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

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

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

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

#rockbox log for 2005-02-15

00:00:04LinusNin fact, iriver rockbox is in a very early stage
00:00:18LinusNso we don't want that many people running it
00:00:24DeadManthere will be plenty of folks explaining how to get it on there in Mystic River no doubt when it is ready
00:00:50Digital007well i got rockbox running on my iriver before i reflashed it back to iriver f/w
00:01:13DeadManI am not going to touch it until RockBox says it's ready
00:01:17 Quit Digital007 ("CGI:IRC (EOF)")
00:01:32DMJCthe firmware is ready, if you like snake and tetris
00:01:33DeadManI want it mainly for gapless anyhow
00:01:37preglowbut wee, i've got an iir filter running
00:01:44LinusNpreglow: nice
00:01:46 Join Digital007 [0] (
00:01:59Digital007well i saw work on the sound output is coming
00:02:03Digital007being started on
00:02:07 Quit Digital007 (Client Quit)
00:02:15DMJCI just want mp3 and ogg support heh
00:02:24DeadManactually I'd take it with MP3 playback only and nothing else right now :)
00:02:29DMJCanything else is a nice addon
00:02:31preglowbut the output volume is foolishly low, i probably need to choose a filter structure more suited to dsp
00:02:35preglowehhh, fixed point
00:02:37 Join Digital007 [0] (
00:03:06DeadManhow loud can you make output before it distorts?
00:03:20preglowDeadMan: that's a rather pointless question, really
00:03:22preglowin this context
00:03:39DeadManwell I mean can you make it louder than iRiver have made it? :)
00:03:40LinusNhow long is a string? :-)
00:03:49Digital007Linus, twice as long as half of it
00:03:50hubblepreglow: there's 3 volume controls in the audio chip.. maybe iriver firmware doesn't max them =)
00:04:03preglowit's all about precision here, not loudness
00:04:17preglowhubble: oh, i'm still writing result to the disk
00:04:41Digital007im just wondering if its possible to stop the bass "thinning out" as the volume is increased
00:04:44DeadManjust give us some extra gain without the need for MP3Gain ;)
00:04:53hubblepreglow: aha. thought you ment iriver fw =)
00:04:58LinusNhubble: did you use our i2c driver to speak to the 1380?
00:05:24hubbleLinusN: no, i'm using the build in i2c controller
00:05:29DeadManDigital007 that is annoying true
00:05:33LinusNhubble: ok
00:05:53Digital007well since the iRiver is software driven, maybe that is just a software issue
00:06:10Digital007was thinking of getting a Boostaroo headphone amp
00:06:23hubbleLinusN: you don't happend to know how L3MODE from 1380 is connected to the CPU?
00:06:35DeadManI want one of those expensive inline amps that are about the size of a gum packet
00:06:47Digital007well the Boostaroo is about that size
00:06:51rasher(looks like uips (ips utility) is redistributable, although it'd be nice to be able to modify it - that's not allowed)
00:06:53DeadMannah it's way too big
00:07:05LinusNhubble: if it isn't in the schematics, no. you want me to find out?
00:07:11DeadManBoostaroo is like 3 times bigger
00:07:19DeadManI will have to show you the amp
00:07:50Digital007but its still pocket-sized
00:07:59hubbleLinusN: no problem, I can check it tomorrow..
00:08:15DeadManthis is like tiny and quality and will run off a single AA
00:08:31rasher <−− help, this man is visually violating me!
00:09:46 Quit Digital007 ("CGI:IRC (EOF)")
00:10:01 Join Digital007 [0] (
00:10:08Digital007why does this keep disconnecting?
00:10:25Bagdermy fault
00:10:32BagderI restarted the web server
00:10:36HClnot loading that site..
00:10:44Digital007lol no probs
00:10:53Digital007linus was gonna send me the bootloader
00:11:11BagderI think the attachments in the wiki work now
00:11:35Digital007cool bag
00:12:49HClthe rockbox logo on the front page is dead
00:12:56HCljust telling..
00:13:27BagderI can get it fine
00:13:43HClyea, works now
00:13:51HClah well.
00:13:55amiconnBagder: Attachments work now, thanks :)
00:14:15DeadManabout that amp
00:14:25Digital007where are these attachments amiconn?
00:14:33DeadManbut it will cost ya
00:14:37BagderI restarted the server before and then it stopped working, I had to fiddle the config to make it start working again. No idea why
00:14:59amiconnDigital007: All wiki attachments didn't download, I got a 404
00:15:06Bagdermight've been an upgrade that brought back a default config or so
00:15:20DeadMansome dudes could build one
00:15:30DeadMansome dudes here rather
00:15:43*amiconn wonders who needs an inline amp
00:16:11DeadManwell I could do with a bit more volume when out and about in noisey enviroments
00:16:19DeadManas I don't wear earplugs
00:16:30amiconnI never use full volume, it's way too loud for me
00:16:41 Join webguest88 [0] (
00:16:48preglowi used my europen firmware iriver out today, and i coudln't hear anything else
00:16:55preglowincluding buses driving past me
00:17:02preglowi don't think i need more volume
00:17:05 Quit webguest88 (Client Quit)
00:17:18preglowand i use a proper headset as well
00:18:01DeadManI'm just a loudness fiend and probably damamged my ears in earlier life heh
00:18:08preglowyes, so have i
00:18:13Bagderamiconn: Rockbox on Archos is way louder then iRiver stock firmware
00:18:18DeadManI like it loud and bassey
00:18:21Bagderthan even
00:19:19amiconnYeah, maybe. However, I hardly ever go above 75% in rockbox when using earphones. Usually settled at 68..70%. That holds for Recorder & Ondio, for player it's even less
00:20:12preglowthe iriver is not loud enough for medium loudness music, though
00:20:43DeadManPump up the Volume Pump up the Volume :)
00:23:04preglowbut with rockbox i'll probably try to code a limiter, so music can be reduced to square waves if you like
00:24:06preglowhell, if i get it my way, rockbox will support plugable dsp modules
00:24:43preglowbut yes, it'll be a while before i can even start thinking about that
00:24:45Digital007Thanks Linus
00:24:46Bagderthat'll be fun
00:24:51[IDC]DragonLinusN: the iriver HP340 has a bdm prepared, too?
00:24:52Digital007Rockbox is working now
00:25:02LinusN[IDC]Dragon: guess so
00:25:16[IDC]Dragona 20 pin connector?
00:26:07[IDC]Dragonlooks like a match
00:26:43[IDC]Dragonif I should ever upgrade, perhaps I should get a 3x0
00:27:05 Join Odie [0] (
00:27:14[IDC]Dragon(although I don't like displays which *require* a backlight)
00:27:14LinusNdefinitely a bdm connector, judging from the silk screen text
00:27:14Digital007What are u going to do for recording? since the iRiver is also software driven for recording?
00:27:21amiconn[IDC]Dragon: To implement colour video playback? ;)
00:27:30LinusNDigital007: ?
00:27:44LinusNyou mean which format, or?
00:27:48[IDC]Dragonfor a jukebox, I'm fine with b&w
00:27:49rasherI appear to have created an ips patch
00:28:04LinusNrasher: nice
00:28:18[IDC]Dragonthe wiggler is just a PAL, right?
00:28:26Digital007as in CBR or VBR, since iRiverbox is still showing the Archos settings
00:28:29LinusN[IDC]Dragon: basically, yes
00:28:32[IDC]Dragonwith known content
00:28:49LinusNyes, you can build one
00:29:06rasherwould be nice to create a graphical interface for it though, that also checked md5 of the result
00:29:07 Quit methangas (" HydraIRC -> <- The future of IRC")
00:29:21LinusNDigital007: we don't know yet
00:29:23[IDC]Dragonyou found a connector, does it come with a cable?
00:29:28Digital007just wondered
00:29:36DeadManeInk shows potential for static or semi static displays in terms of no need for backlighting and battery usage
00:29:38LinusN[IDC]Dragon: yes
00:29:48[IDC]Dragonwhat's on the other end?
00:30:10LinusN[IDC]Dragon: it's just a ribbon strip
00:30:22LinusNi will build an adapter to the bdm connector
00:32:07[IDC]Dragonsomebody complained about button differences H1x0/H3x0
00:32:21[IDC]Dragonbut I don't know which is the bad one
00:35:53DMJC100's joystick is l33t
00:35:55amiconn[IDC]Dragon: The problematic video now plays just fine :)
00:37:07amiconnWanna know how I spotted the wrong sector count calculation?
00:37:26[IDC]Dragonthere's a story to it?
00:37:59amiconnMy sister told me - she got the 'reading past end..' panic with an .mp3. It played fine when played from Ondio 'drive' via USB...
00:38:30[IDC]Dragonsuspicious, indeed
00:38:56amiconnThen it was unlikely 2 totally different cards have the same format error. I rechecked mmc datasheet...
00:39:29*rasher tries to figure out what kind of checks ips do to the result
00:40:19Digital007well iRiverbox is good for making on-the-fly playlists for now, for playback in the iRiver f/w
00:40:42 Join Trevmar [0] (
00:40:43BagderI've never managed to play a playlist with my iRiver
00:41:12Digital007you have to make a folder called PLAYLISTS, put your playlists in that, and at the stop screen, press the A/B button to access them
00:41:24*Bagder faints
00:41:28linuxstbSo intuitive...
00:41:35Bagderthat explains my problems
00:41:35LinusNhow silly
00:41:41Digital007its true
00:41:52[IDC]Dragonhi Trevor!
00:41:56Bagderok, I'll rephrase:
00:42:03BagderI will never use playlists with my iriver firmware
00:42:14rasherFun, I never bothered with playlists.. now I know why
00:42:22Digital007its not like Rockbox or Archos f/w where you can simply select them in the browser
00:42:37rasherwow, .ips is really primitive by the looks of it
00:42:48Bagderthe person who made that choice should be punished
00:44:05amiconnThe other really silly thing about the iRiver fw is that it reads the whole drive at boot...
00:44:19amiconn(at least I was told so)
00:44:33Bagderin fact, it is a crappy firmware
00:44:37preglowyes, it does
00:44:39amiconnImagine an 80 GB disk full of music...
00:44:50rasherI'd rather not
00:44:58preglowi would have exploded with rage if i would have had to wait for the iriver firmware to rescan the disk every time i've used usb the last couple of days
00:45:11amiconnHmm, wasn't there a limit for the total number of songs on the disk as well?
00:45:23LinusNyes, 9999
00:45:37preglowLinusN: that usb support commit of yours has surely rescued my sanity many times over
00:46:33[IDC]Dragonnight folks
00:46:45HCl :P
00:46:56HClLinusN: out of sheer curiousity, is it possible to overclock the iriver
00:47:12preglowi think 140 mhz is the max pll setting
00:47:21 Part [IDC]Dragon
00:47:29 Join webguest45 [0] (
00:47:41LinusNit is probably possible, but i won't try it
00:47:58preglowmuch more fun to work within limits anyway
00:48:01LinusNthe pll can be set to a lot more that 140
00:48:25LinusNwill probably make the cpu a little warmer... :-)
00:48:29preglowthen someone is bound to try it
00:48:49 Quit webguest45 (Client Quit)
00:48:55preglowand i am bound to laugh when to join and say "omg lol my coldfire imploded"
00:49:02 Quit lolo-laptop ("Client exiting")
00:49:06 Quit QT (Read error: 110 (Connection timed out))
00:49:36XShocKhmm. I have some kind of sound out in line out when rockbox is working. all have it?
00:49:51Digital007xShock as in music?
00:49:53Digital007or white noise?
00:50:02amiconnTrevmar: My recent fix for MMC (as well as the one I just committed) don't have to do anything with the mv problems on the 0308 units.
00:50:22preglowXShocK: no sound, just light
00:50:32amiconnTo be honest, I have no clue what's going on by looking at the code only.
00:51:04XShocKnot music not white noise... it is noise but... sounds like saw-signal( don't know english translation).:)
00:51:14amiconnTrevmar: I'd really like to have such a unit here for investigation...
00:53:08 Join elinenbe [0] (
00:54:39XShocKstarting iriver f/w. everything is quite. restart to rockbox. everything is quite.... hmm, cannot make this effect again..
00:55:00quel|outtime to go sleep, cu
00:55:12***Saving seen data "./dancer.seen"
00:55:22elinenberockbox has really come a long way in a short time on the iriver... way to go guys.
00:55:49 Quit quel|out ("KVIrc 'Realia'")
00:55:53HClbootloader helps a lot -> more people can work on it
00:56:18 Join mrmags [0] (
00:57:11elinenbehubble: how is the iriver sound coming?
00:58:13amiconnLinusN: Hmm... what about the... oldplayer flashing?
00:58:31 Join muz [0] (
00:58:58muzhey i was wondering if rockbox would fix a hissing sound i get with my shure headphones when connected to the iriver
00:59:34HClLinusN: any ideas how you're gonna add cpu scaling support?
00:59:45preglowwith a bdm, you can step through code _on_ the unit and dump all registers, yes? :-)
00:59:51XShocKagain that sound coming...
01:01:35XShocKI would guess that that is a left sound buffer from iriver. just a wild guess, because from the last time the sound itself changed a bit
01:02:16LinusNHCl: it will be a set_cpu_frequency() function, with a few predefined frequencies
01:02:24HClok :) sounds nice
01:02:28HCl*can't wait, obviously :P*
01:02:54LinusNworking on it as we speak
01:03:12preglowi'm begining to suspect i'm just plain daft
01:03:20hubble*frustration*.. i did something that broke uda1380 initialization while cleaning the code :(
01:03:33*linuxstb knows the feeling
01:03:47LinusNah, bdm works in 140MHz too...
01:04:04LinusNi feared that it wouldn't
01:04:06hubblebeen sitting for hours trying to solve it.. where's backup when you need it? =)
01:04:15LinusNpreglow: we all suspect that :-)
01:04:25HClthats why cvs is so nice
01:04:29HClyou can roll your changes back
01:04:31*preglow recommends: subversion
01:04:50linuxstbHCl: Obviously too late now, but it's easy to set up a local CVS.
01:04:51preglowLinusN: with good cause
01:04:56HCllinuxstb: yup.
01:05:04HCli should make a local cvs for gnuboy
01:05:10 Quit Digital007 ("CGI:IRC (Ping timeout)")
01:05:15HCli already broke it once today and had to revert to an backup i happened to have
01:05:40*amiconn is frustrated by X11 programming
01:05:43 Quit cYmen ("leaving")
01:06:12amiconnX is sometimes just plain silly. Wth it can't cope with multithreading!?
01:07:27 Quit muz ("CGI:IRC")
01:09:06amiconnI already tried. Doesn't work, at least not reliable, on cygwin.
01:09:31LinusNamiconn: maybe it's cygwins fault?
01:10:12amiconnI don't know. The cygwin X claims to support this multithreading extension (returns true).
01:10:37amiconnUsing this it seems to get a little more stable, however it still crashes after some time
01:11:00LinusNhow nice
01:11:15DMJC-Lhas anyone got cygwin bins for gcc for compilling rockbox?
01:11:29LinusNDMJC: which target?
01:11:31amiconnEither with 'unexpected asynchronous event', or it simply stops responding (and eats all available cpu power)
01:11:54XShocKI got it working
01:12:10LinusNDMJC: eric lassauge has it, but they are built with the wrong binutils, so you can't build rockbox with it...
01:12:13XShocKI got gcc compiling in cygwin and working
01:12:18DMJC-Lah n/m
01:12:33DMJC-Lbesides I just remembered I got ftp/ssh working yesterday
01:12:36LinusNamiconn: wonderful
01:13:01amiconnI guess I'll have to resort to a less real-time approach....
01:13:37amiconnThe timer_tick itself is working properly with setitimer() though
01:14:20amiconnGetting my first approach to work would have allowed sharing much more code between target & sims, too bad...
01:15:38amiconnI think it can be done fully realtime, but having a dedicated X11 handling thread and running the main rockbox thread separately, passing messages by non X11 means
01:16:18amiconnHowever, that might be slightly beyond me..
01:18:28amiconnThis would mean passing all graphics operation between the threads
01:26:52 Join lostlogic [0] (
01:27:39preglowi require sleep
01:27:48 Quit preglow ("l")
01:30:16amiconnLinusN: Do you have an idea why the stack overflow check doesn't work on iRiver?
01:31:32LinusNbecause it is disabled?
01:32:13amiconnYeah, I know it's disabled in cvs. Someone tried enabling it. It crashed...
01:32:13linuxstbI may have made a silly mistake, but I tried enabling it the other day, and the iRiver just crashed immediately.
01:33:59LinusNmaybe one of the stacks wasn't 32-bit aligned?
01:36:44 Join Digital007 [0] (
01:41:56amiconnLinusN: You can take the mmc dump off the server now. Problem solved, finally.
01:42:51*rasher stares blankly at gtk..
01:51:26 Part Patr3ck
01:54:31 Join StrathAFK [0] (
02:02:15 Quit Digital007 ("CGI:IRC (Ping timeout)")
02:04:20 Quit hubble ()
02:04:28 Join webguest78 [0] (
02:05:47 Quit webguest78 (Client Quit)
02:08:09amiconnNite all
02:08:32 Part amiconn
02:10:42LinusNnite all
02:10:51 Part LinusN
02:11:07 Quit elinenbe (" HydraIRC -> <- State of the art IRC")
02:12:41 Quit Hohoman ("( :: NoNameScript 3.81 :: )")
02:13:26 Quit Strath (Read error: 110 (Connection timed out))
02:14:15 Quit Odie ()
02:16:49*HCl bored :/
02:17:10*rasher punches gtk in the stomach and gives up
02:17:22rasherI'm this close to declaring it worse than swing
02:17:32HCl :P
02:18:19HClit can't be worse than windows api
02:18:28rasherit may just be this howto guide thing
02:18:30 Join Lurski [0] (
02:18:52*HCl gets frustrated just thinking about that SendMessage api >.<
02:25:27XShocKgood night
02:27:41 Quit edx (Read error: 110 (Connection timed out))
02:32:22*linuxstb celebrates squashing the last bug in the AC3 decoder.
02:34:43 Quit Trevmar ()
02:37:44 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
02:40:27*rasher gives linuxstb a cookie
02:50:47 Quit mrmags ("Download Gaim:")
02:53:29 Join amx [0] (
02:55:15***Saving seen data "./dancer.seen"
03:16:09DeadManac3 decoder? for what?
03:17:51*rasher prods libmodplug slightly
03:18:49coobthere is never a last bug
03:21:44*rasher looks at and tries to do the same with modplug
03:21:48rashernot having much luck
03:22:03rasherand I'm not even getting to code problems yet :)
03:25:16rasherwhat the
03:25:37rasherah, here we go
03:45:02DeadManOMG you have to check this out hehe You have GOT to watch this
04:02:31 Join QT [0] (as@area51.users.madwifi)
04:08:44 Join edx [0] (
04:17:16 Quit Lurski (Read error: 60 (Operation timed out))
04:20:44 Join ashridah [0] (
04:30:26 Quit XShocK (" HydraIRC -> <- IRC with a difference")
04:36:49rasherit'd be wonderful if I knew what I was doing
04:38:07rasherfor now, I'll just leave it at "modplug needs math functions that I have no idea how to give it"
04:38:13rasherand then do my homework instead
04:38:24rasherprobably a better idea anyway
04:55:17***Saving seen data "./dancer.seen"
05:47:46 Part amx ("Client Exiting")
06:23:35zeDeadMan: heh i used to have that nick, back in the day
06:31:59 Join adiamas [0] (~chatzilla@
06:39:39zemy client froze a sec
06:39:44zedunno how it ended up sending a blank line
06:40:00zeprobably pressed space then enter
06:44:57 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
06:55:20***Saving seen data "./dancer.seen"
07:05:02 Join Lynx_ [0] (HydraIRC@
07:07:33 Join Lynx0 [0] (HydraIRC@
07:12:41 Join Lynx [0] (HydraIRC@
07:12:41***Alert Mode level 1
07:12:41DBUGEnqueued KICK Lynx_awy
07:12:41DBUGEnqueued KICK Lynx_
07:12:41***Alert Mode level 2
07:12:41DBUGEnqueued KICK Lynx0
07:12:41DBUGEnqueued KICK Lynx
07:12:41***Alert Mode level 3
07:13:34 Quit Lynx_awy (Read error: 60 (Operation timed out))
07:13:34 Nick Lynx is now known as Lynx_awy (HydraIRC@
07:13:34DBUGEnqueued KICK Lynx_awy
07:13:34***Alert Mode level 4
07:14:34 Quit Lynx_ (Read error: 60 (Operation timed out))
07:19:59 Quit mecraw ()
07:23:35***Alert Mode OFF
07:24:47 Quit adiamas ("Chatzilla 0.9.66 [Mozilla rv:1.7.5/20041107]")
07:26:13 Quit Lynx0 (Read error: 113 (No route to host))
07:48:24 Join wacky_ [0] (
07:48:35wacky_what's the ON+PLAY equivalent on the iRiver ?
07:51:09 Join rickst131 [0] (
07:52:26 Quit rickst131 (Client Quit)
07:57:51wacky_heh :) I'm looking at sone Plugin pages on the wiki.. and it says ON+PLAY
07:58:19wacky_but I just found it... the ON+PLAY is equal to Play (iRiver's ON equiv.) + SELECT (push the joystick down)
07:59:28wacky_is it normal I can't get Rockbox to read .txt files ? I'm using a default bootloader + some of the latest builds for iRiver
07:59:47ashridahthe button assignments for the plugins haven't really been properly assigned yet
08:00:39 Nick StrathAFK is now known as Strath (
08:04:12wacky_:) .. but is it normal that a directory full of .txt files appears empty to Rockbox ?
08:04:28wacky_is it strictly dependent on the viewers.config file ?
08:04:59HClits in general settinsg
08:05:05HCldisplay settings then file view
08:05:08HClor something.
08:07:39wacky_wow great :) but why isn't it set to 'supported' by default ?! :)
08:08:12 Join rickst131 [0] (
08:08:12 Quit rickst131 (Remote closed the connection)
08:08:58 Join rickst131 [0] (
08:14:11wacky_ok thanks :) that's great!
08:14:12 Quit wacky_ ("bye")
08:32:35 Part rickst131
08:52:52Bagderit should show .txt files simply by playing them
08:54:26 Join webguest37 [0] (
08:55:24***Saving seen data "./dancer.seen"
09:04:43 Join Zagor [242] (
09:04:52Bagderhey Z
09:05:22BagderZagor: the twiki attachments broke yday when I restarted apache
09:05:49BagderZagor: I suspect it was because the twiki.conf file had been tampered with due to an update since we started it the last time
09:05:55Zagorok. i noticed there was a twiki update in the bunch when i ran dist-upgrade yday
09:06:16Bagderjust wanted to mention it
09:06:36Zagoryeah. i'll look into it.
09:06:51BagderI fixed it
09:06:55Zagoraha, ok
09:09:56dwihnowicked wiki?
09:16:07 Nick Lynx_awy is now known as Lynx_ (HydraIRC@
09:16:07DBUGEnqueued KICK Lynx_
09:16:51Lynx_Bagder: Ot: woudl twiki be the best wiki choice for something like the rockbox page? Or did you just take it because it's widespread and perl?
09:17:19Bagderwe like it since we've used it in the past as well
09:17:33Bagderthere might be others that are equilly suited
09:17:45Lynx_but there is no wysiwyg editor for it, right?
09:18:13Bagderbut we are not wysiwyg-editing guys ;-)
09:20:00Lynx_Bagder: i know :) if there just weren't so many wiki engines to choose from...
09:21:56 Join ashridah_ [0] (
09:22:37 Quit ashridah (Nick collision from services.)
09:22:40 Nick ashridah_ is now known as ashridah (
09:25:53 Join quelsaruk [0] (~kvirc@
09:28:28 Join cYmen [0] (
09:30:12 Quit Lynx_ (" bye")
09:33:11 Join bobTHC [0] (
09:34:20bobTHChi floks!
09:34:36bobTHCs/floks/folks ;)
09:37:20bobTHChow are u amigo ?
09:44:41 Join courtc_ [0] (
09:44:52 Quit courtc (Read error: 110 (Connection timed out))
10:09:57quelsarukback right now :D
10:21:55Bagderred build red build
10:22:20quelsarukwho did it?
10:22:24bobTHCmayday mayday ;)
10:22:33Bagderamiconn's commit of rasher's plugin
10:22:34CtcpIgnored 6 channel CTCP requests in 2 hours and 3 minutes at the last flood
10:22:34*quelsaruk holds the gun
10:22:43Bagderme fix
10:27:38 Join Patr3ck [0] (
10:31:41 Join ashridah_ [0] (
10:32:29 Quit ashridah (Nick collision from services.)
10:32:31 Nick ashridah_ is now known as ashridah (
10:37:56linuxstbMorning all.
10:39:06Bagdergood morning
10:40:34linuxstbI'm thinking it would be useful to start putting some codec code in CVS. I'm thinking of an apps/codecs/ directory, with one subdirectory per library, e.g. "apps/codecs/libmad".
10:40:56Bagdersounds reasonable
10:41:20linuxstbI'm thinking that other people will need to look at them and start optimising the decoding.
10:41:27Bagderlinuxstb: mail me your preferred user name and password, and I'll set you up for cvs commit access
10:41:56linuxstbOK, will do.
10:42:50linuxstbI've also written some "viewer" plugins (one each for libmad, libFLAC and liba52) which decode the selected file. I'm not sure if they belong in CVS, but they are the only way to test the libraries on the target.
10:43:21Bagderwe could add them and have them ifdefed coldfire in the SOURCES
10:43:37 Join amiconn [0] (
10:43:49linuxstbAm I right in saying there is no current mechanism for creating different viewers.config for different targets?
10:44:05Bagderbut we should/could fix that
10:44:15linuxstbOr will Rockbox just ignore lines if the .rock file doesn't exist?
10:44:36Bagderit will just not play the file since the .rock won't exist
10:45:01BagderI'd say that is a tiny problem
10:46:01linuxstbI also need to increase the plugin RAM size and plugin stack size for my plugins. I think 512K will be more than enough for the RAM size, but I'm still trying to find the minimum stack size needed.
10:46:45BagderI think we should just increase that size for iRiver for now
10:47:01Bagderwe can work out the optimized size later on
10:47:43BagderGmini will want a larger size as well
10:48:20linuxstbOK, I'll email you my details and then start slowly committing things this evening (real work is calling me now). WIll you be around on IRC this evening?
10:48:38Bagderprobably, at least late euro time
10:49:27linuxstbOK, I'll try not to break any of the builds.
10:49:53Bagderlinuxstb: we should be sure to note exactly what version of each external source we commit, so that we can make proper patches back to those later on, and also know when/how to update them
10:50:29linuxstbYes, I'm aware of that. I'll also include (C) and licensing information.
10:51:09linuxstbMaybe a README.rockbox in each codec subdirectory
10:53:06linuxstbIs remote CVS access done via ssh? i.e. can I give you a key instead of a password?
10:53:51Bagderno, plain pserver
10:54:19Bagderyou can just provide a user name and password via /msg if you want
10:55:27***Saving seen data "./dancer.seen"
10:56:53linuxstbI missed your last message - I've just emailed it to you.
10:57:05amiconnlinuxstb: Iirc there is a mechanism that adapts the viewers.config depending on the target. does that job, I just checked
10:58:41linuxstbDoes it remove lines where there is no .rock file?
10:59:40Bagderright it does
10:59:51Bagderor rather it only adds lines for those that exist
11:00:13linuxstbAre we happy that the codecs belong in apps/ and not firmware/ ?
11:00:24 Join jyp [0] (
11:00:27BagderI think apps
11:00:44Bagderbut I'm not 100% sure
11:01:34Bagderlinuxstb: 'dave' is added now
11:02:33Bagderwe're 17 committers now
11:02:52linuxstbCan you tell how many are active?
11:03:15Bagderperhaps 6-7
11:03:37Bagderthere's a log... wait
11:04:30*quelsaruk hides
11:04:47Bagderuh, it wasn't that easy to understand the log ;-)
11:06:06linuxstbBagder: My CVS access seems to be working fine, thanks.
11:07:35Bagderthe last 2000 commits were done by 13 people
11:07:57Bagder6 people did more than 100 commits
11:08:21Bagderamiconn is #1 with 647
11:12:01linuxstbSo amiconn is the most committed developer?
11:12:23BagderI bet he does many small commits just to impress :-P
11:12:40linuxstbSo makes lots of mistakes that need correcting :-)
11:14:51linuxstbliba52 needs to know the endian-ness of the target CPU. I'm assuming all the targets are big-endian (at the moment), does anyone have any suggestions for testing that in the simulator builds?
11:16:12linuxstbI think I can check endian.h on Linux, and assume WIN32 is little-endian, and Mac OS X (does anyone build it there?) is big-endian.
11:16:59ashridahLinusN has some of the instructions he's written targetting the mac, so it's safe to assume someone potentially does
11:17:50linuxstbI don't think it's urgent - my first commit could just assume big-endian for the targets and little-endian for the simulators.
11:17:51dwihnoIsn't macos endian neutral?
11:17:59dwihnoOr is it the motorola cpu?
11:18:14Nibblerhow can one be endian neutral?
11:18:32linuxstbI think the PowerPC can change it's endiannesss, but I'm not sure.
11:18:49dwihnothink so
11:18:59linuxstbMac OS X definitely runs it in big-endian mode though - I've compiled things on it before.
11:19:05dwihnoNibbler: It's, like, totally balanced ;)
11:19:13Nibblerlol hehe :)
11:19:19Nibblerok, if it can change i understand :)
11:21:27amiconnlinuxstb: You can check for the LITTLE_ENDIAN (or BIG_ENDIAN) #define, but beware than cygwin doesn't define either one.
11:21:47amiconnYou'd need to add the same workaround as I did in apps/dbtree.c
11:22:06Bagderso maybe we should make a global fix for it
11:22:52linuxstbI think it would be useful. For the targets, we know the endianness. For the simulator, we can try and guess, and then default to LITTLE_ENDIAN.
11:36:21linuxstbI have also added an "inttypes.h" into firmware/include - this currently defines int8_t, int16_t, int32_t and int64_t (and the unsigned versions) as "char", "short", "long" and "long long". Can anyone see a problem with committing that as it is? Is it OK in firmware/ or should I make it local to my apps/codecs directory?
11:38:22linuxstbMaybe it would be useful for the tag database to use it as well.
11:48:41amiconnBagder: Sorry that I missed that C89 compatibility problem...
12:03:23 Quit webguest37 ("CGI:IRC (Ping timeout)")
12:30:57Bagderlinuxstb: I think it should be in firmware/include
12:30:59Bagderamiconn: no worries
12:33:26Bagdernow that commit is gonna hurt
12:33:46Bagderor perhaps not...
12:34:25Bagderjyp: include/stdlib.h needs rand() fix too
12:34:30linuxstbOK, I'll add it in firmware/include. Are you talking about my upcoming commit?
12:34:40Bagderno, the one jyp just did
12:37:06linuxstbI've just seen your "C89" fix. I'm not sure, but some of the codecs may use that type of code. Does it need to be fixed?
12:39:00Bagderfor codec work, I'm on a leave-as-much-as-possible-untouched policy
12:40:37Bagderbut I would be slightly surprised if they require c99
12:40:45Bagderseems a bit restrictive
12:47:53linuxstbI've just had a quick look - liba52 and libmad seem to require c99, but libFLAC looks clean,
12:48:48Bagderour current cross-compiler for win32 on Linux is not c99 compliant
12:48:54Bagderbut I guess we should fix that
12:50:10jypBagder: fixed... sorry for the lag
12:51:42quelsarukZagor: did you read the mail list? seems that the browser bug attacks again
12:52:15Zagoryeah, i saw that. haven't had time to test yet though.
12:54:30 Join YouCeyE [0] (foobar@youceye.user)
12:55:31***Saving seen data "./dancer.seen"
13:14:21 Join Patr3ck_ [0] (
13:16:24Zagorjyp: ?
13:16:49jypI wonder what's your opinion on how to fix the latest builds
13:18:06BagderI feared that would happen
13:18:43Zagorjyp: you need to change the rand declaration in plugin.h
13:18:57Bagderno, the problem is that he has changed that
13:19:05Bagderbut now it doesn't match the system's prototype
13:19:20jypActually both problems exist
13:20:05jypFor plugin.h ... does it mean that random is changeable by plugins ?
13:20:06Bagderthe sims shoud probably use -nostdinc
13:20:18Bagdernot changeable, but usable by
13:20:24jypah , ok
13:21:17jypEIther ignore std include, or use the standard random generator ...
13:21:25jypbut it seems weird to mix the two
13:21:35Bagderthe standard one returns int
13:22:08Zagorbut only sims use the standard one
13:22:12BagderI think ignoring the std inc is the correct fix anyway
13:22:18Bagderfor more stuff than this
13:22:33Bagderbut it may make other problems appear
13:27:25 Nick quelsaruk is now known as quel|out (~kvirc@
13:32:01 Quit Patr3ck (Read error: 110 (Connection timed out))
13:34:20 Join R3nTiL1 [0] (~zorroz@
13:35:44amiconnBagder: I wonder why the FM simulator build didn't fail...
13:35:53Bagderyes, that is weird
13:36:18Zagorit never built at all
13:37:10Bagderexactly, there's something wrong with that
13:38:09 Quit lostlogic ("Going to the moon")
13:38:49amiconnBagder: So why is that column always green then? Shouldn't it be empty instead?
13:39:09 Join Sucka [0] (
13:39:15Bagdersilly script
13:42:11Bagderit hasn't built in a long time ;-)
13:42:32Zagorwe could add more columns like that. ipod anyone? ;)
13:42:59linuxstbImagine the number of questions that would generate....
13:43:25linuxstbMaybe in about 6 weeks time...
13:43:45*linuxstb ducks
13:44:45 Quit R3nTiL1 ()
13:45:03Bagderit's been wrong since we dumped the configure update concept
13:46:09 Join R3nTiL_ [0] (~zorroz@
13:46:18BagderI think I fixed it now
13:46:25Bagdernext commit will tell for real
13:48:24jypSo... Anything I can/should do to fix the build ? Is making rand return long a good idea after all ?
13:49:04Bagderits not POSIX
13:49:10 Join Lynx0 [0] (HydraIRC@
13:49:11Bagderthen it returns int
13:49:56Bagderwhich might be a reason to keep it returning int
13:50:09Zagorlong doesn't make much sense. the only places that use it truncates it to below 16 bits anyway.
13:50:19jyphaha ;)
13:50:37Bagderok, so let's keep it int then
13:50:54jypSo, perhaps there's a 16 bits version of the twister somewhere ...
13:51:28jypAnyone know something ?
13:52:19amiconnImho the algorithm itself should stay 32 bit
13:52:30 Nick Lynx0 is now known as Lynx_ (HydraIRC@
13:54:17HCland even then, rand()<<16|rand() -> long
13:54:18jypThe author doesn't seem to provide a 16 bits version, so let's go for your proposal amiconn
13:55:27jypThen again, what about
13:55:50amiconnI would even prefer it returning 32 bits. I have a test plugin that uses all bits returned...
13:56:50jypWhat pugin ?
13:57:28BagderI suggest 'int rand() to remain POSIX on it, and/or possibly adding lrand48() if we need a long-returning rand one
13:57:29amiconnIt's a file system stress test. Writing a looong file with random data, then reading back & comparing
13:58:03jypBagder's solution seems the best
13:58:22amiconnIf we have lrand() in the plugin api, I agree.
13:58:28jypIf I could persuade cut&paste to work ...
13:58:33Bagderour decision to go with POSIX has proven a benefit in the past
13:59:34jypSo, what about "return (y & 0xFFFFFFFE) >> 1; /* by Bj?rn Stenberg */"
13:59:56 Join lImbus [0] (
14:00:07linuxstbIf someone is changing the plugin api, could they add memcmp there as well? I need it for my codec-testing plugins.
14:00:09lImbushi all
14:00:20HClhey lImbus
14:01:45Bagderlinuxstb: ok, adding
14:01:57amiconnlinuxstb: Iiuc you can do that yourself... Welcome in the commit gang, btw. :)
14:03:46linuxstbThanks - I remembered that I could do it myself after I asked the question...
14:04:58jypZagor: Is the 31-bit limitation still applicable? Is it POSIX ?
14:06:12Bagderit returns a value from 0 to RAND_MAX
14:06:27Bagderso there is no fixed bit limit
14:06:47BagderRAND_MAX is set to be at least 32767
14:06:59Bagderwhich is 15 bits... :-)
14:07:10jypAlright; I know this is nitpicking; but while I'm at it ...
14:07:31amiconn says it returns 0..(2^15)-1
14:07:33jypand since I failed with my 1st naive attempt ;)
14:09:12Bagderso we should probably add RAND_MAX in a limits.h to be really nice
14:10:04linuxstbQucik CVS question - what should the $Id: $ line look like in the Rockbox header for a new file I want to add into CVS?
14:10:24Bagderlinuxstb: you can copy one from an existing file, it'll be done right when you commit
14:11:09linuxstbSo it doesn't matter that the filename etc are for a different file?
14:11:24Bagderthe id string is generated by cvs
14:12:12linuxstbI thought that you had to specify which fields cvs generated in that string (I should read the manual).
14:12:35*linuxstb goes away to read CVS manual
14:12:52jyp... linuxstb is brave ;)
14:14:08*linuxstb avoids manual and heads for the FAQ
14:14:52jypBagder: RAND_MAX is defined in firmware/include/stdlib.h already
14:14:58jypshall I move it to limits ?
14:15:10Bagdernah, leave it there
14:15:19BagderI think that is right place
14:15:30Bagderaccording to the opengroup page
14:15:35Bagderand they should know
14:17:30 Quit R3nTiL_ ()
14:18:03*linuxstb now understands keyword substitution in CVS
14:20:12jyp /highfive
14:22:55 Join lolo-laptop [0] (~lostlogic@
14:27:22linuxstbHCl: Does rockboy still need a 700k PLUGINSIZE defined?
14:32:37 Quit cYmen (Remote closed the connection)
14:32:38 Join cYmen [0] (
14:32:49HCllinuxstb: yes
14:32:59HClnot 7mb
14:33:01HClbut 700k
14:34:08linuxstbAnyone object to a 768K PLUGINSIZE for the iRiver (at least during these early stages)?
14:34:40 Join einhirn [0] (
14:35:10linuxstbShould I make it depend on the memory size - i.e. 32MB RAM -> 768K PLUGINSIZE ?
14:37:10 Join ripnetuk [0] (
14:39:29Bagderyes, do that
14:51:09 Quit DMJC-L (Read error: 110 (Connection timed out))
14:51:09 Join webguest98 [0] (
14:55:33***Saving seen data "./dancer.seen"
15:02:32linuxstbOK, done. It should probably be changed so it's not done in 5 different places, but I don't know the build environment that well yet.
15:03:29amiconn'640 KB should be enough for everyone' ;)
15:04:07linuxstbAnyone fancy porting an IBM PC (8088) emulator to the iRiver?
15:04:19ashridahwhat for?
15:04:48linuxstbBecause it can be.
15:04:54 Quit ripnetuk ("Leaving")
15:06:15ashridahlinuxstb: reminds me, you manage to sort out that odd read bug you were getting?
15:08:38linuxstbashridah: I think it was just a stack overflow - amiconn thought of that.
15:09:07ashridahah. painful to track down.
15:10:20ashridahheh. i always enjoyed getting stack underflows while trying to implement stack-based argument passing for subroutines.
15:10:47ashridah(even tho we weren't even coming close to using all of the registers in the app, that wasn't the point, as most uni exercises usually are)
15:14:02 Part Zagor
15:14:53*rasher blushes
15:19:50ashridah seems to be missing from
15:22:42*HCl hopes 100k of extra code is enough to write a dynarec..
15:23:00ashridahwhat's a dynarec anyway?
15:23:11ashridahi really should pay more attention when this stuff is being discussed the first time around
15:23:12 Quit webguest98 (Read error: 110 (Connection timed out))
15:23:19*ashridah shakes his fist "damn you morrowind"
15:26:55lImbusis somebody here around having a minute to get my sims building again ?
15:30:08Bagderwhat's the prob?
15:30:46 Join preglow [0] (
15:30:57lImbuswhen compiling any sim (win32-sim) on cygwin on native windows it complains it's missing the limits.h
15:31:06lImbusin fact, I don't have a limits.h
15:31:39lImbusbut I did not find any such named file in the source tarball neither.
15:32:01lImbussorry, I dunno how to switch off that away-mode in kvirc
15:32:17lImbusah,ok, I'm oficcially back
15:32:38Bagdergotta go, back later
15:33:28preglowanything exciting happened?
15:35:05ashridahlImbus: uh. isn't limits.h part of the system headers?
15:35:48ashridahshould have come as part of the libc development stuff (cygwin or glibc, depending on if you're in win32 or unixalike)
15:36:26ashridahmy copy of cygwin certainly included a limits.h
15:36:44ashridahand i only did a base install plus vim.
15:36:55ashridahand gcc
15:37:13ashridahshould be /usr/include/limits.h
15:37:19ashridah(based off the cygwin root
15:40:40 Join elinenbe [0] (~elinenbe_@
15:41:10rasherUsage of floats in libmodplug does indeed seem to be quite limited
15:41:28preglowrasher: can't imagine why a mod player would need floats at all
15:41:43preglowrasher: how is mod plug on the accuracy, btw?
15:42:01rasherI don't know much else than what I've been told
15:42:14rasherand what I've found by googling
15:42:26rasherthat modplug appears to be loads better than mikmod
15:42:40rasherBut it's not something I can tell apart
15:43:00rasherI don't mave a lot of mods
15:43:29HCli hate computers >.<;
15:43:40preglowyou're not alone
15:43:42rasherI think pretty much all the uses of floats are connected to logarithmic calculations
15:43:45HClfirst the ram of my laptop malfunctions making it crash horribly over and over
15:43:52HClthen the power supply of my main computer dies
15:43:54preglowi don't have a lot of mods anymore either
15:43:58HClthen the spare power supply i put into it dies
15:44:06preglowHCl: you're living the dream
15:44:15rasherpreglow: that might change though :)
15:44:34preglowi just need time to find the good old mods i know and love, and i might be able to give some feedback
15:44:37preglowbut i have no time right now
15:44:43preglowneed to work on school stuff
15:45:10rasherdamn school
15:45:37preglowlast semester now
15:45:52rasherThat's not fair
15:45:55preglowthen i have to work, which'll leave me even less time :PPP
15:45:58rasherI only just started :|
15:50:57lImbusashridah: sorry for not answering, my damned irc-client is once more not notifying me.
15:51:16lImbuswhy did my sims build before, and do not anymore now ?
15:51:54linuxstbrasher: You've probably searched more than me, but here is also libunimod, which I think is another fork from the same code libmodplug is based on. I found libunimod in the source for "timidity".
15:52:26rasherOh, I hadn't noticed that
15:52:32linuxstblibmodplug seems to be C++, libunimod is straight C.
15:52:32rasherI'll have a look
15:52:47rasherWell that's an easy choice then :)
15:52:56linuxstbAre they both based on libmikmod, or is that something else completely?
15:53:08rashermodplug and mikmod are completely seperate
15:53:25rasherI don't think they share a line of code
15:53:36lImbusashridah: there is actually a file called limit.h in the path you mentioned. and it's looking good. but why isn't it found when needed ?
15:54:22rasherlibunimod looks dated though
15:54:27ashridahlImbus: well, it could be referenced as "limits.h"
15:54:39rasherbut.. there probably hasn't been much change in modplug recently
15:54:40ashridahin which case, it'll be looking in a local path for it, not the system's include path.
15:55:02lImbusyes, may be. but how did that change ?
15:55:21linuxstbrasher: OK, I think libunimod has taken a small amount of code from mikmod - so maybe it's a "best of both" solution. But I've only very quickly looked at them.
15:55:24 Join Taxi|3 [0] (
15:55:50rasherlinuxstb: hm, as I read it, it's a straight lift from mikmod
15:57:32linuxstbrasher: Yes, I've just read that. But the filenames are identical in libunimod and libmodpug... (apart from .c or .cpp extensions).
15:58:00linuxstb(I'm looking at libmodplug-0.7 - is that the right version?)
15:58:01ashridahlImbus: well, it depends. where does the build first reference limits.h?
15:58:16rasherExcept there are quite a lot fewer file types aren't there?
15:58:20rasherlinuxstb: that's what I have as well
15:58:27ashridahwhich file is it in, and is it #include "limits.h" or <limits.h> ?
15:58:39lImbusthats in apps/settings.c
15:59:14rasherlinuxstb: I think it's just a case of the filenames being descriptive and hence used for both projects
15:59:23lImbusashridah: and it's include "limits.h"
15:59:40linuxstbrasher: Yes, I think you';re right. So libunimod doesn't appear to be useful then.
16:00:02rasherWell, if we're not going to use mikmod, we shouldn't use unimod either
16:00:19rasherI don't know if it's a problem that modplug is c++
16:00:25preglowsomeone mail the xmplay guy and order him to opensource his engine, please
16:01:58ashridahlImbus: was the file it was in filetree.c or settings.c perhaps?
16:02:07ashridahyou already said settings.c
16:02:17*ashridah points at jyp
16:02:23ashridahhe's the one to poke here, he broke the build
16:03:00jypWhat ?
16:03:05lImbusI just modified it to be <limits.h> but that does not work either
16:03:11ashridahlImbus: it won't
16:03:19ashridahit's referring to a local set of definitions for limits
16:03:24ashridahusing the system one won't get you places.
16:03:50ashridahjyp: your commit to settings.c last saturday has broken lImbus's win32-sim build
16:03:57*jyp draws its katana
16:04:13lImbusyeah, but shouldn't there be a local limits.h then ?
16:04:28lImbusthere is no local limits.h, not even in the source-tarball.
16:05:02lImbuswhile my playersim built up to last week or so, the gmini and iriver sim never built. now all three are not building
16:05:33*rasher has succesfully built iRiversim multiple times in the recent week
16:05:38rasheron linux though
16:06:10jypAnything I can do to help? I'm not sure to understand what the problem is
16:06:49ashridahjyp: in rockbox/apps/settings.c, you've added a #include "limits.h"
16:06:57ashridahwhere is limits.h? :)
16:07:24jypstandard includes ...
16:07:47rasherBuilds For Me[tm]
16:08:21lImbusso it should read #include <limits.h> right _
16:08:35ashridahlImbus: not necessarily.
16:08:56ashridahjyp: if you're including from the system libraries, why are you using "limits.h"
16:09:00ashridahinstead of <limits.h>
16:09:39ashridahin that case, why does using <limits.h> not build for him?
16:09:46ashridahwhen he has /usr/include/limits.h
16:10:08ashridah(this leads me to assume he's building the firmware, which would ignore /usr/include/limits.h, in which case #include <limits.h> isn't applicable here)
16:10:55jypI'm gonna dig into it, hold on
16:13:53lImbusI had that problem with others sims before too, gmini and iriver
16:14:44lImbusand my keyboard has just gone wild
16:15:11lImbusbut only on kvirc, that|s why i|m going to restart
16:15:18 Quit lImbus ("KVIrc 'Realia'")
16:15:30 Join lImbus [0] (
16:17:27 Quit Shulberry (Read error: 110 (Connection timed out))
16:17:49jypI'm already gonna fix the " " to < >
16:24:33 Join ripnetuk [0] (
16:29:42 Join ep0ch [0] (HydraIRC@
16:30:35 Quit ashridah ("sleep")
16:32:13ep0chyou guys looked at DUMB for tracker playback? not sure if it fits the job or not.
16:33:22 Join methangas [0] (
16:34:02rasherep0ch: I'll make a note of that
16:34:31jyplinuxstb, you're 'dave' right ?
16:35:00linuxstbjyp: yes - I'm in the IRC nicks Wiki page.
16:35:23jypOk ;)
16:36:39rasherep0ch: looks like being a tradeoff between accuracy and file format support compared to modplug
16:36:55ep0chrasher: kode54 (HA) did a plugin of DUMB for foobar2000, so maybe get in touch with him.
16:40:23linuxstbAre there any MOD players for mobile phones or PDAs? It may be worth checking what libraries they used, or even looking at their source (if available).
16:41:02 Quit einhirn (Read error: 104 (Connection reset by peer))
16:41:15elinenbewhat is status on sound on the iriver?
16:41:19 Quit ep0ch (" HydraIRC -> <- IRC for those that like to be different")
16:41:22 Join ep0ch [0] (HydraIRC@
16:41:36elinenbewow! 46 people in this channel... not bad.
16:42:23rasherlinuxstb: I'll google around for mod players on pdas
16:42:25linuxstbelinenbe: "hubble" is working on sound, but he's not one of the 46 people at the moment.
16:42:59linuxstbrasher: I'm sure there must be one for the Palm.
16:44:35rasherlinuxstb: I just found one for pocketpc
16:44:42rasherIt's not open though
16:44:57rasherbut by the looks of the format support, I'd say he used modplug or mikmod
16:45:04linuxstbpocketpcs are a bit too powerful. An older Palm could be comparable to the iRiver.
16:45:25linuxstbAnd aren't the old Palms 68k-based as well?
16:45:30 Quit Sucka ("( :: NoNameScript 3.81 :: )")
16:48:00rasherI don't know.
16:48:19Bagderlinuxstb: yes, dragonball iirc
16:48:45rasherI can't find any palm players
16:49:29 Join Sucka [0] (
16:50:58jypQuestion ... Why is plugin.c compiled on the gmini, whereas plugins are disabled ?
16:51:29Bagderbecause the disabling is only a little makefile thing that is checked in the plugins dir
16:52:40jypWould it hurt to make part of it conditionally compiled ?
16:53:10Bagdernot really, no
16:54:24jypI'm still on fixing the warnings & looking for the best method
16:54:46Bagderbut disabling plugins seems like very short-term
16:55:20jypThere's still the "no ram for code" problem
16:55:34***Saving seen data "./dancer.seen"
16:55:42Bagderbut isn't that because your lds file is wrong?
16:57:19jypNo, here's how it is...
16:57:30jypCode & data use separate address spaces
16:57:42jypand code is mapped to flash
16:57:57jypexept for 16k of internal mem
16:58:39 Join DrRick [0] (
16:59:05HClhas anything been reported about rockbox crashing upon shutdown, and it not shutting down at all?
16:59:43rasherHCl: will rockboy run on onmodified rockbox now that PLUGINSIZE is 768k?
17:00:34HClrasher: yea, for now, it will.
17:00:43rasherFor now?
17:00:45HClit might get bigger when/if i build dynarec for it
17:01:00HClat the moment its about 650k
17:01:15rasherWhere can I find it?
17:01:31HClsource on my ftp is pretty recent..
17:02:03linuxstbHCl: Why does it need so much memory? Are there static buffers that could be moved to the MP3 buffer area? (I know this isn't your top priority, I'm thinking long-term).
17:02:11rasheroh jesus!
17:02:14*rasher punches nautilus
17:03:06 Join mecraw [0] (~mecraw@
17:03:11HCli moved the two biggest buffers to the mp3 buffer area
17:03:24HCli tried to move some more, but they ended up crashing rockboy
17:10:47 Join Stryke` [0] (~Chairman8@resnet-241-86.resnet.UMBC.EDU)
17:11:56Bagderso the 650 is actual ram usage and parts of that are buffers?
17:13:01HCli think its mostly code
17:13:16HClbut i can check.
17:13:31DeadManQuestion: Will the RockBox FW for iRIver include an option to flash back to the original FW if needed?
17:13:47Bagderrockbox will not flash at all
17:13:53rasherThe original firmware will still be there
17:13:59DeadManoh yeah sorry forgot :)
17:14:05DeadMansilly me
17:18:06Bagderrasher: done any more work with the binary patches stuff?
17:18:20lolo-laptopwell you can use the original firmware to repatch a non-rockbox-bootloader version...
17:18:28rashernot really, I downloaded that uips and built it
17:18:30rasherworks fine
17:18:48rasherWe can't modify the source to include it in a graphical utility
17:19:07Bagderlicense issues?
17:19:29rasherBut, it's 5-6 years old
17:19:39rashermaybe if we could contact him we could get permission to do so
17:19:46rasherShould I try doing that?
17:19:55Bagdersounds like a good idea to me
17:21:17rasherI'll compose an email then
17:21:53rasherAlternatively, someone could reimplement it
17:21:53linuxstbSorry if this has been discussed before, but why do we need a binary patching program, couldn't we just write a front-end to Linus's mkboot?
17:22:19Bagderyes we could
17:22:30rasherAny reasons not to?
17:23:01Bagdernot if it is the same amount of work
17:23:14rasherI'm guessing it wouldn't be much different
17:23:17BagderI thought using a binary patch format would give us something free
17:23:27rasherwe'd want to add md5 checking to the ips program anyway
17:23:43Bagderthen let's skip the patch concept
17:23:44rasherbecause it doesn't look like the format does much of that
17:24:35linuxstbMaybe only Linus can answer this, but will his mkboot work with every iRiver firmware, including the hacked (different bitmaps) versions?
17:25:02BagderI think it works on all
17:25:12linuxstbIt would be good to have a list of "known good" MD5 checksums of iRiver .hex files though.
17:25:22 Join Hohoman [0] (
17:25:23Bagderwe should do that
17:25:38Bagderthe tool should have a list
17:26:05rashermaybe ability to keep the list up to date with a version on the website
17:26:21linuxstbYou mean every time iRiver release a new firmware? :-)
17:26:42rasherwell I was thinking of md5s of the result
17:27:13Bagderif we just get a simple tool to start with, we can expand it with fancy features later
17:27:20 Quit jyp (Remote closed the connection)
17:27:33linuxstbMaybe each bootloader.bin could come with a list of resulting MD5 checksums
17:27:49rasheryes, a package of some sort like that
17:28:06linuxstbOr just append them to the end of the .bin - the mkboot program will strip them out.
17:28:16rashersounds good
17:28:19 Quit lImbus (" Like's GUI? Then try HydraIRC -> <-")
17:28:24BagderI'd prefer one md5 for each bootloader.bin and one for each known good iriverfm
17:28:51rasherBagder: but there are multiple md5s for each bootloader.bin (depending on what you patch with it)
17:29:08rasher(version.. eu/us)
17:29:20Bagderno, bootloader.bin is still one
17:29:33Bagderyou mean the hex output
17:29:46rasheroh.. I misread
17:30:08linuxstbRemember we are only talking about the bootloader - it's not going to change very often. Maybe just have the bootloader.bin hard-coded into the GUI.
17:30:20Bagderlet's not overdo this
17:30:24linuxstbSo when the bootloader changes, we release a new version of the GUI.
17:31:15rasherand then keep a list of known md5s for iriverfirmwares
17:31:27rasherand for the resulting .hex
17:32:42linuxstbYes - the GUI would just have all the MD5s built-in.
17:33:03Bagderyes, and a maintainer-mode that produces a new set for a new bin
17:33:35BagderI mean if the hex should be md5'ed
17:34:34 Quit BBub (""Fange nie an aufzuhören. Höre nie auf anzufangen." - Tiki")
17:35:18 Join BBub [0] (
17:44:50linuxstbBagder: No, I don't think we even need that - we can just generate the .md5s (of the resulting patched .hex files) as part of the build process for building the GUI - new bootloader.bin means new release of the GUI.
17:45:57Bagderyes, but since we'd accept N different iriverfm-files, we'd need to generate a new list of hex-md5s when the bootloader.bin is updated. That would of course be done and included when a new tool is released.
17:46:29Bagderstill, those are easy things
17:46:52Bagdergetting a windows GUI tool is the bigger part, at least for me
17:48:19linuxstbI've played with wxWindows a little in the past - it's a very easy to use cross-platform GUI that compiles to the native UI for the target. It supports Win32, Unix/GTK, Mac OS X, plus others.
17:48:47linuxstbFor windows,, you can just distribute the program as a single .exe file with everything statically linked.
17:49:17 Nick courtc_ is now known as courtc (
17:49:22linuxstbIf it's still on the to-do list in a few weeks time, I may have a look at that.
17:49:40 Quit ep0ch (" HydraIRC -> <- IRC with a difference")
17:52:40HCllatest rockbox build doesn't do usb anymore
17:52:42HClwhat happened ?
17:54:35linuxstbThe version I'm using (including my plugin RAMSIZE changes) is working - I'll update from CVS and double-check.
17:55:23linuxstbHCl: Are you using a downloaded bleeding-edge build, or your own compile?
17:55:32ripnetuklate entry to conversation: can we have a java applet on the web page to do the mkboot thing?
17:55:41HClwith slight makefile / viewers.config adjustments for rockboy
17:55:56preglowdo we really need an easy way to make patched firmwares yet?
17:56:02ripnetukthere are 2 target 'markets' for mkboot - people like us who are quite happy with a mkboot and md5sum
17:56:12linuxstbripnetuk: Good idea - not Java, but maybe a cgi-bin Perl script?
17:56:17Bagderpreglow: no, it would be for upcoming fun
17:56:21ripnetukand 'end (l)users' who want to just patch their firmware
17:56:31linuxstbOr is that too close to us distributing patched versions ourselves?
17:56:36HClpreglow: i agree.
17:56:47HClwe don't want end users to try it just yet.. do we..?
17:56:56HClthats what linus said anyways
17:57:01ripnetuki think maybe java is better, as we dont want the entire multi-meg firmware going from user to your server
17:57:04preglowi say make a program that contains the patch and is able to do the patching on its own, complete with error checking
17:57:05ripnetukand back
17:57:16linuxstbI don't think anyone is actually planning on starting to code anything yet - we are just throwing around ideas.
17:57:30ripnetukpreglow - that is OK but we would need a windows version for the bulk of the users who dont understand cygwin and so on
17:57:59HClwget blah;unscramble -iriver blah;mkboot blah boot blah2;scramble blah2;md5sum blah2
17:57:59preglowripnetuk: yes, and so?
17:58:03HClits not that a complicated script
17:58:11preglowripnetuk: that was what i meant anyway
17:58:24ripnetukwindows code? in rockbox ;)
17:58:33linuxstbAs soon as Rockbox on the iRiver has sound, people will start to want the bootloader.
17:58:35preglowwe do have a win32 sim, don't we? :P
17:58:42ripnetuki could easily port mkboot to Delphi... very eay indeed
17:58:47ripnetukgotta go... home time
17:58:56preglowand i could easily port it to win32
17:58:58Bagderoops food time
17:59:17HCloh. wait.
17:59:26HClwindows might be fucking up again
17:59:33HCland taking a cached
17:59:37HClwhile i actually updated it on my ftp
17:59:44HCland it simply refuses to download the new one
17:59:47HClwindows is such crap
17:59:54 Quit ripnetuk ("Leaving")
18:01:04linuxstbHCl: Current CVS works fine for me.
18:01:54HClyea, it was windows being a total bitch again
18:02:05HCltaking a .zip out of its internet cache
18:02:09HClwhile the file has obviously been updated
18:02:14HClits such crap
18:02:34rasherI suggest you use a real browser :)
18:02:49HCli don't have much choice.
18:03:00linuxstbThat's what everyone says.
18:03:00HClif i type an url into an explorer window it still uses internet explorer
18:03:02HCland most of the time
18:03:03HClthats what i do.
18:03:10HCli'm not gonna sacrifice userfriendlyness
18:03:37HClthats the reason i stopped using linux for a desktop, cause it was far too much fucking work to set everything up
18:03:38bobTHCuse the firefox
18:03:52HClif firefox can't integrate with windows, no.
18:04:10rasherIt's perfectly integrated for me
18:04:20rasherexcept MSN Messenger *still* opens links in ie
18:04:22rasherbut that's about it
18:04:33HClwhat if you type an url into explorer?
18:04:37HClwhat does it do?
18:04:47rasherit opens it in explorer
18:04:52HClusing firefox?
18:04:53rasherI don't see how you could expect otherwise
18:05:05linuxstbThis may be a silly question, but why don't you type URLs into your web browser?
18:05:08HClyou realize explorer uses internet explorer as backend?
18:05:27HCllinuxstb: because a lot of the time i happen to have explorer windows open and its quicker to type it in there
18:05:48rasherI don't see why you claim that this is a problem with firefox' integration
18:06:00rasheryou are typing an url into one program, and expect another program to be opened?
18:06:06HClrasher: if firefox was properly integrated, explorer would use firefox as backend
18:06:16HClinstead of iexplore
18:06:53HCljust like how any standard media controls in programs are backends of media explorer?
18:06:58HClor should i just give up on trying to explain?
18:07:05rasherWell I understand what you mean
18:07:12HClsorry, i'm not in the best of moods since 2 computers are failing rather badly
18:07:25HCland i can't really be bothered to argue about internet browsers on top of that
18:09:48 Part Patr3ck_
18:12:48 Join einhirn [0] (
18:14:09HClrockboy suddenly crashes when i try to load it with the new memory sizes..
18:14:47 Quit bobTHC ("( :: NoNameScript 3.81 :: )")
18:15:15linuxstbHCl: There was also a change to the plugin API - did you merge those changes?
18:15:39linuxstbBut thinking about it, it shouldn't matter.
18:16:13linuxstbWhat was your plugin memory size set to before?
18:16:32HCl7 mb
18:16:44linuxstbDid you change the stack size?
18:18:11linuxstbI'm not sure what the problem is then. Were the 5 places I changed the plugin memory size the same as your changes?
18:18:43linuxstbObvious question - did you reboot your iRiver to load the updated firmware?
18:18:46HCllet me change.
18:18:52HClotherwise it would ofcourse, still work.
18:18:56HClsince i originally had my own version
18:22:11HClstill not working, i'd say your defines are broken
18:22:20HCli'll add #error tags to the 0x8000 sizes
18:24:42HClyour define in plugin.c is failing
18:24:43HClfor one
18:25:05HClCC plugin.c
18:25:06HClplugin.c:69:2: #error memerror
18:25:15HCl#if MEMORYSIZE >= 32
18:25:15HCl#define PLUGIN_BUFFER_SIZE 0x100000
18:25:15DBUGEnqueued KICK HCl
18:25:15HCl#define PLUGIN_BUFFER_SIZE 0x8000
18:25:15HCl#error memerror
18:25:19linuxstbI tested the ones in firmware/*.lds - but you're right, I didn't test the ones in apps/
18:26:25linuxstbYes, I get the same thing in plugin.c
18:27:34linuxstbI think I need to use MEM, not MEMORYSIZE
18:29:40linuxstbNot sure if that's a Makefile bug though - maybe a better fix would be to change the Makefile to define MEMORYSIZE. Anyone?
18:33:31linuxstbHCl: Does changing MEMORYSIZE to MEM in plugin.c fix your problem?
18:34:42 Join Aison [0] (
18:35:55 Nick Aison is now known as Aison^8km (
18:37:04HCllinuxstb: let me test
18:37:31 Join Tang [0] (
18:38:47Tangi've got another failure to mail the "comparison chart correection"
18:39:35linuxstbWhat kind of failure?
18:43:12Tangmail failure as this:
18:43:15HCllinuxstb: yup, works now
18:43:27linuxstbHCl: Thanks, I'll commit that fix.
18:43:33HClnp :)
18:43:44TangYou are not allowed to post to this mailing list, and your message has
18:44:28linuxstbI guess you need to subscribe to the mailing list then. I don't have anything to do with that part of things though.
18:45:12TangAh a mailing list
18:45:19Tangi'ts not evry pratical
18:45:26CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
18:45:26*rasher creates a wiki
18:45:46TangHum i'm not a Rockbox memeber
18:45:55Tangevene if i did some very minor wiki edit
18:46:01Tangcannot make a full page
18:46:55rasheryou can
18:50:05DiwayI have a "arlex.l" error trying to compile binutils CVS, could anyone help me ?
18:50:14Diwayand yes, I've installed bison
18:55:35***Saving seen data "./dancer.seen"
18:57:19Tangthnaks rasher very nice
18:57:44rasherit's not linked anywhere though
18:57:54linuxstbDiway: What environment are you using? Cygwin, Linux, etc?
18:58:10Diwayoups, right ! linux debian testing
18:58:27TangI geuss will be
18:58:40Tangi' just want to correct
18:59:13linuxstbI'm using Debian unstable. and had no problems at all. When you ran "configure", which configure did you use? I think there's one in the top-level "src" directory, and one in a binutils subdirectory. You should use the top-level configure script.
18:59:34DiwayI tried both, they ran fine but make fails
19:00:02linuxstbWhat does typing "bison −−version" display?
19:00:21Diwaybison (GNU Bison) 1.875d
19:00:32linuxstbSame as me. And "yacc −−version" ?
19:01:01Diway-bash: yacc: command not found, damn !
19:01:12linuxstbMy yacc is just a symlink to bison
19:01:35Diwayok, trying with this...
19:02:13linuxstbSorry, "yacc" is a symlink to a shell script containing two lines:
19:02:35linuxstbexec bison -y "$@"
19:02:49linuxstbAnd that script is called /usr/bin/bison.yacc
19:03:20Diwaylrwxrwxrwx 1 root root 22 2005-02-14 22:03 yacc -> /etc/alternatives/yacc
19:03:49Diwaylrwxrwxrwx 1 root root 21 2005-02-14 22:08 /etc/alternatives/yacc -> /usr/bin/bison++.yacc
19:03:51Diwaydamn !
19:03:53linuxstbYes, and /etc/alternatives/yacc is a symlink to /usr/bin/bison.yacc ?
19:04:20linuxstbmmm. Does bison++.yacc exist?
19:05:07Diwaysymlink to bison.yacc ?
19:05:27linuxstbYes, does /usr/bin/bison.yacc exist?
19:06:00TangWhat's this:
19:06:01TangSupports the VBRI header
19:06:02linuxstbFingers crossed then.
19:06:19TangIs it for good remaining time estimation for VBR files?
19:09:07Diwaylinuxstb> same error :/
19:09:44Tangmade my edits
19:09:45Diwaygot a stange msg in configure:
19:09:49Diwaychecking for bison... (cached) /home/diway/rockbox/build/src/missing bison -y
19:10:17linuxstb"cached"? DId you try deleting your build directory and starting again?
19:10:53Tangfeel free to correct syntax
19:10:58Tangi've to go
19:10:59DiwayI tried a make clean but not deleting, testing :)
19:12:06 Quit Tang ("Chatzilla 0.9.66 [Mozilla rv:1.7.5/20041108]")
19:17:00Diwaylinuxstb> It's working, thanks !
19:17:37linuxstbCool. Don't forget to add the path of your new binutils to $PATH before compiling GCC.
19:19:39 Quit hile ("enmä sitä täälläkään kuuntele")
19:19:54rasherHelp, that finnish guy is scaring me!
19:51:01HCli'm getting the idea cvs update doesn't check out new files?
19:52:40linuxstbIt may not get new directories unless you do "cvs update -d"
19:52:41rasherit does
19:53:34HCllinuxstb: ahh, nice, didn't know that
19:54:23linuxstbI think it's a feature, so if you delete a directory in your local copy, cvs update doesn't keep getting it again.
19:55:50HCli guess
19:56:02HCli really should dig more into how to use cvs
19:57:20linuxstbI think you now know everything I do.
19:57:34HCl P
19:57:35HCl :P
19:57:52HClare there any more requirements before rockboy can be added to cvs...?
19:58:13HCli guess i should get rid of warnings and stuff?
19:58:44linuxstbYes, and also make the build process clean so it doesn't break any other targets.
19:59:05HCllet me alter its makefile a bit
19:59:29 Quit lolo-laptop ("Client exiting")
20:02:48 Join xen` [0] (
20:08:25rasherlinuxstb: oooooh
20:10:06rasherlibmad in cvs :)
20:10:23linuxstbThat's the easy part!
20:11:25rasherWhat do you mean?
20:12:00linuxstbWell, we need to write the new audio/codec system...
20:12:40rasheroh, like that
20:26:12*HCl sighs.
20:26:19HCli absolutely suck at making makefiles :(
20:26:57linuxstbHCl: I'm struggling with the same thing at the moment.
20:27:05HClwho wants to be nice and help me change the Makefile of rockboy to actually work nice?
20:27:09HCl :/
20:27:49preglowi dislike makefiles
20:28:02HClthere should be a replacement system..
20:28:04preglowi would prefer a simpler and smarter system
20:28:12preglowbut i can't find anything that's good
20:28:22preglowlast thing i tried was jam, but that has absolutely no docs
20:28:25HClsomething perl based, probably
20:28:35preglowi've considered writing my own simple system
20:28:44preglowmy major requirement is that it calculates dependencies itself
20:28:50preglowlike scanning for .h files, etc
20:29:08HClits hard to do that for .o / .c files when linking
20:29:11HClbut for .h is possible
20:29:35*HCl sighs.
20:29:45HCllinuxstb: what is the policy here, keep gnuboy as much standard as possible?
20:30:11HClor just modify it so it'll work nice with our current build system?
20:32:17linuxstbMy strategy for the codecs is going to be to commit the original files to CVS (i.e. unmodified from the originals), and then slowly get them to build cleanly - making the minimum number of changes to the original files.
20:32:32linuxstbI think it depends if gnuboy is still actively developed and if you want to merge changes later.
20:32:37HClits not.
20:32:49HClit hasn't been altered in 4 years
20:33:04HClso i guess i'll just go ahead and shape it for rockbox needs?
20:33:19linuxstbI can't think of a good reason why not.
20:33:30HClthat makes things easy, somewhat.
20:34:15 Quit Stryke` (Read error: 54 (Connection reset by peer))
20:44:45 Nick Aison^8km is now known as Aison (
20:45:04 Join DMJC-L [0] (
20:46:25 Join Chamois [0] (
20:47:03 Join lostlogic [0] (
20:48:57 Join Tang [0] (
20:49:01TangHello guys
20:49:03Tang'im abck
20:49:11Tangbut not be late
20:49:25Tangso the comparison chart was it OK?
20:50:46rasherI fixed a colour you didn't change
20:51:43Tangnice thanks a lot :)
20:52:05Tangmaybe there is some mistakes lefts
20:52:18Tangdue to my difficultie to speak english
20:52:26Tangand some things i do'nt know
20:52:35Tanganyway in Wikimode anyone can edit
20:52:41TangSo i've to go
20:52:59Tangwas there some progress today? :)
20:55:37***Saving seen data "./dancer.seen"
20:58:20 Join Stryke` [0] (~Chairman8@resnet-241-86.resnet.UMBC.EDU)
21:03:07Tangbye all guys
21:03:33Tangmmm gonna check on rbx board if Linus has received my remote
21:03:56Tangyeah indeed
21:04:03Tanghe received it today :)
21:09:46Tangi go now
21:09:53Tanggood evening
21:09:55 Quit Tang ("Chatzilla 0.9.66 [Mozilla rv:1.7.5/20041108]")
21:11:59HCli hate makefiles!
21:12:27HClmake[2]: Entering directory `/home/hcl/rockbox/rockbox-devel/apps/plugins'
21:12:27HCl/bin/sh: -c: line 0: unexpected EOF while looking for matching `"'
21:12:27HCl/bin/sh: -c: line 1: syntax error: unexpected end of file
21:12:27HClmake[2]: *** [/home/hcl/rockbox/rockbox-devel/iriver/rockboyobj.o] Error 2
21:12:32HClwth is that supposed to mean :(
21:13:33 Join hile [0] (
21:13:46HClsomeone seriously needs to write a replacement for make...
21:13:47Suckayou used a ' instead of a ` ?
21:13:56Sucka(i dont knowwhat im talking about btw)
21:14:03HCli have no idea, all i did was move stuff around in my make file.
21:14:52linuxstbRemember that a \ at the end of a line continues that line onto the next line. Is that the problem?
21:16:44HClmeh, i got it
21:19:39HCli hate the make system, seriously.
21:19:40HCl :(
21:33:03preglowlinuxstb: will we use most of timer.c at all?
21:34:15 Join XShocK [0] (
21:34:16linuxstbpreglow: I don't know. I haven't looked that closely.
21:35:20HCldoes anybody know the make variable to get all requirements to build a file? (everything behind the :)
21:35:32HCl$< is for just the first one...
21:35:48HCl$@ is for the thing thats getting built (in front of the :)
21:36:05HCl":", not the smiley >.>
21:37:28rasher$* perhaps? $- ?
21:37:31rasherI'm just guessing
21:37:34rasherI know nothing about make
21:38:09linuxstbI'm not a Make expert, so I normally set all the requirements to a variable - e.g. "target: $(OBJS)" and then use $(OBJS) in the command
21:40:27*HCl slaps make
21:42:23HClthat clearly didn't work.
21:42:49HClguess i can't do that.
21:44:42HClokay, i think i built a makefile that'll now put everything into $(OBJDIR)/rockboy/*.o
21:45:18HCli'll try to fix it up some more..
21:52:00HClokay. now its pretty cleanish.
22:02:55HClthe only bad thing is that it has a dummy .c till the plugin system can be adapted to use seperate dirs for each plugin rather than a single .c
22:02:58HCland i'm lagging :(
22:05:07amiconnHCl: Does it compile for the sim now?
22:06:21HClhold on, still testing it building for iriver
22:06:26HCllast time i built it for the sim it crashed
22:06:29HClbut i'll try again after this
22:08:16XShocKI can't understand. Coldfire has its own I2C controller. does the processor itself has special pins for that or this controller can be configured to use pins you want?
22:13:00preglowspecial pins
22:13:10preglowmost pins probably have shared functions
22:14:35XShocKyeah. found that out. :)
22:18:32 Quit methangas (" HydraIRC -> <- IRC has never been so good")
22:20:57HCli found a bug.. i think.
22:21:01HClwhen running a plugin
22:21:06HCland you plug in usb
22:21:12HClhdd accesses crash the plugin
22:22:55HClaside from that, my new makefile stuff seems to work fine
22:23:06rasherare you sure it's not that your plugin is failing to exit itself?
22:23:07HClamiconn: i'll try the sim build in a bit
22:23:27HClit has a check when it can't open a file
22:23:30rasheroh you mean because your plugin is doing hdd access?
22:23:30HClto return giving an error
22:23:43HCli plug in usb while booting the plugin
22:23:50HCland the plugin crashes when it accesses the hdd
22:24:22rasherah, like that
22:28:02 Join sox [0] (
22:29:14soxhoy all
22:31:02 Join hubble [0] (
22:32:06hubbleanyone willing to take a look at the uda1380 code? it worked on sunday, but I broke it yesterday and have spent too many hours trying to fix it :)
22:32:34soxwhat do Daves libmad commits to the CVS mean in practice? Cuz the API for connecting the codec is not ready, is it?
22:34:49Bagderit means the code is in cvs
22:35:22Bagderit is work in progress
22:35:49rashersox: it means that you can (possibly) convert an mp3 file to a wav
22:35:51rasherthat's it
22:36:18 Quit BBub (Read error: 113 (No route to host))
22:36:31soxallright, i understand
22:37:04soxbut its the same code that will - in a close future i hope - add the sound output functionality?
22:37:17soxjust not connected to rockbox yet?
22:37:24preglowit will decode mp3s, yes
22:37:53rasherand no sound output either
22:48:01 Quit Chamois ("CGI:IRC (EOF)")
22:48:06 Join Chamois [0] (
22:51:47 Quit Chamois (Client Quit)
22:55:40***Saving seen data "./dancer.seen"
22:59:53XShocKhubble: I would like to. :)
23:00:22HClamiconn: got any time to look at my latest version? no idea why its not running in the sim..
23:00:32HCli'll pack just the files i modified.
23:02:50 Quit elinenbe (" HydraIRC -> <- \o/")
23:06:22amiconnHCl: Which sim did you try, Win32 or X11?
23:07:19 Quit Sucka ("( :: NoNameScript 3.81 :: )")
23:08:09amiconnDidn't you say you use windows?
23:08:21HCli do, but i develop on linux
23:08:27HClremote vnc
23:08:39HCli can't be bothered to install a complete cygwin environment on my laptop.
23:09:31amiconnI have that, including X11.
23:10:11amiconnImho the win32 simulator has better rockbox "compatibility" than the x11 version
23:10:17preglowi just plain hate cygwin, linux is preferable
23:10:29HClwell, i'd try the win32 version...
23:10:40HClwell, i'm not really fond of cygwin
23:11:06amiconnThe win32 simulator *should* be compilabe in MS Visual C as well, however, I can't verify
23:11:19HCli have that.
23:11:28amiconnThe MSVC makefile most likely needs to be adjusted to my latest changes
23:11:50Bagderspeaking of that, its been a good while since Hardeep was around
23:12:05amiconnI added the multiple UI pic thing, but I can't adjust MSVC
23:12:43rasherI created a featurecomparison chart in the wiki by the way
23:13:01rashersomeone should link that from the frontpage
23:14:15HClmm, whats the url? i wanna see
23:14:26 Join _aLF [0] (
23:14:56rasherit's just the regular one imported to the wiki
23:14:56preglowas for iso8859-1 support, doesn't the iriver os support htat?
23:15:25rasherpreglow: yeah, it sure does
23:15:33rasherI think it even supports utf8 for ogg tags
23:15:40Bagderpreglow: that feature is for "player" as in the Archos player models
23:15:40rasherbut limited charset of course
23:17:08preglowwhy doesn't rockbox support sound while ff/rw, btw?
23:17:34rasherfast forward/rewind
23:18:53preglowfor the xing vbr header you can put in the remark 'only records in cbr'
23:19:13Bagderright, xing makes no sense then
23:19:25preglowwhat't the xing header good for anyway?
23:19:27Bagdersame goes for VBRI
23:19:41Bagderpreglow: finding a position in a VBR file
23:19:51preglowthe words 'xing' and 'mp3' together makes me think of the crappy encoders they had some years ago
23:19:54preglowBagder: ok
23:20:16Bagderxing and vbri are two different approaches to do the same thing
23:21:18preglowrasher: and by pitch control, what do you mean? 1.6 iriver firmware has crude pitch control
23:21:48rasherpreglow: I don't know - I just copied the original verbatim
23:21:51rasherI didn't change a thing :)
23:22:26amiconnBagder: Some russian player users keep asking for cyrillic support. So I'm putting together a table with cyrillic chars, possible direct mappings to lcd hardware chars, and the replacements I could think of. Then I'm going to attach this to the forum thread, and let the russian guys decide about the missing replacements. If they manage to find replacements for all cyrillic letters, I'll make a first (hacky) player rockbox version with 'RockCyrill
23:22:27rasherbut yes, I'd say that should be a "Yes" too
23:22:30preglowwell, won't hurt to correct stuff
23:22:34rasherprobably Yes (limited)
23:22:42rasherand put that in Text as well
23:22:43preglowsounds like shit it does as well
23:22:46preglowsince it doesn't interpolate
23:22:53Bagderamiconn: nice
23:22:53rasherto get rid of the long text
23:22:55preglowso 'limited' indeed
23:23:28rasherare you correcting?
23:23:38rasherapparently not :)
23:23:39*rasher does it
23:23:52preglowno, i'm not
23:23:59preglowi keep forgetting it's a wiki, heh
23:24:16XShocKI as i explored, there is only one cyrillic font. kind a pity since I liked the font by default, and only crynx1 support russian
23:24:29*rasher changes some of the longer explanations
23:24:53preglowlike TEXT folder explanation isn't really necessary
23:25:01preglowit's capable of viewing text, that's the criterion
23:25:09rasherhow about a (limited) there?
23:25:13Bagderyes, the comparison is meant to be simple
23:25:39preglowand for 'boot time' a simple 13-30 will do
23:25:56rasherhigh resolution volume control :|
23:26:19rasherNo but useless for the built-in LithIon Polymer battery -> No (not needed)
23:26:37Bagderor just N/A
23:26:51preglowyes, that's probably better
23:28:13*rasher removes the version information as well
23:29:25Bagderthey are better off below in some explanations if at all
23:29:43rasherI don't think it's important
23:29:54rasherdo they have the feature or not
23:30:04preglowi'd say you'd have to assume we're talking about the latest firmware anyway
23:31:00rasherQueue songs to play next: Yes (but only 1 song in queue)
23:31:02rasherwhat about that?
23:31:11XShocKI think the best way t compare would be something like green-blue-red since it is confusing for me, why "File Delete & Rename " "Partly(only delete)" is also green.
23:31:26Bagderrasher: how about a footnote like on the devicechart?
23:31:54preglowfor the file delete, i'd remove the whole Partly thing and write 'Files only'
23:32:16preglowahh, no, that won't do
23:32:22preglowFile delete only
23:32:39Bagderpossibly make it two lines
23:33:02rasherfile delete:
23:33:05rasherfile rename:
23:39:26Bagdernumber the footnotes in the bottom as well
23:39:30 Quit mecraw (Read error: 104 (Connection reset by peer))
23:39:43rasherjust noticed that :)
23:39:53Bagderbut otherwise it looks just fine
23:40:01 Join mecraw [0] (~mecraw@
23:40:54Bagderwhy is there a question mark above the table?
23:41:14rasherI don't know
23:41:28rasherdoesn't make sense
23:41:35amiconnrasher: " Generates XING VBR header when recording" should be N/A for iriver (?)
23:42:01Bagdersince it only records CBR
23:42:26*HCl stares at the video plugin and wonders whether that can be done on the iriver?
23:42:47BagderHCl: it most certainly can, it just takes lots of work
23:42:47amiconnHCl: I believe it's possible
23:43:25amiconnThe iriver firmware is really only available in _1_ language ?!
23:43:41HClwell, i guess there's not much point in making more grayscales for rockboy
23:43:45rasherthere's a korean version isn't there?
23:43:45Bagderdr grayscale^H^H^Hamiconn can do miracles ;-)
23:43:50HClsince all gameboy games natively support a 4bit color
23:44:31rashershould we have a Games (iRiver) row?
23:44:41amiconnHCl: I'd say the 4-grayscale mode should be supported by the rockbox core on iRiver.
23:44:54HClamiconn: ofcourse o.o
23:45:00preglowamiconn: no
23:45:04preglowamiconn: you can select language from a list
23:45:14amiconnMore grayscales are possible as a plugin lib as on the archos
23:45:18rasherpreglow: but it doesn't change the menu languages
23:45:31rasherit.. does?
23:45:36preglowgimme a sec
23:45:50preglowwhat the hell?
23:46:06rasherI really don't know what it's supposed to do
23:46:09amiconnHCl: The grayscale lib allows for up to 33 grayscales on the b&w-only archos display.
23:46:15preglowapparently, i never actually tried this before
23:46:20rasherTang had written that it was for charset
23:46:23rasherbut that doesn't make sense
23:46:28preglowand must have dreamt of the multiple languages
23:46:34HClamiconn: sweet.
23:46:35preglowrasher: id3-tags
23:46:42rasherYes, but it should just have been "Western European" or something
23:46:50amiconnMy guess is that it's possible to get 49 grayscales out of the iRiver lcd by using a similar technique
23:47:12HCli'm still rather anxious to see how well rockboy will do with 140mhz and less waitstates for ram
23:47:35rasherpreglow: but I think there's a korean version
23:47:39*rasher changes to 2
23:47:45preglowrasher: that's probably where i got it from, then
23:48:07rasherthat's a different firmware though
23:48:23amiconn'boot time from flash' should say '8-12' for archos
23:48:23preglowbut you can change to english
23:48:39amiconnThe archos recorder takes ~12sec, but the player only ~8sec
23:48:48rasherwhat's pre-recording?
23:49:30amiconnalias retro-recording, recording before you actually press the record button
23:49:38rasherthat's a no then to capture the <n> seconds before the press as well
23:50:24Bagderthe Archos site claims it is patented ;-)
23:50:52amiconnThe 'archos' column is in fact a bit too much 'general'.
23:50:54rasherThere should be lots of prior art :)
23:51:00coobbet there's prior art for that
23:51:20Bagderwell, perhaps THEY don't own the patent :*)
23:51:38rashermaybe they made it something like "pre-recording on portable mp3-recording devices"
23:51:52rasherin which case it may be true, but a bogus patent anyway
23:52:04amiconnrasher: The (Recorder) for 'screensaver style demos' can be removed
23:52:10preglowbut btw, who has the mp3 license for my player? me or iriveR?
23:52:19amiconnThere are now some for the player too :)
23:52:40Bagderpreglow: very good question
23:52:45preglowBagder: i think so as well
23:52:56preglowit might hamper us releasing binaries
23:53:23HClwhats the use for pre recording...?
23:53:36BagderHCl: recording anything that you don't know exactly when it starts
23:53:41rasherHCl: if you're recording from tv/radio for example
23:53:46HClthats sweet.
23:53:51rasherit's a very nice feature
23:53:56HClyea, i figure
23:54:15coobwell tivos do it...
23:54:47preglowyou could of course release binaries that only play vorbis ;)
23:55:11HClwhy are iriver games N/A rather than no?
23:55:22preglowi think frauenhofer have said that they wont pursue personal use anyway, but that's no guarantee
23:55:45rasherHCl: because it's "Games (Recorder)"
23:55:58*rasher adds a Games (iRiver)
23:56:02rasherhow many games do we have?
23:56:22HClhaven't checked others..
23:56:27preglowmine sweeper
23:56:32Bagderthe recorder ones should all work on iriver as well
23:56:58DMJCsnake 2
23:57:03*rasher counts
23:57:08DMJCthe games I played sucked
23:57:13HCli'm guessing i have to get rockboy to run in sim before its allowed into cvs?
23:57:14DMJCthey don't use enough screen size
23:57:22Bagderyou are too kind DMJC
23:57:24HClthats cause they haven't been adjusted yet.
23:57:27DMJCI modified snake2 to fix it
23:57:30preglowDMJC: mine sweeper sucks now????
23:57:31HClwe should have a todo list.
23:57:40HClthat people can work on.
23:57:41rasher8 yes
23:57:41DMJCactually minesweeper worked!
23:57:54HClcause i'm generally bored here, while i can obviously help with other stuff, i just don't know what
23:57:56rasher(and solitaire with my patch :>)
23:57:59 Join amiconn_ [0] (
23:58:02linuxstbTODO: Item 1: Write TO DO list
23:58:05HCl :P
23:58:07DMJCI took a crack at rockblox..
23:58:16DMJChad rather hillarious results
23:58:29DMJCnow the blocks fall upwards
23:58:30HClwe should add those patches into cvs, really
23:58:37 Quit amiconn (Nick collision from services.)
23:58:38 Nick amiconn_ is now known as amiconn (
23:58:46HClafter i had some food
23:58:57HCli'll look at it

Previous day | Next day