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

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

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

#rockbox log for 2005-12-09

00:00:54 Quit DreamTactix291 ()
00:01:03b0brLinusN: what is needed to be done to make recording from h1xx work on h3xx? is it a lot of work? isn't the recording hardware the same?
00:01:08 Quit mirak (Connection timed out)
00:01:27 Join DreamTactix291 [0] (
00:01:35LinusNb0br: the mux pin is different, for starters (easy fix)
00:01:47LinusNother than that, i don't really know
00:02:06LinusNi tried to enable recording, just for fun, and it froze the player
00:02:30b0brok, thx
00:03:02 Quit idanm ()
00:03:32b0brpls, who's working on recording on h1xx?
00:03:47LinusNnobody and everybody
00:04:02LinusNanyone who feels like it
00:04:17Moosespecially Andy (thanks for him)
00:05:46b0brhmm sounds mysterious :)
00:11:01markuncan I use a '+' in a filename?
00:12:15markunI combined the 6x13 and 12x13-ja fonts to 6+12x13 and would like to commit it
00:13:44 Part XavierGr
00:14:52andyb0br: sup? :)
00:15:27Midgey34well, I have a class I need to get going to. cya later
00:15:35andyaha.. h300 recording.. actually i'm borrowing a h300 this weekend so I might look into this
00:17:00 Join frederic_ [0] (
00:17:02 Join gromit`` [0] (n=gromit`
00:19:23 Quit gromit` (Read error: 104 (Connection reset by peer))
00:19:56 Quit Paul_The_Nerd ("Chatzilla [Firefox 1.5/undefined]")
00:20:28 Quit petur ("here today, gone tomorrow")
00:21:33 Part andy
00:24:30 Quit b0br ("CGI:IRC")
00:26:02 Quit frederic_ ("Chatzilla [Firefox 1.5/undefined]")
00:26:28 Quit ender` (Read error: 104 (Connection reset by peer))
00:27:59 Quit perplexity (Read error: 113 (No route to host))
00:33:47 Quit Amar ("CGI:IRC")
00:37:21Gaceltime to sleep, cu
00:37:32 Part Gacel
00:42:36 Quit Moos ("Glory to Rockbox")
01:00:48LinusNamiconn: looks like it wasn't the spurious off after all
01:01:02 Join Cassandra [0] (
01:03:26 Join ts-x [0] (
01:10:30 Quit eli_sherer ()
01:11:41***Saving seen data "./dancer.seen"
01:16:39 Quit Kohlrabi ("Leaving")
01:26:41 Quit Rob2222 (Read error: 104 (Connection reset by peer))
01:33:41 Join Benacool [0] (
01:34:48Benacoola little question ;-) ... do the playback speed is supposed to be on the H300 builds ?
01:37:22Benacoolwhat ?
01:37:35LinusNi didn't understand your question
01:37:37Benacoolmy english is so bad ??
01:37:40Benacoolok ;-)
01:38:14LinusNyou want to change the playback speed on the h300?
01:39:12Benacoolthe playback speed control ... is it on the h300 too now ?
01:39:19Benacoolyea it's what I meant
01:39:22Benacoolwhere ?
01:39:26Benacooli don't found it lol
01:39:28LinusNhold ON and use the + and - buttons
01:39:52LinusNin the wps
01:41:35golf7nice linus
01:41:35 Quit Benacool (Read error: 104 (Connection reset by peer))
01:41:37golf7just flashed today
01:41:40golf7im loving it
01:41:45golf7keep it up
01:42:25 Join Benacool [0] (
01:42:44golf7you know if its possible to change the speed, but not the pitch on the h300
01:42:52golf7it might be a hardware limitation
01:43:18golf7like doesnt it have to have 2 mp3 decoders or something
01:43:40Benacoolthey are some distortion when you slow down pitch
01:44:08LinusNgolf7: it takes some serious DSP algorithms to change speed without changing pitch
01:45:25LinusNBenacool: yes, the resampling algorithm has a glitch
01:46:37markunSetting the display boolean in wps_data_load to false is claimed to fix a crash with loading WPS'
01:46:41BenacoolIt's normal (know bug) that when you hold a button for a long time that the song skip ?
01:46:57markunLinusN: do you know anything about it?
01:47:55LinusNmarkun: i saw that, yes
01:47:59markunWith grep I discovered that with this change in filetree.c the display boolean is always false now
01:48:30markunIs it used to dump the WPS on the screen after loading?
01:50:59markunIf it crashes the player maybe it should be removed, or do people find it a useful feature?
01:51:51LinusNwe should remove it
01:52:41markunok, I will then
01:53:06BenacoolLinusN: Is it normal (know bug ;-)) that when you hold a button for a long time that the music skip ?
01:53:17markunbtw what do you think about loadable keyboard layouts?
01:53:32LinusNBenacool: when does it do that?
01:54:24LinusNBenacool: what makes you think that it is the actual holding of the button that is the problem?
01:54:41Benacoolyou know when you search a music file and you scrool down for exemple
01:55:03LinusNthen i believe it is the scrolling that causes it, not the holding of the button
01:55:17Benacoolin the brownser
01:55:23 Quit San||Away (Read error: 110 (Connection timed out))
01:56:53LinusNmaybe i sound like a nitpick here, but saying "it happens everytime you hold a button" isn't very helpful
01:56:55 Quit xmixahlx ("blah blah blah")
01:57:04Benacoolwhen i change the pitch it do teh same thing but the music skip a sec., play a sec., etc but in the browser it stops untill a sopt holding the button
01:57:17Benacoolsorry ^^'
01:57:17 Quit ts-x ("CGI:IRC (EOF)")
01:57:54LinusNit's always better to find a specific situation when it happens, so i know how to reproduce the error on my player
01:58:26LinusNi can tell you that it is the LCD update that causes it, it is currently very slow on the h300
01:58:27Benacoolso the browser is the best place to reproduce it
01:58:50Benacoolthanks ^^
01:58:56LinusNthat's also why the scrolling continues when you release the button
01:59:08LinusNand the wps flickers
02:00:05 Join cool [0] (
02:00:05 Quit Benacool (Read error: 104 (Connection reset by peer))
02:00:18 Nick cool is now known as Benacool (
02:03:48 Join webguest23 [0] (
02:04:01golf7i dont think i can go back to the 3-4 second gap inbetween songs, im loving the crossfade
02:04:07lostlogicis there any followup I have to do after submitting the profiling patch to the sourceforge tracker?
02:05:34webguest23Hi, has anyone noticed that scrolling through files on h1xx is now not eating button events fast enough? (for me, at least). Bleeding edge from earlier today. I *think* it happened first after unicode, but am not sure.
02:05:54LinusNyes, harass the developers in this channel until we cave in and commit it
02:06:35LinusNwebguest23: it is quite likely to be because of the unicode build
02:07:03webguest23It sounded like a likely explanation
02:07:28webguest23But then, I'm not going to jump to conclusions when I really have no idea
02:07:48LinusNwebguest23: which font? is it different if you play music and if you're not?
02:08:17webguest23Hm, let me test..
02:09:19 Join XavierGr [0] (
02:10:38webguest23Font is 6x10
02:10:52webguest23When playing music, the music will start skipping (guessing the buffer empties)
02:11:10webguest23But the scroll will be fine.
02:11:21webguest23When not playing, the scroll will lag behind
02:11:43webguest23Seems reasonable, because of the lower cpufreq, I'm guessing?
02:12:42LinusNwell, it looks like the scroll takes too much cpu
02:13:23webguest23I'm not even using any glyphs outside of latin-1, which I set as my default codepage
02:13:56LinusNi don't think that matters
02:14:17webguest23I don't know if that would even be an issue, I'm kinda guessing those would be loaded when entering the dir?
02:15:04LinusNi think so
02:15:29webguest23It's a large dir - it filled the dir buffer at the default size
02:16:23webguest23It does not seem to happen with the default font.
02:17:22LinusNi believe it takes slightly longer to render a font which height isn't a multiple of 4
02:17:49webguest23Let me try with a non-unicode font
02:17:53webguest23say, snap
02:18:12markunLinusN: the quickmenu is not displayed for non-latin1 users because it uses the internal font
02:19:03webguest23It's not happening with snap, but is with 6x10
02:19:08amiconnLinusN: I don't think so. The font height itself shouldn't matter at all on displays with >= 8 bit depth
02:19:27LinusNamiconn: it's h1x0
02:20:11LinusNmarkun: oops
02:20:13webguest23That doesn't seem to be it though.
02:20:24amiconnwebguest23: Do you have the remote connected?
02:21:22webguest23There's no noticable difference between snap and the default font. With 6x10, it's all wrong.
02:22:11webguest23unifont is no better, it seems
02:23:13markunThe quickscreen looks still ok on a H1x0 with unifont
02:23:35markunshall I change the quickscreen to use the UI font?
02:24:16amiconnIt will look all wrong on archos with any font >8 px
02:24:30amiconnIt's already tight with the default in some languages
02:24:53amiconnThe default fonts should get unicode support, I'd say
02:25:24markunsome languages like korean will never have a small font size..
02:25:45amiconn8859-1 code page part compiled in (or just ascii if we want to save some more), and the full version on disk
02:26:08amiconnThe default could be a bit larger on iriver, imho
02:28:01LinusNtime to sleep
02:28:20LinusNnite all
02:28:24 Part LinusN
02:29:42markunamiconn: happy with the addition of unicode to rockbox?
02:30:03amiconnI think it's a move in the right direction
02:30:19amiconnStill some quirks to iron out of course
02:30:32amiconn(like the default font problem)
02:30:48 Quit actionshrimp ("a bird in the bush is worth two in your house")
02:31:22amiconnI'd say we will need a method to mark possible font/language combinations, to protect users from nasty surprises
02:31:25markunyes, and educating people. Saab will write a wiki page with his experiences, I hope that lowers the number of questions
02:31:32amiconn(like a menu consisting of all spaces)
02:32:02amiconnSame goes for the built-in player 'font'
02:32:09markunyes, that would be a very useful feature.
02:32:19amiconnSome languages won't be possible there, of course
02:32:20lostlogicLinus's commit of the higher max frequency, does that work w/o a new bootloader, or will it also be in v4?
02:32:35amiconnlostlogic: No new bootloader needed
02:33:35markunamiconn: how do I enter the quick menu on the recorder?
02:33:55amiconnThere are 2, on F2 and F3
02:34:58markunyou were right, the quickscreen with unifont is a disaster :)
02:36:05markunWith 6x13 it's still normal.
02:37:00markunwell, almost. The lower row of pixels from the clock is gone
02:39:02XavierGrOMG daily biulds page: Buidl Expected at 1.33.17 Time now: 1.38.00 is that normal?
02:41:06 Quit linuxstb ("CGI:IRC (Ping timeout)")
02:46:06markunXavierGr: I think the compile time differs a lot because of ccache
02:49:26XavierGrcan i use this when i compile?
02:49:44markunI think you can, but I've never used it.
02:55:28XavierGrAh it seems that it will not run on the Devkit.
02:55:40XavierGrwhen I configure it some commands are not found.
02:55:57 Join pengo [0] (i=xtofu@
02:56:12webguest23What *is* the devkit, that cygwin is not, anyway?
02:56:32XavierGrIt is a very light version of cugwin
02:56:54pengois there a way to save an arbitrary bookmark (without the save-on-stop thing) ?
02:57:21pengoand question 2, why isn't it in the menu?
02:57:22XavierGrpress and hold select (or play on the remote) when you are in wps
02:57:32XavierGrthen select Create bookmark.
02:57:41XavierGrno need to be in the menu
02:57:42pengothanks.. lemme try
02:58:11lostlogiccoldfire SRAM == IRAM?
03:00:25markunlostlogic: yes
03:02:24lostlogiccoldfire has no data cache, hence the codec data > IRAM is most effective, and codec code > IRAM is sometimes slower.
03:05:55 Quit lostlogicx (Read error: 110 (Connection timed out))
03:06:20XavierGrarhg I found the missing .exe files and added them to bin folder. But still configure of ccache fails. :((
03:09:35 Join Speedforneed [0] (
03:10:43SpeedforneedIs it just me, or does the H340 (Korean) have an issue with the LCD backlight being to bright when Rockbox is enabled?
03:10:51webguest23I guessing a full cygwin environment would be better suited
03:11:04webguest23XavierGr: ^^
03:11:31XavierGrLCygwin Utilities cool little zip that contains many cygwin .exe programs.
03:11:38XavierGrFor those that run the Devkit.
03:11:44***Saving seen data "./dancer.seen"
03:12:05XavierGrwebguest23: I could use cygwin but it is a major hassle to build the crosscompiler my self...
03:12:32 Join rubberglove [0] (
03:12:53 Quit Benacool (Read error: 110 (Connection timed out))
03:12:57webguest23Why? It's all in the wiki
03:13:26rubberglovehello any and all. i've been getting errors with buildzip today:
03:13:32webguest23And it's not like you have to do it more than once.
03:13:33rubberglove" 9165 [main] perl 2028 fork_copy: linked dll data/bss pass 0 failed, 0x412000."
03:15:50SpeedforneedHmm... I'm guessing not.
03:18:50golf7the bleeding edge builf from today is very good, however the daily build, not good
03:19:22SpeedforneedIf you are talking to me, I use the bleeding edge stuff.
03:20:37 Join Maxime [0] (
03:20:37 Quit Maxime` (Read error: 104 (Connection reset by peer))
03:23:03golf7ok guys i got a question
03:23:15golf7can i use the rockbox USB to update the .rockbox folder?
03:23:20lostlogicgolf7: yes
03:23:22golf7cause isnt it like using some of the files
03:23:27golf7just a thought
03:23:39lostlogicgolf7: if you change the boot stuff, it'll ask you if you want to ROLO afterwards.
03:23:54golf7ROLO = ????
03:24:05lostlogicrockbox loader, quick reboot basically
03:24:30golf7and when i update the .rockbox folder will i have to redo all my settings?
03:24:36golf7cause i spent like 30 minutes getting em all right
03:24:56lostlogicnot if the config layout hasn't been updated
03:25:01lostlogicbut to be on the safe side, save a config file first
03:25:09golf7thanks man
03:25:14golf7just flashed today and i love it
03:25:18golf7just trying rockboy
03:25:25golf7gonna be killin it in class tommorrow
03:25:36golf7my 60 gig h340 with a 2200 mah battery is not ever more beast
03:25:39golf7with rockbox
03:25:48lostlogicwow, that's a hellofa toy :):)
03:25:52golf7hell yah
03:26:00golf7just wish i would have gotten a newertech battery
03:26:11lostlogicaren't they only 2100?
03:26:18golf7i cheaped out and got a cheap ipod battery and not the newertech
03:26:24golf7yeah they are only 2100 but they ARE 2100
03:26:30golf7this one i got rockbox says is only 1300
03:26:35lostlogicno no
03:26:38lostlogicyou set that
03:26:39thegeekyou have to set it
03:26:49thegeekso that rockbox can estimate remaining time
03:26:54lostlogicit can't detect the battery size
03:27:00lostlogic1300 is the stock battery
03:27:06lostlogicso it will estimate time based on stock battery
03:27:18PaulJis there a setting to enable tis rolo thing? because my h320 doesnt ask if i want to reboot after i have copied the new rockboxfolder on it.
03:27:21golf7i knew that
03:27:42golf7im sure its not 2200 though
03:27:51golf7im thinking like 1700-2000 mah
03:27:52lostlogicPaulJ: I may have noticed that you have to be in the root folder when you start USB for it to work right
03:28:01lostlogicor perhaps it just detects when you load the root folder
03:28:03lostlogicnot sure which
03:28:21lostlogicgolf7: batteries always lie ;)
03:28:21webguest23The latter
03:28:23golf7ill just load the .rockbox folder with iriver firmware to be on the safe side till v4 comes out
03:28:27lostlogicwebguest23: ty
03:28:47golf7but hard drives dont, well they kinda do, haha, but 60 gigs is till 55.7 gigs
03:29:13lostlogic60g = 60*10^9B != 60*2^30
03:29:19PaulJthat explayns it, i was never in the root when using usb
03:29:28webguest23Well, it's the computer world's own fault for bastardizing the prefixes
03:29:44golf7i know
03:29:59golf7its so gay
03:30:04golf7they should have been
03:30:07golf71000 increments
03:30:07 Quit rubberglove ("CGI:IRC (EOF)")
03:30:12golf7would have been som much easiery
03:30:16golf7but yet its powers of 2
03:30:17webguest23I'm all for using GiB.. just please don't say the name
03:30:33lostlogicwebguest23: so true!
03:30:55golf7you think its worth it to upgrade to get the 124 mhz back =P
03:31:01golf7on h300?
03:31:06lostlogicgolf7: I'd say yes.
03:31:18lostlogicgolf7: bettery battery life with higher cpu speed, believe it or not.
03:31:28golf7worth an upgrade
03:31:39 Part Speedforneed
03:31:41golf7ive been running this thing since 3:30 pm est
03:31:45golf7and its 9:30 est
03:31:48golf7and it hasnt stopped
03:31:56golf7on like 91% when i started
03:31:57golf7and its on
03:32:02lostlogicplaying mp3s?
03:32:05golf7exactly 50% now
03:32:12golf7LAME −−alt-preset standard
03:32:15lostlogicyou and your giant battery
03:32:38*golf7 hugs his 60 GiB hdd and 2200mah battery in his h360
03:32:58golf7with an inskin
03:33:05golf7i modded the hell out of it
03:33:13golf7only thing left was rockbox, and now i got it
03:33:15golf7and killing it
03:33:20lostlogicgolf7: not GiB ;) GB
03:33:30golf7bigger than yours =P
03:33:45lostlogichehe, I'm not even using half as is, otherwise I'd have upgrade
03:34:13golf7i like that you can output at 0db cause i dj, and i would always have to run the iriver side of the mixer about 2-3 steps higher than the cd player side
03:34:32golf7then when i play vinyl, it is always too loud, and created a bad feedback
03:34:37golf7so that solved that problem =P
03:34:47lostlogichope it doesn't clip
03:35:19 Join Subterranean [0] (
03:35:24golf7doesnt sound like it on headphones
03:35:27lostlogiccan anyone suggest a good way to estimate how large a function would be after compilation?
03:35:36lostlogicgolf7: 0db on headphones!? OUCH?
03:35:41golf7even if it does, ill play with it and get it just below clipping =P
03:35:47golf7yeah, i barely could stand it
03:35:52golf7i stood it for like 2 secs
03:35:57golf7i have some sony ex-51's
03:36:07golf7and i could only stand one in for like 2 seconds
03:36:16golf7ive been listening at -25 db
03:36:22golf7and thats a little loud
03:36:23lostlogicget some Shures, at -40dB they'd be too loud ;)
03:36:36golf7i was debating some shure e2,3,4,5c's
03:36:46golf7but i aint got 100-500 bucks to blow on earbuds
03:36:50lostlogicI have E3 (gotta get them repaired, because I broke them) love them.
03:41:11 Join amiconn_ [0] (
03:41:37golf7dang guys
03:41:46golf7i just realized that rockbox supports anti trickle charge
03:41:53golf7to preserve my new battery =)
03:42:24 Join lostlogicx [0] (
03:42:29golf7however i dont think charging is implemented in h300 yet
03:42:37lostlogiccharging is done in hardware
03:42:45lostlogicanti-trickle just involves turning off the charging hardware
03:43:46golf7i was just reading the manual on the wiki and just realized it though
03:43:49golf7thats pretty cool
03:44:00golf7cause thats what ruined my other battery
03:44:04golf7i charged it ALL THE TIME
03:44:14golf7this one, no charge unless less than 50%
03:44:31golf7and only leave it on there until it gets done, then take it off charger
03:45:05lostlogicLiION doesn't like deep cycling, so charging from 50->100 is a pretty good condition for the battery, but yes, being kept at 100 is also not good.
03:46:42golf7i kept it at above 60% at all times
03:49:39webguest23I don't think Rockbox controls anything charging-related on the iriver platforms
03:50:43lostlogicah, wasn't sure if that feature was non-iriver only. my bad.
03:52:32golf7neither was i
03:52:41golf7i just read it on the manual
03:52:51golf7even if it isnt, good thing for the arcos
03:53:02webguest23Only on the Recorders, too
03:54:44webguest23markun: The playlist viewer is having issues with non-ascii characters it seems? Is this right?
03:56:08golf7ive noticed it takes a while to go back the first directory when in the filetree on rockbox
03:56:12golf7like from the WPS
03:56:22golf7hit NAVI then left
03:56:29golf7and it takes a little while
03:56:45golf7then you can go faster than iriver firmware, maybe i gotta set the hdd spin up or something
03:57:30lostlogicwhat's the difference between IDATA_ATTR and ICONST_ATTR?
03:57:39webguest23golf7: Better yet, use the directory cache
03:57:45golf7true that
03:58:14webguest23It will eat a bit of ram, but browsing is lightning fast
03:58:16golf7will it cut down on boot time?
03:58:39golf7and could it make rockboy any slower =P
03:59:22 Quit amiconn (Read error: 110 (Connection timed out))
03:59:22 Nick amiconn_ is now known as amiconn (
03:59:59webguest23golf7: It will increase boot time slightly
04:00:21golf7but nothing like the iriver firmware, ugghghghg
04:00:33webguest23Not at all, it's really fast
04:00:40Jungti1234markun is sleeping.
04:00:50 Join Shaikoten [0] (n=shaikote@
04:00:58golf7o well, ill holla man, hold it down guys
04:01:23ShaikotenI just came here to give whoever fixed high quality vorbis a huge thanks :)
04:01:57ShaikotenLostlogic I think it was?
04:04:27lostlogicwasn't me.
04:04:47lostlogicShaikoten: they just bumped the cpu speed up so it can handle it, vorbis is still way too slow... I'm working on it ATM
04:07:45ShaikotenCool. Need a hand with testing of it or anything?
04:11:04lostlogicShaikoten: if I ahve a patch, I'll let you know, right now, I'm doing some cache math and stuff so it's a lot of work up front before any reward.
04:12:00ShaikotenI know what you mean, that's why I stay away from coding these days :P
04:12:03golf7im gonna be so happy when i actually learn c++ and C#
04:12:14golf7cause then i will know what yall is talking bout
04:12:59lostlogicgolf7: learn a real language like C or m68k assembly ;)
04:13:12lostlogic(this is the guy who programs Java all day talking)
04:13:48lostlogicso, how much of the code cache do we think I can expect vorbis to be able to hold onto?
04:14:40*Shaikoten eyeshifts.
04:19:02golf7c sounds good =P
04:19:04lostlogicI'll pretend that 80% of the code cache can be held by the active codec for now.
04:19:53lostlogicmake that 70 :-D
04:22:34 Quit Jungti1234 ("bye")
04:22:59XavierGrlostlogic: Your research for vorbis optimizations will be universal (rockbox wise), right?
04:23:48 Quit Subterranean ()
04:24:09DreamTactix291lostlogic: you are the man then :D
04:24:09 Quit pengo ("Curious? Open binary with Qubero")
04:24:31XavierGrlostlogic do you have any knowledge of ASM?
04:24:35lostlogicI'm optimising for coldfire.
04:24:40lostlogicXavierGr: 4 classes or so in college.
04:24:54lostlogicXavierGr: but converting to assembly isn't my first step right now.
04:25:09XavierGr4 classes you mean subjects or hours?
04:25:22lostlogic4 subjects involved assembly, 2 m68k and 2 mips
04:25:44lostlogicdoes that mean I remember it well? no.
04:26:00XavierGr2 m68k? OMG are your classes THA specific??!!!
04:26:12lostlogic61k of data in the vorbis codec, what and how much should go in IRAM?
04:26:40lostlogicXavierGr: the classes weren't all about m68k assembly, they were teaching other fundamentals on m68k hardware and therefore in m68k assembly.
04:27:18XavierGrso assembly is a whole different matter on each target?
04:28:12lostlogicquite different, yes. MIPS is 3 address, small instruction set, m68k is 2 address, large instruction set
04:28:25lostlogicI know nothing of ia32 or AMD64
04:28:50Midgey34lostlogic: you're american right?
04:29:06XavierGrhmm I hope I knew how to optimize code. Optimizing is A-Z for embedded programming....
04:29:10Midgey34where did you go to school?
04:30:39Midgey34is that Illinois or Indiana?
04:30:47lostlogicIllinois Institute of Technology
04:31:15Midgey34ah, that's moderately close to me
04:32:05lostlogicit's a decent school −− very flexible degree programs, can get a CS degree that is _very_ good, or one that is _very_ useless at the same school ;)
04:32:45Midgey34I see, I think I might end up at UofM or Northwestern
04:32:52Midgey34I haven't really decided
04:33:09lostlogic*nod* (I was CE actualy, not CS thougH)
04:33:18XavierGrright stkov error. On filescrolling plugin. What did I do to make the stack overflow?
04:33:24lostlogicI think CS has switched to IA32 assembly.
04:33:48XavierGrDoes CS stands for Counter-Strike LOL :D
04:33:52 Join ts-x [0] (
04:34:24lostlogiccomputer science vs computer engineering.
04:34:46XavierGrI know just kidding...
04:35:01lostlogicbut counter strike is one of the degree programs available at IIT :-D
04:35:11ts-xShaikoten: You still here?
04:36:16XavierGrLoading a theme takes forever!!!!
04:36:45Midgey34XavierGr: part of the time is loading off the HDD
04:36:54Midgey34spin-up time and whatnot
04:37:11XavierGrI know this is also due to many little picture icons.
04:37:25ts-xGuess not...if you check back, I'm interested to know what quality level of ogg you've had success with.
04:38:02Midgey34is amiconn's (is that right?) combined bitmaps patch ever going to be committed?
04:39:05XavierGrI thought it was Firefly's work no?
04:39:19Midgey34it may be, that would make sense
04:41:34ts-xMy files are all soon as I've got confirmation that they'll play without issue, I'm flashing! Rockbox that is :)
04:42:01lostlogicShaikoten: you have q9s that play?
04:43:07ts-xThat's what I'm trying to discern...I'm assuming his post related to LinusN's cpu commit earlier?
04:43:52lostlogicyeah, that should make it better, but it'll still be boosting way too much.
04:43:54ts-x04.01.23 # <Shaikoten> I just came here to give whoever fixed high quality vorbis a huge thanks :)
04:44:36ts-xYou're working on optimization right?
04:44:56ShaikotenLostlogic: Yes, Fantomas - Suspended Animation
04:45:09ShaikotenPlays smooth as butter, even gapless.
04:45:56ShaikotenGeneral question: does USB charging work for H300s yet? (or will it ever?)
04:46:13ShaikotenI don't have an American wall wart for mine, unfortunately.
04:46:40Midgey34Shaikoten: I'm fairly certain it doesn't yet, but it will in the future
04:46:54ShaikotenOK, cool.
04:47:23ts-xThe iRiver wall warts are hard to come by here unfortunatley, I've been trying to pick up a spare for some time
04:47:41ShaikotenI bought my H340 int from a Hong Kong store and had to send it, and my charger back to China for service.
04:47:43DreamTactix291i should get a spare H1xx
04:47:48ShaikotenThey sent my player back, but not the charger.
04:47:49lostlogicts-x: yes, have some theories ATM
04:48:04DreamTactix291i got a spare charger from iriver
04:48:28ts-xlostlogic: Excellent, your work is much appreciated
04:48:31 Quit yngwi (Read error: 104 (Connection reset by peer))
04:48:45DreamTactix291i live in the US and bought my H140 from the UK so i had to get another charger if i wanted to charge my H120 and H140 at the same time
04:49:06DreamTactix291finding UK to US pin adapters here is a pain
04:49:12DreamTactix291US to UK wasn't so bad
04:49:56ts-xIt was hard for me to let go when I ebayed my H120 to purchase an H340
04:50:12ts-xIt's an excellent dap
04:50:30ShaikotenI'm trying to put my Rio Karma behind me.
04:50:39DreamTactix291i didn't really want an H340
04:50:42ShaikotenI love the thing, but it likes breaking too much.
04:50:43DreamTactix291so i fought hard to find my H140
04:51:37ts-xI ran out of space and was faced with that decision
04:52:04ts-xVideo and the prospect of album art swayed me
04:52:18ts-xBut that optical out on the H1xx series is sweet
04:52:23DreamTactix291album art i'd never use
04:52:29DreamTactix291and i have a PMP for video
04:52:37DreamTactix291so i wanted to stick with the H1xx line
04:52:57DreamTactix291plus i really much prefer the design of the H1xx over the H3xx
04:53:03DreamTactix291but that's of course personal preference
04:53:51ts-xI actually like the H3xx control layout better, but I agree the form factor of the H1xx is superior
04:54:06DreamTactix291the H3xx prototypes looked a lot nicer
04:54:19DreamTactix291i was especially fond of the silver one with the red and green LEDs like the H1xx has
04:54:39ts-xThen there's the H10...what have they done
04:54:48DreamTactix291i know :|
04:55:28DreamTactix291the H10 probably wouldn't half as bad as it is if i had better firmware
04:55:39DreamTactix291i mean it is a pretty looking thing
04:55:46DreamTactix291but that's all it is right now
04:55:53DreamTactix291and i hate the concept of MTP too
04:55:55ts-xAgreed, and ogg + video
04:56:09DreamTactix291considering all my lossy audio i make is vorbis
04:56:11DreamTactix291i have to have it
04:56:23DreamTactix291Lancer :D
04:57:03 Nick Lost-ash is now known as ashridah (n=ashridah@
04:58:34ts-xGetting late here in the later
04:58:44 Quit ts-x ("CGI:IRC")
05:00:12 Quit webguest23 ("CGI:IRC (Ping timeout)")
05:02:59lostlogicfirst optimization whack of the day about to get tested on live ogg 224kbps abr torture test @90mhz
05:05:21DreamTactix291fun fun
05:05:23ashridahlostlogic: you're too late. they've already solved the issue requiring the cap :)
05:06:32lostlogicashridah: I know, but I know how it used to behave at 90, so I just reversed Linus' patch for qualitative testing.
05:06:55*ashridah nods
05:06:59ashridahgood to optimise anyway
05:07:26lostlogicI mean afterall I just spent like a week making a profiling patch for _some_ reason, I hope!
05:07:44DreamTactix291and since it's general coldfire optimisations it'll benefit more than just the H3xx
05:08:49DreamTactix291and as a Vorbis user the more optimisations the better for me
05:09:00lostlogicDT, you and me both.
05:09:25DreamTactix291it runs well on the H1xx currently
05:09:28DreamTactix291but it could always be better
05:09:39*ashridah tosses his chips in that pile too
05:09:40DreamTactix291what exactly was the problem on the H3xx?
05:09:50ashridahDreamTactix291: it doesn't run well enough
05:10:03DreamTactix291but i mean what caused the CPU cap anyway
05:10:06DreamTactix291what had to be fixed?
05:10:10ashridahan unrelated bug
05:10:25ashridahwhich, subsequently, has been recently patched.
05:11:17DreamTactix291well what is realtime for Vorbis on the H3xx now?
05:11:22DreamTactix291or better than realtime
05:11:45***Saving seen data "./dancer.seen"
05:12:07*ashridah blinks
05:12:26DreamTactix291i mean what settings are too much for it to handle right now
05:12:32DreamTactix291i'd test myself but don't own the right unit
05:12:35DreamTactix291just out of curiosity
05:13:06DreamTactix291lostlogic: are you interested in getting a vorbis encoder that produces probably the highest bitrate spikes from me?
05:13:15lostlogicno code in IRAM, only DATA leaves about 12k, and is a little bit worse than what is currently in CVS.
05:13:29lostlogicDT: no need yet.
05:14:14DreamTactix291this was an experimental encoder (actually two of them) made about 18 months ago i guess
05:14:35DreamTactix291they threw bits like crazy to see if that would solve some of the issues Vorbis had at the time
05:16:18lostlogic*nod* don't think that's what we want to optimize for...
05:17:08 Nick ashridah is now known as Lost-ash (n=ashridah@
05:17:27 Join Jungti1234 [0] (n=jungti12@
05:18:33Jungti1234How about feel?
05:19:34DreamTactix291well i was asking because i wasn't sure if that would cause CPU usage to spike on decode
05:19:55DreamTactix291you know, as a worst case scenario
05:34:04 Quit PaulJ (".")
05:38:52lostlogicDreamTactix291: it does, but now that we have profiling, causing spikes isn't the only way to test things ;)
05:39:49DreamTactix291well if you're interested in acquiring it just ask
05:40:01DreamTactix291lots of 500-700k frames even for -q6 :)
05:49:19 Quit Jungti1234 ("bye")
05:49:26 Quit hardeep ("[BX] We are BitchX of Borg. You will be assimilated. Using mIRC is futile.")
06:09:38 Join ReKleSS [0] (
06:18:00lostlogic*sigh* I still haven't gotetn it to not skip on whistles during applause, but it's noticeably improved.
06:18:08lostlogic(at 90mhz, I'm sure it's fine at 124)
06:18:52lostlogicthis song required 99% boost before, and is down to 94%, and profiling shows shifts of time-in-function data as well. *parties*
06:19:40lostlogicooh, down to 92%!
06:27:20lostlogicneed to deprofile inlines and get more profiling data, I think
06:30:24DreamTactix291lostlogic: well whistles and applause cause the worst case of bitrate spikes
06:30:39lostlogicyes, I know ;)
06:31:11lostlogicI think I got about a 4% improvement out of it, which makes me pretty happy for now, and I'm able to start tagging things for more improvement.
06:32:34lostlogic3 functions want inline-assemblyficating.
06:32:57lostlogicand now I sleep.
06:39:54 Part Midgey34
06:40:13 Quit RotAtoR ()
06:59:49 Join tvelocity [0] (n=tony@
07:11:47***Saving seen data "./dancer.seen"
07:22:21 Join webguest38 [0] (
07:23:11 Join harmann [0] (
07:23:11 Quit webguest38 (Client Quit)
07:37:31 Quit harmann ("CGI:IRC (EOF)")
07:39:04 Join perplexity [0] (n=joust@
07:39:15 Join Jungti1234 [0] (n=jungti12@
08:09:44 Quit Jungti1234 ("bye")
08:28:48amiconnlostlogic: The difference between IDATA_ATTR and ICONST_ATTR is just to make the compiler happy, there is no other difference.
08:29:49amiconnWhen we had IDATA_ATTR only, it wasn't possible to have both 'const' and non-'const' declared data in IRAM, as the compiler stopped with a section type conflict error
08:31:27amiconnmarkun: While I agree that displaying the wps after load isn't exactly necessary, this 'fix' looks a bit strange to me
08:31:54amiconnSomething in the line of 'this feature is buggy, so remove it'.
08:38:29 Join ender` [0] (i=ychat@
08:43:17preglowooh, h300 frequency is back?
08:49:04 Join YouCeyE [0] (
08:53:20amiconnpreglow: Yup
08:53:45 Join _FireFly_ [0] (
09:05:02*Bagder tortures his machine
09:05:03*preglow was afraid there was more optimising afoot
09:05:56preglowcoffee time!
09:06:04Bagdermmmmmm cooooffeeeeee
09:06:54Bagder3000 connections to local HTTP server suck some memory... :-)
09:11:49***Saving seen data "./dancer.seen"
09:17:38 Join Zagor [0] (
09:18:15 Join aliask [0] (
09:22:03preglowwho would have thought it?
09:23:43*perplexity thinks about USB-OTG driving a simple USB-ETH dongle and running a web server as a rockbox plugin..
09:24:31Bagdernow _that_ is useful
09:25:30perplexityor better still.. a small smb server.. turn the h340 into a small NAS device :)
09:25:59perplexityNow if only we could get it to make the coffee also
09:32:57 Join Zak1392 [0] (
09:33:21Bagderwe should just port rockbox to a coffee machine
09:33:33Zak1392hey guys
09:33:38preglowshould we use java for that? :P
09:33:46*Bagder feels pain
09:34:01Zak1392anybody know why bmp's blink in wps's?
09:42:51Bagderwe think it looks nicer that way
09:43:01Bagderit is a bug
09:43:07 Join Vlad0man [0] (
09:43:14Bagderor no it doesn't it only looks like that to you
09:43:25Bagdertake a pick
09:43:31preglowthree options, you're too kind
09:48:18 Join linuxstb [0] (
09:51:05 Join einhirn [0] (
09:58:29 Quit preglow (
09:58:51Zak1392what causes it?
09:59:33 Quit Vladoman (Read error: 110 (Connection timed out))
10:02:40_FireFly_Zak1392 we can't say anything about it without more information from you which device ?? on which screen the bmp flicker ??
10:03:01SlasheriHmm, in fact on remote wps bmps do flicker considerably
10:03:20Slasheribut on main wps they don't
10:03:57Slasherifor example with iamp rwps and paused playback
10:04:10 Join thomjoha [0] (
10:04:13Slasherithen the paused message blinks
10:04:34_FireFly_the bmp on my remote only flickers when ticking-fix option is set
10:04:36Zak1392on my h340
10:04:43 Nick thomjoha is now known as preglow (
10:05:15Slasheri_FireFly_: ah, then that is probably a problem with the lcd driver
10:05:48Slasheriit might erase the entire screen but that shouldn't be necessary
10:06:49_FireFly_Slasheri i think the "problem" is caused due the reduced transfer-speed for remote-lcd-commands
10:06:50amiconn_FireFly_: Is the flickering BMP a background image, and is/are there scrolling line(s)?
10:07:32_FireFly_amiconn no scrolling lines on the pos where the bmp is set
10:07:36Slasheri_FireFly_: yep, but why that should cause it? I can't think other reason but that the lcd is entirely cleared before redrawing it
10:07:58amiconnSlasheri: The lcd driver doesn't erase anything by itself. However, the problem might have to do with overdraw
10:08:08SlasheriAh, that's true
10:08:23amiconn_FireFly_: Is the flickering BMP located to the right of any text?
10:09:16_FireFly_the bmp is between two texts left is the volume and on the right side it's the battery-value
10:10:12_FireFly_the other side -effect of the remote-ticking fix is that the bmp gets slighty brighter when holding e.g. the vol-up button
10:12:00amiconn'Educated' guessing: An lcd_update seems to happen before the BMP is redrawn, and if the text to the left changes, this causes short periods where the bmp isn't displayed,
10:12:18amiconnso it flickers/gets brighter when the text to the left changes
10:12:23 Join Jungti1234 [0] (n=jungti12@
10:12:46_FireFly_could be
10:15:21linuxstbamiconn: What was your idea for implementing background images? Are you planning on doing it?
10:15:38amiconnIt's on my list.
10:16:14amiconnMy idea is to allow to set a fullscreen background image, which replaces the background colour/shade while in effect
10:16:47amiconnIt would be handled directly in the driver, hopefully with very little overhead
10:18:53linuxstbI guess the main overhead would be the second copy of the framebuffer needed - which I assume will need to go into the slow SDRAM
10:23:02 Join LinusN [0] (
10:24:30preglowis the h300 lcd driver still painfully slow?
10:26:37LinusNamiconn: to enable the fast lcd transfer mode, we need to make sure that the transfers are in 4-word chunks
10:26:42preglowi wonder if ours is equally slow
10:26:54preglowdoesn't look like it's super fast, but then not incredibly slow either
10:26:59preglowhow many pixels does h300 have?
10:27:10Bagder220x176 isn't it?
10:27:14Zak1392220 by 176
10:27:26Bagder77440 bytes for 16 bit
10:27:46LinusNthat's a serious amount of bytes to transfer
10:27:51preglowit is indeed
10:28:57preglowsame as the ipod photo, then
10:29:47preglowthe nano has actually got quite a nice resolution for its size
10:30:07Bagderdid we ever go with that rect-update improvement patch?
10:30:13Bagderfor h3x0
10:30:41Zak1392linus: have you looked into H300_DrmKey.hex?
10:31:44 Join Tandoc [0] (
10:39:32preglow**** 542 [iir.asm 25]: ERROR −−- Possible invalid white space between operands
10:39:44preglowyou want me to tell you have much i love compilers that require you to use whitespace in a certain way?
10:40:57preglowthe motorola 56k assembler does not allow you to use whitespace after a comma
10:41:32LinusNZak1392: no
10:42:01LinusNBagder: no, not yet
10:51:03Lost-ashpreglow: better still, it's an error for a possibility. what if you really meant that? :)
10:55:56 Join Kohlrabi [0] (
11:04:08markunYesterday I was closing some patches, but of some I'm not sure anymore.
11:04:40markunShould we keep the language patches open after commit so other people can reuse them for new language updates?
11:08:34preglowi have never seen a compiler with us unhelpful error messages as this thing
11:08:35 Nick Lynx_awy is now known as Lynx_ (
11:09:14markunamiconn: I could take a look at what caused the error instead of just removing the code :)
11:10:21 Nick Lost-ash is now known as ashridah (n=ashridah@
11:10:51amiconnlinuxstb: No second copy of the framebuffer needed
11:11:53***Saving seen data "./dancer.seen"
11:13:03markunamiconn: what do you think about closing language update patches?
11:17:32linuxstbamiconn: How do you store the background image then?
11:18:06Jungti1234good morning markun
11:22:09 Quit edx ()
11:29:22markunamiconn: the voice string in a language file can be left empty if it's the same as the new string, right?
11:32:21ReKleSSerm... is there a way to string multiple .bdf files into a single .fnt?
11:32:35ReKleSSI've tried sticking them together in the bdf format, but rockbox doesn't seem to like the resulting .fnt
11:33:53markunReKleSS: would be nice to have a tool to do that.
11:39:04markunamiconn: the problem with the WPS loading is probably that the code can crash when it receives a non UTF-8, which I agree should be fixed.
11:40:14 Join DJDD_ [0] (
11:41:02LinusNReKleSS: it's not as easy as concatenating them
11:41:17LinusNyou need to write a script that changes the character codes
11:41:46Jungti1234How did unicode font that I send become?
11:42:19markungulim or dotum?
11:42:42 Join Hooligan [0] (
11:42:46markunAh, I had it already :)
11:43:00markunDid you try it on your H120?
11:43:24Jungti1234I'm going to try now.
11:43:24 Quit Kohlrabi ("Leaving")
11:43:55markunBecause I think the font is not very suited for a small display, but I didn't test it either.
11:44:23HooliganHas Linus said anything about looking at that H300_Drmkey.hex file, by any chance?
11:44:30Jungti1234Do you want small font?
11:45:05LinusNHooligan: i haven't had time to look at it
11:45:41HooliganOh, alright.
11:48:17 Join Cassandra- [0] (
11:50:59Jungti1234I know the website.
11:52:22markunI like the arial unicode font, but I doubt we can distribute it.
11:55:52 Quit Cassandra (Read error: 110 (Connection timed out))
11:55:53 Nick Cassandra- is now known as Cassandra (
11:56:26amiconnmarkun: No, the voice string can (and should) be left empty if the string is not to be voiced, otherwise it has to be filled in
11:57:30markunamiconn: I was looking at the italian language file in the patch tracker and most of the changes were added voice strings that were the same as the new string..
11:57:39 Join Febs [0] (
11:58:03markunI'll leave those out and then commit it so the patch can be closed.
11:58:20amiconnThey have to stay in if the string is to be voiced
11:58:45amiconnThe voice file creation looks at the id: and voice: line _only_
11:59:55amiconnThis is on purpose; not all strings are voiced, and including a clip that won't be used isn't desirable
12:01:44amiconnOf course, no voice: string should be added in a .lang file for an ID where the voice: string in english.lang is empty
12:02:20amiconnBtw, are the voice: lines also converted to utf-8 now? (guess they are)
12:02:38markunyes, they are
12:02:49markunWill it be a problem for the voice generators?
12:02:59amiconnHmm, then I have to adjust my script
12:03:23amiconnHave to find out how to handle utf-8 in vbscript
12:03:57amiconn(or how to tell SAPI the correct encoding)
12:04:23markunHCl: would you like to take a look at your tagdb generator to have an optional 'codepage' argument and store all the strings in UTF-8?
12:05:44Jungti1234oh.. -_-;
12:06:07Jungti1234Cyberbit is not goot.
12:07:17 Quit Zak1392 ()
12:07:20 Quit Hooligan (Read error: 104 (Connection reset by peer))
12:16:53 Quit linuxstb ("Leaving")
12:19:35 Join edx [0] (
12:22:03 Nick solexx___ is now known as solexx (
12:24:12 Quit aliask ("Chatzilla 0.9.69 [Firefox 1.5/2005111116]")
13:10:13 Join Membrillo [0] (
13:11:43 Quit tvelocity ("Leaving")
13:11:57***Saving seen data "./dancer.seen"
13:13:31Membrilloyo firefly, you here?
13:15:51*Bagder can't leave the neuros google group
13:16:14BagderI've already left it on the web interface
13:16:21Bagderbut I still get emails
13:16:32Bagderand the unsub URL in each mail doesn't work
13:18:12 Join webguest66 [0] (
13:18:47 Join webguest12 [0] (
13:18:47 Quit webguest66 (Client Quit)
13:20:06Jungti1234Is no rockbox plan to change interface?
13:21:21 Join Maxime` [0] (
13:21:41preglowBagder: now there's a strategy we can use as well!
13:22:16Bagderyes, very effective!
13:23:48 Quit webguest12 ("CGI:IRC (Ping timeout)")
13:29:16Jungti1234Isn't plan?
13:29:39BagderJungti1234: the plan is to make rockbox good
13:29:44ashridahJungti1234: no offical plans i know of. if you've got a proposal that will work on all platforms rockbox targets, feel free to write it up.
13:30:32Jungti1234I have H100 and H300 only.
13:30:57Jungti1234I don't know other platform.
13:31:21ashridahthat shouldn't stop you if you've got plans for interface updates, all you really need to know are the keys and LCD dimensions, which are on
13:31:27ashridahthe simulator would cover the rest
13:31:31Bagdermost people only have one or two players, only a rare exception have them all
13:32:09Bagderand we even want to make rockbox work for future players
13:32:13Bagderwhich none of us have
13:32:23markunI have one :)
13:32:40Jungti1234Don't program although I have plan.
13:33:13markunJungti1234: write it down or draw some pictures of it
13:33:31Jungti1234hehe.. ok
13:33:46Bagderyes, the more detailed it is, the better it will serve as base for discussions
13:34:16Jungti1234However, I don't know I can do well. :(
13:39:35 Quit Maxime (Read error: 110 (Connection timed out))
13:41:24 Nick ashridah is now known as Lost-ash (n=ashridah@
13:44:43Jungti1234Does rockbox support color BMP now?
13:45:12Jungti1234Is it developing now?
13:45:24BagderI don't know
13:45:34Bagderfeel free to make it
13:45:40Jungti1234But, does it support sometime?
13:45:50Bagderif someone writes the code for it, sure
13:46:16Jungti1234Don't you do it?
13:46:22BagderI don't, no
13:46:25Jungti1234Is it difficult?
13:46:31Bagdernot very, no
13:48:20Jungti1234If I paint interface, it in rockbox can use?
13:48:39Jungti1234paint -> draw
13:49:40BagderI don't understand the question
13:50:18 Quit Nixsos (Read error: 110 (Connection timed out))
13:52:37Jungti1234Can I use interface that I sketch in rockbox?
13:52:54Jungti1234It sometime.
13:53:08markunJungti1234: you can paint the interface in color
13:54:07markunIf people want to implement your interface design and it requires color BMP they will also add that.
13:54:48Jungti1234Then, can I use not text but icon?
13:54:58Bagderyou can use anything
13:55:04Bagderit is a suggestion, isn't it?
13:55:19Bagderan idea of how it _could_ be
13:55:48Jungti1234I understand.
13:56:05Jungti1234I can use anything...
13:56:26Jungti1234OK. I will try.
14:04:36 Join saa[b_r]ider [0] (n=saab_rid@
14:04:49markunHi saa[b_r]ider :)
14:05:23saa[b_r]iderhello markun :)
14:07:00Membrillosomeone might be interested, i got my iPod video today (no! i didnt buy one!, the insurance company gave me it)... I must say, video wise, it kicks my H3xx in the rectum... audio wise... the other way round. its very very very simple audio wise
14:07:18Membrillono customability
14:07:31Bagderit having good video is hardly a surprise
14:07:33saa[b_r]ideryeah, 30 FPS is much more advanced
14:08:06Membrilloi have my videos working at 704x304. Its even good quality working with TV out
14:08:09Lynx_Membrillo: what video formats does it play?
14:09:08saa[b_r]iderI think it only plays MP4
14:09:10MembrilloLynx_: not many. Just MP4 and some other lame one ive never even heard of. There is a lot of emphasis on buying videos from itunes :P there are a lot of converters (similar to iriverter) out there there
14:09:22Membrilloout there though*
14:09:30Lynx_hmm, ok
14:09:42Membrilloits a bugger getting DVDs on though
14:10:05saa[b_r]iderhave you guys seen the new Creative Zen M (their answer to the Video iPod)
14:10:20Membrilloive heard of it
14:10:29saa[b_r]iderit plays divx, xvid, and many other video formats
14:10:37Membrillosounds cool
14:10:44saa[b_r]iderhas 4 hours video playback
14:10:51Bagdermarkun: does the gigabeat play videos any cool?
14:10:57Membrillothe ipod barely lasts two and a half
14:11:13Membrillotop notch LCD though
14:11:14markunIt doesn't play video! :)
14:11:19saa[b_r]ideronly thing is it's fat (compared to the iPod), and it only comes in 30 GB at the moment
14:11:30saa[b_r]iderno 60 GB
14:11:46markunI think the harware and software designers of the Gigabeat had very different ideas.
14:11:53saa[b_r]iderthe iPod is pretty slim, and has a nice LCD as you said
14:12:15saa[b_r]idermarkun: how easy is it to use the + touch pad?
14:12:20 Join PaulJ [0] (
14:13:08markunsaa[b_r]ider: It's not the nicest thing. Maybe as nice as the joystick of the H120
14:13:22Membrillothe ipod video is pretty damn sexy, although the touchwheel has become harder to use with the new version
14:14:18saa[b_r]idermarkun: yeah, i thought so... for example, when you're at the bottom and decide to go left or right, it get's complicated
14:14:45saa[b_r]idermembrillo: I hear the click wheel is small compared to the overall size
14:15:17saa[b_r]iderI also hear that the quality of the new click wheel doesn't compare well with previous generations
14:15:20Membrillosaa[b_r]ider: very much so. the device has got wider, and the touchwheels radius has got smaller
14:15:30Membrilloagreed again
14:15:47Membrilloits a lot less responsive than older models
14:15:56markunThere are 9 sensors on it. Sometimes when I want to press the center it het 'left' by accident.
14:16:23saa[b_r]idermembrillo: doesn't make sense why the would make the radius small...
14:16:31Bagdermarkun: yes, that sounds like the h1x0 joystick ;-)
14:16:46Membrillosave materials. Apple are stingy bastards
14:16:51Bagderthe iaudio x5 joystick is nice though
14:17:06Membrillomy 4G screen spontaneously shattered and they wouldnt replace it
14:17:17saa[b_r]idermarkun: on my H300, clicking up sometimes clicks the navi button (or vice versa)
14:18:02saa[b_r]idermembrillo: that's nasty! well I'm sorry to tell you this, but the video iPod's screen looks much more vunrable :D
14:19:23Membrillowell it ended out ok
14:19:38saa[b_r]iderbtw, black or white?
14:19:47Membrilloi gave it to a friend to use (it worked you just couldnt see anything). His house got broken into and I got a video on insurance
14:19:52Membrillo(video ipod came today)
14:21:15Membrillostupid thiefs stole a broken ipod
14:22:19Lynx_my archos was stolen last week. now i don't know what to buy as a replacement
14:23:14 Quit Membrillo ()
14:24:12saa[b_r]iderLynx_ what are you looking for in your next player?
14:24:45Lynx_saa[b_r]ider: maybe and h340
14:25:30saa[b_r]iderI wish iRiver release a new DAP, something to compete with the Video iPod
14:26:06Lynx_I don't know if i actually want a hd based player again.
14:26:06Bagderwhy does it have to be iriver?
14:26:22*saa[b_r]ider owns an H340, and loves it even more now that Rockbox is installed on it :)
14:27:32XavierGrBagder: It has to be a Rockbox compatible one. And not many are available to buy...
14:27:39XavierGrIs that right Lynx_?
14:27:45saa[b_r]iderBadger: if you're asking me, I like the UMS, drag-and-drop functionality of iRiver players...
14:27:48Bagderyou wimps, you can port it to a NEW platform! ;-)
14:28:00Bagdersaa[b_r]ider: there are many such players
14:28:13Lynx_XavierGr: is what right?
14:28:26markunThe archos gmini of my friend didn't look too bad either.
14:28:42saa[b_r]iderwell the X5 is nice and all, but some of the specs are short of the already 1 year old H340
14:28:47XavierGrthat you want a new DAP that is rockbox compatible and you can find easily to buy it.
14:29:06Bagdersaa[b_r]ider: actually I think the x5 is nicer than h3x0
14:29:13Bagderapart from the rockbox of h3x0
14:29:20Lynx_hmm, the h340 can play videos?
14:29:29LinusNa lot nicer form factor
14:30:03saa[b_r]iderBadger: the X5 looks nice, I agree very much... but the color LCD is inferior, and what's with the audio port being on the side?
14:30:16Bagdersaa[b_r]ider: inferior for what?
14:30:19XavierGrdoes X5 has a 40GB model?
14:30:24Bagderand the sound seems fine enough
14:30:51Bagdersure there are some things that aren't perfect, but what player is?
14:30:56saa[b_r]iderLynx_: yes, the h340 plays videos, but @ 10fps only, and a very specific format
14:31:21Lynx_saa[b_r]ider: ok
14:31:23saa[b_r]iderBadger: hang on, I'll check the X5's specs
14:31:32markunThe X5 only has 16MB memory, right?
14:31:47BagderI'm not sure of all the specs
14:32:15saa[b_r]idervery handy :)
14:32:22saa[b_r]iderthanks markun!
14:33:20markunThe USB ATA bridges can only be used for UMS right?
14:33:52Bagderyes, unless there are more fancy ones
14:34:23Bagderbut both h3x0 and x5 have separate usb host ports
14:34:24markunBecause I told the people in the Gigabeat forum to stop dreaming about USB 2.0 in the Windows Media Player mode.
14:34:37Bagdercoffee time
14:35:24XavierGrBagder: Again?
14:35:42XavierGrYou drink too much coffee!
14:35:50saa[b_r]idera man needs his fix :D
14:36:01markunThe Gigabeat cradle has a 4 port USB 2.0 hub, so I could connect a keyboard and mouse to it maybe :)
14:36:22saa[b_r]iderI never use my H340's cradle
14:37:21markunsome things on the gigabeat player are not possible without the cradle.
14:37:36Lynx_so the x5 could also be attached to a card reader?
14:37:40markunThe USB host connector and the line out are not on the unit.
14:38:00saa[b_r]iderI hate it when they do that... but it helps keep the device more compact
14:38:40markunI made a backup copy of my photos with a card reader connected to the Gigabeat.. but the software is so nice that you cannot view the pictures after transfering them!
14:39:04markunYou have to connect it to a PC and transfer the pictures to a special folder with Toshiba's special software..
14:39:24saa[b_r]iderreminds me of Sony and sonicstage
14:39:25Lynx_markun: well, that's not great but better than nothing
14:39:43 Quit Tandoc ()
14:39:46Lynx_i wonder why so many cameras can't act as usb mass storage
14:40:11saa[b_r]idermarkun: it doesn't support filetree browsing does it?
14:40:14markunI have a small USB card reader in my camera bag.
14:40:19markunsaa[b_r]ider: no
14:40:21amiconnLinusN: Lately I had a look at the X5, even held it in my hands
14:40:45Lynx_markun: i don't like to swap compact flash cards in and out of the camera all the time...
14:40:57saa[b_r]ideryeah, card readers are the best way for transfering files off memory cards
14:40:57amiconnIt's a matter of personal preference, I like the H300 form factor way more than the X5
14:41:50saa[b_r]iderthe X5's LCD is just 160X128, that's what's bad about it... but badger is having coffee :)
14:42:09SlasheriLynx_: Hmm, in fact professional photographers usually have a bunch of CF cards which they use on different situations :)
14:42:22saa[b_r]iderLynx: what camera do you have? Canon?
14:42:42Lynx_saa[b_r]ider: yes, recently bought a 350d
14:43:00Lynx_Slasheri: i always think one of the little pins will break ;)
14:43:08*saa[b_r]ider checking out the 350d
14:43:21amiconnWhat I dislike about the X5 compared to the H300: (1) Side buttons. (2) Joystick (3) Bulkier at the top than at the bottom. The lower display resolution isn't relevant for me
14:43:47SlasheriLynx_: ah, i think that is very unlikely because the card is aligned very tight before it reaches the pins
14:44:23Lynx_Slasheri: i guess you're right. but aren't the cards specified for only so many insert and retrieven events?
14:44:55saa[b_r]iderI never heard that before
14:46:03SlasheriLynx_: maybe, i am not sure about that.. anyway, the connectors are gold plated so they should be able to handle many insertions before breaking. I just know that many people swap the cards regurlarly and i haven't heard yet about that kind of problems :)
14:46:25preglowme neither
14:46:29preglowthey can take quite a beating
14:46:44*amiconn wants a H3x0 now. The other options are all inferior
14:46:52Lynx_ok, so i will buy a new card reader tonight :)
14:47:04Slasherihehe :D
14:47:17Lynx_my old (aldi) one killed 2 cards, for some reason.
14:47:23saa[b_r]iderCF cards are one of the most popular memory cards, if they were vunrable, people would of reported it
14:47:25 Join PaulJ_ [0] (
14:47:33Slasheriamiconn: Hmm, i would like X5 more
14:47:34preglowcan't see a reason for buying a h3x0 when i've got my trusly h120
14:48:07Lynx_amiconn: amazon germany still has 340 for 349,- euro
14:48:10*saa[b_r]ider envies H100 for optical in/out!
14:48:15amiconnSlasheri: Really? One of the two things that annoy me about the H1x0 are also present on the X5, but not on the H3x0
14:48:45Slasheriand buying a h300 for development isn't so interesting at this point because h140 is better music player for me and rockbox already works on h300 and it's discontinued also..
14:48:45amiconnThese things are one reason why I still don't use the H1x0 too often
14:49:14Slasheriamiconn: Hmm, what are those things? The joystick?
14:49:27amiconnYes, the joystick and the buttons on the sides
14:49:35Slasheriah :)
14:50:24amiconnThe H3x0 just has front buttons, as it should be :)
14:50:39Slasherihehe, then it could be a good choise, maybe :)
14:50:40preglowi don't have a preference right there
14:50:42preglowi'd want both, heh
14:50:50preglowdepends on how you're carrying it
14:52:33Lynx_does the h340 come with a remote?
14:52:52Slasheriyes, lcd-less remote i think :)
14:53:01saa[b_r]idercomes with an LCDless remote, yes
14:53:08saa[b_r]iderthe LCD remote is optional
14:56:43*saa[b_r]ider just saw the Canon 350D! very professional!
14:57:13Lynx_saa[b_r]ider: well, it's the cheapest digital slr from canon
14:57:39 Quit _FireFly_ ("Copywight 2004 Elmer Fudd. All wights wesewved.")
14:58:12Jungti1234bye all
14:58:14Jungti1234good night
14:58:20 Quit Jungti1234 ("bye")
14:59:10LinusNi don't like that the x5 doesn't have a bdm port
15:00:08XavierGrand how are you supposed to work on it?
15:00:37LinusNwell, i'll just have to make sure i don't brick it :-)
15:01:28XavierGris that possible during development of the bootloader?
15:01:39XavierGrI mean, how many times did you brick your H100 during development?
15:01:58 Quit PaulJ (Read error: 113 (No route to host))
15:02:26LinusNtwice i think
15:02:55XavierGrI am amazed!
15:03:17XavierGronly twice? excelent
15:03:53XavierGrand lets say that you do something wrong and X5 is bricked, any alternatives?
15:04:07LinusNbuy a new one
15:04:28 Join eli_sherer [0] (
15:04:34XavierGrargh! That's tough...
15:04:41markunSend it back to iAudio
15:06:06LinusNi like the fact that the x5 firmware image isn't encrypted
15:06:24Bagderthat's neat indeed
15:06:40Bagderand a very simple checksum, if the wiki is correct
15:07:03XavierGrso do you think that you can make it? It seems impossible to me to write code with a gun pointing at me that will shoot on a compiler or execution bug...!
15:07:43Bagderno risk, no fun ;-)
15:08:02XavierGrAlways the first time I compile I get like what? 40-50 errors maybe :D
15:08:29XavierGrBagder I don't know about you but I just hate taking risks...
15:08:58Bagderwell, the risk involved here is pretty mild compare to other risks really
15:08:59LinusNBagder: where in the wiki?
15:09:14Bagderlemme check...
15:09:17Bagderah, no
15:09:26Bagderits was the iaudio.c tool
15:09:31Bagderin tools/
15:09:37LinusNoh :-)
15:10:04preglowooh, finally, no longer a newbie on the forums
15:10:05Bagderand I dissassembler an x5 firmware image and ran in hexl-mode
15:10:27LinusNi'm doing that now
15:10:45preglowXavierGr: my status is upgraded to junior member!
15:10:45XavierGrah you mean your title?
15:11:16LinusNBagder: but i can't figure out what the first 4K in the image are for
15:11:36Bagderno, they seem to be data that is tricky to guess
15:12:01***Saving seen data "./dancer.seen"
15:13:41XavierGrpreglow I need 3 posts to get you.
15:14:11*preglow kicks winapi, then strangles it
15:14:17Slasherione way would be to read the flash contents directly from the unit (via bdm on desoldering the chip)
15:14:21 Join Jungti1234 [0] (n=jungti12@
15:15:53LinusNwell, there's no bdm port...
15:16:09LinusNthe flash is bga too...
15:16:29lostlogicthat's how they made it so damn thin?
15:16:53amiconnThere's a H340 with LCD remote on ebay germany...
15:16:58Slasherioh, that's nice
15:17:10LinusNthe x5 pcb is incredibly small
15:17:26SlasheriLinusN: hmm, do you know if there is any traces on the pcb that connect to the bdm port pins?
15:17:32preglowbga = socket for enthusiasts
15:17:54LinusNSlasheri: doesn't look like it
15:17:59Slasheriah :/
15:18:59Slasherione way could be still to replace the bga flash chip with some sort of adapter, so you could easily program the flash by computer and then try code on the device
15:19:07XavierGrhow did they do the development themselves then?
15:19:11Slasheribut that soldering would be hard
15:19:35SlasheriXavierGr: probably they programmed the flash before soldering it on the unit
15:19:43Slasheriand of course they had a modified development device..
15:19:57XavierGrthat's more like it.
15:20:36XavierGrbut then fixing a bricked player on warranty will be impossible for them, right?
15:21:19Slasherimight be..
15:22:52 Quit Jungti1234 ("bye")
15:24:32lostlogicpreglow: after some firstround optimizations that gave _some_ improvement, I'm trying profiling Tremor w/ all the inlines back as inlines and not profiled
15:24:54lostlogicgives both faster running during profiling, and hopefully a different look at how the code runs.
15:25:42preglowlostlogic: how do you do that?
15:27:06LinusNholy fuck, the cpu wasn't bga after all!!!!
15:27:19Slasheriuuh :)
15:27:32SlasheriLinusN: do you have the real unit or just scans of the pcb?
15:27:53LinusNi had merely been looking for test pads
15:28:30LinusN(couldn't figure out how to remove the lcd, but now i have)
15:29:02LinusNthe cpu is under the lcd panel
15:29:26LinusNso i will be able to connect a bdm after all
15:30:59LinusNi am so impressed by how small the damn thing is
15:31:06lostlogicthere's a __attribute__ ((no_instrument_function)) so I stuck it on all of the inlines I wanted to not profile.
15:31:26lostlogicI've also made a NO_PROF_ATTR define for convenience locally, but it's not in the patch I submitted yesterday :(
15:34:21LinusNthe best part is that since it is a 52xx platform, I don't have to investigate the ATA interface
15:34:53lostlogicmmm, bootloader by guesswork! :-P
15:34:55LinusNand probably not the SDRAM or flash interface either
15:35:21LinusNwe need to find out how to program the DA202 FM tuner
15:35:34amiconnLinusN: That reminds me - is the H1x0 rockbox ATA timing optimised already?
15:35:47preglowthen why did you it was bga??
15:35:53preglowinsert 'say'
15:36:00LinusNi just assumed
15:36:14preglowisn't it easy enough to see? :-)
15:36:21preglowbut no
15:36:22LinusNnot underneath the lcd
15:36:23preglowyou're alright, then
15:38:03 Quit gromit`` (Read error: 110 (Connection timed out))
15:38:47lostlogicanyone really good at m68k assembly and want to optimize a couple of functions for me?
15:39:29lostlogicthey get about 10-15 million calls per song, compared to the next most at 1.2 million, so I think they are excellent candidates.
15:40:10 Join _FireFly_ [0] (
15:43:38perplexityWanna post your profile results somewhere lostlogic so we can see what needs to be optimised ?
15:44:27perplexityI can't promise anything as I'm so short on time at the moment I can hardly stop to eat, but I'd love to take a look and see if we can't slim it down a little">
15:47:31lostlogiceach has somewhat of a description at the top, forgot to mention that optimized.profile was also at 90mhz, only no_inlines was done at 124mhz
15:47:47 Quit Febs (Read error: 104 (Connection reset by peer))
15:47:50lostlogicalso, they are all the same song, just different durations, starting from the beginning.
15:48:23perplexityCool.. I'm out for dinner now, but I'll probably have some time a little later tonight to have a squiz.. great work on the profiler btw
15:48:42lostlogiccool and thanks.
15:48:58lostlogicI'm off to work now myself.
15:50:40 Join Rob2222 [0] (
15:54:02 Join Kohlrabi [0] (
15:54:47 Join frederic_ [0] (
15:56:33frederic_Hi, is there a one-step solution to encode an image + cue to mp3s using the −−nogap option? I read about the cuesheet input patch in the wiki but it's outdated and I'm trying to do the same the patch did with "acdir" - it's "successor"
15:56:47frederic_I described in more detail here
15:59:25XavierGramiconn: do you remember a link for a nice pdf document that was a guide to CVS? I remember you gave the link a few days back.
15:59:48 Quit Rob2222 ()
16:01:55crwlfrederic_, use ogg vorbis and forget those odd image + cuefile hacks ;)
16:02:14 Quit saa[b_r]ider ()
16:02:36XavierGramiconn: never mind I think I found it.
16:02:52 Join taladan [0] (
16:03:10frederic_^_^ yeah, lame hacks :P but still: I don't want to encode for a rockbox enabled device :>
16:03:47 Join Rob2222 [0] (
16:04:19*taladan has an Archos FM Jukebox & is thinking about getting the rockbox software for it (just found out about it today), wondering how it handles playlist creation and such...stability issues...stuff like that?
16:04:41Bagdera fresh archos user, that is rare these days ;-)
16:04:54Bagdertaladan: we have lots of docs online
16:05:08Bagderrockbox has worked on Archos for many years and I'd say it is very solid
16:05:33LinusNexcept for right now, that is in the unicode turmoil
16:06:07LinusNi'd recommend installing the 2.5 release
16:07:18taladanBadger - going over the documentation right now. Distressing lack of info on how to flash the Archos (New Paragraph text goes here is all it says) but that's okay.
16:08:26*taladan uses linux, so cherishes the idea of open source firmware driving his mp3 player ;)
16:08:52taladanMy biggest question is:
16:09:03XavierGrtaladan: did you get your archos now, or it has been for a long time.
16:09:27taladanIs there a way to make a playlist up while the unit is plugged into the box via USB instead of having to use the nav-keys?
16:10:04Bagderwhen plugged in to USB, rockbox can't do much
16:10:05taladanXavierGr: I've had the Archos for....4 years? Something like that. Right after Radio Shack started carrying it for the first time (Used ot work for RS)
16:10:27XavierGrpitty that you have been using default fw for so long...
16:10:45Bagderyou're up for an adventure! ;-)
16:10:49taladanBagder: So is there any way to manipulate a playlist (creation, editing and such) in linux that Rockbox will recognize & play?
16:10:57Bagdertaladan: certainly
16:11:07Bagdera playlist is just a plain text file with file names
16:11:50Bagderfind . -type f -printf "/%p\n" > playlist.m3u
16:12:20taladanBagder: I can VI a playlist...good. What format do the names need to be in? And I'm assuming a full pathname is required with the Archos drive main folder being the root of the path
16:12:25XavierGrdoes anyone knows a good simple free hosting site?
16:12:31Bagdertaladan: correct
16:12:40XavierGrfile host I mean.
16:13:10Bagdertaladan: but you can also make them using Rockbox, in case you'd like to
16:13:26Bagderand in comparison to the original...
16:13:27taladanwhat's the /%p\n in that command?
16:13:32Bagderrockbox _flies_
16:13:36Bagderwhen reading playlists
16:14:07_FireFly_taladan: man find ;)
16:14:09Bagdertaladan: it makes an absolute path from your local one, if you for example have your songs in /home/user/mp3 and you run the command in there
16:14:32taladanMakes sense
16:15:04taladan_FireFly_: Yeah, yeah, I know...but sometimes it's just quicker for me to ask the person than dig through the man pages...esp when I've got class in a bit
16:15:30_FireFly_taladan: you can search within the manpages
16:15:41taladanShhhh, don't remind me :p
16:15:49_FireFly_with shift+7 and then search-text
16:16:02_FireFly_when man-page is shown
16:16:08taladanCall it a lack of good human interaction within linux users, hence why I'm creating a LUG
16:16:24taladanThe / lets you search ;)
16:16:30*taladan loves /
16:16:57taladanI didn't realize that firefox uses / as find until yesterday either.
16:17:30_FireFly_taladan: or strg+f ;)
16:17:57*taladan chuckles
16:18:06*taladan goes back to reading documentation.
16:19:02taladanOne question: Putting the firmware on there while I've got music on the jukebox, will that mess it up in any way, or do I need to format it out with ext2/3 and then put the firmware on it first and then re-load my music up to it?
16:19:18*eli_sherer thinks it's going to be along time before a working version of his AVI player will be ready
16:19:25Bagdertaladan: it won't mess up anything
16:19:39_FireFly_taladan: rb support only fat32 as fs
16:20:23_FireFly_why support other fs when all original fws of the supported players uses also fat32
16:20:24taladanThat's good to know then. Wonder why it doesn't support ext2/3? No journalling! *weeps* ;)
16:20:37_FireFly_ext2 has no journalling ;)
16:20:44taladanyes, but ext3 does ;)
16:20:55Bagdertaladan: well, for starters you must have your player using fat or you can'r un rockbox as it is read from disk
16:20:56taladanAnd with tune2fs you can turn journalling on in ext2 ;)
16:21:00Bagder(unless you flash)
16:21:13_FireFly_taladan: then it is ext3 ;)
16:21:27 Join edx__ [0] (
16:21:27 Quit edx (Read error: 104 (Connection reset by peer))
16:21:44taladanyeah, I'll just use it as vfat, that's not a problem...just have to remember that when I take it to my LUG meeting and show it to the folken there
16:22:00 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
16:22:58 Nick edx__ is now known as edx (
16:24:35taladanquestion...what's the 'voice files' all about?
16:24:47Bagderspeaking menus
16:24:55Bagderand more
16:24:57taladanoooooh...for the sight impaired!
16:25:00taladanNeat as shit!
16:25:05Bagderor for using in the car etc
16:25:20*taladan has a blind friend in Atlanta that might be interested in that....
16:25:47Bagderwe have a huge blind user base
16:26:38taladanI can imagine
16:27:38taladanI don't know who wrote this software, but it does my heart good to see someone thinking of the blind like that. If you know who the maintainers are, please pass along my personal thanks to them....Man, that's awesome!
16:27:52Bagderwe are the maintainers
16:28:14taladanheh, good enough then! Thank you guys
16:28:24 Join mdeboer [0] (
16:28:57taladanone note - the link to the voice files via twiki in the PDF manual seems not to exist
16:29:46preglowwho made the voice ui feature again?
16:30:10mdeboeri just compiled rockbox from cvs, and when running in the uisimulator (on linux/x11), when i try to run a plugin (tried several), it freezes saying [Loading]
16:30:18mdeboerany idea?
16:30:30Bagdermdeboer: you did make install, right?
16:30:31preglowi think the software codec voice ui approach needs polishing
16:30:36preglowcurrently, it's very laggy
16:30:53mdeboerah, i see some work. logo for example. but calculator doesn't
16:31:07BagderI don't know why that would happen
16:31:33mdeboeri build for H3xx
16:32:21Bagderthen it might be a color lcd problem
16:32:29Bagderthere's no color lcd support in the x11 sim
16:32:35mdeboeri see
16:32:50mdeboerbut rockbox itself does do color lcd on the h3xx?
16:32:59Bagderits a simulator flaw
16:33:13Bagderand only in the x11 version too
16:33:18mdeboeri see...
16:33:27mdeboeris anyone working on this problem?
16:33:34Bagdernot that I know
16:34:04mdeboerso the windows version does do color?
16:34:21mdeboera port to a crossplatform library would be nice...
16:34:32Bagderyes indeed
16:34:34mdeboerinstead of having to seperate versions.. SDL comes to mind...
16:34:35taladanhmmmm....this isn't good.
16:34:43taladanI just got a bunch of scsi1 (0:0): rejecting I/O to dead device
16:34:43taladanFAT: bread failed in fat_clusters_flush
16:34:47_FireFly_maybe Cassandra is working on somehting
16:34:53Bagdershe is?
16:34:54taladanin dmesg and it dropped my mount....
16:34:57Bagderit would be great
16:35:01mdeboeror i might run the uisimilator under wine ;-)
16:35:17Bagdermdeboer: it is said to work pretty well
16:35:35_FireFly_Cassandra: i don't know it correctly but she is trying code somthing with wxwindows
16:35:43mdeboerbut i suppose an sdl port would be feasible. i'll have a look at the code.
16:35:47mdeboerwxwindows... hmm.
16:36:02LinusNsdl sounds like a good solution
16:36:09preglowyeah's switched from sda1 to sdb1...
16:36:14taladanon its own
16:36:17preglowi've used it a quite a bit, and it works fine
16:36:19Slasherisdl should be quite universal
16:36:19taladananyone know why it would do that?
16:36:20mdeboeror maybe fltk
16:36:24preglowand is agreeable to work with
16:36:37Slasheritaladan: is battery fully charged?
16:36:37preglowwe don't need any widgets
16:36:43preglowso no need to use wxwindows or fltk, really
16:36:47mdeboerok. than sdl would be good indeed.
16:36:49taladanshould be, I left it to charge overnight
16:36:58taladanvia USB
16:37:04Bagderit would be really neat to go with a portable lib to get rid of all the win vs x11 related diffs
16:37:10mdeboerso the windows code base would be the most up to date to create an sdl port off?
16:37:14LinusNi'd love it
16:37:17taladanmy charge port is scragged on it for some reason, so it only charges via the USB now
16:37:20amiconnmdeboer: You can run the win32 simulator under wine
16:37:39Bagdermdeboer: they're pretty similar except for the color support, afaik
16:37:46Bagderuse the one most similar to the target lib api
16:37:54Bagderor both
16:37:56Bagderor none
16:38:29mdeboeri'll have a look at the code, and see if i can find the time.
16:38:54mdeboeri have no SDL experience, but it would be a nice project to get some
16:40:04preglowsdl is nice
16:40:13preglowpretty fast to get into it
16:40:26mdeboeryes, i've been looking at documentation and sample code
16:41:01preglowdoes anyone know how the hell one uses scrollbars with winapi? :/
16:41:18preglowhow i hate winapi
16:41:28preglowa great big steaming turd of an api
16:42:04mdeboeryeah, i'm now looking at the uisimilator win32 code, and my stomach is revolting already ;-)
16:42:20preglowoh, i vomit regularily
16:42:31preglowit's a cleansing ritual i have when working with winapi
16:42:37mdeboernot that low level X11 is such a pleasure to look at...
16:42:42preglowno, no it's not
16:42:47preglowbut i can ignore that, since i don't need to use it
16:43:17Bagder"And remember: X Windows is to graphics hacking as roman numerals are to the square root of pi."
16:43:37Bagderxscreensaver quote
16:44:16LinusNhmmm, i think i have found how to use dma to write to the h300 lcd
16:45:02LinusNand also for the ATA
16:45:15Bagderthis can get fun
16:45:42dwihnopreglow: what winapi stuff are you doing?
16:46:58 Quit frederic_ ("Chatzilla [Firefox 1.5/undefined]")
16:47:25 Join Cassandra- [0] (
16:48:05preglowLinusN: ooo, will the ata dma work for h1x0 as well?
16:48:19preglowdwihno: control app for a dsp platform thing
16:48:25preglowLinusN: that'd be splendid!
16:48:26LinusNbut it won't buy us that much
16:48:39preglowLinusN: well, at least we wont have to boost the cpu when reading anymore?
16:48:52mdeboeram i seeing correctly that at least _some_ color support intent has been mmade in lcd_x11.c?
16:48:59LinusNpreglow: do we?
16:49:11Bagdermdeboer: it supports greyscale, for the h1x0 simulation
16:49:30taladanquestion...Can Rockbox be made to play files in subdirectories of the current directory?
16:49:31Slashericurrently we boost the cpu to get much higher read performance
16:49:35LinusNboosting the cpu increases the memory bandwidth, so it is still a good idea to boost
16:49:54Bagdertaladan: yes
16:49:55mdeboerBagder: any idea what the problem is with full color support?
16:50:04Bagdermdeboer: it just isn't done
16:50:19taladanBagder: Does it automatically, or ... ?
16:50:28dwihnopreglow: win32api of xwin?
16:50:37Bagdertaladan: no, you need to enable it with an option and then it'll traverse down the dir hierarchy
16:50:56taladanI never could figure out how to get the old os to do that
16:51:47preglowdwihno: winapi
16:51:55preglowLinusN: well, you tell me :)
16:52:06 Join Moos [0] (
16:52:10preglowLinusN: we do currently, yes, if you were asking
16:52:21 Quit Cassandra (Read error: 113 (No route to host))
16:52:22 Nick Cassandra- is now known as Cassandra (
16:52:23dwihnopreglow: Erhm. winapi? :)
16:52:31preglowdwihno: win32api
16:53:06dwihnopreglow: aaah, okay. hardcore win32api or some MFC/STL/WTL stuff?C
16:53:59preglowhardcore win32api, i wouldn't call mfc win32api, heh
16:54:43dwihnoSome people have different opinions :)
16:54:55dwihnoTo be honest, I never learned any wrapper stuff
16:55:24 Quit Zagor ("Client exiting")
16:55:37preglowit's torture
16:55:43preglowslow-winded torture
16:55:48dwihnopleasure is just a higher form of pain.
16:56:15dwihnoor was it the other way around ;)
16:56:46dwihnoAll those messaging queues and stuff really drives you mad sometimes.
16:59:51 Join NicoFR [0] (
17:02:24 Quit mdeboer ("Leaving")
17:07:10dwihnoIs a 25% gain reasonable from a 25% mAh increase (battery-wise)?
17:11:31*eli_sherer updated the site with new versions of his projects...enjoy! (eli_sherer/">
17:12:03***Saving seen data "./dancer.seen"
17:12:55LinusNeli_sherer: you'll write a new clock plugin?
17:13:13LinusNand a new calendar too?
17:13:16eli_shereryup...if an RTC on the H300 will be found...
17:13:35LinusNi take it you don't like the existing clock then?
17:13:47eli_shererall colorful with bitmaps and analogic clock...the source is already written
17:14:13eli_shereri havn't seen the exisiting clock cause i have an H300
17:14:49LinusNyou can try it in the simulator
17:15:37eli_sherersomehow my sim doesn't run plugins...:(
17:16:16 Join saa[b_r]ider [0] (n=saab_rid@
17:16:17LinusNwhy is the source zipped?
17:17:04eli_shererit's not an organized website it's just made with frontpage so it's for my organization of things
17:17:27LinusNwhere is plugbmp.c?
17:18:17eli_shereri forgot about the bitmaps as
17:18:18t0masno RTC chip found on the H300?
17:18:33LinusNt0mas: of course it is found
17:19:11t0masjust no driver written?
17:19:15t0masor no datasheet found?
17:19:51LinusNeverything is there, we just have to adapt the rockbox time functions
17:21:00LinusNeli_sherer: i must say that your plugins look good
17:21:08XavierGrOh Hi t0mas! (my hope in commiting the battery_bench plugin) any news in it? :D
17:21:28eli_shererLinusN: havn't you seen them at action yet?
17:22:01LinusNno time
17:22:31eli_shererso wait a few seconds for me to upload the correct zip files
17:22:45eli_shererthat's's up
17:22:53 Quit saa[b_r]ider ()
17:23:03 Join saa[b_r]ider [0] (n=saab_rid@
17:23:07 Join Midgey34 [0] (
17:23:25 Join Mongey| [0] (
17:23:45 Join webguest21 [0] (
17:23:56NicoFRhi eli
17:24:12NicoFRI get a sitorted picture with some BMPs
17:24:15Mongey|nice plugins eli_sherer
17:24:15 Quit webguest21 (Client Quit)
17:24:22eli_shererwhich ones?
17:24:44NicoFRoriginal file is :
17:25:32NicoFRit's a 24 bit/pixel BMP
17:25:39eli_shererinteresting i'll look into that...and when you open other bitmaps it's good?
17:25:51NicoFRI haven't tryed much
17:26:02NicoFRbut yes
17:26:04eli_shererdid you try it with my new version...
17:26:08Mongey|NicoFR; nice pics where did you get it
17:26:14NicoFRthat was 0.45
17:26:51NicoFRMongey|: A friend sent me the picture as an icon for the player... the color was wrong on the buttons so I changed it
17:27:32NicoFRI have a PNG version if you want...
17:27:47Mongey|yes please
17:29:26Mongey|how did he make it
17:30:06NicoFRHe didn't... he found it somewhere, but he didn't tell me where
17:32:43eli_shererthat's really weird...suggestion any onw??
17:34:26NicoFRI suupose it has nothing to do with the aspect ratio not being the same as the screen...
17:36:41Mongey|hmm will rockboy ever reach full speed
17:37:51LinusNMongey|: maybe, if you optimize it enough
17:39:06*Mongey| starts learning how to code
17:40:27lostlogicxshould definitely get it running in profiler :-D :-P
17:40:36 Quit t0mas (Read error: 104 (Connection reset by peer))
17:42:55*eli_sherer goes over the rockboy code to find how to show color on the h300...
17:45:26XavierGrLinus: Is there any rule on what happens if I get a stov error?
17:45:46LinusNthen you have overflowed the stack
17:45:51XavierGrI mean what' triggers that error , what makes the stack overflow?
17:46:08LinusNeither by allocating a huge array on the stack, or by overwriting a global array
17:46:48LinusNeli_sherer: could you have a look at the rockbox bmp loader and adapt it to load color bmp.s?
17:47:00XavierGryes I have a huge array, if you remember it is on the jpeg filescroller.
17:47:55XavierGrit will show an stack overflow once I reach the 103rd file when scrolling.
17:49:32eli_shererLinusN: sure... but your bitmaps are stored in arrays and color bitmaps will need more buffer size which will slow thing down
17:50:44 Join steveb [0] (n=dave@ircguide/director/about/fish/about/uk/steveb)
17:53:34XavierGrHa I just had a brilliant idea on how to make the jpeg scroller independent from large arrays and buffers! (I will tell you just in case I am wrong)
17:55:14XavierGrWhat if... instead of storing a LARGE buffer with all jpeg filenames on memory, we will just make a temporary file (in current folder) with all filenames in there. (and a pointer to where we are in that list.)
17:55:46XavierGrThen every time the user changes file we just change the pointer and no need for big arrays. What do you think?
17:56:11XavierGrAnd we don't have to worry about performance either, on way or the other the disk will have to spin to load the picture.
17:57:43 Quit Strath (Read error: 104 (Connection reset by peer))
17:58:02 Quit Kohlrabi ("Leaving")
18:03:20 Join Strath [0] (
18:03:28taladanas in /the/ Linus?
18:04:33 Join dpassen1 [0] (
18:04:49XavierGrwhat you mean taladan?
18:04:51taladanah, okay...different linus
18:04:57amiconneli_sherer: My idea concerning the bmp loader is that it should support all important bit depths (1, 4, 8, 16, 24) for input regardless of output format, and output the following formats, slectable by a parameter
18:05:10*taladan thought that was Linus Thorvalds for a second there. *laughs*
18:05:18XavierGrYou thought Linus Torvalds?
18:05:29XavierGrmany have thougth that.
18:05:42taladanwas so happy I almost peed m'self ;)
18:05:43XavierGrI think I was one of them... ;p
18:05:56*taladan chuckles
18:06:09amiconn(1) monochrome. (2) native main lcd format if different from 1. (3) native remote lcd format if different from 1 and 2
18:06:21taladanWell, now that's settled, i have to head in to class. Thanks for the help guys, it's working beautifully!
18:06:46 Part taladan ("Leaving")
18:07:06amiconnIt could work very similar to bmp2rb, but without the huge intermediate buffer
18:07:18 Quit Mongey| (Read error: 104 (Connection reset by peer))
18:11:52LinusNgotta go, cu
18:11:55 Part LinusN
18:13:24eli_shereramiconn: my point exactly...
18:13:48eli_sherergot some color on the gameboy but it's not the right
18:15:31XavierGrwell someone that knows about optimizing and feels like it should run rockboy with lostlogics profiling system!
18:15:56XavierGrAfter a good optimization colour can be added.
18:16:17amiconneli_sherer: Afaik, almost all gnuboy palette code is disabled in rockboy for speed
18:19:22 Part Sando
18:20:11eli_shereri just tried it ...and it's not disabled..
18:22:05amiconnMkay, then not everything is disabled, but some parts are
18:23:15 Join DangerousDan [0] (
18:24:29eli_shereri managed to get rockboy support 16 colors...about the other 240 colors you need to guess the palette as it mentioned by the README cause no palette load process been made by the program...
18:24:56eli_shereri played with the 16 colors until it maked sense...
18:25:00XavierGrha southpark! it's been some time since I last heard the "mkay" phrase. hehe
18:25:38XavierGrdoes game boy colour has 256 colours? I thought it was only 16.
18:26:56XavierGrnah now that I think of it 16 colours are too low.
18:27:18eli_shererwell it's has a changing palette so i guess so...
18:27:49amiconnXavierGr: ?
18:28:28 Join Febs [0] (
18:29:51 Nick Lynx_ is now known as Lynx_awy (
18:30:58XavierGramiconn: you said "mkay". "Mkay is a popular phrase from "Southpark" series.
18:31:08amiconnAh ok
18:31:20*amiconn never watched Southpark
18:34:11eli_shererXaviarGr: just that out and your right...16 colors it i did it woohoo color game boy!!!!
18:34:59XavierGreli_sherer: Now do us a favor and optimize it a little! :D
18:35:59NicoFReli_sherer: do you think you could make it fullscreen on the H300 ?
18:36:00 Quit DreamTactix291 (Read error: 104 (Connection reset by peer))
18:36:32 Join DreamTactix291 [0] (
18:38:21 Join TCK [0] (
18:38:24eli_shererXaviarGr: the source is so messed up...i dont know where to begin...a
18:38:54eli_shererNicoFr: about the full screen it's an optimization issue...the lcd updates only the small screen
18:39:04eli_shererso a small screen saves as speed...
18:39:32eli_shererbye everybody i promise to work on it tomorrow...gotta give my gf some
18:39:43XavierGrI know, optimization is a very hard (and serious) aspect of programming, I hope I could help you...
18:39:54XavierGrlucky you!
18:40:31 Quit eli_sherer ()
18:42:07 Join Acksaw [0] (
18:42:32AcksawI got Z-Lined on :@
18:43:42Acksawanyway to get by this? I really wanna go onto the MR channel
18:45:51 Join bagawk [0] (n=lee@unaffiliated/bagawk)
18:55:27XavierGrwhat's that?
18:56:00 Quit NicoFR ()
18:56:22Acksawno idea
18:57:48 Join Amar [0] (
19:01:02 Join edx__ [0] (
19:05:06XavierGrare there any functions to search into a file or I must load a buffer first?
19:05:15 Join markun_ [0] (
19:05:28Acksawalmost full speed roms
19:06:10XavierGre.g I want to count a specific number of linefeeds and read that line. Is that possible with a file fucntion or I will have to load data to a buffer first?
19:07:03_FireFly_XavierGr: read_line
19:07:05_FireFly_XavierGr: read_line
19:07:31XavierGrah yes I remember that from radio.c
19:08:11_FireFly_and now also in the wps_data_load fn ;)
19:08:29XavierGrso read_line synat is: file_descriptor, and the other 2 arguments?
19:08:47Acksawanychance of sound on roms?
19:09:09_FireFly_XavierGr: it's the same as for read
19:09:14 Join sanitarium [0] (
19:09:14XavierGrAcksaw it needs major optimizations before that,
19:09:21_FireFly_fd,buffer, buffer_size
19:09:38sanitariumcan you compile the rockbox source on Kubuntu?
19:09:55_FireFly_why it shouldn't ??
19:10:04sanitariumok, all done in konsole?
19:10:10 Quit markun (Read error: 113 (No route to host))
19:10:11_FireFly_no gui
19:10:22XavierGr_FireFly_: So if I want to read the 10th line from a file?
19:10:24_FireFly_who needs a gui anyway to compile it ;)
19:10:33_FireFly_XavierGr: with a loop
19:10:47sanitariumwhat is the best rockbox compiler?
19:10:56XavierGrah so there is no control of it, I will have to read all the file
19:11:02sanitariumty FF
19:11:04_FireFly_sanitarium: under windows cygwin is used
19:11:10XavierGr(or until I reach the line I want)
19:11:16sanitariumi am using kubuntu
19:11:17_FireFly_XavierGr: yes
19:11:21XavierGrme too thanks Fire_Fly.
19:11:57sanitariumDamn, i used to be able to compile things, but i forgot
19:12:03sanitarium"bash: gcc: command not found"
19:12:07***Saving seen data "./dancer.seen"
19:12:17_FireFly_then you need to install it ;)
19:12:47Acksawive been glined on
19:13:01_FireFly_sanitarium: for rb you need also an cross-compiler if you want to compile it for your device
19:13:09sanitariumthats what I want
19:13:14sanitariumAcksaw, why?
19:13:33Acksawa mate told me to join a channel to talk to him
19:13:42Acksawthis channel was a trojan channel which glined me..
19:13:46Acksawno idea how long for
19:13:54sanitariumprobally 24 hours
19:13:56Acksawits not funny..
19:14:02Acksawits been 32 hours
19:14:23XavierGrwhat is glined anyway?
19:14:26_FireFly_sanitarium: for the cross-compiler :
19:14:26Acksawill just play full speed roms for a bit
19:14:55sanitariumcan you play them now?
19:15:01XavierGrAcksaw rockboy isn't fullspeed yet...
19:15:34Acksawseems to me..
19:15:44Acksawi put the cpu freq up
19:15:57Acksawand its playing a hell of a lot faster
19:16:13sanitariumdid you mod the firmware or is it in the settings?
19:16:32Acksawdebug - cpu freg - boost 1
19:16:37XavierGrwell it is faster than the 90mhz you had before but it isn't fullspeed yet.
19:16:56Acksawand my t-dimen sitll has come..
19:17:11AcksawXavier: I have no idea what glined is.. google might help
19:18:14 Quit edx (Read error: 110 (Connection timed out))
19:18:38Maxime`glined = "banned from server"
19:18:40 Join webguest76 [0] (
19:19:06 Nick edx__ is now known as edx (
19:19:09webguest76XavierGr: Better use a fixed size for your entries in the file, and seek to a position in the file, instead of reading the entire file
19:19:41webguest76Just like you'd use the in-memory array
19:20:35 Join _sanitarium [0] (
19:20:47XavierGrwebguest76: I cant use fixed size in the entries.
19:21:00webguest76Why not?
19:21:02XavierGrThey are file names which I split with linefeeds
19:21:52webguest76Pad them with \0 instead and make each entry be the same size
19:22:01XavierGrokay lets say that I put all the strings together, then how could I extract a filename without knowing length and position on the file?
19:22:28webguest76or seek
19:22:30webguest76or whatever it is
19:22:45_FireFly_webguest76: but he can't know how long a filename is
19:23:06webguest76I don't see why not, if he pads them such that each entry is the same length
19:23:26XavierGrso you are saying to make an entry padded with spaces? I will need 360 spaces
19:23:47_FireFly_that is waste of space for me
19:23:47XavierGr(worst case scenario 260 is the biggest filename possible)
19:23:48webguest76Why 360?
19:24:06XavierGr260 sorry
19:24:14webguest76Well, I'd think wasting a few kb of disk space is better than having to read through the entire file each time you skip
19:24:23XavierGrwell something less than 260
19:24:41XavierGrI will try both and see.
19:25:00XavierGrit is more easy right now to test with the readline.
19:25:07_FireFly_read_line ;)
19:25:15XavierGrbest would be to let me search for a linefeed in the file
19:25:17amiconnAcksaw: Are you saying that rockboy runs faster when you boost the cpu manually beforehand?
19:25:41XavierGramiconn: I don't think so.
19:25:43_FireFly_XavierGr: look at the sources of read_line
19:26:00XavierGrI got the same speed last time I tested it.
19:26:32XavierGr_FireFly_: I got how it is working. Or do I need to know more on that?
19:26:37_FireFly_this fn reads on char form the file an checks if it is an new-line or not if not then the char gets into the buffer otherwise the function returns the number of read chars
19:26:49_FireFly_XavierGr: no afaik not
19:27:52XavierGramiconn: what's your opinion, read_line or seek?
19:27:58Acksawamiconn: Yes
19:28:22 Quit markun_ ("leaving")
19:28:23amiconnAcksaw: That's... strange, it shouldn't happen. Rockboy boosts the CPU itself
19:28:36 Join sanitarium_ [0] (
19:28:43Acksawdoesnt for me..
19:29:00_FireFly_XavierGr: you need an buffer to hold the searched filename
19:29:36_FireFly_why not then use read_line which this buffer and after the loop the buffe holds the searched filename/filepath
19:29:41sanitarium_damn, i am a linux n00b
19:29:45sanitarium_how do I install a tar.gz i downloaded? Do I extract it?
19:30:04XavierGrAcksaw I just tested, I get the same speeds with or without manual boost.
19:30:38amiconnXavierGr: If this is for your temporary file name buffer, I'd say fixed field length and using lseek() is the simpler solution
19:30:57webguest76I win!
19:31:05_FireFly_XavierGr: int i=0;
19:31:05_FireFly_char buffer[261];
19:31:05_FireFly_while(read_line(fd,buffer,260) && i <10)
19:31:05DBUGEnqueued KICK _FireFly_
19:31:05_FireFly_ i++;
19:31:10XavierGr_FireFly_: I will use a loop and hold the search in the buffer. Eventually the loop will go to my specified line and store there the filename.
19:31:48amiconnVariable line length will almost certainly be slower
19:32:03PaulJ_amiconn: this would explain why the people were able to play rockboy when the buttondriver acted weird at 124MHz on h3x0
19:32:37amiconn...and if it's for buffering a list where you want to go back and forth, variable line length will be even more complicated
19:33:05amiconnPaulJ_:^ When 124 MHz wasn't stable for H3x0, it was capped at 90 MHz instead
19:33:09XavierGrok then I will try with seek once I get this running with read_line first.
19:33:39_FireFly_yeah at least it depends what it should be done
19:34:23PaulJ_amiconn: yes, but before the cpu was capped the people were playing rockboy without problems
19:35:00webguest76There's no way rockboy is not boosting on h3x0.
19:35:23webguest76#if !defined(SIMULATOR) && defined(HAVE_ADJUSTABLE_CPU_FREQ)
19:35:34 Quit bagawk ("leaving")
19:35:43webguest76That's what decides whether to boost or not in emu.c
19:36:14amiconnYes, since the platforms with fixed CPU frequency have no cpu_boost()
19:37:23 Join linuxstb [0] (
19:37:50 Quit sanitarium (Read error: 110 (Connection timed out))
19:39:30XavierGris there a maximum filename length in FAT32. (without paths, only filename) (of course I mean lower than the known 260).
19:39:45 Join xmixahlx [0] (n=xmixahlx@
19:40:12amiconn255 chars
19:40:14linuxstbRockbox uses MAX_PATH
19:40:29linuxstb(which is 255 - as amiconn said)
19:40:37amiconnNo, MAX_PATH is 260
19:40:46XavierGryes I know that, that's why I asked filename only (without path)
19:40:49linuxstbIs it?
19:41:15XavierGryes it is 260
19:41:19webguest76To make room for: X:\FILENAME\0 ?
19:42:14 Quit _sanitarium (Read error: 110 (Connection timed out))
19:44:01 Quit edx (Read error: 110 (Connection timed out))
19:44:35amiconnHmm, I just noticed that MAX_PATH might need adjustment for unicode
19:45:06amiconnWe store all internal strings as utf-8 now, that is also paths
19:45:47amiconnVFAT allows up to 255 characters per file name, which means 255 UCS2 codes
19:45:50webguest76That could get ugly quick
19:46:06amiconn...which might translate to way more than 260 bytes
19:50:36preglow512 bytes
19:50:50preglowucs2 is always two bytes per char, afaik
19:51:35preglowbut then again, we probably never save it in that representation anyway
19:51:49 Quit TCK ("I shall not cease!")
19:51:54preglowutf8 can be six bytes max per char, yes?
19:53:22amiconnWhen converting from UCS2 only, then utf-8 is max 3 bytes/char
19:53:36 Join _sanitarium [0] (n=sanitari@
19:54:38amiconnFor full unicode, it's 4 bytes/char maximum
19:57:14 Join DrMoos [0] (
19:57:14 Quit Moos (Read error: 104 (Connection reset by peer))
19:57:52 Nick DrMoos is now known as Moos (
19:58:46 Join yngwi [0] (
19:59:07Acksawi wish there was a script that would write what ive listend to on my iriver
19:59:14Acksawthen upload it to my audioscrobbler
19:59:39preglowfind someone to code it :]
19:59:49CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
19:59:49*preglow suddenly remembers he has forgotten his audioscrobbler account completely
20:02:20preglowhah, there's even a rockbox users group
20:04:56 Join TCK [0] (
20:05:30 Quit perplexity (Read error: 113 (No route to host))
20:05:50 Quit Acksaw ()
20:06:35webguest76It was discussed on the mailinglist once, months ago. Iirc, the audioscrobbler guys don't like clients that run offline
20:07:08preglowthat's correct
20:07:16preglowit breaks their spam protection
20:09:39 Join arkascha [0] (
20:09:46 Quit sanitarium_ (Read error: 110 (Connection timed out))
20:09:50FebsWhat is with the recent fascination with overclocking Rockbox? (
20:10:18FebsYou guys are trying to optimize it to use as little CPU as possible, and people want to boost the CPU manually?
20:10:31linuxstbPeople want to start fires?
20:14:12 Join Jungti1234 [0] (n=jungti12@
20:14:28preglowthe sh processors might even take the beating
20:14:48preglowthe coldfires will not act their names
20:14:59preglowthat is, they will in part ;)
20:15:05FebsHa ha. Same question on Mistic River:
20:19:39ghode|afkhow safe is etting rbx continue playing when the bat. level is 0 on the h140?
20:19:58ghode|afkmissing an L there ><
20:20:26 Quit yngwi ("Chatzilla [Firefox 1.5/undefined]")
20:21:10preglowsafe enough
20:21:11 Quit Amar ("CGI:IRC (EOF)")
20:21:20preglowmine has run out a couple of times, and it handles it okish
20:21:30preglowit fails when trying to spin the disk up, eventually
20:21:38preglowbut i wouldn't do it if i could avoid it
20:21:39ghode|afki'm just wondering how it will affect the battery?
20:21:49preglowthe battery _shouldn't_ be affected
20:21:50lostlogicxLiION doesn' tlike deep cycle
20:21:56preglownor does lipoly
20:22:02preglowbut the hardware should cut it
20:22:02ghode|afki had mine playing on 0 for about 10mins
20:22:29ghode|afkkept thinking it was it was going to explode :p
20:23:12preglowno, usually it just goes on, eventually tries several times to spin up the harddrive, then goes down like a ton of bricks
20:23:27ghode|afkbtw are there any plans to remove some of the bloat from rbx? i remember the good old days of 800k downloads :/
20:23:37preglowwell, if you can tell us where the bloat is
20:23:45ghode|afkthemes imo should not be included...
20:23:48preglowdo you mean the bloat _in_ rockbox?
20:23:51preglowor in the zip?
20:23:52lostlogicxunicode was teh big increase recently
20:23:55ghode|afknah not code wise
20:24:04preglowi think wpses should be seperately packaged
20:24:11preglowbut that's an issue bagder has to deal with
20:24:12ghode|afkyes like voice files
20:24:38ghode|afkunicode can go as well!
20:24:40webguest76Aren't fonts much larger?
20:24:49preglowsome of them are
20:24:57preglowno, i think unicode is really nice
20:25:02webguest76Well, I meaent fonts/ vs. wps/
20:25:05preglownot usable to everyone, but like the voice ui, a real selling point of rockbox
20:25:13preglowthat shows we not only care about what the mainstream wants
20:25:29ghode|afkyeah but you dont include voice files with the rbx builds
20:25:35preglowi know, and i agree
20:25:35 Quit Febs ("CGI:IRC (EOF)")
20:25:38preglowi just can't fix it for you
20:26:01amiconnghode|afk: Including voices would increase the .zip size way more than unicode
20:26:02ghode|afki wouldnt expect you too. but thought i'd mention something.
20:26:22ghode|afki am not asking for voice files to be included though
20:26:51amiconnOne voice file is ~1.4MB, and they don't compress much as they mainly consist of mp3 data
20:26:56preglowamiconn: seems people are starting to see through my crappy resampler!! :>
20:28:30amiconnHaha, overclocking the Ondio...
20:29:31preglowi don't know what people are hoping to get out of it...
20:29:31amiconnIt should be possible, but not in software. The xtal would need to be changed (and that would probably cause USB to stop working, as it's the same xtal
20:29:44amiconn12 MHz
20:29:58 Join bazz [0] (
20:30:02preglowperhaps people are starting to look forward to realtime transcoders, heh
20:30:10preglowand btw, anything happening on the wav codec front?
20:30:26amiconnno :(
20:30:39bazzis there a way to get the .hex file out of the .exe file for the iriver firmware on linux? wine seems to hang after i click the Upgrade Firmware button
20:31:23lostlogicxbazz: I haven't found one :(
20:31:36webguest76didn't the most recent one come as a plain zip?
20:32:32bazzyeah, the zip just has an exe in it
20:32:34webguest76Yeah, but istr 1.66 not being like that
20:32:40webguest76could be my imagination
20:32:54webguest76or is this for h3x0?
20:33:25bazzohh wait, wine just came back to life and succeeded :)
20:33:47webguest76Hrm, looks like rasher has them in hex form, if you're desperate:">
20:34:01webguest76and brave, I guess
20:34:08 Join _arkascha [0] (
20:34:19bazzwell, looks like wine did it fine, just had to be patient
20:35:07 Quit Jungti1234 ("bye")
20:37:32 Quit _sanitarium ("Konversation terminated!")
20:37:53bazzif i don't use fwpatcher then my md5 sum shouldn't match any of the ones on the iriverboot page right?
20:38:01preglowhas anyone at all got a 56000 instruction reference? :/
20:38:31bazzalso, is it normal the my new.bin is twice the size of H300.bin?
20:41:03 Quit arkascha (Read error: 110 (Connection timed out))
20:42:43webguest76I believe the md5sum should match
20:43:04webguest76The size growing is normal.
20:43:25bazzhow come it says on the page then: if you used fwpatcher....
20:43:39 Join ich [0] (
20:44:15ichhi! does anybody know when the bootloader v4 for iriver h300 will be released?
20:44:28webguest76bazz: If you used one of the attached bootloader(-version).bin [...] your new.hex (or ihp_120.hex if you used a fwpatcher) should have one of these MD5 sums:
20:44:37bazznm, i forgot to scramble, it matches now
20:48:08 Join mirak [0] (
20:48:08 Quit ich (Client Quit)
20:48:13mirakhello :o
20:49:29bazzwow, it feels nice to be listening to music on my h340 with rockbox :)
20:53:46 Quit _arkascha (Read error: 110 (Connection timed out))
20:57:29*linuxstb likes the idea of a LaTeX manual:
20:57:47 Join hshah [0] (
20:59:11preglowme too
20:59:20 Join MrStaticVoid [0] (
20:59:29preglowcassandra was planning to do that
20:59:36webguest76That sounds really handy
21:00:14preglowlatex's output beats the shit out of most things
21:00:28preglowand for those of us who actually like dabbling in typography, that's nice
21:01:05 Quit hshah (Client Quit)
21:01:34 Join DrMoos [0] (
21:02:04 Quit Moos (Read error: 104 (Connection reset by peer))
21:02:37 Nick DrMoos is now known as Moos (
21:04:01 Join solexx_ [0] (
21:05:09preglowbut ok, i think we should try to get that going soon
21:05:33preglowwe need to do some quick discussions on how we're going to partition the manual into files (to avoid a cvs mess), then just commit it
21:05:41preglowone long file isn't very nice in the long run
21:05:55linuxstbAlso, do we want to use a pre-processor to produce target-specific manuals?
21:06:03preglowyes we do
21:06:08preglowanother nice thing about using latex
21:06:20preglowi think that might be easily doable from within latex itself
21:06:24preglowit does have conditionals support
21:06:55*linuxstb goes to find his LaTeX book
21:07:36preglowi just bought the second edition of the latex book, too bad it's a 100 miles away
21:07:55webguest76preglow: it looks like this effort is using a file pr chapter, afaics
21:07:56*linuxstb returns with The Book
21:08:06preglowwebguest76: which is probably fine
21:09:27preglowthat's what i've always used to do
21:10:08merbananlinuxstb: "the not so short" is a good read also
21:12:11***Saving seen data "./dancer.seen"
21:12:31 Quit xmixahlx ("blah blah blah")
21:16:05 Join muesli_- [0] (
21:17:29 Join jorbond [0] (
21:17:54 Quit bazz ("Leaving")
21:18:23 Quit jorbond (Client Quit)
21:18:33_FireFly_ho ;)
21:19:17muesli_-jojojo ;)
21:20:25mirakhello darling
21:20:35 Quit solexx (Read error: 110 (Connection timed out))
21:21:07_FireFly_that's nice: There are only 10 types of people in the world: Those who understand binary, and those who don't ;)
21:21:25*muesli_- doesnt ;)
21:22:19XavierGrwhy when a stkov error occurs the disk kicks so much?
21:24:27 Join LinusN [0] (
21:25:39Drumr|awayhello, n00b question here.... one. does screen cap work on h300? two. if so, how is it activated?
21:26:09XavierGrFireFly:This is one of the best sites arounds :)
21:26:59Drumr|away:) thank you
21:27:38XavierGrOne of my bests is: Thanks, and DO call again.
21:29:09_FireFly_also nice is "the administrator from hell" :)
21:29:14Drumr|awaysomething weird just happened... i went to boot to rockbox, and it hung in the start screen with battery info being the last bar of info showing... then in about 10 secconds resmed normaly
21:29:17muesli_-not interesting for anyone?
21:29:28 Join _arkascha [0] (
21:37:15saa[b_r]idermuesli: checking it out...
21:42:06 Join gromit` [0] (n=gromit`
21:44:41saa[b_r]idercan someone explain to me what the relation between CP1252 and ISO-8859-1 is?
21:45:08LinusNis there one?
21:45:50preglownot much
21:46:00preglowwhy would knowing that relation help you in any way anyway?
21:46:32saa[b_r]iderum.. I'm not sure to be honest.. but CP1252 is the "page code" for english, and ISO-8859-1 is "charachter encoding"
21:47:17saa[b_r]iderfor english as well
21:47:20_FireFly_boths are charakter encoding
21:47:39saa[b_r]iderare they just different names for the same thing?
21:47:51saa[b_r]iderlike two different standards?
21:48:11_FireFly_not at all afaik cp1252 has more charakters defined
21:48:25saa[b_r]iderI'm writing up a simple Unicode Guide for the wiki, and I want to use the right techincal names... that's why I want to understand
21:48:55saa[b_r]iderwell both 1252 and 8859-1 are for most western languages, not only english...
21:52:03_FireFly_there is a compersation of voth
21:53:04 Quit linuxstb ("CGI:IRC")
21:53:11saa[b_r]iderthanks firefly :) I should've checked wikipedia... only I became lazy to check it ever since it became blocked in china
21:53:30saa[b_r]iderbut I have my ways around :)
21:53:38 Join linuxstb [0] (
21:54:04mirakI built a coldfire gcc, on a x86 ubuntu, I want to know if I can use the binary I created on another x86 box
21:54:16mirakI mean the gcc binaries
21:54:33preglowthat's a distro issue
21:54:42preglowi have successfully done so myself
21:54:45mirakthat's also an ubuntu
21:54:52mirakso it's ok then
21:55:01_FireFly_mirak: also x86 ??
21:55:02preglowthen you should be fine
21:55:10 Quit `3nergy (Read error: 104 (Connection reset by peer))
21:55:19mirakwelll actually one box is an amd64 but it's x86 ubuntu on it
21:55:34_FireFly_then ut should be no problems
21:55:37preglowas long as it compiles to 32 bit binaries
21:55:40mirakI tried amd64 ubuntu, but you need to chroot for some stufs
21:55:52miraklike flash, that's not handy
21:55:52preglowi use amd64 ubuntu, it behaves like a normal ubuntu
21:56:13preglowoh, if it's one thing i don't miss, it's that
21:56:17mirakpreglow: I had troubles with firefox and some weird crash of some applications
21:56:29mirakpreglow: flash ? :)
21:56:36preglowevery time i see an empty frame where flash should be, i praise he lord
21:56:39preglowthe, even
21:57:05mirakI use flashblock. But some usefull sites like online city maps, use it :-/
21:57:08LinusNpreglow: except when you go to, where the entire page is blank
21:57:09XavierGrok I have a problem: Lets say that I have a huge list with filenames on a file (no option to load it into memory) which is the best way to sort that list alphabetically?
21:57:38LinusNXavierGr: sort.rock
21:58:00preglowLinusN: oh, that's ok, i praise the lord whenever i'm prohibited from visiting that site too
21:58:10XavierGrah yes bravo you always come handy.
21:59:11 Join jlo [0] (
21:59:19preglowthere he is
21:59:29preglowso, have you tested your new crossfeed proposal?
21:59:39XavierGrhehe :D
22:00:05mirakdid iriver used assembly for the ogg and mp3 codecs ?
22:00:12preglowmirak: almost certainly
22:00:13jlohello all
22:00:32mirakpreglow: you can't play real time with just C code ?
22:00:37preglowmirak: i would not be surprised to discover that the bigger parts of the mp3 codec is assembler
22:00:47preglowmirak: 'course you can, but it will always be less efficient
22:00:58 Quit _arkascha (Read error: 110 (Connection timed out))
22:01:00preglowmirak: unless you have The Perfect Compiler
22:01:05mirakwell if it could save battery it would be greeat
22:01:26preglowwell, it does
22:01:35preglowour mp3 decoder was slow as hell before we started using asm
22:01:36mirakpreglow: C code compiled will always be biger
22:01:39preglowand iram, but that's another matter
22:02:11mirakwhat's iram ?
22:02:16preglowfast internal ram
22:02:32preglowone of the biggest, if not the biggest, optimisation technique we employ
22:02:40mirakhow much is there ? 64kbyte ?
22:02:46preglow48kb for codecs
22:02:49jlopreglow : I tried but came to some strange things so I returned to your values to measure that gaves those curious curves
22:02:50preglow48kb for rockbox
22:03:01 Quit muesli_- (Read error: 104 (Connection reset by peer))
22:03:16preglowjlo: the crossfeed that is in rockbox now has to go, or at least the highpass filter
22:03:24preglowwhich is a bad joke
22:03:31mirakpreglow: swapping iram data to slow ram is slow I guess
22:03:41preglowmirak: well, yes, of course
22:03:42XavierGrLinus: sort.rock will not run out of memory even with a very long file list, right? (like 2000 lines/ 30 chars each)
22:04:40mirakpreglow: still wonder how they managed to include a xvid decompressor ...
22:04:53mirakan image would be bigger than the iram
22:05:36preglowit's not necessarily the finished product that should be kept in iram
22:05:42preglowwhatever data sees the most accesses should be in iram
22:05:55jlopreglow : the highpass has a too high FC and should also be shelving (limited to about +4 dB, not more)
22:05:58_FireFly_if they fit
22:06:15preglowjlo: the highpass filter isn't really highpass
22:06:24mirakpreglow: since mpeg is per block ...
22:06:26preglowjlo: ok, so there's no way around not having shelviung filters?
22:06:55preglowmirak: but yes, i wont argue that implementing xvid is a good feat, it'll be quite some time before you see that in rockbox
22:06:59LinusNXavierGr: sort.rock depends on being able to load the entire file, so it uses the entire audio buffer
22:07:25LinusNif you want so sort files larger than the available memory, google for "merge sort"
22:07:29preglowjlo: shelving filters will be somewhat slower to implement
22:09:26jlopreglow : the highpass is only to keep a balance between low freqs and high, bass is mainly correlated so it adds at +6dB and treble is uncorrelated so it adds at +3dB, that's why you have to push up treble by about 3 dB
22:09:44preglowjlo: yeah, that's true
22:10:13preglowso the filtering isn't done to do a rough simulation of filtering done by the head?
22:10:14 Join ts-x [0] (
22:11:51mirakpreglow: I am interested in trying to do that, but that's a long way ^^
22:11:56 Quit MrStaticVoid ("Lost terminal")
22:11:59jlopreglow : it's the lowpass function to filter like the head, that's about -12dB at higest freqs
22:11:59XavierGrLinusN: I see that as a last resort in sorting. I would like to make jpeg.rock to be able to scroll files without stopping playback (when I include the don't stop playback patch) (of course that is for small pics that can be handled without the need of the audio buffer.) But anyway a filelist should never be more than 1MB (at most).
22:13:00mirakI have seen it's possible to use tags for C in editors, anyone is using that ?
22:13:41jlopreglow : sorry but the lowpass has also to be shelving !
22:13:48_FireFly_mirak: ??
22:14:02mirak_FireFly_: mmm ?
22:14:08mirak_FireFly_: it's called etags
22:14:18ts-xI installed Rockbox on my H340 for the first time's a thing of beauty! Thank you all for your efforts.
22:14:21preglowjlo: but ok, what i'd like before embarking upon this is good test results, at least
22:14:30_FireFly_mirak: do you mean for syntac-highlighting??
22:14:35mirakit's used to have references on types etceetera in the EID
22:14:51mirak_FireFly_: more like what eclipse is doing for java
22:15:05mirakyou have a variable, eclipse can tell you the type, where it's declared, etceetera
22:15:06preglowts-x: nice to hear you like it
22:15:17jlopreglow : I looked at dsp.c, I'm no programmer but where are the Fc of filters ?
22:15:28preglowjlo: there are none, the coefficients are hard coded
22:15:36mirak_FireFly_: I remember it can be used with VI and various editors
22:15:49mirakI hate vi so I haven't tried
22:17:05jlopreglow : what I'll do is take a Yamaha DME24 or 64 (DSP engine) and make simulations on it, then I'll tell you if it works before you do too much work
22:19:24mirakpreglow: I am not sure what is the fasted mpeg codec
22:20:01mirakdon't know if theora, the vorbis video codec is fine
22:20:17ts-xA few oggs are still skipping with the newest H3xx build. I've got a q9 example I can provide to any interested dev.
22:20:32 Part LinusN
22:20:56preglowjlo: that sounds great
22:21:02preglowjlo: samples would be cool as well
22:21:05jlopreglow : when you do those dsp function, do you test it before on wavewarp or some other mathlab/audio programms ?
22:21:22preglowjlo: ordinarily, i would have modelled it either in sync modular or reaktor5
22:21:37preglowi only have sync modular available, though, and that can't load samples
22:21:42preglowit's great for realtime testing
22:22:16jlopreglow : samples is a good idea, h
22:23:09jlopreglow : tell me if you know which samples would be interese
22:23:46preglowold rock would probably do
22:23:55preglowlucy in the sky with diamonds by the beatles is a nice example
22:24:03preglowsince it actually is annoying as hell to listen to without crossfeed
22:24:03 Join thegeek_ [0] (
22:24:04 Join chiller_ [0] (
22:24:18jlostrawberry fields also !!!!
22:24:21preglowhahah,. yes
22:24:26preglowmarkun has some nice samples you could ask for
22:24:47 Quit ReKleSS (
22:24:47 Quit jlo (
22:24:47 Quit DangerousDan (
22:24:47 Quit Strath (
22:24:47 Quit saa[b_r]ider (
22:24:47 Quit Cassandra (
22:24:47 Quit PaulJ_ (
22:24:47 Quit DJDD_ (
22:24:47 Quit talaus (
22:24:47 Quit ze (
22:24:47 Quit HCl (
22:24:47 Quit thegeek_ (
22:24:47 Quit solexx_ (
22:24:47 Quit Drumr|away (
22:24:47 Quit Shaikoten (
22:24:47 Quit Lynx_awy (
22:24:47 Quit merbanan (
22:24:47 Quit crwl (
22:25:31NJointhegeek_ [0] (
22:25:31NJoinjlo [0] (
22:25:31NJoinsolexx_ [0] (
22:25:31NJoinDangerousDan [0] (
22:25:31NJoinStrath [0] (
22:25:31NJoinsaa[b_r]ider [0] (n=saab_rid@
22:25:31NJoinCassandra [0] (
22:25:31NJoinPaulJ_ [0] (
22:25:31NJoinDJDD_ [0] (
22:25:31NJoinReKleSS [0] (
22:25:31NJoinShaikoten [0] (n=shaikote@
22:25:31NJoinDrumr|away [0] (
22:25:31NJoinLynx_awy [0] (
22:25:31NJoinmerbanan [0] (
22:25:31NJointalaus [0] (
22:25:31NJoinze [0] (
22:25:31NJoincrwl [0] (
22:25:31NJoinHCl [0] (
22:25:51preglownetsplits, my favourite feature of irc
22:26:18jlowelcome back to all
22:27:33 Join _CoCoLUS [0] (
22:28:29 Join novimon_ [0] (
22:28:29 Join crash__ [0] (
22:28:31 Join _pill [0] (
22:29:13 Quit ts-x ("CGI:IRC (EOF)")
22:29:40 Join XavierGr_ [0] (
22:32:08jlopreglow : just to know, when you do a filer or the like, do you know exactly how long it takes through ? that means that if you add signals like in crossfeed are you sure both signals are timed ?
22:32:21CtcpVersion from freenode-connect!freenode@freenode/bot/connect
22:32:22***Server message 477: 'logbot #RockBox :[freenode-info] If you're at a conference, please contact freenode staff to make sure we've made special allowance for many users coming into our network from a single internet address ( ). Private messages from unregistered users are currently blocked, except to network staff, services and participating registered users ( )... Thanks!'
22:33:41preglowjlo: no time delay as such
22:33:45preglowjlo: there will be a phase shift, though
22:33:55preglowjlo: limited to 180 degrees for first order shelving filters
22:34:49jlopreglow : so only the phase of the filter itself
22:34:51 Join muesli_- [0] (
22:34:55preglowbut this wont be a problem, will it? after all, you're not going to add the filtered signal to itself
22:35:08preglowthen it would be a problem
22:36:09jlopreglow : no problem, but I thought that if there would be one or some samples delay, then you have to compensate
22:39:13jlopreglow : for shelving filters, I think you add an allpass to a highpass ie, then maybe you have to compensate if both filters don't take same time
22:39:18 Join VikraMarkA_ [0] (
22:39:26preglowthere might be problems, yes
22:39:31preglowbut we'll be using iir filters
22:39:45preglowand with first order filters, there is no delay as such, just a phase shift
22:39:53preglowwhich you just have to live with
22:39:57jlothat's fine
22:40:10 Join NibbIer [0] (
22:40:15 Join Bagder_ [0] (
22:41:12 Join Vladoman [0] (
22:41:25 Quit YouCeyE (Read error: 110 (Connection timed out))
22:41:25 Join mirak_ [0] (
22:41:32 Quit phaedrus961 (Read error: 110 (Connection timed out))
22:41:38 Quit Nibbler (Connection timed out)
22:41:40 Quit crash_ (Connection timed out)
22:41:47 Quit kkurbjun (Read error: 110 (Connection timed out))
22:41:54 Join phaedrus961 [0] (
22:41:55 Quit vHs (Read error: 110 (Connection timed out))
22:42:05 Quit novimon (Read error: 110 (Connection timed out))
22:42:07 Quit CoCoLUS (Read error: 110 (Connection timed out))
22:42:07 Quit pill (Read error: 110 (Connection timed out))
22:42:08 Join Jungti1234 [0] (n=jungti12@
22:42:14 Quit Vlad0man (Read error: 110 (Connection timed out))
22:42:18 Quit mirak (Connection timed out)
22:42:19 Quit Bagder (Read error: 110 (Connection timed out))
22:42:25 Quit golf7 (Read error: 110 (Connection timed out))
22:42:30 Quit _FireFly_ (Read error: 110 (Connection timed out))
22:42:40 Quit DreamTactix291 (Read error: 110 (Connection timed out))
22:42:48 Quit chiller (Connection timed out)
22:42:48 Nick chiller_ is now known as chiller (
22:42:59 Nick Bagder_ is now known as Bagder (
22:43:13 Quit TCK (Read error: 110 (Connection timed out))
22:43:15 Join vHs [0] (
22:43:20 Quit thegeek (Connection timed out)
22:44:38 Quit linuxstb ("CGI:IRC (Ping timeout)")
22:45:11 Quit Midgey34 (Read error: 110 (Connection timed out))
22:45:14jlopreglow : is it easier to do a filter and then do its complementary (1-filter) instead of highpasss and lowpass ?
22:46:17 Join RotAtoR [0] (
22:47:18preglowjlo: of course, but you can't use that approach for all filters
22:47:45preglowjlo: for example, i don't think you can calculate a low shelf, then just do 1 - filter and suddenly have a high shelf
22:48:30jlopreglow : sure this won't work but for pure high or lowpass
22:49:29preglowdepends on the phase of the filter, primarily
22:49:30preglowbut yeah
22:49:37preglowwe can't use that if we have to use shelving filters
22:49:54preglowyou can do a shelving filter as 1 + lowpass/highpass, of course
22:50:11lostlogicxpsst preglow: merge profiling patch (or are you not going to merge any of it until you write a non-java parser?) :-D
22:50:22preglowlostlogicx: i want to try it first
22:50:41preglowperhaps i will tomorrow, i need to go home soon, and want to be around to fix up the commit if i break a few builds
22:50:46 Quit XavierGr (Read error: 110 (Connection timed out))
22:51:04preglowthe perils of not having a computer where you currently live :/
22:51:17lostlogicxthat's insane man... don't know how you live.
22:51:23preglowit's temporary
22:51:35jlopreglow : but a shelving lowpass is for example lowpass + attenuated signal, no?
22:51:47preglowi've relocated for some time to work for some people
22:51:51preglowand i'm at my office right now
22:52:02preglowjlo: should be
22:52:07lostlogicxpreglow: gotcha −− I have at least 1 computer wherever I go.
22:52:17preglowlostlogicx: yeah, i usually do as well
22:52:30 Quit ReKleSS (
22:52:43NJoinReKleSS [0] (
22:52:47preglowthough it is refreshing not to have one as well
22:52:50preglowat least i get some reading done :-)
22:53:16lostlogicxpreglow: yeah, that is nice once in a while :)
22:54:09lostlogicxwith my current optimizations, boost ratio is about 50% on 'normal' Q7 oggs with 124mhz... 50% is still pretty hard on battery life I'd imagine though.
22:54:28preglowso, what have you done?
22:54:38 Join linuxstb [0] (
22:56:42webguest76That's quite an improvement, isn't it?
22:57:07lostlogicx1) ripped out all of the ICODE_ATTRs in Tremor previously, 2) gave a bunch of lookup data that was still not in IRAM ICONST_ATTRs, 3) moved the largest frequently called symbols into IRAM to avoid cache thrashing. the remaining top 50% of symbols (in terms of calls) should all fit in cache.
22:57:46preglowwell, just remember that using iram for data should provide the best boost
22:57:50lostlogicxstill have about 4k of IRAM left, I think, but moving more code in there probably wouldn't help.
22:57:52preglowwe can fix the cache issue in other ways
22:57:55lostlogicxpreglow: all IRAMable data is IRAM'd
22:57:55preglowbut anyway
22:58:05preglowhow much of an improvement have you done?
22:58:19preglowthat is, what was the boost rate before opts?
22:58:23lostlogicxboost ratio wise at 90mhz, I got from 99 to 94% boost.
22:58:45lostlogicxI'll do some better tests this evening (playing single full songs on different optimizations)
22:58:47preglowwhich is pretty decent for just a redistribution of iram
22:59:12lostlogicxyah, but still skips on the really bad parts of q7s at 90mhz :(
22:59:43lostlogicxoh, question for you:
23:00:05lostlogicxI get a area type conflict or some such when I try to move a couple of const arrays to IRAM
23:02:42preglowdo you use ICONST_ATTR?
23:03:43lostlogicxgah, I must have, when it needs to be IDATA_ATTR, I thought I'd tried both, but apparently not.
23:04:34 Join `3nergy [0] (
23:04:38 Quit jlo ()
23:10:26XavierGr_one last question: can I delete a line from a file?
23:10:39XavierGr_or characters
23:12:12***Saving seen data "./dancer.seen"
23:16:10linuxstbXavierGr: No. The only operations you have are read, write, seek and (I think) truncate.
23:21:39lostlogicxup to ba50 out of c000 bytes of IRAM used in Tremor.
23:22:34preglowlostlogicx: you're starting to come close to what i use in my local copy of libmusepack, then ;)
23:22:38lostlogicxand why does extracting take 60-90s in windows and 5-10 in linux?
23:22:40preglowi think i've got something like 48 bytes free
23:22:44lostlogicxhaha, nice
23:23:07lostlogicxso, is there a 'no cache' symbol that can be used to tell coldfire not to cache certain functions?
23:23:14lostlogicxbecause I could use that to pretty good effect too.
23:23:14ender`lostlogic: do you extract directly to the player?
23:23:17preglowlostlogicx: no
23:23:19lostlogicxender`: yes.
23:23:34ender`linux by default caches writes to USB devices, Windows doesn't
23:23:34preglowlostlogicx: what we can do is localise functions that are important to cache
23:23:47preglowlostlogicx: like group them all as tightly together as we can
23:23:59lostlogicxender`: I was including unzip;cd;umount /mnt/xd in the time.
23:24:30lostlogicxpreglow: sounds good, how do I do this in code?
23:24:32preglowlostlogicx: windows is about as fast as linux for me
23:24:48ender`it's faster to write something from cache than to read from disk, then write again
23:24:50preglowlostlogicx: you can't, yet, we need to intruduce a new section for it
23:26:02ender`simply try this: connect the player on Windows, go to Device Manager, double-click the player's HDD, find the checkbox that enables caching, reboot Windows (required) and try again (you'll need to unmount the drive through the Unplug/eject hardware icon in the system tray then)
23:26:51amiconnI would never enable write caching for a removable drive
23:27:16ender`it does wonders in terms of speed
23:27:18amiconnI got burnt once by doing this; once is enough
23:27:35lostlogicxpreglow: gotcha −− I really wish there was just a "don't cache me" flag in this case, because from profiling, I know that there are some pretty big symbols which are called every few ms which we don't want to polute the cache.
23:27:37ender`i'm careful enough to always use the unplug icon
23:27:43lostlogicxender`: ty
23:28:02lostlogicxamiconn: being used to linux, I know the risks and don't kill myself with them.
23:28:04amiconnSince I have write caching disabled, I can just rip out the USB cable
23:28:41amiconnI like to be able to simply disconnect, lengthy unmount procedures are cumbersome
23:29:11linuxstbIs there a windows equivalent to the umount command?
23:29:16preglowlostlogicx: we can only tell the cache what address ranges we don't want it to cache
23:29:17linuxstb(i.e. a command-line command)
23:29:23preglowlostlogicx: and i don't know how far we can use that mechanism
23:29:49 Quit Jungti1234 ("bye")
23:30:01webguest76I don't seem able to disable write caching for my iriver on windows for a non-priviledged user
23:30:09preglowit never was enabled for me
23:30:11preglowdisabled per default
23:30:14preglowwhich suits me just dandy
23:30:27webguest76seems to be enabled here
23:30:35webguest76And I can't turn it off. It's infuriating.
23:30:51amiconnWrite caching for removable devices is disabled by default in WinXP as long as they are 'pure' USB MSD
23:30:56lostlogicxpreglow: *nod* thanks for the information.
23:31:27webguest76amiconn: I'd like to explain my installation of WinXP that.
23:31:28amiconnwebguest76: If you disable it once as admin, it should be disabled for everyone.
23:32:17amiconnHowever, windows remembers this setting per device, so if for some reason your installation sets it to enabled by default, you'll have to configure each device individually
23:32:23amiconn(or find the global switch)
23:32:36preglowgreat, now my ipod earbuds are broken as well
23:32:43preglowi have no means to listen to music anymore
23:33:28preglowdamned piece of shit junk
23:33:38lostlogicxpreglow: :(
23:33:42webguest76Hrm, if I go to the properties for this device, it shows a greyed-out screen where "Optimize for performance" is ticked
23:34:01preglowbut ok, i need to go
23:34:02preglowlaters all
23:35:41amiconnwebguest76: Hmm, then for some reason your system thinks this is a non-removable device
23:35:45webguest76However, if I login as an admin-account, I see "Optimize for quick removal" enabled. Dear Windows: wtf.
23:36:07amiconnAh, yes, you need to be admin as I said earlier...
23:36:29webguest76Well yes, but does setting in the greyed-out screen matter?
23:36:37webguest76Because there, it's set to the wrong option
23:36:43amiconnI'm not sure...
23:37:51webguest76Is there anything like /proc/mounts in windows, where you can see what settings mounted devices have?
23:38:12webguest76Is there anything in Windows like /proc/mounts, rather
23:38:58webguest76See, this is what I don't like about Windows - trying to shield me from what's going on, and being damn confusing about it to boot
23:40:41 Join edx [0] (
23:42:20webguest76Guess I'll just do the unmount ritual
23:43:41amiconnThis one's interesting:;en-us;811392
23:45:27webguest76guess I'll see what that one says
23:45:58webguest76Oh, you can't just get it, it seems
23:48:50webguest76"safe surprise removal" tee-hee
23:49:44 Join San [0] (
23:54:14Santhe cygwin copmpiler doesn't work for me
23:54:26SanKYL3 is helping me and I followed the steps
23:54:32webguest76Hm, that dskcache tool only shows the ide harddisk
23:54:44webguest76Oh well.
23:54:55webguest76San: what's the error?
23:56:52Sanany ideas?
23:57:08Sani tryed rebooting and searching for the files, but nothing
23:57:48webguest76I'm fully ready to blame the devkit for this.
23:58:22webguest76Have you tried searching for cygwin1.dll to see if you have multiple versions installed?
23:58:23Bagderdon't use rbconf to start with
23:58:33Sanwhat do I use?
23:58:43webguest76Well, it doesn't seem related to rbconf, does it?
23:58:49Bagdernot really
23:58:54webguest76San: ../tools/configure
23:58:56Bagderbut using it hides what it does

Previous day | Next day