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

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

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

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

#rockbox log for 2006-02-22

00:00:10virtualballi like both
00:00:22virtualballbut doesnt iPL have more apps and stuff?
00:00:39virtualballall i heard about on rockbox is a gb emulator
00:01:10Shadowarrior13Yeah, the controls suck though
00:01:18Shadowarrior13And iPL isn't working :P
00:01:27amiconnShadowarrior13: There's dd for windows as well
00:01:43Shadowarrior13Well, I don't know what it is :>
00:01:50Bagdervirtualball: Rockbox seems to focus more on the music playback
00:02:04JdGordon.. and games
00:02:08virtualballya but i dont need music playback :-P...its an IPOD haha
00:02:37virtualballJdGordon, what other games?
00:02:58BHSPitLappyJdGordon: bejeweled, brickmania, a bunch others
00:03:00JdGordongb/gbc emulator.. doom, breakout, tetris, snake...
00:03:14virtualballwow...all work with Rockbox?
00:03:18virtualball*on iPod?
00:03:23Bagderand a faster lcd driver ;-)
00:03:23BHSPitLappyon 5g
00:03:27Shadowarrior13Doom got fixed?
00:03:29virtualballOMG i found home
00:03:31BHSPitLappyit's fast as hell
00:03:45BHSPitLappyyou can listen to music while playing bejeweled
00:04:07 Quit ender` (" printk("; corrupted filesystem mounted read/write - your computer will explode within 20 seconds ... but you wanted it so!\n)
00:04:10virtualballBHSPitLappy, yay i found a place that works with my iPod lol
00:04:26Shadowarrior13I thought doom didn't work..
00:04:45BHSPitLappyvirtualball: still doesn't support HFS though.
00:04:55virtualballdamn lol
00:04:55BHSPitLappyvirtualball: you HAVE to convert...
00:05:15virtualballi love mac lol
00:05:24BHSPitLappydon't quote me on that... I'm not positive
00:05:30virtualballbut obviously...and surprisingly...linux doesnt haha
00:05:35BHSPitLappybut I'm pretty damn sure you can't run rockbox from hfs
00:05:48 Join yeahx [0] (
00:05:50BHSPitLappylinux will boot from fat32
00:06:05virtualballya i know
00:06:26BHSPitLappyi meant hfs
00:06:36BHSPitLappyactually from nothing... it doesn't even use that partition
00:06:38virtualballya i know, but it seems harder for me
00:06:51BHSPitLappyvirtualball: the installation is the same method here as there.
00:07:13virtualball...shut up lol
00:07:38virtualballwell all that aside...i think rock box is amazing lol
00:09:11virtualballso does anyone have a 30 gb video iPod in here lol
00:09:18josh_Anyone know how fast rockbox runs the iPod's CPUs?
00:10:06Bagderat 75Mhz methinks
00:10:09Paul_The_Nerdjosh_: 75 mhz, and only one is awake. There's a patch that enables the "boosting" so that it slows to 30mhz and boosts to 75 as necessary, but I don't think it's on the tracker yet.
00:10:31BHSPitLappyvirtualball: I did, yesterday.
00:10:51virtualballya but you dont have it anymore :-(
00:11:04BHSPitLappyI had it just yesterday, installed linux and rockbox on it
00:11:07BHSPitLappyand it was a WINpod
00:11:12linuxstbvirtualball: Yes, I've got a 30GB 5g... I'll upload my mbr for you.
00:11:26virtualballlinuxstb...i love you haha
00:11:42BHSPitLappylinuxstb: it'll work from hfs?
00:12:24linuxstbBHSPitLappy: Will what work from hfs?
00:12:42BHSPitLappylinuxstb: virtualball has a HFS ipod, not fat32
00:13:14virtualballBHSPitMonkey, it has to be from a fat32 iPod
00:14:05linuxstbBHSPitLappy: This is to convert an HFS ipod into a FAT32 ipod.
00:14:27virtualballyes BHSPitMonkey, i am converting...
00:14:35BHSPitLappylol k
00:14:55 Nick DT291 is now known as DreamTactix291 (
00:15:29linuxstbvirtualball: OK, it's at the bottom of the Mac OS installation page now.
00:16:03 Quit yeahx_ (Read error: 110 (Connection timed out))
00:16:10 Join yeahx_ [0] (
00:16:49JdGordonany1 wanna have a guess as to why my player freezes after closing a file? (that was written to)
00:17:04JdGordon.. in my plugin
00:17:25BagderJdGordon: and it works fine in the sim?
00:17:42JdGordonhavnt tried it in the sim :p
00:17:48BHSPitLappywhatcha got cookin' now, JdGordon ?
00:17:50 Join amiconn_ [0] (
00:18:03 Quit amiconn (Nick collision from services.)
00:18:03 Nick amiconn_ is now known as amiconn (
00:18:12JdGordonBHSPitLappy: fixing up my timetable viewer.. seen as uni starts next week.. ill actaually need it
00:19:20JdGordon.. seems to be working now..
00:20:30 Join Fogerty [0] (
00:23:27virtualballok im using newfs_msdos now, should it be taking awhile?
00:24:03 Part LinusN
00:24:39linuxstbI've no idea... Can you see a disk icon spinning on your ipod?
00:24:54linuxstb(top-left corner)
00:24:59virtualballya, its in disk mode, so its in b&w
00:25:16virtualballya the spinny thing is there
00:27:29 Quit yeahx (Read error: 110 (Connection timed out))
00:27:36virtualballwow...its still loading lol
00:27:45 Join ashridah [0] (
00:31:07krmathisnewfs_msdos might take a while. I guess it took ~1 minute on my 4GB nano
00:31:11virtualballok done now when im installing the bootloader, should there be a seek eror?
00:31:40linuxstbread the full message.... But I should probably delete that message.
00:32:15BHSPitMonkeyvirtualball: diskmode is a lot slower than appleos disk mode, FYI
00:32:44virtualballoh thanks...BUT UIT WORKS!
00:32:54linuxstbOf course it works.....
00:33:10virtualballi love you guys! not gay btw
00:34:53JdGordonhmm.. why isnt fd = rb->open(buf,O_WRONLY|O_CREAT); opening a file if the file doesnt exsist?
00:35:00JdGordonit works fine if the file exsists tho
00:35:30amiconnI'm not sure what's the most efficient method for computing the drawing optimisation to use :/
00:35:47amiconnThis backdrop stuff complicates such things a lot...
00:36:25*amiconn is working on memset16() and related drawing optimisations
00:36:34 Quit krmathis ()
00:39:44JdGordonany1 see the problem?? am i missing a flag or something?
00:40:04linuxstbJdGordon: What error is rb->open returning? (the value of fd) /
00:40:32JdGordon<0... i dunno coz there is no debug output :'(
00:41:19 Join tvelocity [0] (
00:42:46linuxstbThat means you are not using an absolute path.
00:42:59linuxstbi.e. it must start with "/"
00:43:05JdGordonah ok
00:43:30linuxstbYou can look at the implementation of open() in firmware/common/file.c
00:43:56virtualballok last question i promise :-D
00:44:48virtualballwhat do i do to get to the games?
00:44:59JdGordonplugin browser...
00:45:06JdGordonthanx linuxstb
00:45:06virtualballok thanks lol
00:45:20virtualballbtw, any themes for the 5G yet?
00:45:56JdGordoncept, it doesnt show the new file if dir cache is on, unless i change the file filter from supported to all and then back ?
00:46:52amiconnWell, dircache adds complexity and therefore I'd expect some bugs. You have been warned
00:47:09*amiconn did say this over and over again back when dircache was added
00:47:11 Join midkay [0] (
00:47:21JdGordonhaha ok
00:47:21*amiconn never uses dircache for exactly these reasons
00:47:30 Quit tvelocity ("Ex-Chat")
00:47:33*XavierGr is glad that dircache exists
00:48:28*Paul_The_Nerd is currently using a Nano a lot, where dircache is mostly irrelevant.
00:48:50*amiconn doesn't care much about dircache because he generally doesn't browse a lot
00:49:05*amiconn is more concerned about the voice bugs
00:49:08*midkay is using dircache because it's delicious.
00:49:49amiconnWhen voicing dirs with mp3 clips the disk has to spin anyway
00:50:33*midkay has voice disabled
00:50:36JdGordonah, finally can do my timetable :'(
00:50:39JdGordonfucking uni
00:51:48XavierGrvoice is troublesome but it must be fixed one way or the other... I wish I could help, but playback code.... ehm ... *cough cough*
00:52:41lostlogic:( yeah yeah voice bugs −− try to hunt for the initialization ones, I'll try to fix the various weirdness with voice during playback.
00:56:08virtualballok my last last question ;-)
00:56:24JdGordonvirtualball: ok last question i promise
00:56:27JdGordontoo bad :p
00:56:37virtualballok second last lol
00:57:13virtualballi downloaded a theme and opened it from the theme browser and it says settings loaded, but its still blue
00:57:36JdGordon:'( uni sux... i have a shit timetable coz it wont let me go in a prac group because it clashes with a lecture..
00:58:04Paul_The_Nerdvirtualball: Themes change the While Playing screen and/or font. Not all change the background in the menus.
00:58:13virtualballaww lol thanks
00:58:30Paul_The_Nerdvirtualball: And at the moment, they don't have the power to change the bg color, just replace it with a fullscreen image (which could be a solid other color, if they really wanted I guess.)
00:58:39virtualballand why is my battery mete saying its dead...Oh and is this bad for the battery life like iPL?
00:58:59 Join moozooh [0] (n=moozooh@
00:59:24linuxstbBattery status isn't implemented it for any ipod.
00:59:27Paul_The_Nerdvirtualball: The battery meter isn't hooked up yet. Rockbox isn't really optimized for battery life at all yet, so it'll last less per charge than Retail OS.
01:00:14virtualballok thanks and LAST one i promise...sorry for waisting your time, but where do i get the games from?
01:01:40Paul_The_Nerdvirtualball: They're built in. In the 'browse plugins' option in the menu
01:01:55virtualballi click it and it doesnt open
01:02:06Paul_The_NerdTry using Right
01:02:19Paul_The_NerdNext Track rather
01:02:29virtualballnope does nothing
01:03:10moozoohdo you see the list of plugins?
01:03:13Paul_The_NerdSo, when you choose Browse Plugins, nothing happens?
01:03:34midkaycan anyone point me to some info on how backdrops work?
01:04:07Paul_The_Nerdvirtualball: You do have the .rockbox folder copied over to it, right?
01:04:08virtualballshould there be a folder thats called plugins? cuz there isnt haha
01:04:19Paul_The_NerdIn that is a folder called Rocks
01:04:22Paul_The_NerdThat's where the plugins are.
01:04:36virtualballnope lol lemme redownload
01:04:40midkayah, i think i got it
01:04:55Paul_The_NerdAs in /.rockbox/rocks/
01:05:03virtualballnope lol
01:06:14linuxstbmidkay: If you want to change the backdrop permanently, you need to copy the bmp file into /.rockbox/backdrops/
01:06:49virtualballwhats your CVS?
01:06:50midkaylinuxstb, and then in a config file do: backdrop: /.rockbox/backdrops/whatever.bmp?
01:08:04 Join webguest37 [0] (
01:08:55linuxstbmidkay: Yes.
01:09:06virtualballwhere would i get the rocks folder, i cant find it in the CVS?
01:09:15midkaylinuxstb, makes sense. thanks. :)
01:09:47midkaylinuxstb, should the bmp file be 16-bit or 24-bit?
01:09:51linuxstbvirtualball: Just download, and then type "unzip -d /Volumes/IPOD/", or whereever your ipod is mounted.
01:10:02virtualballok thanks
01:10:12linuxstbmidkay: 24-bit
01:10:30midkaygot it, thx
01:11:12moozoohbtw, what does h300 lcd do with colors it cannot reproduce?
01:11:18*amiconn is teh silly
01:11:54amiconnFound that my fillopt check was sub-optimal for a long time. More binary code than necessary
01:12:48linuxstbmoozooh: What do you mean?
01:13:18midkaylinuxstb, for the WPS backdrop the tag is: %X|filename.bmp| - do you know if it should be like %X|backdrops/whatever.bmp| ?
01:13:50linuxstbIt should be stored in the same directory as all your other WPS bitmaps
01:14:17moozoohwell, as far as i know, it supports up to 262144 colors, but that's far from 16.7 millions provided by full 24-bit palette.
01:14:19midkayah, a subfolder in .rockbox/wps with the same title as my wps file?
01:15:16linuxstbmoozooh: All colours are simply converted to the nearest colour displayable on the LCD.
01:15:24linuxstbmidkay: Yes.
01:15:37Paul_The_Nerdvirtualball: If you build from CVS, you should always do a make zip after your make, and just use the zip it creates. Very definitely the easiest way.
01:15:42moozoohah, i see, thank you.
01:15:46***Saving seen data "./dancer.seen"
01:15:55midkayand it works. :D thanks!
01:16:33midkayexcept.. it seems when i go to the WPS and then back to the file browser i lose the filebrowser backdrop..?
01:16:40virtualballok thanks Paul, but im just downloading todays daily build
01:17:09linuxstbmidkay: That shouldn't happen. Rockbox should switch between them.
01:17:58midkayk, a reboot seemed to fix it.
01:18:18 Quit slimx (Remote closed the connection)
01:23:16JdGordonumm.. how do u goto the next file in the jpeg viewer without exiting if your zoomed in?
01:23:33golf7on h300
01:23:56JdGordonah, thanx
01:24:07 Quit ashridah ("Leaving")
01:24:14virtualballok how do i get to the Apple OS?
01:24:32JdGordonits a bit slow... some indication that the next pic is loading would be nice..
01:26:00JdGordonoh, linuxstb, the colour settings has a mnd of its own !! sometimes it loades the correct colours, and sometimes it doesnt :p
01:26:33linuxstbWhat colours are displayed when it doesn't work?
01:28:36JdGordonblue on black
01:28:51JdGordonactually.. i tink its only when u reset the colours is doesnt stick..
01:29:25JdGordonanyway, the sun is up so i cant play with rb :p
01:30:44 Quit Genre9mp3 ()
01:31:22 Quit matsl (Remote closed the connection)
01:31:50 Quit perl|fbi (Read error: 110 (Connection timed out))
01:32:06virtualballdoes anyone know how to goto the Apple OS? :-P
01:32:22Paul_The_Nerdvirtualball: Hold menu while booting.
01:32:26Paul_The_NerdBefore the apple logo appears
01:33:47virtualballok thanks
01:34:14BHSPitLappyvirtualball: how dare you ask me in #ipodlinux as well
01:34:20BHSPitLappynow I wish I hadn't answered you.
01:45:47RotAtoRanyone know why the sdl ipod sims in cygwin would crash on startup immediately after displaying the rockbox logo?
01:45:55RotAtoRnone of the other sims seem affected
01:47:03amiconnDoesn't happen here...
01:47:35RotAtoRhmm, they were working last i tried, but no luck after updating from cvs today
01:48:57amiconnThe sdl sim has another problem though which I just found
01:49:17amiconnIf you give it a cmdline option which it doesn't understand, it silently quits
01:49:39amiconnIt should print usage info
01:49:47RotAtoRyeah, i noticed that too when i mistyped "background"
01:50:11amiconnYeah, I keep trying to use −−backdrop
01:50:18amiconn(for whatever reason)
01:51:41amiconnDid you 'make clean' or equivalent?
01:52:05 Quit muesli- (Read error: 110 (Connection timed out))
01:53:16*amiconn is rebuilding sdl iPod sims
01:55:31virtualballI still cant find the rocks folder, where could i dl them from?
01:57:30virtualballnever mind i found it lol
02:00:44 Quit Febs ("CGI:IRC (EOF)")
02:07:09amiconnBagder: Shouldn't the columns with removed builds disappear from the build table when the last content is scrolled out at the bottom?
02:11:17amiconnRotAtoR: Both iPod sdl sims I usually build are still working...
02:11:25amiconn(nano and 4g greyscale)
02:12:17RotAtoRsigh, thanks for checking
02:12:20 Join midkay_ [0] (
02:13:45 Quit hyarion (
02:13:45 Quit darkless (
02:13:45 Quit qwm (
02:13:45 Quit preglow (
02:13:45 Quit markun (
02:13:45 Quit Fogerty (
02:13:45 Quit Paprica[sleep] (
02:13:45 Quit erus` (
02:13:45 Quit saa[b_r]ider (
02:13:45 Quit DreamTactix291 (
02:13:45 Quit lostlogic (
02:13:45 Quit steveb (
02:13:45 Quit ansivirus (
02:13:45 Quit t0mas (
02:13:45 Quit Kiwiclx (
02:13:45 Quit pabs (
02:13:45 Quit Kingstone (
02:13:45 Quit slarti (
02:13:45 Quit dwihno (
02:13:45 Quit Hadaka (
02:13:45 Quit jmg (
02:13:45 Quit moozooh (
02:13:45 Quit Siku (
02:13:45 Quit Slasheri (
02:13:45 Quit midkay (
02:13:45 Quit yeahx_ (
02:13:45 Quit TCK (
02:13:45 Quit Paul_The_Nerd (
02:13:45 Quit mirak (
02:13:45 Quit solexx (
02:13:45 Quit Grackshrike (
02:13:45 Quit ismo (
02:13:45 Quit tipi^ (
02:13:50NJoinmoozooh [0] (n=moozooh@
02:13:50NJoinFogerty [0] (
02:13:50NJoinPaprica[sleep] [0] (
02:13:50NJoinerus` [0] (
02:13:50NJoinsaa[b_r]ider [0] (n=saab_rid@
02:13:50NJoinDreamTactix291 [0] (
02:13:50NJoinhyarion [0] (
02:13:50NJoinSiku [0] (
02:13:50NJoindarkless [0] (n=darkless@
02:13:50NJoinlostlogic [0] (
02:13:50NJoinqwm [0] (
02:13:50 Join steveb [0] (n=steve@about/cooking/nakedchef/risotto/steveb)
02:13:50NJoinansivirus [0] (
02:13:50 Join t0mas [0] (n=tomas@unaffiliated/t0mas)
02:13:50NJoinpabs [0] (
02:13:50NJoinmarkun [0] (
02:13:50NJoinKiwiclx [0] (
02:13:50NJoinKingstone [0] (n=sg1@
02:13:50NJoinslarti [0] (n=tmartin@gentoo/developer/slarti)
02:13:50NJoindwihno [0] (n=dw@
02:13:50NJoinSlasheri [0] (
02:13:50NJoinpreglow [0] (
02:13:50NJoinjmg [0] (
02:13:50NJoinHadaka [0] (
02:13:52***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! ( )'
02:14:04NJoinyeahx_ [0] (
02:14:04NJoinTCK [0] (
02:14:04NJoinmirak [0] (
02:14:04NJoinsolexx [0] (
02:14:04NJoinGrackshrike [0] (
02:14:04NJoinismo [0] (
02:14:04NJointipi^ [0] (
02:14:22 Quit erus` (Remote closed the connection)
02:16:41 Join erus` [0] (
02:17:25lostlogicamiconn: want to test some minor changes to the buffering code surrounding voice that seem to make it better at least?
02:23:10lostlogicwell for anyone who uses voice and is interested: I only use voice when I'm messing with it, so I'm not a good test subject.
02:25:56virtualballim sorry, but rockbox is AMAZING! haha i opened brickmania and am like..WOW! lol
02:26:12 Quit moozooh (Read error: 131 (Connection reset by peer))
02:27:01 Quit Shadowarrior13 ()
02:28:32virtualballbejewled is also amazing ^_^
02:32:11 Nick josh_ is now known as josh_test (
02:32:32 Nick josh_test is now known as josh_ (
02:36:05 Join yeahx [0] (
02:47:31 Join Magnum616 [0] (
02:48:38 Quit Kohlrabi ("Fast alle Menschen sind Regenwürmer")
02:50:28 Quit yeahx_ (Read error: 110 (Connection timed out))
02:53:39 Join DT291 [0] (
02:53:39 Quit DreamTactix291 (Read error: 104 (Connection reset by peer))
02:54:00 Quit mikearthur ("Konversation terminated!")
02:55:22 Nick DT291 is now known as DreamTactix291 (
03:02:48 Join Paul_The_Nerd [0] (
03:02:54 Part Magnum616
03:03:03 Join Shadowarrior13 [0] (
03:07:18 Join aaronfg [0] (
03:14:30aaronfgwho updates the main page? There have been a bunch of things added (like sound on the ipod video, the beginnings of the 3g port, etc) that aren't even shown in the News section
03:14:34 Quit virtualball (Connection timed out)
03:15:11aaronfgPerhaps the News could be it's own wiki page?
03:15:47***Saving seen data "./dancer.seen"
03:16:38Paul_The_Nerdaaronfg: The "news" page is like, the biggest landmarks mostly. Since the changelog tells most everything anyway, and all.
03:16:52 Quit erus` (Remote closed the connection)
03:17:38aaronfgI figured as much. It's just I see a bunch of people who show up here or in the forums and don't even know stuff like the 5g having sound, etc
03:17:56aaronfgand I"m wondering if that's because there is no real easy list of updates
03:18:04aaronfgthe cvs changes are so all encompasing
03:18:25Paul_The_Nerdaaronfg: Well, you could always create a wiki page for major changes.
03:18:29lostlogicThe best parts are the "oops" and "fix red builds" commits :)
03:18:54midkay_a wiki page would be cool..
03:19:20aaronfgPaule_The_Nerd: true. I may do that later then.
03:20:32pyroWhat is the general concensus about how RockBox compares with IRiver firmware as far as battery life is concerned?
03:20:43Paul_The_Nerdaaronfg: Besides, it's not even officially released for any iRivers yet, let alone iPods. So really, the status on iPod is "It's not done yet." for all iPod targets.
03:21:03Paul_The_Nerdpyro: Well, it depends on which iRiver target you're talking about.
03:21:12pyroI used RockBox today for the first time, and the battery was dead by the end of the day. It usually takes me 4-5 days of normal usage with the Iriver. It makes me think I may have even left it on by mistake
03:21:25lostlogicit's definitely still significantly worse on H3x0
03:21:30lostlogicit's about the same on H1x0
03:21:51pyrook, just making sure it wasn't just me, or me leaving it on by mistake
03:22:14pyroany thoughts as to why?
03:22:15lostlogicpyro: the more features you use (like eq, replaygain, games, etc.) that you use the worse it will be
03:22:44pyroyeah. I was just using it to listen to music. No games, no excessive pushing buttons, no EQ, etc...
03:23:04lostlogicpyro: at this point, it is thought that at least part of the problem is the extra CPU power used updating the big LCD framebuffer, there may also be wasted power in the stuff Linus is working on WRT component power control
03:23:46Paul_The_Nerdlostlogic: I thought I'd read somewhere someone had actually tried disabling the display completely during playback (no updates) and that battery life was still not even close.
03:24:04pyroI got about 1/4 of the battery life of Iriver I think. I'd be interested if someone has attempted to gather any data on this
03:24:13*Paul_The_Nerd may have misread something, and doesn't even remember where he saw that, so I'm more trying to find out if that was just one nut posting something.
03:24:15golf7misticriver forums
03:24:21lostlogicpyro: Rob222 spent some time wiht it...
03:24:22golf7they did battery tests
03:24:32golf7and rockbox got like 10 hours, no screen
03:24:38golf7iriver got like 22 or 23
03:24:53lostlogicPaul_The_Nerd: I believe amiconn did it with backlight off and backlight on, and there was a difference −− but even backlight off, there are still updates to the framebuffer, just not to the screen
03:24:54golf7this is with a 2200mah ipod battery, like i have in mine =P
03:25:18golf7cause the WPS still calculates the variables even with the screen off
03:25:35pyrolostlogic: what is WRT?
03:25:37lostlogicgolf7: more than that, the off-screen framebuffer is still completely updated −− the only thing turned off is the writes to the physical LCD
03:25:45lostlogicWith Regard To
03:25:50Paul_The_Nerdlostlogic: Okay. Well, yeah, most people don't even have the backlight on most of their music listening time, so I'd hardly expect that to be the primary cause.
03:26:48 Join |Josh| [0] (i=user@unaffiliated/josh/x-000000001)
03:29:16pyroPerhaps I could plot current from a power supply for a couple hours with RB and Iriver. This might help to tell if it's just the average draw which is higher or certain peaks. Not sure if this is useful info though
03:30:50golf7lost: yeah
03:30:53golf7thats what i meant
03:31:04golf7like everything is being updated except the physical screen
03:31:37 Join erus` [0] (
03:32:01lostlogicpyro: since the CPU in the two is the same, we pretty much know that it's 1) because we use the CPU a lot more and 2) anything else −− what would be useful would be power draw graphs for both H3x0 and H1x0 over time.
03:33:10Paul_The_NerdCould it be a "power leak" of some sort? I mean, like powering hardware like the USBOTG stuff that's not actually being used / needed?
03:33:55pyroSo your saying the LCD is constantly updating some memory buffer somewhere, even when the LCD backlight isn't on?
03:34:08 Join BHSPitLappy2 [0] (
03:35:19 Part |Josh|
03:35:26lostlogicpyro: that's how most video drivers work −− you write to a frame buffer and then copy that buffer to the LCD, because you need X frames per second on the LCD, but you want any program to be able to write display data _somewhere_ any time it wants to.
03:36:16lostlogicPaul_The_Nerd: it's possible
03:36:31lostlogicPaul_The_Nerd: hard to tell until we get CPU usage down in the ballpark of where it is on the H1x0 series, I'd think
03:37:31golf7im almost positive its because everything is updating even though the physical LCD isnt on
03:37:51golf7a funtion to stop the updating while its off, and restart it when it comes back would be nice
03:37:54Paul_The_Nerdlostlogic: Has anyone tested its battery life in a completely idle state? (Just showing the menu, no playback/etc)?
03:38:07golf7just sitting there, nope
03:38:10pyroso why would it not affect the H1x0 series?
03:38:17golf7true that pyro
03:38:19golf7i have no idea
03:39:26 Quit saa[b_r]ider ()
03:40:27pyroPaul_The_Nerd: That actually would be an interesting test. What do you mean by just showing the menuy though? WIth the LCD on constantly?
03:40:41lostlogicpyro: because the H1x0 is a 4bpp framebuffer, and a smaller size in pixels to boot instead of a 16bpp larger buffer
03:41:10lostlogicgolf7: you can't just stop that −− because then when you wake up, you don't know what state the LCD should be set to.
03:41:28lostlogicgolf7: no single function knows what exactly to display all over the screen without a framebuffer holding that information
03:41:39Paul_The_Nerdpyro: Just disable Idle Power Off, turn on the unit with cleared settings (no backdrop enabled) with a full battery, then see how long it lasts.
03:42:25pyroSo we woudl expect the idle test to be almost identical to each other - comparing Iriver and RockBox f/w
03:42:37golf7that is totally true
03:42:39lostlogicpyro: I would expect
03:44:17pyroI'll see if I can rig up a power supply to plot voltage and current and play around with it. Me helping developing software isn't working out to well anyways : ) To much to get acquinted with initially.
03:44:40lostlogicpyro: rockbox is a small codebase... you should try on a large one sometime :-P :-D
03:44:54pyrono thanks!
03:45:51pyroI've only done small Microchip microprocessors. Very small. But I have a lot to learn with this project so I'm sticking with it. ttyl...
03:46:20lostlogicpyro: cool, thanks :)
03:46:39midkay_golf7.. that name rings a bell. :)
03:48:59golf7where you been lately =P
03:48:59 Quit BHSPitLappy (Nick collision from services.)
03:49:32golf7i used to hang out in there ALOT bout 3 years ago
03:49:38 Join Arrogant [0] (
03:50:42midkay_golf7, oohhhh, are you serious?
03:50:51 Nick midkay_ is now known as midkay (
03:51:08midkaywow, it's been, like.. yeah... a _long_ time. :)
03:54:03aaronfgdoes this color setting patch work on 5g ipods?
03:54:56midkayaaronfg, yes, it should.
03:55:54Paul_The_NerdWhen I told it to reset colors it set my BG and FG to black. Ouch.
03:56:45golf7yeah thats where i hung out
03:56:54golf7BetasIRC is what its called now i think
03:56:57golf7last time i went there
03:56:59*Paul_The_Nerd also doesn't like the button mappings on iPod, but yeah.
03:57:12golf7laste i went there, they all rejoiced, and were like golf, yo mayne, whats good
03:57:37 Join yeahx_ [0] (
03:57:57lostlogicsloccount is telling me 233kloc in rockbox, sound right?
03:58:51aaronfghmm, I'm getting "malformed patch at line 14" for that colour patch
04:01:23Paul_The_Nerdaaronfg: I got that too, had to strip out the /apps/gui/SOURCES section, and just modify that manually to get it to work
04:04:19aaronfgI'm not sure I follow. You deleted the whole part from line 7 -20?
04:05:12lostlogicsloccount tells me 1mloc in the repository I work from at my job... yep, rockbox is comparatively an easy codebase to navigate and learn
04:05:42Paul_The_Nerdaaronfg: Erased the first 18 lines
04:06:56aaronfgPaul_The_Nerd: ok, that seemed to work. Do I manually add those 18 lines to the bottom of the SOURCES file though?
04:07:23aaronfg#the ifdef part, i'm assuming
04:08:55Paul_The_Nerdaaronfg: No, actually find:
04:08:55Paul_The_Nerd#ifdef HAVE_LCD_COLOR
04:08:55DBUGEnqueued KICK Paul_The_Nerd
04:08:55Paul_The_Nerdin the sources file, and add a new line after that one that has 'gui/colour_picker.c' without the 's
04:10:47aaronfggot it
04:11:30 Quit yeahx (Read error: 110 (Connection timed out))
04:12:14aaronfggot an error
04:12:33aaronfgoh wait
04:12:38aaronfgi left in the "+"
04:12:52 Join Kyle_ [0] (
04:13:25Kyle_Should I be able to compile the SDL simulator under cygwin?
04:13:52aaronfgPaul_The_Nerd: new error - "no rule to make target 'gui/colour_picker.c', needed by 'home/aaronfg/rockbox-devel/build/apps/dep-apps'
04:14:48Paul_The_Nerdaaronfg: Does the file rockbox-devel/apps/gui/colour_picker.c exist?
04:15:07Paul_The_NerdAlso, have you done a make clean?
04:15:16aaronfgthat would be a problem
04:15:26 Join mymomthelush [0] (
04:15:50Paul_The_Nerdcolour_picker.c should've been created when you applied the patch
04:16:22 Quit `3nergy (Read error: 104 (Connection reset by peer))
04:16:36aaronfgaha. i found it
04:16:44aaronfgin the root of my rockbox-devel dir
04:17:34 Join BHSPitLappy [0] (
04:17:46 Join YouCeyE [0] (
04:17:54Paul_The_NerdBut the rest of the patch said it applied properly?
04:18:04aaronfgmake is runnign now
04:22:54 Join `3nergy [0] (
04:22:57pyroHow many targets are like the H3x0 in the fact that you can't see the LCD at all unless the backlight is on?
04:25:04Paul_The_Nerdpyro: I think it's unique.
04:26:43Kyle_why are there compiler warnings when its compiling anything that has to do with the remotes?
04:26:50pyroBeause when battery is < 10% the LCD Timeout gets automatically set to 2sec. Which is to short for the H3x0. You can't see anything cause the screen goes black.
04:27:08Paul_The_NerdKyle_: What warnings?
04:27:39Kyle_MAKE in plugins/bitmaps/mono
04:27:39Kyle_make[3]: Nothing to be done for `all'.
04:27:39Kyle_MAKE in plugins/bitmaps/native
04:27:39DBUGEnqueued KICK Kyle_
04:27:39Kyle_make[3]: Nothing to be done for `all'.
04:27:39Kyle_MAKE in plugins/bitmaps/remote_mono
04:27:39***Alert Mode level 1
04:27:39Kyle_make[3]: Nothing to be done for `all'.
04:27:41Kyle_MAKE in plugins/bitmaps/remote_native
04:27:43Kyle_make[3]: Nothing to be done for `all'.
04:27:45Kyle_MAKE in plugins/bitmaps/native
04:27:47Kyle_make[4]: Nothing to be done for `all'.
04:27:49Kyle_MAKE in plugins/bitmaps/mono
04:27:51Kyle_make[4]: Nothing to be done for `all'.
04:27:53Kyle_MAKE in plugins/bitmaps/remote_native
04:27:55Kyle_make[4]: Nothing to be done for `all'.
04:27:57Kyle_MAKE in plugins/bitmaps/remote_mono
04:27:59imphasinglinuxstb: I got pengo working
04:27:59Kyle_make[4]: Nothing to be done for `all'.
04:27:59Paul_The_NerdYou don't need to paste them all
04:28:01Kyle_MAKE in plugins/bitmaps/mono
04:28:03Kyle_make[3]: Nothing to be done for `all'.
04:28:05Kyle_MAKE in plugins/bitmaps/native
04:28:07Kyle_make[3]: Nothing to be done for `all'.
04:28:09Kyle_MAKE in plugins/bitmaps/remote_mono
04:28:10josh_Paul_The_Nerd: the iPod video is also pretty darn hard to see the screen on when the backlight's not on
04:28:11Kyle_make[3]: Nothing to be done for `all'.
04:28:13Kyle_MAKE in plugins/bitmaps/remote_native
04:28:15Kyle_make[3]: Nothing to be done for `all'.
04:28:15josh_it's not *impossible*, it's just hard.
04:28:18Paul_The_Nerdkyle_: Also, those aren't exactly warnings.
04:28:19Kyle_oops.. sorry didnt know it would paste that mucg
04:28:39Paul_The_Nerdjosh_: I've found that if I set a white background, the Nano's plenty readable with the backlight off.
04:28:58aaronfgwhat do you have to press to save the config file when you "Write .cfg file" ?
04:29:12josh_Paul_The_Nerd: nano != video
04:29:33Paul_The_Nerdjosh_: I know. But the LCDs are very similar.
04:29:44josh_yeah, I was just giving my experience :-)
04:30:39Paul_The_NerdKyle_: Those just mean exactly what they say: Nothing to be done for 'all'
04:34:13 Part Kyle_
04:34:28 Join moozooh [0] (n=moozooh@
04:34:30 Quit BHSPitLappy2 (Connection timed out)
04:35:02 Quit Strath (Read error: 104 (Connection reset by peer))
04:36:55 Join Strath [0] (
04:37:40***Alert Mode OFF
04:41:32 Join yeahx [0] (
04:48:15 Join perldiver [0] (
04:48:20BHSPitLappylinuxstb: ping?
04:51:20 Quit midkay (Read error: 104 (Connection reset by peer))
04:51:27 Join midkay_ [0] (
04:53:47 Quit mymomthelush ("Trillian (")
04:54:20 Quit yeahx_ (Read error: 110 (Connection timed out))
05:00:34 Quit webguest37 ("CGI:IRC (EOF)")
05:00:41 Quit Shadowarrior13 ()
05:01:44 Join Rob2222_ [0] (
05:10:59***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! ( )'
05:11:50 Join DT291 [0] (
05:12:12bagawkmidkay_: Hello
05:15:48***Saving seen data "./dancer.seen"
05:18:53 Quit Rob2222 (Read error: 110 (Connection timed out))
05:22:28JdGordonwhich of the BHSP 's wanted to see the coour chooser widget thingy?
05:22:45 Quit DreamTactix291 (Read error: 110 (Connection timed out))
05:28:12 Join ashridah [0] (
05:31:09 Join virtualball [0] (
05:31:39virtualballHi I have one last question lol
05:32:05JdGordonhahah ur back :D
05:33:56virtualballcan the iPod play doom?
05:34:00virtualballya im back :-P
05:34:43BHSPitMonkeyJdGordon: we're the same person
05:34:59JdGordonah ok
05:35:11JdGordonvirtualball: dunno.. i tinhk so
05:35:13BHSPitMonkeyvirtualball: I know for sure that it can in Linux ;)
05:35:24virtualball...grrr lol
05:35:35virtualball^directed at BHSPitMonkey lol
05:35:44BHSPitMonkeyvirtualball: from where you are right now, you have very little left to do
05:36:23BHSPitMonkeyI might be kind enough to assist you in , if you're cooperative :P
05:37:24JdGordonBHSPitLappy: ok, well my patch works enough if you want to patch and compile it yourself... or i can send u the rockbox.iriver for the h300...
05:37:32Paul_The_Nerdvirtualball: The Rockbox version of Doom doesn't work on the iPod yet.
05:39:48 Nick DT291 is now known as DreamTactix291 (
05:43:22BHSPitMonkeyPaul_The_Nerd: are you an, uh, "authoritative" figure in Rockbox? :P
05:43:37BHSPitMonkeybecause I have a question about logo representation
05:47:15Paul_The_NerdNaah. I'm just someone who tends to leave a channel window open whenever he's at a computer working on anything.
05:47:40Paul_The_NerdWhat do you mean "logo representation", out of curiosity?
05:47:52 Quit Arrogant ("I AM QUIT HAVE A NICE DAY")
05:52:26 Join Jungti1234 [0] (n=jungti12@
05:52:51Jungti1234What is this?
05:53:33Jungti1234It looks like source code of PMP-100.
05:54:16Paul_The_NerdIt's supposed to be. I'd been lead to understand that it's not all of it though.
05:57:27JdGordon:'( my h300 looks like i spilt milk on it... but i cant print a colour inskin to make it look good coz i dont have a colour printer...
05:59:02Jungti1234I don't have colour ink
06:00:01JdGordonwell.. i actually have 2 colour printers... 1 with no ink (and the shops dont sell ink for the model anymore) and the other is a foto printer and h3tool doesnt print the right size for it..
06:01:02 Quit aaronfg ("User pushed the X - because it's Xtra, baby")
06:02:14 Quit virtualball (Read error: 110 (Connection timed out))
06:03:53 Join ModernExecutive [0] (
06:03:57 Part ModernExecutive
06:04:20 Join ModernExecutive [0] (
06:04:59 Part ModernExecutive
06:05:37 Join aridese [0] (n=tim@
06:05:41aridesehi guys
06:06:45aridesehow is ipod 3g going?
06:15:21 Quit Paul_The_Nerd ("Leaving.")
06:18:35 Quit imphasing (Read error: 110 (Connection timed out))
06:20:57aridesehm, why won't my 3g ipod charge off of USB
06:21:53JdGordonprob not supported yet
06:25:04 Quit midkay_ (Read error: 104 (Connection reset by peer))
06:25:07 Quit aridese (Remote closed the connection)
06:25:14 Join midk [0] (
06:29:12 Quit ashridah ("Leaving")
06:29:16 Join ashridah [0] (
06:40:56 Quit RotAtoR ("zzzzzzzzzz")
06:44:38 Quit TCK ("well, if you say so.")
06:55:04JdGordonany fotochoppers here?
06:56:22JdGordonye, i found an awesome pic i want to make an inskin from.. but its square and looks shite when its shrunk to fit.. i need some1 who can make it rectangular and still look awesome..
06:59:18Jungti1234I don't understand.
06:59:44Jungti1234Keep rules of spelling.
07:00:24BHSPitLappyJdGordon: exactly what is a "photochopper" :P
07:00:36Jungti1234JdGordon: you want this? -
07:00:38ScoTTiei think he means is there anyone here good with photoshop
07:00:39JdGordonu havnt heard that term?
07:00:50BHSPitLappybut I have photoshop open :D
07:01:32BHSPitLappyJungti1234: if I do it, will you give me an "I owe you one"
07:01:38BHSPitLappydamn tab completion.
07:02:27JdGordonok, neva mind.. i tihnk im happy...
07:02:47Jungti1234JdGordon: Don't you have Photoshop?
07:03:22JdGordoni have no photoshop skills
07:03:42BHSPitLappyI am leet
07:04:05BHSPitLappyI did a handful of the build table thumbnails, without having the devices :P
07:04:09Jungti1234Especial skill doesn't need using it.
07:04:21BHSPitLappyJdGordon: what's the picture?
07:04:33JdGordonim trying to find it on the net again..
07:04:53Jungti1234I can make it and give to you.
07:05:08Jungti1234Back image?
07:05:21BHSPitLappyJdGordon: actually, why do you care? I thought your printer doesn't work?
07:05:32JdGordoni got a friend who will print if for me :D
07:06:18JdGordonthat pic looks .ok.. in h3tools.. but the man is squeezed.
07:09:03BHSPitLappywould the pic go on the front or something?
07:09:19JdGordonfront and back...
07:09:40JdGordonactually if u could get rid of the man for the front id be willing to have your kids....
07:10:00 Quit moozooh (Read error: 104 (Connection reset by peer))
07:10:07JdGordoni mean... i would be extreemly apreciative..
07:10:11BHSPitLappy... "Miranda"?
07:10:59BHSPitLappythat's your username...
07:11:03CtcpIgnored 2 channel CTCP requests in 18 minutes and 44 seconds at the last flood
07:11:03*BHSPitLappy looks suspicious
07:11:10JdGordon? thats my client...
07:11:54JdGordonah, thats the Ident id.. not my freenode username
07:12:29BHSPitLappyusually that's where a login or etc gets shown
07:12:37BHSPitLappylike root, etc
07:12:57 Quit Jungti1234 ()
07:13:12BHSPitLappyguess he couldn't live up to the responsibility he took on
07:14:48BHSPitLappyhe said he'd make it
07:14:50BHSPitLappythen /quit.
07:15:51***Saving seen data "./dancer.seen"
07:16:46BHSPitLappyif you can find a good template for the back...
07:19:04JdGordon ?
07:21:53JdGordoni want it to look like for the front...
07:22:06JdGordonif i could figure out how to print that the corretc size id use it...
07:24:44JdGordonBHSPitLappy: h320 psd
07:25:54 Join DJ_Dooms_Day [0] (
07:29:03Paprica[sleep]linuxstb, here?
07:29:10 Nick Paprica[sleep] is now known as Paprica (
07:31:06 Join damaki__ [0] (
07:34:57BHSPitLappyso should the sides be black
07:35:07BHSPitLappy@ JdGordon
07:35:38JdGordoni found a cd with photoshopCS on it and im trying to get it to run...
07:43:24BHSPitLappyI think I'm done...
07:44:08BHSPitLappyJdGordon: got a VNC client?
07:44:34BHSPitLappyvnc password jdgordon
07:44:41BHSPitLappyjust to show you what I got
07:44:56 Quit damaki_ (Read error: 110 (Connection timed out))
07:45:09JdGordonthats awesome
07:45:16JdGordonu cant make the back full size?
07:45:37BHSPitLappywithout losing anything??
07:45:53JdGordon:( ye, i guess so
07:46:33JdGordonis the crud quality from vnc?
07:46:55BHSPitLappyso like that?
07:47:40JdGordonmy mate is gonna cream himself when he sees this :D
07:48:20JdGordoncould u try getting rid of the text on the back? or get the whole "in flames" there?
07:48:24JdGordonyour a legend :D
07:51:02 Join moozooh [0] (n=moozooh@
07:54:35 Join damaki_ [0] (
07:54:51JdGordonBHSPitLappy: thats awesome
07:55:09JdGordondont worry about the name tho...
07:57:32JdGordonmorning amiconn
07:58:48midkmorning amiconn, just today tested out jpeg viewing in color (yeah, i'm a bit late..) :) nice work!
07:58:49 Join segphault [0] (n=chatzill@
08:00:53BHSPitLappyJdGordon: of course, this is being released CC ;)
08:01:33amiconnmidk: thx, just it wasn't only my work
08:01:59BHSPitLappyJdGordon: Creative Commons license :P
08:02:04JdGordonamiconn: could u get it to preload the next picture in the background while your viewing?? speed things up..
08:02:12midkamiconn, i just know it went from not working to working and it was due to you ;)
08:06:26 Join damaki [0] (
08:08:46Papricaamiconn, the resume/sel resume is garbage in the player?
08:08:48 Quit damaki__ (Read error: 110 (Connection timed out))
08:10:01warewolf| . |
08:10:05warewolf| .|
08:10:21midk|. |
08:10:22warewolf| . |
08:10:37midk| .|
08:11:27amiconnPaprica: I still don't get what you mean...
08:11:34amiconn(and I would be *really
08:11:57amiconnsurprised if you ported brickmania to the *player*
08:12:32midkspeaing of "resume". whatever happened to "ask resume" on boot? now it's just either yes or no.. :(
08:12:39amiconnIf you mean the menu on recorder/ondio, the 'quit' item has a squeezed left side of the 'q'
08:13:14Papricawait, i'll give you a pic from the sim
08:13:32warewolfI couldn't get rockboy to play two pubdomain gameboy games
08:15:23*BHSPitLappy strangles firefox with a dead hooker's arm
08:15:33BHSPitLappyI'm sorry, how terribly inappropriate...
08:16:53amiconnPaprica: Ah, yes. Never checked the 'resume' item in active state...
08:17:03Papricaoh =\
08:17:15amiconnThe inactive state is okay, just the text striked through in white
08:17:16Papricai dont know why it happens
08:19:42 Quit damaki_ (Read error: 110 (Connection timed out))
08:20:07amiconnHmm, I don't know how this menu is working internally...
08:20:41amiconnThe garbage even changes depending on where the cursor is
08:21:11amiconn..and the cursor isn't displayed correctly either in this line
08:22:18Papricait's not the cursor, i try to view the bitmap alone and the same thing is happen
08:24:10JdGordonhell ye!
08:28:43JdGordonactually.. BHSPitLappy, could u lighten it up a bit... i tihnk it will be too dark like that unde the cover..
08:31:57amiconncolour plasma can be made a *lot* faster
08:32:06BHSPitLappyit IS fast on my nano
08:32:13BHSPitLappya little slow on the 5G, but meh
08:33:07Bg3rBHSPitLappy wow, even license included :)
08:33:17BHSPitLappyBg3r: my first time! :D
08:33:40*Bg3r really should inskin his player too
08:34:16Bg3rBHSPitLappy where did you get the mask from ?
08:34:27BHSPitLappyJdGordon linked it
08:34:51Bg3r <= this one ?
08:35:35 Quit moozooh (Read error: 131 (Connection reset by peer))
08:36:55JdGordonumm... how do u print it the correct size without importing it into h3tool?
08:39:27 Join moozooh [0] (n=moozooh@
08:40:51BHSPitLappytrial and error? ;)
08:42:00 Join Paul_The_Nerd [0] (
08:42:09 Part Paul_The_Nerd
08:45:01 Join needleboy [0] (
08:45:50 Join Paul_The_Nerd [0] (
08:48:55 Join webguest46 [0] (
08:54:09 Quit webguest46 ("CGI:IRC")
08:54:54 Quit Rob2222_ ()
08:55:40 Join Rob2222 [0] (
09:05:57 Join einhirn [0] (
09:09:20 Quit needleboy ()
09:09:52JdGordon:( it didnt work...
09:09:59JdGordonh3tool screwed it up...
09:10:39 Join needleboy [0] (
09:12:00 Join DT291 [0] (
09:15:33JdGordonBHSPitLappy: your a freeking legend :D
09:15:49JdGordonhad to cut up your jpg tho so it would print correclty
09:15:53***Saving seen data "./dancer.seen"
09:16:22BHSPitLappyis it on?
09:16:47JdGordonnot yet... did a test print here, have to wait till i an get to a mates to print it in colour
09:19:38 Join perpleXa [0] (
09:24:11 Quit DreamTactix291 (Read error: 110 (Connection timed out))
09:24:43 Join damaki_ [0] (
09:24:55BHSPitLappybe sure to take pictures.
09:24:59JdGordonwill do
09:34:24JdGordonfear my leet mspaint skillz :D ive added fire on the side of the screen :D
09:34:31JdGordondinner.. bbs
09:37:15 Join damaki__ [0] (
09:40:07 Quit Higgy_ ("Trillian (")
09:40:13 Quit damaki (Read error: 110 (Connection timed out))
09:40:54BHSPitLappyoh dear
09:41:17midk.. yeah..
09:41:18 Quit XavierGr (Excess Flood)
09:42:42 Join krmathis [0] (
09:45:00 Join damaki [0] (
09:45:16 Join XavierGr [0] (
09:50:32 Quit damaki_ (Read error: 110 (Connection timed out))
09:53:58Papricawhich target is this?
09:53:59Paprica#elif (LCD_WIDTH == 160) && (LCD_HEIGHT == 128) && defined(HAVE_LCD_COLOR)
09:55:44 Quit needleboy (Read error: 110 (Connection timed out))
09:57:49 Join petur [0] (
09:58:59Papricasmall screen @@
09:59:06midksurprisingly low-res for a color screen..
09:59:30midkgreat minds think alike!! :)
10:00:06 Quit damaki__ (Read error: 110 (Connection timed out))
10:05:21 Join damaki_ [0] (
10:06:56*petur searches eBay for 'time' but finds no relevant offers :-\
10:07:21midkpetur, how much are you willing to pay? ;)
10:07:39peturthat, is a good question...
10:08:32Bg3rpetur any progress ?
10:09:46peturnone... am at the stage of examining the disassembled code, as it doesn't work by using only the config values from them...
10:10:03Bg3r:( fsck ...
10:11:05perpleXaAR+RANLIB /home/perpleXa/rockbox/build/librockbox.a
10:11:05perpleXamake: *** /home/perpleXa/rockbox/bootloader: No such file or directory. Stop.
10:11:05perpleXamake: *** [all] Error 2
10:11:19perpleXatried to build the bootloader for the mini
10:11:50perpleXaany fix for this?
10:12:07peturBg3r: also, in their code they only seem to access the DC, where I would expect the HC (for usbotg), so either there's something wrong with the way we think it's connected, or I don't understand how the isp1362 works :(
10:12:20midkperpleXa, did you checkout rockbox-devel?
10:12:25midkthat's all i can imagine is the problem
10:12:35perpleXai'll try that, thx
10:12:37midki think only rockbox-devel package has the bootloader/simulator stuff
10:12:41Paul_The_NerdperpleXa: Or just cvs co bootloader
10:12:52Paul_The_NerdIf you've already got the normal rockbox package
10:13:14Bg3rpetur very very strange ... :(
10:13:18peturBg3r: I assume one uses DC to act as a device and HC to act as a host, but now I'm not so sure anymore...
10:13:41Bg3rpetur i've read the isp1362's datasheet and this *must* be correct
10:13:41Papricawhy the configure ask for that: Build (S)DL version? (S)
10:13:41perpleXaok Paul_The_Nerd, did that
10:13:52perpleXait doesn help :/
10:14:06perpleXai'll try the devel package
10:14:08PapricaPaul_The_Nerd, have you try the brickmania for nano?
10:15:06peturBg3r: then why are there only accesses to the DC in the code? Need to study their code a bit more (and learn m68k assembler along the way)
10:15:08Paul_The_NerdPaprica: You can still hit (W) for Win32, it's just hidden because it's not longer officially supported.
10:15:24midkperpleXa, i have "rockbox" and "rockbox-devel" checked out.. only -devel has the bootloader folder, so that should do it.
10:15:40Paul_The_NerdPaprica: Also, yes, I've tried brickmania on the nano. It seems to work fairly well, though you might want to either speed up movement from the scrollweel, or disable it to force people to use the arrows.
10:15:41peturBg3r: unless Linus disassembled a US version, but I don't think so
10:16:25Papricathe speed of the game is ok?
10:16:50Paul_The_NerdPaprica: Basically, if you try to play with the arrows, if you change directions too suddenly it may think you're scrolling instead of pressing the buttons and move you the wrong way. But the paddle moves too slow with the wheel to use it that way.
10:16:55Paul_The_NerdEverything else is fine, in terms of speed, I think.
10:17:03perpleXaare there any themes out for the mini?
10:17:05Bg3rpetur hm
10:17:39midkperpleXa, no, mini support is very recent..
10:17:49 Nick midk is now known as midkay (
10:18:00Paul_The_NerdperpleXa: What's the resolution of the Mini?
10:18:15perpleXawell, i'm too lazy to make my own, spent some minutes yesterday but it sucked :)
10:18:30perpleXapretty small, eh :)
10:19:24 Quit damaki (Connection timed out)
10:19:30ScoTTieim trying to get keypad working on the mini
10:19:32midkay:) none of the existing themes will work directly then..
10:19:37 Join IcyStorM [0] (
10:19:47ScoTTiebut i really dont have much clue what im doing ;)
10:19:49midkaycould be adapted, though.
10:19:49perpleXai know, midkay
10:19:57Paul_The_NerdperpleXa: Yup, someone will have to make themes for them then. Or at the very least, adjust Recorder themes to fit.
10:20:03 Join safetydan [0] (n=dan@
10:20:13ScoTTieperpleXa: what gen iPod mini do you have ?
10:20:27midkayperpleXa, that kind of sucks.. archos recorder display is way smaller and iriver display is way bigger - nothing will work "pretty well" until it's adapted. :)
10:20:28JdGordonpetur: quick q about the colour... is it stored like this 0..0rrrrrggggggbbbbb (as in the bits) so to get the next valid colour just increment the r or g or b section?
10:21:22peturI have no idea - never looked at that part of the code ;)
10:22:41perpleXathat fucking apple software killed 3gb of my data yesterday
10:22:52 Part XavierGr
10:22:55Paul_The_NerdJdGordon: What do you mean "to get the next valid colour?"
10:23:03 Part krmathis
10:23:04perpleXai uninstalled ipl, the apple os rebooted and deleted the partition table
10:23:42perpleXathe files were still there, i've simply recovered them, but it was a shock anyway..
10:23:55peturJdGordon: but yes, my idea is to try to operate on the native data as the easiest way to make sure you only use valid values...
10:24:07midkayperpleXa, ouch.. well, at least you got them back..
10:24:22Paul_The_NerdIf you're using 0-31 for R, 0-63 for G and 0-31 for B, then wouldn't it just be B+(32*R)+(2048*R) to create the packed value, or am I mistaken?
10:24:34perpleXayeah, i leeched some tools and recovered them :D
10:24:51midkayhaha, exactly what i'd've done ;)
10:24:57perpleXathats why i hate this fucking apple software.. on its first start it deletes all files on the ipod
10:25:16midkaythat's pretty weird..
10:25:19JdGordonPaul_The_Nerd: no, its not.. but if rbg is limmited to the 64/32 values then would just incrementing the bits work?
10:25:24JdGordondw.. that made no sense
10:25:28JdGordonill just play...
10:25:36 Join krmathis [0] (
10:26:00IcyStorMWhat is the standard keys for RockBox on the iPod Video
10:26:12ScoTTieperpleXa: still having compile problems ?
10:26:13midkayIcyStorM, standard keys?..
10:26:33Paul_The_NerdJdGordon: Incrementing the bits should work, but I'm pretty sure that would too.
10:26:46perpleXaScoTTie: i'm still downloading the devel package.. with an amazing speed of 56k!
10:26:48IcyStorMI meant Rockboy :P
10:27:06Paul_The_NerdIcyStorM: Rockboy doesn't yet have an A or B button as far as I know.
10:27:10ScoTTieperpleXa: out of interest did you make the RSS TCL script ?
10:27:22perpleXayeah, thats me :)
10:27:28midkayPaul_The_Nerd, you can "set" them, though, or at least it seems to allow you to...
10:27:51Paul_The_Nerdmidkay: What do you mean "set" them?
10:27:53ScoTTielinuxstb: if i was going to add mini keypad support, would i just need to add a new irq() function in system.c ?
10:27:59Paul_The_NerdOh, wait, misread that
10:28:03ScoTTieand some new read functions
10:28:31IcyStorMBut in iPodLinux you just touch to emulate a button. Touch upercorner A and so on.. Not the same here?
10:28:41JdGordonHouse MD time... back in an hour :D
10:28:48Paul_The_NerdIcyStorM: Nobody's added that functionality yet.
10:29:04IcyStorMSo its unplayable ?
10:29:07 Quit quobl (Remote closed the connection)
10:29:08midkayIcyStorM, touching the pad to emulate a button? really?
10:29:10ScoTTieperpleXa: &
10:29:24IcyStorMCuz there werent enought buttons
10:29:31midkayawesome! can't wait for something like that.. as-is it's pretty much unplayable
10:29:39perpleXathx ScoTTie, but its almost finished =)
10:29:59midkayyeah, either that or a load of combos.. play+left for A, select+menu for start, or whatever, etc...
10:30:00 Quit yeahx ()
10:31:08perpleXathere are rumors, throwing it agains the wall help too
10:31:49perpleXaoh man, i'm definately awake too long
10:31:56 Join mikearthur [0] (
10:32:23perpleXathats horrible, after 3 days you see shadows everywhere, everything seems to move :<
10:33:12IcyStorMBut it looks like its not possible to even bind scroll right / scroll left
10:33:13IcyStorMwhy ?
10:33:14ScoTTiesleep deprevation can make you go crazy.
10:33:32midkayIcyStorM, it doesn't seem to be implemented..
10:33:42midkayyou could bind it from the source file, though...
10:34:20IcyStorMHere you guys go
10:34:23Paul_The_NerdIcyStorM: Because Rockboy is older than the iPod port and hasn't been updated much for them. Remember, the iPod port is still fairly new, and gameboy support is hardly a priority.
10:34:42Paul_The_NerdBut, if you have a great interest in it, feel free to pitch in and submit a patch to the tracker
10:34:51perpleXayet another gameboy emulator?
10:35:03linuxstbIcyStorM: The scroll left/right events are handled differently in Rockbox to the other buttons, so Rockboy's current method of reading the button status won't work with them. As Paul_The_Nerd says, it's not a priority.
10:35:05IcyStorMIt is for iPod Linux you can see the control options
10:35:14perpleXaoh nvm
10:35:28IcyStorMTouch there, touch there :P
10:36:58midkayIcyStorM, well, it isn't for rockbox. :)
10:37:44IcyStorMNo but maybe you guys here at rockbox can make the same control options
10:37:48IcyStorMSo it would be playable
10:37:56linuxstbScoTTie: I'm not sure what needs to be done for the mini keypad... If I was you, I would take the keyboard.c from ipodlinux, and delete all the code in that file that doesn't apply to the mini, and then try and copy it into Rockbox, using the existing 4g pad driver as an example.
10:38:17linuxstbIcyStorM: As we've said, Rockboy isn't a priority at the moment.
10:38:30ScoTTieive done that pretty much
10:38:32linuxstbBut we would welcome a patch...
10:38:52ScoTTiebut no dice, so im thinking i might of missed something somewhere else in the source tree
10:39:30 Nick krmathis is now known as krmathis|autoawa (
10:39:51*perpleXa smells
10:40:10perpleXaactually i'm sitting here since 3 days, i only stood up 2 times.. :/
10:40:26perpleXato get some cigarettes
10:41:45midkayhaha. can anyone say 'obsessed'? :)
10:42:06Kingstoneare people interested in building an installer for windows for ipod for rockbox?
10:42:26midkayKingstone, i was considering writing a script for it, actually..
10:42:39Kingstonebatch, you mean?
10:42:53perpleXalinuxstb: someone shoudl update the CUSTOM_WPS_FORMAT file in the package with the content from the wiki!
10:43:00midkaynot a script per-se.. i mean a la the 1.0<->1.1 converter i wrote for ipodwizard
10:43:26Kingstonewhat are you talking about?
10:43:46midkayhaha. per-se... kind of like "exactly".. not a script exactly.
10:43:56midkayKingstone, aren't you an ipodwizard mod?
10:44:12Kingstonebut what converter?
10:44:26midkaysomeone stickied my converter app in the nano and 5g forums, i guess it wasn't you
10:44:30Kingstoneand what does that got to do with rockbox
10:44:41midkaynothing, just an example.. never mind.
10:44:45Kingstoneconverter for..?
10:44:48Kingstoneconverts what
10:44:58midkaythemes.. between 1.0 and 1.1 compatibility
10:45:08midkayjust renaming a series of bitmaps
10:46:32midkayanyways. i could do a quick app to install rockbox on ipods, sure..
10:47:00perpleXaactually its no problem
10:47:12perpleXawith the source of make_fw and that other tool
10:47:16perpleXai forgot the name..
10:47:23midkayperpleXa, what?
10:47:23perpleXabut you guys know what i mean :)
10:47:33perpleXamaking an installer
10:48:08midkayi mean installing the bootloader..
10:48:27midkaymaybe Kingstone was talking about installing the .rockbox folder and rockbox.ipod file..
10:48:58perpleXaeven the bootloader is easy to install, since the most work is already done
10:49:07perpleXaonly thing you have to put around is a gui
10:49:50midkayperpleXa, well, you have to find your ipod number, copy the original bootloader off it, combine that with the rockbox bootloader, extract the broadcom thing in some cases, and then copy it over..
10:50:03midkaylots of command typing, it would be really nice to hit a button and have it all done at once..
10:51:00perpleXawell, bit annoyance is always cool =)
10:51:12Kingstonemidkay, are you using ipodwizard..?
10:51:14perpleXasince windows click-and-finish
10:51:15midkayperpleXa, ah, you like that? well, don't use any scripts then. :)
10:51:23midkayKingstone, yes.. :)
10:51:30perpleXai'm happy when i get something that can go wrong :D
10:51:41Kingstoneon next version you can directly edit the ipod firmware, without a need for updater
10:51:42midkayperpleXa, haha, as you learned when you almost lost 3gb of data?
10:51:50Kingstoneso it will be good for guys with other firmwares
10:51:51midkaydidn't you describe that as "a shock"? :)
10:51:56midkayyou like that kind of shock? :)
10:52:05Kingstoneradioshack? ;p
10:52:08midkayKingstone, _awesome_!
10:52:18IcyStorMIsnt it possible to bind the menu to Center key and Menu
10:52:18midkaythe annoying process of write, downgrade, update...
10:52:30perpleXaKingstone: btw, the current version crashs when i open some firmware parts
10:52:32midkayIcyStorM, not in rockboy..
10:52:40mikearthurI hear album art is _almost_ working for 75x75px bmps called cover.bmp
10:52:44mikearthuris that in current cvs?
10:52:48perpleXanot actually the current one
10:53:09Paul_The_Nerdmidkay: It's possible, with source changes.
10:53:18linuxstbmikearthur: There's a work-in-progress patch on the patch tracker. I believe it works.
10:53:20perpleXayeah Kingstone in that dropdown, when i click on several items it simply crashs
10:53:23midkayPaul_The_Nerd, of course, but not from the actual app..
10:53:29perpleXabut i guess my version is a bit outdated anyways
10:53:31IcyStorMAny guy here that can change a bit in the source to make Rockboy playable
10:53:41mikearthurlinuxstb: cool, is it a patch for all versions?
10:53:48 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
10:54:04linuxstbmikearthur: It will work on any Rockbox player with a colour LCD.
10:54:16Paul_The_NerdI don't think it restricts you to 75x75 does it?
10:54:46mikearthurlinuxstb: I've got a iRiver h140 and want my 4-color art :D
10:54:51linuxstbI think the size is hard-coded at the moment, unless that was changed in the very latest patch.
10:55:05midkayIcyStorM, i suppose i could..
10:55:11midkayyou're on a 5g?
10:55:24linuxstbmikearthur: It could be working on grayscale lcds, I'm not sure.
10:55:29midkaywhat keys did you want bound to what?
10:55:38Kingstonewell v1.2 alpha i didn't release it.. wibu beta tested it and i guess he posted it
10:55:42IcyStorMIs it possible to make it like the Linux build
10:55:44 Join ModernExecutive [0] (
10:55:47 Part ModernExecutive
10:55:56IcyStorMlook that one
10:56:03IcyStorMIs that too much work?
10:56:07IcyStorMto much
10:56:07linuxstbIcyStorM: No - the rockbox button driver doesn't support "taps" at the moment.
10:56:12Paul_The_NerdIcyStorM: It can't be done that way yet, because of the way the button driver is handled at the moment.
10:56:13midkayIcyStorM, again, it's not a priority.. and no, i can't do that. but if you give me key combos you want, i can do that.
10:56:37IcyStorMclick: Menu and Select, Go back to rockboy menu
10:56:56IcyStorMThe center key = select
10:57:29IcyStorMIs it possible to scroll then?
10:58:14IcyStorMMenu button = A
10:58:25IcyStorMCenter= B
10:58:34midkayyou just defined center as "select"
10:58:46IcyStorMThat was a fault
10:58:48IcyStorMim sry
10:58:50IcyStorMCan you fix it
10:58:58linuxstbIf MENU is A, then what is used for Up?
10:59:07midkayum, yes..
10:59:08IcyStorMscroll left is used for up
10:59:52perpleXawoudlnt it be better to buy an gameboy on ebay, for $3.. :P
11:00:05IcyStorMI have a gameboy but I wanna play gameboy on my iPod
11:00:23Paul_The_NerdMaybe just wait until a better method is available for button mapping then?
11:00:27Paul_The_NerdIs it such a rush?
11:00:34IcyStorMI have nice keys
11:00:39IcyStorMHey Midkay
11:00:49midkayIcyStorM, yes?
11:00:56IcyStorMOk i type the buttons in a txt and send you ok ?
11:01:01midkaysounds best
11:03:56 Join imphasing [0] (
11:05:43IcyStorMHmm I cant send cuz im not registered
11:06:11midkayIcyStorM, just list in a PM window
11:06:28IcyStorMI cant send PM's either
11:07:31IcyStorMOH W8
11:07:33IcyStorMI forgot one key
11:07:41IcyStorMBut I know what to bind it to dont worry
11:08:13IcyStorMLol i didnt forget it, its in the file
11:08:16IcyStorMDid you download it?
11:09:20linuxstbmidkay: Rockboy currently uses the button_status() function to check the current status of all the buttons (if they are pressed or released). The scroll events are different - button_status() doesn't include them.
11:09:37midkayIcyStorM, doing it
11:09:46midkaylinuxstb, argh...
11:09:54midkaymaybe i can change that..
11:10:12IcyStorMI hope that it would be really nice
11:10:29IcyStorMI dont know how I can credit you guys. Im a poor 14 year old boy from sweden :P
11:10:35linuxstbDon't forget that Rockboy is an emulator - it emulates the gameboy hardware, and therefore needs access to buttons that are either pressed or released. I'm not sure if the scroll events will fit into that...
11:10:51linuxstbIcyStorM: Learn C and help out...
11:11:12midkaylinuxstb, hmm.
11:11:21IcyStorMIts a nice theory, linuxstb
11:11:26IcyStorMMaybe I should
11:11:42IcyStorMlinuxstb Wouldnt scroll at all time be same as holding a button ?
11:12:13midkayIcyStorM, well, i can't do the scroll right now.
11:12:17midkayso we need alternates for up and down.
11:12:31midkaywhat about Play+Rewind = down, Play+Ffwd = up, or something?
11:12:44linuxstbIcyStorM: I'm sure something can be done, but I'm just saying it's not as easy as just changing the button mappings.
11:14:19IcyStorMNo i apreciate that
11:15:14IcyStorMMidkay I think that Center + Back and Center + Fwd would be better, Because play is bound to pause and you would pause the game every time you wanted to go up or down
11:15:15IcyStorMSo try that
11:15:16IcyStorMOk ?
11:15:42midkayIcyStorM, no, for example center+menu = menu and center = b.
11:15:57midkayi set it up so that a single press of center is b, but if it's held and then menu is pressed, it goes to the menu.
11:15:58***Saving seen data "./dancer.seen"
11:16:10midkayso any key can be used in conjunction with another, even if it's already mapped.
11:16:37Papricasomeone with recorder/ondio?
11:17:17IcyStorMBut I want Center + rewind = down and center + back = up
11:17:27perpleXabbl, bye
11:17:29 Quit perpleXa ("( :: NoNameScript 4.01 :: )")
11:17:34IcyStorMits easiser to press center and rewind at same time cuz they are next to each other
11:17:47IcyStorMI can do it with one finger :P
11:17:54midkaycenter + fwd = up, you mean?
11:18:04IcyStorMBut maybe
11:18:09IcyStorMnever mind..
11:18:13IcyStorMYeah i think so
11:18:20IcyStorMOr w8
11:18:22IcyStorMI have a new idea
11:19:15IcyStorMIf we skip RockBoy menu with menu and center
11:19:19safetydanAnyone tried the SDL sim under Windows lately? Does sound work?
11:19:30PapricaSomeone with recorder/ondio? =\
11:19:33midkaysure, alright. let's make menu, like.. center+rewind..?
11:19:36IcyStorMMidkay do you have msn / icq we can take it there ?
11:19:49midkayIcyStorM, both. or 343603744
11:19:53IcyStorMThen we are done we can send it here to everyone else
11:20:36Paul_The_Nerdmidkay: Is that an ICQ number?
11:20:45midkayPaul_The_Nerd, yes :)
11:20:53Paul_The_NerdMaaaan... Haven't seen one of those in ages
11:21:05Paul_The_NerdThey've added a few digits since I used it.
11:21:07midkayPaul_The_Nerd, haha, it's for those people who still want to use it. :)
11:21:24Paul_The_NerdI still sign on every day, but my list is long emptied.
11:21:39midkaywe can add one more to each others' list... :)
11:21:56Paul_The_NerdFeel free. 5143014
11:22:16*petur hugs his jabber client :P
11:22:38 Join damaki [0] (
11:23:07Paul_The_NerdI use the Google jabber server.
11:23:27*Paul_The_Nerd has found the need to use most IM protocols at one point in time or another, because his friends can't all just pick one.
11:23:53peturare they linked to the rest of the jabber network yet?
11:24:05Paul_The_Nerdpetur: No clue.
11:24:18Paul_The_Nerdsafetydan: I'm building the SDL/Win sim right now, I'll let you know.
11:24:19 Quit petur ("CGI:IRC (EOF)")
11:24:31 Join petur [0] (
11:25:01peturthere are gateways to most other IM, they work great
11:25:18safetydanPaul_The_Nerd, thanks
11:25:42Paul_The_Nerdpetur: I just use GAIM, it works pretty alright for me. is a good gateway
11:28:02Paul_The_Nerdsafetydan: I get about 1 second of audio, then the sim exits
11:28:12safetydanUpdated to latest CVS?
11:28:28safetydandarn, it works well under Linux
11:28:38safetydannow I'm really going to have to reboot in to Windows to try and fix it
11:28:47safetydanthanks for testing
11:29:01Paul_The_NerdTried it with flac and mp3, just to be safe. Same behaviour for both
11:29:47IcyStorMlinuxstbWhere do I found C tutorials ? Im a big newb but I really wanna help you guys
11:29:50safetydanyeah if there's a problem it's usually codec independant
11:30:27IcyStorMI have good computer knowledge and a little bit Html knowledge. No programming knowledge at all
11:30:41safetydanIcyStorM, google for "learning c" brings up a few places to try
11:33:21linuxstbIcyStorM: The first thing to do is to follow the instructions on the Wiki and compile Rockbox yourself from source. You can then try changing things - looking at some of the simpler plugins would be a good place to start.
11:35:06 Join perpleXa [0] (
11:35:20perpleXabullshit that is..
11:35:23 Quit damaki_ (Read error: 110 (Connection timed out))
11:35:58perpleXaScoTTie: making progress with the keyboard? :P
11:36:09 Quit segphault (Read error: 104 (Connection reset by peer))
11:36:33ScoTTieits not even reading anything at all
11:36:46perpleXanoticed that
11:37:09perpleXaonly thing i could do was plugging it on my pc
11:37:16 Join segphault_ [0] (n=chatzill@
11:37:19 Nick segphault_ is now known as segphault (n=chatzill@
11:37:28ScoTTieatm its set to read as though the iPod's a 4G one
11:38:17perpleXagimme the part of ipodlinux that contains the mini keyboard handler and i'll try to fix itr
11:38:51perpleXai hope it wont be that hard..
11:39:11ScoTTieyou will need to add it
11:39:22ScoTTiethe buttons arnt compatible with any other model afaik
11:40:36*safetydan wonders where these reports of the default EQ frequences being backwards are coming from
11:40:57perpleXawell but it must already been done
11:41:07perpleXacos ipl runs on my mini
11:41:09ScoTTieit is done in iPL
11:41:13perpleXaeven with keys
11:42:15Paul_The_Nerdsafetydan: I dunno, my LS and HS are low and high respectively...
11:42:15 Join tvelocity [0] (
11:42:43ScoTTiesec ill up my diff aswell
11:45:44safetydanPaul_The_Nerd, yeah I just did a settings reset on my h120 and the defaults are the right way around... (I'd already verified the code)
11:46:34 Join webguest13 [0] (
11:46:37 Quit tvelocity (Client Quit)
11:47:10ScoTTieperpleXa: &
11:47:29 Quit perldiver (Read error: 110 (Connection timed out))
11:48:41safetydanHrmm... needleboy's not around. I suspect something in the H300 optimised build
11:50:00 Join marevalo [0] (
11:50:09midkaylinuxstb, are you around?
11:50:15perpleXaweird changes that you've made ScoTTie.
11:50:20perpleXa-#elif (CONFIG_KEYPAD == IPOD_4G_PAD)
11:50:20perpleXa+#elif CONFIG_KEYPAD == IPOD_4G_PAD
11:50:25ScoTTieyeah ignore them
11:50:27 Join tvelocity [0] (
11:50:40ScoTTiei had added IPOD_MINI_PAD
11:51:03 Part tvelocity ("Αποχώρησε")
11:51:06ScoTTiethen i removed it and just used defined(APPLE_IPODMINI) instead where needed
11:51:37ScoTTiesince most app's use the PAD def as what the button layout is.
11:51:47Paul_The_Nerdsafetydan: Isn't lovely how any time there are bugs in that, they're assumed to be in the original code?
11:53:43 Join LinusN [0] (
11:53:46 Part LinusN
11:54:05 Join LinusN [0] (
11:55:37linuxstbmidkay: Yes.
11:56:06midkaylinuxstb, apparently the rockboy method of scanning for keys also doesn't allow key combos.. any way around that?
11:56:36 Nick krmathis|autoawa is now known as krmathis (
11:57:20perpleXaif i just wouldnt be so lazy :/
11:57:38ScoTTiefix it, fix it, fix it, fix it, fix it, fix it..
12:01:20linuxstbmidkay: I don't know. That's why the ipod keymapping is the way it is now - it's not a trivial thing to change.
12:01:43 Join Jungti1234 [0] (n=jungti12@
12:02:51midkayoh well, thx
12:05:39JdGordonlets play.. spot the parse error game....
12:05:49JdGordoncygwin says parse error before the ;
12:05:52JdGordonhere is the line....
12:05:54JdGordonsnprintf(buf,32,"RGB: 0x00%02x%02x%02x",RGB_UNPACK_RED(recombined_val),
12:05:54JdGordon RGB_UNPACK_GREEN(recombined_val),RGB_UNPACK_BLUE(recombined_val));
12:06:19perpleXathat ,
12:06:29perpleXaignore what i said lol
12:06:33JdGordonoh found it... i can blame linuxstb :D
12:06:37ScoTTiecouldnt it be on a line before that ?
12:06:43JdGordonsilly man put ; at the end of the #defines...
12:06:54perpleXadidnt see that second brace
12:07:14 Part petur
12:08:19Paul_The_NerdJdGordon: Right now, when you tell it to reset the colors, it sets the BG to Black and the FG to black. Which then results in everything being fairly unreadable. ;-)
12:08:35JdGordonye, found that.. fixed it
12:09:09JdGordonoh crap.. this almost works...
12:10:53Jungti1234You have screenshot? -
12:10:53linuxstbJdGordon: Just testing you.....
12:11:05JdGordonwow.. way to much free time perpleXa
12:11:26JdGordonlinuxstb: it was an accident? it looks like if coz u got extra ;'s in the code
12:12:48linuxstbJdGordon: Yes, just a copy and paste mistake - those calculations were originally in the code itself, and then I created macros from them.
12:13:56Papricalinuxstb, when i do "cvs commit file" the change will see in the first page?
12:14:14linuxstbPaprica: Yes, but not immediately. I'm not sure when it's updated.
12:14:18perpleXaonly if you have access to change files
12:14:36Papricaoh =\
12:14:48marevalohi all
12:14:48linuxstbIf you subscribe to the cvs mailing list, you should get an email quite quickly.
12:15:13Papricaok so i will do it with message =]
12:15:14marevaloI've just finished the port of GNUChess 2 to RockBox
12:15:21linuxstbmarevalo: Nice :)
12:15:33marevaloit works almost on any target (at least at the uisim)
12:15:40perpleXasome working keys would be nice :P
12:15:57marevaloI'm trying to add it to the patchmachine
12:16:02marevalobut I can only add one file
12:16:18linuxstbEither create a zip file, or attach the files one at a time.
12:16:20marevaloI added the .patch file but I have another .zip with the new files
12:16:43linuxstbWere you logged into sourceforge when you uploaded your first patch?
12:16:57LinusNmarevalo: c0000000000l!!!
12:17:10marevaloI have no usable user
12:17:21JdGordonchess against the cmop? or another person?
12:17:25linuxstbmarevalo: That's the problem I think.
12:17:30marevaloagaint the comp
12:17:39linuxstbDo you have any screenshots?
12:17:48marevalook, I will try to recover an old user
12:17:54Jungti1234hehe.. good JdGordon
12:17:56marevalook, I will post som screens
12:18:18JdGordonnice (y)
12:18:23linuxstbDid you have any problem with memory usage? Does it require audio playback to be stopped?
12:18:25JdGordontoo bad the cmop will shit on me :D
12:18:49IcyStorM Chip8 games for those who wants
12:19:08marevalowell it works with music, of course :-D
12:19:16 Nick krmathis is now known as krmathis|autoawa (
12:19:19marevaloI had to take <64k from the buffer
12:19:20Papricamarevalo, sweet =]
12:19:30marevalofor the trasposition table
12:19:44linuxstbWhich buffer do you take that from? The plugin buffer?
12:20:51linuxstbDo you mean rb->plugin_get_buffer() ?
12:20:58marevalolet me see
12:21:24JdGordonis the sim 565swapped? or unswapped?
12:21:28marevalottable = rb->plugin_get_buffer( &buffer_size );
12:21:56linuxstbThere is going to be a problem on the Archos players - the plugin buffer is only 32KB.
12:22:17marevaloI can switch it off
12:22:22 Join Membrillo [0] (
12:22:28marevaloit can play without transposition table
12:22:49marevaloand the others ?
12:22:49linuxstbWhat does the transposition table do?
12:22:56linuxstbThe others have 512KB.
12:23:08linuxstbThis includes the space for the plugin itself (code and data).
12:23:19marevaloit saves positions and it evaluations for not needing to evaluate them if the same position is found later
12:23:28linuxstbplugin_get_buffer() returns a pointer to the unused space in the plugin buffer.
12:23:49marevaloso it should work without the ttable
12:24:26marevalotry to get some screens...
12:24:35linuxstbBut anyway, these are minor details - I'm looking forward to testing it.
12:24:52linuxstbJdGordon: The sim is the same as the target.
12:25:20JdGordonive almost got the sliders working showing the native rgb and the 24bit values :D
12:25:49 Part krmathis|autoawa
12:26:31 Quit Membrillo (Client Quit)
12:26:50linuxstbmarevalo: How big is the "gnuchess.rock" file itself? Have you compiled it for any actual targets (not the sim) ?
12:27:21Papricanice screendump of brickmania menu
12:27:40Papricawhat happend to the screendump save? =\
12:27:54linuxstbWhich target is that from?
12:27:58marevaloyep, and played on my h120 : 31694 bytes
12:28:05marevaloI'm on it
12:28:09JdGordonhahaha, nice
12:28:43linuxstbmarevalo: It looks like it probably won't fit in the Archos 32KB plugin buffer anyway.
12:29:15marevalowell, search is recursive, it may take some stack too
12:29:49marevaloanyway I made 64x64 pixels boards, because I want to play it on my h120 remote (remote playing not yet working)
12:29:56linuxstbplugins use the stack from the main thread - that's not included in the 32KB/512KB.
12:30:11marevalook, I can make som adjustements
12:30:20Paul_The_NerdLooks nice
12:30:28 Part IcyStorM
12:30:28marevaloI'm a debian guy, I like the leave nobody behind motto
12:30:44linuxstbHow many different board sizes have you implemented?
12:30:57linuxstbIs there a large one for the iPod video (320x240 LCD) ?
12:31:01marevalo64 128 176 240 I think
12:31:15marevalothe bigger pieces are from eboard
12:31:43marevaloI have tested it on the Recorder UI sim, if that matters
12:31:44linuxstbIt's always nice when somebody turns up out of the blue with a brand new plugin...
12:32:02marevalowell rockbox gave new life to my player
12:32:11marevaloit's just giving back
12:32:38Jungti1234oops Paprica, What's that?
12:33:11linuxstbmarevalo: It could be useful if you installed the cross-compiler for the Archos targets. It will tell you if your plugin is too big to fit in the available plugin RAM.
12:33:37marevalowell I'm fed up with source forge
12:34:04linuxstbWe all are - we are about to move the bug/patch/feature tracker to
12:34:25marevalowell, if someone can upload the zip
12:34:32marevaloand probably fix it
12:34:49linuxstbI'll have a look now.
12:34:59marevaloI've just started to work with rockbox this weekend at my parents home
12:35:20marevaloand I'm not quite used to cvs and patch :-P
12:35:25Paul_The_Nerdsafetydan: I'm not sure if he realized you meant a non-optimized build when you said a daily, since when you asked about version he mentioned that he'd tried several versions of the optimized from the sounds of it.
12:36:14marevaloI wanted to upload it because I will try to get some menus, opening books (I will need some memory more) and file mangling
12:36:39marevaloand it will be unstable for a week or so, I think it is quite stable and usable now
12:36:59linuxstbmarevalo: zip file now attached to your patch thread.
12:37:16JdGordonlinuxstb: do u still have that txt file with the values from the test program u did for the colours? im getting odd values and dunno if i should be
12:37:38linuxstb (I think)
12:37:50linuxstbThat's it....
12:37:54*Paul_The_Nerd grins
12:37:58Paul_The_NerdI was looking at it earlier.
12:38:24linuxstbBut compiling that rgb.c program would give a similar output.
12:39:00perpleXahmm the bootloader is doing quite funny stuff on the mini
12:39:09safetydanPaul_The_Nerd, probably, but it sounds like he's using the eq setting files from the optimised build and those are in the wrong order
12:39:17perpleXamirroring the display and such :)
12:39:24 Join Kohlrabi [0] (
12:39:28marevalolinuxstb: do you need anything more from me?
12:39:38perpleXawith the apple software i mean
12:40:29JdGordonhmm... i dont think that actually helps :p.. green shouldnt go to (8bit)0 at (native)32 should it?
12:40:36linuxstbmarevalo: I'm about to test it on an ipod now..
12:40:42Paul_The_NerdperleXa: That's something that happens on all grayscale iPods at the moment, I think.
12:41:09Paul_The_NerdJdGordon: If it starts at 0, there should be no native 32. It should end at 31.
12:41:17Paul_The_NerdOh, wait, Green
12:41:21Paul_The_NerdGreen should be 0-63, right?
12:41:28marevalolinuxstb: ok, I tried it on the uisim for the ipod color and seems to work fine, the ipod video will be good chess playing machine
12:41:47JdGordonfor some reason it goes 0->31 then the 8bit value goes back to 0
12:42:18Paul_The_NerdWhat's the 8-bit val at 31?
12:42:31perpleXaPaul_The_Nerd: looks funny
12:43:08Paul_The_NerdAnd the variables, they are unsigned?
12:43:09perpleXawell the bad thing is, the apple software is now pretty f*cked and rockbox doesnt respond :'(
12:43:25JdGordonye, unsigned char
12:43:55 Quit webguest13 ("CGI:IRC")
12:44:26 Quit Jungti1234 ()
12:44:50Paul_The_NerdJdGordon: Do you have any idea what's doing it then?
12:45:23JdGordonbut i just saw abother bug.. which could be related...
12:45:24linuxstbmarevalo: I can't seem to exit on the ipod. But don't worry, I can look at the button mappings and fix it.
12:45:57linuxstbOne question - is eboard a GPL'd program? i.e. is it OK for us to use the bitmaps?
12:46:26marevalolinuxstb: yep there aren't many buttons on the ipods, I implemented the cascadinf for rows for the ipod users :-D
12:46:42marevalolinuxstb: I don't really like the wheel ;-)
12:46:44Paul_The_NerdJdGordon: If you want to pastebin it, I (and maybe others) could look it over. Not sure that I personally could help, but more eyes can make a difference.
12:46:54JdGordonye, will do
12:47:19marevalolinuxstb: the buttons asignements are copied almost verbatim from the bejeweled code
12:47:23linuxstbmarevalo: I will probably change the button mappings to use MENU/PLAY for up/down, instead of the scroll wheel. I need to change that in other plugins as well - e.g. sudoku.
12:47:34JdGordoncrap :p found the bug
12:47:40JdGordoncopy/paste bug AGAIN :'(
12:47:49JdGordongot the wrong bit mask for green..
12:47:51linuxstbmarevalo: Yes, bejewelled was a good choice.
12:47:59marevalolinuxstb: as you wish, I have no ipod to try, and figured up the assignements
12:48:16linuxstbmarevalo: I have to go now, but if no-one else has done it by then, I'll try to commit it this evening.
12:48:27marevalook, big thanks
12:49:04JdGordonits working
12:49:11perpleXaJdGordon: you should even write more and paste less :D
12:49:11*JdGordon feels stupid
12:49:22linuxstbOne issue though - including a .c file from another .c isn't good.... :) I think it would be better to merge them, even though they are very big.
12:49:45marevalolinuxstb: as you see fit better
12:49:48JdGordonyour not talking to me r u?
12:49:53linuxstbJdGordon: No, sorry...
12:50:05marevalolinuxstb: as you commit it I will use it later
12:50:10Paul_The_Nerdlinuxstb: Out of curiosity, have you ever seen a situation in which including a .c file didn't work, but copying and pasting the code into right where the #include was did?
12:50:10JdGordonhaha no.. plame perpleXa, i saw his higlight and just kept reading :p
12:50:19JdGordondidnt c your name between the lines
12:50:23linuxstbPaul_The_Nerd: No, but it just feels wrong...
12:50:40amiconnIncluding .c files is evil
12:51:02JdGordonlinuxstb: last chance.. we want to store the 8bit values to disk right?
12:51:19marevaloby the way, coding for rockbox is really really fun, congrats to all of you
12:51:32Paul_The_Nerdlinuxstb: Yeah, I agree it feels wrong. I did it in a project I was working on in school once, so that it was easier for me to hand off parts of it to teammates (with the intent to paste the final version in) and when one day our build stopped working, the professor complained a very large amount about the includes, pasted the code in, and our problem vanished (it was actually some strange drawing bug) and insisted that the inclu
12:52:01linuxstbJdGordon: That's my opinion, and I think others agree....
12:52:06Paul_The_NerdIt drives me nuts to this day not knowing the real cause of the problem
12:52:07JdGordonok, fine
12:52:16JdGordonshould be ready with a new patch in a min
12:52:23*Paul_The_Nerd agrees about storing the 8-bit values.
12:53:40 Nick marevalo is now known as marevalo_away (
12:54:08 Join petur [0] (
13:03:41 Quit Rob2222 ()
13:04:12PapricaCC chessbox.c
13:04:12PapricaIn file included from chessbox.c:38:
13:04:12DBUGEnqueued KICK Paprica
13:04:12Papricagnuchess.c:2251: warning: function declaration isn't a prototype
13:04:12Papricagnuchess.c:2318: warning: function declaration isn't a prototype
13:04:12***Alert Mode level 1
13:04:12***Alert Mode level 2
13:04:12Papricachessbox.c:161: warning: function declaration isn't a prototype
13:04:13***Alert Mode level 3
13:04:13***Alert Mode level 4
13:04:13Papricachessbox.c:288: warning: function declaration isn't a prototype
13:04:15PapricaLD chessbox.elf
13:05:21linuxstbPaprica: Yes, I get those warnings as well. I'll fix them before committing to CVS.
13:05:36Papricaohh commit =]
13:05:39linuxstbAlso, the plugin is too big for the Archos devices, so I'll have to disable that for now.
13:05:42Bg3rso what, a new plugin ? (chess ?)
13:05:48 Join IcyStorM [0] (
13:05:50linuxstbBg3r: Yes - it's on the patch tracker.
13:05:56Bg3rvery good :)
13:06:12linuxstbFeel free to commit if you want, I won't be able to do it until this evening.
13:06:15Bg3rlinuxstb btw, amiconn has made a rockboy.ovl for archos
13:07:33linuxstbThere are three problems with the patch though: 1) It #includes a .c file; 2) gcc gives a few warnings; and 3) It's too big for the Archos (but it works in the sim).
13:08:17linuxstbBut yes, using amiconn's ovl method should work. Unless another solution can be found.
13:09:08JdGordonstill no debug with the sdl sim?
13:09:52Bg3rlinuxstb what does the .zip file contain ?
13:11:21linuxstbThe rest of the patch - the new files and the bitmaps
13:14:14***Alert Mode OFF
13:14:46amiconnHow big is the plugin?
13:14:50amiconn(including .bss)
13:16:02***Saving seen data "./dancer.seen"
13:16:11Bg3ramiconn give me 1-2 mins
13:17:00Bg3rin fact, how can i gather this info?
13:17:06amiconnCheck the .map
13:17:33Bg3rcan i tell you the coldfire size ?
13:19:10amiconnYes. SH1 will usually end up a tiny bit smaller
13:19:58amiconnCode sizes usually relate SH1 < coldfire < arm, with the difference between coldfire and SH1 being considerably smaller than between arm and coldfire
13:21:20 Join damaki_ [0] (
13:21:21Bg3rokay, what to look in the .map file for ?
13:22:01linuxstbThe .bss is 0x93e0 bytes
13:22:24linuxstbThe plugin itself is about 0x6a00
13:22:41amiconnCheck .plugin_start_addr (should be 0x32f80000 for iriver) and .plugin_end_addr
13:22:52amiconnThen subtract start from end
13:23:03linuxstbPlus according to the author there is another 64KB table that is assigned from the plugin buffer. But that's just needed as a speed optimisation.
13:23:17 Nick marevalo_away is now known as marevalo (
13:23:24marevalohi, i'm back
13:23:28 Join muesli__ [0] (n=muesli_t@
13:23:46marevaloyes, after commit I will fix for the archos
13:23:55marevalo(once I build the toolchain)
13:24:09linuxstbplugin_end_addr is 0x9207fcc, and start is 0x91f8000
13:24:51marevaloHow can I detect 32K targets ?
13:24:58amiconnTwice as much as we have...
13:25:42linuxstbIt seems a shame to have to stop playback for 32KB...
13:25:44Bagdermarevalo: PLUGIN_BUFFER_SIZE
13:25:47marevaloand how can I know that I'm within limits, I would really like it to work on the smaller targets
13:26:07amiconnWe can make it an .ovl for archos
13:26:11linuxstbmarevalo: If you try to cross-compile it for the real target, gcc will tell you.
13:26:21marevalook, will try it through this week
13:26:37amiconnThat requires it to be a multi-file plugin with its own dir
13:27:04marevaloI am a chess engine coder myself (although I'm using gnuchess this time) so I will probably be able to fine tune the engine itself
13:27:07amiconnIf you do this, the rest is fairly trivial
13:27:17amiconnI can do that; I already did so for rockboy
13:28:01amiconnIn fact I "invented" the .ovl method, and the new plugin loader/header is very similar to the .ovl mechanism
13:28:17linuxstbBut that would still stop playback?
13:28:21marevalook, do whatever you want with the code, I will checkout it in a couple of days and keep working on it for the new features (trying to keep compatibilty)
13:28:50linuxstbmarevalo: Was the source originally in lots of files, or is "gnuchess.c" how it was originally?
13:29:14marevalolinuxstb: it's code from the eighties, it was in one file
13:29:45marevalowell, it was splitted, it seems that some editor at the time could edit only 64k text files
13:30:16marevalothat's also the reason for some warnings, it was in k&r C
13:30:39Bg3rlinuxstb should i work on commiting it (without the .ovl part of course, but in separate dir) or someone other is willing to do it now:)
13:30:40marevaloI tried to fix almost all of them, but it seems I missed someone
13:30:58amiconnI should make the overlay loader a plugin lib component
13:30:58linuxstbBg3r: Please do - I'm about to go out.
13:31:09amiconnMakes it even easier to create overlays
13:31:42marevalowhat are the requirements/implications of the ovl method ?
13:32:33amiconnThe plugin itself needs almost no work
13:32:40linuxstbIIUC, The main implication is that it will stop audio playback - the plugin is loaded into the audio buffer instead of the plugin buffer.
13:32:51amiconnIt will just be linked to a different address, the header structure is the same
13:33:19amiconnThe only thing to care about is when the plugin allocates memory dynamically
13:33:39amiconn...and the remaining plugin buffer is too small
13:33:58linuxstbThat's the case here - the plugin gets 64KB from the plugin buffer.
13:34:11amiconnThen the allocation has to be done from the audio buffer, and care has to be taken that the plugin doesn't overwrite itself
13:34:18marevaloit's nice to play chess while playing, and it doesn't need sound
13:34:35marevalocan it be made optional
13:34:41amiconnlinuxstb: Could be easier to allocate this mem statically then
13:34:57marevaloI mean, I want to be able to hear music while playing chess
13:35:19linuxstbmarevalo: Is there a reason you used rb->plugin_get_buffer(), instead of just defining a static array of 64KB?
13:35:34amiconnmarevalo: You still can on units with large enough plugin buffer
13:35:43marevaloheh, with the static array it gave me some problems in the gnuchess code
13:35:58marevaloand I didn't have te time to check it
13:36:20 Quit damaki (Read error: 110 (Connection timed out))
13:36:21linuxstbYou could do something like "unsigned char my_storage[65536];" and then "my_pointer = &my_storage[0];"
13:36:23marevalomaybe I have forgotten some C, I'm a python coder mainly now
13:37:00marevalowell, if fact this solution is commented out in the code, but gave me problems
13:38:14amiconnlinuxstb: Why not just my_pointer = my_storage; ?
13:38:44linuxstbBecause I'm confused about arrays and pointers...
13:39:57linuxstbAre you saying that that is equivalent?
13:40:21amiconnthe symbol of a one-dimensional array represents its address
13:40:34amiconn..which is the same as the address of the 0-th element
13:40:38marevaloaaaahh, the joys of C
13:40:44CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
13:40:44*amiconn likes pointers
13:43:08marevaloheh, my first C program was a Chess engine, with all the tree stored with pointers, I now enjoy pointers :-D
13:43:37linuxstbDo you have any idea how deep the recursion is? The main stack is 8KB IIUC.
13:44:00marevalowell, there is a hard code limit of 39 half moves
13:44:07linuxstbamiconn: Is the Archos stack the same size?
13:44:09JdGordongah im sick of this bloody code... finaly got it :D
13:44:36 Join Mongey [0] (n=mongeyc@
13:44:38marevalosorry, 29, so 28 recursive calls of "search"
13:46:10marevalobtw, don't remove comments as they are quite useful, they are the hooks for the UNIX frontend and quite useful for adding features
13:46:21 Join ender` [0] (i=ychat@
13:48:14amiconnlinuxstb: Yes, main thread stack is 8KB on all targets
13:48:15 Quit Furious_G (Read error: 104 (Connection reset by peer))
13:48:49 Join Furious_G [0] (
13:50:58JdGordonlinuxstb: it works 100% now.. except if the colour u want to load (from the config file, so in 24bit) is not a correct colour... specifically.. the default bg colour doesnt work... ?any idea how to get the closest match?
13:52:31peturconvert and then the reverse?
13:52:40JdGordonye, but thats yuk :p
13:52:43JdGordongonna have to tho
13:52:46linuxstbThat's your only choice.
13:53:07*JdGordon thinks its funny the default colour isnt really valid :D
13:54:51ScoTTiewhere is the irc() function called from ?
13:55:56 Nick marevalo is now known as marevalo_away (
13:56:10linuxstbScoTTie: It's called when an interrupt occurs.
13:56:59linuxstbThe code that sets up the interrupt handlers is in firmware/crt0.S I think.
13:57:05ScoTTieim trying to debug it using lcd_puts(), but nothing is ever printed to screen..
13:57:18 Join Nico_P [0] (
13:57:21ScoTTieim guessing something is messed with the irq's
13:57:34ScoTTieunless lcd_puts() wouldnt work in there for some reason
13:57:44linuxstbAre you calling lcd_update() afterwards?
13:58:02ScoTTiewhat ive got there should work
13:58:27ScoTTieso im thinking the button problem for mini's is abit deeper that i original thought
13:59:24linuxstbIt's a bad idea to call the lcd functions from an interrupt handler - but I think I remember preglow saying he did it when he was developing the button driver.
14:00:05linuxstbThe way I debugged the audio driver interrupt handler was just to increment a global variable every time it was called, and then display that value in one of the screens in the debug menu.
14:00:26ScoTTiei cant really open the debug menu though :p
14:00:32linuxstbTrue :)
14:00:45JdGordon look good?
14:00:53LinusN"make clean" doesn't work as expected with the new bitmap system...
14:00:57linuxstbYou could add some debugging code to apps/main.c
14:01:20linuxstbLinusN: You mean it remakes the bitmap libraries?
14:01:24Paul_The_NerdJdGordon: I think so, yes. :)
14:01:43LinusNit should remove them
14:01:50linuxstbThe problem is that now the dependencies file itself depends on the bitmap libraries.
14:02:07linuxstbIt should remove them in the end though?
14:02:36LinusNmaybe it does, i didn't check
14:02:56linuxstblostlogic's suggestion was to simply change the clean: target to be something like rm -r apps firmware *.[oa]
14:02:57LinusNapps/bitmaps/Makefile doesn't even have a "clean" target
14:03:27LinusNlinuxstb: sounds fair to me
14:04:51Bg3rlinuxstb: heelp :)
14:05:34Bg3rnow ld complains about the chessbox_pieces...
14:05:54JdGordon finished!
14:07:20amiconnlinuxstb: It is possible to call lcd functions from an interrupt as long as you take care not to call *certain* lcd functions from outside the isr at the same time
14:07:33amiconnIn fact that's how the grayscale lib works
14:07:57linuxstbBg3r: For which target?
14:08:08preglowlinuxstb: any luck with cpu boosting?
14:08:21LinusNJdGordon: how does the color picker look with a larger font?
14:08:22Bg3rlinuxstb it doesn't matter :)
14:08:30Bg3ri'm building it for h300
14:08:32linuxstbpreglow: No - I've not been able to deduce anything apart from the fact that it sometimes hangs when changing speed.
14:08:43linuxstbBg3r: Have you tried a make clean?
14:09:22JdGordonLinusN: looks fine
14:09:24Bg3rlinuxstb trying now
14:09:25linuxstbBg3r: The problem may be that you need to adjust the Makefile in the subdirectory to link against the bitmaps.
14:09:46JdGordongrr.. the patch didnt add the colour_picker.c have to redo it :'(
14:09:46linuxstbThis will be the first "subdirectory plugin" to use the external bmp files.
14:09:53Bg3rlinuxstb probably the last one :)
14:10:13Bg3rhow should i do this ?
14:10:34Paul_The_NerdJdGordon: You could just upload the colour_picker.c file alongside it to the tracker
14:10:48JdGordonhe's a thinker! :D
14:10:49Paul_The_NerdWith a note saying "Put this in apps/gui/" or something
14:12:14linuxstbBg3r: Firstly, you will have to simply add the -lpluginbitmapsnative lines in the Makefile. But I'm not sure how you can take care of the dependencies.
14:12:54JdGordonLinusN: as long as the font isnt HUGE it looks fine... CourB14 is too big.. but the rest is ok
14:13:36JdGordonhmm... actually jackash doesnt look ok :( so ill fix that
14:13:36linuxstbYou should probably edit apps/plugins/Makefile to pass the value of $(LINKBITMAPS) to the chess Makefile.
14:14:55linuxstbAnd then also change apps/plugins/Makefile to make $(SUBDIRS) depend on $(BITMAPLIBS) (line 189)
14:16:19Bg3ri'm lost :)
14:16:29*safetydan posts long winded explanation of the EQ to misticriver...
14:16:35safetydanHopefully someone will correct me if I'm wrong
14:16:50*HCl yawns
14:17:02*JdGordon kicks self
14:17:47linuxstbBg3r: I've got to leave now. If you don't work it out, post a patch with your current changes, and I'll pick it up this evening.
14:18:02Bg3rlinuxstb okay, 10x
14:18:40Paul_The_Nerdsafetydan: I'm curious why they were writing the files from highest to lowest in the first place.
14:20:32 Nick perpleXa is now known as \o (
14:20:49Bg3rhow could i add a new dir to cvs ?
14:21:33JdGordonLinusN: is a screeny with a larger font... it sort of falls apart witht he huge fonts?
14:21:54JdGordonthe colour and sliders will always show.. just the text can disapear
14:22:44Paul_The_NerdMaybe don't allow customizing of the font?
14:22:50 Nick \o is now known as perpleXa (
14:29:06safetydanPaul_The_Nerd, for some reason people just thought the defaults were wrong and assumed the EQ went from high to low
14:29:56 Join hands0me [0] (
14:31:36 Join seanyseansean [0] (
14:32:02Paul_The_Nerdsafetydan: That is so incredibly strange.
14:32:14JdGordondamn the mario music is annoying :p
14:33:24JdGordonwho is developing rockboy atm?
14:33:33Papricasomeone know why screendump from the simulator look like this?
14:33:36 Join tvelocity [0] (
14:33:56seanyseanseanhello - if anyones bothered then i'm also getting the data abort error with my 5g ipod, with all the bleeding edge builds from the last few days
14:34:28 Quit perpleXa ("( :: NoNameScript 4.01 :: )")
14:34:51Paul_The_Nerdseanyseansean: Have you tried resetting the settings in a working version before installing a new one?
14:35:33seanyseanseani can't get a working version - lost them! i did delete the rockbox directory and exe before copying the new one over
14:35:39 Quit DJ_Dooms_Day ("Trillian (")
14:36:05hands0mehello :-) I'm also getting 'Data abort at 0003FF58' error with latest build on my nano when trying to play some music
14:36:19hands0meno resetting
14:36:26hands0mejust fresh install
14:36:28seanyseanseana bleeding build from yesterday only crashed once playing a tune, todays crashes at startup just after the rockbox logo
14:36:49Bg3rwtf cvs add on a directory writes directly to the repository
14:37:45Paul_The_Nerdlinuxstb: Do you know what key to press on boot to reset settings on an ipod?
14:37:56ashridahBg3r: ah, the joy of cvs
14:38:30ScoTTiecvsdo isnt it ?
14:38:32LinusNBg3r: "cvsdo"
14:39:04Bg3rLinusN if only i know where to find a ebuild for it ...
14:39:15LinusNa what?
14:39:16preglowanyone know of any way to detect a carry condition in c? :>
14:39:46 Quit tvelocity (Remote closed the connection)
14:40:14Paul_The_Nerdhands0me, seanyseansean: Just erasing the .rockbox directory and rockbox.ipod doesn't remove saved settings. You'll need to go to and get an Older build that works, go into Manage Settings, and choose to reset the settings, then copy over a new build and see if it works. I know there's a button combo you can use while booting instead of doing this, but I'm not sure what it is for iPod.
14:40:32seanyseanseanok, will try that now
14:40:36hands0meok thanx
14:40:51Bg3rLinusN a gentoo "package" more likely a rule where to get the source from, what it depends on.. where to "install" the result etc
14:41:09LinusNBg3r: so there is no ebuild pkg for cvsutils?
14:41:14Papricaaffff why the simulator screendump is garbage =\\\\\\
14:41:43JdGordoncan any1 with some cvs access make 1 minor change to plugins/rockboy/lcd.c line 1022 and make it snprintf(frameout,sizeof(frameout),"FPS: %d \t %d \t Batt:%d%",options.fps, options.frameskip,rb->battery_level()); so it shows the battery level int he status bar.. ?
14:42:17JdGordonmight need another % after the Batt:%d%
14:43:30 Join Rob2222 [0] (
14:45:19Bg3rLinusN afaics there's no such thing ...
14:45:55lostlogicBg3r: LinusN: where would I get the source, and what's involved in installing cvsutils? (I have gentoo CVS still)
14:46:51Nico_Pmy sim crashes...
14:46:55Bg3rlostlogic i don't know ...
14:47:01Nico_Pbuilt from cvs
14:47:06Bg3rmarkun knows for sure :)
14:47:06Nico_Pwith make clean
14:48:42Nico_Phow can i get a previous CVS version of the files that cause the prob ?
14:49:02Bg3rlostlogic: maybe this?
14:49:05JdGordongnite all
14:49:13lostlogicI found the homepage actually:
14:49:23JdGordonany chance of my patch being commited while im asleep?? :p
14:49:30 Quit JdGordon (Read error: 104 (Connection reset by peer))
14:50:07seanyseanseanPaul_The_Nerd: I reverted back to the 19/02 build (worked fine), cleared the settings then installed the latest bleeding build. It gets further, ie it doesn't crash at startup, but it just hangs when selecting a tune. Sometimes it gets as far as displaying the now playing screen, sometimes not. Does that help?
14:51:10Paul_The_Nerdseanyseansean: Okay, couple more questions. Are you using the rockbox_default Theme or a custom one? Also, what format is the music in?
14:53:55seanyseanseanI was using a third party one (can't remember where from) but that's been switched to the default theme since I started getting problems. The music is all in musepack format, I can try an mp3 if you like?
14:53:58lostlogicman, I haven't made an ebuild in years.
14:54:05Bg3rany objections to commit the chessbox plugin (it still has undefined references, caused by missing dependencies in the Makefile), but without adding it to the plugin's Makefile?
14:54:13Bg3rlostlogic hehehe :)
14:54:30safetydanPaprica, that screenshot is definitely broken. Was that the SDL sim?
14:54:35amiconnPaprica: What sim is that (type, simulated target)?
14:54:48Paul_The_Nerdseanyseansean: It'd be helpful. It's odd, because hands0me said he was using a Nano, and my nano seems to be having no problems with my music.
14:55:07seanyseanseanok give me a minute i'll find an mp3
14:55:10PapricaSDL sim, target nano/h300
14:55:22 Join tvelocity [0] (
14:55:35amiconnPaprica: Could you put the .bmp somewhere?
14:55:49Bg3ramiconn ? LinusN ?
14:56:00Papricawhere do you want?
14:58:17seanyseanseanpaul: ha, i've rebooted and now i'm getting the data abort at bootup again! I'll put the buld from the 19th on again, reset settings, run the latest bleeding with an mp3 and get back to you :D
14:58:29amiconnScreenshot is working fine here btw (SDL, h300 target)
14:58:44Nico_Pamiconn: and sound ?
14:58:51amiconnNo idea
14:58:51Nico_Pmy sim crashes when playing a file
14:59:45Nico_Pyes the recent commit caused the prob
15:00:17Nico_Pi downloaded versions prior to that change for sound.c and uisdl.c and no it works
15:00:24Nico_Pno -> now
15:01:56Paul_The_Nerdseanyseansean: Okay. Once you get into a bleeding edge build successfully, before trying to play music, reset setting in it again, then turn it off and back on for me, if you could.
15:02:30seanyseanseani will, just installing the bleeding build now
15:02:55Paul_The_NerdAnd you said you had a 5g?
15:04:42seanyseanseani do, a 30gb. Hey i've just reinstalled the bleeding build after the 19/02 and clearing settings and guess what? Data abort at bootup again!!
15:05:00amiconnNico_P: I checked it with gdb: sdl_audio_callback calls memcpy() with an insanely high length
15:05:06 Quit hands0me (Read error: 110 (Connection timed out))
15:05:06seanyseanseanData abort at 0003FFF4C
15:05:18amiconnIt looks like this is a negative number...
15:05:29seanyseansean3FF4C sorry, too many F's theree...
15:06:14Paul_The_Nerdlinuxstb: You have a 5g right. Is it working properly with recent builds?
15:06:49Bg3rPaul_The_Nerd i think linuxstb isn't here atm
15:07:08Paul_The_NerdBg3r: I can never remember how long ago I saw someone speak. =/
15:07:39 Join hands0me [0] (
15:07:49seanyseanseanHave you got a debug build I could use? I might look at setting up the toolchain at home myself if I get time later this week anyways.
15:08:51lostlogicBg3r: LinusN: alright, cvsutils gentoo package created in Gentoo unstable, it'll be available within 30m on the rsync servers.
15:09:18Bg3rlostlogic 10q very much :)
15:09:50Paul_The_NerdAnyone seen an error like this before: make[2]: *** No rule to make target `<<<<<<<', needed by `/home/DarkkOne/rockbox-devel/5g/apps/plugins/dep-plugins'. Stop.
15:09:57Bg3rPaul_The_Nerd yep
15:10:05Bg3rin the SOURCES probably
15:10:31Paul_The_NerdMust've missed it
15:10:51Paul_The_NerdYup. Bloody bitmaps ones never seems to merge right.
15:11:23 Quit Mongey ()
15:11:50Bg3rso, no objections to add the chessbox to the cvs (without being built atm)
15:12:40amiconnsafetydan: The error happens in line 157 of uisimulator/sdl/sound.c
15:12:55amiconnI can paste the complete backtrace from gdb on request
15:13:22 Join LGMS [0] (
15:13:28 Join DreamTactix291 [0] (
15:13:59LGMSis there someone which have compile a simulator for H3XX player today?
15:14:17 Join webguest03 [0] (
15:15:59LGMSIs Dan here?
15:16:04***Saving seen data "./dancer.seen"
15:16:20LGMSBecause I have always trouble with sound in the H3XX simulator with SDL
15:19:35 Quit LGMS ("CGI:IRC")
15:19:42Bg3rmarevalo_away are you here ?
15:22:10 Quit ashridah ("Leaving")
15:22:10 Quit midkay (Read error: 104 (Connection reset by peer))
15:22:20 Join midkay [0] (
15:22:59peturBg3r: read you own question :P
15:23:05 Part Paul_The_Nerd
15:24:40Bg3rpetur :)
15:25:31 Quit DT291 (Read error: 110 (Connection timed out))
15:26:09*Bg3r just commited a non-compiling plugin ...
15:26:11 Join perldiver [0] (
15:27:05Bg3rand Miguel's last name definitely needs correction :(
15:27:12 Nick marevalo_away is now known as marevalo (
15:27:42marevaloBg3r: i'm back
15:27:46Bg3ryep, i see
15:28:02marevalowas out eating
15:28:12Bg3rokay, i'm with CP1251 here...
15:28:15marevaloany question ?
15:28:35marevaloare we talking about charsets?
15:28:38Bg3rand i'm pretty sure your last name doesn't look as it shouldbe
15:28:57marevalothe e with acute accent
15:30:17Bg3r <= it's okay now, yep ?
15:32:13marevalothere is another spanish name wrong on this file
15:32:45marevaloJosé Maria Garcia-Valdecasas Bernal should be Jos Mara Garca-Valdecasas Bernal :-D
15:35:18Bg3rmarevalo doesn't it look okay with UTF-8 encoding ?
15:36:03marevalowere, here not
15:36:42marevalodon't really know what is the default charset for source files on rockbox
15:37:22marevaloI supose viewcvs does the recoding itself, doesn't it?
15:37:33 Join midkay_ [0] (
15:38:43Bg3rmarevalo sources should be in 8859-1
15:38:53Bg3rbut i think credits are in utf-8
15:39:35marevalook, so viewcvs is showing this file "wrongly"
15:39:58Bg3rk, and *now* your name looks wrong ...
15:39:59marevalobtw, it seems that anjuta is naughty about tabs and spaces
15:40:21 Join Dick0 [0] (
15:40:21marevaloheh, so that seems to be the case
15:41:30Dick0Hi there, any developers around?
15:41:57marevaloBg3r: any hint for the next round of patches ? (for the commiter to have it easier)
15:42:01LinusNPaprica: check out the latest cvs for a screendump fix
15:42:29Dick0I think I've found a problem with saving dynamic playlists.
15:42:51LinusNi just *hate* text-mode file handling
15:43:06LinusNwindows. burn in hell!
15:43:21Bg3rmarevalo the problem is with the bitmap dependencies
15:43:33Dick0When I go to the save playlist screen, it defaults to /dynamic.m3u. If I delete this using the record button it will then only allow me to enter 3 characters.
15:43:34 Quit muesli__ (Read error: 110 (Connection timed out))
15:43:53 Quit seanyseansean ("CGI:IRC")
15:44:40LinusNDick0: that bug should be fixed in the daily builds
15:44:57Dick0OK I'm using CVS 060212
15:45:00marevaloBg3r: so what is the problem ?
15:45:18Bg3rthe problem is that i'm not a Makefile expert *at all*
15:45:27Dick0I'll try a later version tonight
15:45:46LinusNDick0: i fixed it on Feb 16
15:45:54Dick0OK Thanks.
15:46:02Bg3rlinuxstb told me that he will look at it later...
15:46:10 Quit Dick0 ("CGI:IRC")
15:46:21marevaloBg3r: well, you now more than me ;-)
15:46:48Bg3rhaha i just took the Makefile from the databox plugin :)
15:46:51 Join Mongey [0] (n=mongeyc@
15:47:23Bg3rthe problem here is that this is the first "subdirectory" plugin which uses bitmaps ...
15:47:43marevalothat why I included gnuchess.c in such a crude way
15:48:03 Join [IDC]Dragon [0] (
15:48:25*[IDC]Dragon noticed chess activities
15:48:46safetydanamiconn, thanks, but it should be fine. The description of a call to memcpy with big negative number should be sufficient. I'll just need to boot in to Windows to test.
15:48:48preglow[IDC]Dragon: you know arm well, yes?
15:48:52[IDC]DragonI went with the idea for a chess plugin, too
15:48:55marevalohoho, I'm thinking about a new problem, I want to use the Archos pieces for iRiver's remotes, and don't really know how to do it without including two times that .bmp
15:49:03[IDC]Dragonpreglow: well
15:49:06preglow[IDC]Dragon: gcc always seems to try to group ldr instructions together, is that beneficial or something?
15:49:14safetydanHas anyone tried the sound in the SDL sim under Linux? Or am I the only whom it works for?
15:49:28preglowsafetydan: i can give it a go
15:49:35[IDC]Dragonperhaps to resolve data dependency
15:49:40amiconnsafetydan: I can try tonight in debian on vmware
15:49:48[IDC]Dragonavoiding a delay slot
15:50:04preglowthe imdct_l arm opt in libmad also seems to do so
15:50:05safetydanamiconn, preglow: thanks
15:50:11preglowso i'm just wondering if it's something i should do too...
15:50:13marevalo[IDC]Dragon: yes, gnuchess2 for rockbox
15:51:06[IDC]Dragonmarevalo: we could have a huge opening lib on disk
15:51:19[IDC]Dragonbut have little CPU and memory
15:51:25marevalo[IDC]Dragon: but not in memory :-P
15:51:45preglowany opening libs available for free?
15:51:47[IDC]Dragonstrange hardware, yes
15:52:11marevalo[IDC]Dragon: don't worry, in a couple of months you will be able to fill your HD with gigs of pgn games :-P
15:52:38marevalopreglow: GNUChess comes with a very basic one, it's still not integrated
15:53:17 Quit midkay (Read error: 110 (Connection timed out))
15:53:19marevalo[IDC]Dragon: pgn o png ?, don't remember the format, my days of computer chess ended years ago before this plugin
15:53:40[IDC]Dragonthe small set of pieces looks very strange
15:53:54[IDC]Dragonperhaps we should use the grayscale lib
15:54:03[IDC]Dragon(and make it slower)
15:54:08marevalo[IDC]Dragon: 8 pixels , the better I could do
15:54:24marevaloin the end I had to go for non-realistic pieces
15:54:44[IDC]Dragongrascale would allow antialiased gfx
15:54:57[IDC]Dragonscaled down larger bitmaps
15:55:16amiconngrayscale will eat CPU...
15:55:20marevalowell, I tried to stay portable with the smaller machines
15:55:23preglowsafetydan: seems to work fine
15:55:36amiconnWe could switch off display during calculation, but that will probably feel odd
15:55:38[IDC]Dragonthat's what I meant with slower
15:55:50*Bg3r just exoected this reply about the CPU usage from amiconn ;)
15:55:56marevaloone day my H120 will be one of this "obsolete" machines and I would really like people to not forget me at that moment
15:56:18amiconnmarevalo: Did you do 8x8 pieces for archos?
15:56:30[IDC]Dragonmy first 8 bit computer had Cyrus Chess
15:56:35amiconn10x8 pixels would look better because of the pixel aspect
15:56:42marevaloamiconn: yep, I plan also to use them on the remote
15:56:44[IDC]Dragonon a 16kB cartridge
15:56:49safetydanpreglow, thanks. hopefully it's something obvious under Windows then. I think it's probably the SDL audio buffer size being too small.
15:56:50preglowsafetydan: peak meter doesn't work
15:57:01[IDC]Dragonit beat me all the time
15:57:06marevalo[IDC]Dragon: mine too, a CBM VIC-20
15:57:11amiconnmarevalo: The iriver remote does have square pixels
15:57:25[IDC]Dragonwhich is no wonder, I suck at chess
15:57:28 Join ModernExecutive [0] (
15:57:37*amiconn remembers Psion chess and voice chess on ZX Spectrum
15:57:39safetydanpreglow, nothing to do with the sim since it just uses the existing pcm buffer code...
15:57:56marevaloamiconn: yep, I noticed that (the 10x8 option) looking at bejeweled, thats why I don't asume tiles to be square on the code
15:58:02[IDC]Dragonvoice chess, now that would match our audience
15:58:35*[IDC]Dragon remembers battlechess
15:58:36marevaloamiconn: but making the 8x8 pixels left me exhausted, gimme some time and will also make the 10x8
15:58:37amiconnThe ZX Spectrum only has a 1-bit (!) speaker output, directly driven by a Z80 I/O pin
15:58:53safetydanpreglow, actually the peak meter looks like it's #ifdef'd out for the sim\
15:59:03amiconnIt's a miracle that the voice was somewhat understandable...
15:59:06marevaloI like doing graphics pixel by pixel, but sometimes it's hard
15:59:44 Part ModernExecutive
15:59:48marevalo[IDC]Dragon: believe me, I have also made plans for a voice enabled chess, it's on the queue
16:00:02amiconnNow that can be sth for Rotator - make bejeweled work simultaneously on main & remote for iriver
16:00:10marevalocan plugins be voice enabled ?
16:00:14amiconnNot yet
16:00:29[IDC]Dragonyou can bring your own voice clips
16:00:35preglowsafetydan: it seems keeping a button pushed for a long time when cycling settings and such crashes the sim
16:00:39[IDC]Dragonand play them from a plugin
16:00:43amiconnThere are ideas for supporting localisation and voice in plugins with l10n v2
16:00:50marevalook, well, I promise chessbox will work on the iRiver remote, that's my point for starting the port
16:00:55[IDC]Dragonbut I don't know if that's possible for non-Archos
16:01:35preglowsafetydan: luckily, i can't reproduce it with gdb
16:01:52marevaloI did the 8 pixels for that, didn't think gnuchess would fit on the archoses
16:02:16 Join XavierGr [0] (
16:02:26amiconn[IDC]Dragon: Back then we did a chess match ZX Spetrum, Psion Chess vs. Atari 800XL with ?? Chess
16:02:47amiconnThe spectrum needed up to 25 hours per move at the highest level
16:02:48[IDC]Dragonusing human net (tm) ?
16:03:08amiconnThe spectrum lost (after 40+ moves)
16:03:35[IDC]Dragonwalkover net
16:03:42marevaloheh, computer chess is quite interesting, I lost interest in chess but still find computer chess interesting
16:03:52amiconnActually meet-at-scholl-and-tell-move net
16:04:04Lynx_now we need a go plugin :)
16:04:26marevaloyes, that's really interesting (now I play more go than chess)
16:04:41[IDC]Dragonwith a match lasting a month
16:04:45marevalostrategy vs. tactics
16:05:25Lynx_go is frustrating for a beginner
16:06:34safetydanpreglow, that's handy when gdb does that. more things to look at when I get home
16:07:22marevalook, amiconn Bg3r , when should I come back and expect chessbox building for adding more features ?
16:07:31preglowLinusN: btw, isn't there a way to avoid the use of functions when doing the i2c bitbanging? i can think of one way involving macros, but it wouldn't work when you have more than one generic bitbanged driver in use at once...
16:07:50Bg3rmarevalo i think tomorrow ?:)
16:08:01preglows/functions/function pointers/
16:08:12marevaloheh, well, I still have real life work piled on my table
16:08:37marevalobut definitely will be back this week, regards and thanks for the help !!
16:08:50LinusNpreglow: the whole point is to have more than one
16:08:57amiconnpreglow: My idea was defining inline functions in .h files
16:09:10amiconn(more readable than macros)
16:09:14 Quit tvelocity ("Αποχώρησε")
16:09:30preglowamiconn: i guess that would work
16:09:54 Quit marevalo ("Abandonando")
16:09:55preglowbut i'll suffer from the same problem of not having more than one i2c driver at once, yes?
16:10:19LinusNone point of the generic i2c driver is to have one function for all i2c chips, even if they are on different busses
16:11:13preglowstill, it's nice
16:12:58LinusNif we do it with macros/inline functions, the code size will increase somewhat, and it might be trickier to debug
16:13:25LinusNbut i guess it's manageable
16:14:36preglowwhole point is that it'd be nice to be able to use these also for i2c drivers where performance is important
16:15:11LinusNpreglow: sure, but which platform would that be?
16:15:19preglowdunno, wouldn't the mas count?
16:15:38preglowi thought that was i2c and pretty heavy on the bus to boot
16:15:47LinusNye sit is
16:16:04LinusNbut my goal wasn't to replace those
16:17:06 Part IcyStorM
16:17:20preglowwell, ok
16:17:25preglowthen i guess it doesn't matter too much
16:18:14LinusNthe mas i2c routines are tight and well-used/debugged
16:18:29LinusNso i wouldn't want to touch them if it isn't necessary
16:19:12LinusNbut if we can find a good way to make the generic ones work with inline functions, i'm all ears
16:19:24preglowi'll just let amiconn explain his plan :>
16:19:33preglowi don't think the code size would increase much, btw
16:19:37preglowat least not on coldfire
16:19:47LinusNprobably not
16:19:55preglowthe call overhead on that platform alone is enough to fill each of the functions
16:20:10preglowfor arm, however, it'll probably be a bit larer
16:21:54peturI think he meant something along the lines of this: C file has defines for the IO operations, H file contains the actual I2C driver code. This is in fact including code in a C file, but in this case (performance req) I think it is acceptable
16:22:19preglowwhat, now even including code in a _c_ file is bad?
16:22:47preglowor do you mean including a c file...
16:23:09LinusNthat could work
16:23:13petura H file with inlined functions...
16:23:27peturthat use those defines
16:24:51Mongeyanyone know how to fix a broken windows bootloader?
16:25:11peturfdisk /mbr
16:25:32Mongeybut where do i put fdisk /mbr ?
16:25:33 Join seanyseansean [0] (
16:26:10peturboot from install CD, then go to recovery console?
16:26:28Bg3ri think it's even
16:26:35Bg3ror similar
16:26:54peturcould be, haven't done it in quite some time...
16:26:58LinusNbtw, both the h300 and the x5 pcf50606 drivers work with the generic i2c driver :-)
16:27:29Mongeyoh..... any updates for the h3xx, or any plans?
16:28:13LinusNi'm working on the x5 at the moment
16:28:37Mongeydid you get the new h3xx
16:28:43LinusNbut i have a new h300 now, so i can continue working on the battery drain issue
16:29:17preglowdo anyone know what the odds are of the ipod nano functioning without its battery? i don't really know much about power supply/charging circuits
16:29:21Lynx_there is a battery drain issue?
16:29:45LinusNLynx_: yes, rockbox eats a lot more battery than the iriver firmware
16:29:59Lynx_ok, haven't noticed really
16:30:16 Quit seanyseansean (Client Quit)
16:30:30amiconnI wouldn't say a lot, but it's more
16:30:46Mongeyhmmmmm does repeating the same song eat the battery?
16:30:46amiconnIriver claims up to 16 hours with stock battery. I get around 12 hours
16:30:46Lynx_btw, would it technically be even possible to have to h300 run off usb power while connected via usb, transferring data?
16:31:30LinusNnot without violating the usb spec
16:32:35Lynx_hmm, ok. because transferring a lot of data really runs the battery down
16:33:03Mongeywill (sometime in the future) there be an alarm, fm recording and schedualed recording, on the h3xx?
16:33:45peturpossibly, yes
16:33:52LinusNMongey: yes, yes and yes - if someone codes it
16:37:35peturbtw, forget what I wrote about the generic i2c, it wouldn't link....
16:37:40Mongeydoes repeating the same song eat the battery?
16:38:02preglowof course
16:38:06preglowlike everything else
16:38:21preglowit eats just as much battery as playing it the first time did
16:39:18peturdoes it spin the disk again then? no code in place to see the file is still in mem?
16:39:28preglowthere is some code for that
16:39:31preglowbut it doesn't always work
16:40:28 Join TCK- [0] (
16:41:07LinusNpetur: it could link, if the function names were generated with a prefix, defined by a macro
16:42:22LinusNtime to go
16:42:27LinusNcu around
16:42:29 Part LinusN
16:42:42 Quit Mongey ()
16:43:05peturspoiled by namespaces I am ;)
16:49:24 Join tvelocity [0] (
16:50:38 Join Thus0 [0] (
17:00:27 Join SereR0kR [0] (
17:03:11 Quit Zagor ("Client exiting")
17:04:26 Quit TCK- (Read error: 110 (Connection timed out))
17:07:13 Join yeahx [0] (
17:10:33 Quit safetydan (Read error: 104 (Connection reset by peer))
17:11:12 Join safetydan [0] (n=dan@
17:11:24safetydanlousy crashy Linux
17:15:27 Join seanyseansean [0] (
17:15:35 Join Rondom [0] (
17:16:07***Saving seen data "./dancer.seen"
17:23:20 Quit seanyseansean ("CGI:IRC")
17:23:24 Join seanyseansean [0] (
17:25:07Nico_Pwow get_metadata() seems to be called A LOT of times
17:26:13Nico_Pit was called 17 times when i loaded the first of 3 files in a dir
17:26:47Nico_Perr.. maybe loess
17:26:55Nico_Pbut still more than 3 times :p
17:27:38 Quit tvelocity ("Αποχώρησε")
17:27:55preglowit loads for all the files it can, i think
17:28:15Nico_Pyes it's logicla, but why several times per file ?
17:29:04 Nick SereR0kR is now known as AfK|SereR0KR (
17:29:34Nico_Pwhen i load a file with navi, it's called 3 times, one for each file : all good
17:29:49 Quit hands0me (Read error: 110 (Connection timed out))
17:29:59Nico_Pbut then i press NEXT and then it goes all crazy, getting called several times per file
17:30:07Nico_Pwhereas it shouldn't be called at all
17:30:09 Quit seanyseansean ("CGI:IRC")
17:31:48 Quit AfK|SereR0KR ("XChat Aqua")
17:33:50preglowseveral times per file it should not be called, afaik
17:34:06 Quit safetydan ("Leaving")
17:38:16Nico_Pin this log i reproduced what i was saying : play the first file of a dir containing 3
17:38:36Nico_Pthis made get_metadata be called 3 times
17:38:43Nico_Pthen audio played normally
17:39:08Nico_Pand then i press NEXT, and all the remaing calls to get_metadata followed immediatly
17:39:20Nico_Pi quit the sim when audio started playing normally
17:42:42 Join sethians [0] (
17:43:57preglowperhaps lostlogic has some idea of why this happens
17:44:01preglowor slasheri
17:45:11Nico_Pwell it's certainly due to repeat mode being on (repeat all)
17:45:23Nico_Pwithout it, it's much more efficient
17:45:30Nico_Pbut still, it's weird
17:47:32 Nick Nico_P is now known as Nico_away (
17:54:55 Quit Thus0 (Remote closed the connection)
17:55:21 Quit Rondom ("I'm leaving on a jetplane, don't know when I'll be back again...")
17:57:35lostlogicafter you hit next, it fills the audio buffer completely, in this case those 3 songs fit in the buffer several times each, so it calls get metadata on each of them
17:57:39lostlogicat least that is my guess
17:57:39 Quit sethians ("CGI:IRC (EOF)")
17:59:01peturwhich reminds me of my question to preglow earlier about reloading files if they're already in memory
17:59:34lostlogicplayback engine isn't smart enough for that
18:01:15 Join bluey [0] (
18:06:01 Join seanyseansean [0] (
18:07:54 Quit seanyseansean (Client Quit)
18:09:20 Part petur
18:12:02 Join DT291 [0] (
18:16:21 Quit Rob2222 (Read error: 104 (Connection reset by peer))
18:17:09 Join Rob2222 [0] (
18:22:30 Join webguest53 [0] (
18:22:55 Quit webguest53 (Client Quit)
18:25:18 Quit DreamTactix291 (Read error: 110 (Connection timed out))
18:33:26 Part [IDC]Dragon
18:39:41 Quit bluey ("Leaving")
18:43:21 Join DreamTactix291 [0] (
18:45:08 Quit midkay_ (Read error: 104 (Connection reset by peer))
18:45:20 Join midk [0] (
18:46:58 Join damaki [0] (
18:49:43Papricai cant do "extern static struct font font_ui"
18:51:31amiconnstatic and extern are mx
18:51:56Papricablah =\
18:52:32lostlogicPaprica: ther should be no need for the static keyword there −− the real varialbe the is external will already be defined _non static_
18:54:06Papricaok, so i dont need the static? (i delete static struct font font_ui from font.c)
18:56:41 Quit DT291 (Read error: 110 (Connection timed out))
18:56:58 Join webguest35 [0] (
18:57:19 Nick webguest35 is now known as dpassen2 (
19:00:24 Quit damaki_ (Read error: 110 (Connection timed out))
19:04:05 Quit mikearthur (Read error: 104 (Connection reset by peer))
19:04:12 Join mikearthur [0] (
19:05:36 Join DrumRBoy320 [0] (
19:06:19DrumRBoy320hey, im having aproblem booting RB on my 5th gen
19:06:29DrumRBoy320im gtting this error...
19:06:30DrumRBoy320Data abort (next line) at 0003FF4C
19:07:04DrumRBoy320thst in black text with rockbox blue around it, and the rest of the screen is black, that happens just after bootloader shows "rockbox Loaded"
19:07:19preglowseveral people speaks about this
19:07:20DrumRBoy320and this is after i installed the latest "bleeding edge"
19:07:32preglowsomeone has probably added some code that is not properly access aligned
19:07:37DrumRBoy320should i revert to the daily build?
19:08:02preglowi'll give it a try, give me a sec
19:08:19DrumRBoy320ok, im going to the daily, rather than the bleeding now
19:09:14 Join |Beowulf| [0] (
19:09:21 Join Bger [0] (n=Bager@
19:10:59preglowi don't get any data abort on my nano
19:13:53Papricaarr no way to do "extern static struct font font_ui" is other way?
19:14:38amiconnYou *cant* declare a static struct extern. Remove the 'static' keyword
19:14:57DrumRBoy320...shit, even with the daily build, its no good
19:15:01Papricai ask if i can do it in other way =\
19:15:07amiconnstatic means the symbol is invisible outside the current translation unit
19:16:09***Saving seen data "./dancer.seen"
19:17:59lostlogicPaprica: if it's really an external symbol then by definition it's not static −− it's globally visible
19:18:00 Part dpassen2
19:18:07 Join dpassen2 [0] (
19:18:19*Bger just realises that his "unfinished" commit has raised discussion on the mail list ...
19:18:31 Join Lear [0] (
19:19:59LearAnyone tried the sdl simulator lately? Sound doesn't work too well on cygwin, it seems...
19:20:15preglowDrumRBoy320: bleeding downloaded from also works fine on nano
19:20:17lostlogicLear: last I heard from safetydan he'd just managed to get sound working... dunno what platform though
19:20:28preglowworks just fine here
19:20:30Learthat's what I wanted to test...
19:20:41LearInteresting, I get a segfault here...
19:20:53preglowhe said he had some suspicions, though
19:21:00DrumRBoy320preglow: my 5g 60 gig isnt working even with the daily...
19:21:17DrumRBoy320seems as if the 60 gig gets pinned with the worst bugs
19:21:26XavierGrdamn what an idiot I am.
19:21:42XavierGrIt seems that the H300 cradle has the same ground for both ports
19:21:46*preglow sees he should make an optimised memset16
19:21:51DrumRBoy320now im putting the 21st's daily on
19:21:54DrumRBoy320testing that
19:22:09amiconnLear: It crashes with an insanely high length call to memcpy() - check with gdb, I think you'll see the same
19:22:20amiconnsafetydan already knows about it
19:22:28LearNo, I saw an insane pointer - I think. :)
19:22:39DrumRBoy320if that dosent work, im deleting he rockbox directory entirely, and loading the 20th's buil
19:22:49preglowDrumRBoy320: when does it crash?
19:22:55LearAh, len was bad too. I have an idea why...
19:23:33 Join LinusN [0] (
19:23:56LinusNheh, my x5 lcd says "Hello world" :-)
19:24:13preglowwhat did you do to make it that far?
19:24:23lostlogicLinusN: cool!
19:24:29DrumRBoy320OK, the 21st build works fine
19:24:34XavierGryeah nice indeed
19:24:41LinusNwell, i used the generic i2c to port the pcf50606 driver
19:25:13LinusNand ripped the lcd settings from the x5 fw
19:25:49DrumRBoy320...has anyone made any 5g wps's yet?
19:25:58LinusNthey use a similar mechanism as the h300 for the lcd backlight
19:26:02LearBtw, anyone got working stdout from the sdl sim? I got it in a file first, so I turned off the redirect, and now I don't even get that. :/
19:26:40LinusNLear: it seems to be a little more complicated than that
19:27:01LearOn Linux too?
19:27:30preglowLinusN: well, how'd you run the code? bdm? bootloader light?
19:27:51LinusNthe x5 has its own bootloader
19:27:57preglowooh, lucky you
19:28:06LinusNif it finds a firmware file, it flashes it
19:28:22LinusNso you can't really brisk it
19:28:32preglowwell, does it have a disk mode as well?
19:28:40preglowthen all is well
19:28:54LinusNso i haven't yet needed the bdm
19:29:07lostlogicsilly iRiver Hxx0 the only players without any brick protection
19:29:16preglowonly and only
19:30:28 Join Moos [0] (
19:30:54MoosHello here! −− very nice commit day today :-)
19:31:23LinusNLear: no, this is a windows specific problem, stdio works fine in linux
19:31:32 Join petur [0] (
19:32:27MoosLinusN: Félicitations for your X5 LCD progress
19:33:23preglowanyone know how to replace tabs by spaces while still retaining formatting?
19:34:06lostlogicpreglow: ask astyle to do it for you :-P
19:34:53Learpreglow: what, you have an editor without such basic functions? :)
19:35:57 Join safetydan [0] (
19:36:04 Quit dpassen2 ("CGI:IRC")
19:36:29safetydanstdio does sort of work in SDL under Win32. Look for the files stderr.txt and stdout.txt
19:36:37lostlogic:%s/\t/howevermanyspacestheyrepresent/g is usually pretty close too ;)
19:37:38preglowLear: i have a basic brain that doesn't know how to do such things...
19:37:47preglowi use vim, and i'm willing to bet it can do it
19:37:57preglowlostlogic: that doesn't retain formatting
19:38:07Learsafetydan: I know, but it is a bit of a pain to have to do that...
19:38:20lostlogicpreglow: ah, so these aren't just leading tabs...
19:39:07LearHm... On windows, is WinMain always called first, even for apps compiled with Cygwin? Seems like it...
19:39:30preglowlostlogic: no, this is the nasty kind of tabs
19:39:46 Join lodesi [0] (n=moi@
19:41:59 Quit safetydan ("CGI:IRC (Ping timeout)")
19:42:21lostlogicpreglow: try :retab
19:42:23lostlogicthat's what gool says
19:42:35lostlogicgoogle even
19:42:43preglowlostlogic: thanks, works like a dream
19:42:57preglowi just ripped a memset out of linux
19:42:59preglowwill commit now
19:43:05lostlogicahahahah, nice work
19:43:07preglowmemcpy does some macro magic i can't get to work
19:43:20preglowi'll then hack it to be memset16 as well
19:44:20lostlogic*sigh* and yet people still doubt the powers of vim.
19:44:51 Quit YouCeyE ("Leaving")
19:45:34preglowyes, they do :/
19:45:38 Join damaki_ [0] (
19:46:40|Beowulf|hi - is make clean meant to rebuild the bitmaps?
19:47:10preglowif i take code from someone who's licensed the code under the gpl, i don't need to retain the small gpl note some people keep in the file header, no?
19:47:24 Join webguest80 [0] (
19:48:56 Quit webguest80 (Client Quit)
19:50:16 Join webguest39 [0] (
19:50:55webguest39ne1 using archos jukebox on a mac?
19:51:10webguest39i have rockbox running great
19:51:11webguest39love it
19:51:13ender`preglow: AFAIK, you have to keep the GPL note
19:51:55webguest39can't mount the thing on osx 10.3.9 on a g4 but i have it working on a powerbook g3
19:52:05webguest39is there a usb driver or something?
19:52:18preglowender`: oh?
19:52:20webguest39i can't remember how i got it working on powerbook
19:52:31preglowender`: there already is a gpl note in the top of the file i'm pasting into anyway
19:52:40preglowender`: and i don't see the reason of having two gpl notes in one file
19:52:44ender`in that case, you can skip it
19:52:57ender`but add the author's copyright to the current header
19:53:07webguest39oops sorry if i have come in on the middle of something
19:53:16ender`webguest03: don't worry
19:53:18preglowender`: i've added the copyright, yes
19:53:29ender`that should be sufficient then
19:57:03preglowi commited the wrong bloody file
20:00:31merbanancvs admin
20:00:31 Quit midk (Read error: 104 (Connection reset by peer))
20:00:41 Quit webguest39 ("CGI:IRC")
20:01:17 Join midk [0] (
20:01:25 Quit damaki (Read error: 110 (Connection timed out))
20:01:38preglowcvs admin?
20:01:50preglowthink i'll just revert it
20:02:37preglowthere is a cleaner way to back out a change instead of just recommiting the unchanged file?
20:02:38 Quit mirak (Remote closed the connection)
20:03:21merbanancvs admin
20:03:30 Quit erus` ()
20:03:49preglowcvs admin then what?
20:03:53merbananbut it should be used with care
20:04:00preglowok, then i'll just not use it
20:04:10merbanan-o I think, but don't take my word for it
20:04:31Kingstonewe will we will rockbox rockbox you ;p
20:05:34preglowwhy does cvs refuse to commit the old memcpy..........
20:06:37amiconnLear: If you compile a win32 app, WinMain() is called unless you specify it to be a win32 console app
20:07:04amiconnDon't ask me how to do this with cygwin gcc...
20:07:37Learamiconn: A but Cygwin was the issue here... :)
20:07:48preglowamiconn: -mconsole?
20:07:54 Join SereR0kR [0] (
20:08:44 Join erus` [0] (
20:08:46preglowok, how do i bloody make cvs commit an older revision of a file? it seems to know its older and refuses to commit it...
20:09:22amiconnYou need to make it a newer version with the changes reverted
20:09:23Bgerpreglow easy
20:09:25Mikachumake a copy and update and replace it and commit should work
20:09:34Bgerexactly :)
20:09:41Mikachui noticed if the current playlist is empty, i can't insert a folder shuffled, the menu only has Insert
20:09:51Mikachuis that on purpose and if so, why?
20:09:54*Bger learned this with the CREDITS's issues today
20:10:08amiconnLear: preglow is right, there is a -mconsole option ...
20:10:09 Join ModernExecutive [0] (
20:10:15 Part ModernExecutive
20:10:42preglowyes, there is
20:10:43amiconn'Create console application'
20:10:49preglowthere's also an -mwindows option
20:10:58preglowi use them all the time with mingw
20:11:07preglowand sdl is sensitive to which you choose
20:11:35ender`you can easily switch between console/gui application by flipping 1 byte in the compiled EXE
20:11:42preglowamiconn: btw, the arm memset seems to make pretty good use of the arm flags
20:11:43 Join safetydan [0] (
20:11:52safetydanyeesh cygwin is slow to compile
20:12:12preglowmakes one compare and uses several different flag combinations
20:12:15ender`cygwin is extremely slow
20:12:20safetydanAnyway, Windows SDL sim users, please try changing line 201 of uisimulator/sdl/sound.c
20:12:22amiconnSo we should try to build the sdl sim with -mconsole ?
20:12:26safetydanwhere it says 512, change it to 2048
20:12:31safetydansound plays fine for me then
20:13:09Learsafetydan: so it doesn't crash, you mean? :)
20:13:18Lear(even without the change...)
20:13:38safetydanno, no crash for me
20:13:58safetydanbut sound stutterred very badly
20:14:05Learodd, me and amiconn get strange len values in the audio playback callback...
20:14:18Learand I get segfaults too.
20:14:21safetydanyeah I saw that in the logs
20:14:26safetydanbut try the buffer size thing anyway
20:14:37safetydanit could be related
20:14:50Leardid that earlier, made no difference then. but that was before the latest fixes to always fill len bytes...
20:15:14LearI'll try that later though; currently rebuilding sdl, and that is slow going. :)
20:15:20safetydanthose fixes are likely to make all the difference I hope
20:15:30safetydan512 is way too small for a lot of sound cards
20:16:04preglowoh yes
20:16:09preglowespecially if single buffered
20:16:41Learwith a latency of ... 512 samples. :)
20:17:22*preglow realises that converting memset to memset16 might not be trivial..
20:18:34Learsafetydan: btw, don't you need to loop to be sure to fill len bytes? (I've no idea what sizes pcmbuf guarantees.)
20:19:41 Quit erus` ()
20:19:48|Beowulf|with 2048 samples I still get a seg fault at sound.c:157
20:21:30Mikachuso this is probably what most would call an ugly hack, but "#include "gnuchess.c"" in chessbox.c makes it work :)
20:21:47 Join erus` [0] (
20:22:43safetydanLear, pcmbuf makes no gurantees AFAIK...
20:22:58 Join webguest64 [0] (
20:24:03t0masyour memset change...
20:24:04safetydanI'm going to have to get a better set up for testing Windows builds... my XP install is just not set up for development
20:24:14t0masdid that build on your local workstation for ipod targets?
20:24:32amiconnI didn't change memset.c
20:24:35t0masthe main build system had a problem with it
20:24:44t0masno, memset16
20:25:29amiconnNo, this was indeed memset, but it wasn't me
20:25:30 Join muesli__ [0] (n=muesli_t@
20:25:36amiconnblame preglow ;)
20:25:37webguest64this may have been covered before but isn't the problem with the bitmaps in the chessbox because you need to name the bitmaps with the target's screen resolution not the bitmap resolution
20:25:45t0masamiconn: the only change in that commit was yours...
20:25:53t0masand the distributed buildsystem saw no errors
20:25:59t0maswhile the main system had that error
20:26:25amiconnI didn't commit anything memset related for the last hours
20:26:54t0mashave any idea why it's complaining there?
20:27:27preglowit should complain
20:27:44preglowthe whole commit was botched, i had the arm assembler it something it would choke on
20:28:20*t0mas wonder why it didn't break
20:28:27preglowto your build page
20:29:38preglowit didn't build the file it should have
20:29:39preglowthat's why
20:29:54preglowyou built memset.c when SOURCES now clearly states that memset_a.S is to be built
20:30:05t0mashow fast was your commit? :)
20:30:16preglowhahah, fast?
20:30:19 Quit lodesi (Read error: 110 (Connection timed out))
20:30:21preglowit was just another commit
20:30:47t0masmaybe it was a little to late...
20:31:44 Join lodesi [0] (
20:32:23 Join infamis [0] (
20:33:00preglowi just got a data abort
20:33:05t0maspreglow: common/memset_a.S:23:6: #if with no expression
20:33:08preglowit seems there's some funkiness going on in the replaygain code
20:33:52t0masfor FM
20:33:57t0masondio, and rec
20:34:04preglowi'm such a bloody idiot
20:34:13 Part infamis
20:34:17t0masok, then it's not my fault :)
20:35:07preglowi wonder why the hell the other builds don't pick up on it
20:35:12preglowit went straight through
20:35:47 Quit Kohlrabi ("Fast alle Menschen sind Regenwürmer")
20:38:08preglowamiconn: looks like you've broken arm builds with memset16
20:39:39preglowdata abort everytime something calls fillrect
20:40:03amiconnThen there's something wrong with framebuffer alignment...
20:40:42preglowyour code doesn't handle unaligned surfaces?
20:41:22amiconnlcd_framebuffer is of type fb_data, so it should automatically be aligned
20:41:37preglowi guess it's not too far-fetched to require alignment by 2 for memset16
20:41:46amiconnmemset16.c is derived from memset.c
20:42:36amiconnPerhaps I did overlook something in there.
20:42:44preglowbut yeah
20:43:01amiconnI could use my test_mem.rock with memset16 mod on archos; it should complain the same way as arm
20:43:02preglowwhat's your intent? can memset16 write to any pointer or does it require some alignment?
20:43:13amiconnIt requires word alignment
20:43:20preglowok, then i'll do that as well
20:43:25preglowcan i have your test plugin?
20:43:31preglowsome hard numbers would be cool on arm as well
20:43:56amiconnNote that you need to modify the plugin api in order to test memset16
20:44:08amiconnThis test is disabled in the source (#if 0)
20:44:19preglowbut yeah, requiring alignment by two will simplify the code very much
20:46:04amiconnIt has 2 functions for each mem* function, one for testing correctness and one for measuring performance
20:46:26amiconnA maximum of 6 tests can be hooked to the menu
20:46:55amiconnPerformance test for memset16() isn't implemented yet
20:49:03 Join bluey [0] (
20:52:59Learsafetydan: still crashes here, but now I actually got some sound first...
20:53:49safetydanwell we're getting there
20:54:13safetydanso much for a nice simple cross platform API taking care of everything
20:55:15LearI'm starting to wonder if playback maybe is faster than decode in the simulator; that could explain the bad buffer lenghts (possibly negative values)...
20:55:54amiconnpreglow: Your score: 210
20:56:32amiconnhigh score: 5140
20:56:52preglowlousy performance by me
20:57:56preglowdoesn't seem like my memset16 worked that much better
20:58:10PapricaSomeone try brickmania on archos targets?
20:59:53lostlogicLear: if the computer has a really big audio buffer so the pcmbuf in rockbox keeps feeding it as fast as possible, that could potentially lead to weird... depends how the sim handles the situation where callback_for_more doesn't have anything
21:01:01preglowamiconn: it actually does seem like the address i'm passed isn't aligned at all
21:01:05Learwell, runtime performance in rockbox seems a bit bad to me in general (e.g., it takes a while after start of playback before anything happens), but I might have found a problem unrelated to pcmbuf...
21:02:14lostlogicamiconn: btw, have you retested your various voice bugs since my commit yesterday? I've not had any problems with it except for the known and hard to fix "voice doesn't play if the pcmbuffer was nearly empty when the sample tried to start" problem
21:02:29 Quit webguest64 ("CGI:IRC")
21:02:41preglowamiconn: forget that, i now have it working
21:02:59amiconnlostlogic: Voice seems to be back at the state it was before your pcm buffer changes
21:03:41amiconnMeaning it's still far from perfect though...
21:03:48amiconnI got no hang today
21:03:50 Join Jue_ [0] (
21:04:13lostlogicamiconn: ok, still don't know how the changes I made yesterday could have beeen related to some of the stuff you were seeing :-\ but at least it solved some issues
21:04:32lostlogicamiconn: other issues will exist in playback.c, which we already know needs to be eaten.
21:04:58 Quit Rob2222 ()
21:05:34lostlogic(really all I changed was to insert 1/8s of space as a minimum ahead of the voice / beep insertion point to keep the DMA from reading over where we are inserting while we're doing it)
21:07:31 Quit bluey ("Leaving")
21:07:50 Join Rob2222 [0] (
21:09:11amiconnpreglow: What was the problem?
21:09:22preglowjust me acting the fool as usual
21:09:39preglowso yeah, it's got to be the c code that's wrong
21:09:44preglowmy asm code works perfectly well
21:10:11Mikachui should read here before i cvs update
21:10:17Learsafetydan: seems like pcm_data_size wraps to negative...
21:10:18lodesiPaprica: just tested brickmania on an archos recorder
21:10:29lodesiPaprica: works perfectly, thx :)
21:10:46Papricaohh good to hear
21:10:48preglowamiconn: what does 50000000*0 bytes mean, really?
21:10:53preglowamiconn: in the test_mem
21:11:13safetydanLear, well now that is interesting
21:11:14amiconnOnly the time_* functions use *0 size
21:11:28lostlogicLear: safetydan: I changed the pcmbuf to use mostly unsigned math with very precise bounds checking, this is probably a good idea for the sims too.
21:11:38amiconnIt does just that: call the function with a length of 0
21:11:40lostlogic(and by mostly I mean almost exclusively)
21:11:54 Join saratoga [0] (
21:11:58preglowamiconn: ahh, so 5000000 calls with length 1
21:12:07 Quit saratoga (Client Quit)
21:12:12preglowamiconn: and then four times?
21:12:26amiconnThis (1) checks that the function doesn't choke because of it and (2) tells you the overhead if you compare with 500000000*1
21:12:44preglowamiconn: the first pass is always a second slower than the subsequent ones here
21:13:08amiconnIt tests all possible alignments
21:13:10 Join saratoga [0] (
21:13:17preglowahh, so those are alignments, yes
21:13:24amiconnlongword, longword+1 etc
21:13:28safetydanhah, now SDL sim audio doesn't work for me at all
21:13:36amiconnOn iriver there'll be 16 passes
21:13:41preglowamiconn: then i wonder why this memset was slower for unaligned acceses...
21:13:45preglowamiconn: i mean _faster_
21:13:46amiconn(to test the line alignment stuff)
21:13:52ScoTTiekeypad kinda works :)
21:14:42ScoTTie(on ipod mini)
21:14:47saratogaDid slimx ever get the 3G ipod build to boot?
21:15:04preglowamiconn: for small sizes it's actually faster for non-aligned sources, yes, and i think it's because of the early termination it does in the non-aligned case
21:15:25amiconnEarly termination?
21:16:11 Join bluey [0] (
21:16:13***Saving seen data "./dancer.seen"
21:16:31 Join Cassandra [0] (
21:16:46preglowamiconn: yeah, just look at the source, there's a handler at the top for non-aligned addresses that skips to the very end of the function if you haven't got enough for it to run
21:17:12amiconnAh, yes. I do a similar thing in the SH1 and coldfire versions
21:17:39amiconnIt seems the iPod ram access is quite fast then
21:17:41CassandraWell, that's interesting.,1558,1928324,00.asp?kc=ETRSS02129TX1K0000532 - this man is crying out for Rockbox. (I mailed him the website link.)
21:18:32BgerCassandra: did you read my response ?
21:18:33amiconnpreglow: Is the data cache copyback or write thorugh? I guess the former...
21:18:58preglowamiconn: i don't think i know
21:19:04preglowwrite through sounds the most likely, though
21:19:18CassandraBger, response to the article? I didn't see it.
21:19:24Learsafetydan: after the fourth call to sdl_audio_callback, pcm_data_size is negative...
21:19:27BgerCassandra response in the mailing list
21:19:55Bgeri'm Hristo Kovachev
21:20:16CassandraBger, yeah. Sounds to me like what you did was perfectly sensible, since you knew there was someone going to pick it up.
21:20:24*amiconn has a speed optimised colour plasma.rock now :-)
21:20:28saratogathe guy writing that article is sort of dim
21:20:29CassandraI was more raising it as a general issue, really.
21:20:47saratogahow can you manage to find lossless audio without realizing that you can convert between lossless formats without reripping
21:21:01lostlogicsaratoga: just what I found myself wondering
21:21:10Bgermaybe, but my commit was another reason for it too, i suppose
21:21:11amiconnRenders directly in fb_data format, and directly to the framebuffer. Conversion & copy stage eliminated, double speed on H3x0
21:21:16CassandraMight be hard to convert to Apple Lossless, given how locked down it is.
21:21:38saratogaor one of the dozen apps that do it
21:21:49saratogadbpoweramp, etc
21:22:18saratogaapple was nice enough to provide a COM interface for it, since they realized no one will ever care about QT
21:22:23safetydanLear, thanks, that's useful information... now to figure out what to do with it :)
21:23:20 Join Massa [0] (n=Massa@
21:23:56Learprobably the threading simulation that causes problems; in the Win32 simulator, you could have decent playback or low cpu usage, not both...
21:24:22CassandraCOM? QT?
21:24:23MassaHi everybody!
21:25:03preglowamiconn: hmm
21:25:03Learbut the "four calls" is with a zero delay in yield(), with the standard (?) 1, there's only one call before a negative size.
21:25:24preglowamiconn: asm memset and c memset are about equally fast on arm when address is aligned
21:25:46preglowamiconn: asm memset is always faster for non-aligned addresses, not surprisingly
21:25:51*Cassandra played about with Rockbox EQ for the first time the other day. It's pretty cool. What are the other numbers (that you get to by pressing PLAY, is it?) about though? Is there a wiki page?
21:25:52 Quit Massa (Client Quit)
21:26:21preglowCassandra: theres frequency, q (number describing bandwidth) and gain
21:26:31lostlogicpreglow: whatever happened to giving a graphical representation of the EQ?
21:26:44Cassandralostlogic, it's in there.
21:26:46amiconnpreglow: C memset has some opts as well, but these are only invoked when the destination is aligned right from the start
21:26:58preglowlostlogic: got more time-consuming than i planned, so postponed until i have time for more than just small work
21:27:20 Join Massa1 [0] (n=Massa1@
21:27:23preglowamiconn: and surprisingly, c memset is faster for some sizes when aligned
21:27:35Massa1Hi everybody!
21:27:39amiconnIt seems iPod won't profit from line writes (assuming that the asm version would do that if it would help)
21:27:40preglowamiconn: most notably 16
21:27:42saratogaCassandra: Apple's codecs are generally accessed through Quicktime, however iTunes for Windows includes access via the standard Win32 COM interface too, so its very easy to use their AAC/ALAC encoders from any ripper/encoder
21:27:49preglowamiconn: the asm version does so
21:27:54Cassandralostlogic: Oh, sorry, thought you meant the interface.
21:28:14lostlogicCassandra: he is going to make us a frequency vs boost graph like you see in winamp or other equalizers to give a better idea what the sliders are doing
21:28:18preglowamiconn: i think the cache makes it a bit hard to see if line reading/writing is done
21:28:19CassandraAh, right. So you'd have to have iTunes installed, but it's doable.
21:28:19amiconnOn coldfire, asm memset is always faster than C memset because the C version doesn't burst
21:28:35amiconnpreglow: Not if the cache is write through
21:28:38saratogayeah, the only way to get their codec is from itunes or quicktime
21:28:50Learsafetydan: adding a < 0 check (and then setting pcm_data_size to 0) fixes the crashes, but it doesn't sound very good. Strange echoes and clicks and whatnot...
21:28:54amiconn(for writing)
21:29:03preglowamiconn: then it isn't write through
21:29:19preglowamiconn: afaik, if it was write-through, we wouldn't have cache coherency problems between the two cores
21:29:24amiconncopyback means even better performance, but also more coherency problems...
21:29:33safetydanLear, those echoes and clicks are what I fixed (at least for Linux) in that last round of changes
21:29:45amiconnwrite through may still cause coherency problems
21:29:51Massa1Is some dev here with knowledge about what is allowed in which thread?
21:29:58preglowamiconn: well, yeah, but not as severe case one core changes data (and writes it to mem), but the other core already cached the data before
21:30:20CassandraI really need to read up on the basics of audio processing. Bass enhancement for Rockbox is something I'd really like to see (SRS-WOW was the one thing the H120 did really well) but I really know nothing when it comes to digital sound processing.
21:30:31Learsafetydan: I guess they happen for the zero size calls; really should do a memset there (better try that in fact).
21:30:47preglowamiconn: anywho, ipod nano uses 0.12 secs for the big block write at 30mhz
21:30:58preglowwhat league is coldfire in at 45mhz?
21:31:19preglowwrong, that was the 160 bytes write...
21:31:55amiconn5 MB memset ?
21:31:55Massa1Especially if it's O.K. to make a call to settings_load_config in audio_load_track?
21:32:09preglowamiconn: 5mb memset clocks in at 0.32 sec
21:32:48amiconnHmm, slower than the 31250*160 bytes ??
21:32:53Learmassa1: should be safe aside from the fact that settings_load_config tends to _stop_ audio playback currently... :)
21:32:58preglowamiconn: yup, apparently
21:33:09amiconncache... it's indeed copyback
21:33:19preglowthe 1024 block write clocks in at 0.07
21:33:24Massa1Lear: that's bad :(
21:33:37preglowamiconn: but yeah, how does this compare to coldfire throughput?
21:33:43amiconncoldfire @45 MHz manages 5MB in 0.22s
21:33:52amiconn(asm version)
21:34:01preglowso it's a tiny bit faster
21:34:05LearMassa1: I was actually planning to look at that today, but I have no Win32 simulator with decent sound at the moment. :)
21:34:06Massa1Lear: I try to write a patch which loads configuration settings dependen of the directory where the track resists
21:34:17amiconnC version needs 0.52s (aligned) or 2.06s (unaligned)
21:34:25LearMassa1: suspected something along those lines...
21:34:50Massa1Lear: the patch is almost finished - but it crashes my device and also the simulator
21:34:54saratogaamiconn: This is just to copy 0s to a 5MB block of memory?
21:35:28LearOh well, back to get the sound in the old Win32 simulator working again, but that'll be for some other day...
21:35:46amiconnArbitrary byte values
21:36:06Massa1Lear: and it seems to crash not in my code, it crashed somewhere later in a memory function :(
21:36:06preglowwouldn't be faster if it was just 0s anyway
21:36:07saratogathats pretty slow though
21:36:28Massa1and I have absolutely no idea what's wrong :-(
21:36:29amiconnpreglow: You can try to verify data cache size with time_memset()
21:36:34saratogai didn't realize memory on the Ipod was so slow
21:36:56yeahxit gets the job done
21:37:16amiconnpreglow: It means main mem on iPod and iriver are about equally fast (you measured at 30MHz, correct?)
21:37:35CassandraHmmm. audible wants people to pay $2 / podcast. I really don't see that happening.
21:37:36amiconn..but the iPod profits from data cache
21:37:56yeahxIm surprised audible are still around
21:38:14preglowamiconn: correct
21:38:21LearMassa1: load track is executed in the audio thread, iirc, maybe loading settings from there is a bad idea...
21:38:23yeahxI got a url for 2 free books but I have to sign up for a free trial, I dont like where that is heading
21:38:25Cassandrayeahx, they make a fair bit from eBooks, I believe.
21:38:30preglowamiconn: but yeah, how do i do that using timing test?
21:38:34CassandraErm, audiobooks.
21:38:42yeahxat those prices ofcourse they di
21:38:51safetydanLear, I think I can reproduce this under Linux now... takes a while but eventually pcm_data_size goes negative
21:39:07yeahxI got a few in the past whenever they had a "any book for $2.50" deal
21:39:19amiconnpreglow: Add some more steps between 1024 and the large one, with the same total of 5MB of data, perhaps doubling the data size for each
21:39:44Massa1Lear: do you have any suggestion where I can put in a check and loading which should occur for each played file and also when loading a playlist
21:39:48amiconnThe step that doesn't fit in the cache any more will be significantly slower
21:40:10amiconn{ 2441, 2048},
21:40:21amiconn{ 1220, 4096},
21:40:28amiconn{ 610, 8192},
21:40:38amiconn{ 305, 16384},
21:42:02Massa1Or is there a documentation which tells me which functions are called in which thread?
21:42:32amiconnLoading settings during playback is still broken for swcodec
21:43:00Papricawhy is there question mark in the title?
21:43:13lostlogicbecause that's a nonexistant wikiword
21:43:23lostlogictell the wiki not to wikiword it, which I forget how to do exactly
21:43:37Bgeruse <nop> before the word
21:43:47Massa1amiconn, what exactly is broken?
21:43:58amiconnWell, it would crash
21:44:13amiconnThat's why loading a .cfg currently stops playback
21:44:17Massa1Oh, that's exactly what happens here ;)
21:44:21Massa1But why?
21:44:22amiconnIt's a quirky workaround...
21:44:39LearMassa1: Not really. Possibly about the same place where some WPS-related variables are updated, unless that's called from an interrupt or something. :)
21:44:48Nico_awayis there a function to check the existence of a file ?
21:44:53 Nick Nico_away is now known as Nico (
21:45:04amiconnI dunno, probably something with the swcodec playback engine
21:45:28 Quit Bger ("[BX] For a good time, call 1-900-4BitchX")
21:45:30amiconnSlasheri wanted to look into this...
21:45:35lostlogicamiconn: you are Not a Fan of the swcodec stuff eh?
21:45:50Massa1So there is no chance to create some kind of automatic configuration loading during playback?
21:45:54Learaudio_set_crossfade does stop audio, but it should restart it too...
21:46:15amiconnlostlogic: I admit I don't understand how this engine works...
21:46:31LearNot right now at least; that config loading needs fixing first, so that it only stops/restarts the sound when it really has to.
21:46:52Papricahow could i add BrickMania page to here
21:47:14amiconn..and I know that it's not exactly the most stable playback engine..
21:47:15lostlogicamiconn: *nod* I think I understand well enough to give a high level overview, maybe I'll write that today
21:47:38amiconn..but it's currently the only swcodec engine we have
21:48:12Niconobody knows ?
21:48:13Massa1Hmm, I want to create the possibility to have different configurations for different directories
21:48:26lostlogicNico: in rockbox? dunno... it probably has a stat() function...
21:48:31preglowtest_memset16 was certainly interesting
21:48:35preglowit zipped along until len: 8
21:48:35Nicostat() ?
21:48:40preglowthen almost grinded to a halt
21:49:12amiconnThe test_* functions stop if they find an error and wait for a button press
21:49:13lostlogicNico: you on a unixy system? man 2 stat
21:49:23Nicono i'm on windows
21:49:25preglowamiconn: but they don't display there's an error?
21:49:36Massa1e.g. to have different EQ settings for directories which contains classical music
21:49:44amiconnThat's what the beginning and end of the area are displayed for
21:49:47 Join midgey34 [0] (
21:49:53Nicowhat i need is a function which will return true if a file exists and false if not, on rockbox
21:50:10preglowamiconn: memset16 is BUGGY! :-)
21:50:13amiconnFeel free to add more detailed display; I had the archos LCD in mind when creating the plugin
21:50:15Nicomaybe by trying to open the file i can get the same result...
21:50:23Massa1or to switch of shuffling for listening books
21:50:39amiconnIn fact archos was the only suppoted platform back when I first created this plugin
21:50:41 Nick Nico is now known as Nico_away (
21:51:14amiconn2004 in the header is correct
21:51:30preglowamiconn: looks to me like your memset16 test doesn't only use alignment 2
21:51:36Massa1O.K. something else....
21:52:21Massa1Since a few days I have a problem with the playback on my H340 device
21:52:37midgey34hmm... on the sdl sim in windows the remote background changes to green...
21:53:00Massa1When the device switches from one track to another, the tracks plays but I don't hear anything
21:53:00Learwhen backlight goes off, yes.
21:53:03midgey34and stop is mapped isn't mapped as enter on the numpad
21:53:05amiconnpreglow: You men test_memset16()? I does align properly...
21:53:27midgey34wow... that last sentence made sense...
21:53:38Massa1When I then do a rewind and start it again the sound is coming...
21:53:40preglowamiconn: yah, it does
21:53:54midgey34stop is mapped as enter, but no the one on the numpad (on the right)
21:53:55amiconnWhat do you mean then?
21:57:59safetydanmidgey34, the remote background colour is supposed to be a green colour
21:58:11safetydanwhen the backlight goes off anyway
21:58:45 Join slimx [0] (
21:58:50midgey34alright, I wasn't sure
21:59:42slimxhy all, i have good news for the iPod 3g port
21:59:55slimxthe button driver does work
22:00:08slimxand rockbox seems to be useable now
22:00:37Massa1amiconn: it crashes in common/memcpy.c:96 with a SIGSEGV
22:00:46slimxi'll submit a patch
22:00:54lostlogicslimx: kickass
22:01:30Massa1*aligned_dst++ = *aligned_src++;
22:01:46 Join solexx_ [0] (
22:02:12 Quit Moos (Read error: 104 (Connection reset by peer))
22:02:13preglowamiconn: still no idea about memset16.c?
22:02:14amiconnMassa1: bt ?
22:02:23amiconnpreglow: What idea ??
22:02:24Massa1would it be possible to trace it back and find out why it crashes?
22:02:26CassandraHmm. Is anyone else having trouble with the bleeding edge builds?
22:02:30preglowamiconn: of why it crashes ipods...
22:02:58amiconnI have no idea. I can't imagine where it could be passed an unaligned address.
22:03:11saratogaslimx: awesome
22:03:15Massa1amiconn: when calling settings_load_Config inside the audio thread
22:03:30amiconnbacktrace ??
22:04:01Massa1well, not really useful
22:04:43amiconnpreglow: 'dst' is an fb_data*, so it should always be aligned
22:04:44Massa1somewhere in sdl_audio_callback.. - but it also happens on the real device :(
22:05:04 Quit SereR0kR ("XChat Aqua")
22:05:35safetydanMassa1, the audio playback in the SDL sim is flaky under Windows
22:05:42safetydanwell... not so much flaky as broken
22:05:45amiconnMassa1: I told you that loading settings during playback is broken. I also mentioned that I don't know where, or why. Someone with good knowledge of the playback engine needs to fix it
22:06:04CassandraMy nano is giving me a data abort at 0003ff00 when I start Rockbox.
22:06:10Massa1amiconnn, who would be able to do this?
22:06:20Learmassa1: if load config is done in the codec (or audio) thread, it will send a stop playback to that thread, but it is busy with other stuff, so it will continue processing for a while. This could explain it..
22:06:35Massa1safetydan, it also happens on the real device, not only in the simulator
22:07:28CassandraDoes anyone else with a Nano have a working current build?
22:07:59 Join Moos [0] (
22:08:00safetydanMassa1, the crash at common/memcpy.c is most likely specific to the sim
22:08:01Massa1Lear, you mean that the real loading of the settings happens a bit later?
22:08:45CassandraMy suspicion would be that the load settings function is reseting the playback engine in some way, which isn't surprising if you think about it.
22:08:46Massa1safetydan, maybe - when I use this code on my real device it just switches the device off :)
22:08:57amiconnand memcpy itself is most probably not the reason. Very probably it gets passed weird parameters
22:09:31Mikachuarm still broken in cvs?
22:09:40Learmassa1: no, loading is probably
22:09:55Learoops, sorry, shouldn't have said that. :)
22:10:10 Part LinusN
22:10:27*amiconn wonders what's located at 0x0003ff00 on nano
22:10:45amiconnMy tells me it's arab_lookup() ...
22:10:47Massa1Cassandra, would it make sense to define some parameters which could be set without affecting playback and therefore resetting is not necessary?
22:11:33amiconnAll parameters should be settable without restarting playback (immediately)
22:11:46Learmassa1: only a few affect it. Those need to be found and fixed.
22:12:09LearPossibly not having an immediate effect (e.g., for certain buffer sizes)...
22:12:35Massa1Lear, any suggestion for some problem candidates?
22:12:48amiconnThe limiit settings currently require a reboot to take effect anyway
22:12:50Learaudio_set_crossfade is the primary one, it seems.
22:13:07CassandraMassa1, it'd be sensible to only do the resetting if something requiring it had changed, but I'm afraid my suggestion is purely hypothetical at this point. You'd need to find out what was causing it first.
22:13:33Massa1Hmm, what happens if you load a configuration file which only contains a few settings?
22:13:44amiconnCassandra: Do you have a .map file of that very build?
22:13:47Massa1Will it reset the others or just change the included ones?
22:13:59Learmassa1: by loading the config in audio_load_track, audio thread gets stalled, waiting for codec to stop.
22:14:21Learmassa1: audio_set_crossfade is always called, regardless of what settings was in the file.
22:14:29amiconnCassandra: I'd rather want a splash() saying that some parameters only take effect after restarting playback. Would be less disturbing, imo
22:14:44Massa1Lear: why?
22:14:57Cassandraamiconn: but more complex to code.
22:15:00Learwhy what?
22:15:19Massa1Lear: why is audio_set_crossfade called always?
22:15:26CassandraMassa1, I don't know from memory. You'd need to look at the config reader functions.
22:15:43Leareasier that way. No need to worry about which setting was set, and what to call for that setting.
22:16:23LearConfig loader reads file, sets fields in global struct, then calls settings_apply().
22:16:59 Join Henrico [0] (n=henrico@
22:17:04Massa1And in settings_apply it always calls audio_set_crossfade for SWCODEC...
22:17:58Massa1would it help to build a "new_settings" structure and check what has changed in settings_apply?
22:18:03 Quit solexx (Read error: 110 (Connection timed out))
22:18:24Learno, lots of settings, lots of code, I'd wager...
22:18:35peturuse dirty flags?
22:19:25Learmaybe dirty flags for the few cases that needs it (perhaps only crossfade)...
22:19:29amiconnKISS!! =:-)
22:19:33Massa1or do store the last setting in a static variable in each setting function (e.g. audio_set_crossfade)
22:20:03Massa1amiconn, and what if KISS kills :)
22:20:11peturamiconn: an array of flags, one for each setting that has changed
22:21:02amiconnMost settings don't really need it. Waste of resources, imho
22:21:50amiconnThe functions which really need to check for changed values should handle that themselves,
22:22:13amiconni.e. not use global_settings directly, but use the settings_apply() method
22:22:29amiconn...but of course only if absolutely necessary
22:24:16Massa1amiconn: but in settings_apply all the settings according to global_settings are again set - also audio_set_crossfade...
22:24:22Cassandraamiconn, erm, sorry. It's the daily downloaded off the web. Want me to build locally so I can get the map file?
22:26:56safetydanSo why is pcmbuf_play_start returning negative pcm buffer lengths to pcm_play_data?
22:28:02Learit isn't, afaik (if you're talking sdl sim...)
22:28:15safetydanit is here, or at least that's the problem I'm chasing
22:28:46Learhappens at the code "pcm_data_size -= datalen" in sdl_audio_playback.
22:28:55Massa1Now I added a "static in last_enable" to audio_set_crossfade and only do the work when the parameter "enable" differs from "last_enable"
22:29:08lostlogicpcmbuf_play_start should call pcm_play_data, not the other way around
22:29:16 Join IcyStorM [0] (
22:29:22safetydanlostlogic, by returning I meant passing
22:29:38Massa1And it does currently _NOT_ crash my simulator : (although I don't have an idead what I break with that change)
22:29:47safetydanLear, I've rewritten the SDL playback code (again) and now this is the error I'm getting. It only happens after a while (around 20 to 30 seconds).
22:29:57Learmassa1: not good enough: crossfade length might have changed...
22:30:08IcyStorMI am experiencing problemems with the new builds / bleeding edge
22:30:27lostlogicah −− could pass corrupted values if pcmbuf_play_start is called when it shouldn't be (when there are no buffers to play)
22:30:34lostlogicI should add a check for that into the function itself
22:31:32safetydanlostlogic, I think it's these two lines in pcmbuf.c
22:31:34safetydan last_chunksize = pcmbuf_read->size;
22:31:34safetydan pcmbuf_unplayed_bytes -= last_chunksize;
22:31:40Massa1BTW, is it good behaviour to directly read from global_settings?
22:31:58safetydanif last_chunksize is larger than pcmbuf_unplayed_bytes things go wacky
22:32:11Massa1Isn't it a problem when more than one thread accesses it without a mutex?
22:32:20lostlogicsafetydan: but that's _already_ a broken condition
22:32:22safetydanerr.. ignore that last
22:32:35Learmassa1: nope, not with cooperative multitasking, in general...
22:32:37lostlogicpcmbuf_unplayed_bytes by definition contains at least last_chunksize, unless the system's broken
22:32:49amiconnMassa1: Yes, that's the idea. settings_apply() should apply settings for all modules which don't use the global settings
22:33:00 Quit preglow (Read error: 110 (Connection timed out))
22:33:14amiconnThe module itself would then check whether the new value is actually different from the old one
22:33:45IcyStorMI am experiencing problemems with the new builds / bleeding edge
22:33:49amiconnapp code is allowed to read global_settings directly. firmware code is not.
22:34:25amiconnBagder: Are the .map files of the automated builds available somewhere?
22:34:28LearSome modules only read the settings during init/setup (like replaygain, equalizer...).
22:35:00Bagderamiconn: they are, if you figure out the path to the build dirs ;-)
22:35:04IcyStorM"Data abort at 0003FECC" How can I do to get it working again? What is the problem ?
22:35:18Bagderlemme see...
22:35:19Massa1amiconn: sorry, I could not follow your explanation. You mean that settings_apply() should given a complete list of changed parameter and only call the correspoinding module functions if the settings differ?
22:35:51amiconnsettings_apply() would apply all settings, as it is now
22:36:07amiconnThe module itself would check whether the value has actually changed
22:36:22Massa1Lear: equalizer? So it will recall it's init/setup when loading a new configuration?
22:36:23 Part IcyStorM
22:36:40amiconnIt stores the value internally anyway, so it can be compared in the settings function
22:36:41lostlogicso playback should remember the previous crossfade settings (if it's important to do so) and only stop/start playback if they've changed
22:36:47slimxpatch submited :)
22:37:29Bagderamiconn: for example, you can get the .map from the bleeding ipod video build here:
22:37:41CassandraI think we want the Nano one.
22:38:26Learmassa1: yes, to recalculate filter parameters.
22:38:30Massa1amiconn: you mean for example to store all settings for crossfade as static variables in audio_set_crossfade (enable/disable, fade-settings, ...)?
22:38:58lostlogicMassa1: the values are mostly stored already, so just have to be compared
22:39:00CassandraAmicon: Looks like it's in memset16
22:39:24 Quit Lear ("Chatzilla 0.9.70 [Firefox]")
22:39:38amiconnMassa1: no need to store them again in audio_set_crossfade(). The crossfade code has them stored anyway
22:40:18 Quit Jue_ ("Verlassend")
22:41:35t0masamiconn? why do you need a .map file from a build?
22:41:37t0mas<amiconn> Bagder: Are the .map files of the automated builds available somewhere?
22:41:50Massa1lostlogic, amiconn: where are they stored? audio_set_crossfade uses a parameter (enable/disable) and the global_settings...
22:43:24Massa1... and it already checks if the pcm_bufsize has been changed - if not, it doesn't stop the playback...
22:43:24 Quit saratoga ("CGI:IRC (EOF)")
22:44:26 Quit Kiwiclx (Read error: 104 (Connection reset by peer))
22:48:18amiconnCassandra: I would prefer a more exact location. I.e. exact address, .map file, and rockbox.bin for disassembly.
22:48:23amiconnI can't test that myself
22:50:29amiconnIn fact I can't see a single place in memset16.c that could cause misalignment
22:52:12CassandraWell the map file and the binary would be in the bleeding edge build.
22:52:43CassandraThe exact address was 0003ff00
22:53:00amiconnThis was with latest bleeding edge?
22:53:02CassandraSorry, the map file is where bagder said it was.
22:53:39 Quit lodesi (Read error: 110 (Connection timed out))
22:53:42CassandraUnless it's recompiled since I installed.
22:54:20 Join lodesi [0] (
22:54:21Bagderthat URL points right into the build dir so it will change at next build
22:55:18CassandraLooks like the build is 2100.
22:55:33CassandraSo it'll be overwritten when the current build completes.
22:55:48 Quit webguest03 ("CGI:IRC (EOF)")
22:56:11Cassandra(The build I'm using is 2100 or thereabouts - bit difficult to tell since it only flashes up.)
22:57:09CassandraSomeone else seemed to have a similar problem, but they left channel before I could ask them what platform they were using.
22:57:45CassandraBagder, d'you think it'd be worth including a map file in the daily and bleeding zip files?
22:58:18Bagdersince this is the first time it has come up, I'd have to say no
22:59:43amiconnI don't get this...
22:59:48Mikachuwhat should i revert to get rb working on nano again?
23:00:00amiconniPod shouldn't crash within memset16() by all means
23:00:07CassandraMikachu, go back to the last daily build?
23:00:39Mikachu 36 files changed, 407 insertions(+), 125 deletions(-)
23:00:46amiconnIt's only used in lcd-16bit.c so far, and that only calls it with fbdata* as argument, which should always be aligned
23:01:11Cassandraamiconn, that would suggest that perhaps it isn't always aligned.
23:01:17lostlogicMassa1: It would take a fair ammount of refactoring in playback.c to deal with it gracefully, but that's refactoring that needs doing any way, because there's duplicated code all over the audio_set_crossfade function.
23:01:35Mikachui get a data abort when i boot now
23:01:42CassandraIf it helps, Rockbox is setting the backdrop before crashing, which may give some indication of where the memset16 is being called.
23:01:43 Quit yeahx ()
23:01:46amiconnCassandra: 0x0003fff0 isn't in memset16() for the current bleeding edge
23:01:51 Nick Nico_away is now known as Nico (
23:02:17lostlogicMassa1: I can think of some ways to do it with less pain, but won't be able to look at it really for some time
23:02:17Massa1lostlogic: well, my knowledge abotu the rockbox code is very limited - I'm still in the learning phase :)
23:03:19petursmall asm question inbetween: what does this do exactly: movem.l d7/a4-a5, (sp)
23:03:30Massa1So I will not touch a lot of things which I don't understand, but I'll do some more research work ;)
23:03:49CassandraMikachu, we're trying to chase down the bug now.
23:03:51Massa1... when I find spare time to do it - which is not really much :(
23:03:55lostlogicpetur: it moves d7, a4, a5 to the location of the stack pointer
23:04:05safetydannow I think I've finally go the SDL sound sorted
23:04:13Mikachui had an old firmware dir lying around (from yesterday)
23:04:26Massa1thanks to everybody - I'll have a deeper look in how the settings are handled, mabye I find something out...
23:04:27lostlogicyou can move any set of registers in a single command, and if possible a burst write will be used (if the destination is aligned)
23:04:32safetydanwhoever suggested looping until we get all the data from callback_for_more (Lear?) should be thanked
23:04:35 Join TCK- [0] (
23:04:41Massa1Bye for now!
23:04:43 Join AppleGuy [0] (
23:05:31 Quit Massa1 ("ChatZilla 0.9.61 [Mozilla rv:1.7.12/20050915]")
23:05:56AppleGuyI want to ask, if it's possible to record sound with Rockbox on an Ipod 5 G?
23:06:17Cassandraunlikely as of yet.
23:06:45CassandraIn the future, almost certainly.
23:07:05AppleGuyis it a soft or hardware problem? or is the video not able to record?
23:08:01 Quit AppleGuy (Client Quit)
23:11:43amiconnArgh, me silly...
23:11:52amiconnThere *is* a bug in memset16....
23:12:28CassandraWell, you found it, which is good.
23:12:32 Join Shadowarrior13 [0] (
23:12:52amiconnCassandra: Could you test this for me?
23:13:02safetydanAnyone with Windows want to try a "guaranteed" fix for audio in the SDL sim?
23:13:19amiconnWanna build yourself or should I prepare a test build?
23:13:29CassandraI'm happy either way.
23:14:02amiconnOkay. In memset16.c, change line 22 to read:
23:14:07amiconn#define UNALIGNED(X) ((long)X & (sizeof(long) - 1))
23:14:22amiconni.e. replace LBLOCKSIZE with sizeof(long)
23:14:53 Join Jungti1234 [0] (n=jungti12@
23:15:43amiconnI'll double-check the disassembly meanwhile
23:16:17***Saving seen data "./dancer.seen"
23:16:42amiconnThis arm asm looks quite exotic, but also quite elegant...
23:18:16 Join RotAtoR [0] (
23:20:53amiconnHah, I like this: orrr3, r1, r1, lsl #16
23:21:17amiconnval = (val << 16) | val; in one instruction...
23:21:18|Beowulf|safetydan: what's the fix?
23:22:04safetydan|Beowulf|, apply this patch, rebuild, and tell me if it works
23:24:17CassandraHmmm. It's no longer necessary to periodically 'make tools' any more, is it?
23:24:52Cassandraamiconn: elegant. ;)
23:25:24amiconnCassandra: Does that change fix the crash?
23:25:31CassandraNever mind. Cruft in my build dir.
23:25:51Cassandraamiconn, nope - I didn't build the build that crashed.
23:26:17BagderCassandra: tools are built "automatically" these days
23:26:19amiconn? :confused:
23:26:43|Beowulf|safetydan: working so far :)
23:26:48Cassandraamiconn: Don't worry. Just having build troubles that were nothing to do with your patch.
23:26:52|Beowulf|small error on make: sound.c:158: warning: signed and unsigned type in conditional expression
23:26:55*amiconn just scored 30
23:27:14amiconnplasma.c needs proper iAudio button mappings...
23:27:16 Join preglow [0] (
23:28:17safetydan|Beowulf|, thanks, I'll commit it once I fix that warning
23:28:26preglowbut yeah, i'm not the only one who has an ipod that crashes whenever fillrect is called, no?
23:28:58amiconnCassandra:"> <== test build
23:30:26Cassandraamiconn, looks like that worked.
23:30:37Cassandra(I got my build working.)
23:30:49Cassandra(make clean just wasn't clean enough, apparently. :( )
23:31:28 Quit Strath (Read error: 104 (Connection reset by peer))
23:32:14Cassandrapreglow, might be the problem that amiconn's build just fixed.
23:32:14 Quit midk (Read error: 104 (Connection reset by peer))
23:32:16 Join Strath [0] (
23:32:41 Join midk [0] (
23:34:36amiconnpreglow: The alignment check in memset16.c was wrong. I just committed the fix
23:34:36 Quit midk (Read error: 104 (Connection reset by peer))
23:34:43 Quit lodesi ("Leaving")
23:35:26 Quit Siku ()
23:35:40 Join midkay [0] (
23:37:46 Quit Henrico ("[BX] The birds kept calling his name, thought Caw")
23:39:03 Join Mongey [0] (
23:39:14 Part Mongey
23:39:36|Beowulf|safetydan: just had a crash in pcmbuf_callback
23:39:46CassandraYou know, there are some good looking plugins for colour targets.
23:40:29safetydan|Beowulf|, had it been playing for a while?
23:40:55|Beowulf|yes, over 10 minutes - just re-ran in gdb, happened much more quickly then
23:41:05|Beowulf|#0 0x00422a31 in pcmbuf_callback (start=0x569368, size=0x569378)
23:41:05|Beowulf| at pcmbuf.c:150
23:41:05|Beowulf|#1 0x00438aa0 in sdl_audio_callback (udata=0x0, stream=0x4290bd0 "",
23:41:05DBUGEnqueued KICK |Beowulf|
23:41:05|Beowulf| len=44112) at sound.c:174
23:42:34safetydanokay, well it's better than it was before anyway
23:43:56safetydanI'll commit it anyway since this is probably the right approach
23:43:57amiconnpreglow: You've committed memcpy_a.S for arm as well, but it's not used in SOURCES?
23:44:00 Join Kohlrabi [0] (
23:44:21 Join CoasterMaster [0] (
23:45:00CoasterMasterI've got an interesting bug to report
23:48:05 Quit safetydan ("Leaving")
23:49:12preglowamiconn: i'm pretty sure i reverted it
23:49:36preglowno, apparently i haven't
23:49:49CoasterMasteron the latest daily build, if I go into some folders, I get a "Data abort at 0003FF80" error. Some folders work just fine, but if I start to play a song, it throws the same error. Today's daily build, 2 GB iPod Nano
23:50:13preglowamiconn: there's a reason, i can't get memcpy_a to work on arm, it uses some strange macro magic i don't understand
23:50:13amiconnKnown, fixed. Get the next bleeding edge
23:50:32CoasterMasternifty, thanks
23:51:08 Quit bluey ("Leaving")
23:51:22CoasterMasterAhhh, I see it in the change log for the current build
23:51:31CoasterMasterwhich will be finished in ~30 mins
23:51:36preglowwhenever i cvs up now, cvs fetches an old version of memcpy_a.S for me
23:51:54Cassandrapreglow, that's kind of it.
23:52:00preglowhow the hell do i make it stop
23:52:00 Join matsl [0] (
23:53:33amiconnYOu did checkout a specific version before? Then you've set the sticky flag for memset_a.S
23:53:50preglowyes, so it seems
23:53:51preglowfixed it
23:53:53amiconnYou need to remove that, but I can't remember how atm
23:54:00preglowjust delete and checkout again
23:54:18amiconnThat's cheating ;)
23:54:42preglowi'm ok with cheating cvs
23:54:49preglowit's such a cheesy bastard anyway
23:55:05 Join PhoenixSpirit [0] (
23:55:17PhoenixSpirithi all
23:55:40PhoenixSpiritis anybody there with an iRiver H120/140 ?
23:56:31 Quit petur ("here today, gone tomorrow")
23:56:51PhoenixSpiritI found a bug in rockbox interface on H120, though others may have find it before as well
23:57:00Kohlrabitell us :)
23:57:30PhoenixSpiritwhen you're using the LCD remote and switch to a screen, what would not fit on the small LCD on the remote
23:58:14PhoenixSpirityou can't go back to the playlist/wps using the remote buttons, only by unlocking your main unit's hold switch and pressing stop on it
23:58:38PhoenixSpiritdid you notice that as well ? is it fixed now in the latest builds ? :)
23:58:41amiconnSome screens are not yet adjusted to the multi-screen gui
23:58:58PhoenixSpiritsorry, I'm rather new to this rockbox stuff, but already love it and am curious about it in every way :)

Previous day | Next day