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 2006-04-06

00:04:36 Part stripwax
00:04:36 Quit Chamois (Read error: 104 (Connection reset by peer))
00:04:50 Join Chamois [0] (
00:05:12 Join [TCK] [0] (
00:06:01 Quit Chamois (Client Quit)
00:06:58 Join WVU2034 [0] (
00:07:47WVU2034guys i just have a quick question for the x5 when i go to load the rockbox file in the gui it says rolo loading then executing then does nothing. any suggestions ?
00:08:00 Join luckz [0] (i=alcyon@
00:08:26Bagderso don't
00:08:37luckzquick question.. can I manually query a non-standard tag field from a .wps, like disc/discnumber or so?
00:08:37 Quit imphasing (Read error: 104 (Connection reset by peer))
00:09:04*Bagder runs off to bed
00:09:37 Join imphasing [0] (
00:09:56 Quit Kohlrabi ("Fast alle Menschen sind Regenwürmer")
00:10:46 Join |Unknown| [0] (
00:10:56WVU2034anyone ?
00:11:21preglowdon't load it
00:11:23preglowit's not supported yet
00:12:04WVU2034so how do i get the wps files to work
00:12:14luckzWVU2034: to work how?
00:12:32luckzhow to browse them or..?
00:12:42WVU2034well theres a rockbox file
00:12:43|Unknown|hey is Rockbox like a big browser?
00:12:56WVU2034and i click on that and it wont load on the x5
00:12:56luckzWVU2034: player?
00:13:06luckzthat file doesnt do anything
00:13:15WVU2034o no
00:13:15luckzi took a few hours to find the actual menu, WVU2034 :P
00:13:30luckzjust hold down the stick
00:13:31 Quit Mark_ (Read error: 110 (Connection timed out))
00:13:51luckzthat gets you a menu with the first option being playist
00:13:54|Unknown|where do I put .wad files?
00:14:06luckzthen press the REC button
00:14:09luckzand navigate from there
00:14:54WVU2034let me try it
00:15:15|Unknown|where do I put .wad files?
00:16:00|Unknown|. . .
00:16:47preglowwiki page...
00:16:50|Unknown|wheres this secret
00:17:06preglowwhat the hell, have we been overrun by people who don't like reading web sites?
00:17:13WVU2034thanks luckz i appreciate the help
00:17:21Mikachuand people who want to play doom
00:17:36WVU2034luckz what theme works the best for u
00:18:22|Unknown|put the doom file in a folder called games/doom/
00:18:41|Unknown|but when i go to Plugins on my iPod and select doom it says Sorry, Not enough base words
00:18:44|Unknown|or something
00:18:54sharpeyou need the required wad too.
00:19:00 Quit PaulJ (".")
00:19:18 Join RaeNye [0] (
00:19:34|Unknown|i did, i put the wad in games/doom/
00:19:57sharpesure it's the required one?
00:20:40|Unknown|i downloaded the one from the rockbox site
00:21:31RaeNyeDoes it matter whether you call it /games/doom/ or /Games/Doom/ ?
00:21:40 Quit swampcow (Read error: 104 (Connection reset by peer))
00:22:04sharpe|Unknown|: rockdoom.wad ?
00:22:48|Unknown|well lemme check
00:23:13 Quit petur ("zzz")
00:23:19|Unknown|279 k
00:23:26RaeNyeIs any dev alive here? I've just posted my first patch and I want it commited :))
00:23:49sharpefeature freeze...
00:23:52lostlogicRaeNye: there are dozens of patches on the tracker, is there a particular reason that yours takes first priority?
00:23:59RaeNyeit's a bugfix
00:24:16sharpeyay bugs
00:24:37lostlogicRaeNye: link?
00:24:50RaeNyethe unstoppable sleep timer one (#5001)
00:25:17|Unknown|so um anyone?
00:26:06*RaeNye points at
00:26:43*amiconn spotted a bug in the iriver button driver
00:27:07*sharpe spots a bug outside my window.
00:27:45amiconnThe iriver button driver wouldn't pick up a main unit button press as long as a remote button is pressed (goes for buttons hooked to adc)
00:28:19*RaeNye is being spotted by bugs from outer space
00:28:22 Quit Rob2222 ()
00:29:23 Join Rob2222 [0] (
00:30:53|Unknown|hey is Rockbox suppose to when you go into it show you all your files?
00:31:38*preglow needs the norwegian equivalent of "tag" :/
00:31:43lostlogicRaeNye: I approve of your patches in a general sort of way. Is there any way to avoid rounding either direction unless the value is actually changed in the menu?
00:32:10preglow|Unknown|: hey, is you coherent sentence make
00:32:11lostlogicpreglow: keyword, metadata?
00:32:45preglowhmm, metadata might work
00:33:37RaeNyeI guess the only way would be to make the sleep timer value count seconds instead of minutes, updating step size to 300=5*60
00:34:09luckzWVU2034: I use a modified S3T_003
00:34:29luckzI added current time, genre, track number and info about the next track
00:34:35preglow"commiting tagcache" <- THE worst sentence to translate to norwegian
00:34:39RaeNyeeither that or a ugly patch in sleep_timer that saves the value beforehand and resets it on set_int returning false.
00:34:41luckznot used that extensively
00:34:41|Unknown|hey can you use IpodWizard without losing Linux?
00:35:17RaeNyeThe problem of this last patch is that you lose the time spent inside set_int's menu
00:36:44luckzdoes playcounting work? if so, how?
00:37:23lostlogicRaeNye: applied your patches as is, as they are better than the previous behavior for sure. Thanks.
00:37:54RaeNyeBTW, does anyone here uses a dark background? if so, do you have a problem with hard-to-see splash messages (e.g., "Shutting down...")?
00:38:11RaeNyelostlogic: no, thank YOU :)
00:38:12preglowlostlogic: remember CREDITS
00:38:17lostlogicpreglow: I did!!!
00:38:23preglowi haven't gotten all mails yet
00:38:27lostlogic*bites your finger*
00:38:30luckzI mainly had a problem with not being able to read anything while in a car using a dark background, not fully solved that yet due to lack of time
00:38:34*preglow kicks ass
00:38:53lostlogicRaeNye: yeah, that's a problem, because the splashes and a few other places don't respect all color settings properly
00:38:58lostlogicfile a bug? :)
00:39:03RaeNyenow if only I could run the UI sim on linux instead of uploading it to the real thing every time.... <sob>
00:39:18lostlogicRaeNye: SDL doesn't work for you?
00:39:32lostlogicRaeNye: I never use the sim myself, but then I may be a bit masochistic
00:39:33RaeNyealsa conf.c segfaults
00:39:47lostlogicRaeNye: try using esd or oss sound?
00:40:10RaeNyeI ran the precompiled win32 sim, but my build environment is gentoo
00:40:35RaeNyehow? i have alsa oss emulation
00:40:46lostlogicRaeNye: not sure, I'm no SDL expert, it was just a thought.
00:40:55|Unknown|Do you run iBoy through Linux, or Rockbox? or Will it be located in IPL Bootloader?
00:40:59*preglow starts wondering if anyone will ever use this lang file, then reconsiders
00:41:24lostlogicpreglow: reconsiders as in stops working on it or stops wondering?
00:41:38lostlogicpreglow: I'm sure your sister will use it... if you give her your nano... if you have a sister...
00:42:05RaeNyeI want to check whether the problem persists with a uisim that was compiled somewhere else. Anyone has a linux build somewhere over the net?
00:42:30lostlogicRaeNye: why not compile the sim yourself?
00:42:43RaeNyedid so. segfaults
00:43:20RaeNye(I want to know whether to blame my runtime env)
00:43:29DrumRBoy320|Awaydoes rockbox for the ipod understand chargers and external batteries?
00:43:40 Quit BHSPitLappy (Connection timed out)
00:43:55 Join BHSPitLappy [0] (
00:43:58lostlogicDrumRBoy320|Away: it understands AC charger and USB charge, don't know about batteries
00:44:18luckzso nobody has a clue whether I can in any way show custom tags?
00:44:37lostlogicand on that note, I bike home (just to make preglow jealous again)
00:44:58preglowlostlogic: i've got a half-sister, and she rather fancies my nano...
00:45:11preglowquite luckily she's in england, so she never sees it
00:45:12preglowi win!
00:45:22preglowoh, and yeah, she doesn't speak norwegian
00:46:05*RaeNye is away
00:48:20 Quit thegeek (Read error: 113 (No route to host))
00:50:02|Unknown|hey whats the difference of these 2 files?
00:50:29luckzI'd assume the former to be newer
00:50:35luckzthat's the cvs, no?
00:51:10luckzbasically the second was done at 6 AM GMT, AFAIK
00:51:40luckzthe former gets dynamically updated every time somebody changes something
00:51:49 Part raf82
00:55:01|Unknown|which is better to use?
00:55:44midkaythe very latest..
00:56:15scottderdid the line out for nano get changes....adjustign colume used to adjust output levels on the line out
00:56:19scottderbut now it doesn't
00:58:48preglownow it's always fixed
00:59:00preglowthe way the good lord intended it to be
00:59:00scottderAh cool :)
00:59:10*scottder buys preglow a beer
00:59:14midkayline out on nano? :o
00:59:17*preglow explodes in glee
00:59:32scottdermidkay: yup through the doc port
00:59:37preglowfinally all this rockbox coding pays off in beer!
00:59:45midkayscottder, ah.. nice. :)
01:00:00*scottder hands preglow white can that simply says "BEER" on the side
01:00:17scottderI picked up one that is just a line out....quite nice
01:02:49 Join quobl [0] (n=quobl@tor/session/x-9526b2fe7f6c6054)
01:02:53goffascottder: that's not very nice... you should at least go for something that hasn't been filtered through a human :)
01:02:54 Quit bluebrother^ ("Leaving")
01:03:08 Join harbel [0] (
01:04:36*scottder breaks out the Guinness
01:04:42goffathat's more like it
01:04:44goffahe he he
01:06:25Zoide777I just created a thread in the forums for those of you who might be interested in getting recording to work on the iPod:
01:06:42Zoide777It's just the start of a project, but I guess at least it's a start
01:06:46*preglow fetches his trusty bottle of gammel dansk to not feel completely left alone
01:06:54macncheesewere there tests yet with the h140 for playing video?
01:07:15Zoide777(though if someone realizes they can hack it up alone in a couple days, by all means go ahead :) )
01:07:46Zoide777the more eyes, the better...
01:08:06Zoide777(and hopefully more finger for typing code too!)
01:12:37macncheeseone more small question, how reliable/precise is the battery time specified in rockbox on h140?
01:13:56preglownot at all
01:14:07preglowit's pure guesswork based on vague uncertanties
01:14:10macncheesethat's what i figured heh, it's saying i have 6 hours left
01:14:16macncheesewhere im more thinking i have about 1 hour left
01:14:21macncheesealthough i hope it's right :]
01:14:38preglowwhat time you have left is depending on so many factors it's not funny
01:14:47preglowwhat codec, how many disk access, if you browse around a lot
01:15:08macncheeseyea, it's a bit off though
01:15:09macncheese(or i am)
01:15:27macncheesei bet playing gameboy games didnt quite help the battery either :O
01:16:18***Saving seen data "./dancer.seen"
01:16:18 Join nave7693 [0] (
01:16:49preglowand a sure bet it'd be
01:17:18 Join unknownkind [0] (
01:17:24macncheeseonce the plasma screensaver kicks in though, you know it has all been worth it
01:17:57macncheesehaha there's a chip-8 emulator, that's brilliant
01:18:04macncheese(i'm quite new to rockbox, as you might notice)
01:18:30macncheesewell.. i've been following it from even before they started considering the iriver series, but only installed it on my iriver yesterday :)
01:18:32unknownkindHey why do I get these to errors while loading the modules for linux? Warning - Send text input selected. Make sure a keyboard is attached or select a different input method, Error Unable to connect to MPD
01:18:52Zoide777I think there is a mistake in the IpodHardwareInfo wiki (
01:19:05Zoide777it says that CPU boosting is working fully, but afaik it doesn't work on 4g
01:19:09Zoide777(grayscale at least)
01:19:10macncheeseit's a wiki, go edit it :]
01:20:00amiconnZoide777: You could test whether boosting works on your 4g
01:20:07amiconnThat would be interesting in fact
01:20:46macncheeseim assuming the base wads it can't find when starting up doom are the base wads that came with the original doom, and werent included for copyright reasons?
01:21:17Zoide777how do i do that?
01:21:19 Join ashridah [0] (
01:21:28macncheeseor are free wads still available?
01:21:39Zoide777unknownkind: you're in the wrong forum
01:21:44Zoide777unknownkind: mpd is for ipodlinux
01:21:55Zoide777unknownkind: try asking at #ipodlinux
01:22:00amiconnZoide777: Do you have a develepment environment installed so you can build yourself?
01:23:16macncheesenevermind, figured it out :o
01:24:10amiconnOkay, so open firmware/export/config-ipod4g.h , go to line 97 and remove the // in front of #define HAVE_ADJUSTABLE_CPU_FREQ
01:24:22unknownkindWhats Sorry, you have no base words mean?
01:24:22amiconnThen rebuild and install
01:24:34 Quit |Unknown| (Read error: 110 (Connection timed out))
01:24:49preglowunknownkind: what go perhaps #ipodlinux then whatever help get
01:25:02amiconnIf it doesn't work, this build will probably crash. You'll then need to revert to your previous build
01:26:10 Quit RaeNye ("CGI:IRC (EOF)")
01:26:44amiconnNote: You need to unzip the full build when changing this #define, because it changes the plugin/codec api
01:31:17sharpehmm... cpu emulator works...
01:31:24midkaysharpe, so what use is it?!
01:31:36sharpecorrect basic/kernal/char set loaded..
01:31:38 Quit Farpenoodle ("Large sacks of cheese")
01:32:00 Quit Zoide777 ("CGI:IRC (Ping timeout)")
01:33:22preglowsharpe: porting something?
01:35:04 Quit obo (Read error: 110 (Connection timed out))
01:37:10 Join Zoide777 [0] (
01:37:22 Quit Zoide777 (Client Quit)
01:37:33preglowsure sounds like it
01:37:44preglowwhich automatically qualifies the entire effort as a Good Thing
01:37:50lostlogicheck yeah
01:38:14*amiconn wonders how a c64 emulation could be controlled
01:38:17 Join Zoide777 [0] (
01:38:35lostlogicprobably use it in boot-a-rom mode
01:38:39lostlogicto play joystick based games
01:38:42preglowor boot-a-tape
01:39:09*amiconn wants the boing demo ported
01:39:10lostlogicI only had floppy drives and cartridges on my c64 :-P
01:39:17*preglow wants second reality ported
01:39:21lostlogicI just want qbert and frantic freddie
01:39:25preglowlostlogic: i only had tapes and a cart, heh
01:39:45sharpeyeah... i was thinking like have a rom file to load... then it'd start execution from that...
01:39:58sharpeyeah, tis the c64.
01:40:07preglowsharpe: porting a specific emu?
01:40:14lostlogicsharpe: remember homework is overrated... (that's what you're not doing, right?)
01:40:18preglowsharpe: what?
01:40:28sharpeno emulator at all, actually.
01:40:32preglowlostlogic: not only is homework overrated, good grades are too
01:40:48lostlogicpreglow: that's right, I know well that D is for Diploma (assuming you use the ABCDF scale)
01:40:52sharpetis not homework for the next... few days, until i have to actually start it.
01:40:59Zoide777amiconn: something's wrong. the build hangs at 100% cpu after "gcc -g codepages.c codepage_tables.c -o codepages"
01:41:00preglowlostlogic: that's what we use (most of the time)
01:41:20preglowthe norwegian government has had a hard 10 years deciding what to use
01:41:32Zoide777you people are a terrible influence! leave poor sharpe alone :D
01:41:54amiconnZoide777: Check your system time, time zone, and the codepages.c and codepage_tables.c time stamps
01:41:55lostlogichehe, in the US, some schools use ABCDF, some use ABCDE, some use 54321, and some now decide that there is no such thing as not passing :-\
01:42:13Zoide777amiconn: what about them?
01:42:20sharpehmm... i'm not sure, but... is the kernal supposed to eventually just stop execution at 0x48FF...
01:42:22preglowi've had "Lg, Ng, G, Mg, Sg", "123456", and "ABCDE"
01:42:29amiconnSuch effect can be caused by a source file with a future time stamp, causing make to rebuild again and again
01:42:31sharpeelement names... :-D
01:42:34preglownothing like extreme diversity
01:43:08 Join Zoide777_ [0] (
01:43:09 Quit Zoide777 (Client Quit)
01:43:33sharpeit seems as if it stops after 16 opcodes...
01:43:58sharpeer, 22
01:44:18XavierGrall-right another productive day start for me. :D
01:44:35lostlogicpreglow: Lg = Lotsa-good, Ng = Nice-n-good, G = Good, Mg = Moderately-good and Sg = Slightly-good? :-P
01:44:52preglowlostlogic: haha, at least you got the adjective right
01:44:59preglowlostlogic: it's other way around
01:45:17preglownokså, noe, nothing, meget and særdeles
01:45:18Zoide777_amiconn: i now tried saving the config-ipod4g file with the date from april 1, but it still hangs
01:45:24preglowwhich i suspect makes no sense for you
01:45:31lostlogicbagh, the ABCDE ordering on the last set threw me, punk.
01:45:32preglowplus displays several weird characters
01:45:50sharpewhat should i implement in the c64 emulator? or, as of right now, 6502 emulator?
01:46:04preglowlostlogic: leading people astray makes me feel superior, sue me
01:46:14 Quit WVU2034 (Read error: 104 (Connection reset by peer))
01:46:24amiconnZoide777_: Hmm, did you check codepages.c and codepage_tables.c as well?
01:46:26sharpeideas anyone? :)
01:46:34lostlogicsharpe: joystick input primarily
01:46:46Zoide777_amiconn: what should be the relation between their date and that of config-ipod4g?
01:46:50sharpei mean, other than that...
01:46:58 Quit unknownkind ()
01:47:08amiconnZoide777_: YOu mean if you revert that definition, it works?
01:47:38 Quit macncheese ()
01:47:49Zoide777_amiconn: i haven't tried without making changes. what i mean is: what is it that i should look for when checking codepages.c and codepage_tables.c?
01:48:04amiconnthe file date....
01:48:12Zoide777_well, yes, but what about it
01:48:14amiconnIt should not be in the future
01:48:21Zoide777_relative to when?
01:48:25amiconn...relative to your system time
01:48:26 Nick Zoide777_ is now known as Zoide777 (
01:49:03Zoide777oh! i noticed my time was in am instead of am
01:49:08Zoide777i mean, am instead of pm
01:49:08sharpeoh, wow. okay...
01:50:11sharpejoystick #1 input is at address 0xdc01
01:51:00Zoide777amiconn:now it's building correctly
01:51:39sharpeand i've no idea how color ram works, i guess i get to find that out
01:51:52 Join webguest91 [0] (
01:51:55 Quit aegray (Read error: 113 (No route to host))
01:54:43Zoide777i've been getting this error when building cvs: "cp: connot stat `/home/rockbox-bleeding/apps/tagnavi.config': No such file or directory
01:54:54Zoide777i mean, when doing "make zip"
01:55:06amiconnYou're not buildinmg from a cvs checjout?
01:55:06 Join JdGordon [0] (
01:55:19Zoide777no, from the cvs source file on the website
01:55:37amiconnIt seems there's a newly introduced file missing in the tarballs.
01:55:41XavierGrI am in the middle of abattery benchmark for the H300 and I can say that the results still seems disappointing. The test started at 15:40 (3:40 pm). It still plays but the battery meter reports another 2 hours of juice. That means that it will barely reach 13 hours of playback.
01:56:21XavierGr15:40 GMT -2
01:56:30Zoide777amiconn: what would that affect?
01:57:02amiconnXavierGr: H300 still has that mysterious power problem. Probably only solvable with measuring equipment and some digging in disassemblies of the stock firmware
01:57:12amiconnZoide777: That you can't 'make zip' ...
01:57:23 Quit webguest91 ("CGI:IRC (Ping timeout)")
01:57:27Zoide777well, it does make the zip after that
01:57:37Zoide777and it works on the target (except now that i tried cpu boosting)
01:57:59Zoide777amiconn: The cpu boosting build gives a prefetc abort error on boot
01:58:07 Quit carini (Remote closed the connection)
01:58:28amiconnpreglow: Why would cpu boost cause a prefetch abort?
01:58:31linuxstbYes, I get that sometimes with the CPU boosting.
01:58:37Zoide777amiconn: Prefetch abort at 60005FFE, and the display is shifted to the left by one letter. Also, there is a 3 on the right side of the screen
01:58:54XavierGrah and I forgot to say that this is with a custom 1900mah battery
01:59:21amiconnNon-present tagnavi.congig means you can't use the tagcache
01:59:23preglowamiconn: how?
01:59:53amiconnpreglow: iPod 4g greyscale with boosting enabled
01:59:56preglowahh, 4g
01:59:58preglowi don't know
02:00:00amiconnsee a few lines above
02:00:02Zoide777amiconn: oh, and no wonder i didn't notice the tagcache not working... i only noticed the the tagnavi missing on sim builds where i didn't use tagcache
02:00:04preglowit sounds like the ram is misconfigured
02:00:20preglowthe prefetch abort is a tricky exception to understand
02:00:26preglowsince it depends so much on internals
02:00:39amiconnpreglow: What is at 60005ffe? That looks like some PP register?
02:00:51 Quit [TCK] ("well, if you say so.")
02:00:55Zoide777preglow: could that ram misconfiguration affect things other than cpu boosting?
02:01:10lostlogicXavierGr: backlight use percentage? type of music?
02:02:02preglowamiconn: it is
02:02:12linuxstbamiconn: I've just added tagnavi.config to FILES
02:02:18preglowamiconn: that's why you got the prefetch abort, the pp core tosses one if it's about to read from a port
02:02:22amiconnlinuxstb: thx
02:02:31preglowor any io reg of any kind
02:02:32sharpemmm, pizza.
02:02:46preglowZoide777: sure, but i don't think that's what we're seeing
02:02:50XavierGr128kbps mp3. Left in a drawer, I just checked once the battery level. So the backlight usage is pretty low. (2-3 times of 15 seconds each)
02:03:00preglowZoide777: i think perhaps some internal controller is panicing when the pll is reconfigured
02:03:19amiconnpreglow: I don't think that can be true. Sure you need to read from "I" IO registers...
02:03:21preglowit would be very nice if linuxstb could do some extensive tests with changing around udelay values... :)
02:03:25lostlogicXavierGr: damn.
02:03:28preglowamiconn: yeah, but not code
02:03:41linuxstbpreglow: I wouldn't say I've done extensive tests, but I tried a few different values.
02:03:57amiconnpreglow: So 'prefetch' implies code?
02:04:01preglowamiconn: yes
02:04:10lostlogic"execute my damn IO ports!"
02:04:11preglowamiconn: at least i'm fairly certain
02:04:12XavierGrI am in awe, with the same battery rockbox could output 25 hours on H100
02:04:31Zoide777do you need me to test changing something to see if it no longer gives the prefetch abort?
02:04:55amiconnlinuxstb, preglow: The cpu frequency is programmed once even when boosting is #ifdefed out, correct?
02:05:50lostlogicXavierGr: power drain bugs are great that way :(
02:06:18lostlogicXavierGr: ipod video gets ~6 hours on rockbox in my experience compared to nearly 20 on apple firmware :(
02:06:46preglowlinuxstb: yeah
02:06:49*amiconn can't compare to apple fw at all
02:06:51XavierGr20 on apple firmware?
02:06:56preglowamiconn: yeah
02:07:07lostlogicXavierGr: I've not used the apple firmware, that's just what I've heard...
02:07:21amiconnlinuxstb: A suggestion for an interesting experiment:
02:07:24preglowamiconn: you can see the routine that's used for cpu freq programming when boosting is ifdeffed out in system.c
02:07:28XavierGrwell apple is known for the false battery readings
02:07:34preglowamiconn: seek for the cpu freq define
02:07:43XavierGrI can't really believe that it would achieve 20 hours
02:07:50amiconnInstead if setting the frequency just once, do the following:
02:07:50XavierGrmaybe 16
02:08:26amiconnipod_set_cpu_frequency(); sleep(1); ipod_set_cpu_frequency();
02:09:02amiconnI would expect it to panic too, at least sometimes
02:09:47Zoide777is there a way of shutting down the ipod while it is plugged into the wall without having it reset right back on?
02:09:52preglowi also got some weird errors sometimes when i just started coding cpu freq chaning
02:10:04preglowbut i can't duplicate them now :/
02:10:19preglowbut i do _nothing_ out of the ordinary even now when it works, so godknows
02:11:00linuxstbZoide777: No - it seems to need to be on to charge. At least, that's how the apple firmware works, so Rockbox does the same.
02:13:06linuxstbamiconn: Trying your suggestion causes an immediate *PANIC* Stkov...
02:13:24Zoide777linuxstb: actually, you're right. i had forgotten that even with the apple os, if you try to turn it off while plugged in it goes into the "Charging" screen with the big battery icon
02:13:27amiconnYou got *Stkov* from that??
02:13:44linuxstbamiconn: I'm guessing it's too early to call sleep()
02:13:55amiconnHmm, maybe
02:14:02amiconnTry a simple delay loop instead
02:14:16amiconn..or just leave the sleep() out
02:14:17preglowtoo early?
02:14:18linuxstbDoing that now. We have a udelay() function on the ipods.
02:14:47amiconnYeah, and ugly inl() and outl() thingies...
02:15:03preglowamiconn: i'm going to weed out those i can (which aren't many) :/
02:15:10linuxstbChanging the sleep to udelay and it seems to work fine.
02:15:19preglowalso, i'm not convinced all the new GPIO defines we have got are truly gpio
02:15:31preglowlinuxstb: hmm
02:15:46preglowit almost seems like we're switching back to pll too soon
02:15:49preglowbut you've tested that
02:16:52linuxstbI'll try again now. What would suggest for the frequency change udelay? 4000?
02:17:08kkurbjunlinuxstb, were you having trouble with doom in the latest cvs build?
02:17:15linuxstbOnly on the 5g.
02:17:27kkurbjunwhat problems were you having?
02:17:37preglowlinuxstb: something outrageously high
02:17:40linuxstbThe top 80% of the screen has brown lines all over it.
02:18:05midkaylinuxstb, i had that too..
02:18:06linuxstbpreglow: OK, trying 100000
02:18:13midkaylinuxstb, it's not quite new.
02:18:20amiconnpreglow: It could also be that we're chaging frequency too soon, or we don't disable the pll before switching
02:18:25midkayi had it for the first time almost 24 hours ago..
02:18:26preglowamiconn: we do
02:18:36preglowamiconn: afaik, at least, it's not easy to say with ipl source code
02:18:43kkurbjunlinuxstb, could you try erasing your default.dfg and see if that has any effect?
02:18:48amiconnOn coldfire, the cpu must not be clocked from the pll during a frequency change
02:18:52preglowamiconn: the comments you see in system.c is basically just me guessing from prior knowledge about how pll systems work
02:19:18kkurbjunI don't see why the 5g would be having trouble, it works fine in the sim
02:19:21preglowamiconn: but afaik, we switch to xtal before reclocking
02:19:41amiconnThe sequence is switch cpu to input clock - change frequency - wait for relock - switch back to pll output clock
02:21:09preglowyes, that's just what i think we do
02:21:19preglowwe just udelay for a specific time, though
02:21:23preglowand hope relock is done by then
02:21:43*linuxstb needs to find some files which are not FLAC...
02:21:58Zoide777linuxstb: i have maaany wavpacks...
02:22:04preglowlostlogic: well said
02:23:04linuxstbI think it's working....
02:24:20preglowgod, please
02:24:21preglowlet it be
02:24:50Zoide777would this fix the prefetch abort errors?
02:25:18scottderconver them to ogg
02:25:29scottderoggenc can usually take flac files directly too :)
02:25:39preglowi sure as hell hope this works
02:25:47preglowi can think of no other explanation as to why it doesn't
02:26:04sharpewell, it should take... not much effort to implement the joystick in the emulator..
02:27:14sharpeit seems like all it is, is changing bits on a byte in memory... depending on which directon, or if the button is pressed.
02:30:49 Quit ashridah ("meanwhile, to uni")
02:31:01XavierGris there any possibility that a file in linux is writable but not readable?
02:31:18XavierGr(I mean it doesn't make sense that way)
02:32:03kkurbjunyes, depending on permissions
02:32:23XavierGrbut if you are going to write to file you need to read it first so...
02:32:40lostlogicXavierGr: why? you can just write new data to it without reading it
02:33:26XavierGrso if a I have a test.txt and have only write permission
02:33:39 Join ze__ [0] (
02:33:40XavierGrdoing vi test.txt will output an error rihgt?
02:34:16lostlogicmost likely
02:34:20lostlogicwhy don't you test it?
02:34:49XavierGryeah better test it. I just read a tutorial on file permissions and the idea came to me
02:34:53 Quit midkay ("Leaving")
02:34:54preglowlinuxstb: no further news?
02:36:46linuxstbYes, it seems to be boosting nicely.
02:37:00linuxstbBut is there a bug in Rockbox at the moment that leaves the CPU boosted?
02:37:10linuxstbI have to go into the debug screen and reset the boost count.
02:37:26linuxstbAh, spoke to soon. It's just froze...
02:38:58preglowhow freeze?
02:39:41linuxstbIt just froze on the WPS, with the backlight still on.
02:40:52amiconnThere's a bug in rockbox that *sometimes* leaves the cpu boosted at boot
02:41:06amiconnI bet it's the tagcache code
02:41:19linuxstbAnd it's just frozen again...
02:41:48linuxstbBut before the freeze, it was happilly playing wavpack and boosting at around 40%, with lots of frequency changes.
02:41:59amiconnMaybe we should disable interrupts during a freq change?
02:42:38preglowamiconn: ahhh
02:42:39preglowamiconn: good idea
02:42:48linuxstbBut the FIQ is still going...
02:42:53preglowlinuxstb: disable that too!!!
02:43:02preglowbut no, there's no reason we should
02:43:09preglowthe clock is locked to 24mhz
02:43:13preglowall should be well
02:43:21preglowsomebody bloody give portalplayer a ring before i implode
02:44:08lostlogichear hear
02:45:01 Quit ze (Read error: 110 (Connection timed out))
02:45:01 Nick ze__ is now known as ze (
02:46:22linuxstbkkurbjun: I tried setting the sound frequency to 44.1KHz in Doom earlier today, and it seemed to work fine on my 4g ipod.
02:46:56lostlogicman, I am just dreading trying to wade through the file buffer filling code to figure out why with my new 'corrected' swcodec stuff it doesn't play on resume (starting a new song, and seeking within a song work)
02:47:47preglowlostlogic: whisky
02:47:51preglowlostlogic: that'll do the trick
02:48:23lostlogicyou make a valid point
02:48:28lostlogic*adds whisky to his beverage
02:48:36linuxstbdisabling the interrupts didn't help...
02:48:42preglowi didn't expect it would
02:49:12linuxstbAh, and now my battery is dead...
02:50:37 Quit imphasing (Read error: 104 (Connection reset by peer))
02:50:46preglowlostlogic: at any rate you surely need to fearless which probably calls for some strong drink when we're talking about that particular portion of code
02:50:52preglowinsert "be"
02:51:12preglowi always read my sentences after having pressed "enter", before? never!
02:51:22 Join imphasing [0] (
02:51:32amiconnWah! My H100 remote does indeed tick when connected to the H300
02:51:47preglowamiconn: isn't there some hardware mod for that problem?
02:51:51lostlogichehe :) feeling better already... I'm really tempted to just commit in it's _very_ broken state just so everyone else can feel my pain while I work on it :)
02:52:06amiconnXavierGr: see, you're right. It depends both on remote & main unit...
02:52:15XavierGramiconn: see I told you it may tick
02:52:48Zoide777lostlogic: commit! commit! :D
02:53:00XavierGrstrangely enough my original remote (as I said earlier) doesn't tick on my H300 (while it ticks on the H100) or it ticks so low that I can't hear it.
02:53:09preglowmy h100 remote doesn't tick the least bit
02:53:36lostlogicZoide777: you are asking for a world of pain for you and everyone else running rockbox on swcodec
02:53:54amiconnIt's a faint ticking here, and it's independent of the volume setting
02:54:12XavierGramiconn try this:
02:54:53XavierGrgo to filetree and move the navi button to the right/lef and hold it.
02:55:23XavierGrit should scroll files until it will reach the end of the list and the ticking will continue to be heard. (as long as you push the button)
02:55:31amiconnWell, it's the remote lcd update
02:56:11amiconnIf you put the cursor on a long line in the file browser so that it starts scrolling, and you've set a not too low scroll speed, it will make a hiss
02:56:16preglowamiconn: tried slasheri's anti-ticking?
02:56:23amiconnWith low scroll speed it ticks
02:56:35amiconnpreglow: Need to make a build with ticking fix enabled
02:57:35preglowamiconn: ahh, right, it's not enabled by default?
02:57:43amiconnNot for H300
02:58:19amiconnMy H140 doesn't tick with the very same remote
02:59:28XavierGr#define HAVE_REMOTE_LCD_TICKING on config-h300.h
03:00:56preglowthen why did i just bloody translate a settings string for it?
03:01:36XavierGrwell hooray. then amiconn can make some improvements on the remote now that he can hear it
03:06:56preglowlinuxstb: ok, so we'll conclude the udelay hack was a flaming wreck, then?
03:12:28linuxstbI'll try again tomorrow... It wasn't a total disaster.
03:15:15preglowbodes well, at least
03:15:20preglowwhat with the crashing
03:16:22***Saving seen data "./dancer.seen"
03:24:21XavierGrhmm copy paste works on vmware only if the user run vmware-toolbx& and the program is running
03:24:40XavierGrthis is something that the manual or FAQs I've read don't mention it.
03:25:18XavierGrI wasted hours and hours trying to figure why it won't work only to discover it by chance.
03:25:26*XavierGr goes to update the wiki
03:25:45 Quit Strath (Read error: 104 (Connection reset by peer))
03:25:53 Join Strath [0] (
03:27:01sharpewow, that may be what's wrong...
03:27:43amiconnXavierGr: 'reduce ticking' does what it says, it reduces the ticking
03:27:52amiconnIt's still there, but really faint
03:28:12sharpeanyone else think something may be wrong if the cpu emulator jumps to 0xE2FC instead of FCE2? :D
03:28:12amiconnStock firmware does also tick
03:28:46XavierGrstrange: you just added the define in h300-config?
03:29:05XavierGryes stock firmware ticks too, it isn;t a rockbox issue.
03:29:53lostlogicI might hafta start hurting myself soon, the buffering logic is beginning to make sense :-\
03:30:03preglowamiconn: does it seem slower in any way?
03:30:10preglowamiconn: rockbox remote navigation, that is
03:30:28preglowahh, i remember people complaining about that
03:30:29amiconnThe ticking fix causes an overall slowdown
03:31:04XavierGramiconn: if the only thing you did was add the define then I did something seriously wrong when my cousin asked my a modded build with it enabled
03:31:27amiconnE.g. when you scroll a long list, more cpu cycles are used to redraw the remote lcd
03:31:36preglowahh, yeah, but do you notice it?
03:31:43amiconnThis makes it more likely for playback to skip....
03:31:57amiconnpreglow: Not during navigation, no
03:32:23amiconnXavierGr: You need to reset settings. The settings block changes with the ticking fix enabled
03:32:51XavierGrany ideas on this to make it better? What about those other alternatives you were saying?
03:33:03amiconnpreglow: I want to get the ticking fix to work better, while drawing less cpu power
03:33:29amiconnWill still be more than with the fix disabled though
03:34:20amiconnMy extra button filtering for H300 seems to work well
03:34:52preglowamiconn: please don't invest too much time into that :)
03:36:17amiconnIt annoys me when things don't to what they should.
03:36:44amiconn..and we need to get things ready for release
03:37:10preglowahh, sure, but few people uses the remote anyway. it would be better if you had a look at the playback engine ;)
03:37:41 Quit damaki (Read error: 110 (Connection timed out))
03:38:42 Join aegray [0] (
03:39:54preglowis bluechip drunk?
03:41:09 Quit Daishi ("Client exiting...")
03:41:37 Part nave7693
03:42:48scottderWhat about Optifrog support?!?!?! :)
03:42:50scottderJust kidding
03:43:43preglow_OPTIFROG_ ?
03:43:49preglowyou're drunk too
03:44:07XavierGrBluechip is surely drunk
03:44:48XavierGr"Honour" is British word. <−−- What's that supposed to mean?
03:45:05goffaheh... it means it has an extra u
03:46:18scottderLimeys need to learn English ;)
03:46:28goffahe he he
03:46:35 Join midkay [0] (n=midkay@rockbox/developer/midkay)
03:47:24scottdernext paycheck I may throw a few bucks in the Paypal tip jar for Rockbox :)
03:47:33preglowyou don't start fucking reciting custom made poetry on a mailing list unless you're smack high on crack
03:47:43preglowscottder: :-P
03:47:54preglowscottder: would make us all a happy bunch that would
03:48:11scottderNonono....if yer gonna stick yer tongue out
03:48:15goffacustom made poetry? who, what, where?
03:48:49scottderI think preglow is drunk
03:49:02lostlogicI'm working on getting that way :-P
03:49:04preglowgoffa: on the dev-list
03:49:07goffai put a measly $50 in the tip jar.. and then i made an offer that aparently they could all refuse
03:49:10preglowscottder: i'm just slightly drunk...
03:49:49lostlogicbagh, issues with track skipping not always skipping, but initial seek works.
03:50:14XavierGrlostlogic: about the 'pops' 'clicks:
03:50:18preglowscottder: still, don't keep my alcoholic tendencies from making you donate
03:50:36goffayeah.. tip money=beer money :P
03:51:00preglowi wish i had direct access to the donation money
03:51:07lostlogicgoffa: there was i ndeed a lot of free beer at devcon2006... ::shiftyeyes::
03:51:18XavierGrit seems to me that they occur when I press /right/left to change track (though not always) while selecting another track from the filetree is nearlt glitchless.
03:51:41lostlogicXavierGr: if I can get track skipping to NOT SUCK I think my next commit will help with such things
03:51:51XavierGrok nice :)
03:51:55preglowlostlogic: and the very best of luck to you
03:52:11sharpewell, i figured out what was wrong
03:52:16lostlogicpreglow: cheers, good lookin' out on the whisky => better coding mindset thought :)
03:53:01goffalol... my mom is telling my sister not to smuggle cigarettes across state lines
03:53:41goffathey are cheaper in north dakota... and montana has a law that cigarettes need to have a tax stamp
03:53:47preglowlostlogic: ahahaha, i gave up on coding four hours ago
03:54:01preglowand you'll all benefit from my doing so
03:54:14lostlogicpreglow: it's only 20:53 here, so I've still got some code left in me :)
03:55:17lostlogicdamn, buffer filling causes skips still, but I think that's going to be a case of "I can commit with that problem"
03:55:28preglowpwned by gammel dansk
03:56:30Zoide777amiconn: how's grayscale lib going?
03:56:31scottderpreglow: Paypal won't let me pay in bottles of jack Daniels...sorry
03:56:39scottdermaybe Mad Dog 2020
03:56:42linuxstbscottder: You'll have to persuade the author to open source Optimfrog first...
03:56:57goffahe he he... wonder what the exchange is $$ to jack
03:56:58scottderlinuxstb: hehe I was just kidding :)
03:57:23preglowscottder: i've got plenty of alternate payment methods handy!
03:57:49lostlogicpreglow: glasses?
03:58:13 Quit harbel ()
03:59:01preglowlostlogic: no, i've got good eyesight
03:59:12preglowlostlogic: what were you asking, really? :)
03:59:30scottderpreglow: could give a whole new meaning to "feature bounties"
04:00:12 Quit Fitzsimmons (Remote closed the connection)
04:00:34lostlogiclostlogic: alternate methos of receiving payment in whisky
04:00:49preglowscottder: a new revolution in open source, this is
04:01:01scottderlostlogic: talking to yourself is a sign you've had too much already
04:01:03preglowlostlogic: you just passed drunk hurdle level 1
04:02:50preglowoh, but we're really a decent bunch of people
04:02:58preglowthat's why you should donate
04:03:35 Quit Zoide777 ("CGI:IRC (EOF)")
04:03:44lostlogichmm... file reading is a synchronous operation? no DMA? :(
04:03:47lostlogicthis saddens me
04:04:20preglowsomeone (probably amiconn) was planning dma
04:04:22 Join Zoide777 [0] (
04:05:25scottderI've been playing around with vorbis....encoded an album at Q3, still sounds pretty good to me
04:05:52lostlogic*nod* if file reading was non-blocking, the problems with CPU resource collision between codec and buffer threads would go byebye
04:06:11lostlogicnot that this helps me solve my "skipping changes tracks, maybe" problem.
04:06:24preglowscottder: q3 should be just fine
04:06:34preglowscottder: vorbis is actually faster than mp3 on ipod
04:07:46preglowlostlogic: what's the time?
04:08:27scottderI've been pretty happy with battery long as i make it through my shift, it's all good :)
04:08:35preglowscottder: haha
04:08:49preglowscottder: as in ogg will give you more out of your battery than mp3 now
04:09:29lostlogicpreglow: 21:08
04:09:49scottderSweet :)
04:10:06scottderespecially at q3 :)
04:11:07preglowapproximately nine hours behind me, then
04:11:35*preglow rests his case
04:12:00Galoisscottder: if you haven't already, you might want to try the aotuv vorbis encoder. It has better sound quality than the standard encoder, especially for quality between -2 and 2.
04:12:07sharpeahh!! swapping bytes in a word...
04:12:13Galoiseven at q3, there should be some slight improvement
04:12:46preglowsharpe: what aout it?
04:12:57lostlogicthere is an aotuv useflag for the libvorbis package in Gentoo Linux (courtesy of yours truly) ;)
04:12:59scottderGalois: that's what I use :)
04:13:20sharpepreglow: it doesn't like me.
04:13:22Galoisin the sept. 2005 guruboolez test, aotuv at q3 beat lame mp3 at 128k for rock/pop
04:13:51preglowsharpe: rockbox has macros for doing it...
04:14:08sharpeoh my god.
04:14:13preglowoh your gox
04:14:14lostlogicGalois: the code as of sep 2005 should be in the libvorbis mainstream at this point, I believe...
04:14:21lostlogiccan't remember exactly when they last merged
04:14:26sharpeyeah, my god. not yours. hah! get your own! :D
04:14:38sharpepreglow, what would those macros be?
04:14:56Galoisthey merged aotuv beta2 last year, and haven't synced since. In the last irc meeting there was talk of syncing up again.
04:15:10Galoisthat meeting was only a couple of weeks ago, I believe
04:15:18lostlogicGalois: ah, my mistake.
04:16:09preglowsharpe: swap32 and wsap16
04:16:47sharpepreglow: thank you.
04:16:50Galoisoops, five weeks ago
04:18:25pussfellerwhats this data abort at 0001642c
04:18:40pussfelleris that a ref to a place in memory thats usefull at all
04:19:03 Join Doomed [0] (
04:19:22Doomedthere is still a problem with track numbering in tagcache
04:19:29preglowsharpe: oh, you're welcome
04:20:05Doomedjust letting u guys no
04:20:22preglowDoomed: we know
04:20:30scottderOggEnc v2.81 (Lancer 20060302(SSE) based on aoTuV
04:20:53Doomedidk if it was attempted to be solved since last i was in here
04:20:54Galoishey lancer is not too bad
04:20:55pussfellerwell tagcache crashes on me when I scroll up and down real fast
04:21:01GaloisI still use ordinary aotuv, though
04:21:09scottderI rip to flac then transcode with that to ogg
04:21:13scottderVERY fast
04:21:19scottder23x for me most of the time
04:21:45Galoisaudio encoding benefits a lot from 64-bit
04:21:52scottder32bit AMD 3000+
04:23:50scottderDoes the job for me
04:24:01 Join warthawg [0] (
04:24:05scottderI am sure I'll go 64bit next time I upgrade my windows box
04:24:11GaloisI'd use the 20060331 version if I were you
04:24:48warthawgif i download a newer version of rockbox than i already have installed on my nano, do i need to go thru the entire install again, or can i just copy the .rockbox directory over?
04:25:08Doomedjust unzip
04:25:09Galoiscopy both the .rockbox dir and the rockbox.ipod file
04:25:17Doomedthe zip to ur nano
04:25:19warthawgtks, guys
04:25:47scottderGalois: was that to me?
04:26:17Galoisscottder: yeah
04:26:27sharpehow does one get 0x6c60 out of 0xfce2...
04:26:55scottderAHh didn't notice there was a new version out
04:26:56scottderthanks :)
04:29:34sharpeokay, very strange...
04:31:29sharpei wonder if this works.
04:34:56preglowi need to sleep
04:35:20sharpethen you should probably sleep.
04:35:55sharpesomething still doesn't like me :)
04:36:57sharpe0xECF2, should be 0xF2EC... however, it apparantly is now 0x6C60
04:37:07preglowrockbox code has had years to learn how to be hostile
04:37:16preglowand new code learns quick
04:37:34sharpeand it wasn't even rockbox's that did it first, when i tried, i got similar results.
04:37:53preglowlike i said
04:37:54preglowit learns fast
04:38:06sharpepolymorphic code!
04:38:20preglowwelcome to the wonderful world of embedded coding
04:39:19scottderpreglow: code darwinism? :)
04:40:28Doomedwhat the hell, its not showing me my db
04:40:33sharpehmm, i wonder if this will work.
04:41:28sharpedoubt it, but still...
04:42:50sharpewell, back to where i started, except instead of a void pointer, now it's a char pointer.
04:45:04XavierGr'+' = addition then '-' = ?
04:45:07preglowsolid progress
04:45:11XavierGrI don't remember the word
04:45:29XavierGrok 10x
04:45:50scottderwerid. that 20060331 version con't convert flac on the command line :(
04:46:24Doomedi dont even get the tagcache menu
04:50:17sharpestill no success, oh well, sleep time.
04:52:24Zoide777Doomed: did you reboot after forcing a tagcache update?
04:52:40Zoide777(even after it finishes the forced update, you have to reboot so that it will "commit" the new db)
04:53:46Doomedyeah i just unzipped the same cvs
04:53:46Zoide777(or at least that's how it works for me)
04:54:29Zoide777hmm.... i don't know what it could be
04:55:11Zoide777maybe one of the other 94 people in the channel can help you? (though I'd say 90% of them never say anything... very creepy)
04:56:08 Quit hardeep ("BitchX: its not for the tabkey impaired!")
04:56:28midkayDoomed, what's up?
04:56:40Doomedi fixed it
04:57:36Doomedi think i did
05:00:38midkaygood _enough_..
05:00:51scottderweird I grabbed the latest lancer version dbpoweramp is showing it as an version from 2003
05:01:16GaloisI trust ogginfo on such matters
05:01:16scottdertool = Xiph.Org libVorbis I 20030909
05:01:38 Join nave7693 [0] (
05:02:11nave7693hi rockbox is not making for me
05:03:27nave7693gcc uclpack.o src/libucl.a -o ../uclpack
05:03:27nave7693src/libucl.a: could not read symbols: Archive has no index; run ranlib to add one
05:03:27nave7693sh-elf-ranlib is called on libucl so I am wondering what's up
05:07:25nave7693anyone alive?
05:09:22 Join scf_ [0] (
05:10:55 Join FalloutMan [0] (
05:11:12FalloutManhey guys
05:11:42FalloutMandoes anyone know how to build a playlist on an ipod?
05:12:36midkayFalloutMan, a custom one, or one of a directory? or how to create one, or what?
05:13:01FalloutMani dunno, i just need guidence from step A onward
05:13:41midkayso you want to create a playlist of tracks you select or what?
05:14:50midkaylocate the first song you want, hold SELECT on it and choose "playlist" -> "insert". it will begin playing. press select while it's playing to get back to the browser, keep locating the songs you want, holding select and "insert"ing until it's got the tracks you want... then go to menu - playlist options - save playlist (gotta double check that last bit)
05:15:03midkayyou can reorder tracks from the playlist viewer also..
05:15:30FalloutManthanks a bunch
05:15:36midkay"save current playlist", yeah.
05:15:56lostlogicdon't we have a user's manual that has stuff like this in it?
05:16:15FalloutManthe manual is so confusing
05:16:26***Saving seen data "./dancer.seen"
05:16:54midkaylostlogic, we have way too many :)
05:17:04midkaya wiki manual, a 2.4 manual, a latex manual..
05:17:28lostlogicwell the latex manual is the most up-to-date, yes?
05:17:35midkaylostlogic, right..
05:17:40midkaybut also the hardest to access, i think.
05:17:50lostlogicah, someone was talkinga bout doing weekly pdf builds ofit?
05:17:56midkayyou have to compile it yourself, i think.. so most people who need it can't really get to it.
05:18:02midkayso they don't know where to look..
05:18:04XavierGrH300 stopped at 6:16am
05:18:09XavierGrfrom 15:40
05:18:19midkayXavierGr, doing/playing what?
05:18:26 Quit Rob2222 (Read error: 113 (No route to host))
05:18:33XavierGr128kbps with 1900mah battery on repeat
05:19:40 Quit DrumRBoy320|Away ()
05:20:29XavierGrbattery_bench didn't report anything
05:20:34XavierGrnot a single measurement
05:21:18XavierGrat least I managed to time it myself
05:24:01midkayXavierGr, well, that's quite impressive, i think. :)
05:24:23 Join stevenm [0] (
05:24:28XavierGrI don't think so
05:24:34XavierGrthe battery is 1900mah
05:24:35stevenmHey all
05:24:42stevenmWhat's the whole deal with the gl thing?
05:24:47XavierGrH100 managed 25 hours on rockbox
05:24:58XavierGrwhile H300 just managed 14:40
05:25:09midkayXavierGr, color screen...
05:25:20goffaXavierGr: what coodec... and did you mess with it while it was going?
05:25:23midkaythough it shouldn't make that huge a difference, 15 hours is definitely nothing to complain about..
05:25:38XavierGrthe screen was not activated (except 2-3 times) during benchmark
05:25:42DoomedH100 needs a better greyscale for doom
05:26:05midkayDoomed, as though it should be running doom at all..
05:26:10XavierGrmidkay: again, the test is with a 1900mah bettery
05:26:16XavierGrnot with the default
05:26:19Doomedi no, im just saying
05:26:27midkayXavierGr, my archos recorder couldn't pull nearly that, i think, with those batteries.
05:26:36 Quit unexterminatable (Read error: 113 (No route to host))
05:26:36midkayand its LCD is considerably smaller, it's much slower..
05:26:40midkaybut the buffer is a killer, 2mb..
05:26:49XavierGryes but iriver firmware can make it play more.
05:26:58goffaall i have to say about iaudio x5l is 35 hours .. my ass
05:27:00XavierGrso something is wrong as Linus has said
05:27:03goffaits close to 20 or so
05:27:06 Quit imphasing ("Woo!")
05:27:13midkayXavierGr, how much more?
05:27:16goffabut seems to be the same with rockbox or iaudio firmware
05:27:39XavierGrmidkay: at least 14-15 hours with a normal battery
05:27:54goffahard to do an exact comparrison though... because i have a bunch of mpcs on the player now
05:27:56midkayhow is that more? looks the same-or-slightly-less.
05:27:57XavierGrbut my next test will be on default firmware
05:27:57 Join imphasing [0] (
05:28:06midkayoh. never mind.
05:28:12midkayi'd like to hear your results..
05:28:21goffaand iaudio firmware does't support mpc
05:28:30goffabut i suspect mpc being in between mp3 and ogg
05:28:34XavierGrI will keep them updated along with the other on IriverRuntime in the wiki
05:28:42midkayXavierGr, cool.
05:28:56nave7693midkay: how long does your archos last in a full charge
05:29:16midkaynave7693, god, it's been too long, and my batteries are too old to say..
05:29:27midkaywhen they were new, 2050mah i think.. probably 11-12 hours..
05:29:43 Quit scf_ (Read error: 113 (No route to host))
05:29:45midkaymaybe less, really can't remember..
05:29:57midkaysounds about right though.. or what should be expected.
05:30:05nave7693mine goes 3~4 with brand new 2500 energizers
05:30:06 Join Rob2222 [0] (
05:30:11midkaynave7693, wtf?
05:30:16nave7693yea wtf.
05:30:25XavierGrRob are you here?
05:30:28midkaywell, i haven't used it lately. it actually seems to be way _slower_ than it used to be..
05:30:29 Quit imphasing (Client Quit)
05:30:38midkaybefore all the iriver/iPod/iAudio ports started..
05:30:46midkayloading dirs, buffering, navigation..
05:31:08 Join imphasing [0] (
05:31:29goffathat's scary...because rockbox is lightyears faster than the original iaudio firmware
05:31:35nave7693newer versions just seem to last shorter and shorter...
05:32:04scottderKnow whats great
05:32:11 Join Arrogant [0] (
05:32:15scottderfinding a co-workers embarassing MySpace page :)
05:33:06goffalol... you should have preluded that with new keyboard $20... internet connection $70... finding your coworker's embarassing page on myspace... priceless
05:33:22 Join corevette [0] (
05:33:32goffafollowed by the url :)
05:34:04 Join webguest18 [0] (
05:35:08nave7693sigh seems like it's time for me to ditch my trusty archos and go with a newer player
05:35:37Doomedi like my H120
05:35:37 Quit corevette (Client Quit)
05:36:25midkaynave7693, agreed..
05:36:31 Quit Galois (Remote closed the connection)
05:36:32goffai'm tempted to get a second 340 so i can pack around more tunes
05:36:44midkaynave7693, do you agree that it seems to browse, load and generally _go_ slower than it used to?
05:37:09 Join Galois [0] (
05:37:41 Join imphasin1 [0] (
05:38:15 Quit imphasing ("Oi.")
05:38:25 Join [C]Boomstick [0] (
05:38:32[C]Boomstickhello :)
05:38:58 Quit linuxstb (
05:38:58 Quit kkurbjun (
05:38:58 Quit pussfeller (
05:38:58 Quit xmixahlx (
05:38:58 Quit novimon_ (
05:38:58 Quit Bg3r (
05:38:58 Quit luckz (
05:38:58 Quit dwihno (
05:38:58 Quit CoCoLUS (
05:38:58 Quit crash (
05:38:58 Quit vHs (
05:38:58 Quit dark (
05:38:58 Quit steveb (
05:40:34NJoinCoCoLUS [0] (
05:40:40nave7693midkay: that's true, but maybe it's because I am using unicode fonts
05:40:47 Join crash__ [0] (
05:40:50 Join dark [0] (
05:40:52***Server message 505: 'logbot- :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
05:41:11 Part FalloutMan
05:41:21nave7693midkay: it has to be slower because of the added features
05:41:21midkaynave7693, it's possible..
05:41:22 Join novimon [0] (
05:41:28NJoindwihno [0] (n=dw@
05:41:33 Join steveb [0] (
05:41:33midkaynave7693, which? there isn't anything majorly new.
05:41:34***Server message 505: 'logbot- :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
05:41:41NJoinvHs [0] (
05:42:57nave7693midkay: there's support for other players now, when rockbox becomes not specific to archos it adds overhead
05:43:53midkaynave7693, stuff that doesn't apply to the archos is, for the most part, not compiled for the archos.
05:43:58midkaymost of it the archos can't even handle.
05:44:18nave7693as long as there's shared code... I guess
05:44:26midkaythat's not really how it works.
05:44:46nave7693some things can be done differetly that get generalized
05:44:56midkayi'm not sure if you're familiar with rockbox's code, but the compiler simply doesn't compile stuff that doesn't pertain, as long as it's told not to, and it most often is.
05:45:14midkaymost often is told not to, that is..
05:45:24XavierGrah shit. I know why the benchmark didn't work
05:45:27 Join imphasing [0] (
05:45:30 Quit imphasin1 ("Oi.")
05:45:43midkayXavierGr, why?
05:45:43nave7693and somebody please fix the build lol
05:45:44XavierGrI got a version of the plugin with a bug that relates ti the new tagcache
05:46:01XavierGrthis bug was fixed today but I forgot to update
05:46:05 Join warthawg_ [0] (
05:46:10midkaynave7693, 'fix the build'?
05:46:26warthawg_where is the games folder?
05:46:33 Quit warthawg ("my work here is done")
05:46:54midkaywarthawg_, it doesn't exist unless you make it, if you're talking about for doom.
05:47:08 Join corevette [0] (
05:47:10warthawg_yes, i am.
05:47:18midkaywarthawg_, where are you reading about it?
05:47:22nave7693midkay: I can't build the CVS for archos
05:47:25warthawg_in the forum
05:47:45nave7693midkay: actually its uclpack
05:48:00midkaynave7693, rockbox seems to be able to compile itself just fine..
05:48:08midkaywarthawg_, you need to create it yourself. that's not the best source of info anyways..
05:48:21nave7693midkay: can you update and try make veryclean and make?
05:48:32warthawg_tks, midkay
05:48:38midkaynave7693, alright.
05:48:42nave7693oh make sure you run configure to update the makefile
05:48:48warthawg_just create /games/doom/addons and drop the wad(s) in there?
05:48:58midkayif it's an 'addon', yes..
05:49:12midkaynave7693, standard recorder, right?
05:49:18nave7693midkay: yes
05:49:24warthawg_i guess i don't know if it is or not
05:49:31midkaynave7693, works for me.
05:49:36midkayor seems to..
05:49:56midkayyep, worked.
05:49:57NJoinlinuxstb [0] (
05:50:01NJoinxmixahlx [0] (n=xmixahlx@
05:50:07nave7693midkay: does it compile uclpack fine?
05:50:25nave7693what ar and ranlib does it call on libucl.a?
05:50:26midkaynave7693, it was probably already compiled from before. but there were no errors..
05:50:48nave7693make veryclean should clean everything so you recompile all tools
05:50:48NJoinpussfeller [0] (
05:51:16NJoinkkurbjun [0] (
05:51:36midkaysh-elf-ranlib libucl.a
05:52:11nave7693what does it use to link
05:52:19nave7693to produce uclpack?
05:52:57nave7693can you type "which gcc"?
05:53:33midkayit's 3.3.6
05:53:40midkayfor this..
05:53:47nave7693so it's the sh1 gcc
05:53:59midkayyes, that's what you need for recorder building..
05:54:13nave7693midkay: in my machine it is sh-elf-gcc
05:54:16 Quit [C]Boomstick ("Knowing and not doing are equal to not knowing at all.")
05:54:22midkaynave7693, well, mine too..
05:54:31nave7693which the makefile should really use
05:54:32midkay'sh1 gcc' == 'sh-elf-gcc', no?
05:54:40midkaysame for me.
05:54:45nave7693but "gcc" != "sh-elf-gcc"
05:54:53nave7693gcc is the host's compiler
05:54:57midkayUsing sh-elf-gcc 3.3.6 (303)
05:55:05nave7693says on mine too...
05:55:07midkaynave7693, ... right....?
05:55:22midkaywhy did you ask about gcc if you wanted to know about sh-elf-gcc?
05:55:47midkaythis is the vmware image..
05:56:06midkaybe back in ~5 minutes.
05:56:10nave7693because the makefile is calling gcc
05:56:20nave7693which it sould really call sh-elf-gcc
05:58:32 Quit corevette ("CGI:IRC (EOF)")
06:01:52midkaynave7693, not for me..
06:02:01webguest18how do i get to the menu where u can select which files to show on a g5?
06:02:14midkayhold menu, or menu - general settings - file view - show files.
06:02:56nave7693midkay: maybe your gcc is linked sh-elf-gcc
06:03:00nave7693mine is not
06:03:05midkaynave7693, probably.
06:03:11midkayi've never had a problem. is this recent or what?
06:03:29webguest18how do i get it to show all files?
06:03:37midkaywebguest18, from where?
06:03:47midkaypress play (downwards) from the quick screen..
06:05:13lostlogicwell I'm going to do my famous late evening playback breakage and go-to-sleep commit act.
06:06:40nave7693man this is annoying.
06:07:08nave7693some of the tools are meant to be built with the host's gcc and others built by sh-elf-gcc
06:07:24nave7693otherwise they can't be executed by the host
06:07:33lostlogicnave7693: and the build system should detect and use the right ones, if you have sh-elf-gcc and host-gcc in your path
06:07:51nave7693lostlogic: uclpack is not using the right one
06:08:03lostlogicswcodec people, please test, and email, /msg, or flyspray me the new bugs.
06:08:12lostlogicnave7693: have you re-run configure recently?
06:08:28nave7693lostlogic: I rerun it on every cvs update
06:08:31 Quit midkay ("Leaving")
06:08:57lostlogicnave7693: that is quite odd... what host and target are you building?
06:09:46nave7693host is x86 windows/cygwin, target is sh-elf
06:09:56lostlogicI mean what player specifically
06:10:04nave7693archos recorder 2MB
06:10:46nave7693the problem apparently is not in rockbox itself but the uclpack tool
06:10:57lostlogiccan you paste the results of gcc -dumpversion and sh-elf-gcc -dumpversion here?
06:11:13nave7693it is getting linked by the host's gcc which does not know about the results by sh-elf-ranlib.
06:11:17nave7693sure just a sec
06:11:59lostlogicalso ranlib -v and sh-elf-ranlib -v (jus tthe first line)
06:12:22nave76933.4.4 for host and 3.3.6 for sh-elf-gcc
06:12:39nave7693GNU ranlib 2.16.91 20050610
06:12:51nave7693GNU ranlib 2.16.1
06:12:57nave7693first one is host
06:13:16lostlogicdamn, I was hoping to see an old ranlib on one of those to point a finger at :(
06:13:31nave7693you don't find that with redhat's stuff :P
06:13:56lostlogicon my build server (which is one of the CVS Build machines) I have 3.4.5, 3.4.5 and 2.16.1, 2.16.1, but I can't imagine that'd matter.
06:15:26lostlogicw00t! look at that green build table, I was thinking I might be the one to break that streak with this latest commit ;-P
06:16:04nave7693does the build server do a complete build every time?
06:16:29lostlogicyeah, and I just ran clean builds of 3 targets separately from the build system, no problems.
06:16:39lostlogicbut these are all linux machines, not cygwin\
06:17:07nave7693can you see what gcc is getting called to link uclpack?
06:17:14nave7693I mean which gcc
06:19:19lostlogiclemme see
06:20:47 Join RoC_MM [0] (
06:21:40lostlogicthe command run for uclpack is "gcc" whichis hostcc
06:22:49 Quit warthawg_ ("my work here is done")
06:23:28nave7693so your gcc knows about sh-elf libs...
06:23:59lostlogichmph. I guess so.
06:24:13lostlogicbut I think it's more likely a binutils issue than GCC
06:24:30lostlogicanywho, I must sleep, nothing seems to be burning down as a result of my latest commit, so it's a good time for it.
06:24:39nave7693ok see ya
06:24:44lostlogicgood luck :)
06:27:20 Join hcjake [0] (
06:27:21 Quit hypercooljake (Read error: 104 (Connection reset by peer))
06:28:04 Nick hcjake is now known as hypercooljake (
06:28:13 Part hypercooljake ("Leaving")
06:37:45 Join afruff23 [0] (
06:38:42afruff23in the WPS, does the X5 support the hold switch? Battery indicator? Battery charging?
06:40:53afruff23more imprtantly what feature should all WPSes include?
06:41:13afruff23play status indicator, volume indicator, battery indicator, progress bar, artist, title
06:41:24afruff23repeat and shuffle status
06:41:28afruff23what else?
06:41:47 Quit webguest18 ("CGI:IRC (EOF)")
06:41:47 Quit Zoide777 ("CGI:IRC (EOF)")
06:42:44 Quit afruff23 ("Did anyone see my lost carrier?")
06:45:33 Quit nave7693 (Read error: 110 (Connection timed out))
06:57:30 Join luckz [0] (i=alcyon@
06:57:44JdGordonWOW... this is interesting...
06:58:26dpassen1JdGordon: very interesting
06:58:38JdGordonhave appl lost the plot??
06:59:23JdGordonhaha, reading the site.. 2 snaps @ windows...
06:59:30JdGordonWindows running on a Mac is like Windows running on a PC. That means itll be subject to the same attacks that plague the Windows world. So be sure to keep it updated with the latest Microsoft Windows security fixes.
06:59:58dpassen1They seem to be getting closer and closer to a hardware company
07:03:33 Quit RotAtoR ("zzzzzz")
07:03:33 Join JBGood [0] (
07:05:01 Join B4gder [0] (
07:06:17 Quit BHSPitMonkey (Read error: 104 (Connection reset by peer))
07:07:08BHSPitLappyoh dag
07:16:24RoC_MMgood point JdGordon
07:16:29***Saving seen data "./dancer.seen"
07:16:58*JdGordon would definatly consider an apple laptop now tho.. unless the price difference is rediculous..
07:16:58*B4gder almost responded to the trolls but managed to resist
07:17:09JdGordonwhich trolls in particular?
07:17:16B4gderthe dev list trolls
07:17:27JdGordonmove the list to the forums and u can deal with this crap..
07:17:41 Quit jbauman (Read error: 110 (Connection timed out))
07:18:04B4gderI prefer mailing lists to any forum
07:18:16Galoispfft, weak. You should get an apple laptop and run linux on it.
07:19:22JdGordonosx is linux..
07:19:37B4gdera big almost
07:19:41B4gderits *nix though
07:19:42Galoisyeah except for the "free" part
07:20:02JdGordonu only pay for the gui.... the backend is all DSB
07:20:14B4gderno, but BSD-like
07:20:28dpassen1Its a modern *nix desktop, nothing wrong with that.
07:20:49Galoisthe typical slashdot flamewar would then continue by pointing out lack of basic features such as virtual desktops...
07:22:13B4gderbut slashdot is very apple friendly
07:22:20Mikachuand then someone would point out there is a shareware program that lets you have virtual desktops, but it costs like $50
07:22:56Galoisthere's a couple of free virtual desktop programs that by the authors' own admission resort to "devious tricks" to get it working
07:26:11 Join solexx [0] (
07:33:53 Join Bg3r [0] (n=Bager@
07:34:07Bg3ramiconn ?
07:34:35XavierGrmorning Bg3r
07:34:44Bg3rmorning :)
07:40:41 Part RoC_MM ("Leaving")
07:41:48 Join midkay [0] (n=midkay@rockbox/developer/midkay)
07:48:01XavierGr Bg3r: I did a bettery bench for my H300 today
07:48:20XavierGrBad thing thing was that I forgot to update my build :(
07:48:43XavierGrBattery_bench didn't capture a single measurement
07:49:16XavierGrbut at least I kept a timer myself so I managed to see how long it managed to play
07:49:27XavierGr14:40hours with a 1900mah battery
07:49:46XavierGrH100 did 24:55 with the same battery and album.
07:51:35Bg3roh :(
07:52:17XavierGrI think 10 hours is too much :(
07:59:02XavierGrI mean the difference between H100 runtime and H300
07:59:30B4gderwas that with the same cvs version on both units?
07:59:57XavierGrH100 was much earlier
08:00:33XavierGr 2005-11-15
08:00:41XavierGrwas the build for H100
08:01:00XavierGrand now I can't compare them. I have only 1 1900mah battery
08:02:11 Join nave7693 [0] (
08:02:20 Part nave7693
08:06:24 Join Paul_The_Nerd [0] (
08:09:01Bg3rXavierGr yes, it's too much
08:14:23amiconngood morning
08:18:06 Join mikearthur [0] (
08:18:12Bg3ramiconn morning
08:18:17Bg3rsorry for yesterday ...
08:22:03Bg3ri thought that i could fix the .langs today ...
08:22:34amiconnI just happen to know the .lang system rather well, because I already dealt with it a lot
08:23:01Bg3rhm, i think that this should be noted somewhere too (the changes to the <source>...)
08:23:01amiconnSo I usually know what implications a non-standard change might have
08:23:25Bg3ri'll add it to the wiki
08:24:12amiconnWell, usually, when you add or deprecate string (as a coder), you only ever change english.lang, and leave all others alone
08:24:58amiconnChanging an ID is quite exceptional (although it doesn't break backwards compatibility)
08:25:23amiconnYou could just have changed the meaning, and let the ID alone
08:25:58 Join _FireFly_ [0] (
08:26:40 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
08:27:39 Quit Arrogant ("Leaving")
08:28:01 Join einhirn [0] (
08:28:37 Join ender` [0] (i=ychat@
08:29:07 Quit stevenm ("Connection reset by beer")
08:30:52Bg3ramiconn yep, but imho it's not correct
08:31:49amiconnWell, even after my .lang rework, there are still some IDs where the usage doesn't match the ID name fully. The IDs are only used internally, the user will never see them
08:33:54amiconnSome typos stayed for years, just to avoid touching all .lang files for no user-visible effect (one ID used LENGHT instead of LENGTH)
08:34:39amiconnI corrected it now because I had to touch everything anyway
08:43:35Bg3rwhat about removing them and just specifying the model ?
08:43:43Bg3rremoving one of them
08:44:30Bg3rand how should i do this ?
08:44:38Bg3ri don't want to make a new mistake ...
08:44:40amiconnThat's something I already saw during rework and knew it will be solvable with langv2
08:46:04amiconnWell, in this case the correct translation for both player and the rest are known, so this can be applied to all .lang files even if you don't know the language
08:46:28amiconnBtw, this pair is related to a second one
08:47:04Bg3rto _SAVE_PLAYER
08:47:09Bg3rand _SAVE_RECORDER
08:47:32amiconnOh, and the _BATTERY_ part should be replaced when touching it anyway, since as you might have noticed that it doesn't match anymore
08:49:09amiconnThis would be the first actual use of langv2
08:58:05 Join linuxstb_ [0] (
08:59:42 Join Zagor [0] (n=bjst@rockbox/developer/Zagor)
08:59:43 Join petur [0] (
09:00:13 Quit linuxstb (Read error: 110 (Connection timed out))
09:06:01 Join Coldtoast [0] (
09:06:07Coldtoasthi all
09:06:26Coldtoastthere's a prob with the h120 if nobody knows
09:06:31Coldtoastcan't ffwd
09:06:58 Join gtkspert [0] (
09:07:01Coldtoastyou do it once and when you release, it doesn't resume playing
09:07:16B4gderI think lostlogic is aware
09:07:33Paul_The_NerdColdtoast: How old is the build you're trying with?
09:07:44Coldtoastbetter than yesterday at least. last night, it'd just lock the player up and you HAD to hit the reset button
09:07:54Coldtoastit's the absolute latest CVS build
09:07:59Paul_The_NerdThe most recent CVS change says it fixes seeking, but I don't know if that's your bug, or a different seeking bug.
09:08:05Coldtoastit's not fixed
09:08:22Coldtoastjust downloaded and installed it, oh, 3mins ago
09:12:14 Quit Lynx_ (Read error: 104 (Connection reset by peer))
09:12:18 Join Lynx_ [0] (
09:13:11ColdtoastI'd download the latest daily but it's going to have the same bug. heh
09:13:34Coldtoastso, for the first time in ages, I'll have to use the iriver fw
09:13:37*Coldtoast shudders
09:14:05Bg3rColdtoast then just switch to the daily build from yesterday
09:14:07Paul_The_NerdYou could download an older daily
09:16:31Coldtoastany idea when the bug turned up?
09:16:32***Saving seen data "./dancer.seen"
09:17:27Bg3rB4gder, amiconn: is this all i should do for the first real usage of langv2 features:
09:18:20Bg3rthe stupid doesn't have syntax highlight for diffs
09:18:22Paul_The_NerdColdtoast: the playback system's undergoing alot of changes. It could be that most recent commit, or it could be something that was agitated and has been around for a while. But I'm sure it'll get worked out soon enough. :)
09:18:54B4gderBg3r: looks fine to me
09:19:17Bg3rB4gder what about the lang version ?
09:20:18B4gderI'm not sure
09:20:25B4gderstrictly speaking, yo'd have to bump it
09:20:41B4gderbut I think we could also just leave it for our cleanup work
09:20:52B4gderand bump it once
09:21:17B4gderwith langv2, the lang version is becoming a lot more complicated
09:21:28Bg3rbecause ?
09:21:40B4gderbecause it is much easier to break an old language
09:21:57B4gdereven for a specific target only
09:22:16B4gderlike you can set "player: none" on lots of strings
09:22:27B4gderthat the players don't use/have
09:23:10B4gderand if you do that for example on a target that doesn't have feature X
09:23:17Bg3rshould i also remove the old LANG_...[PLAYER|RECORDER] from the translated .lang-s
09:23:24B4gderand then one day you add feature X and remove the "none", it breaks
09:23:42B4gderBg3r: no need, the -u option will do that for translators
09:23:47Bg3rah :)
09:23:55Bg3rthis is good :)
09:24:12amiconnBg3r: During the rework, I would just deprecate old strings which are now combined under 1 ID, to avoid both bumpding the version every time during the rework, and at the same time avoid incompatibility
09:24:47amiconnThen we'll do a final round, removing all deprecations and bumping version once
09:25:13amiconnThe settings save failure messages should even be combined into one.
09:25:31amiconnNo need to start with the dreaded ...1 ...2 etc business again
09:25:41Bg3ramiconn and how to do this ?
09:25:42amiconnWe *do* have splash, for quite a while
09:25:49Bg3rah, yes
09:26:03Bg3rso, just use the splash ?
09:26:06B4gderright, it even works on the player! ;-)
09:26:22amiconnB4gder: Why the ;-) ?
09:26:35Bg3rhow do i deprecate a string ?
09:26:56*B4gder cuts off the ":-)"
09:28:14amiconnIt definitely does work on the player, although it needs a bit of fixing
09:28:37amiconnBg3r: Using splash() also avoids having to handle the remote separately
09:29:20LinusNare splashes spoken?
09:30:17Bg3rLinusN no, but these strings are with empty voice: strings
09:30:39Bg3ramiconn yep
09:31:36B4gderbut I guess making the splashes speak shouldn't be that hard
09:31:46amiconnLinusN: Please test my new button code on X5. I don't expect problems though
09:34:58 Join Zoide777 [0] (
09:35:36amiconnB4gder: Making splashes speak has some ugly implication. They must not even try to speak in some cases
09:37:00*JdGordon needs to set up more computers in my room... winter has come :'(
09:37:06B4gderyes it struck me too
09:37:44peturdownunder, remember?
09:37:48Bg3rJdGordon australia ?
09:38:16peturbut still early to call it winter
09:38:29JdGordonits like 15C outside... thats winter
09:38:42Bg3rJdGordon here it's about 15 too
09:38:48HClmove to europe till the winter has passed
09:38:58Bg3rbut i'd call that spring ...
09:39:03*HCl sighs @ work
09:39:10JdGordonye, coz u guys like the cold ...
09:39:22HCl15C is cold
09:39:25*B4gder has 2C and thinks its spring
09:39:26*petur slaps HCl for reminding him of it
09:39:33HClpeople here are weird
09:39:36peturthe work that is
09:39:38HClit was 15C the other day
09:39:43HCland they said it was hot O.o
09:40:39Paul_The_NerdIt's about 21C here right now. At night.
09:41:02JdGordonnice (y)
09:41:17JdGordonthats much nicer than bloody 15C
09:41:30Paul_The_Nerd15C is almost exactly where I like it.
09:42:29JdGordonwanna swap?
09:42:49Paul_The_NerdI dunno about you, but I've got no money and I don't trust my luck swimming. ;-)
09:50:53 Join bluebrother^ [0] (
09:54:10Zoide777is anyone here familiar with the wm*.c files?
09:54:20Zoide777I am trying to understand the conversion between the ipodlinux and rockbox code
09:54:57Zoide777For example ipodlinux has a line "ipod_i2c_send(0x1a, 0x45, 0x50);", while Rockbox uses "wm8975_write(LOUTMIX1, 0x150);"
09:55:49Zoide777or ipodlinux has "ipod_i2c_send(0x1a, 0x1f, 0xff);" while Rockbox has "wm8975_write(RESET, 0x1ff);"
09:56:17Zoide777i'm not sure what's going on with the hex numbers
09:57:15Bg3rZoide777 the wm8*** chips are documented... see the DataSheets page in the wiki
09:57:48_FireFly_Zoide777: i think the defines LOUTMIX1 equels to 0x1a,0x45 and Reset to 0x1a,0x1f
10:01:23Zoide777Bg3r: thanks, i'll see what i can find there
10:01:24Zoide777_FireFly_: but then it doesn't make sense that "ipod_i2c_send(0x1a, 0x45, 0x50);" becomes "wm8975_write(LOUTMIX1, 0x150);" ... shouldn't it be "wm8975_write(LOUTMIX1, 0x50);" instead?
10:01:44 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
10:01:54 Join einhirn [0] (
10:02:29petursome feature of the chip got enabled afaik
10:02:36peturcheck the datasheets
10:11:57 Join RedBreva [0] (
10:18:42linuxstb_Zoide777: The i2c writes to the WM codecs consist of a 16-bit value. The top 7 bits are the register address, and the bottom 9 bits are the data value to write to that register. The IPL code in your example writes the 16-bit value 0x4550 to the wm8975. This equates to register 0x22 and value 0x150 - which is what Rockbox does.
10:18:52 Nick linuxstb_ is now known as linuxstb (
10:22:25*petur had no idea at all
10:22:46Zoide777linuxstb: wow, i would have never figured that out hehe
10:23:17linuxstbIt's explained in the datasheet for the codec - but it wasn't immediately obvious to me either.
10:24:24Zoide777oh! the "Register Map" on p. 50
10:24:43linuxstbThat will help...
10:27:08linuxstbBut the hard part for recording isn't the initialisation of the codec, you will need to use the FIQ (fast interrupt) to read the incoming data from the codec, and incorporate it into Rockbox's general recording code.
10:27:30crashdis this for recording on 3g?
10:29:22linuxstbNot specifically - it's for all ipods.
10:31:08linuxstblostlogic: The 5g still detects the wall charger as a USB connection. You're the second person to try and close that bug report....
10:31:50peturmost of the recording code is usable, just the dma handler needs to be replaced with that FIQ thing [I think]
10:32:29linuxstbYes, it will be a similar job to the playback code.
10:32:32peturand add init code and gain setting of course
10:32:56Zoide777linuxstb: when you mention 16-bit writes, which of the 3 arguments in ipod_i2c_send corresponds to what?
10:33:34linuxstbThe first value is the i2c address of the wm codec (0x1a), and the second and third values are the high and low bytes of the 16-bit value.
10:34:28Zoide777ah, ok thanks
10:36:58 Quit Xerion (Read error: 104 (Connection reset by peer))
10:37:09 Join Xerion [0] (
10:37:51 Quit RicII (Remote closed the connection)
10:38:19 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
10:38:28 Quit ze (Read error: 110 (Connection timed out))
10:41:33Bg3rB4gder how do i deprecate strings ?
10:42:36B4gderset them to be ""
10:44:30Bg3rk, 10x
10:45:24B4gderof you set them to 'none' (without quotes) they will be completely left out from the strings
10:45:59B4gderwhich is how lots of player strings should be dealt with
10:46:04 Quit RedBreva ("CGI:IRC (EOF)")
10:46:07B4gderand radio strings for targets without radio etc
10:46:12Bg3ryou mean <dest> ?
10:46:36Bg3r(for all)
10:46:47B4gderwell, setting all to none makes no sense ;)
10:46:58Bg3ryes, it makes
10:47:08Bg3rin cases like LANG_BUFFER_STAT_RECORDER
10:48:00B4gderif it won't create output for any target, it could just be completely removed insted?
10:48:40Bg3rbecause amiconn suggested to deprecate them first
10:48:54B4gderyes, but deprecating them is to set ""
10:49:06B4gdersetting them to none removes them
10:49:40B4gderI should perhaps add another keyword for deprecate
10:49:44B4gdermight make it more readable
10:50:45Bg3ryeah, and some notes in won't hurt ;)
10:51:41Bg3ror maybe even moving it to LangFiles
10:51:47Bg3ror something like this
10:55:10 Quit amiconn (Read error: 110 (Connection timed out))
10:55:11 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
11:00:42 Join damaki [0] (
11:01:28 Join Nico_P [0] (
11:03:51*XavierGr learns linux playing with X commands in the new vmware image. :P
11:04:09Bg3rXavierGr go go :)
11:04:52XavierGrI must update the wiki with the info I learnt
11:06:12XavierGrThe thing that I don't get. What's the difference between .xinit and .xsession?
11:07:03XavierGrI think I got it now. /.xinit runs if you type xinit before X has started
11:07:19 Quit DJ_Dooms_Day ("Trillian (")
11:07:37XavierGrwhile /.xsession runs all the commands you put there and exits (so if you don't star a WM it will exit)
11:11:01 Join DJ_Dooms_Day [0] (
11:16:16B4gderBg3r: renamed and edited
11:16:16 Join pkern [0] (
11:16:28Bg3rB4gder very good :)
11:16:36***Saving seen data "./dancer.seen"
11:17:35pkernHm. The Rockbox website is unclear about it, but does the Rockbox bootloader actually work on iPod minis? I tried the 4G bootloader (as I remembered it from ipodlinux to be very similar to the mini ones) and the background light indeed flashes, but only with an apple within.
11:17:55B4gderpkern: you need the mini bootloader
11:18:09pkernB4gder: I thought that and I saw build logs, but where could I get it?
11:18:23pkernb4gder: It's not included in the zip, isn't it?
11:18:39 Quit _FireFly_ ("Leaving")
11:19:31pkernB4gder: I'm sorry, thanks.
11:20:19 Join DrMoos [0] (
11:20:57linuxstbpkern: Do you know if you have a first or second generation mini?
11:21:10pkernlinuxstb: IIRC it's 2g
11:21:22pkernIt's a Silver 4GB Mini
11:21:23linuxstbThat's perfect then. Only the 2g is known to work well.
11:21:37pkernlinuxstb: At least ipodlinux did once run on it ;)
11:22:21pkernTime to put some audio onto it. Perhaps it does work in an usable way...
11:22:28amiconnpkern: There's a dedicated mini2g rockbox bootloader & rockbox build
11:22:45pkernamiconn: Well, I took the mini2g rockbox build from the webpage.
11:22:45amiconnIt definitely does work.
11:23:15pkernWell, I get the screen, it's just in the FAQ that audio is not yet continous.
11:23:15 Quit DJ_Dooms_Day (Read error: 104 (Connection reset by peer))
11:23:21amiconnYou also need to use the mini2g bootloader. Other bootloaders won't work
11:23:35pkernamiconn: Yep, it started :D
11:24:13Paul_The_Nerdpkern: If you're referring to IpodFaq, the portion is in response to some people having problems with audio. It often depends on the bitrates or whether you're using the Equalizer, or show the Peakmeters, or other things.
11:24:30amiconnPlayback should be working fine, as long as you don't try mp3 with the eq enabled
11:24:50pkernI mainly want to play OGG ;)
11:25:00Paul_The_NerdWell you're in luck
11:25:04amiconnPeakmeter is no problem on the greyscale ipods. It might be on the colour ones
11:25:09Paul_The_NerdOgg tends to work pretty well.
11:25:26Paul_The_Nerdamiconn: Peakmeter has caused problems even on H120s sometimes recently. There's something up there. =/
11:25:40pkernHm. I get overmodulation or whatever that's called.
11:26:06pkernThe OGG is weird. (Wesnoth one)
11:26:08pkernSo thanks a lot (=
11:26:08Paul_The_Nerdamiconn: There were reports of playback skips in the H120 forum at least, and when I recommended disabling the peakmeter it was fine. They *may* have been using other cpu intensive things simultaneously though
11:30:21pkernHm. 203kbit/s OGG playing fine. Well, Rockbox very much exceeds my expectations.
11:30:42pkernSo thanks for your work. :D
11:31:52 Join thegeek [0] (
11:32:59amiconnB4gder: We don't want any lang IDs to be non-present on any target
11:33:17B4gderI do
11:33:37amiconnAll that should happen is set them to empty, which means wasting one byte (the \0 )
11:33:42 Join ashridah [0] (
11:33:47B4gderI know your position on that
11:33:47amiconnWe already discussed that a while ago
11:34:26*petur notices gl has threatened *again* to leave :)
11:34:46amiconnThe bad thing with missing ids is that if someone implements a feature and enables a string for a target where it wasn't enabled before, that'll screw up string order for this target
11:34:51LinusNpetur: i'm sooo scared
11:34:55B4gderamiconn: I know
11:35:02amiconnSo we would be forced to bump the binary lang version
11:35:11amiconnI would prefer to avoid that mess
11:35:12B4gderbut I think its worth that hassle
11:35:44B4gderits not like people get random language files and use anyway
11:35:56amiconnIt's not about random .lng files
11:36:03amiconnIt's about older .lng files
11:36:14B4gderpeople don't use older either
11:36:19B4gderthey use the ones we ship
11:36:27 Join markun_ [0] (
11:36:31B4gderI mean in real-world
11:36:35 Quit ashridah (Remote closed the connection)
11:36:42amiconnWhat happens if people forget to replace their .lng file, just replacing rockbox.whatever?
11:36:55B4gderthey mess up
11:36:59B4gderthey do things like that today
11:37:01B4gderand suffer
11:37:07B4gderso they learn not to
11:37:09 Quit pkern ("I'm very glad about Rockbox, thank you. (:")
11:37:17B4gdereverything else breaks hard when you do those things
11:37:20*amiconn is a bit disappointed
11:37:35B4gderabout what?
11:37:42Paul_The_Nerd Yeah, as it is just only ever replacing the rockbox.whatever as an update method is asking to encounter troubles in the end.
11:38:03B4gderthe breakage of bad updates is not due to lang files
11:38:08Paul_The_NerdAren't the installation instructions "extract the .zip on your player"?
11:38:10amiconnWe have a number of measures to avoid severe messup in case of errors like this. The plugin api versioning comes to mind...
11:38:19amiconnWhy not use a similar policy here?
11:38:30B4gderI'm not against that
11:38:36B4gderI'm just against your way of solving that
11:38:58B4gderI'd rather have a more strict policy
11:39:05Paul_The_NerdIt seems the plugin API's method would work pretty well.
11:39:28B4gderits just that the language files are more complex than the plugin API
11:39:36B4gderor they could be at least
11:40:12amiconnWell, languages work different. For plugin api, we need to enforce a minimum version, because otherwise a plugin might call a function not yet present, and crash
11:40:24 Quit markun (Read error: 60 (Operation timed out))
11:40:36Mikachuis the byte only wasted on disk or in ram too?
11:40:41Paul_The_NerdWouldn't you need a minimum version with languages too?
11:40:52amiconnHowever, for .lng, we want to support older files, and just use the english strings for things not yet present in the .lng
11:41:01B4gderMikachu: both, and its more than a byte
11:41:18Mikachuyeah i guess it needs some padding etc
11:41:24amiconnYeah, in RAM it's 5 bytes
11:41:41amiconn(the \0 and a pointer)
11:41:47B4gderand on the player, I bet there's a 100 strings it doesn't want/use
11:41:50Mikachuwhy do you load the string if it's not used anywhere in the build? shouldn't that be pretty easy to avoid?
11:42:12 Join RedBreva [0] (
11:42:23 Quit RedBreva (Client Quit)
11:42:40B4gdernow where's reaching the culprit ;-)
11:43:04B4gderwe load them because they are in the .lng file
11:43:21B4gderand they are in the .lng file because the .lang file says so
11:43:38XavierGrwhich command do I have to type in Linux to see running processes?
11:43:42B4gderand the .lang file says so, because if we _ever_ add the string we don't want to break backwards compatiblity
11:43:50 Join ashridah [0] (
11:43:54B4gderXavierGr: ps
11:44:13Paul_The_NerdI don't understand why breaking backwards compatibility is so horrible?
11:44:28B4gderthat's my position too
11:44:42Paul_The_NerdI mean yes, people will have to make a choice between using a new build in english, or the one they've already got in their native tongue.
11:45:19Bg3rthe best would be to not allow loading of older lang file ...
11:45:25Paul_The_NerdIt also provides more incentive for people to update translations, if a bit draconian in its methodology.
11:45:37Mikachuis the lang file built by regular configure+make?
11:45:46Bg3rMikachu .lng is
11:45:49B4gderMikachu: make zip builds them
11:45:55Mikachuoh.. silly :)
11:47:31linuxstbXavierGr: There is also the "top" command - which is an interactive version of ps. Press "q" to quit it.
11:48:19XavierGrnice thanks
11:49:03XavierGralso the kill command works for me only with the -9 option
11:49:23B4gderXavierGr: it depends on the program you try to kill
11:49:46B4gdermost signals can be caught and dealt with by the app itself
11:50:00Mikachuwhy is the swedish lang file called svenska, but the japanese is japanese?
11:50:13B4gderMikachu: to confuse readers
11:50:25LinusNwouldn't the voice files suffer from incompatible lang changes?
11:50:36MikachuB4gder: ah, okay
11:50:52B4gderMikachu: I guess its just because different people adding them
11:50:57amiconnLinusN: yes, exactly
11:51:09B4gderdon't the voice files use the id ?
11:51:23amiconnThey use the numeric id the same way as .lng
11:51:24B4gderor do the use the same ordered thing as strings?
11:52:05amiconnSo, if the numeric IDs shift because of left-out strings introduced later, voice clips will also be messed up
11:52:14LinusNi think keeping the voice file compatible is a good thing
11:52:22 Quit bluebrother^ ("leaving")
11:52:23amiconn...and unlike .lng, we don't include .voice in the .zip
11:53:24Mikachuis english included in the binary? i haven't updated the lng file since i first installed but new strings appear anyway
11:53:32amiconnMikachu: I think the rule is to write the language name in the target language, unless it's impossible with latin1
11:53:44B4gderMikachu: english is always built-in
11:54:32Mikachuwhat is wallisertitsch?
11:54:56XavierGrMikachu maybe we will have to rename the lang file to 'nihongo'
11:56:09amiconnI'm not sure whether zip can handle unicode
11:56:23Mikachuthere is probably someone that has a zip implementation that doesn't
11:56:38Mikachuas long as it doesn't strip arbitrary 8bit bytes from filenames it should be fine
11:56:55B4gderI'd hate 8bit letters in file names in our tree
11:57:16Mikachumine works fine
11:57:25Mikachu% unzip
11:57:25Mikachu extracting: にほんご
11:57:46B4gderbecause it would force me to make sure all my systems use utf8 or what not
11:57:57B4gderand they don't now
11:58:03amiconnThey would need to be true unicode filenames, not just utf-8 without zip or other tools not knowing about it
11:58:16amiconnMight also be a problem with cvs...
11:58:24Mikachuwhat do you mean by"true unicode"?
11:59:03 Nick DrMoos is now known as Moos (
11:59:14Mikachua filename by itself doesn't have an encoding
11:59:48Mikachuit's just a couple of bytes \ {'/'}
11:59:52amiconnI mean that the tools need to know about it, and the filenames need to be in the native unicode encoding *on the file system*
12:01:14Mikachuin general, filesystems don't have encodings
12:01:31 Quit Coldtoast ()
12:01:32Mikachuthey only store said bytes in a filename field...
12:01:40amiconnThat's not always true
12:02:01amiconnAs mentioned yesterday, VFAT filenames are *always* UCS-2 encoded
12:02:23amiconnThe OS translates that to/from whatever the current locale uses
12:02:52Mikachuthat's up to the filesystem driver though, not zip
12:03:02crashdjust wanted to headsup, as we cant submit bugs for 'new targets', searching on ipod is the fucked, you can search once, but once you let go and try and skip through again rockbox hangs
12:03:19crashdat least on the 5g
12:03:29Mikachuof course you can put ipod bugs in the tracker...
12:03:34Mikachuthat's what the target field is for
12:03:38Paul_The_NerdYou mean "next track" or the actual search function?
12:03:42crashdMikachu: i understand that.
12:03:46crashdbut Please do not submit bug reports for very early Rockbox target ports!
12:03:52crashdand it is still an early target, non?
12:03:55Paul_The_NerdMikachu: Until the new tracker, you weren't supposed to put bugs up for targets other than those that had an official release.
12:03:57crashdPaul_The_Nerd: searching through a track
12:04:00crashdas in, ff and rw
12:04:05Paul_The_Nerdcrashd: That's "seeking" then. :)
12:04:16Mikachuin this case, it's not related to ipods, but lostlogic's swcodec hackings
12:04:17crashdahh, sorry, i do agree that seeking is a better word ;)
12:04:29Paul_The_NerdYeah, that's an all-SWCODEC bug at the moment
12:04:31Mikachuhe knows it doesn't work well yet, i think
12:04:37crashdah, ok, fair enough
12:04:46Paul_The_NerdBut yeah, it's okay to post ipod-related bugs now I think
12:05:02Mikachuif a devel doesn't want to watch it, he can exclude ipod targets from the search
12:05:06Mikachuor she!
12:05:45Mikachuor would be able if flyspray had a sensible search function
12:06:03Paul_The_NerdYeah, it should probably be changed to something as simple as "Don't report bugs until Audio works on your target" rather than as it is now.
12:07:08linuxstbOr just until daily builds are available.
12:07:23B4gderthe thing is that people start to file bugs like crazy on things we just haven't done yet
12:07:31 Join ep0ch [0] (
12:07:35B4gderlike the "remote on x5 doesn't work!"
12:08:00Paul_The_NerdThat is true.
12:08:04B4gderbut its hard to explain that in a few words
12:08:10linuxstbBut even bug reports like that still have some use IMO - so things are documented somewhere.
12:08:27B4gderhm, yeah you might be right
12:08:53Paul_The_Nerd"Please, don't file bug reports until Audio is functional, search carefully for your bug before filing a report, and try to make sure in advance it's not a case of 'the feature isn't done yet'"
12:09:16Mikachuso if english is compiled in, what happens if i select english.lng, all strings are stored twice?
12:13:06*petur joins B4gder
12:14:27 Join theli_ua [0] (
12:17:51 Join ze [0] (
12:26:27Febs_Look like gl.tter is taking his toys and going home.
12:27:45 Join mnhnhyouh [0] (
12:27:52 Quit mnhnhyouh (Client Quit)
12:28:06 Join anrdewmel [0] (
12:28:36anrdewmelwondering if anybody here is working on making plugins work for iPod? i
12:28:51anrdewmeli have found a couple of problems I would like to flag
12:29:40Mikachuwould doing something like this make sense for ipods?
12:29:44Paul_The_NerdWhich plugin are you having problems with?
12:29:55Mikachumaybe i should wait for lostlogic
12:29:59anrdewmelthe jpeg viewer and the stopwatch
12:30:18anrdewmeli have edited the wiki to show what is and isnt working
12:31:27Paul_The_NerdSo, for the jpeg viewer it's mainly the fact that only limit types of jpegs are supported?
12:31:46 Join Jungti1234 [0] (n=jungti12@
12:31:58anrdewmelnope, there is no way to terminate it, or move to the next or previous image
12:32:17Paul_The_NerdI can move to previous/next image...
12:32:25Mikachuselect+left/right and select+menu
12:32:26anrdewmelhmmmm, how are you doing that
12:32:32anrdewmelahhhhhh, bugger
12:32:45anrdewmeland to terminate?
12:33:21Mikachumy button change thing seems to work
12:33:38Paul_The_NerdWhat does it do?
12:33:38anrdewmelhmmm, looks like some editing is in order, thanks
12:33:51Zoide777Mikachu: does it prevent the flood of scroll events? can you scroll long lists without the sound skipping?
12:33:55Mikachuhrm, it probably breaks repeat now that i think :)
12:34:22anrdewmelI will edit the wiki ....
12:34:48 Join damaki_ [0] (
12:35:10Mikachuhm, it didn't
12:35:12Mikachuthat's odd
12:35:54Paul_The_NerdWait, so it works AND it doesn't break things?
12:36:10Mikachuyes, but it should break repeat
12:36:13Mikachuso that means it doesn't work
12:36:35Paul_The_NerdWhen working is defined by breaking something else... interesting. :)
12:37:31*Mikachu adds piezo ticking to hear when the button drivers if statement is true
12:37:47Mikachuah, constant sound
12:38:08Mikachuhah, of course, i'm stupid
12:38:13Mikachutick−− won't stop at zero
12:39:08 Join _FireFly_ [0] (
12:40:42Mikachunow it sounds right, but the buttons only work if i scroll at the same time i press a button :P
12:41:22Paul_The_NerdWell that is unfortunate.
12:41:37Paul_The_NerdThough it would prevent accidental volume adjustments.
12:41:43Paul_The_Nerd"Scroll only while select is held"!
12:41:50Mikachuscrolling always works
12:42:03Paul_The_NerdButtons only while scrolling
12:42:07Mikachuthose events are sent from the interrupt handler
12:42:08Paul_The_NerdWell, nobody wants THAT
12:42:14Mikachuit's obviously not on purpose
12:42:25Paul_The_NerdI know.
12:43:21Paul_The_NerdI'm just giving you a hard time because the forum is slow at the moment. :)
12:43:55linuxstbPaul_The_Nerd: People have run out of Doom questions?
12:44:56Paul_The_NerdAt the moment. It's rather astonishing.
12:45:16Mikachuhm, apparently button_tick needs to run twice to register a button, /me adds a tick++;
12:47:35Mikachuheh, now i managed to make it tick only when i release the finger from the wheel, it's somewhat unsettling
12:49:39 Join damaki__ [0] (
12:50:02 Quit damaki (Read error: 110 (Connection timed out))
12:55:16*Paul_The_Nerd realized he finally found something proper for his forum sig.
12:55:35JdGordonu dont _have_ to have a sig u know...
12:55:48Paul_The_NerdYes, I know. And I didn't want one, until now.
12:56:05Paul_The_NerdBut I realized it would be the perfect place to keep that "How to ask questions the smart way" link LinusN posted.
12:56:57Febs_Is that the article by Eric Raymond? I've had that in my Mistic River sig for a long time. Unfortunately, the people who most need to read it never will.
12:57:29Paul_The_NerdFebs_: Yes, that one. I know they never will, but at least it's there now.
12:59:27XavierGrPaul where is that link?
13:01:35 Quit anrdewmel ("Trillian (")
13:02:15JdGordonany1 know how much 256mb ddr laptop ram would go for 2nd hand?
13:03:06XavierGrwow big Guide on just to tell you how to ask. Imagine the answer :p
13:06:26 Quit damaki_ (Read error: 110 (Connection timed out))
13:07:57solexxhm. does anyone else still have problems seeking in sw codec platforms?
13:08:19Paul_The_Nerdsolexx: Many people have reported that it's still messed up, yes.
13:09:40amiconnMikachu: What are you trying to do?
13:11:35Mikachuamiconn: skip most of button_tick when i know no button is pressed
13:13:24 Join MarcoPolo [0] (
13:14:32 Join cismo [0] (
13:16:39***Saving seen data "./dancer.seen"
13:17:31 Quit Nico_P (Remote closed the connection)
13:19:13MarcoPoloi get some weird error with current cvs :
13:19:26MarcoPoloUndefined subroutine &main::No Info called at genlang line 418, <LANG> line 2490.
13:19:58amiconnMikachu: Hmm, I don't get it? button_tick is there to determine which buttons are pressed. How can you know no button is pressed before processing this very information?
13:20:04amiconnWhat target, btw?
13:20:15Mikachuamiconn: on ipods, so i know in the interrupt context
13:20:39amiconnWell, for the 4g iPods, button_tick() is almost empty anyway
13:21:07Mikachuone if is less than 5 :)
13:22:09amiconnThere's just 2 for iPod 4g and up
13:22:16amiconnIt all boils down to
13:22:47amiconnbtn = int_btn; then check for flip and apply, then apply debounce
13:23:33amiconnline 1211 and lines 1293..1306
13:23:36Mikachui guess so
13:24:45amiconnThe debounce check *might* be unnecessary for iPod, depends on whether there is hardware button debounce
13:25:04amiconnBut it doesn't hurt to do it.
13:25:26 Quit ashridah ("Leaving")
13:26:20Mikachui also noticed POLL_FREQUENCY is HZ/100, doesn't that mean the first if statement is always true?
13:27:34amiconnAh, you're talking about button_tick...
13:27:46 Join damaki_ [0] (
13:27:46amiconnYes, this first if is always true nowadays
13:30:27amiconnIt could be removed, together with the POLL_FREQUENCY definition, if we don't want to go back to slower polling one day
13:30:27 Quit thegeek (Read error: 104 (Connection reset by peer))
13:31:00amiconnBut you can't easily skip parts of that code without breaking the release/repeat handling
13:31:10amiconnThis is quite optimised as it is
13:31:21Mikachui noticed
13:31:42Mikachu(that you can break it)
13:34:07Zoide777arrrggghhh.... ipod recording stuff is driving me *craaaazy*
13:34:29Zoide777I keep getting undefined references to a bunch of functions in recording.c
13:34:46Mikachua lot of stuff is probably ifdeffed out for ipods
13:35:04Zoide777even though they are all in pcm_record.c and I am doing #include pcm_record.h
13:35:34Zoide777Mikachu: the ifdefs *are* a pain...
13:36:10 Quit MarcoPolo (Read error: 104 (Connection reset by peer))
13:36:22Mikachuhm, it seems the hold button actually generates an interrupt
13:36:42 Join MarcoPolo [0] (
13:36:56 Quit Jungti1234 ()
13:38:25 Quit cismo_ (Read error: 110 (Connection timed out))
13:38:54peturZoide777: does your ipod config file have #define HAVE_RECORDING ?
13:39:05Zoide777petur: yes
13:39:51peturalso check SOURCES and FILES that the file is compiled for your target
13:42:50 Join Nibbler [0] (
13:43:28 Quit damaki__ (Read error: 110 (Connection timed out))
13:43:40Zoide777petur: that might be it! pcm_record wasn't being built as far as the SOURCES file was concerned
13:44:24Zoide777well, now i have new errors, but at least i got out of the rut i was in for so long!
13:44:26Zoide777thanks petur
13:46:00peturrecording is about the only part of RB I know a bit of :)
13:47:47Paul_The_Nerdlostlogic: Your "fixes seeking" thing apparently didn't.
13:48:01lostlogicPaul_The_Nerd: oh?
13:48:04Mikachuum, how do i put an int in a string in rockbox? "%i" and "%d" didn't work
13:48:16lostlogic*seeks a bit*
13:48:30lostlogicsure seems to work
13:48:37Paul_The_Nerdlostlogic: I've got reports of swcodec seeking not working properly on H120s and various iPods...
13:48:38lostlogiceven when disk reload is needed
13:48:41Mikachuer, yeah, forgot to copy the plugin
13:49:15Paul_The_NerdFF/RW is freezing players for many people.
13:49:34lostlogichahahhahaha, I think I may have broken the opposite case, instead of it not working when disk reload is needed, it might not work when disk reload is _not_ needed now :(
13:49:51Paul_The_NerdSuch things happen though
13:50:06lostlogicalthough I haven't had a freeze yet, just playback fails, stop and resume works which is better than the old behavior here.
13:50:35 Quit quobl (Remote closed the connection)
13:50:45 Join quobl [0] (n=quobl@tor/session/x-1d0a05e4582d9e49)
13:50:58Paul_The_NerdWell, these were ordinary people saying "It freezes" so that could just mean that playback halted.
13:51:10lostlogicon the upside, the playback.c code is beginning to make sense in a basic sort of "this can reasonably work" way :)
13:51:30lostlogicPaul_The_Nerd: yeah, that's my guess, pending further reports of _freezes_
13:51:42Mikachulostlogic: not that you care, but that last queue_wait thing broke headphone detection a bit :)
13:52:02lostlogicMikachu: which what how?
13:52:10lostlogicheadphone detection? we have that?
13:52:16Mikachuno, it's a patch on the tracker
13:52:26Paul_The_Nerdlostlogic: Actually, I remember one of them saying specifically "At least it no longer locks up my player after this update" so I'm actually guessing it's just playback halting.
13:52:29lostlogicoh, if anything I did broke it, then it's done wrong
13:52:37ep0chlostlogic: yeah playback hasn't hung rockbox for me yet, but when i FF sound stops so i have to press pause and play to get playback to start again
13:52:47Mikachuit polls for the headphone in the playback thread i think
13:52:51lostlogicep0ch: great, useful information!
13:52:58lostlogicMikachu: right, and that is _wrong_.
13:53:09Paul_The_NerdEven *I* know that's wrong...
13:53:29lostlogicplayback thread is designed for serialization, which is why I had to do the stuff I did last night, it can't be used for polling.
13:53:35Mikachuyeah, we want to poll somewhere else and put an event on the playback queue! (i guess)
13:53:42lostlogicMikachu: exactly :)
13:54:14ep0chpreglow: how's the new crossfeed? I want something to play with :D
13:54:15Mikachuwhen i say playback*, it seems i mean audio_*
13:54:34lostlogicMikachu: yep, they seem synonymous to me.
13:54:45Mikachuso at least you don't have both then :)
13:54:50preglowep0ch: try to have it done today
13:55:07 Join thegeek [0] (
13:55:13preglowbut no promises, i've also got a bunch of work
13:55:40XavierGrwork? what is that preglow :P
13:55:49XavierGryour work is here! :D
13:56:12Zoide777what are DSR1, DCR1, DAR1, BCR1, IPR? they are present in mcf5249.h, for example, but not in pp5020.h
13:56:31preglowZoide777: they are completely different architectures...
13:56:32peturthey are for DMA
13:57:07peturas said earlier, that's the place where the platform specific code must come
13:57:09Zoide777hmm.... the situation does not seem very friendly....
13:57:16Zoide777i see
13:57:37 Join webguest22 [0] (
13:57:50Mikachulostlogic: do you have an idea where it would be logical to poll it? button_tick()?
13:58:09peturthe init for recording and he handling of data reception must be written for the ipods, the rest is reusable
13:58:35 Join webguest42 [0] (
13:58:39 Quit webguest22 (Client Quit)
13:58:57 Nick webguest42 is now known as obo (
13:59:21oboMikachu: guess I've got some work to do
13:59:24Mikachubtw, it seems the button driver happily sends repeat events when hold is on
13:59:45peturhold repeat? :D
14:00:02B4gder"hold is still active hold is still active hold is still active hold is still active" ;-)
14:00:05Mikachui accidentally pressed button_left when turning on hold earlier today and it went seeking
14:00:20Mikachuuntil i turned off hold, and then playback froze :)
14:01:18 Join damaki__ [0] (
14:02:56Mikachuobo: i have a pretty good idea of how to do it now
14:04:14oboMikachu: move it into button.c??
14:06:39lostlogicMikachu: could make headphone insert/remove a button event actually and then handle it on the WPS thread (or any thread who wants to do soemthign with it)
14:06:55*B4gder awakes
14:08:21Bg3rB4gder prv
14:09:08lostlogicany swcodec people care to comment on performance changes from my commit last night? New skips during buffering? Better or worse boost levels during play?
14:09:13 Quit MarcoPolo (Remote closed the connection)
14:13:52XavierGrdpkg seems evil
14:14:20ep0chI haven't noticed any skipping during buffering
14:15:50 Join Kanashimi [0] (
14:16:25peturlostlogic: your mail will tell if he actually left, because it's the type of mails he likes to jump on ;)
14:16:59lostlogicpetur: if you say so *shrug*
14:17:05 Quit damaki_ (Read error: 110 (Connection timed out))
14:18:29KanashimiHi there, just thought I'd hop on and let you know that the current version of Rockbox for IHP-120 ( just hard locked twice when I was trying to fastforward an audio book. All the controls froze completely and would not respond in any manner. The title of the file still kept on moving though. The lock
14:18:34Kanashimihappened after I stopped fastforwarding and the player should have started playing the file again.
14:19:13Paul_The_Nerdlostlogic: See, all night. :)
14:19:14KanashimiI had to do a hard reset from the hardware button to free up the player again.
14:19:43Bg3rhah lostlogic your comment about B5032 ...
14:19:56Bg3rbetter ask "with today's daily build"
14:20:40Paul_The_Nerd"yes, but I haven't updated in a while."
14:21:10peturbetter ask "with todays bleeding edge"
14:21:24KanashimiSame audio book fastforwards and works just fine with version CVS-051125.
14:22:05 Quit Febs_ (Read error: 110 (Connection timed out))
14:22:21peturthe last fix is only in todays bleeding edge build, not the daily *I think*
14:22:47ep0chKanashimi: try bleeding edge, there was an update to playback code at 4am
14:22:56Paul_The_NerdKanashimi: This is a known bug, but thanks for reporting it. At the moment the playback code is being reworked, and this is resulting in problems like this here and there. Hopefully it should be fixed soon, and we look forward to hearing back from you if there are further problems once it is. :)
14:23:07Paul_The_Nerdep0ch: That 4am update is actually the one that *causes* it. :)
14:23:16Bg3rPaul_The_Nerd :)
14:23:27ep0chwhat time is the dailybuild made?
14:23:33Bg3rbtw, what happened with the theme contest ?
14:23:39peturwith development now going on 24/24, what's the use of the dailies?
14:23:50LinusNBg3r: it hasn't been announced yet
14:23:56Paul_The_Nerdep0ch: Well, okay, it may fix the hard freeze, but it replaces it with a soft freeze.
14:24:14Paul_The_Nerdpetur: Archival copies so that when something breaks people can get one from 24 hours ago.
14:24:40KanashimiPaul_The_Nerd: Ah, alright then, and sure. Thanks for the info, I'll let you know if I have any problems with the new version when it comes out :) Oh btw.. another question, what's with the volume level being reported in negatives in that build, too ? :)
14:25:00peturhey long time no see :P
14:25:11peturthey're in dB now
14:25:19petur0 is max
14:25:19Paul_The_NerdKanashimi: Volume levels have been in negatives for a long time. It's a measurement of attenuation vs the base level of the file, as that's more useful functionally for people who need it.
14:25:40B4gderwe're just a negative bunch!
14:25:50KanashimiHeh, alright :)
14:26:31Paul_The_NerdKanashimi: The main advantage of that is that if you use Rockbox on an older device, then move to an iPod, if you were listening at -32 before, you'll be listening at it now too.
14:26:49Paul_The_NerdKanashimi: As well, you also know that 0 is line out level (since some devices go into the positives, though this may result in clipping)
14:27:00KanashimiAnyways, a thanks for you all for working on creating rockbox and porting it over to the iriver's too. I've been using the firmware for a while now, and in specific enjoyed some of the functions like bookmarks and so that the original firmware doesn't provide.
14:27:20*petur goes back to MFC coding :(
14:27:37LinusNspeaking of bookmarks, it looks like rename() doesn't work when dircache is enabled
14:27:39KanashimiPaul_The_Nerd: Ah hmm *nods some* so it's a more standard metric in that regard.
14:27:48 Join damaki_ [0] (
14:28:07Zoide777could someone lend me a hand w/ the ipod fifo & dma stuff for recording?
14:28:09Paul_The_NerdKanashimi: Much moreso than just a percentage, yes. 100% would then mean different things on a target that can go from -83 to 0 than one that goes from -78 to +6
14:28:27preglowpetur: sweet lord, mfc :/
14:28:28Zoide777it looks almost impossible for someone like me who's not very familiar with how it works
14:28:37preglowgive this man a valium
14:29:01Zoide777and one for me too, please!
14:29:06*petur passes the valium in overdose on to the customer
14:29:19KanashimiPaul_The_Nerd: Hmm, indeed, I see what you mean. Well thanks for clearing it up for me *small smile and a nod*. I'll just use the old version of rockbox I have for the moment and check up on a update in a few days or so.
14:29:27preglowZoide777: going to ruin my lovely asm optimised fiq handler, are you? :/
14:29:31Paul_The_NerdHm. With WMA on the horizon, and MP3 recording an assumed likelihood, is there anything else from the H1x0 stock firmware Rockbox lacks?
14:29:45Paul_The_NerdKanashimi: I would actually suggest waiting until May 1st. :)
14:29:58Zoide777preglow: maybe yes, maybe no. i'm not too sure what i'm doing! :P
14:30:04KanashimiPaul_The_Nerd: Oh ?
14:30:08Paul_The_NerdKanashimi: There's a lot of changes in the works at the moment, but 3.0 *should* be available at or near that date, and hopefully it will be "stable"
14:30:45preglowZoide777: don't worry about the asm handler, revert it to use the c handler instead, and i'll fix it when i can
14:31:12KanashimiPaul_The_Nerd: Ah.. ok. Any major changes in specific ?
14:31:31 Quit _FireFly_ ("Leaving")
14:31:34Zoide777preglow: I don't quite understand what you mean by that... i'm trying to port ipod recording code from ipodlinux (audio.c) to rockbox's pcm_record.c
14:32:18Paul_The_NerdKanashimi: Well, Tagcache will hopefully be fully functional (or at least usably consistent). That's a database so you can find tracks by Artist, Album, or Genre. As well, there's a new language system, so it might actually say "Navi = Yes" in those Yes/No screens. :) Plus of course, bugfixes and optimizations, and Doom.
14:32:22Zoide777the problem is that pcm_record.c is structured very differently, and it relies on variables like DCR1, BCR1, etc. that are not used by pp5020 from what i've been told
14:32:44preglowZoide777: you probably can't use dma for ipod recording naywya
14:32:48preglowanyway, too
14:34:23preglowbecause we don't know how
14:34:47KanashimiPaul_The_Nerd: Mm, alright :) Thanks again for the info, I'll need to run off now though. Sso take care everyone.
14:34:59 Quit damaki_ (Read error: 104 (Connection reset by peer))
14:35:36 Join damaki_ [0] (
14:36:05 Quit Kanashimi ("-")
14:36:40Zoide777preglow: can't we just use the ipodlinux code? for example there is a function in their audio.c called "i2s_pp5020_rec_dma"
14:42:06peturZoide777: porting != copying
14:42:34peturyou'll obviously need to change it to fit into our recording framework
14:43:02 Quit damaki__ (Read error: 110 (Connection timed out))
14:43:40peturthere's an init function that configures the dma currently, there you need to put the code to start it
14:43:54preglowZoide777: they don't use dma
14:44:06preglowZoide777: the function is named dma because they're doing soft dma
14:44:15preglowZoide777: at least that's the way it is for playback code
14:45:34peturZoide777: the current code uses a circular buffer consisting of chunks of 8KB and a pointer to the chunk that is currently being filled from hardware
14:46:01 Join SereR0KR [0] (
14:46:40webmindhmm, how do you turn on an ipod in deep sleep ?
14:46:53lostlogicYeah, and I'm goign to do another fairly big commit to the audio stuff pretty soon, but I think we're past the downside of the functionality ladder.
14:47:12Mikachuwebmind: menu
14:47:52lostlogichmph, the filebuf is never filled up completely, only up to how many bytes it had space for when the fill process was initiated.
14:47:57webminddoesnt seem to work...
14:47:58amiconnlostlogic: That odd playback stop behaviour is still there.
14:48:37lostlogicamiconn: ok
14:48:40 Join _FireFly_ [0] (
14:48:45Paul_The_Nerdwebmind: Did you have some sort of error before it entered deep sleep?
14:49:01Zoide777petur: and what are dcr1, etc? are they like registers?
14:49:03Paul_The_Nerdwebmind: Try holding Menu + Select until it reboots. If that doesn't work, are you sure the battery is charged?
14:49:11lostlogicamiconn: doesn't happen on ipod that I've seen :-\
14:49:17 Quit XavierGr (Read error: 104 (Connection reset by peer))
14:49:35webmindno error that I know
14:49:47webmindcould be that the battery died while it was paused
14:50:02Paul_The_Nerdwebmind: Well, if menu+select doesn't reboot it and wake it, then it needs charging.
14:50:03webmindit was in my pocket for 1 day without checking
14:50:09peturZoide777: they're from the current implementation that uses dma
14:50:11webmindseems so
14:50:35Paul_The_NerdYou have to hold Menu + Select for several seconds (8? 10?)
14:50:49peturZoide777: that's all code that needs to be replaced with proper handling for your platform
14:50:54webmindsome better battery support for the nano would be lovely :)
14:51:09amiconnIf I am playing music on my H300 and hit Stop, music stops immediately for ~0.1 sec, then continues for ~0.5 sec, then stops
14:51:11Paul_The_NerdIt's coming eventually. :)
14:51:27Zoide777i see, so the ipod code in ipodlinux doesn't use real dma, but the non-ipod code in rockbox does use dma
14:52:02amiconnThis is with voice enabled, and *sometimes* (if voice decides that it will be working), I can hear a short snippet of my music again when the first voice clip is played
14:53:41Zoide777one thing i don't understand is the way the 3 letter variables appear to behave like functions. for example, the DMA transfer is started by the strange line "DCR1 = DMA_INT | DMA_EEXT | DMA_CS | DMA_DINC | DMA_START;" and is stopped by "DCR1=0"
14:53:43lostlogicamiconn: I continue to hate voice.
14:53:47amiconnlostlogic: Regarding your question about ide dma last night: No, rockbox doesn't use dma for ide transfers, on any platform
14:54:01lostlogicamiconn: k, thanks
14:54:02Zoide777do they behave more like sentinels?
14:54:19 Join swampcow [0] (
14:54:47amiconnOn sh1 (archos) it's not possible due to the endian swap. On iriver it might be possible, but we only have 2 dma channels, which are currently reserved exclusively for playback + recording
14:54:59lostlogicZoide777: they are control registers.
14:54:59peturZoide777: its a bunch of bits that get set in a register to start/stop DMA
14:55:01 Join JBGood25 [0] (
14:55:09 Quit B4gder ("time to say moo")
14:55:31amiconnFurthermore it might end up slower to use dma, as it doesn't use memory bursts in the required operating mode
14:56:18lostlogicamiconn: good to know, so no non-blocking IO for me :(:)
14:56:30amiconnI have no idea whether ide dma is possible on ipod
14:56:40 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
14:56:52amiconnDisk IO is always blocking, even with dma.
14:57:16amiconndma would just let other threads run during the transfer
14:57:20Zoide777I'd just like to port recording to rockbox, no matter the mechanism :O
14:57:34Zoide777tooooo much lost sleep already hehe
14:57:43lostlogicamiconn: right...
14:58:06linuxstb_Zoide777: IMO, the best place to start will be to try and understand the playback code - comparing what's in Rockbox with what's in IPL. The same principle applies to recording, just that the transfer of data is in the opposite direction.
14:59:04JdGordonhaha.. uni is gonna kill me :) ive been hogging one of the main remote-access unix boxes for nearly the last 4 hours (my program has been @ 99.9%cpu usage since it started :) )
14:59:19Zoide777i noticed that the ipodlinux pseudo-dma implementation uses constants such as "DMA_READ_OFF" that are set to "/* special locations in fast ram */"
14:59:37 Quit JBGood (Read error: 110 (Connection timed out))
14:59:38JdGordonchange that to nearly 5 hours :)
14:59:55Zoide777for example 0x40000004, 0x40000008, 0x4000000c, 0x40000010. are those in use in rockbox?
15:01:07linuxstb_For playback, IPL uses an 80KB ringbuffer in playback. Podzilla or MPD will put audio in this buffer, and the "fake DMA" function runs on the second processor (the COP) and is constantly trying to empty that ringbuffer. Communication via the main CPU and the COP is via those variables in IRAM.
15:01:20linuxstb_In Rockbox, we do everything on the main CPU.
15:01:45linuxstb_(replace "in playback" in the first sentence with "in IRAM")
15:03:27linuxstb_Zoide777: Have you tried using the recording feature in ipodlinux?
15:04:16 Part LinusN
15:05:30linuxstb_But time for lunch. bbl.
15:06:15amiconnlinuxstb: Do you know why ipl doesn't use dma?
15:06:20Zoide777so, would it work if I just made those constants in rockbox and pretty much pasted over code from "i2s_pp5020_rec_dma" in audio.c (
15:06:59Zoide777well, i *really* gotta go to bed now... thanks for all the help
15:07:23Zoide777feel free to have a try at the recording porting :D ...
15:07:40 Quit Zoide777 ("CGI:IRC 0.5.7 (2005/06/19)")
15:10:44 Join man_in_shack [0] (
15:10:46*man_in_shack waves
15:11:03*JdGordon shoots..
15:11:34man_in_shackfine then
15:11:37*man_in_shack vorbises
15:12:08man_in_shackwondering if rockbox supports 4G ipods. one page says yes, another says no
15:12:33Paul_The_NerdWhich one says "No?"
15:13:16man_in_shack"iPod Color/Photo, the iPod Nano, and the iPod Video"
15:13:35Paul_The_NerdLater in that page it actually talks about it specifically working on the 4G
15:13:58Paul_The_NerdI'm not sure but I think Apple's firmware may still have graphical issues with the 4G bootloader though
15:14:19Paul_The_NerdSo, basically, with 4G Rockbox you have to choose, you can't have both properly. I think. It may have been yet another fix I missed.
15:15:17man_in_shackwell if i end up getting an ipod, i will want to be sure of that before installing rockbox
15:15:33man_in_shackcos i want a portable ogg/mp3 player :D
15:15:41JdGordonrockbox works fine
15:15:44Paul_The_NerdWell, why would you want to go into retail then?
15:16:03Paul_The_NerdRockbox itself works just fine on 4G.
15:16:28man_in_shackwhat you mean retail?
15:16:40***Saving seen data "./dancer.seen"
15:16:47peturapple fw
15:17:12Bg3rBagder ?
15:17:18man_in_shackwill i still need apple fw for downloading files to the ipod?
15:18:31Paul_The_NerdRockbox is primarily filetree based, and iPods have a hardware disk mode.
15:18:32man_in_shackwhat's this about rockboy?
15:18:54Paul_The_NerdWhat about it?
15:18:58amiconnPaul_The_Nerd: The graphical issues with the rb bootloader on 4g grayscale + mini 2g are fixed
15:18:59man_in_shackwhat is it?
15:19:06man_in_shackamiconn, w00t
15:19:08Paul_The_Nerdamiconn: thanks. I'll update the faq then
15:19:23amiconnI fixed that 3 days ago or so, and made new bootloaders
15:19:36man_in_shackipod is looking like a promising option
15:20:05Paul_The_Nerdamiconn: I thought I remembered you making some changes to the display code, but I wasn't sure if they were a fix to it. :)
15:20:41amiconn Check my 00:55 commit :)
15:21:37Paul_The_Nerdamiconn: Now that I see it, I remember reading that too. Ugh. One day I will actually be able to remember things. :)
15:21:46man_in_shackwhat sort of battery life can i expect from rockbox compared to say ipodlinux and apple'
15:21:51man_in_shackapple's estimates?
15:22:30 Join Kohlrabi [0] (
15:23:40Paul_The_Nerdman_in_shack: Well, on the 2G Mini I believe ~75-80% of Apple's estimate was seen (though it was not a new battery.) Is that about right amiconn?
15:23:41 Quit midkay (Read error: 104 (Connection reset by peer))
15:23:51 Join midkay_ [0] (n=midkay@
15:24:23Bg3ramiconn okay, what should i do with these _PLAYER and _RECORDER things ?
15:25:09Bg3rif i deprecate them and make new LANG_IDs... it's like i haven't done nearly anything
15:25:39Bg3rs/i haven't done nearly anything/i have done nearly nothing
15:26:55JdGordonwho was complaiing about the text viewer menu yesterday?
15:27:13 Part obo
15:27:47amiconnJdGordon: That was me
15:28:06JdGordonwhat about it annoys u?
15:28:39lostlogicI want to add "queue_post_single" to the queue functions in the kernel, both for scrolling and now for my audio buffer changes
15:28:51lostlogicany complaints?
15:29:18Bg3ramiconn ?
15:30:02amiconnlostlogic: ?
15:30:14JdGordontisk tisk.. some1 let tabs into viewer.c!!
15:32:07lostlogicamiconn: it will post this event to the queue, only if it is not the same as the last event posted to the queue.
15:32:14amiconnlostlogic: I don't see why this would be needed as a general function. It somehow contradicts the purpose of a queue
15:32:37amiconnThe button repeat is special, and it's already taken care of
15:33:10lostlogicamiconn: but it's not taken care of right... you _should_ be able to queue up a sequence of actions, including one or more scroll events
15:33:14lostlogicand you cannot do that currently
15:33:22lostlogicwith queue_post_single you could
15:36:20JdGordonamiconn: what about the viewer menu do u not like?
15:38:59amiconnI don't like that it pops up when pressing Stop/Off
15:39:08amiconnThat's unintuitive.
15:39:16JdGordonwhere else would u put it?
15:39:27amiconnIt should be called by the appropriate Menu button for the target
15:39:35amiconnStop/Off should exit
15:39:56preglowbtw, i can't find the ^M being talked about in the mp3_encoder.c cygwin bug mail
15:39:58JdGordoni dont know if this will be accepted, but im changing the settings to be settable from the menu, so that could free up buttons so the menu could have a designated button..
15:42:05amiconnThat's something I was planning to do for a long time.
15:42:22JdGordon:) would it be acceptable for pre-3.0?
15:42:54preglowi think so
15:42:54webminduhm... is there a way to enforce charging, or do a hardreset or something on a nano?
15:43:11 Join muesli__ [0] (n=muesli_t@
15:43:18webmindafter a disconnect it doesn't give me anything anymoreafter I copnnect it to usb
15:44:25Paul_The_NerdHolding Menu+Select performs a hard reset.
15:44:39webmindit was locked :)
15:44:53webmindand for some reason it then doesn't go to usb mode either
15:45:02amiconnShould be done better sooner than later
15:45:27amiconnThese cryptic combos are annoying, as I never remember the right ones
15:45:56amiconnThe settings should be put in the menu, and made persistent via config file
15:46:22amiconniirc they already are persistent (???)
15:48:13Bg3ramiconn please, reply to my question: should i mark _PLAYER and _RECORDER LANGids as deprecated and create new ones ?
15:48:42lostlogicamiconn: ok, I'll do it a different way, but we still need a better soln for wheel.
15:49:37amiconnBg3r: Hmm, either that, or bump the lang version properly.
15:49:47linuxstb_amiconn: I can only guess that the IPL people don't know how the DMA works - either because the Apple firmware doesn't use it, or because they haven't r.e.'ed it yet.
15:49:54preglowamiconn: would you care to check if mp3_encoder has any bugs on cygwin? i can't find any carriage returns in the file
15:49:59amiconnIf you go for the latter, you should also modify the strings in the other .lang files properly
15:50:25Bg3ramiconn imho it's better to bump the ver...
15:50:37Bg3ryes, i'll do this too
15:53:17 Join marevalo [0] (
15:53:28amiconnWe still need to solve the ID numbering problem for voice, esp. since we don't bundle voices...
15:54:22Bg3rbut there will be new voice clips for v3.0, yep?
15:56:48 Quit SereR0KR ("XChat Aqua")
15:57:21marevalohi, all
15:57:34marevaloi'm adding opening book to chessbox
15:57:54marevaloand get this on the recorder build:
15:58:01marevalo /storage/H120/sh/lib/gcc/sh-elf/3.4.5/../../../../sh-elf/bin/ld: region OVERLAY_RAM is full (/storage/H120/rockbox-opening/build-recorder/apps/plugins/chessbox/chessbox.elf section .bss)
15:58:22marevaloprobably it's just to big
15:58:24preglowthen it's too big
15:58:29 Join quobl_ [0] (n=quobl@tor/session/x-c353cdad54e7e66f)
15:58:37preglowi don't know how big the overlay can be, though
15:58:44linuxstb_I think you'll need to allocate the memory dynamically, rather than having a static buffer.
15:59:08marevalowell, the point is that dinamically or statically I will be using this memory
15:59:17marevaloit's the opening book
15:59:25linuxstb_How big is it?
15:59:41marevaloI will try to check it
15:59:51marevalohow could I tell you?
16:00:52marevalo27566 2006-04-06 15:56 opening.o
16:01:06marevalo43260 2006-04-06 15:56 gnuchess.o
16:02:13 Join Nico_P [0] (
16:02:35linuxstb_The Archos targets only have 2MB RAM in total. Is gnuchess now requiring a lot of memory?
16:03:18Bg3rmarevalo on the archos targets the plugin size is max 32kb
16:03:21marevaloI think the problem is the overlay, I remeber something about smaller limits on the archoses
16:03:21Paul_The_NerdAnd the iFP-790 will only have 1mb, if it ever resumes progress.
16:03:51marevaloI will try to estimate the size increment with this version
16:04:21 Join MarcoPolo [0] (
16:05:38linuxstb_My undertanding is that the overlay is a small plugin which loads the main plugin directly into the audio buffer. So the limit should be around 1.5MB. Your changes shouldn't be changing the overlay plugin.
16:05:43 Join RoC_MM [0] (
16:07:08marevalowell, by the moment I'm just checking memory use increment on this version
16:07:42marevalobut the opening book is just: unsigned short OBook[241][56];
16:08:23 Join bluebrother^ [0] (
16:09:43marevaloa H120 build, with openings:
16:09:56marevalo62692 2006-04-06 16:03 rockbox-opening/build-h100/apps/plugins/chessbox.rock
16:10:00marevaloand without:
16:10:08amiconnmarevalo: If the overlay ram on recorder isn't big enough, it is possible to extend it. I just made it a bit bigger than what was needed
16:10:15marevalo35560 2006-04-06 16:07 rockbox-all/build-h100/apps/plugins/chessbox.rock
16:10:19amiconnOpening adds quite some code + data
16:10:57marevalowell, openings didn't add any code, it was still there and compiled (silly me)
16:11:03amiconnIt's not only the binary size that counts. You need to add the .bss size. Check the .map fie
16:12:06marevalowhat should I check?
16:12:23 Join linuxstb__ [0] (
16:12:49linuxstb_amiconn: So the "overlay RAM" is a fixed-sized part of the audio buffer?
16:13:46amiconnIt is calculated downward from audiobuffer_end
16:13:58marevalowell, I tried to decrease the memory size of the opening book, but my C foo wasn't quite enough
16:14:04 Quit linuxstb (Read error: 110 (Connection timed out))
16:14:10amiconnThat's because an overlay is linked exactly like a plugin, just with a different start address
16:14:26marevaloan opening is something like this:
16:14:29marevalo */
16:14:29marevalo {
16:14:29marevalo 0x0C1C , 0x3424 , 0x0615 , 0x392A , 0x051A , 0x3D22 , 0x0B13 , 0x3E2D ,
16:14:29DBUGEnqueued KICK marevalo
16:14:29marevalo 0x0112 , 0x332B , 0x0226 , 0x372F , 0x262D , 0x3B2D , 0x1223 , 0x2D3B ,
16:14:32 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
16:14:33marevalo 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 ,
16:14:34marevalo 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 ,
16:14:36marevalo 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 ,
16:14:38***Alert Mode level 1
16:14:38marevalo 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 ,
16:14:40marevalo 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000 , 0x0000
16:14:42marevalo }
16:14:52marevaloso there is some memory overuse
16:15:11linuxstbamiconn: OK, it's clear to me now. How big is it currently?
16:15:51amiconnmarevalo: In your recorder build dir, check apps\plugins\chessbox\
16:16:16amiconnOVERLAY_RAM should start at 0x091d8000 currently
16:16:32amiconnCheck the address of _plugin_end_addr
16:16:55marevalo0x00000000091fb4ec _plugin_end_addr = .
16:17:14amiconnAha, so it's just a bit too big
16:17:24marevalojust a bit?
16:17:27amiconnIs that with some things disabled you would like to enable?
16:17:38amiconnYes, less than 16KB
16:17:53 Join aXp [0] (
16:18:06aXphi all
16:18:07marevalowell, I can save more or less 50% of the memory the opening book is currently using
16:18:28marevalobut I would need some help with my faulty C
16:18:57aXpi now installed rockbox on my ipod nano but i cant use the itrip for the ipod nano(griffin) anyone has it got working?
16:18:58amiconnI'm not talking about disabling things. I rather want to estimate how much we should increase the overlay area for chessbox on archos
16:19:17Paul_The_NerdaXp: No.
16:19:26amiconnSince it is still fairly small, we could just give it some more than absolutely necessary
16:19:27linuxstbaXp: No, Rockbox doesn't support accessories that communicate with the ipod.
16:20:00marevalowell, apart from this change (openings) the other thing I should add is just a little bit of code before my target for the freeze
16:20:00RoC_MMThe MONSTER line of FM transmitters work.
16:20:40marevalothink on your time is already on the code, I just should find a clever way to activate it on a uni-thread environment
16:20:55marevalobut it shouldn't be so much code
16:21:44marevaloso 56 or 64k should be more than enough
16:22:25marevaloI can tell you that I will try to use also any remaining memory for the transposition table once the features are finished
16:23:31amiconnmarevalo: Open apps/plugins/chessbox/, and change line 11 to
16:23:35amiconn#define OVERLAY_LENGTH 0x30000
16:23:55amiconn(192 KB instead of 128KB, i.e. 64KB more)
16:23:56JdGordonok, ive moved all the viewer options to the menu.. should i remove all the key combos to set the otions and set 1 key to show the menu? or keep the combos and change 1 to show the menu?
16:24:19amiconnPlease remove all option combos
16:24:31JdGordon.. done and done
16:24:31amiconn(but not the scrolling)
16:24:39***Alert Mode OFF
16:24:58JdGordonye, ive added a new option to change the default to scroll by line or page..
16:25:10JdGordondefault being page
16:25:21amiconnmarevalo: You could also change the comment at the top. It still says 'rockboy'
16:26:19marevalobuilt ok
16:26:22 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
16:26:23 Join JBGood [0] (
16:26:37marevaloshould I add this change to the finished patch?
16:32:16 Quit JBGood25 (Read error: 110 (Connection timed out))
16:32:37 Join warthawg [0] (
16:35:11amiconnmarevalo: yes
16:35:48lostlogicbagh, my new buffering stuff seems great... until it starts miss-buffering and causing codec failure all over the map. but now, to work.
16:35:56marevaloamiconn: ok, thanks, will send a patch this afternoon once checked on a couple of targets
16:39:42linuxstbAre there any Etymotic ER4P owners here? Cassandra?
16:40:03 Quit MarcoPolo (Read error: 104 (Connection reset by peer))
16:41:40 Quit ep0ch (Read error: 110 (Connection timed out))
16:42:08 Join goffa_ [0] (n=goffa@
16:46:22 Quit RoC_MM ("Leaving")
16:50:00 Quit JdGordon (Read error: 104 (Connection reset by peer))
16:52:23 Join damaki__ [0] (
16:53:45CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
16:53:45*petur wonders if there's an STL guru lurking around
16:55:34warthawgi may look like i am playing doom on my nano, but in reality, i am researching
16:56:23Mikachui tried to run a timedemo but i got a data abort instead
16:57:57linuxstbCan you replicate it?
16:58:06goffa_he he he.. playing doom seems to be the thing to do around here lately
16:58:36Mikachuhm, didn't kkurbjun say he fixed the menu select thing on ipods?
16:59:03Mikachumy graphics are very corrupt and stuff too, before it crashes
16:59:08Mikachuso maybe i have a strange build
16:59:48preglowpetur: as in c++?
17:00:23peturyeah... was wondering if one is allowed to compare iterators with reverse_iterators
17:00:54peturbut already changed it to a normal one to be safe ;)
17:01:22nudelsee if there's an operator< defined which takes both as an argument
17:02:14peturnevermind - no time to experiment
17:02:20nudelfair nuff :)
17:02:29nudeli'd probably do the same thing
17:02:42peturworkload ;)
17:03:20Mikachuhrm, now that i think, it's strange that doom ever worked here...
17:03:27Mikachuhow much ram does it need, and where does it allocate it from?
17:03:32linuxstbI was just thinking the same - with your tiny audio buffer.
17:03:38linuxstbThe audio buffer.
17:03:52Mikachuis 26MB enough?
17:04:03linuxstbI thought your plugin buffer was 26MB?
17:04:06preglowshould be ample, i think
17:04:16Mikachuyes, i was thinking about changing it to use the plugin buffer
17:04:44linuxstbYes, as preglow said, I'm sure 26MB is ample.
17:05:07preglowwhy the hell is your plugin buffer so big?
17:05:29Paul_The_NerdI was about to ask something similar, though with less invocations of afterlives. I was curious what benefit there is since all the plugins fit anyway.
17:05:50Mikachui wanted to view a large jpeg while listening to music :P
17:05:59 Join DJ_Dooms_Day [0] (
17:06:02Mikachuwell, now it says allocated 26000kB zone buffer, but the same thing happens
17:06:07preglowhighres map of the world?
17:06:14Mikachubut i'm not going to make you run around and look for things when i've changed so much
17:06:35 Quit Nibbler (Remote closed the connection)
17:06:39Mikachupreglow: the jpeg plugin needs to fit the whole uncompressed image
17:06:42Paul_The_NerdOdd, I think Doom only requires 4mb
17:06:47preglowMikachu: know
17:06:52Paul_The_Nerd8mb for Doom 2..
17:07:01Paul_The_NerdIf I remember well.
17:07:03preglowit doesn't require much
17:07:12preglowi ran it on machines with waaaay less then 32 megs ram
17:07:41Paul_The_NerdIt's either 2/4 or 4/8 I think
17:07:46 Quit damaki_ (Read error: 110 (Connection timed out))
17:07:53Mikachui'll just wait passively a couple of days
17:09:12preglowMikachu: sounds like you'll be doing more looking at the progress bare than viewing pictures anyway
17:09:30Mikachuyeah, it takes a while to decode
17:10:05Mikachui guess i have a big plugin buffer mostly out of spite
17:10:12 Quit darkless (Client Quit)
17:10:26preglowso that's why you made that plugin buffer clearing patch
17:10:33Mikachuyes :)
17:10:42Mikachuit took like 10 seconds to load a plugin before
17:10:50Mikachubut it still speeds up loading with 512k buffers
17:10:56Mikachuor maybe that was placebo
17:11:56 Join Nibbler [0] (
17:12:25Mikachui cvs up -D'ed to a couple days ago and it seems to run fine
17:13:43preglowi can't think of any reason we shouldn't commit it
17:13:53lostlogicSlasheri: ping?
17:15:13Slasherihi :)
17:16:25lostlogicSo, I'm totally abusing the playback engine... and I have a question... In design, what is the purpose of the track_count variable?
17:16:41***Saving seen data "./dancer.seen"
17:17:05Slasherihmm, i will check the code if it has a real meaning.. just a moment
17:17:22 Quit bluebrother^ ("Leaving")
17:17:52lostlogicSlasheri: thanks :) I think I've finally got a handle on the _buffer_ handling, but the track struct handling is still difficult :-\
17:18:12Slasheriyeah, it's used to prevent the ring structure overflowing.. and when clearing the buffers so correct track entries are cleared
17:18:34lostlogicSlasheri: ok, I think it may be my next victim (muahahaha)
17:18:57Slasheri:D i hope you do some extensive testing after changing the code..
17:19:33Slasheribecause even the smallest changes like to break some parts from the engine :)
17:19:50Slasheriso you are thinking about converting that ring structure to a dynamic list?
17:21:24lostlogicSlasheri: not yet... but I prefer to do things in a mathematically correct way, then see what I've forgotten (when Paul_The_Nerd and ep0ch report) and then fix my math and try again...
17:21:25 Join webguest54 [0] (
17:21:41lostlogicit's impossible to test every corner case and weirdness, but it's possible to mathematically capture them all :)
17:22:19lostlogicSlasheri: it's broken in CVS already, and will probably continue to be so for the next week or so as every major change will fix some things and break others until it restabilizes.
17:22:38Slasheriwell, yeah. but you are going to remove that track_count variable?
17:22:54lostlogicSlasheri: it is likely that it will fall victim to a refactoring.
17:23:12lostlogicSlasheri: but now that I think of it, that might not happen until the tracks become a linked list.
17:23:44Slasheriyep, it shouldn't.. and probably not even after the linked lists are implemented
17:23:57Slasheribecause it's always good to know how many tracks are currently loaded in the buffer
17:24:30lostlogicI can't think of a good reason with linked lists to know how many tracks are in memory, only that there are tracks, and whether or not each track is in the ringbuffer or not.
17:25:42 Quit Paul_The_Nerd ("Leaving.")
17:25:44lostlogicbut any way, I'm not there yet. right now I need to figure out why my work to fix some track skipping errors has decided to break buffering completely.
17:26:49lostlogicI was hoping you'd tell me that track count was a magical variable and that would tell me that it was the probable cause for my current effedupness
17:27:57Slasherihehe, there shouldn't be any magic with that :) unless you have already touched it
17:28:16Slasheribut current implementation absolutely requires that track count
17:28:21lostlogicnope, haven't done anything to it yet.
17:29:01Slasherior maybe not "absolutely" but it makes things easier and cleaner
17:31:28 Quit JBGood (Read error: 110 (Connection timed out))
17:31:30lostlogicyah, I understand how it is used now.
17:34:14 Join JBGood [0] (
17:42:28 Quit theli_ua (Remote closed the connection)
17:44:24webguest54hello guys
17:44:55webguest54I installed the bleeding edge build and playback engine is completly crazy
17:45:01lostlogicSlasheri: what happens if you skip backwards to a track that was set to use the track earlier than itself in the buffer's codec? Force buffer refill to get the codec?
17:45:24lostlogicwebguest54: yes, that is likely, it's undergoing a complete sandblasting.
17:45:29webguest54every few seconds HD activity.... + no skiping possible, are those problems for all?
17:46:05webguest54lostlogic: will you fix this *soon* or need do dowgrade?
17:46:06lostlogicI've been rewhacking major parts of it. I think the version that is in CVS right now will be buffering far far too often.
17:46:13Slasherilostlogic: yep, buffer refill is forced if that happens
17:46:30webguest54just few seconds buffered
17:46:45lostlogicwebguest54: soon is relative, I use the software daily and the version that I'm working on is worse than CVS so I'm actively working on it, but I'd guess I won't have another commit until afternoon (CDT)
17:46:55webguest54all the buffer isn't used, it's why needed HD activity so often
17:47:06lostlogicwebguest54: yes, of course.
17:47:28webguest54ok, I can wait then, thanks for work on it
17:47:37lostlogicwebguest54: cool
17:48:13webguest54will be the playback engine and pcm buffer, better after those reworks? I assume yes
17:48:24lostlogicthat is the hope, of course :)
17:48:34webguest54that's great then :)
17:48:36lostlogicI'm not touching the pcmbuffer at this time though.
17:48:46lostlogicI already abused it a few months ago
17:49:39*webguest54 will not use his iriver until you fixe the problem
17:50:00amiconnMikachu, preglow: Buffer clearing patch? What's that do?
17:50:19lostlogicI'll do my best, however I am currently using my iPod with the issues *shrug* just hafta fiddle with it sometimes to keep it playing.
17:51:23webguest54lostolgic: cheers for this *no-payed* hard works
17:51:55lostlogic*hides under his desk* I'm being payed... sorta... in that I'm at my _paying_ job doing this work :-P
17:52:04webguest54time for me to go and keep my DAP in my bag since I can't use it with this build and no usb cable here
17:52:24webguest54lostlogic hahaha XD, great plan
17:52:42 Quit webguest54 ("CGI:IRC")
17:55:50linuxstbamiconn: It was something to do with not memsetting the entire plugin buffer. It's here:
17:56:34 Join theli_ua [0] (
17:58:48marevalook, the Chessbox opening book patch is at the tracker:
17:59:18marevalono more boring games ;-)
17:59:42 Quit theli_ua (Remote closed the connection)
18:00:03amiconnlinuxstb: Hmm, while I don't think this is absolutely necessary (I never noticed a delay resulting from the buffer clearing), it shouldn't hurt to include this. The change is minimal
18:00:16 Quit Nibbler (Connection timed out)
18:00:18amiconn(and it could in fact be applied to the overlay loader as well
18:01:18lostlogicwhat happens on a double close in rockbox? ie fd = open(); close(fd); close(fd);?
18:01:42 Join JBGood25 [0] (
18:01:53amiconnThe second close() will find the file descriptor already closed and terminate
18:02:15amiconnNot closing a file is far worse
18:03:12amiconnLeaking file descriptors, and their number is fixed...
18:03:32lostlogic*nod* is there any reasonable chance that the same file descriptor would have been assigned to another caller to open() in the meantime and thereby causing another application to get a closed descriptor suddenly?
18:03:48lostlogic(I'm planning to prevent this nonetheless, but was just wondering)
18:05:11 Quit petur ("-poof-")
18:06:51amiconnHmm, now that you mention it.... that might happen
18:07:23 Join _Lucretia_ [0] (
18:07:31amiconnYou would then close the other thread's descriptor, with the effect that the other thread's read(), write() etc calls will al fail
18:08:26Mikachuamiconn: i've had the patch applied for a month or two, no problems
18:08:43 Join Nibbler [0] (
18:09:10amiconnlostlogic: (with EBADF)
18:09:18 Join Lamed [0] (
18:10:01Lamedhello everybody
18:10:31lostlogicso basically, it's important to 1) not leak descriptors and 2) not double close if another thread has run in the meantime
18:11:22amiconnThe mpeg thread on archos has a simple mechanism to ensure this
18:12:10LamedI'm gonna be off an internet connection for a long while. can someone suggest me a way of downloading everything from rockbox's site? (the www package is obsolete btw)
18:12:13amiconnThere's a global fd variable (the mpeg thread opens only one file at a time) which is set by open(), and reset to -1 after close()
18:12:30amiconnSo, closing means
18:12:35lostlogicamiconn: yeah, same in the swcodec, except I think I abused it this morning.
18:13:48Kyomi|offHey you all
18:13:52 Nick Kyomi|off is now known as Kyomi (
18:13:54 Join tianjing_ [0] (
18:13:58KyomiI have a question about the Doom game
18:14:06KyomiWhat happened to the strafe key?
18:14:34 Quit JBGood (Read error: 110 (Connection timed out))
18:14:42linuxstbYou should now be able to define your own keys in Doom.
18:14:51Kyomi*goes to update it*
18:15:04 Quit _FireFly_ ("Leaving")
18:15:07Mikachui'm trying to find the commit that broke it here
18:15:13luckzis %ik for displaying discnumber in the current CVS build?
18:15:16 Join RotAtoR [0] (n=e@rockbox/developer/rotator)
18:16:08 Quit tianjing (Read error: 110 (Connection timed out))
18:18:01KyomiI'm guessing Doom is updated in the experimental build now too?
18:18:09KyomiOne more question
18:18:16lostlogicI don't think anyone is updating the experimental build any more...?
18:18:23MikachuKyomi: do you mean cvs build?
18:18:30KyomiIs there an option like... "always strafe"?
18:19:53linuxstblostlogic: I think paulheu has taken over the H300 experimental build.
18:20:05lostlogicah, my mistake.
18:20:10Kyomi :)
18:20:18Mikachufound it
18:20:25linuxstbSomeone is also doing a h1x0 "optimized build"...
18:20:31Mikachu* d_englsh.h, doomdef.h, m_menu.c, m_misc.c, rockdoom.c: Fix up configurable keys, edit exit string, more work needs to be done on menu keys
18:20:54Mikachudoes 0063bec8 tell you anything?
18:21:37 Quit swampcow ("told you...")
18:22:39KyomiOh well..
18:22:40lostlogicalright, latest CVS is much less broken, although skipping backwards seems to be unreliable for some ungodly reason. it's starting to feel under control.
18:23:12Mikachubut that doesn't make any sense... why would that break the graphics?
18:23:23Mikachui wonder if cvs update -D "clock" is my time or cvs server time
18:23:46Mikachulooks like it's my time though, so it is that commit
18:23:49linuxstbYou can specify a timezone in "clock"
18:23:53linuxstb(I think)
18:24:40*preglow kicks gcc
18:24:45Mikachui don't suppose i can make cvs diff a directory between two dates?
18:25:09linuxstbYou can do it easily with the web interface.
18:25:19linuxstb(but only on single files...)
18:25:24Mikachuyeah, i looked at those
18:25:36linuxstbI guess you just have to download two copies and do a diff.
18:25:48preglowamiconn: doesn't exactly look like the 64 bit issue will see a quick resolution
18:25:55linuxstbMikachu: What's the problem you're trying to find?
18:26:07Mikachulinuxstb: data abort and broken graphics in the two seconds before that
18:26:37lostlogic diff [-kl] [rcsdiff_options] [[-r rev1 | -D date1] [-r rev2 | -D
18:26:37lostlogic date2]] [files...]
18:26:43amiconnpreglow: How do you know? I didn't get any mail so far regarding this bug...
18:26:50lostlogicwhy wouldn't you be able to diff between two dates?
18:26:51Mikachuahah, there i go
18:27:01Mikachui just assumed it wouldn't work :P
18:27:03preglowamiconn: because the only guy who seems to be interested in fixing it doesn't seem to have time to fix it
18:27:19lostlogicthat would be one shitty SCM tool.
18:27:36Mikachucvs is not in so high regard with me
18:28:17lostlogicCVS may not be the be-all, end-all it used to be, but it's still a functional and robust tool...
18:28:47lostlogicI still gotta find some time to learn and play with git myself.
18:28:53luckzhow do I know if some patches already made their way into the cvs?
18:29:03preglowluckz: they should be closed in the tracker
18:29:14lostlogicluckz: patch −−dry-run < patchfile to see if it says "reversed"?
18:30:26luckz <- so this one isnt in. is there any easy way to get it? is quite necessary for me unless I want to re-tag a huge lot of music for portable use.
18:31:07lostlogicSlasheri: let me know what you think of my last 3 or 4 commits to playback.c, combined they are a pretty major rehack, but I think they pave the way for us to stabilize this thing for 3.0.
18:33:07Lamedguys: any chance of recommanding me something so i could download everything from (html, that is mostly)
18:33:23Mikachuwhy would you want to?
18:34:03amiconnpreglow: For some reason I start to dislike gcc. I have now filed 2 bug reports, which seem to be ignored more or less.
18:35:34Lamedmikachu: i'll be off an internet connection for a while, and i want to be able to view the wiki and stuff for information
18:35:46linuxstbluckz: That seems a reasonable patch to me, but I think it's up to Slasheri to add it to tagcache - tagache is his baby. But remember that we are currently in a feature freeze in preparation for releasing 3.0.
18:36:02preglowamiconn: they aren't known for their ability to fix bugs
18:36:15MikachuLamed: wget -r should work
18:36:34preglowamiconn: you could join their irc channel and ask if someone wants to fix it, i've done it, but like i said, only one guy seems to care
18:36:39preglowperhaps more people nagging will help
18:37:24preglowtheir dev channel is
18:37:50 Quit linuxstb ("Client Exiting")
18:37:55Lamedmikachu: how does wget handels files? what will i end up with..?
18:38:16MikachuLamed: ? you'll end up with lots of files
18:38:25Mikachuyou may want to do wget -kr now that you mention it
18:41:21Slasherihmm, that disc number patch looks ok, but i don't know what others like about committing that during the feature freeze
18:41:56Lamedmikachu: yeah, I should... but will it download .pdf s, .exe 's ?
18:42:00luckzSlasheri: how would I go about using that and on my player?
18:42:26 Join bluebrother^ [0] (
18:42:31MikachuLamed: -A html,htm
18:43:00Mikachulinuxstb__: this is the commit that breaks doom for me,
18:43:05Slasheriin fact both of the patches looks good. But the tagcache patch requires some modifying to apply it to the latest cvs
18:43:16Mikachulinuxstb__: updated all files to newest cvs and reverted that one, and everything works (except opening the menu of course)
18:43:26Slasherias the tagtree codes has changed a bit
18:43:51lostlogichmph, I've managed to reduce playback.c by 12 functional lines, but overall increase by 20 lines, must break that bad commenting habit I've developed.
18:44:07luckzwhat tags are %rr and %rp actually? %rp at least doesn't seem to be the same as %PLAY_COUNTER% in foobar.
18:45:14 Quit marevalo ("Abandonando")
18:45:18Lamedmikachu: thanks! i've found a windows wget binary build and it's working rather nice :) but i'm very sceptic if this would ever finish in a reasonable time :/
18:45:29Mikachuheh, yeah it'll take a while
18:46:28luckzSlasheri: I assume that means I should either develop insane coding skills or just give up on it for now, correct?
18:46:46MikachuLamed: also note links won't be made relative until it finishes, so if you press ctrl-c they're unchanged
18:47:48Slasheriluckz: well, that requires only very little code changes.. i think you can probably find out what to do if you look at the code carefully and what has changed and how =)
18:48:14 Quit aXp (""need to save the world ..."")
18:48:55 Quit thegeek (Read error: 104 (Connection reset by peer))
18:52:45luckzwhat's easiest and less hasslesome for coding on windows - cygwin, colinux or vmware?
18:52:57 Join webguest83 [0] (
18:53:59Lamedmikachu: kinda silly, it's downloading the huge irc logs... and it stays as a background procces if i'll ctrl+c it. :/ what's the -c switch any good for?
18:54:15Mikachu-c should resume broken downloads
18:54:20Slasherilostlogic: well, with a quick look through four past diffs combined, i didn't see anything too major. Just some comments were funny and scary, like "wind the buffer...", but then i saw there wasn't actual code changes but just the comments =)
18:54:27MikachuLamed: it would also be a good idea to start in /twiki and give -np
18:54:41Slasherilostlogic: but of course something might be still broken
18:55:41Lamedmikachu: my bad, ctrl+break puts in in the background. there's no way to skip to next file isnt' it?
18:55:41 Quit webguest83 (Client Quit)
18:55:48MikachuLamed: i doubt it
18:56:03MikachuLamed: i think you can give patterns of filenames to ignore though
18:56:34 Join RoC_MM [0] (
18:57:51Lamedmikachu: wget -r -np
18:57:53Lamedfor now
18:58:08Lamedi can exclude extentions with -R
19:07:17Slasherilostlogic: hmm, the latest cvs seems to be very broken by the playback..
19:07:20 Join borges_ [0] (
19:07:36amiconnLamed: IRC logs are just ~90MB
19:07:37Slasheripause is very laggy and playback continously stops/skips
19:07:48 Join RedBreva [0] (
19:08:04Slasherieven after i pressed stop, i still get some continous random noises
19:08:18Slasheriand suddenly playback starts for five seconds until buffer runs out! :D
19:09:04Slasheriit seems that pause doesn't respond at all while buffering..
19:09:31 Join Paul_The_Nerd [0] (
19:10:05borges_hi there, just wondering: can you make the ipod nano louder with rockbox than it actually is with the default firmware?
19:10:24Paul_The_NerdI think so, yes.
19:10:24Lamedamiconn: i don't have any use of them, not that bored after all
19:10:32preglowborges_: yes
19:10:37Paul_The_NerdThough going above 0 can result in clipping
19:10:45Slasheriand pressing stop causes really weird effects.. it takes over 10s for the playback to actually stop (i have a large crossfade buffer)
19:10:56preglowPaul_The_Nerd: retailos is capped at far lower than 0db
19:11:25Mikachuthey released a new version where you can set a limit, but i don't know how high the max is there
19:11:30Paul_The_Nerdpreglow: I was just clarifying on the 0 thing before there's any "Rockbox distorts my music" posts. :)
19:11:51Paul_The_NerdMikachu: The limit you can set is downward from the current Retail cap, I believe
19:11:59Slasheriwell, it seems there are still a long way for a stable playback now... =)
19:12:44 Join actionshrimp [0] (
19:13:06borges_so the volume limit isn't in the hardware or sth., right?
19:13:40preglowborges_: correct
19:13:48Mikachulinuxstb__: heh, i just needed to remove defaults.dfg
19:13:56borges_cool. i think i'm gonna buy the nano with intent of playing around with rockbox. from what i've seen this really seems to be what i'm looking for :D
19:14:00Lamedhow can i tell the size of a site's folder? say /documentation?
19:14:50Paul_The_Nerdborges_: That is the best way to do it. The Nano with Rockbox is exceptional if you don't go into it with a desire to see a reproduction of how Apple works. Everything Retail does is there, and a lot more, but it doesn't have the polish to the interface some people expect.
19:14:54MikachuLamed: download it and check...
19:15:08Slasherilostlogic: there seems to be even buffering errors.. (errorneus glitches during playback)..
19:15:28 Quit quobl_ (Remote closed the connection)
19:15:31MikachuLamed: does it have an index?
19:15:50preglowPaul_The_Nerd: we aren't exactly owning retailos in the database department yet
19:15:57Lamedmikachu: DocsIndex?
19:16:16MikachuLamed: hm, i guess you could download just the http header for each file referenced
19:16:21Mikachubut i don't think wget has a switch for it
19:16:33Lamedwget can spider
19:16:43***Saving seen data "./dancer.seen"
19:17:00Lamedbut i think it doesn't knows the file size until the time it finishes downloading it
19:17:20Mikachuit knows the file when it starts downloading if you look closer
19:17:21borges_nah, i don't care about the default interface.
19:17:33borges_i just love (and need) those extra codecs and features like directory structure, mass storage etc.
19:17:49borges_plus i like to customize things like *i* want them to be.
19:18:16Paul_The_Nerdpreglow: Yeah, but other than "Play All" buttons at each level, we basically offer what they do.
19:18:27Lamedmikachu: i've looked. are you certain?
19:18:45Paul_The_Nerdpreglow: Besides, I said "everything it does" is there. I consider that more of an interface difference, since from the filetree you can do a lot more anyway.
19:19:18 Join subwayjared [0] (
19:19:39MikachuLamed: hm, but it looks like rockbox' web server doesn't tell sizes
19:20:03preglowlostlogic: wasn't there some patch to make gcc use addq and subq more often?
19:20:16Paul_The_NerdQ: Should rockbox be able to play all WAV files now, or are there still unsupported things?
19:20:32crashdLamed: it cant guesstimate how big a website will be, as it doesnt know until it's spidered it, and by then you have it all alreayd, if you get me
19:20:38Lamedyepp, that's what i ment. but somebody here outa know this kind of stuff won't they?
19:20:50MikachuBagder would know
19:20:53crashdwell, if you ask whoever the web admin is, he could find out
19:21:11crashdthere was a 'nicely' in there somewhere ;)
19:21:44Lamedcrashd: i think wget can get file sizes if the server tells them, but doesn't seems to support them iiuc
19:22:33crashdLamed: works fine here
19:22:40crashdLength: 2,332,711 [application/zip]
19:22:47crashdfor the CVS build of ipodvideo rock
19:22:53crashdmight be your win32 build tbh
19:22:59Lamedcrashd, what switches?
19:23:03crashdtry another one, there's various ones based on different branches
19:23:06crashdwget <url>
19:23:30Lamedcrashd: acctually i've thought about it. it was compiled in 1999
19:23:30crashdas i say, there is 2 main branches of wget, either that or they changed their interface at some point
19:23:46 Quit pussfeller ("Leaving")
19:23:46crashdone does a text percentage bar, and one doesnt
19:23:52crashdtry getting a newer build :)
19:23:53crashdwhat ver is it ?
19:24:19crashdhas a dec 2005 build
19:24:52crashdhmm, that doesn't seem to work
19:25:02Lamedyepp, i've just g00gled it myself
19:25:30crashd1.9.1 is the version on my bsd box
19:25:43 Join pussfeller [0] (
19:26:10crashdhmm, might be something to do with windows then, rather than the build
19:26:20crashdwhat are you trying to do, anyway ?
19:26:22 Quit imphasing ("New xterm!")
19:26:55 Nick t0mas is now known as t0mas| (n=Tomas@rockbox/developer/t0mas)
19:27:06 Join imphasing [0] (
19:28:16 Quit JBGood25 (Read error: 110 (Connection timed out))
19:28:29Lamedcarry out as much information as i can from the site before i'm going off the cyberweb for a week, so i could read and learn
19:28:45crashdfrom the wiki?
19:29:36Mikachuthe html pages don't have sizes
19:29:37 Quit t0mas| ("Doei")
19:29:41Lamednew wget tells me the file size, but it won't work -r <recursively> and i still can't tell a complete folder size ( ex)
19:29:43Mikachupresumably because they're generated
19:30:13Lamedmikachu: apperently they do
19:30:14Lamed20:28:24 (33.47 KB/s) - `' saved
19:30:20crashdarent the wiki pages actually static, just you can edit them via a 'dynamic interface'
19:30:23Mikachuthat's a speed
19:30:33Lamedmaybe it's just stated after the download was completed
19:30:39Mikachuthat doesn't help you a lot
19:30:53crashdwhy dont you just set it spidering and leave it?
19:31:00crashdit's only text anyway, shouldnt take too lon
19:31:06Paul_The_NerdCorrect me if I'm wrong, but you shouldn't have to adjust horizontal and vertical position of the image on an LCD monitor should you?
19:31:09crashdor ask the webadmin to tar up the wiki directory and get it all in one hit
19:31:17MikachuPaul_The_Nerd: i don't think so
19:31:19crashdthus reducing the load you're creating by using websuck :)
19:31:28Paul_The_NerdMikachu: That's what I thought. I'm trying to figure out why I can on mine.
19:31:33MikachuPaul_The_Nerd: since there's no electron beam, it should know where the image is going
19:31:53Mikachuespecially if it's dvi stuff
19:32:11Paul_The_NerdWell, it's not dvi. I suppose that might be it.
19:32:37Mikachumaybe it's possible for the graphics card to send the data so it's unaligned then
19:32:41Paul_The_NerdIt's an old bugger. I actually have it hooked up to my PS2 by way of this horrible little adapter, so it's even worse.
19:33:03Lamedcrashd: i _am_. it's currently harvesting the wiki. i was wondering how and if it can be done.
19:33:26Mikachui guess google does it every day
19:33:32Mikachu(probably not day)
19:34:08 Join raf82 [0] (
19:34:43 Quit nkh (Read error: 110 (Connection timed out))
19:34:57Lamedfor a different subject: how do i attend files under VMware?
19:35:28crashdMikachu: google spiders run a weekly basis, or at least they used too
19:35:37crashdthat doesnt mean a site will get reindexed once a week tho
19:35:40Lamedi've downloaded a cvs tree. what then?
19:36:09crashdwhat do you want to do?
19:37:41Lamedanyone is using the VMware machine?
19:37:57*Paul_The_Nerd mutters.
19:38:42Paul_The_NerdI hate this monitor. Just because
19:38:49Paul_The_NerdLamed: What exactly are you having problems with?
19:39:07crashdwhois Llorean, on the forums?
19:39:12 Join RedBreva_ [0] (
19:39:26Lamedwhoops, sorry. i'm trying to edit files with an external editor.
19:39:36Paul_The_NerdCrashd: Before I answer: Negative or positive reason for asking? ;-)
19:39:51crashdno reason, ive always wondered, as he reminds me of you Paul, and ive never seen a Llorean on irc
19:39:54crashdis all =]
19:39:57Paul_The_NerdLamed: The vmware machine can be seen shared on your network, and accessed as \\Debian usually
19:40:36Paul_The_NerdYeah, 'tis me. I signed up for the forums before I started regularly hanging out in here, but I *always* use this on IRC channels.
19:40:45crashdfair deal, makes sense then
19:40:50 Quit RedBreva ("Chatzilla 0.9.72 [Firefox]")
19:41:11Paul_The_NerdAnd of course I'm something else again on the patch tracker, simply because that came from my Sourceforge ID.
19:42:00 Nick linuxstb__ is now known as linuxstb (
19:42:20linuxstbamiconn: I read in #ipodlinux-dev that the 3g also suffers from the mirrored display issue in RetailOS - so it seems your fix was needed for that as well.
19:43:32 Quit _Lucretia_ ("Leaving")
19:44:20 Join theli_ua [0] (
19:45:16 Join t0mas [0] (
19:45:23 Quit t0mas (Client Quit)
19:45:52Lamedhey! look at that! thanks Paul!
19:46:44Mikachulinuxstb: you saw that i only had to remove the dfg file?
19:46:53amiconnlinuxstb: Is that with the ipl bootloader?
19:47:22linuxstbamiconn: Yes, with ipodloader2.
19:47:43amiconnThey should switch format, like I did for rockbox...
19:47:52linuxstbI've pointed them in the direction of your changes. Not sure if anyone will implement it though.
19:47:52Mikachuwhen they get to 3, they can call it tripodloader
19:48:17*linuxstb groans
19:48:31 Join wefds [0] (n=muesli_t@
19:48:38linuxstbIt can already triple-boot Rockbox, IPL and Apple.
19:49:10crashdive never had much luck with loader2 though
19:49:12*amiconn can't
19:49:35 Join Bizmo [0] (
19:49:43Bizmohow can i apply a patch to my 5g ipod?
19:49:49amiconnipodlinux worked exactlly once for me, subsequent rebbots just hung.
19:50:07crashdipl works ok, but trying to boot it from loader2 just ended in tears, hanging on the menu, hanging on the loading bar
19:50:11crashdgeneral hanging, really
19:50:32*amiconn doesn't really care
19:50:49crashdgrumpy :>
19:51:11amiconnI don't need ipl on my ipod. That doesn't mean ipl is useless, the sources are valuable for rockbox
19:51:14Paul_The_NerdBizmo: You apply the patch to the source code, and then compile a 5G build. Have you read the wiki pages relating to how to compile your own build?
19:51:31Mikachuamiconn: but it would be better if they joined rockbox instead :)
19:51:32 Quit Bizmo (Client Quit)
19:51:33Lamedmazal-tov, i got the wiki!
19:51:58preglowMikachu: you'll be thankful they exist when you one day want to run apache on your ipod!
19:52:14 Join Bizmo [0] (n=shadowsw@
19:52:15crashdsurely squid, get some nice webcaching going on
19:52:16Mikachuyes, i will stream mp3s over the lineout attached to my computers serial port
19:52:28BizmoHow can I apply a .patch file?
19:52:34Mikachuread said wiki page
19:53:08Paul_The_NerdBizmo: There's a wiki page entitled I believe "Working With Patches" that covers how to use them.
19:53:23 Join Rondom [0] (n=Rondom@
19:53:33 Join t0mas [0] (
19:53:41linuxstbBizmo: Basically: 1) Install a development environment; 2) Compile an unmodified version Rockbox to test your development environment; 3) Apply your patches to the source code and compile your own patched version of Rockbox.
19:54:11Paul_The_Nerdpreglow: I bet an iPod could actually handle running a dedicated Quake 1 server actually. Which would be a really neat trick at a LAN party.
19:55:15 Quit ender` (Read error: 104 (Connection reset by peer))
19:55:30 Join ender` [0] (i=ychat@
19:56:04Lamedplease remind me: how do i open a new process that doesn't takes over the terminal?
19:56:33Mikachualso, you can probably press alt-f1-6 to switch vts
19:56:51ender`screen -DR and you get virtual terminals
19:56:59linuxstbMikachu: Removing the .dfg file also fixed the graphics corruption I was having in Doom on my 5g. Very strange.
19:57:24Mikachulinuxstb: my guess is the menu config things messed up the offsets
19:57:36*Mikachu tries to sound like he knows what he's talking about
19:57:52Lamedi've thought this won't work over a virtual machine
19:58:03 Join ep0ch [0] (
19:58:17 Quit muesli__ (Read error: 110 (Connection timed out))
19:58:28 Quit Bizmo (Read error: 104 (Connection reset by peer))
19:58:52linuxstbMikachu: Maybe. But I can't imagine what config setting would have caused the video corruption. But anyway, I think we should add a version number to the .dfg file.
19:59:27ender`Lamed: there's a program called screen that lets you run multiple sessions at once
19:59:54Lamedit's not commonly installed i suspect?
20:00:19ender`usually isn't, though it's present on all my boxes
20:00:55linuxstbLamed: Try (as root) "apt-get install screen"
20:01:15Lamedwhat man page should i look into for managing processes?
20:01:49 Join thegeek [0] (
20:03:06 Join quobl_ [0] (n=quobl@tor/session/x-31eefcdcba0885f8)
20:03:57 Join nave7693 [0] (
20:04:27Mikachulinuxstb: i removed that 2second sleep on startup and it pretty much starts up instantly...
20:04:37Mikachui wish i'd done that before i did the cvs bisect though
20:04:55linuxstbWhich 2 second sleep?
20:05:05Mikachuthe one that says Rockdoom v0.90
20:05:32*ender` wants doom for h120 :)
20:05:34linuxstbThe rb->splash() ?
20:05:38Mikachuyeah, sorry
20:05:40linuxstbender`: You've got it.
20:05:44lostlogicSlasheri: latest CVS seems fine here when just playing, but breaks when skipping is that what you were talking about buffering errors?
20:05:47lostlogicpreglow: yes, there is such a patch
20:05:53preglowlostlogic: found it too
20:05:57Lamedlinuxstb: i'm not interested in screen. i'm trying to stop using cygwin, and vmware has virtual terminals, and i'm trying to use only ubuntu linux but i can't get it to connect a pptp connection :(
20:05:57linuxstbender`: Yes, it's been working for a few days.
20:06:00preglowlostlogic: not in main tree yet, i see...
20:06:07lostlogicpreglow: yeah, they suck.
20:06:14ender`great, is it in the CVS or separate?
20:06:18Lamedwhat man page should i look into for managing processes?
20:06:35linuxstbUsing the grayscale lib. Only problem is that it's a bit dark at the moment.
20:06:45ender`cvs -z3 up && cd build && make
20:07:45linuxstbLamed: I don't think there is a single man page. Your shell does some things (try "man bash"), and other commands related to processes are "ps", "top", "kill", "killall"
20:08:12Lamedwill l00k
20:08:12Mikachuthere is probably some good startup docs in ldp
20:08:14ep0chlostlogic: FF is still a bit flakey, i have to pause and unpause to get playback after FF
20:08:34lostlogicep0ch: how's track skipping?
20:08:50lostlogicep0ch: yeah, I'm actively trying to figure out what the deal with seek_complete not being called is :-\
20:09:30Lamedgreat, now that i know how to get a web... this is too much info for one time.
20:09:40 Quit Rondom ("I'm leaving on a jetplane, don't know when I'll be back again...")
20:09:48lostlogicSlasheri: if you call the changes my last 4 commits have made to playback.c not major, I question your sanity...
20:11:22 Part raf82
20:11:43linuxstblostlogic: Whats the problem with seek_complete? I remember that function was added relatively recently, so if not all the codecs use it, it's a bug and they should be fixed.
20:11:50 Join raf82 [0] (
20:12:06lostlogiclinuxstb: hmm, I wonder if that's the problem, I was looking for a cause in the playback code, but it could be in the codecs.
20:12:11lostlogiclinuxstb: thanks!
20:12:36ep0chlostlogic: skipping seemed to work with sound glitches, but now i think i've skipped so many times its gotten confused
20:13:03Mikachui found a real use for the headphone patch, i can just put my hand in my pocket and pull the cord out, and rb will shut down after 4 minutes, i don't have to fiddle with the lock and find the play button and turn hold back on etc
20:13:05lostlogicep0ch: let me know if you can determine how you confuse it, the sound glitches I am also working on.
20:14:16 Join damaki_ [0] (
20:14:36lostlogiclinuxstb: :( all the codecs call it, the problem is still mine :-P
20:15:25Paul_The_Nerdlinuxstb: You shouldn't get him all hopeful like that. Poor fellow.
20:21:38ep0chlostlogic: play/pause seems very dodgy while buffering
20:22:20 Join bluey [0] (
20:22:34lostlogicep0ch: ie pause doesn't pause and unpause doesn't unpause?
20:25:02ep0chlostlogic: it does pause and unpause eventually, but it has a fit on the way :)
20:25:28lostlogic*tries it out*
20:25:45ep0chlostlogic: sometimes it sounds like it is rapidly pausing/unpausing all by itself, and the sound flutters
20:26:23lostlogicoh my, so it does
20:26:48ep0chi think that was what slasheri might have been talking about
20:27:31lostlogicgah, hell if I can think of what I've changed that would impact pause behavior, but obviously...
20:27:51 Quit Hadaka (Read error: 110 (Connection timed out))
20:28:22lostlogicfirst pause works, but after that all hell breaks loose.
20:28:52lostlogicactually, I wonder if this and the seeking artifacts stem from the same problem wherever that is.
20:30:06lostlogiccan anyone tell me the design purpose of the mp3_pause_done() function?
20:30:12 Quit damaki__ (Read error: 110 (Connection timed out))
20:30:42 Join obo [0] (
20:32:58preglowno :/
20:34:12preglowi wonder why the hell get_replaygain_int returns 0 for a decibel value of 0
20:34:19preglowit should return fixed point 1
20:39:33 Join harbel [0] (
20:40:04luckz20:30:07 < lostlogic> can anyone tell me the design purpose of the mp3_pause_done() function? <- fadeout?
20:43:02Lamedhow do i update lang strings to lang2?
20:45:37lostlogicwe don't fade on pause.
20:45:47preglowlostlogic: yes we do
20:45:48luckzmy player does..
20:45:58Lamedisn't there a script to generate lang 2 files out of old lang files?
20:46:00luckzso, uhm, unless your rockbox is entirely different..
20:46:01preglowthere's a setting for it, even, which i always turn off
20:46:24lostlogicwhere's that done!?
20:47:05luckzlook at the options ;)
20:47:21luckzmaybe it's related to the crossfeed settings, not sure.
20:47:32preglowit's not
20:47:34preglowit's in playback
20:47:35amiconnNo, it's totally different
20:47:52preglowand i suspect it's been in rockbox for ages
20:47:55amiconnFade out is done with the audiocodec's volume control, and also works on archos
20:48:03lostlogicok, I have never seen any code in playback.c, pcmbuf.c or pcm_playback.c to handle fading out on pause
20:49:46amiconnmp3_pause_done() is related to fade out, in that it tells whether playback has really stopped
20:50:43preglowcould anyone on cygwin check if mp3_encoder compiles?
20:50:56amiconnIt exists because the MAS has an internal buffer of a fixed length, which we can't flush programmatically and causes it to continue playing for some time after we stop feeding it mp3 data
20:51:58amiconnThis time can be more than a second for very low bitrates. When doing fade out, after we reach silence, we need to wait for the MAS to run out of data before switching the volume back to normal
20:53:11lostlogicamiconn: ah, fade out is done with volume, no wonder I've not seen it, thanks!
20:56:13 Join jbauman [0] (
20:56:46 Quit Nico_P (Remote closed the connection)
20:56:57 Quit solexx (Read error: 110 (Connection timed out))
21:07:46 Join solexx [0] (
21:09:20 Join SereR0KR [0] (
21:14:08 Quit borges_ (Read error: 104 (Connection reset by peer))
21:16:26preglowanyone wanna test some crossfeed?
21:16:47***Saving seen data "./dancer.seen"
21:17:17warthawgno thanks, i just ate
21:17:31 Join JBGood [0] (
21:23:06CtcpPing from t0mas!n=tomas@rockbox/developer/t0mas
21:23:06***Server message 505: 'logbot- :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
21:23:33luckztest how preglow?
21:23:50 Quit jbauman (Read error: 110 (Connection timed out))
21:25:05 Join webguest09 [0] (
21:26:09preglowluckz: as in test the new one
21:26:10preglowbut forget it anyway
21:26:10luckzwhere are new WPS tags actually defined?
21:26:18preglowi just remember i haven't tested it on target yet :-)
21:26:26lostlogicluckz: gwps-common.c
21:27:09luckznow I understand :)
21:27:57 Join Fitzsimmons [0] (
21:28:21luckzwas metadata.c/.v abandoned in favour of tagcache?
21:30:05 Quit mirak (Connection timed out)
21:31:11Slasheriluckz: no
21:31:17RoC_MMpreglow, I don't wnat to bug you, but how can I keep track of battery indicator progress on 4G grey?
21:31:27RoC_MMis there anything more I can provide?
21:31:28Slasheritagcache uses metadata.c to get the real metadata
21:31:37Slasheri(while it builds the db)
21:31:47luckzSlasheri: well, the albumartist patch changes metadata while the discnumber one changes tagcache
21:32:06preglowlinuxstb: didn't you do some measurements on the 4g now?
21:32:13Slasheriluckz: yep.. the albumartist patch needs to add an entry to tagcache also
21:32:21Slasherisame way the discnumber patch does
21:32:52preglowamiconn: shouldn't be too hard to have battery readout support soon now, then?
21:32:52luckzso I'll just go to all places changed by either and put in the necessary stuff for both, yes? their function is pretty alike, after all.
21:33:16Slasheriamiconn: btw, what do you think if the first entry in file browser would be <Browse Tags> and in tagcache browser <Browse Files>?
21:33:22SlasheriI have almost implemented that now..
21:33:24ender`where does doom expect it's WADs?
21:33:34Mikachuender`: it expects its wads in /games/doom/
21:33:37crashdSlasheri: i'd like that ;o
21:33:39luckzender`: apparently in Games\Doom or so
21:33:52amiconnSlasheri: That sounds less intuitive to me than just a virtual sub-dir
21:33:58luckzis doom supposed to work on the X5 in any way?
21:34:05Slasheriamiconn: hmm, yeah.. but that would work for both ways
21:34:08crashdi guess though, if you can choose it via a menu
21:34:08ender`ok, i've got doom1.wad in there, but it doesn't seem to work
21:34:09luckz(it gets compiled for it, after all)
21:34:19crashdand there is a way to change on-the-fly, via a virtual dir as amiconn suggests
21:34:23Mikachuender`: do you have rockdoom.wad too?
21:34:27amiconnpreglow: I would like to see discharge curves of other ipods taken with battery_bench and continuous playback from fully charged to low-bat shutdown
21:34:44amiconnSlasheri: That would look like a circular link...
21:35:11preglowamiconn: yeah, i just need to fix the ata.c hack first, then i'll do a rundown
21:35:12Slasheriamiconn: hmm, true.. maybe the virtual sub-dir would be better then
21:35:24amiconnvirtual sub-dir also works both ways: just leave it to get back to fs
21:35:56amiconnpreglow: Nice.
21:35:59Slasheriamiconn: however, that has the one benefit that after rebooting unit, the last browse mode would be recalled
21:36:08luckz21:32:14 < Slasheri> luckz: yep.. the albumartist patch needs to add an entry to tagcache also <- the point also was more that not one patch changes both metadata and tagcache while the other doesn't but that both ONLY change either one and that that confuses me
21:36:11amiconnWe could also use a curve from a 5g...
21:36:44RoC_MMpreglow, should I just keep asking you from time to time to see the progress of the ata.c hack and the calculations needed and to keep it on your mind?
21:37:07Slasheriluckz: hmm, do you have a link to the patch that changed the metadata also?
21:37:10amiconnSlasheri: Hmm. We could remember the last dir instead. Would be even more general
21:37:22Slasheriamiconn: hehe, agreed too
21:37:27amiconn...but requires to stop using the config sector
21:37:34preglowRoC_MM: 4g battery should work soon
21:37:38preglowyou need to bug linuxstb, not mee
21:37:41preglowi don't even have a 4g least to do it in a clean way
21:38:02luckzSlasheri: <- only tagcache || <- only metadata
21:38:20RoC_MMWhat is linuxstb working on preglow?
21:38:32amiconnpreglow: If we have the curves, and also the standard capacities, correct battery readout for all ipods should be simple
21:38:54amiconnWe already know the scale factor is identical (it has to, since it is the same chip)
21:39:21Slasheriluckz: ah, interesting
21:39:31Slasheriluckz: probably that is to support ogg tags too
21:39:46 Join hardeep [0] (i=hardeeps@otaku.freeshell.ORG)
21:39:49amiconnThe 100% limit is also the same. The 0% limit is a bit different depending on generation and storage medium
21:39:51Slasherias id3.c supports only the id3 tags
21:40:09lostlogicbagh, this pause -> unpause -> all effed up during disk buffer is not easy to reproduce.
21:40:10amiconnWhat would also be useful is a charge curve
21:40:41 Join stripwax [0] (
21:40:45 Join PhR3aK [0] (
21:40:49luckzSlasheri: which method seems better, easier to implement, potentially less problematic etc?
21:40:54 Quit webguest09 ("CGI:IRC")
21:41:13RoC_MMIt's not linear?
21:41:16 Quit actionshrimp ("( :: NoNameScript 4.01 :: )")
21:41:31PhR3aKi have a question, when i update rockbox, should i remove all the rockbox files on my x5 or should i overwrite them?
21:41:42RoC_MMoverwrite them PhR3aK
21:41:46 Join afruff23 [0] (
21:41:47luckzPhR3aK: I just overwrite them, keeps your settings
21:41:57Paul_The_Nerdluckz: Deleting them will still keep your settings.
21:42:05luckzPaul_The_Nerd: where are settings saved?
21:42:21Paul_The_Nerdluckz: In their own sector, elsewhere on the disk. They aren't in an actual file.
21:42:23afruff23does the X5 have software charge control or monitoring
21:42:30 Quit goffa_ ("leaving")
21:42:42afruff23for power
21:42:52Paul_The_NerdUsually it's safe to just overwrite, but every now and then you end up with spare files. Like bejeweled.rock after the rename.
21:43:11 Join goffa_ [0] (n=goffa@
21:43:26PhR3aKyes thats what i meant, in the and i got too many trash in my rockbox folder
21:43:30PhR3aKso better delete?
21:44:01luckzdid you add custom themes etc? still have them on your comp?
21:45:17PhR3aKis someone working on the power management for the x5? because the battery life really isnt that good even with the "L" version
21:45:22afruff23does anybody know if the X5 has software charge control or monitoring
21:45:55DBUGEnqueued KICK PhR3aK
21:45:55PhR3aK(afruff23): how long lasts your x5 with the rockbox fw?
21:46:09afruff23don't keep track
21:46:15luckzafruff23: I dont think that just repeating yourself will get you anywhere; I also dont understand your question
21:46:24lostlogicep0ch: is it true that the screwed up pause crap only happens during initial buffer fill?
21:46:43ender`ouch, doom on h120 can certainly cause a headache
21:46:49afruff23 %bc "c" if the unit is currently charging the battery (only on targets that have software charge control or monitoring)
21:46:57afruff23taken from CustomWPS page
21:47:16afruff23I want to make an X5 WPS with this
21:47:19afruff23is it possible?
21:47:27ep0chlostlogic: seems so
21:47:38 Part stripwax
21:47:46lostlogicep0ch: yay for at least narrowing down WTF.
21:47:54hardeepI'm totally for the virtual sub dir idea for tag database
21:48:05luckzafruff23: I dont know if you can even charge the X5 while using it. at least I havent done so successfully.
21:48:19Paul_The_Nerdafruff23: Can you charge while in the retail firmware? If so, you'll probably eventually be able to charge in Rockbox.
21:48:24hardeepespecially if other modules (e.g. playlist) could use the virtual path directly
21:48:27Slasheriluckz: i am going to do the sub dir idea. But not sure if i do it during the feature freeze :)
21:48:52hardeepthat way you wouldn't have to insert each track into the playlist
21:49:06hardeepwhen in tag database
21:49:29Paul_The_Nerdafruff23: By "charge while in retail" I mean, can you charge while doing other stuff (playback, browsing)
21:49:54 Nick Kyomi is now known as Kyomi|off (
21:50:02amiconnSlasheri: How's chunked browsing coming along?
21:50:13afruff23basically, when rockbox is on?
21:50:18luckzSlasheri: how long do you think it'll approximately take until I in some way can obtain a then-current rockbox build with those features?
21:50:20PhR3aK(Paul_The_Nerd) yes you can
21:50:32Paul_The_Nerdafruff23: I didn't say anything about rockbox...
21:50:38Slasheriamiconn: needs more work, but probably in a few days when i get enough time to do it
21:50:45afruff23oh, charge in retail firmware?
21:50:47luckzPaul_The_Nerd: if I understood the manual etc correctly, you cant charge in retail
21:50:51luckzbut that is just me
21:51:08Paul_The_Nerdafruff23: Then eventually you'll be able to charge in Rockbox probably, so the battery charging tag should work (once that's implemented, that is)
21:51:08luckzas it tends to then throw you OUT of the firmware and into a "charging..." screen, IIRC
21:51:28afruff23no, it doesn't
21:51:36ep0chSlasheri: any idea what a "-2" error code is on boot with dircache enabled?
21:51:42Slasheriluckz: well, if not during the feature freeze, probably very soon after it (just make sure you remember to mention that to me again) :)
21:52:02afruff23the battery icon animates while charging in the original firmware
21:52:08Slasheriep0ch: unfortunately the error code doesn't tell much :/ you have two possibilities:
21:52:38Slasheritoo many files (dircache ram size limit is reached) or too deep directory structure (over 6-8 levels)
21:52:54PhR3aKhm, any idea why rockbox hangs sometimes? it happens after a time, i recognize it often when crossfading stops working, and when i try to swich tracks, the x5 hangs
21:53:00ep0chSlasheri: yes too many files - i tried 30,000 sid files :)
21:53:00Slasheriep0ch: if you want the exact error, you may check with logf enabled
21:53:05Slasherihehe :D
21:53:28Slasheriyep, that is too much for dircache to handle
21:53:30ep0chboy wasn't rockbox happy when i tried to delete them as well
21:53:30goffa_haven't had that happen to me PhR3aK but then i don't use crossfade
21:53:47Slasheriwe would need some sort of compression to handle that much of files
21:53:54PhR3aKcould it be that crossfading is not this stable?
21:54:01ep0chno worries
21:54:09afruff23I bleive it puts a lot fo strain on the processor
21:54:13afruff23to use crossfade
21:54:18PhR3aKyeah that could be
21:54:26PhR3aKbut i love this feature :)
21:54:34goffa_its always annoyed me
21:54:43theli_uai'm slightly confused ... why does 'make clean' makes sources unusable? (for example it removes 'firmware' dir)
21:54:47Slasherilots of changes to the playback engine has been made recently
21:54:58luckzI got rid of crossfading because it takes annoyingly long to just skip a few tracks
21:54:59 Quit SereR0KR ("XChat Aqua")
21:54:59Slasheriand in its current state it's quite buggy..
21:55:08luckz(it should not crossfade on manual skip IMO)
21:55:15afruff23Could the hold switch for the X55 be used for somethign other than hold?
21:55:18luckzI really like crossfades on track end though
21:55:18Slasheribut now going to sleep, night
21:55:18goffa_it took me a while to disable it on pause
21:55:34goffa_then i found it
21:55:35afruff23since the buttons and joystick never get accidentally pressed in most pockets
21:55:40hardeepluckz: i believe that is configurable
21:55:49luckzafruff23: uh, in mine they do. I thought that too, at first.
21:55:58 Join robin__ [0] (
21:55:59PhR3aKhm... crossfading works for my when skipping tracks, too
21:56:06luckzhardeep: you can make it crossfade only when in shuffle mode, but that still does that on manual skip
21:56:07robin__where do I get the wad files for doom ?
21:56:13luckzrobin__: hy legal reasons
21:56:23hardeepi thought midk added an option for crossfade on skip
21:56:32robin__hy ? means :) ?
21:56:45Mikachurobin__: there's a shareware wad on the wiki page
21:56:45luckzrobin__: that this is not #warez, if I understand correctly
21:57:01Paul_The_Nerdrobin__: I found them on a copy of the game I bought for $10 at walmart
21:57:03robin__ok, but thought the demo wads was okay
21:57:09afruff23you can use free, legal WADs made by people
21:57:11Mikachui can reveal that it's not very hard to find on google
21:57:19robin__no okay, it's doom1 ?
21:57:26Mikachurobin__: yes, the demo wad is free
21:57:27Paul_The_Nerdrobin__: Well the demo wad is linked to in the wiki at PluginDoom...
21:57:55luckzso did anybody actually try doom on x5? or not?
21:58:10afruff23memebrs report it to be working
21:58:19PhR3aKyep they sad it works smoothly
21:58:26afruff23not smotthly
21:58:26luckzusing the /games/doom/ folder there too?
21:58:32afruff23they said there is no enter button
21:59:12PhR3aKi mean it doesnt lag...
21:59:17luckz <- what game is that?
21:59:21afruff23don't know aboyt that
21:59:26Mikachuluckz: bubbles
21:59:44 Quit Kohlrabi (Nick collision from services.)
21:59:53 Join Kohlriba [0] (
22:00:07afruff23my mesage was directed to phr3ak BTW
22:01:07PhR3aK(afruff23): Is your x5 getting hot, too? i experienced this today
22:01:31warthawgis a wad an ascii file?
22:01:34warthawgor binary?
22:01:35PhR3aKi mean while using rockbox
22:02:06PhR3aKmine is getting kind of hot on the back
22:02:16 Join Naked [0] (
22:02:33robin__hmm I have created a /games/doom/ directory and placed doom1.wad there
22:02:33 Nick Naked is now known as Hadaka (
22:02:37robin__but it can't find the wad file?
22:02:38afruff23a lot of people have been notiving that as well
22:02:50Mikachurobin__: please read the wiki page!
22:03:03afruff23directed at phr3ak
22:03:17robin__Mikachu: sry :D
22:03:40 Quit RedBreva_ ("Leaving")
22:03:52subwayjaredsomething cool to add to the "car adapter mode" would be a switch that turned off all effects/eq and turned the volume all the way up for the best "line in" sound
22:06:22PhR3aKafruff23 hm... and another thing is, that the voltage (normally about 4V) is getting lower and lower at every start (you can see the voltage in the boot screen)
22:06:22afruff23line out*
22:06:22 Quit tianjing_ (Connection timed out)
22:06:22ep0chis it just me or is .rockbox starting to have too many dirs inside it?
22:06:23afruff23over tiem batteries lose voltage
22:06:23afruff23but I don't know how often
22:06:23DBUGEnqueued KICK afruff23
22:06:23afruff23sicne they build up resistance
22:06:46subwayjaredline out on the mp3 player, line in on the stereo yes
22:06:53afruff23you won't see anymore extra dirs for a while
22:06:59afruff23(feature freeze)
22:07:01Doomedsomebody fuxxed up the tagcache play (i guess thats what ill call it), and they took off the "Add to playlist option"
22:07:13ep0chdoom dir will go in there
22:07:43Doomedo nvm
22:07:45 Quit damaki_ (Remote closed the connection)
22:08:09afruff23subwayjared, yes but you were referrign to the player itself, the greatest line-out may not be mean a great line-in
22:08:14ep0cha common 'settings' dir would be nice, instead of having them scattered around
22:08:17 Quit theli_ua ("")
22:08:19afruff23may not mean*
22:08:20robin__and how do I play fullscreen ? (there is some border), and get back to the menu ?
22:08:23 Join damaki_ [0] (
22:08:34CtcpIgnored 2 channel CTCP requests in 20 minutes and 28 seconds at the last flood
22:08:34*robin__ kicks myself for not reading wiki
22:08:40Mikachurobin__: go to the menu and options, and make change Screen Size
22:08:51Mikachuwow, great sentence
22:09:00*Mikachu escapes to kitchen
22:09:18*Doomed sees if the new cvs
22:09:21scottderRTFW :)
22:09:46sharpei are back here!
22:10:00afruff23how doe we find out the exact changes to a wiki topic
22:10:50afruff23I knwo about the recent changes thign but it doesn't say where exactly in the topic
22:10:50afruff23how do*
22:10:50 Join JBGood25 [0] (
22:11:21 Join webguest83 [0] (
22:11:28robin__lol the sounds
22:11:30robin__are wicked :)
22:11:36robin__there is no music, that's right?
22:12:34 Join muesli__ [0] (n=muesli_t@
22:12:40 Quit webguest83 (Client Quit)
22:13:39afruff23robin, what palyer are you using?
22:14:00Doomedanybody no why this is happening?
22:15:03robin__afruff23: ipod nano
22:15:33Doomedwtf, can anybody help me real quick
22:16:27warthawghas the rapture begun?
22:16:36robin__what's the status of the gbc emulator for the ipod?
22:16:39robin__ipod nano
22:16:55 Quit novimon (Read error: 110 (Connection timed out))
22:17:40Doomedwhen i play a song, then i change to another song, the first song keeps playing
22:17:55Doomedthe screen changes then the music abruptly stops and the new song starts
22:18:53Paul_The_Nerdrobin__: It does not work yet.
22:19:00Doomedand why was the add to playlist option taken out of tagcache
22:19:31luckzI assume there won't be a rockboy for X5 any time soon, due to the LCD things?
22:19:34 Quit JBGood (Read error: 110 (Connection timed out))
22:20:15Doomeddoes this happen to anybody else
22:20:20ep0chDoomed: playback code is being reworked
22:20:26afruff23LCD things?
22:20:33afruff23you eman small LCD?
22:20:43Doomedso thats why eh
22:20:48 Join lee-qid_ [0] (
22:21:00luckzafruff23: the X5 LCD works different, technically, and thus you cant simply make rockboy work in a second or two. read the feature request stuff for more information.
22:21:47afruff23luckz: link
22:21:50afruff23I couldn't find it
22:22:39 Quit wefds (Read error: 110 (Connection timed out))
22:23:17harbelCould anyone point me to a tutorial on how to make rockbox themes, I searched, but couldn't find anything
22:23:23Doomedwhats the estimated time itll be finished?
22:23:48 Join Lamed_ [0] (n=Shachar@
22:24:01Paul_The_NerdDoomed: There are no time estimates, really
22:24:18Doomedits really annoying
22:24:39Paul_The_NerdWell, it'll be done when someone does it. If you want to know when it's done, be the one working on it. :-p
22:24:46ep0chcan you get an older build?
22:25:06*amiconn just discovered horizontal scrolling in oscilloscope is broken on greyscale ipods after the lcd driver flip
22:25:18*amiconn goes to fix the scrolling routines
22:25:52RoC_MMOlder builds are available on the website.
22:26:01RoC_MMunder "daily builds" click "older" under your player
22:26:35RoC_MMWhat is this LCD driver invert thing on 4G ipods?
22:27:35afruff23that might not be what you're looking for
22:28:30harbelNope, I'm looking for how to make WPS files, moreso
22:28:36harbelmore so*
22:28:49afruff23go to the customWPS page
22:29:24Doomedyeah i no where they are
22:29:30Doomedi would really like to help
22:29:35Doomedbut i have no clue where to start
22:29:53afruff23doomed: are you talking about WPSes?
22:30:05Doomedno, talking about helping code
22:30:36Doomedeh i guess ill get an old version
22:31:49harbelThanks aruff23
22:32:35afruff23I don't know how to change font and font color though. Do you jsut change them in general settigns?
22:32:57harbelYou can set them in the .cfg file for the theme
22:33:52afruff23but there is no wiki on this format
22:34:01afruff23it mentions that thmes also use this format
22:34:08afruff23but they don't mention what variables
22:34:16Paul_The_NerdManually save your settings, and then all variables are in the .cfg
22:34:21Paul_The_NerdThen trim the ones you don't want changes.
22:34:41 Quit PhR3aK ("get satisfied! :: (Gamers.IRC) ::")
22:35:01afruff23can you change the font color for the WPS and not the menus?
22:36:36afruff23if I "open" a cfg file with only the foreground color variables, will my other settings change? (i.e. id3 tag priority, scroll speed, etc.)
22:36:58 Quit Lamed (Read error: 110 (Connection timed out))
22:38:03afruff23to reword my question better, if a cfg file is applied and it only contains some but not all variables, will the unincluded variables be affected(i.e. go back to default values)?
22:38:11luckzI dont think so.
22:38:16 Quit subwayjared ()
22:38:23luckzI had a huge problem because one config file changed my background and that made everything unreadable.
22:38:27luckzI had no other background file.
22:39:02luckzI think a "reset .cfg-changed things to standard values" option wouldnt be all that horrible. an easy-to-find one, that is
22:39:36harbelJust put the hold switch to the on position when starting up rockbox
22:40:07harbelIt resets settings, if I'm not mistaken
22:41:13 Join Febs [0] (
22:41:16afruff23does anybody know the mAh ratign of the X5L battery?
22:41:27Paul_The_NerdIt's in the x5 forums, isn't it?
22:41:34afruff23I couldn't fidn it
22:41:44afruff23you meant rockbox.rog forums
22:41:47afruff23I'll check
22:43:02Paul_The_NerdI'm basically always going to mean the Rockbox forums.
22:43:47afruff23yea, I'm used to the forums
22:43:56 Join RedBreva [0] (
22:44:15 Quit harbel ()
22:44:24goffa_afruff23: 2250 mah
22:44:27goffa_is what linus says
22:44:31goffa_i think he's pretty close
22:44:37goffa_some claim that its over 3200
22:44:41goffa_hard to guage
22:44:49 Quit Lamed_ ("Chatzilla [Firefox]")
22:44:52goffa_and really isn't important as the meter isn't accurate
22:45:09afruff23yea, I found hat topic
22:45:30goffa_ah ok
22:45:53afruff23is 2250 good?(compared to other rockbox players)
22:46:21Paul_The_NerdThe H120/320 are 1300.
22:46:32goffa_actually i think the x5l has one of the stronger batteries if not the strongest
22:46:41amiconnSome say the H300 is 1400mAh
22:46:41Paul_The_NerdThough many people have replaced 'em with 1900-2200 ones.
22:47:16afruff23are there any higher cpacity batteries available in a size that would fit mp3 players
22:47:29afruff23I am aware the price would be high
22:48:12goffa_probably.. it all depends on the form factor of the case, how badly you want to void your warranty, etc
22:48:13Paul_The_NerdWell it depends on the mp3 player
22:48:54goffa_i get pretty good life out of my x5l.. like 20 hours
22:49:04afruff23anybody know the iPod 60 gig 5g's mAh?
22:49:08goffa_that was with default firmware or rockbox
22:50:08goffa_all of my songs are ogg q6, mp3 vbr (around 256), and mpc extreme
22:50:34goffa_ogg seems to give me the most battery life...
22:50:48afruff23odd, ogg is usually the msot intensive
22:51:00afruff23what bitrate is q6?
22:51:02goffa_i think the smaller the size the better battery life you can expect.. of course the higher the compression lower the sound quality
22:51:09Mikachui think mp3 is the worst optimised one
22:51:17preglowafruff23: vorbis doesn't use bitrate settings...
22:51:25preglowthat's the point of the quality setting
22:51:38afruff23there's no AVerage bitrate?
22:51:53Mikachuyou _can_ encode with an abr
22:51:57preglowq6 is _very_ roughly average to 192kbps
22:52:04preglowbut that's very roughly, and there's no rate control
22:52:14Mikachubut it's not recommended
22:52:43goffa_and ogg uses more cpu, but less disk access... and the disk access is what drains the battery
22:52:48goffa_at least in my experience
22:52:54goffa_flac really sucked the juice
22:52:56Mikachusome guy made the encoder work with very low bitrates, like 4 or 8, and encoded a dire straits song with it
22:53:09Mikachuit's like 250kB for 5 minutes and you can actually hear what they sing and some of the instruments
22:53:18goffa_ha ha ha
22:53:27Mikachuspeaking of bitrates
22:53:29 Join fiftyfour123 [0] (
22:54:24crashdi want that song!
22:54:42goffa_looks like ipod batteries vary in size.. from 600 to 1600
22:54:59goffa_according to that anyway
22:56:24Paul_The_Nerdslasheri: Is there a particular reason why when selecting a song in an album it would take 8 seconds for playback to start in tagcache, or is that just an "I haven't gotten around to making it create playlists quickly" thing?
22:56:29preglowMikachu: haha, yeah, that was experimental, however
22:56:40preglowsultans of swing at 1kbps
22:56:47Mikachui know i have the ogg somewhere
22:56:50Mikachui just can't remember the name
22:57:00goffa_ha ha ha.. sultans of vomit
22:57:02goffa_at that bitrate
22:57:40preglowsultans of swing > *
22:57:48afruff23goffa; there are no iPod video 60 gig specs on that page
22:58:32goffa_afruff23: from that page
22:58:37goffa_ iPod 4th Generation (4G) Battery
22:58:37goffa_Battery for the Apple iPod 4th Gen (4G) 20 GB, 30GB, 40GB & 60GB. Battery specification : Li-Ion, 3.7v, 750 mAh
22:58:41goffa_Compatible with the following part numbers :
22:58:43goffa_616-0183, 616-0206, AW4701218074 , ICP0534500
22:58:46Mikachui found it :)
22:59:32afruff23that says 4G
22:59:35afruff23not 5G
22:59:52afruff23ipod video=5G
23:00:21 Quit RoC_MM ("Leaving")
23:00:32MikachuVorbis stream 1:
23:00:32MikachuTotal data length: 199244 bytes
23:00:32MikachuPlayback length: 5m:46.293s
23:00:32MikachuAverage bitrate: 4.602895 kb/s
23:01:23Mikachuyou could stream this comfortably on a 28k modem
23:01:44crashdMikachu: send ;)
23:01:48crashdor url, heh
23:01:53sharpeserial port would be just dandy.
23:01:55Mikachucrashd: you already autoaccepted it :P
23:01:59crashdoh, heh
23:02:01crashdi went for a smoke
23:02:03goffa_i'd estimate around 910
23:02:09crashdand then did the irssi 'chase the window' game
23:02:09goffa_for the 5g
23:02:15crashdMikachu: so i did!
23:02:16Mikachucrashd: alt-a?
23:02:25goffa_based on
23:02:33goffa_and taking 14hrs vs 20 hrs
23:02:36crashdMikachu: alt-winnumber over and over, i cant do with having acts
23:02:46Mikachucrashd: try holding alt-a instead
23:02:57goffa_thats probably not exact.. but its ballpark
23:02:58crashdi remapped the alias
23:03:02crashdbut i dont remember to what O.o
23:03:21Mikachuor /script install chanact
23:03:50 Join harbel [0] (
23:03:54*Paul_The_Nerd sighs.
23:04:54Paul_The_NerdThere is something clearly wrong when someone encounters a problem, you ask if they followed all the instructions, and they say "What guy does?"
23:05:12crashdyeah, some kind of language barrier there
23:05:42 Quit RedBreva (Remote closed the connection)
23:05:45Mikachucrashd: what do you think of the quality? :)
23:06:20crashdit's like listening to it through a wall
23:06:32crashdbut you can identify the instruments
23:06:35Mikachuit is better than mp3 at 4kbits/s
23:07:04crashdi bet you could tweak it to make it sound better
23:07:07amiconnHmm, interesting: Just removing data_end++; and changing the loop condition from data < data_end to data <= data_end shaved 16 bytes off the binary
23:07:47Mikachui found an inline function in tremor that when un-inlined made the codec 10kB smaller (it was 110kB before)
23:07:57Mikachui wonder which is better
23:08:52 Part harbel ("Leaving")
23:09:44 Join harbel [0] (
23:09:49preglowamiconn: weird...
23:10:17preglowMikachu: only one way to find out
23:10:34amiconnAnyway, one step closer to grayscale lib: lcd_blit() in cvs...
23:11:54amiconn...and I know the LCD refresh rates for 4g gray and mini 2g. They are a bit lower than on archos and H1x0, but that lowers the CPU load a bit
23:13:08Mikachupreglow: and what is that?
23:13:10amiconnUnfortunately the grayscale lib on ipod will suck almost 50% CPU power (at 75MHz), similar to what it draws on archos
23:13:28amiconnOn H1x0 it needs less than 10%, and we don't need to boost...
23:13:54 Join |Unknown| [0] (
23:15:47preglowamiconn: ouuuch
23:15:53preglowamiconn: why so high?
23:16:00|Unknown|where can i get tagcache?
23:16:04amiconnThe slow lcd write...
23:16:13afruff23tagcahe is in the CVS
23:16:14Paul_The_Nerd|Unknonw|: It's built in.
23:16:31afruff23it's already included if you have a relatively new rockbox build
23:16:50***Saving seen data "./dancer.seen"
23:16:56|Unknown|hum is there a way to access music then without accessing ipod_control and selecting music and different folders
23:17:21Paul_The_Nerd|Unknown|: Enable tagcache and let it index your music. Tagcache is explained more fully at the TagCache wiki page.
23:17:34Doomedu have to force it to use tagcache
23:17:55Mikachuat gunpoint
23:18:07afruff23will this be lower quality than a dedicated line-in?
23:18:14preglowanyone wanna test new crossfeed?
23:18:41afruff23the link I posted was to a discuassion of using the iPod's headphone prot as a line-in port
23:19:26Mikachuor zoide speaking and some guy saying lol
23:19:27amiconnpreglow: So, using the grayscale lib on ipod *will* make mp3 playback skip as hell for sure :-(
23:19:43ep0chpreglow: meeeeeeeeeeeee :)
23:20:08preglowep0ch: you can do your own compiles, yes?
23:20:13preglowamiconn: how lovely
23:20:14Paul_The_Nerdamiconn: Remember, you have the second core right? Won't playback eventually be done over there?
23:20:47*amiconn decides PP should be bashed for not releasing docs
23:20:55amiconnPaul_The_Nerd: riiight
23:21:21amiconnpreglow: Would it be possible to run the grayscale isr on the cop?
23:21:24Paul_The_Nerdamiconn: By eventually, I mean somewhere around 2010 or so, but it still means that playback + grayscale is feasible. Just not, y'know... likely any time soon.
23:21:55*amiconn doesn't know anything about the communication between the 2 cores
23:22:53preglowbeware your ears, sounds may be distorted until you start adjusting settings for some reason
23:23:01preglowamiconn: sure
23:23:17preglowamiconn: but i don't think we should do that
23:23:30preglowamiconn: eventually we'll move audio/dsp to it anyway so that core 1 will have more free cycles
23:23:32lostlogicI really wish I understood why pause -> unpause during initial buffer fill causes insanity.
23:23:39 Join ashridah [0] (
23:24:09Paul_The_Nerdlostlogic: It's one of those fundamental laws of the univers. Y'know "What can go wrong, will go wrong" and "The work Brandon picks shall never be easy"
23:24:22amiconnlostlogic: Why is there a distinction between initial fill and subsequent fills, btw?
23:24:39lostlogicamiconn: this is one of the ways I've considered solving the problem :-P
23:24:48amiconnThe hwcodec engine doesn't distinguish these
23:24:59lostlogicamiconn: but I'd rather understand why it causes a problem first.
23:25:18 Join Daishi [0] (
23:25:56preglowep0ch: lemme know what you think
23:26:00Mooslostlogic: did you have the strange behaviour when you stop playback and resume, no audio and when you go to debug audio menu, all is freezed?
23:26:00 Quit quobl (SendQ exceeded)
23:26:05lostlogichmph, now I can't get it to screw up any more, this is very frustrating, and I'm obviously missing some mode of interaction between the two threads.
23:26:31lostlogicMoos: yes, sometimes when playing from an offset I've had that.
23:26:37ep0chpreglow: will do, just copying over the new build
23:26:57 Part raf82
23:27:06Mooslostlogic: you really have destoyed playback man : )
23:27:47amiconnpreglow: From looking at the lcd writing code, it seems there is something like an interface controller between the cpu and the actual lcd controller
23:27:59preglowamiconn: wouldn't surprise me
23:28:04Mooslostlogic: ohoh, I didn't test your last fix yet
23:28:07amiconnThis controller handles the timing requirements of the lcd controller
23:28:16afruff23 Audio ADC/DAC driver 50% Playback works, recording not done
23:28:17lostlogicMoos: don't expect much new in the latest, that was mostly non-functional changes.
23:28:21afruff23from X5 port page
23:28:31afruff23does recording necessarily mean using hte Mic?
23:28:35amiconn...but: somehow the timing is apparently much slower than what the lcd controller can handle
23:28:37Mooslostlogic: ok
23:28:39ep0chpreglow: do you have any recommended settings for crossfeed?
23:28:40preglowamiconn: damn, it's got _SIX_ dma channels
23:28:47preglowep0ch: the defaults should be pretty nice
23:28:55lostlogicMoos: yes, I expected I'd severely break playback when I started this, now the goal is to fix it. and in the end we'll have a cleaner and more robust (and a bit faster) system.
23:28:56afruff23in other words, are there definite plans to support the mic on the X5?
23:29:24amiconnI suspect there's some configuration register for this interface controller, where the timings can be adjusted
23:29:25Mooslostlogic: we are in bugs hunting, no? : )
23:29:28preglowep0ch: check that page, it's from jlo, who designed the crossfeed
23:29:40preglowwe really should find out how to use the dma
23:29:43lostlogicthis was on the 3.0 release schedule :-P
23:29:47amiconnNow if I only knew this register :-(
23:29:53Paul_The_Nerdafruff23: There are always plans to support all hardware, it's just a matter of how long it'll take.
23:30:00ep0chpreglow: cool
23:30:29afruff23"Datasheets are only available to ALi's development partners and customers who have signed non-disclosure agreements. Unfortunately, datasheets cannot be provided to end-users. ALi will review requests from other developers and grant them on a case-by-case basis. "
23:30:44afruff23taken from X5 hardware components page talking about USB chip on X5
23:30:50 Quit Febs ("CGI:IRC (EOF)")
23:30:51Mooslostlogic: but why didn't made patches and ask for testers instead of commit broken code? :p
23:31:00afruff23has any developer cotnacted these guys for a datasheet
23:31:07ep0chpreglow: no option to change the delay?
23:31:18Paul_The_Nerdafruff23: Since there's a non-disclosure agreement, information from the datasheet may not be able to be used anyway.,
23:31:19afruff23to help engineer USba nd USB-OTG
23:31:20preglowep0ch: nope
23:31:52preglowamiconn: the pp spec sheet says it's got display controllers
23:32:11amiconnpreglow: The lcd controller can handle cycle times of 600ns at the lowest voltage. It's a 16bit i/f, so this would allow ~3MByte/s transfer rate.
23:33:00lostlogicMoos: I'm encouraging people to help me.
23:33:20lostlogicMoos: and because I don't have my own SCM system and want revision control and tracking on the various steps of the process.
23:33:33lostlogic(that is I'm too lazy to set up a local repository for versioning)
23:33:33preglowwe're in dev anyway
23:33:36amiconnHowever, we get 156fps on mini at 75MHz, which means only 580KByte/s
23:33:54Mooslostlogic: hehe :)
23:34:01Mikachu"naked: Use this attribute on the ARM, AVR, C4x and IP2K ports to indicate that the specified function does not need prologue/epilogue sequences generated by the compiler. It is up to the programmer to provide these sequences." <- what does that mean?
23:34:08amiconnSo where's the slowness factor of 5 coming from?
23:34:10Mooslostlogic: if you need I test something, just ask
23:34:11preglowbesides, now lostlogic has all the testers he could ever want :P
23:34:34lostlogicexactly, I just hafta look for people yelling my name in the channel :)
23:35:01preglowep0ch: i guess i could hack up a version that has the delay control as well
23:35:14*scottder yells random things at lostlogic
23:35:17preglowi'm just sceptical to include it since i don't think it'll have very much impact, and will make my code slower
23:35:22Moospreglow: new crosfeed coming?
23:35:34preglowMoos: i just posted the patch here
23:35:51ep0chpreglow: i say don't do it if its going to make it slower
23:36:11preglowep0ch: it might, i'm coding the assembler optimised routine now, and it doesn't look like i'll have any registers to spare
23:38:50 Quit Doomed ("( :: NoNameScript 4.01 :: )")
23:40:06ep0chpreglow: crossfeed seems good, without the attenuation loss its easier to compare the difference with and without crossfeed now
23:41:00preglowep0ch: yup, be careful you don't make it clip now, though
23:41:04preglowep0ch: since that's possible
23:41:07Mikachuare the interrupt handlers written in assembler?
23:41:18preglowMikachu: on ipod, only the fiq handler is
23:41:24preglowthat's the most critical one anyway
23:41:36Mikachuthere is an "interrupt" attribute in gcc
23:41:41amiconnpreglow: Hah, this is nice: In arm asm, a conditional return is just one instruction, including the register restore...
23:41:45Mikachu"The compiler will generate function entry and exit sequences suitable for use in an interrupt handler when this attribute is present."
23:41:55preglowamiconn: yeah, that's really sweet
23:41:57MikachuPermissible values for this parameter are: IRQ, FIQ, SWI, ABORT and UNDEF.
23:42:07preglowamiconn: and including a thumb/arm state change
23:42:16amiconnldmeqia sp!, {......., pc}
23:42:46preglowthe ldm and stm instructions are really handy
23:43:58preglowsweet god, gcc is so incredibly wasteful with the registers
23:44:48amiconnWhat still confuses me is the operand order
23:45:09preglowit's quite logical, really
23:45:14preglowit's in algebraic order for most thing
23:45:26Mikachupreglow: are you supposed to add strings in the middle of the lagn file?
23:45:27preglowadd r0, r1, r2 -> r0 = r1 + r2
23:45:37amiconnMikachu: No, never
23:45:48Mikachuthat's why one of my menus had YES = PLAY in it then
23:45:53preglowi tried to follow the logical layout of amiconns lang update
23:46:32preglowi question how useful the lang file backwards compatability is
23:46:45Moospreglow: crossfeed is now so far better, in one usable state for use it always for me now : )
23:47:23*Moos is playing with settings