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 2005-07-15

00:00:11SlasheriamaroK's current engines doesn't have true gapless at all :/
00:00:12Bagdernow ipodlinux has a fine player ;-)
00:02:20Bagderwell, volume support feels a bit urgent
00:02:55MirfleI've updated the hebrew patch ( to current cvs, but I'm having trouble with patching.
00:03:01Slasherifortunately i always use the external amplifier for that ;)
00:03:16 Quit ansivirus (Read error: 145 (Connection timed out))
00:03:39BagderMirfle: what's the problem?
00:03:43Mirflesince this: "diff -u oldfile newfile > patch" only does one file,
00:03:44SlasheriBagder: btw, i think crossfade should work when we return true value from the pcm_is_playing() stub
00:04:17Mirfleand this: "diff -u olddir newdir > patch" won't patch the subdirs
00:04:38BagderMirfle: 1) you can append more diffs to the same patch or 2) use diff -r to do it recursively
00:05:35 Quit Harpy (Read error: 110 (Connection timed out))
00:06:00Mirflehow do you append the diffs? just with a word processor?
00:06:09Bagder>> output
00:06:12FebsCan I make a suggestion for the General Settings −−> Playback −−> Repeat options?
00:06:15Bagderinstead of > output
00:06:28Mirfleoh, thanks
00:07:23FebsCurrent options are Off, One, All and Shuffle. "All" is ambiguous. It's not clear whether it means all files in the directory or on the player.
00:07:24 Join tiegs [0] (
00:07:37FebsPerhaps Off, Song, Directory, and Shuffle would be more intuitive?
00:07:45Bagdernot really
00:07:53Bagderrockbox always works on playlists internally
00:08:02Bagdera dir is a playlist too
00:08:24Slasheri|#if CONFIG_HWCODEC == MASNONE && !defined(SIMULATOR)
00:08:24Slasheri| pcmbuf_crossfade_enable(global_settings.crossfade);
00:08:39SlasheriBagder: that #if should be fixed from settings menus also
00:08:40FebsUnderstood, but from an end user's perspective ...
00:08:42Slasherithen it should work
00:08:53Slasherinights everybody ;) ->
00:08:54BagderSlasheri: but I could change the setting in the sim?
00:09:07BagderFebs: I see your point
00:09:08SlasheriBagder: yes, but it doesn't really change anything :D
00:09:17BagderSlasheri: ok ;-)
00:09:25Slasheri(before that !defined(SIMULATOR) is removed)
00:09:28BagderFebs: but I don't know what the best fix is
00:10:34FebsThe reason I make the suggestion comes from a problem that someone posted on the forum today, where he couldn't figure out why "Move to Next Folder" wasn't working, and as I tried to replicate his problem, I
00:10:46BagderI read it
00:11:11Mirfleanother patch question: haw do I add a new file to the patch?
00:11:20FebsAnyway, just a suggestion.
00:11:57BagderMirfle: diff -uN existing nonexisting
00:12:12Bagderthe -N being the key
00:12:49BagderFebs: yes, but any repeat in such a situation is odd
00:12:51Mirflewhat nonexisting file am I supposed to write?
00:12:58BagderFebs: whatever it would be called
00:13:10BagderMirfle: a name of a file that doesn't exist!
00:13:22Mirfleohh, duh o.o
00:14:30Coldtoasthey. so did you guys decide if youwere going to add the limiting option? as in what limits what to avoid clipping?
00:15:31 Join webguest53 [0] (
00:16:33 Quit webguest53 (Client Quit)
00:17:22ashridahColdtoast: they've already got volume limiting.
00:17:32ColdtoastI know
00:17:51ashridahaah, misread.
00:20:34 Quit Febs ("Chatzilla [Firefox 1.0.5/20050711]")
00:26:00Bagderadded anims to the plugin wikis
00:27:25amiconnEek, getting into trouble with the timer....
00:28:21amiconnI need to have 2 timer ISRs for the same timer interrupt, and switch between them...
00:28:33 Quit Moos (" HydraIRC -> <- Leading Edge IRC")
00:29:11Bagdersleep time
00:29:41MirfleBagder: you got mixed up, it should be nonexisting existing, now, trying to apply the patch i
00:30:11Mirflei'm being warned about deleting nonexisting files :)
00:34:54 Quit Coldtoast ("Peace and Protection 4.22")
00:43:46 Join TCK [0] (
00:45:19 Part bobbers
00:46:50 Quit thegeek (Read error: 104 (Connection reset by peer))
00:48:05Mirfledoes compiling leave a log file?
00:54:08MirfleYey! It worked :)
01:00:23 Quit rooomish2 (Read error: 104 (Connection reset by peer))
01:01:13 Join KasKaraK_ [0] (
01:03:42 Join Cassandra [0] (
01:04:04CassandraAnyone here done the H140 battery mod?
01:06:42CassandraOr come to think of it, anyone know how to get the iRiver innards out of the case?
01:07:52Mirflecan anyone here with cvscommit access review my new hebrew patch?
01:08:15***Saving seen data "./dancer.seen"
01:09:35ashridahthere's four screws on top, 4 on bottom, and one on each side. after that, it should be relatively easy
01:09:42ashridahhope you've got the right screwdriver :)
01:13:50 Join jpegreen_ [0] (
01:14:39jpegreen_Hey all does anybody know the page on the site that lists the plugins and their compatability to which players?
01:16:15Cassandraashridah: The plug seems to be hidden under the circuit board.
01:16:34CassandraI think I've found the four screws holding the board down now.
01:18:48 Quit KasKaraK (Read error: 110 (Connection timed out))
01:20:19 Join asdsd____ [0] (
01:20:26 Part asdsd____
01:24:39 Quit ashridah ("uni")
01:28:38 Quit Mirfle ("Chatzilla 0.9.68a [Firefox 1.0.4/20050511]")
01:40:07 Join austriancoder [0] (~austrianc@
01:41:54*Cassandra swears at this disassembled H140.
01:42:11CassandraI can't unplug the damn battery because the power inlet is in the way.
01:45:42 Join memmem [0] (
01:46:11memmemHi austriancoder, I found the checksum in the M3 and X5 firmware files.
01:46:38 Quit lodesi (Remote closed the connection)
01:53:28CassandraCool. Upgrade successful. I'll leave it to charge overnight and screw it together properly in the morning.
01:54:13austriancodermemmem: could you give me some more infos?
01:55:23 Quit cYmen ("zZz")
01:56:49memmemThe byte at offset 0x102b contains the 8-bit sum of all the bytes starting with the one at 0x1030.
01:57:46memmemI've patched the original firmware to record 1024 MB (instead of 256 MB); will test this over night.
01:58:00austriancoderah fine
01:58:11austriancoderif it worked, let me know it...
01:58:28austriancoderwe will need this for our fwpatcher
01:58:39memmemStill 7h10 to go...
01:59:19CassandraLithium batteries are so tiny these days.
01:59:27*austriancoder downloads a x5 firmware bin
01:59:41memmemAnd yes, the patched firmware is accepted and boots (only if the checksum is patched correctly).
02:00:52austriancoderok.. so you also know how to calcualte the checksum?!
02:01:14memmemSee above.
02:01:47austriancoderah ok..
02:02:06*austriancoder opens the bin with a viewer
02:02:06 Quit ]RowaN[ ()
02:05:59austriancodermemmem: have you a bdm?
02:06:22 Quit silencer (Nick collision from services.)
02:06:23 Join silencer [0] (
02:06:25 Quit silencer (Nick collision from services.)
02:06:52jpegreen_can any one point me to the sites page for plugins and their compatibility to which players?
02:09:59austriancoderwhat happens if the checksum is incorrect?
02:10:20CassandraAmazingly, it's linked directly off the wiki home page.
02:10:58memmemIf the checksum is incorrect, the firmware file is silently ignored and then deleted.
02:11:09jpegreen_Cassandra: Thanks
02:11:25 Join silencer_ [0] (
02:12:05jpegreen_Cassandra: I am amazed that I couldn't find it ... Ughhhhh (Napolean Dynamite sigh)
02:12:22austriancodermemmem: good.. have you done a patching routine yet?
02:13:02memmemaustriancoder: no, so far, I've patched only one byte and the checksum.
02:13:15 Quit hicks (Remote closed the connection)
02:14:40austriancodermemmem: ok... i will hack a little checksum-updater
02:15:05*memmem goes to bed
02:15:08 Quit memmem ("ERC Version 5.0.3 $Revision: 1.726.2.17 $ (IRC client for Emacs)")
02:18:13jpegreen_Does anyone know if there are any .wps files for the studio 10 player?
02:23:13 Quit tiegs ("CGI:IRC (Ping timeout)")
02:25:30CassandraGuess not. You could be the first to make one.
02:26:20jpegreen_Cassandra: I Think I will give it a try... Its been awhile since I hav ewritten any C
02:27:05CassandraWPS files aren't written in C.
02:27:53jpegreen_Cassandra: Thanks again for the slap in the right direction... Good looking out
02:31:45jpegreen_Cassandra: I have another question? Is it possible for the Studio 10 to have a clock plugin?
02:35:27jpegreen_Cassandra: I didn't see one listed on the plugins page but I am not afraid to try and make one if it is possible
02:36:19CassandraIt doesn't have a clock, I'm afraid.
02:36:50jpegreen_Cassandra: Ugh oh well it would have been nice
02:39:10 Join thegeek [0] (
02:45:59 Part jpegreen_
02:49:43 Part floam ("Adios")
02:58:38 Quit austriancoder ("using sirc version 2.211+KSIRC/1.3.12")
03:01:21 Part KasKaraK_
03:08:18***Saving seen data "./dancer.seen"
03:13:39 Join CheeseBurgerMan [0] (~youshould@
03:20:19 Join Rori [0] (
03:21:02RoriHey this is pretty cool. Someone did a SACD to DTS 5.1 conversion of Jeff Wayne's War of the Worlds remastered.
03:30:05 Quit Sucka ("a bird in the bush is worth two in your house")
03:46:10 Quit CoCoLUS (Read error: 110 (Connection timed out))
04:03:42 Join StrathAFK [0] (
04:05:36 Join QT_ [0] (as@area51.users.madwifi)
04:09:48 Join TCK- [0] (
04:10:21 Quit TCK- (Client Quit)
04:15:47 Quit QT (Read error: 110 (Connection timed out))
04:22:10 Quit Strath (Read error: 110 (Connection timed out))
04:43:28 Join jamesshuang [0] (~james@
04:57:12 Join Maxime [0] (
04:57:13 Quit Maxime`Mrn (Read error: 104 (Connection reset by peer))
05:04:29 Join amiconn_ [0] (
05:08:19***Saving seen data "./dancer.seen"
05:23:00 Quit amiconn (Read error: 110 (Connection timed out))
05:23:00 Nick amiconn_ is now known as amiconn (
05:38:09 Join MrStaticVoid [0] (
06:30:56 Nick CheeseBurgerMan is now known as CBM-away (~youshould@
06:47:00 Quit TCK (Read error: 104 (Connection reset by peer))
06:47:11 Quit jamesshuang (Read error: 110 (Connection timed out))
07:08:22***Saving seen data "./dancer.seen"
07:15:33 Join webguest06 [0] (
07:15:54 Join LinusN [0] (
07:16:31 Join jpegreen_ [0] (
07:17:05jpegreen_Hey all
07:17:36jpegreen_LinusN: I want to thank you again for all your help yesturday.
07:17:44LinusNyou're welcome
07:18:53*LinusN gets a cup of hot black goo, it says "coffee" on the machine
07:20:11jpegreen_LinusN: I was wondering if it was possible to create a .wps file for the Jukeboxstudio 10
07:21:05LinusNoh yes
07:21:40LinusNit's basically the same thing, except you only have 2 lines
07:23:05jpegreen_LinusN: Cool I guess I know one of my new pet projects I just hadn't seen one on the plugin DL pages or in the forums. Is it possible to create a clock plugin for it too
07:23:51LinusNthe studio doesn't have a real time clock chip
07:24:05jpegreen_Ughh... Oh well
07:37:02 Join justmejg [0] (
07:38:00 Join Coldtoast [0] (
07:39:35amiconnGood morning
07:45:58amiconnLinusN: I have a problem concerning timers.
07:46:04LinusNtell me
07:46:21amiconnI want to make plugin_(un)register_timer() functional
07:47:01amiconnHowever, the coldfire only has 2 timers, and timer00 is already taken by the tick, so the timer1 must be shared between backlight and plugin timer
07:47:01 Join ashridah [0] (
07:47:13 Nick QT_ is now known as QT (as@area51.users.madwifi)
07:47:25amiconnI don't know what is better:
07:47:37amiconn(1) Have one ISR, and dispatch in this ISR
07:47:52amiconn(2) Have 2 ISRs, and rewrite the vector
07:48:15amiconn(3) Have 2 ISRs and use 2 different vectors by assigning different levels
07:49:58LinusNi'd say (1), but it will get nasty in any case
07:50:38amiconnThe simplest is (3), as long as we don't use up all autovectors...
07:50:43amiconnAlso, I think that plugin_register_timer isn't that clean a solution, because it fiddles with the timer in application code...
07:51:25amiconnJörg added it back then to avoid fiddling with the hardware in a plugin....
07:51:56LinusNi think a more generic timer function is in place
07:52:05amiconn(1) would (imho) require a firmware/apps/timer.c
07:52:21amiconnOtherwise it is very unclear where the ISR should reside
07:52:59amiconnErm, firmware/drivers/timer.c of course
07:53:22LinusNi was confused for a minute there...
07:53:45amiconnThen there's the question what to do on the archos.
07:54:01amiconnThe SH1 has 5 timers (which is A Good Thing, btw)
07:55:08LinusNi think a generic function is in order, like set_timer(int us, void (*callback)(void));
07:55:24LinusN"call this function in x us"
07:55:44LinusNor even "each x us"
07:55:53 Quit jpegreen_ (Read error: 110 (Connection timed out))
07:56:11LinusNthen it's up to rockbox to handle the timer allocation
07:56:19LinusNand the plugin doesn't care
07:56:20amiconnLinusN: We need cpu clk resolution on archos
07:57:24amiconn...for the video playback, as .rvf defines the frame time as # of cpu clocks, based on 11.0592 MHz
07:57:42amiconn..which already created a problem on the Ondio, btw
07:59:19amiconnvideo.rock on the Ondio has to recalculate the frame time for 12 MHz clock (*625/576), but this causes roudoff errors
08:00:07amiconnNot that big a problem with videos on the Ondio, since you won't typically have very long clips on it, but on a HD unit there may be a problem
08:05:24amiconnreboot, brb
08:05:28 Part amiconn
08:07:52 Join B4gder [0] (
08:08:44 Nick StrathAFK is now known as Strath (
08:09:13LinusNi wish the iriver had a scf5250 too :-)
08:09:50pabswish in one hand...
08:09:52*pabs whistles
08:10:00LinusNmemory to memory dma, 128kbyte sram
08:10:46 Join amiconn [0] (
08:10:52LinusNwelcome back
08:11:08 Join Harpy [0] (
08:13:17amiconnLinusN: On a related topic, it seems the grayscale lib is very possible on the iriver.
08:13:53amiconnIf I calculated correctly, the LCD has an internal repeat rate of 72 Hz in 4-grey mode, and 144 Hz in monochrome
08:14:11amiconn(I think I need to switch to monochromo for the gs lib)
08:14:12B4gderbtw, I vote we remove the gmini builds from the cvs build table
08:14:40amiconnTransferring the frames will take < 10% CPU at 120 MHz
08:15:23LinusNB4gder: good idea, let it rest in peace until someone digs it up again
08:15:36amiconnThe downside is that I will most probably have to boost for the same reason as the backlight fading requires it
08:15:41B4gdermy thinking too, nothing has happened for months in there
08:16:23amiconnB4gder: Do you say that because you would otherwise have to fix the gmini sim? ;)
08:16:32B4gderyes :-)
08:17:36B4gderactually, that was what made me think
08:17:41amiconnI'd vote for fixing it ;)
08:18:58amiconnLinusN: Your timer function idea might be the way to go. We could add a time_base value, allowing to select µs or 11.0592 MHz base (on archos)
08:19:43amiconnI also thought about a high precision timer for a while, allowing to request timings that lead to a non-integral number of cycles
08:20:43amiconnThat would solve the Ondio problem perfectly, but I'm still not sure what to do with varying CPU clock...
08:23:44 Quit justmejg (Read error: 110 (Connection timed out))
08:27:13LinusNi want to release bootloader v4
08:27:27LinusNany objections?
08:29:51B4gdernone from me
08:31:34LinusNno buzz at boot, correct rec+on check, low battery warning, allows usb mode even when ata init fails
08:34:28 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
08:37:11amiconnAnd then 2 clicks at boot instead of one? (There is a click at boot at least on line out, and I believe this is because of the uda reset)
08:37:50amiconnSame thing happens on archos with mas reset...
08:39:21LinusNamiconn: yes, but the second click can be removed in rockbox
08:44:48amiconnLinusN: Could you please check the LCD wait states? I dunno if you calculated them or measured to make sure the transfers are within specs, but my calculations yields the following:
08:45:24amiconn120 MHz: 8 wait states instead of 9 (perhaps a remnant from 140 MHz operation?)
08:45:52amiconn48 MHz: 1 wait state instead of 2.
08:47:27 Quit webguest06 ("CGI:IRC (EOF)")
08:50:29LinusNamiconn: how do you calculate this?
08:51:01LinusNat 120MHz i get 16,673336
08:51:17amiconnWe need to stay above 200 ns cycle time, and min. 60 ns inactive / active time each
08:51:31amiconn120 MHz has 8,33333 ns cpu cycle
08:52:04LinusN = 11,9952 clocks
08:52:22LinusN12 clocks - 3 = 9 wait states
08:52:35amiconnThe loop uses 7 cycles in other instructions that the lcd write, so inactive time is 8.33333 * 7 = 58.33333 ns - inactive time (almost) correct
08:52:40LinusNamiconn: i'm talking about bus clock
08:52:47amiconnYes, wait a mom
08:53:13amiconn200 ns / 8.3333 ns = 24 cpu cycles
08:53:28amiconn24 - 7 (inactive time) = 17
08:53:56amiconn17 cpu clocks -> 9 bus clocks (rouned up!)
08:54:03amiconn-> 8 wait states
08:54:36amiconnSame with 48 MHz
08:54:57amiconncpu cycle is 20.83333 ns
08:55:20amiconn200/20.83333 = 9.6 -> 10
08:55:34amiconn10 -7 (inactive time) = 3
08:55:44amiconn3 cpu clks -> 2 bus clks -> 1 wait
08:56:47LinusNin fact, we can set the wait states much lower
08:57:12amiconnI don't get the -3 in your formula, but if this means the inactive time (in bus clocks and rounded down) it is still off by 1
08:57:30amiconnbecause 9 bus cycles means 8 wait states
08:57:46LinusNa bus cycle is minimum 3 clocks
08:57:57LinusNand the wait state is added to that
08:58:20amiconnThen we can set it even 2 lower than my calculations yield
08:58:39 Join einhirn [0] (
08:58:40amiconn(I assumed a bus cycle is 1 bus clock + wait states)
08:59:26amiconnHmm, perhaps this isn't correct... I forgot the pipelining
08:59:38LinusNmy calculation is based on the 200ns cycle time in the data sheet, but if we include the inactive time in the lcd driver, we can decrease it to the 160ns minimum low-pulse width
09:00:01amiconnminimum low pulse is 60 ns
09:00:14LinusNi mean 60
09:00:18amiconn...and minimum high pulse is 60 ns as well
09:00:54amiconn..but now I'm no longer sure we guarantee that....
09:01:14amiconn...because of the write pipeline
09:01:17LinusNmeaning that we can set it to 1 wait state at 120mhz if we want
09:01:55amiconnIt may be that the write cycles are back-to-back with 9 wait states, completely shadowing the loop
09:02:00LinusNas long as we keep the 200us cycle time
09:03:18LinusNanother reason to bring up the LA ...
09:03:40amiconnyup :)
09:04:10LinusNhow much do you know about the write pipeline?
09:04:38amiconnIf this is the case, the best solution would be to have different write loops, to let the cpu core dictate the timing
09:04:55amiconnThe we colud reduce the wait cycles to 50% low pulse with
09:05:58LinusNi must confess i didn't know there was a write pipeline
09:06:36amiconnI'm not that sure, but it seems write pipelining can completely shadow instruction executing if the write target is slow, like the SDRAM
09:07:20amiconnI think I observed this when writing asm memset(), where some changes didn't affect speed at all
09:08:25***Saving seen data "./dancer.seen"
09:08:54LinusNFor all STORE operations (except MOVEM), certain hardware resources
09:08:54LinusNwithin the processor are marked as ?busy? for two clock cycles after the final DSOC cycle of the
09:08:54LinusNstore instruction. If a subsequent STORE instruction is encountered within this 2-cycle window, it
09:08:54DBUGEnqueued KICK LinusN
09:08:54LinusNwill be stalled until the resource again becomes available. Thus, the maximum pipeline stall
09:08:54LinusNinvolving consecutive STORE operations is 2 cycles. The MOVEM instruction uses a different set
09:08:59LinusNof resources and this stall does not apply.
09:09:12amiconnHmm, it seems there is no write pipeline, but the instruction fetch pipeline and oep may have that effect
09:11:11amiconnNow I think this is really a case for the LA...
09:11:24LinusNso we *could* decrease the lcd waitstates considerably
09:11:38amiconnDoes the write stop execution until the bus cycle is completed, or not?
09:12:18LinusNnot, provided the pipeline has something to do
09:13:34amiconnYour quote means cpu clocks, and we don't have 2 consecutive store operations.
09:15:19LinusNthat's just about everything it says about write pipelines and stalls
09:15:45LinusNso i assume that there is no write pipe
09:17:51amiconnNo, but this is interesting too (from the exception processing description):
09:17:52amiconnThe ColdFire processor uses an imprecise reporting mechanism for access errors on operand writes.
09:17:52amiconnBecause the actual write cycle may be decoupled from the processor’s issuing of the operation, the
09:17:52DBUGEnqueued KICK amiconn
09:17:52amiconnsignaling of an access error appears to be decoupled from the instruction that generated the write.
09:18:08amiconnThe "decoupled" being the key here
09:18:18 Nick Lynx_awy is now known as Lynx_ (
09:19:51amiconnAlso, from the cache's DBWE description:
09:19:52amiconnFor the ColdFire CPU, reporting access errors on operand writes is always imprecise and
09:19:52amiconnenabling buffered writes simply further decouples the write instruction from the signaling of the fault.
09:20:10amiconn*always impresice*
09:21:14amiconnSimilar in the ACR description...
09:21:44LinusNi hate them for not telling how it really works
09:23:14LinusNpage 83 in the coldfire 2 user's manual is interesting
09:24:04LinusNbut it probably only refers to instruction fetches
09:25:10amiconnIf it works like I think, we could lower the waits to get a ~50% duty cycle, but then need a 24-cpu-cycle loop for 120 MHz, and a 10-cycle loop for 48 MHz
09:26:24amiconnThis would also mean that we now obey the cycle length, but not the minimum high pulse width
09:26:27LinusNyes, i think so too
09:27:16amiconnWhen hooking up the LA, you could also check the MFDR(2) values... ;)
09:27:32LinusNyes i will
09:28:02LinusNmaybe put up a wish list for me in the wiki
09:28:12LinusNso i don't forget anything
09:28:23amiconnBtw, we are slightly above LCD specs on archos too
09:29:11amiconnAh, no
09:29:44amiconnForgot the implicit 2 wait states for port access....
09:30:58amiconnWe're right at specs on Ondio. The cycles alternate; 8 cycles - 6 cycles (including the wait states)
09:31:23CtcpIgnored 2 channel CTCP requests in 8 minutes and 58 seconds at the last flood
09:31:23*amiconn is totally confused now :/
09:36:05 Join justmejg [0] (
09:42:14 Join Chamois [0] (
09:52:21justmejgHello anyone who may know the answer to this would make my day... When Creating a .wsp file and using the %ac tag is there anything special that you need to do to make it work?
09:53:40justmejgAlso does that tag work when creating a .wsp file for the Studio series of players
09:56:29LinusNi'm not sure it works for the studio
09:57:05LinusNare you using the very latest daily build?
09:57:41 Join webguest26 [0] (
09:57:43LinusNthen i guess it doesn't work on the studio
09:58:32justmejgOh well, I was also thinking that a %aj Full Justify option would be a good thing do you think that it would be possible?
09:59:04justmejgOn any of the players
10:14:30B4gderthe justify things don't work on player/studio
10:17:22justmejgThats ok I have a studio 10 but am about to get an AV420 will the Rockbox firmaware work on it ?
10:18:20justmejgDumb question I will look it up on my own...
10:19:04 Join Mirfle [0] (
10:22:44 Join cYmen [0] (
10:23:44justmejgOkay so I looked it up on my own... and the answer was no... but would you know if there is any chance that there may be a firmaware for the AV series in the future?
10:27:17justmejgOn another note... what would one have to do to create a %pf style item for the .wps that looks like that but has the track length static on the left and the time remaining counting down on the right?
10:28:08justmejgfor the studio series?
10:29:19 Join ep0ch [0] (
10:31:18justmejgWould that even be possible?
10:41:33 Quit Mirfle ("Chatzilla 0.9.68a [Firefox 1.0.4/20050511]")
10:48:04 Nick justmejg is now known as jpegreen (
10:50:21jpegreenB4gder: I want to write my version of the %pf what would I have to do to make that happen?
10:56:10 Part jpegreen
10:56:40 Join lodesi [0] (
11:08:26***Saving seen data "./dancer.seen"
11:09:51amiconnB4gder: The 'bounce' scroll animation is outdated - no greyscale... it doesn't have the left / right margin corrections
11:11:45amiconnLinusN: The link for the SCF5250 user's manual says SCF5249...
11:15:07LinusNboot loader v4 is released
11:15:09amiconnLinusN: That clock idea floating around in the forums - it seems that it is possible to configure the PLL so that the CPU clock is _lower_ than the base frequency
11:15:46amiconnDo you think this will save power, or would the PLL eat more power than what is saved by lowering the CPU clock?
11:16:14amiconn...compared to the PLL being inactive at 11 MHz
11:17:34LinusNwell, we'll have to try it
11:18:27amiconnWe could switch off pretty much everything - ata power, lcd, and even the sdram controller when running completely from iram
11:19:02LinusNwould be an interesting experiment
11:19:13LinusNbut the clock will drift like hell
11:19:26 Quit ze (Read error: 110 (Connection timed out))
11:19:37amiconnIt should be the same accuracy as the xtal itself
11:19:42LinusNwhat would you base it on? the tick?
11:20:06amiconnI would even lower the tick to once per second
11:20:33LinusNyes, but as soon as you use the player, the tick will drift because of the cpu boosting
11:20:55amiconnAh, you mean while using it - yes it will drift
11:21:13amiconn...but max. 10 ms per boost/unboost
11:21:35LinusNit can be good for relative stuff, like wakeup in 8 hrs
11:21:49LinusNbut it will never be a good rtc replacement
11:22:12ep0chsubtract 5ms each time a boost/unboost happens?
11:22:25ep0chor add
11:22:33LinusNep0ch: the pll lock time is unknown
11:22:50LinusNwe only know that it is never longer than 10ms
11:23:20LinusNif only we had a timer that counted external oscillator cycles
11:24:43amiconnYes, I would want that for the timer stuff I mentioned as well...
11:26:42LinusNjust another of the many coldfire design flaws
11:27:23 Join Lurkski [0] (
11:28:19LinusNwho was working on remote control wps and stuff?
11:28:28LurkskiI'm learning C right now and was hoping someone here could answer a quick question
11:28:52 Join tucoz [0] (
11:29:13tucozHi, looks like the hebrew/arabic patch is updated:
11:29:16amiconnLinusN: Iirc there is a patch, but I think it is the quick-hack type
11:29:31LinusNamiconn: aha
11:29:37LinusNtucoz: yes
11:29:43amiconnMy idea was to use a second ui thread for the remote
11:29:56LurkskiWhen defininf variables, specifically character arrays, must this been done in the main function before any of the rest of the body of the code? For some reason I thought if I defined an array it would be good from that point on.
11:30:11amiconn...but I don't think I will work on that in the near future
11:30:19 Join ze [0] (
11:30:36amiconn(1) Too many other ideas (2) Remote isn't exactly top pri for me
11:31:42LinusNLurkski: C99 allows declaring variabled in the middle of a code block, but it is still common practice to declare them in the top of the function
11:31:56tucozbye bye
11:31:59 Part tucoz
11:32:35LurkskiLinus N: Oh I see, is C99 a branch off of C?
11:32:48LinusNit's the latest version of the C specification
11:34:04Lurkskicool, just wanted to make sure my compiler wasn't on the fritz. Thanks for the help Linus. This project is what inspired me to learn C so thanks in so many ways! Goodnight.
11:36:14 Quit Lurkski ("CGI:IRC (EOF)")
11:45:21lodesiamiconn: i'm trying to build my own overlay plugin
11:45:32lodesiit runs fine on the sim
11:45:45lodesibut i get a black screen o the recorder
11:46:02lodesiwith red led blinking..
11:46:09lodesioverlay like rockboy
11:46:29 Quit Coldtoast (Read error: 104 (Connection reset by peer))
11:46:47lodesii thouht i would need the overlay as i'm usnig bi amount of ram :)
11:47:27lodesibut i don't undertand at all how this thing works :/
11:47:44lodesijust did a copy/paste from rockboy
11:48:34lodesiany idea?
11:48:44 Join hicks [0] (
11:49:14LinusNi don't know how the archos rockboy works
11:52:50LinusNok, i see now
11:53:25LinusNlodesi: what is your problem?
11:53:53LinusNdid you use for linking?
11:54:09lodesii think so
11:54:20lodesii reused the makefile of rockboy
11:55:44LinusNso your plugin just loads the overlay and calls it?
11:56:17lodesibut there is more code i the overlay
11:56:27LinusNhave you checked the .map file?
11:56:57lodesimmh.. do'tthink so, where is it?
11:57:59LinusNthe rockboy map file is in $(builddir)/apps/plugins/rockboy/
11:58:08lodesiit seems linked agaist my plugin
11:59:47LinusNSlasheri: u there?
12:07:20LinusNlodesi: any clues?
12:09:34lodesinot really..
12:10:53lodesido i need to change the .entry/.text/.data ... entries, or should they stay like they are in rockboy?
12:11:49LinusNi think they can remain
12:12:20LinusNlodesi: can you send me a zip of your code?
12:12:52LinusNlinus at haxx dot se
12:13:29 Join Lear [0] (
12:15:14lodesitgz is fine?
12:16:08lodesiwell.. sent a tgz :)
12:18:17 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
12:24:59lodesiyou will need this ->
12:25:11lodesiif you want to try it in the simulator
12:25:35lodesijust untgz it in the root folder of the archos-sim
12:26:01lodesigoing to lunch
12:26:04*lodesi is away: OccupƩ
12:31:27 Join Coldtoast [0] (
12:32:21 Join austriancoder [0] (~austrianc@
12:32:38SlasheriLinusN: hi, now i am :)
12:32:47LinusNSlasheri: hi
12:33:07LinusNi was investigating a problem with the playback code
12:33:25LinusNit doesn't skip bad files
12:33:47Slasheriah, yes.. that should be fixed
12:33:55LinusNplay a wav file with a bad header, and it will retry forever
12:34:01Slasherii think load_track function could do the skipping if it finds the metadata bad
12:34:36LinusNi'm not sure what to do with last_peek_offset
12:37:47SlasheriHmm, it should contain index to the next unbuffered track
12:38:17SlasheriThat index is for example decreased when we call playlist_next(1)
12:39:00SlasheriI think you should also call playlist_next(1) and decrease the pointer if it was increased
12:39:21Slasheri(i cannot check that before i get to home)
12:39:38LinusNi'll try it
12:41:37 Nick austriancoder is now known as ac_away (~austrianc@
12:43:51 Quit Lear (Read error: 113 (No route to host))
12:47:07 Join memmem [0] (
12:48:44memmemHi austriancoder, do you have a BDM?
12:50:35LinusNlodesi: i don't see any immediate errors in the target build, but the sim build didn't compile (cygwin)
12:50:49LinusNi haven't run it on target yet
12:50:56LinusNlunch time
13:08:30***Saving seen data "./dancer.seen"
13:13:34 Quit edx (Read error: 110 (Connection timed out))
13:14:06 Join matsl [0] (
13:20:43 Join edx [0] (
13:22:53 Quit mdeboer (Remote closed the connection)
13:24:00 Join einhirn [0] (
13:25:14 Quit B4gder (
13:25:55NJoinB4gder [0] (
13:37:28ac_awaymemmem: hi
13:37:45 Join CoCoLUS [0] (
13:41:54 Join Nuxator [0] (
13:42:33Nuxatoris linus here?
13:42:56 Quit ac_away ("using sirc version 2.211+KSIRC/1.3.12")
13:43:06 Join austriancoder [0] (~austrianc@
13:43:20memmemHi austriancoder, do you have a BDM?
13:44:12austriancodermemmem: look at private chat with me ;)
13:47:42LinusNNuxator: i'm here
13:47:56Nuxatori've got a question
13:48:16Nuxatoris it possible to add a key to force usb disconnect?
13:48:39NuxatorStill my usb disconnect problem
13:49:09LinusNreresh my memory
13:49:18LinusNit doesn't notice when you unplug?
13:49:22Nuxatormy ihp140 doen't detect that i unplug it (in rockbox no prob in iriver)
13:49:34LinusNonly in rockbox?
13:49:44Nuxatorand more strange
13:49:45LinusNthen we must find the bug
13:50:00Nuxatorif i switch off my computer with ihp plugged
13:50:16Nuxatorit detec usb disconnet when pc switch off
13:50:41Nuxatorhmm will be hard since it semms that i'm the only one having it
13:51:06Nuxatorbut will do anithing you suggest to find out why
13:51:57LinusNenter the view i/o debug menu
13:52:50LinusNcheck the GPIO1_READ value
13:53:23Nuxatorwhen i plug it?
13:53:33LinusNyes, with and without
13:54:06Nuxatorwithout 079B7660
13:54:44Nuxatorwith 079B76E0
13:55:32LinusNlooks ok
13:57:00Nuxatori puged i unpuged several times and it seems to work (changing from 60 to E0 and back to 60)
13:57:13Nuxator+l +l
13:57:30LinusNso there must be something happening when we activate the usb chip
13:58:08Nuxatorharder to debug
13:58:44LinusNi guess i should find out more about how the usb chip is connected
13:58:56Nuxatorit occurs in both rockox and bootloader usb mode
13:59:02LinusNmaybe the gpio1_read method isn't reliable
13:59:22Nuxatorseems to be ok since value change
13:59:46LinusNyes, but it might be different when the usb chip is powered
14:00:07Nuxatormaybe a good test would be to add a key pressed detection where gpio1_read is to see if it's gpio prob or another problem
14:00:50LinusNNuxator: do you have a dev environment?
14:01:09LinusNso you can build your own rockbox?
14:01:16Nuxatori tried
14:01:29Nuxatorbut prefer downloading bleeding edge
14:02:24Nuxatorbut i think i can build it
14:04:03Nuxatorwhat do you suggest?
14:05:21LinusNoutput the gpio1_read value in the usb screen
14:05:37LinusNscreens.c, function usb_screen()
14:05:49Nuxatorok i'll try that
14:05:58LinusNright after while(usb_wait_for_disconnect_w_tmo(&button_queue, HZ)) {
14:06:16Nuxatorbut i tried to read rockbox code and i don't understand it
14:06:16LinusN char buf[32];
14:06:16LinusN snprintf(buf, 32, "%08x", GPIO1_READ);
14:06:24LinusN lcd_puts(5, 1, buf);
14:06:24LinusN lcd_update();
14:06:25Nuxatorok thanks
14:08:03LinusNi can build it for you if you want
14:08:06 Quit einhirn (Success)
14:08:23Nuxatori think it would be faster and foolproof
14:08:30LinusNhang on
14:12:10*austriancoder will go now
14:13:07 Quit austriancoder ("using sirc version 2.211+KSIRC/1.3.12")
14:13:49 Quit memmem ("ERC Version 5.0.3 $Revision: 1.726.2.17 $ (IRC client for Emacs)")
14:14:21Nuxatorok E0 when i plug
14:14:35Nuxatorstay at E0 whan unplugged
14:16:56LinusNok, so now we know that
14:17:51Nuxatorreally strange bug
14:18:03Nuxatormaybe an hardware bug on some few h140
14:18:17Nuxatorbut i wonder how iriver manage usb
14:18:24LinusNthat's what i want to know
14:18:34LinusNit is possible that rockbox is doing it wrong
14:18:50Nuxatorbut why i'm the only one to complain....
14:19:14Nuxatori think it shoulb be something hardware specific
14:19:19LinusNyes, but the iriver fw handles it well
14:19:28LinusNso we should do it too
14:23:48Nuxatorok , i think i have to give you an email so you can mail me if you have anything to test for this bug.
14:24:09Nuxatordo you want my email?
14:24:16LinusNemail me: linus at haxx dot se
14:24:19LinusNi have to go
14:24:22LinusNcu later
14:24:30Nuxatorok thanks for your great work
14:24:35LinusNyou're welcome
14:24:36 Part LinusN
14:28:43 Join RotAtoR [0] (
14:29:23 Quit webguest26 ("CGI:IRC")
14:30:25 Quit Nuxator ("(have some work waiting)")
14:31:09 Join elinenbe [0] (~elinenbe_@
14:31:37elinenbeblah, blah blah. blah... blah... blah. blah!
14:39:43CassandraHmm. I seem to have buggered my iRiver's internal mike while upgrading the battery.
14:40:20CassandraOne of the wires came off, and I resoldered it, but the soldering job is very messy and connects the 5v wire to the outer casing of the mike.
14:40:36CassandraIs that likely to be the cause, or have I blown it up?
14:44:08 Quit ep0ch (" Want to be different? HydraIRC -> <-")
14:45:12 Join eleom1 [0] (
14:45:53 Quit eleom1 (Client Quit)
14:56:05 Join Sucka [0] (
15:03:16 Quit B4gder ("go go go")
15:08:31***Saving seen data "./dancer.seen"
15:12:00 Join bumi [0] (
15:12:29bumihey everyone
15:12:46bumican anyone here try compiling a patch in the firmware for me?
15:13:09bumithis one
15:13:16bumifor the iriver ihp120
15:27:01crwlshouldn't be difficult
15:28:40bumihey crwl
15:28:44bumican you please do it for me?
15:28:51bumii dont have the development kit
15:28:54crwlok, i patched it against today's cvs... at least it compiled
15:29:01crwli don't have the time to check it, but i can send it to you
15:29:01bumiand me and command promped go together
15:29:20crwldcc send?
15:29:22bumithink it will do damage to my player?
15:29:26bumiDCC send
15:29:46bumiis it allready in with the firmware?
15:29:49bumii mean
15:29:54bumiits a .hex file?
15:30:02bumibecuse as i said i dont have a SDK
15:30:11bumiso i cant combine it with the firmware
15:30:26bumiand i dont the the boot flusher will let me switch to this firmware
15:35:06 Join mdeboer [0] (
15:36:37 Join austriancoder [0] (~austrianc@
15:37:09bumihey crwl you still here?
15:37:48bumianyone else here wanna help me out?
15:39:15*austriancoder has no time
15:42:58*HCl watches the first season of airwolf
15:44:46 Join crwl_ [0] (
15:46:01 Join memmem [0] (
15:46:05 Quit crwl (Nick collision from services.)
15:46:09 Nick crwl_ is now known as crwl (
15:46:21 Quit memmem (Client Quit)
15:47:25 Join memmem [0] (
15:59:53CassandraHmm. On the plus side, 27h 30m est. battery life.
16:00:06CassandraOn the minus side, I destroyed the internal mic.
16:02:06ashridahwait. how did you fuck up the internal mic? isn't the jack for the battery completely on the other end of the player?
16:02:39austriancodera wiki guru here?
16:02:40CassandraYes, but I broke the wire connections while doing the maintenance.
16:02:40ashridah(you probably aren't going to miss it, btw, since it gets noise from the hard drive in some cases, and the player comes with an external mic anyway :)
16:03:06Cassandraashridah: I know. Still a little annoying.
16:03:17CassandraI resoldered it, eventually, but it's still not working.
16:03:26CassandraI think the heat from the iron f***ed the mic.
16:03:37CassandraOr my soldering was sucky.
16:03:47CassandraAbout equally likely, really.
16:04:22CassandraPartially sighted, poor hand-eye co-ordination and soldering do not go well together.
16:04:35CassandraPlus, being a girl, I'm genetically incapable of soldering. :)
16:11:34ashridahnonsense. i got outsoldered by several of the girls doing engineering with me
16:12:07CassandraWas being tongue in cheek.
16:12:31CassandraMy sister used to run an electronics company.
16:12:39CassandraI'm actually the less geeky one. :)
16:15:25 Join Mirfle [0] (
16:16:04CassandraI really can't solder very well though.
16:16:21CassandraAlthough part of the problem is that I only have a crappy iron.
16:19:38ashridahheh. i used to use a gas powered one.
16:19:51ashridahmy friends used to get very worried by that :)
16:22:09zeashridah: heh, you mean butane right?
16:22:36ashridahyeah, whatever. it burned, the iron got hot
16:22:45ashridahand my friends wouldn't go near it. pansies :)
16:22:51zeheh why not
16:22:57zethey go near lighters don't they?
16:23:06zethey burn butane too :p
16:23:16ashridahactually, none of them smoked.
16:23:27ashridahbut lighters don't usually burn constantly
16:23:32zenor lit candles or incense or fires?
16:23:47zetrue.. they say not to keep them lit for >30 seconds
16:23:50ashridahwe covered the 'pansy' bit, didn't we?
16:24:06zei had to use a bic for light once
16:24:18zeit melted
16:24:29zejust around the top
16:24:29ashridahyeah, they're not all that durable
16:24:39zeenough for that metal part to slide all a funky way
16:24:43zeand make the whole thing inoperable
16:24:45 Join DomZ [0] (
16:25:27*ze has recently, finally, got a photon microlight :p
16:26:29ashridahalthough i once found a really brilliant lighter. the kind that burns against a grill, so the grill heats up and that lights ciggarettes, instead of the flame doing it.
16:26:48ashridahpity the refill valve at the bottom never worked
16:26:56ashridahthat thing was great for heatshrink
16:30:52zei'm still trying to find or work out how to invent a nice portable source of temperature-controlled hot air
16:31:03zefor, uh... various purposes ;p
16:41:11 Join cYmen_ [0] (
16:41:15 Quit cYmen_ (Client Quit)
16:43:49 Quit ashridah ("sleep")
16:50:32 Quit bumi ("( :: NoNameScript 3.81 :: )")
16:51:27 Quit DomZ ("CGI:IRC")
17:01:05 Join xen` [0] (
17:08:35***Saving seen data "./dancer.seen"
17:12:47 Part Mirfle
17:13:07 Quit bill2or3 (Remote closed the connection)
17:18:14*lodesi is back (gone 04:52:08)
17:23:23 Quit memmem ("ERC Version 5.0.3 $Revision: 1.726.2.17 $ (IRC client for Emacs)")
17:39:05 Join Naked6 [0] (
17:39:26Naked6has anyone here tried rockbox on thier iriver
17:41:29crwlno, never!!
17:41:47crwlmany of us have
17:42:25Naked6okay, well i was just making sure its not at a point where it might screw up my iriver
17:44:01crwlno, it's probably not doing that, unless you screw up the bootloader installation process (which you will only have to do once)
17:46:05Naked6should I install the h-120 version if I have the h-140
17:47:11 Join jamesshuang [0] (
17:47:27jamesshuangok... getting a bit annoyed here, but anyways... how do I get a plugin to compile?
17:47:43jamesshuangwhat file defines what plugins compile for each platform?
17:50:50crwlNaked6, they're the same
17:51:49Naked6thanks crwl, im gonna try my luck installing it.
17:53:53jamesshuangthe rockbox people should really do a better job of documenting their API... entering programmers like me are completely LOST in a jungle of directories, files, and configs
17:54:23jamesshuangif I can't even get a simple plugin to compile, how am I going to learn enough to do bigger things, like actually make valid contributions to rockbox?
17:58:04 Quit lodesi (Read error: 113 (No route to host))
17:59:33Naked6crwl, when patching the firmware i tried to patch the newest 1.6 for the h140. the patcher did not recognise this firmware.
17:59:41Naked6what do you suggest I do?
17:59:48CassandraThe newest is 1.65.
17:59:58CassandraDownload it from the Korean site, not the US one.
18:00:18Naked6thankyou cassandra
18:00:44CassandraIt's linked from the main page of the wiki.
18:01:46Naked6crap... which symbol is korea
18:02:19Naked6nevermind, its the left one
18:03:26Naked6it is very difficult for me to navigate this korean site
18:03:40CassandraTry the European one then.
18:04:05CassandraI believe the firmwares are also linked off the Rockbox wiki page on flashing the iRiver.
18:04:49jamesshuangCassandra: I know WHICH ones are compiled, but let's say I programmed a new one... how do I get this new one to compile?
18:05:11 Join asdsd____ [0] (
18:05:37amiconnjamesshuang: apps/plugins/SOURCES
18:05:45DBUGEnqueued KICK asdsd____
18:05:45asdsd____1010,00hey guys
18:05:57amiconnThis uses the gcc preprocessor to decide about models
18:06:23asdsd____1010,00hey amiconn
18:06:32***Alert Mode level 1
18:06:32asdsd____1010,00whats the status of the iriver's firmware?
18:07:08jamesshuangTHANK YOU amiconn!
18:07:19 Join lodesi [0] (
18:07:22jamesshuangexactly what I was looking for :)
18:07:39jamesshuanglittle things like this should probably be documented somewhere...
18:08:17Naked6i think they record everything that happens in this room... it would just be an ass load of information to sift through
18:08:46jamesshuangand weird little quirks, like why is the iriver stuff in the "recorder" folder?
18:09:14CassandraWell there's a wiki page, RockboxArchitecture. You could probably add it there.
18:09:43amiconnjamesshuang: The recorder folder is for historical reasons.
18:10:00amiconnThe first 2 devices rockbox ran on were the archos player, and the archos recorder
18:10:19amiconnThe former has a charcell based display while the latter has a bitmap based display
18:10:35amiconnSo 'recorder' is often equal to 'has bitmap display'
18:11:17amiconnIt's not trivial to move files in cvs. You will lose revision history, or at least make it non-obvious
18:11:21Naked6okay so I upgraded my firmware with the patched hex... it turned off.
18:11:30***Alert Mode level 2
18:11:30asdsd____1010,00hey guys hows the work coming for the h3xx?
18:11:35Naked6I still need to get rockbox on it... should i try and do that now?
18:11:47***Alert Mode level 3
18:11:52amiconnasdsd____: Pls configure your client not to use colour codes.
18:12:03Naked6oh sweet its working
18:12:07***Alert Mode level 4
18:12:07asdsd____1010,00i can't! trillian doesn't let me
18:12:09jamesshuangamiconn : exactly why we probably need better documentation on this stuff :)
18:12:09Naked6its magical
18:12:20Cassandraamicon: I find it wise to turn them off in my client.
18:13:01amiconnCassandra: Yes, I could probably do that, but they still clutter the log
18:13:07HClso don't use trillian for irc.
18:13:11HClget a decent client
18:13:19***Alert Mode level 5
18:13:29CassandraWell, people could spend time documenting how to develop for Rockbox.
18:13:39CassandraOr they could get on with developing it.
18:13:42***Alert Mode level 6
18:13:42asdsd____1010,00u learned how to overirde other people's color codes because of me
18:13:54***Alert Mode level 7
18:13:54asdsd____1010,00u and preglow
18:13:55thegeekyou bastard asdsd____d
18:13:58CassandraWhy yes, yes I did.
18:14:11***Alert Mode level 8
18:14:11asdsd____1010,00why do u gotta call me a bastard?
18:14:16HCli wish freenode had +c or +C or whatever it is :/
18:14:33thegeekit seems quite obvious
18:14:44HClit does have +c
18:14:52HCldon't know how i do color in irssi to test though..
18:15:03***Alert Mode level 9
18:15:03asdsd____1010,00anyways i just came to check the status of rockbox on the h3xx, any progress yet?
18:15:14HCland there won't be for a long time
18:15:18CassandraWell, I believe he's spewing it.
18:15:18HCli wish people would stop asking..
18:15:32***Alert Mode level 10
18:15:32asdsd____1010,00just think of people asking as ur motivational factor
18:15:37***Alert Mode level 11
18:15:37asdsd____1010,00anyways im off
18:15:39***Alert Mode level 12
18:15:39asdsd____1010,00c'ya guys
18:15:41 Part asdsd____
18:15:45CassandraMaybe we should get Linus to change the topic to "No, we're not working on H3xx Rockbox."
18:15:46HClpeople using colored text in irc
18:15:57HClis only making me more unmotivated to put effort in h3x0
18:16:10HClset the channel +c
18:16:14HClso they can't use colors
18:16:23 Nick Lynx_ is now known as Lynx_awy (
18:16:27amiconnI just set 'strip codes from server windows' in HydraIRC to true
18:17:40CassandraI set all the strip codes options to true.
18:17:50HClbah, +c doesn't work..
18:18:01CassandraDoesn't seem to have affected the channel monitor, unfortunately.
18:18:39jamesshuangx_x I suck at C programming
18:18:47jamesshuangI can't get this code transplant to work...
18:18:50 Join BBub [0] (
18:19:04jamesshuangwhy is my plugin telling me it's an incompatible version all of the sudden? so odd...
18:19:25austriancodera coding question
18:19:26HClprobably because you didn't recompile properly?
18:19:32austriancoderi have an char 0xffffffb6
18:19:41austriancoderhow can i access only the b6?
18:19:52HClhow can that be a char?
18:20:14austriancoderoh.. i mean only 0fffffb6
18:20:18austriancoder ffffffb6
18:20:27HCla char in memory is only 1 byte.
18:20:36HClif you want the lower 8 bits you and it with 0xFF
18:21:10austriancoderprintf("Old checksum: 0x%x\n", byte);
18:21:21austriancoderHCl: what should i do?
18:21:40amiconnaustriancoder: byte is defined as char byte; ?
18:21:45HClprintf("Old checksum: 0x%x\n", (unsigned char) byte);
18:22:15amiconnbetter still: printf("Old checksum: 0x%02x\n", (unsigned char) byte);
18:23:27austriancoderaustriancoder@neptun ~/rockbox/rockbox-devel/tools $ ./iaudio x5_fw.bin x5_fb_new.bin
18:23:27austriancoderOld checksum: 0xb6
18:23:27austriancoderNew checksum: 0x0
18:23:54austriancodernow i must add the calculation of the new checksum
18:25:40***Alert Mode OFF
18:26:50Naked6rockbox is not as loud as the normal iriver firmware... or is that just mine?
18:27:24CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
18:27:24*austriancoder is happy as the checksum calculations seems to be correct ;)
18:28:14jamesshuangmmm... finally getting some info from devs... thank you all!
18:28:37jamesshuangI can't get it to work, but meh.... I suck at C programming
18:29:40Naked6sweet makes prefect sence
18:29:45 Quit jamesshuang ("CGI:IRC")
18:30:39Naked6so... that rockbox thing i try to load.. thats not supposed to work yet right?
18:32:51austriancoderwith which tool can i compare two binary files?
18:33:01 Join Nibbler [0] (
18:34:51 Join webguest24 [0] (
18:35:19webguest24cmp - compare two files
18:35:34webguest24diff - compare two files
18:35:45amiconncmp is for binary
18:35:53 Part webguest24
18:36:42austriancoderthanks guys...
18:36:57austriancoderi will do the first iAudio x5 related commit now ;)
18:37:53HCl :p
18:49:43 Quit lodesi ("Leaving")
19:05:36Slasheribtw, cvsdo is the magic command to create patches without write access to cvs repository. Just cvsdo add somefile and cvsdo diff should work :)
19:05:37 Quit Naked6 (Read error: 104 (Connection reset by peer))
19:05:51 Join XavierGr [0] (
19:06:34XavierGrHi all!
19:06:39Slasherihi XavierGr :)
19:06:48 Join hardeep [0] (hardeeps@norge.freeshell.ORG)
19:07:28XavierGris amiconn here?
19:08:00Slasherimaybe, ~30min. idle :)
19:08:38***Saving seen data "./dancer.seen"
19:09:23XavierGrI am impatient about remote development and I wanted to discuss some things with him. I want a little direction as to how I can make another thread for the remote
19:10:03XavierGrthe hack I utilize for the remote is hard to keep up for the updates and has some major flaws
19:10:45XavierGralso I could make the settings for the remote scrolling but again I am lost when I stare the code for the settings
19:12:07 Join spiralout [0] (
19:14:37austriancoderhow is the guy, who adds new devices to rockbox build system?
19:15:32hardeepaustriancoder: Bagder, LinusN or Zagor
19:16:19austriancoderah ok.. so i will wait for one of them
19:18:31XavierGrSlasheri: Do you know why in the simulator whenever I push to play a song the wps doesnt appear? (I dont have install libmad)
19:18:40 Quit matsl (Read error: 110 (Connection timed out))
19:19:05SlasheriXavierGr: Hmm, have you installed the codecs?
19:19:17SlasheriAnd can you do cat /dev/urandom >/dev/dsp as normal user?
19:20:42Slasheriat least you have to copy the codecs to use the wps
19:20:58Slasheri(just make zip && cd archos && unzip ../
19:24:28XavierGrI did a normal compole of the simulator
19:24:47XavierGrmake install after the compile
19:27:02XavierGrwhen I move to another menu I get codec failure
19:27:29XavierGrwhere can I put the codecs?
19:31:19XavierGrI just moved the codecs from a traget daily build to the simulator codecs folder but this crashed the uisimulator.
19:31:45XavierGr(I though so) how can I compile the codecs without libmad?
19:34:54SlasheriYou can't do that, you should use the codecs made for the simulator target
19:35:03Slasheridon't use make install but make zip
19:35:20Slasheriafter that just unzip in the archos directory and everything should work
19:35:43SlasheriHmm, why don't you want to compile libmad?
19:36:08XavierGrcurrently I dont have it.
19:36:16XavierGrcan I make the wps work with out it?
19:36:22Slasheriif you have rockbox sources, you have it..
19:36:38Slasherino you can't
19:36:58XavierGroh then I will downlaod libamad from cygwin
19:37:31Slasherijust cvs co rockbox-devel and you have everything you need
19:37:39Slasherimkdir sim
19:37:40XavierGrjust I remember that older version (very old) workd without libamad
19:37:46Slasheri../tools/configure and select simulator
19:38:04Slasherithen make && make zip && mkdir archos && cd archos && unzip ../
19:38:11Slashericd .. && ./rockboxui
19:38:12Slasherithat's it
19:38:44XavierGrI have already the latest source of rockbox and I have typed ../tools/configure
19:38:54XavierGrI will try the latest line you say
19:39:32Slasheriyou should use the make && make zip and extract the zip in archos directory
19:40:19XavierGrI thought that make install did that automatically
19:40:27Slasherimake install does nothing i think
19:40:34Slasheribut i am not sure
19:40:36Slasherijust do it manually
19:40:54XavierGrwell it does install .rockbox directory for me
19:41:18Slasheriah yes, it should do the right thing..
19:41:40XavierGrso I need the libmad sources right
19:41:59Slasheriyou have the sources because you have the rockbox sources
19:42:07Slasheriyou don't need anything
19:42:21XavierGrbecause I remember that in the compile procedure of the simulator somewhere it says "libmad not found"
19:42:50Slasheribe sure to check that you have the most recent cvs snapshot, cvs update on root dir
19:43:04Slasherijust rm -rf the simulator build dir and do it again
19:43:49Slasherior better delete the whole rockbox directory, and check out rockbox-devel again
19:47:24XavierGrI do that every time. Whenever I new daily build comes out I remove all the contents of the rockbox directory. Then copy them there, make the tools
19:47:29XavierGrmake a dir sim
19:47:35XavierGrand then make the sim
19:48:17XavierGrI use gcc 3.3.1
19:48:25Slasherihmm, weird..
19:49:25Slasheriplease paste the exact error message you get while compiling
19:50:09XavierGrI will make a txt file with the context of the compilation procedure
19:50:50XavierGrthe compile procedure ends succesfully it is just that I can see somewhere "libmad not found"
19:51:35XavierGrand I get a lot of "basename: not found"
19:52:16 Join solex_ [0] (
19:53:37 Join MrStatic1oid [0] (
19:53:39XavierGragain no codecs in the .rockbox directory
19:54:04XavierGrthe strange thing is that I can compile fine for the targets
19:54:50 Quit MrStaticVoid (Read error: 110 (Connection timed out))
19:56:23 Join Stryke` [0] (
20:03:50 Quit solex (Connection timed out)
20:04:36XavierGrhow can I add a .c file to be compiled?
20:05:26XavierGrlets say I want to compile a remote-tree.c file, All I have to do is add it to the apps/makefile
20:06:19austriancoderadd it to SOURCES file
20:07:26austriancodernothing more.. leave the makefile untouched
20:07:34XavierGrok thanks! :)
20:09:37 Quit alxcm ("LAN PARTY YAY")
20:09:44XavierGrcan I include a .c file?
20:12:27austriancoderwhat want you to do?
20:13:25XavierGroh god I am so ignorant...well I want to do a test .c file to run from the simulator
20:14:14XavierGrI want to call the lcd_remote_clear_display();
20:14:35XavierGrto see how I can manipulate the remote from a different .c file
20:14:45 Join wacky_ [0] (
20:14:51wacky_hey Daniel, what's the format for ~/.curlrc ?
20:15:16XavierGrthat way I will copy the code from tree.c and change the lcd_ part to remote_lcd to make the remote work.
20:15:36austriancodershould work
20:16:13XavierGronly problem is...
20:16:33XavierGrI have a function called test in remote-tree.c
20:16:44wacky_you guys working on the remote support ?!
20:16:51wacky_I'd love you if you would !
20:17:04XavierGrI want to call this function from the original tree.c file. Do I need a header for that?
20:17:17XavierGrwacky_: I am trying bu I am a real program newbie
20:17:19austriancoderyes you need a
20:17:24austriancoderheader file
20:17:37XavierGra remote-tree.h file?
20:18:18austriancoderfor instance
20:19:16austriancoderi am away now
20:19:29XavierGrok thanks anyway!
20:22:55 Quit Maxime ()
20:23:08 Join Maxime [0] (
20:23:13 Join Lear [0] (
20:27:36 Join bazz [0] (
20:28:17amiconnSlasheri: 'make install' for the simulator does the exact same thing as your 'make zip' & unzip to <simdir>/archos/
20:28:45bazzi'm trying to compile the simulator from cvs and i'm getting the following build error: plugin.c:122: error: `lcd_framebuffer' undeclared here (not in a function) i'm poking about myself but i don't know the codebase at all and i figure someone here can probably fix it much faster
20:29:23amiconnCurrent cvs does compile cleanly here
20:29:43amiconn...for all targets and sims I'm usually building
20:31:00Slasheriamiconn: yes, i noticed that when i looked the Makefile :)
20:31:13XavierGramiconn: what exactly do you mean by making a remote thread. Do you mean to make a different .c file for the remote that it will be called whenever we want to use the remote?
20:32:01amiconnNo, I mean a separate thread. I think several source files are needed, each containing the remote-lcd equivalent of a main function remote-tree.c, remote-wps-display.c etc
20:32:48XavierGryes thats what I though, said it wrong
20:34:31amiconnHmm, sound support is *almost* working on cygwin-x11
20:34:46amiconnIt plays the first snippet (about a second), then stops
20:35:49amiconnThe only difference between cygwin and linux is that the header is sys/soundcard.h, not linux/soundcard.h
20:37:06bazzamiconn: what version of gcc are you using?
20:37:21amiconngcc 3.4.4 (on cygwin)
20:38:00XavierGrI use 3.3.1 is that wrong?
20:38:09amiconnI'll try my debain vm now, mostly to see whether sound is working there
20:38:39bazzhrmm, i'm using debian and compiling with 3.3.5 and getting that error
20:38:41amiconnI don't know. Cygwin used to use 3.3.1 for a long time, and it always worked
20:39:29amiconnI'll check what gcc my debian install uses (only use it from time to time, so I don't remember)
20:39:36bazzcool, thanks
20:39:55amiconnHappens to be 3.3.5
20:40:14bazzheh, and does it compile for you?
20:40:58amiconnupdating cvs....
20:41:43bazzactually, looks like the real error is: plugin.c:122: error: initializer element is not constant
20:42:18amiconnmake running....
20:43:49amiconn....built successfully, without any warning
20:44:06XavierGrcya later
20:44:07amiconn(h120 sim of course, including sound)
20:44:08bazzdang, i wonder what's up with my system
20:44:13 Quit XavierGr ("CGI:IRC 0.5.4 (2004/01/29)")
20:44:31bazzhrmm.. i'm trying the h340 sim, wonder if that makes a difference, lemme see if i can build the h120
20:45:20amiconnh340 won't work
20:45:43amiconnSound support doesn't werk in my debian vm :(
20:45:45bazzwell, i know it's incomplete, but should it compile?
20:46:12amiconnSee : H300 sim being RED....
20:46:40LearHmmm.. I try to run cvs update, but get no response. Login went fine, so what gives?
20:46:49bazzahh okay
20:46:56amiconnOf course lcd_framebuffer undeclared... can't be because we don't know the hw lcd format yet
20:46:57bazzyeah, i just build the 120 sim no problem
20:47:13bazzerr built that is
20:47:56amiconnI'd love to hack away on the lcd driver (would need to get hold of a H3x0 first), but that requires to find the lcd driver chip type, and of course a bootloader
20:48:10amiconnI can't do that myself b/c I don't have a bdm...
20:48:44LearIt sends load of data; wonder why? Didn't use to be this slow...
20:49:15bazzyeah, i'm interested in helping out, someone suggested playing with the sim first, do you think the 120 sim is a decent starting point?
20:52:40amiconnMaybe the sim is a good starting point, especially since (iiuc) you don't seem to have a target running rockbox today.
20:53:29amiconn...although I started totally different, with some target-only work... but then rockbox was working on my target device
20:53:48amiconnArchos recorder v1... and I improved some asm routines.
20:54:31Learamiconn: sound like me, only I had a player... :)
20:54:48bazzthat's cool... guess i'll start digging though the sim code then :) any suggestions on where would be a good place to start / things that need to be worked on that i might be able to do not knowing the codebase?
20:55:41 Quit Chamois (" HydraIRC -> <- 100,000+ downloads can't be wrong")
20:56:36amiconnLear: Those were the days where rockbox ran on 4 platforms only....
20:57:35LearAnd during those days, Rockbox actually had malloc... :)
20:58:13amiconnThen you started earlier than me
21:00:27amiconnActually I like rockbox not having malloc()
21:02:40Learslasheri: new dsp.c seems to be working, but I won't check it in today (need to clean up changes to playback.c)
21:08:40***Saving seen data "./dancer.seen"
21:13:02 Quit Rori (Read error: 110 (Connection timed out))
21:19:03 Quit wacky_ ("changing servers")
21:19:16SlasheriLear: nice, what kind of changes you have done to the dsp? :)
21:22:40LearTried to simplify the code, made more use of iram (makes it noticably faster). The flow is pretty much the same (though there are API differences).
21:23:38 Quit MrStatic1oid ("leaving")
21:24:14 Join Rori [0] (
21:25:52SlasheriLear: good :)
21:29:16Learslasheri: btw, when calling the resampling code, the current dsp.c calls upsample if the following is true: "(dsp.frequency > NATIVE_FREQUENCY)". Is that really correct?
21:30:23SlasheriHmm, it should be because we need to upsample when the song frequency is higher than 44.1k
21:31:39Learwell, that sound like downsampling if you ask me (i.e., down to fewer samples per second)... :)
21:32:49LearAlso, upsample starts with some code that could generate several samples, compared to downsample, that starts with code that procudes one sample (before entering the main loop).
21:34:07Slasheriinteresting.. you should try switching the functions =)
21:34:59LearI did that, but it isn't easy to hear any difference. Both MP3 and Ogg sounds good when using switched functions though.
21:35:48Slasherioh, if it works i think that switch was correct
21:43:40 Join Febs [0] (
21:48:02 Join LinusN [0] (
21:50:32austriancoderhi linus
21:51:28austriancodershall i make a mcf5250.h or can i use the mcf5249.h for iaudio?
21:52:17austriancodershould we add an CPU_COLDFIRE flag to use emac and those cpus e.g. for codec optim.
21:54:06LinusNi think you could make a new file, mcf5250.h
21:54:41LinusNit could include mcf5249.h and just add/change the stuff that differs
21:55:12LinusNaustriancoder: CPU_COLDFIRE sounds like a good idea
21:56:07amiconnHmm, what package(s) do I need to install to enable sound support in debian??
21:56:45austriancoderLinusN: could you manage the CPU_COLDFIRE thing?
21:57:11 Join Moos [0] (
21:57:18bazzamiconn: i think alsa-base will give you most of what you need
21:57:43LinusNaustriancoder: i can set up the include file stuff, but not today
21:58:25austriancoderi am not in a hurry
22:00:30austriancoderin the scf5220 datasheet on page 40 there is a pin table.. for instance, pin 42 - IDE-DIOW/GPIO32 - I/O - IDE DIOW ...
22:00:52austriancodercan i asume now that the hdd stuff is running over GPIO32?
22:02:04LinusNi think you can assume that it works just like in iriver
22:02:14 Quit RotAtoR ()
22:02:44amiconnThis is strange... the x11 sims starts to play and plays about a second....
22:02:53LinusNthe mcf52xx has tha same ide controller
22:03:00amiconnsim_sound_play() is called about 5 times
22:03:18amiconnThen the .mp3 is opened again, and this very moment playback stops
22:04:10amiconnIt lists the id3v2 contents again, and sim_sound_play() is no more called
22:04:54austriancoderLinusN: cool.. that means that ide support is done for iaudio - only need to test it...
22:05:49austriancodercan somebody add the iAudio X5 to the build system? I dont have the bdm yet.. but i hope that i comes every day
22:07:48LinusNaustriancoder: bagder is your man
22:08:55austriancoderLinusN: merci
22:10:15LinusNaustriancoder: do you have an idea about which lcd it has?
22:10:50austriancoderLinusN: at the moment not.. i hope i will find a little print or something
22:11:32*HCl watches more airwolf
22:11:36HClairwolf rocks
22:11:37HCl :p
22:12:14austriancoderLinusN: memem told me that the iAudio M3 seems to be in some points same as the M3.. so maybe we get the M3 also working.. memem is living near by me, so yeah
22:13:06LinusNwell, it has a 5249, so...
22:14:09austriancoderthe firmwares of m3 and x5 are in many points equal.. also the checksum thing is equal
22:14:20LinusNnot surprising
22:15:06austriancoderso the starts are quite good to get also m3 working
22:19:04austriancoderLinusN: so you are doing the CPU_COLDFIRE thing and makeing a mfc5220.h file the next time?
22:26:43 Quit hardeep ("BitchX-1.0c19 -- just do it.")
22:27:59 Quit Maxime (Read error: 104 (Connection reset by peer))
22:29:30 Join Maxime [0] (
22:46:51 Quit bazz (Read error: 110 (Connection timed out))
22:47:17Bagderso what is needed for an x5 build?
22:52:06austriancoderwhat do mean?
22:52:25Bagderin what way does it differ from say a h120 build
22:53:01*amiconn spots Bagder
22:53:11HClprobably all the different chips?
22:53:23Bagderso how would I add it to the build system?
22:53:27HCland lcd..
22:53:29HCl :x
22:53:45amiconnBagder: How is the audio stuff in the sim supposed to work correctly?
22:54:00austriancoderBagder: if you tell me what files i need to modify or add, i will do it by myselfe
22:54:21amiconnBecause, it doesn't work at all in my linux VM (but currently trying to install alsa - lets see whether I get it running)
22:54:29Bagderaustriancoder: first you need to write a firmware/export/config-iaudiox5.h
22:55:02Bagderaustriancoder: and make sure firmeare/export/config.h includes that if the proper define is set
22:55:06amiconn...and if I try it in cygwin (only difference being that it is sys/soundcard.h instead of linux/soundcard.h) it plays ~1 second, then stops
22:55:11amiconnSee the irc log...
22:55:45Bagderwell, I don't know much about sound stuff on linux or elsewhere
22:56:04Bagderthis works for me and Slasheri said it works for him
22:56:14amiconnI mean, what does it log?
22:56:38amiconnBecause, it tries to start playback. It opens the .mp3, lists the id3v2 content
22:56:54Bagderall that is done outside the pcm play thread
22:56:56Slasheriamiconn: Hmm, maybe you don't have a fast enough machine.. :D
22:57:00amiconnThen it starts playing; sim_sound_play() is called about 5 times
22:57:12Slasheriit barely runs realtime even on my 2.8 GHz P4
22:57:25LearBagder: Hm... Maybe I should look at win32 audio playback then...
22:57:26amiconnThen it opens the .mp3 a second time (!), listing all the id3v2 contents again
22:57:43Bagderamiconn: that's what it does
22:57:44amiconn...and it does no more call sim_sound_play()
22:57:52LearBrb, still running Firefox 1.0.4... :)
22:57:54Bagderthat's not a sim artifact afaics
22:58:01 Quit Lear ("Chatzilla [Firefox 1.0.4/20050512]")
22:58:22Bagderdon't confuse codec playback with the production of actual sound
22:58:57 Quit spiralout ("ChatZilla [Firefox 1.0.4/20050511]")
22:59:11austriancoderLinusN: i2c is the same for all coldfire cpus, so we should change this too
22:59:14amiconnBagder: It does produce sound, but only ~1 second
22:59:47amiconnaustriancoder: Doesn't the coldfire have multiple i2c buses?
22:59:48SlasheriHmm, enabling some kind of logf for simulator should help solving that..
22:59:48Bagderamiconn: I can't but to say that it needs debugging then
22:59:59BagderSlasheri: logf works fine in the sim
23:00:09SlasheriBagder: really?
23:00:18BagderI made it entirely using the sim
23:00:21Slasherioh, is it enabled by default?
23:00:31austriancoderamiconn: it has multiple, but the way to access it is the same on all coldfire cpus
23:00:36 Join memmem [0] (
23:00:47SlasheriHmm, is there a way to enable it?
23:00:57BagderSlasheri: the same way you do it for target
23:01:00amiconnaustriancoder: Yes, but it may be that components are hooked up to different buses
23:01:05SlasheriBagder: ah, nice :)
23:01:10 Join bazz [0] (~nick@
23:01:17BagderSlasheri: just enable sim too in that menu
23:02:02austriancoderamiconn: ah ok
23:02:08memmemaustriancoder: the best way to extract the low-order 8 bits is to use chksum & 0xff −− unlike casting to char, that will also work on machines where a byte does not have 8 bits.
23:02:17SlasheriBagder: oh, it worked :)
23:02:24 Join Lear [0] (
23:02:58austriancodermemmem: ok.. the rest is ok?
23:03:29memmemaustriancoder: What rest?
23:03:35 Part LinusN
23:03:56SlasheriHmm, it doesn't print logf output to stdout/stderr..
23:04:16Bagderlogf() works the same way in the sim
23:04:26Bagdersince I needed that to develop it
23:04:29austriancodermemmem: forget.. my fault
23:04:33Slasheriyep, it's nice that works too
23:04:44BagderSlasheri: but it should probably log to stderr as well
23:04:48Slasheribut maybe it could _also_ print the information to stdout
23:04:51Slasheriyes :)
23:06:23austriancodermemmem: ide interface on x5 is the same as on h1xx
23:06:49amiconnSlasheri: Speed might be an issue, but I don't think it should (Pentium M 1.5 GHz here)
23:07:06Bagderaustriancoder: I'll add X5 to the configure script for you if you want?
23:07:49amiconnBagder: Could you also add the linux sound thingy to the cygwin x11 case?
23:08:04Slasheriamiconn: Hmm, try reducing sleeps from the kernel loop if you can find them
23:08:09Bagderamiconn: I'll try!
23:08:11amiconn(remember to replace linux/soundcard.h with sys/soundcard.h)
23:08:42***Saving seen data "./dancer.seen"
23:09:59Bagderlinux can in fact use the sys version too
23:10:03BagderI'll convert to that
23:10:20memmemaustriancoder: Probably yes. Plain vanilla ColdFile IDE interface. See MCF5249UM.PDF (or its counterpart for the MCF5250).
23:12:01austriancodermemmem: could you look in ida if you find something like a reset vector to start the boot loader?
23:12:18 Join muesli- [0] (
23:12:50memmemThe reset vector in the exception vector table is 8, which is not inside the firmware file, but probably in flash memory.
23:14:23austriancoderto what must i set LCD_DEPTH if i have 260000 colors?
23:14:51Bagderbut I guess some things will break then
23:14:51austriancodermemmem: hmmm.. we must find a way to port the mkboot to support iaudio fw
23:15:00amiconnI think about introducing an additional #define
23:15:10austriancoderBagder: then we need to fix it ;)
23:15:13amiconnLCD_COLOR or such
23:15:17memmemaustriancoder: Haven't looked at mkboot yet.
23:15:19 Quit Maxime ()
23:15:24Bagderyes, we need a color define
23:15:29austriancoderamiconn: would be nice
23:15:32amiconn..because the depth alone doesn't tell whether it's colour or greyscale
23:15:40Bagderaustriancoder: first configure fix for x5 coming up
23:15:53 Quit Coldtoast (Read error: 104 (Connection reset by peer))
23:15:55austriancoderi will set it to 18.. until there is an other way
23:16:05amiconnThe epson LCD (which might be the one in the H320) has an 8-colour mode, so if we'd use that, LCD_DEPTH would be 3...
23:16:07Bagderaustriancoder: does iaudio work in any way like the scramble tool for iriver?
23:16:14Bagderthe too iaudio I mean
23:17:18memmemaustriancoder: The entry point of the firmware file seems to be 0x10000, that's at offset 0x1030 of the firmware file.
23:17:40memmemBadget: The iAUDIO firmware files are not scrambled.
23:17:52austriancodermemmem: could you try to hack mkboot?
23:18:03memmemI don't dare without BDM.
23:18:15austriancoderyou dont need to try it
23:18:15Bagdermemmem: neither is the rockbox.iriver file
23:18:30austriancoderonly changeextend mkboot code
23:19:03austriancoderBagder: iaudio tool does only update the fw checksum, so that the original fw accept it
23:19:14memmemaustriancoder: I'll have a look. No promises, though.
23:19:30Bagderok, let me rephrase my question
23:19:39Bagderdo you have a tool that creates the final binary for you?
23:19:46austriancodernot yet
23:19:51austriancoderwe are working on it
23:19:53Bagderall existing platforms run the scramble tool for that
23:20:02Bagderdon't let the name fool you
23:20:16Bagderthe iriver version doesn't actually scramble anything
23:20:26austriancoderwhat does the tool if the firmware is not scrambled?
23:20:33Bagderadds a checksum
23:20:36Bagderand a model name
23:20:45Bagderin the iriver case
23:20:59Bagderthe checksum is for the bootloader
23:21:09austriancoderso i will later extend the scramble tool
23:21:10Bagderwell, the model name too actually
23:21:18austriancoderand remove the iaudio tool
23:21:27Bagderyou can probably re-use the exact same method
23:22:27austriancodersure.. would make thinks easier
23:22:37austriancoderBadger: my config stuff is in cvs
23:23:00austriancoderwhat must i do now ? extend configure script?
23:23:19BagderI'll do that
23:23:30austriancoderah fine ;)
23:23:30Bagdercommit in one min
23:26:04austriancodermemmem: about mkboot hacking: so we must jump from 0x10000 to the addres where the bootloader will start
23:26:11memmemaustriancoder: I had a look at tools/mkboot.c; it needs to be rewritten completely for iAUDIO.
23:26:29memmemaustriancoder: Yes, we have to patch the code at address 0x10000 (offset 0x1030).
23:26:55austriancodermemmem: is it not possible to extend the mkboot to run with iriver and iaudio?
23:26:56memmemaustriancoder: It's kinda hard without knowing what's going on in the lower 64 KByte...
23:27:17memmemWell, we could reuse the I/O code, but nothing else.
23:27:39austriancoderi thout we could do it this way
23:28:05Slasheriah, simulator is great! i was able to crash the playback and traceback with gdb :)
23:28:09memmemBTW, mkcode.c does not check for buffer overflow.
23:28:14BagderSlasheri: yay
23:28:16Slasherii think i found a bug, but i will fix it tomorrow
23:28:38austriancoderdetect if we want to run mkboot with an iaudio fw.. if this is so, we could use our stuff, else if it is a iriver fw, we use the iriver stuff
23:28:43Bagderamiconn: does the configure work now for cygwin x11 sim?
23:29:11 Quit bazz (Read error: 110 (Connection timed out))
23:29:12memmemaustriancoder: Well, if you want to do it the complicated way...
23:29:34austriancodermemmem: hmmm.. g
23:29:38memmemmkcode.c also doesn't check for I/O errors. It's better to rewrite it from scratch.
23:29:58austriancoderwe will also use the scrable tool later
23:30:09memmemI hope that one looks better ;-)
23:30:24Bagderif/when you do write a mkboot version, try making it modular enough so that the functions can be used in other tools
23:30:36Bagderlike in fwpatcher
23:30:46Bagderjust an idea
23:31:03memmemWell, we could wrap opening a file in a class. Otherwise, there's not much to make modular.
23:31:14amiconnBagder: Yes it does, thanks :)
23:31:26Bagdermemmem: no, but the original mkboot code still managed to be cumbersome to reuse
23:31:41Bagderamiconn: goodie
23:31:48amiconnPerhaps the configure script could tell whether it enabled sound support... It does tell about numerous other things, like sed, gcc version...
23:31:58Bagderah, right
23:32:01Bagderme fix
23:32:14memmemIt's better to not reuse the mkboot code for iAUDIO as everything reusable (such as opening files) is trivial anyway.
23:33:00austriancodermemmem: ok... but we will use the scramble tool
23:33:16memmemWill have a look at that one.
23:33:30Bagderplease do, it makes sense
23:33:43Bagderyou can even reuse the checksum code from the iriver bootloader
23:33:44austriancodermemmem: i have started work on scramble yet
23:33:45Bagderfor yours
23:34:02memmemWhat does mkboot actually do? Overwrite some existing firmware code with the boot loader? Or use some unused space?
23:34:40austriancodermemmem: it addes the boot loader - i think - to the end of the firmware and changes that the bootloader is started
23:34:57Bagdermemmem: mkboot appends the bootloader to the existing firmware
23:35:30Bagderthere is unused space, yes
23:35:47Bagderafter all, there is 2MB flash IIRC
23:36:36memmemSo we have 502308 bytes left (plus some apparently unused bytes included in the original firmware file).
23:37:26amiconnPlenty of space, I'd say
23:37:37memmemI found the code for opening and reading files, so the boot loader could use the original firmware's functions for that. Requires hooking after initialization, though.
23:37:59Bagderthat sounds... error prone to me
23:38:08amiconnI wouldn't do that
23:38:22amiconnImho rockbox should be as independent as possible
23:38:26memmemOn the other hand, it could be useful for loading test programs.
23:38:39*amiconn waits for the day he can flash rockbox on his iriver, without iriver firmware
23:39:44 Quit Febs ("Chatzilla [Firefox 1.0.5/20050711]")
23:39:48memmemIs the bootloader position-independent code or has it to be compiled for a fixed location (to be computed from the original firmware size)?
23:40:00Bagdermemmem: it is built to a fixed address
23:40:06Bagderwhy use the size?
23:40:45memmemApparently, there's no header telling the size of the firmware. So to find out where to put the bootloader, we have to look at the size of the firmware file.
23:40:50Bagdersee firmware/
23:41:15memmemI confess that I haven't looked at any rockbox code yet ;-)
23:41:21memmemDitto for the docs ;-)
23:41:32Bagderyes, but the bootloader is independent of the original firmware
23:41:41Bagderthus it can't differ based on specific firmwares
23:41:43muesli-just ordered harry potter vol 6 :D
23:42:30Bagdermemmem: this low-level stuff is not much documented anyway, sources is the way...
23:43:01 Quit thegeek (Read error: 104 (Connection reset by peer))
23:43:16 Join thegeek [0] (
23:43:25 Join LinusN [0] (
23:43:39LinusNmkboot patches the reset vector
23:43:47LinusNat address 0
23:44:08LinusNit doesn't change anything else than that
23:44:19memmemWhy not make the bootloader position-independent? (Haven't tried -fPIC for the m68k yet.)
23:44:29LinusNmemmem: why would we want that
23:44:37memmemFor iAUDIO, we cannot patch the reset vector as it's not in the firmware file.
23:44:55memmemLinusN: to simplify the build process. No need to choose a location.
23:45:14Bagderis that to simply really?
23:45:19LinusNyou still need a separate boot loader for each target
23:45:37memmemIt will no longer depend on the firmware version.
23:45:57Bagderrockbox bootloader doesn't depend on that
23:46:03LinusNit doesn't depend on the iriver firmware version today
23:46:28memmemI thought unused space were used for the bootloader. What space is unused depends on the firmware version.
23:46:40Bagdermemmem: not really
23:46:45Bagderthey all follow the same approach
23:46:55LinusNthe last part of the flash is free in all versions so far
23:47:06memmemI guess it's time to read the code & the docs...
23:47:13LinusNmemmem: guess so
23:47:16LinusNgotta go
23:47:20 Part LinusN
23:49:17memmemI see: the bootloader uses a fixed address and hopes that it's not used by the original firmware.
23:49:28Bagdernot hopes
23:49:30Bagderwe know
23:49:52Bagderfor all those we've checked at least
23:50:07Bagderwhich is a significant amount of firmwares
23:50:44memmemWith PIC, mkboot could find out whether there's enough space left. If there's enough space, just append the bootloader as-is and patch the code to invoke it.
23:51:14Bagderbut there has been no reason for doing that
23:51:33memmemReduce the size of the firmware file? :-)
23:51:33BagderKISS you know
23:52:07Bagderwhy would it reduce the size by any significant size?
23:52:11austriancoderBagder: will use scramble −−iaudio ... not scramble −−add=x5
23:52:23Bagderaustriancoder: why?
23:52:47Bagderwon't you use the very same algorithms?
23:52:55memmemBadger: By not wasting space between the end of the (largest) original firmware and the beginning of the bootloader.
23:53:10Bagdermemmem: that space is not visible in the file size anyway
23:53:24austriancoderas far as i have seen, the header is filled with the checksum... but the checksum in the iaudio fw must be at a spez version
23:53:25memmemThe problem is that the bootloader needs memory to load rockbox into.
23:53:33Bagdernever mind
23:53:52Bagdermemmem: why is that a problem?
23:53:58Bagderit has 32MB ram
23:54:03memmemDon't call boot.ids KISS ;-)
23:54:10Bagderit sure is
23:54:25Bagderaustriancoder: scramble makes the checksum and header for your bootloader
23:54:45Bagderanyway, proceed as you see fit
23:54:50memmemIf the bootloader is not at a fixed address, it could happen to live at the address to be used for rockbox to be loaded at.
23:55:13Bagdermemmem: yes, but you'd have to PIC it at a place where the loaded rockbox is not
23:55:22Bagdernot just "anywhere"
23:55:43memmemExactly. That's a drawback of using PIC for the bootloader but not for rockbox proper. Perhaps a showstopper.
23:56:10Bagderloaded rockbox is currently ~200K
23:56:54Bagderif you can't fit the PIC code in 32MB minus those fixed 200K, then you have probbs :-)
23:57:12memmemOh, that's pretty small. Probably fewer graphics, animations, and sound than has the original iAUDIO firmware.
23:57:23Bagderno doubt
23:57:40Bagderbut rockbox also loads a bunch of things in runtime
23:58:17Bagderbut it is not very graphical, which I guess the original firmware is
23:58:18austriancoderBagder: ah... to scramble
23:58:44austriancoderscramble is used to make a valid fw? So it adds a valid header to it?
23:58:45memmemIt's way too graphical. I'd prefer a keyboard. That shows my age...

Previous day | Next day