.b.pppool.de)
17:02:13 Quit muesli_- (Read error: 104 (Connection reset by peer))
17:03:47 Quit akaidiot ("( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de )")
17:05:02 Part Kaggen
17:06:33***Saving seen data "./dancer.seen"
17:07:14 Quit Febs ("CGI:IRC (Ping timeout)")
17:10:05 Join mashalla [0] (i=mashalla@p5498C774.dip.t-dialin.net)
17:12:09 Join akaidiot [0] (n=nope@c-f744e255.363-1-64736c11.cust.bredbandsbolaget.se)
17:12:51 Join lostlogicx [0] (n=lostlogi@temporal.lostlogicx.com)
17:13:19lostlogicxperplexity: any profiler luck by chance?
17:21:56 Join bobwise [0] (n=45dda986@labb.contactor.se)
17:29:47 Join BBub [0] (i=belzebub@dslb-084-059-201-017.pools.arcor-ip.net)
17:37:51 Quit mirak (Remote closed the connection)
17:41:47 Join PaulJ [0] (n=PaulJ@vpn-3041.gwdg.de)
17:45:59 Part ep0ch|
17:47:07 Quit bobwise ("CGI:IRC (EOF)")
17:52:24 Nick Lynx_ is now known as Lynx_awy (n=lynx@tina-10-4.genetik.uni-koeln.de)
17:55:21 Quit mashalla (Read error: 110 (Connection timed out))
18:00
18:03:07 Part LinusN
18:06:26 Join ]RowaN[ [0] (n=54093c61@labb.contactor.se)
18:06:57]RowaN[this playback speed option for the iriver.. does it alter the pitch too?
18:12:34 Join DangerousDan [0] (n=Miranda@newtpulsifer.campus.luth.se)
18:16:52 Join Moos [0] (i=DrMoos@m79.net81-66-158.noos.fr)
18:26:35 Join tvelocity [0] (n=tony@ipa152.4.tellas.gr)
18:30:46 Quit ]RowaN[ ("CGI:IRC (EOF)")
18:30:53 Join _FireFly_ [0] (n=FireFly@p54A45DFD.dip.t-dialin.net)
18:31:34muesli-is there a limitation of wav-recording?
18:33:47 Join DreamTactix291 [0] (n=DreamTac@adsl-149-140-175.bna.bellsouth.net)
18:34:31_FireFly_muesli-: afail the free disk capazity
18:34:34_FireFly_afaik
18:34:50_FireFly_and after ca 2gb a file-split
18:35:12muesli-sounds awesome
18:35:47muesli-what about "knackser" are they hardware related?
18:36:18_FireFly_the only "knackser" which i know are only on the start of recording
18:38:00 Part PaulJ
18:39:44muesli-ok :D
18:51:54 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
18:54:16 Join PaulJ [0] (n=PaulJ@vpn-3041.gwdg.de)
18:56:58wubblaHi
18:57:11wubblahum. Where's LinusN? :-)
18:57:39wubblaI saw that there has been progress on the dead-bug?
18:58:29 Join lamed [0] (n=d4b3395e@labb.contactor.se)
18:59:11lamedhere's a stupid one: how do i change playback speed on the h1x0?
19:00
19:00:11lamedanyone please?
19:00:19lamedi'm in a bit of a hurry
19:00:32_FireFly_look at the forum
19:00:45lamedk
19:00:48Mooshold play key+up/down
19:03:00lamedon wps that is?
19:03:46Moosyes
19:03:48lamedhere's the problem - the cvs update/build didn't went through like expected. i'm still running 11/25.... thanks all.
19:04:49Moosthis playback speed changes seeting, are activated yesterday
19:06:17lamedyes, i dont know what went wrong.
19:06:34***Saving seen data "./dancer.seen"
19:06:41_FireFly_error messages ??
19:06:54Moosmake clean ;)
19:07:02_FireFly_has cvs updated printed some lines with a C on start ??
19:07:19lameddon't worry _firefly_, i'll check.
19:07:21_FireFly_or do a reconfigure
19:08:12lamedi'm corrently working for the computer to start working.
19:08:29_FireFly_;)
19:09:49lamedthat's the second reset within two minuts. not my computer.
19:10:11 Quit Wett ()
19:11:02lamedkill bill gates.
19:11:10_FireFly_;)
19:12:07lamed(connecting the usb cable caused the computer to... reset!)
19:13:20t0masthat might be a feature...
19:13:47markunAnyone feeling like doing some decrypting?
19:13:52t0masif your reset switch is broken... and you're using windows < 2000
19:14:06 Quit DJDD_ ("Trillian (http://www.ceruleanstudios.com")
19:14:15t0masmarkun: decryption of what?
19:14:38markunI have some system files from my Gigabeat. One of them give quite a regular image when viewing it as a 256x.. 8bit grayscale image
19:15:27lamedi think rb checks for rockbox.iriver at /.rockbox folder before at the root. might be misleading as it still builds the file outside the folder.
19:15:38t0mashm...
19:15:40lamedpretty positive on that
19:15:50t0maslamed: then change the order?
19:15:52markunThis is the image: http://130.89.160.166/rockbox/FWIMG01.png
19:16:03t0maslamed: or better... ask Linus to change it
19:16:14markunthe actual file is the same .DAT
19:16:40_FireFly_markun: it looks very colorful :)
19:16:43lamedtOmas: then if you had an old rockbox.iriver at the firmware folder, and you've just upgraded, you won't get updated.
19:16:58markun_FireFly_: I'm color blind :)
19:17:10lamedyes, thanks for the speed _firefly_, se ya all later.
19:17:19t0mashm... markun? can you make a screenshot of what you see?
19:17:36t0masI don't see anything in it...
19:17:52markunt0mas: Don't you see the stripes?
19:18:04markunMaybe zoom in..
19:18:16_FireFly_it is only a carbage of small color pixel :)
19:18:22t0masmarkun: ah, see it
19:18:44_FireFly_fine lines i can see
19:19:16_FireFly_markun: should the file be an image or what ??
19:20:10markun_FireFly_: I don't know what the image should be, There are 3 firmware files, the first is very small, and the last doesn't show any patterns.
19:20:24markunMaybe they are encrypted the same way
19:20:51_FireFly_maybe the first one is a bootloader
19:21:15_FireFly_are the first 2 bytes of each file the same ??
19:21:58markunno, they aren't
19:22:33 Part Polo_o
19:23:50_FireFly_markun: why do you think it is encrypted ??
19:24:30markunI was just suspecting it
19:24:40t0mason what is it running?
19:24:49_FireFly_gigabeat
19:24:56t0masyeah... but what CPU?
19:25:20MoosARM
19:25:50 Join tucoz [0] (n=81b17b04@labb.contactor.se)
19:26:03tucozmarkun, have you tried running strings on that file?
19:26:14Sananything major new in the daily builds?
19:26:17markunyes, no luck
19:26:20_FireFly_running strings ??
19:26:23tucozI see.
19:26:42t0mas_FireFly_: a gnu utilitie for displaying strings in an executable or something like it
19:26:42tucoz_FireFly_: Yes, that dumps all strings in a file
19:26:52_FireFly_ahyes
19:27:25t0masmarkun: do you know the gcc target for that cpu?
19:27:27markunThere are so accidental strings like g7DeO or something
19:27:34t0masthen you can just try to disassemble it with gdb...
19:27:40t0masif that works... it's not encrypted
19:27:40tucozIf you run strings on the x5 firmware, you see that at least the strings are unencrypted
19:27:58tucozor if you run strings on the unscrambled iriverfw
19:31:29 Quit muesli- (Read error: 110 (Connection timed out))
19:32:10 Join webguest23 [0] (n=d444e303@labb.contactor.se)
19:35:20 Join LinusN [0] (n=linus@labb.contactor.se)
19:37:08 Quit webguest23 ("CGI:IRC (EOF)")
19:37:19LinusNregarding the search order for rockbox.iriver, it looks in .rockbox first because that's where it will be when we finally change the distribution archives
19:37:58 Join ghode [0] (n=garudin@host-84-9-104-109.bulldogdsl.com)
19:38:00LinusNwe haven't changed it just yet, because we want to wait until people have installed the v6 bootloader
19:40:15NixsosLinusN: yesterday i reported the 'disk spinning - no backlight' bug on my H340 Intl, just wanna let you know that the Hold trick works! So do you already have an idea about how to fix the bootloader?
19:40:41LinusNnot yet, i have tried a few fixes, but none of them helped
19:41:00LinusNNixsos: do you have the bootloader installed?=
19:41:05Nixsosyeah
19:41:23LinusNcan you tell me if it behaves differently with the remote attached?
19:41:34LinusNthe lcd remote
19:41:36Nixsosi dont have the remote :(
19:41:42Nixsossold it with my H140
19:41:51LinusNbummer
19:47:14markunLinusN: are you good at decrypting?
19:47:22LinusNnot very
19:47:28LinusNgigabeat?
19:47:32markunyes: http://130.89.160.166/rockbox/FWIMG01.png
19:47:59LinusNwow
19:48:06 Join Febs [0] (n=40326e83@labb.contactor.se)
19:48:06LinusNthat's a pretty even distribution
19:48:42LinusNhave you tried different image widths?
19:51:03LinusNalso, do you have access to some other image files?
19:51:10LinusNprevious updates
19:51:33LinusNthen you might be able to see if there is some common data
19:51:43LinusNlike a header or something
19:52:09LinusNyou should also try to study some ARM acritecture to find some cribs
19:52:30LinusNthe exception table is a nice crib
19:52:38LinusNfor example
19:53:04 Quit _FireFly_ ("Leaving")
19:53:20LinusNmarkun: still there?
19:54:06 Quit ghode|afk (Read error: 110 (Connection timed out))
19:56:50markunback again
19:57:17 Quit hshah_ (Read error: 110 (Connection timed out))
19:57:32wubblaLinusN: hoi! you're there
19:57:55markunLinusN: thanks, I'll try you suggestions later
20:00
20:00:46 Quit DreamTactix291 (Read error: 110 (Connection timed out))
20:01:55wubblahmm...
20:02:08wubblawhat about this Hold trick?
20:02:38Sanput hold on
20:03:29wubblaand it allows a dead-syndrome affected to boot rockbox?
20:05:48 Join muesli_- [0] (i=muesli_t@Bc148.b.pppool.de)
20:08:37Sanno
20:08:40Sanwell..yes
20:08:49Sanput hold on, and hit the reset button twice
20:08:55Sanshould work
20:09:00Sanit worked for many people
20:09:56*amiconn held an iPod today for the first time, and also an iAudio X5
20:10:11amiconnI'm not impressed at all...
20:10:16tucozhehe
20:10:47amiconnThe iPod nano is _really_ thin, but that's about it
20:11:02tucozI would like to look at a gigabeat in real life. I think those look nice
20:11:25amiconnAll the iPods already looked scratched in the shop
20:11:31tucozyes, the nano is really thin. So is the ipod video
20:11:54SanI would love to see a H300 in real life
20:11:55tucozfor real? that is not very good. Imagine what they would look like after some time of use.
20:11:57amiconnThen this wheel thingy... it's irritating
20:11:58Sanoh, wait....I have one
20:11:59San:P
20:12:26amiconn...and the nano screen is too small for the size of the device, imho
20:12:43amiconnThe X5 has a "strange" (at best) form factor
20:13:08tucozIsn't it quite similar to the irivers?
20:13:23amiconnIt's thicker at the top than on the bottom
20:13:43amiconn...and then it has a tiny joystick, imho on the wrong side
20:13:58amiconnand side buttons, like the H1x0 :(
20:14:21 Join Lear [0] (n=chatzill@h73n11c1o285.bredband.skanova.com)
20:14:25*amiconn wants buttons, _front_ buttons
20:15:16tucozI think my next player will be a gigabeat or a nano. Depending on how small the gigabeat is. I really would like to have 320x240 resolution.
20:15:48amiconnFor what?
20:16:13markuntucoz: the firmware sucks, but maybe that's even better :)
20:17:28amiconnmarkun: This seems to be quite common for DAPs... :/
20:18:05markunLinusN: 512x.. still gives a nice pattern. At 1024x... the pattern is shown in 2 columns, so I guess it's 512: http://130.89.160.166/rockbox/001b.png
20:18:21tucozamiconn: that is for all my vga needs :)
20:18:43amiconnI don't need a hi-res display on a dap
20:19:35amiconn...and no colour either
20:20:47tucozhehe, probably not me neither. But, it has a pretty nice processor and could probably be used as a gaming device as well. Then again, there are of course portable gaming devices for that. :)
20:22:55tucozamiconn: you should port rockbox to a nano. No screen what so ever ;)
20:23:01tucoz*shuffle
20:23:26amiconnCould be interesting, really
20:23:40tucozI read that you had practiced on some led-debugging recently. Impressive I must say.
20:23:43amiconnWith a working voice UI...
20:24:09tucozyes, rockbox on a shuffle would make it usefull.
20:24:35amiconnWell, the LED debugging wasn't really difficult by itself, only a little time-consuming
20:24:47tucozhehe, I find stuff like that impressive
20:25:06amiconnI 'blinked' byte by byte, separated into 4-bit groups == 1 hex digit
20:25:55tucozBut, embedded development is like that from time to time. Isn't it?
20:25:58amiconnIt doesn't take long until you are able to 'read' the number right away
20:26:17amiconnLike, -.-. is 0xA
20:26:43tucozso - is a 1?
20:28:36amiconnYes, - (long blink) is a 1, . (short blink) is a 0
20:29:04tucozAnd now the ucl-extractor is functional?
20:29:08amiconnyup
20:29:40amiconnNormally, you can avoid doing such low-level output with some kind of debugging support
20:29:58amiconn...but in this case, the gdb stub wouldn't have helped me
20:30:10amiconnLinus tried it, and it worked fine
20:30:52amiconnThe problem wasn't directly in my code, but an ooold bug in RoLo
20:30:52tucozok, great. Then were one step closer to unicode.
20:31:06tucozYes, I read the logs and the commit message
20:31:09tucozNice catch
20:31:25amiconnBtw, my most-wanted number was: .... -..- .... .... .... ..-. .... ....
20:31:30tucozhaha
20:32:25tucozback to work, see you
20:32:31 Part tucoz
20:34:36Slasheriamiconn: hehe, that kind of led debugging is always more interesting than gdb.. in fact, i have almost never used any debuggers but direct hardware methods :)
20:35:08amiconnYeah, gdb wouldn't have helped me here...
20:35:32amiconnFixing the RoLo bug also fixed RoLo'ing the archos Ondio firmware
20:36:37amiconnIt puzzled me for quite a while why this didn't work. When RoLo'ed, the firmware started, but as soon as I tried to play some music, it froze, and after a couple of seconds the unit switched itself off
20:37:34amiconnRoLo didn't set the stack pointer where the loaded image wanted it to be, but left it wherever it was
20:38:04amiconnThe archos Ondio firmwares want a quite high stack pointer (although the firmware itself is smaller than rockbox)
20:38:51*amiconn summons [IDC]Dragon
20:39:27amiconnImho we should have new flash images containing bootbox with fixed RoLo...
20:41:56LearHm... Should libfaad really include /lib/codeclib.h? That drags in a lot of stuff, causing these frequent, full, libfaad rebuilds (which doesn't use the compile rule in the makefile, btw)...
20:43:18LearIt seems to be done to enable the use of DEBUGF/LOGF, but I don't know if it is actually used...
20:44:17LearNope, not used that I can see.
20:56:19PaulJdoes rockbox on h300 indicate when the battery is fully charged?
20:57:13 Join iobound [0] (n=iobound@221.80-203-3.nextgentel.com)
20:57:31wubblaLinusN: ?
20:58:13ioboundhi guys, i had a problem with gdb not picking up debug symbols from .rock files on the win32 sim, but i fixed it by replacing the %.rock makefile rule with the one from x11-simulator (using gcc -shared)
21:00
21:00:44LearHe, most codecs are built using -O only...
21:01:11 Quit CBM-away (Read error: 104 (Connection reset by peer))
21:02:13 Quit solexx (Read error: 104 (Connection reset by peer))
21:02:15 Join Sandking [0] (n=jacek@ogorek.akron.net.pl)
21:02:44 Join solexx__ [0] (n=jrschulz@d083054.adsl.hansenet.de)
21:03:22 Quit tvelocity ("Leaving")
21:06:35***Saving seen data "./dancer.seen"
21:07:50 Quit solexx__ (Remote closed the connection)
21:07:51 Join solexx_ [0] (n=jrschulz@d083054.adsl.hansenet.de)
21:13:59 Join bobwise [0] (n=45dda986@labb.contactor.se)
21:16:06 Join dropandhop [0] (n=a@gateway.symphony.org)
21:19:55 Join paranoid_android [0] (n=546d8349@labb.contactor.se)
21:23:45dropandhophey!
21:23:50 Quit einhirn ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
21:23:50dropandhopis slasheri around
21:23:58dropandhopor any tick-fix experts?
21:24:13Slasherihehe, hi
21:25:52dropandhophey!
21:26:02LearHm.. Maybe that -O comment should be some, rather than most... :)
21:26:06dropandhopjust wanted to draw your attention to some threads in the forum
21:26:12dropandhopnot sure if you have followed them
21:27:00paranoid_androidi'm just curious,can we asist the int.mic which is in h3xx series to create a voice to text -a -like program or take prepared one and port it to rockbox,if it were possible ,it would have so many uses ( change a tag,or type a text)
21:27:55dropandhopthere is this one: Remote Ticking Poll http://forums.rockbox.org/index.php?topic=1922.0
21:28:21 Quit akaidiot (Success)
21:28:51dropandhopand at the end of this one: Remote Ticking Fix http://forums.rockbox.org/index.php?topic=1872.0
21:30:18 Join Wett [0] (n=Wett@d02v-62-34-192-173.d4.club-internet.fr)
21:31:01 Join miner49er [0] (n=root@82-32-84-226.cable.ubr01.chap.blueyonder.co.uk)
21:31:06 Quit Nixsos ()
21:32:39miner49erHi there :-) I've just updated to CVS and now when rebuild I get a problem with UCLPACK. Any ideas as to why this might be happening?
21:32:55 Quit bobwise ("CGI:IRC (EOF)")
21:33:05LearDevkit?
21:33:19miner49er I'm running Debian
21:34:01 Join DreamTactix291 [0] (n=DreamTac@adsl-149-140-175.bna.bellsouth.net)
21:35:11 Quit Sandking ()
21:35:27 Quit lamed ("CGI:IRC (Ping timeout)")
21:35:34Slasheridropandhop: if you can hear clearly audible ticking from the remote, while on wps and volume is turned to 0, that is the hardware problem
21:36:26dropandhopslasheri- i understand
21:36:38dropandhopbut it seems that there is something else underlying
21:37:45Slasheriit might seem like that, but rockbox can't itself cause that ticking. By default it just communicates faster with the remote lcd and that makes the ticking more audible than it is with iriver fw
21:38:28Slasheriit's not possible to fix the issue completely by software without some compromises
21:40:46 Join hshah [0] (n=hshah@hirenshah.plus.com)
21:41:35dropandhopahh
21:41:37dropandhopgotcha
21:41:43dropandhopso my question for you is
21:42:04dropandhopanyway you can make your great patch more efficient
21:42:12dropandhop(not chew up so much cpu)
21:42:21Slasheriyes, that is possible to do
21:42:30dropandhopthats great
21:42:38dropandhopas it seems that there are more and more people that need this fix
21:42:47SlasheriHowever, it would require double frame buffering and would cause lcd update speed to be much slower
21:42:49dropandhopand there is question about straining the battery life with the cpu
21:43:16dropandhopgot it....is it faster than iriver fw at the moment?
21:43:50Slasheriyes. But with ticking fix enabled, it's not so fast
21:44:11SlasheriAnother way would be to use timer interrupt to drive the lcd, but we don't have any free timers for that :/
21:44:35Slashericpu has only two timers and those are used by the threading itself and lcd dimming
21:45:14amiconnThe threading doesn't use the timer
21:45:18 Quit Febs ("CGI:IRC (Ping timeout)")
21:45:31amiconnThe kernel does, for the tick value and the tick functions
21:46:05amiconnI don't think that it would help performance if we double buffer
21:46:08Slasheriamiconn: oh.. i just again forget our system is non pre-emptive :)
21:46:17Slasheriso it could be possible to steal that kernel timer..
21:46:21Slasherior make it faster
21:46:23amiconnHow should that help?
21:46:26dropandhophmmmm
21:46:30dropandhoptrickey
21:46:47Slasheriamiconn: yield() is too long, but with timer that wouldn't be a problem
21:46:48amiconnIn order to keep up, the timer would need to run _really_ fast
21:46:54Slasheriyes..
21:47:10Slasherias it does with the backlight dimming already
21:47:12amiconn...and then all the context saves/loads will most likely eat more CPU
21:47:31amiconnThe backlight dimming is slow compared to the remote LCD transfer
21:47:38amiconnI mean _S_L_O_W_
21:49:59 Join Jamjam [0] (n=chatzill@user-1391.l4.c1.dsl.pol.co.uk)
21:50:42Jamjamanyone around?
21:50:42 Quit paranoid_android ("CGI:IRC (EOF)")
21:53:57 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
21:54:27 Join xmixahlx [0] (n=xmixahlx@64.122.111.98)
21:56:42Slasheridropandhop: Hmm.. i could try to implement a better solution for the ticking fix when i have time. Ticking is annoying indeed.. But the solution will be a compromise
21:56:46Slasherinow sleep ->
21:56:47amiconnSlasheri: The original delay loop is 12 cpu cycles, the ticking fix extends this to 126 cycles at 45 MHz resp. 354 cycles at 124 MHz
21:57:02amiconnNow tell me how to handle this with an ISR
21:57:16Slasheriamiconn: Ah, that's quite fast still
21:57:27Slasheriwe could transfer several bytes at time..
21:57:28amiconnSave all registers, clock out the bit, restore all registers
21:57:30Slasheribut not too much
21:57:39amiconnThen it would still tick I think
21:57:42Slasherinot just one bit :)
21:57:54Slasherino, if we limit the rate carefully
21:58:07Slasheribut that will be hard indeed..
21:58:19amiconnEven though it's quite fast, it's already too slow for some oggs to play properly
21:58:49amiconnI think it might help if we make the intervals between 2 bits pseudo-random
21:58:57amiconn...but that's to be tested
21:59:24Slasherihmm, maybe. Unfortunately it seems that the clk pulse causes the most ticking
21:59:51amiconnThat's not important...
22:00
22:00:11Slasherii tried separately output data to clk and data lines, and both will cause ticking.. but clk is much louder
22:00:14amiconnI mean, the interval between each two bit changes on the lcd lines should be made pseudo-random
22:00:35 Join Bger [0] (n=Bager@83.142.21.199)
22:00:53Slasherihmm, true. That could reduce the problem
22:00:55 Join Nixsos [0] (n=Nixsos@i62009.upc-i.chello.nl)
22:01:01amiconn...perhaps doing the clk and data change in quick succession _only if_ they change in the opposite direction
22:01:27 Quit Kohlrabi (Nick collision from services.)
22:01:28dropandhopim not as good as you guys...but does it help to observe how iriver fw does it?
22:01:32 Join Kohlriba [0] (n=Kohlrabi@dslb-082-083-140-116.pools.arcor-ip.net)
22:02:00amiconnThat would be a nice little state machine
22:02:04Sanhey
22:02:10Sanwhat does chip-8 do?
22:02:33Slasheribut cu tomorrow, really have to go now :/ ->
22:03:47Jamjamits a game emulator type thing
22:03:53Jamjamlook it up in the wiki
22:03:59dropandhopslasheri - thanks sooo much for thinking about this!
22:04:04dropandhopthis is suuuch a huge thing!
22:04:26Jamjamooh, what was it? i did see the start of that
22:05:52 Part LinusN
22:06:32miner49ersomeone mentioned being able to play music through the simulator a while ago - is that still possible?
22:06:50ioboundit seems my dumb plugin manages to load files fine but it goes into an infinite loop just afterwards in some resampling method
22:07:10ioboundminer49er: mp3 files play fine in the win32 simulator for me
22:09:24 Join arkascha [0] (n=arkascha@xdsl-213-168-118-232.netcologne.de)
22:14:35 Join ender1 [0] (i=ychat@84.52.165.220)
22:15:13 Quit ender` (Connection timed out)
22:16:35wubblawow
22:16:44wubblarockbox booted on my h320!
22:17:31wubblathanks to Matthias Pesch's guide on http://www.rockbox.org/twiki/bin/view/Main/IriverBoot
22:21:04lostlogicDoes anyone know of a simple console tremor audio player?
22:21:50LearTremor, no, vorbis, yes...
22:22:15lostlogicwant a way other than the rockbox simulator to profile tremor, so it has to be tremor
22:22:29LearYou're not the first. :)
22:22:51LearBut tremor includes a simple decoder example, which is good enough (for decoding to a file at least).
22:23:21LearBut modifying ogg123 to use tremor rather than libvorbis/libogg should be fairly easy...
22:23:35lostlogicwas hoping not to have to figure out how to do that.
22:24:04LearUse the decoder example then, good enough for the purpose, and need little or no changes.
22:24:17lostlogictanks!
22:25:06 Quit Lear ("Chatzilla 0.9.68.5.1 [Firefox 1.5/undefined]")
22:27:35 Quit Jamjam (Read error: 104 (Connection reset by peer))
22:28:09 Nick ender1 is now known as ender` (i=ychat@84.52.165.220)
22:29:47 Join Paul_The_Nerd [0] (n=paulthen@cpe-66-68-93-2.austin.res.rr.com)
22:34:21 Join San||Away [0] (n=Test@212.2.163.159)
22:38:12 Quit dropandhop ()
22:47:48ioboundaaargh, the simulator doesn't support pcm_play_data ? :(
22:51:24 Join _aLF [0] (n=Alexandr@mut38-2-82-67-66-128.fbx.proxad.net)
22:51:56 Join Febs [0] (n=40326e83@labb.contactor.se)
22:52:34 Quit San (Read error: 110 (Connection timed out))
22:57:47ioboundif my plugin enters an infinite loop, is there an easy way to abort? (h300)?
22:58:05iobounddo i have to insert a pin in the "reset" thingy
23:00
23:00:31 Quit San||Away (Read error: 110 (Connection timed out))
23:00:36 Join webguest09 [0] (n=45dda986@labb.contactor.se)
23:00:46Rickiobound: yeah, that's about the only way I think
23:00:56*Rick has done that before too (:
23:01:01ioboundok
23:01:23ioboundi got as far as i could on the simulator with a libdumb plugin but since it doesn't support pcm_*() i guess i have to hit real hardware now
23:01:43Rickah
23:01:48 Quit DreamTactix291 (Read error: 110 (Connection timed out))
23:01:54Bagderor fix the sim ;-)
23:02:55ioboundok, uh.. when i build for m68k, i get linking errors, complaining about missing "memset", "memcpy".. do i have to #define memset myplugin_memset and implement it as "rb->memset()", or is there an easier way?
23:03:20ioboundi guess the simulator just linked with msvcrt silently :(
23:03:51Bagderwell, the equivalent of "msvcrt" anyway
23:03:59MoosHello Bagder, did you got a bit of time for the makefile thing for unicode?
23:04:17Bagderyou mean for the compress thing
23:04:22ioboundactually i'm pretty sure its msvcrt, when i load "dumb.rock" in dependencywalker, it actually shows msvcrt imports in use
23:04:28MoosBagder: yes :)
23:04:39Bagdernope, not yet
23:04:49Moosok
23:04:57 Quit muesli_- (Read error: 110 (Connection timed out))
23:05:34iobounddammit, i need to implement pow(), floor() and exp() too. apparently there's some floating point code in libdumb.
23:05:45Bagder!
23:06:01ioboundis it worth the effort? gcc didn't complain, did it assemble for an FPU? will it crash when running on the h300?
23:06:33ioboundsomeone mentioned (very slow) fpu emulation was available in gcc here earlier, anyone know anything more about that?
23:06:37***Saving seen data "./dancer.seen"
23:07:36Bagderthat's what you get
23:07:57Bagderwhen using float/double
23:08:07ioboundoh ok, so it automatically assembled it for emulation?
23:08:17Bagderyes
23:08:24ioboundok
23:08:46iobounddoes anyone know of any free pow() implementations i could use :)
23:11:50 Join mordov [0] (n=Mordov@171.80-202-208.nextgentel.com)
23:12:06solexx_I think the latest daily build introduced a bug
23:12:39solexx_I have some low bitrate mono mp3 files
23:13:17solexx_and the right channel is very, very quiet
23:13:41solexx_it's almost inaudible
23:13:43amiconnBagder: I was fiddling with a perl script to convert the .ucl file to source
23:13:54solexx_other files play fine
23:13:59amiconnIt actually works :) but now I've got a different problem...
23:14:16amiconnsolexx_: That's probably caused by the changes in the resampler
23:14:18 Join muesli_- [0] (i=muesli_t@Bc132.b.pppool.de)
23:14:18amiconnpreglow?
23:14:54solexx_amiconn: who is Thom?
23:15:15Bagderpreglow is thom
23:15:29solexx_thx
23:15:37amiconnI can't use sizeof() on a char[] array which is defined in another source file...
23:15:40 Join BirdFish [0] (i=bfisher@mail.affidavitmaker.com)
23:15:56 Part BirdFish
23:16:27Bagderamiconn: nope, but since you did it with perl you can probably just create a define for the size, can't you?
23:16:45amiconnHmm, you mean in a .h file?
23:17:03Bagderwell, anyway you want
23:17:31 Join DocBrown [0] (n=508fdb72@labb.contactor.se)
23:17:38*solexx_ goes to sleep
23:17:50 Join webguest07 [0] (n=d444e303@labb.contactor.se)
23:26:20 Quit dpassen1 ()
23:29:45 Join len0x [0] (i=user@du213-130-141-194.as15444.net)
23:33:55 Part len0x
23:36:11 Quit Moos ("Glory to Rockbox")
23:36:54 Quit Strath (Read error: 104 (Connection reset by peer))
23:39:03ioboundit seems my gcc compiles "dst->resampler = src->resampler;" to use memcpy(), and then ld fails because there's no "memcpy()" in rbx (?). any hints?
23:39:35 Join Strath [0] (n=mike@dpc674681214.direcpc.com)
23:39:35ioboundmaybe -ffreestanding helps?
23:39:54Bagderwhat are you doing?
23:40:30ioboundi'm trying to write a plugin for playing .s3m and .xm files with libdumb
23:40:37*preglow is back
23:40:42preglowamiconn: what?
23:40:49 Quit miner49er ("using sirc version 2.211+KSIRC/1.3.11")
23:41:13amiconnpreglow: Read back at ~23:12
23:41:19amiconnPerhaps a resampler issue?
23:41:22Bagderiobound: but don't the plugins get built with options like that?
23:41:59ioboundi don't know? maybe?
23:42:13Bagderso again, what are you doing?
23:42:27ioboundtrying to link libdumb with my "dumb.c" into dumb.rock for m68k
23:42:31Bagderyou build plain rockbox with your own plugin?
23:42:41ioboundmore or less, yes
23:42:56Bagderthen it uses -ffreestanding already
23:43:51 Join Jungti1234 [0] (n=jungti12@58.77.81.144)
23:43:51Bagderbut I guess 'dumb' isn't built with it
23:44:03Jungti1234good morning
23:44:07ioboundi'm checking by removing the "@" in dumb's makefile
23:44:17preglowamiconn: could i have the file in question?
23:44:31preglowno wiat
23:44:35preglowi have a mono file
23:44:44amiconnask solexx_
23:44:45ioboundbadger, actually, it DOES use -ffreestanding, but it still fails at link time because it wants to use memcpy() for that struct assignment
23:44:53 Quit Bger (Read error: 104 (Connection reset by peer))
23:45:13BagderI just doubt that is the single place in whole rockbox that does that
23:45:30preglowmy mono files sounds just fine
23:45:32amiconnYou are right
23:45:45ioboundif you want I can paste parts of the "make all" output
23:45:48amiconnSome places had that problem too, e.g. the jpeg viewer
23:45:54preglowouch
23:45:57preglownot with resampling
23:46:01preglowi've broken mono resampling, it seems
23:46:19Bagderamiconn: and it didn't work?
23:46:44amiconnJust replace the operationm in question with an explicit rb->memcpy()
23:46:56ioboundyeah unfortunately i don't have any "rb" in the scope..
23:47:01Bagderright, that's probably the easiest route
23:47:14amiconnThe real problem is that gcc requires even freestanding targets to provide 4 functions
23:47:16ioboundactually i remember reading something about this somewhere, there's an additional gcc option which should fix it
23:47:28amiconnmemcpy(), memmove(), memset() and memchr()
23:47:55 Quit webguest09 ("CGI:IRC (EOF)")
23:48:06amiconn...but rockbox doesn't do so
23:48:12amiconn...except for the core
23:50:37 Quit muesli_- (Read error: 110 (Connection timed out))
23:53:32amiconnBagder: Hmm, now I have the problem to add a proper dependency
23:54:03amiconn...to make sure the perl script gets executed and the .h file generated before trying to compile thedecompressor.c
23:54:15amiconn*decompressor.c
23:54:34Bagdermake sure decompressor.c depends on the .h file
23:54:48amiconnHmm, but how?
23:54:58amiconndecompressor.c doesn't occur in the makefile
23:55:11Bagderyou need to add it
23:55:18amiconnHmm :/
23:55:34amiconnAnd what about the .h file?
23:55:47amiconnIn fact this is similar to lang.c / lang.h
23:55:54 Quit ghode ()
23:55:56preglowmarkun: didn't the gigabeat upgrade file also contain an unencrypted updater?
23:55:59amiconnso if I would understand how this works...
23:56:33Bagderamiconn: in the lang case, the object file is made to depend on the lang file and the genlang script
23:56:44Bagderthe lang.o object file
23:56:50 Quit Wett (Read error: 110 (Connection timed out))
23:57:07amiconnYeah.. but what about the source files that depend on lang.h ?
23:57:58Bagderthat is taken care of by the automatic dependency "generator"
23:58:08 Join muesli_- [0] (i=muesli_t@Bbca0.b.pppool.de)

Previous day | Next day