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

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

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

#rockbox log for 2008-11-12

00:00:11preglowdoesn't help to have a nice, shiny standards document if you don't read it
00:00:18preglowor just blatantly disregard it
00:01:57gevaertsIsn't that standard behaviour in computers?
00:02:00 Join HBK- [0] (
00:03:03 Join saratoga [0] (i=9803c6dd@gateway/web/ajax/
00:03:27saratogapreglow: yeah, he didn't know anything about programming and pestered me to port it for him to some random device i've never heard of
00:03:45preglowi assume you're hard at work, then? :P
00:03:50 Quit HBK (Read error: 104 (Connection reset by peer))
00:04:45saratogaeventually he seemed to have got that hit and took off, after i told him no about 5 times
00:05:13 Quit Thundercloud (Remote closed the connection)
00:05:39Zagorehhhee. do you want to know why my build is so fast? it doesn't build all plugins... :-)
00:05:45BigBambihaha :)
00:05:52preglowi just disabled doom myself :V
00:07:48Lloreanpreglow: You left pacbox and rockboy in?
00:08:00gevaertsZagor: I was going to mention that we should do a thotough comparison before committing :)
00:08:00preglowjust disabled them as well, heh
00:08:32Zagorsomeone might want to actually install a build and see if it works too :)
00:09:19saratogai'm afraid to do stuff like that because I'll almost certainly forget and end up commiting the changed SOURCES file
00:10:15 Quit bluebrother ("leaving")
00:10:52 Quit jgarvey ("Leaving")
00:11:25preglowi always check the diff so goddamn carefully i pick it up anyway
00:11:34preglowi have to, or i'd start commiting all kinds of shit
00:12:11 Quit saratoga3 ("CGI:IRC (EOF)")
00:14:53 Quit kronflux ("Leaving")
00:15:20 Join SUSaiyan [0] (
00:16:27 Quit tyfoo2 (Read error: 60 (Operation timed out))
00:18:07 Quit kugel ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
00:19:38 Join tyfoo [0] (
00:19:40 Quit bertrik ("Leaving")
00:23:23 Join LambdaCalculus37 [0] (
00:26:40 Quit bodymind_ (Read error: 110 (Connection timed out))
00:27:18 Join bodymind_ [0] (
00:27:34 Join DerDome [0] (
00:28:48 Quit DerDome (Client Quit)
00:29:45 Join DerDome [0] (
00:31:29 Join stripwax__ [0] (
00:35:21Zagorwe have four different sound.h ...
00:35:57 Join bodymind [0] (
00:37:02 Quit robin0800 (Connection timed out)
00:37:04Zagorthree of which are included in the sim builds
00:37:06 Quit faemir (Read error: 104 (Connection reset by peer))
00:37:32Zagoryay, ordered include paths!
00:37:49 Quit tyfoo (Connection timed out)
00:37:57 Quit bodymind (Client Quit)
00:38:49 Join skipper [0] (
00:39:22 Quit DerDome ("Leaving.")
00:40:23preglowoh lord
00:40:25preglowtwice an idiot
00:40:34preglowapparently i'm not capable of even writing out correct bitmasks now
00:41:38 Join faemir [0] (
00:42:57 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
00:44:43 Quit faemir (Remote closed the connection)
00:44:49 Quit l403 (Read error: 113 (No route to host))
00:45:40 Quit massiveH ("Leaving")
00:47:52 Join l403 [0] (n=l@
00:48:17 Quit ender` (" If the art world was ran by the RIAA, you would have to lease special glasses if you wanted to see a painting. -- Rodney Ca")
00:52:11 Quit culture (Read error: 110 (Connection timed out))
00:54:51 Quit bodymind_ (Read error: 110 (Connection timed out))
00:54:56 Join DerDome [0] (
00:55:55 Quit DerDome (Client Quit)
00:56:11 Join DerDome [0] (
00:56:16 Quit DerDome (Remote closed the connection)
00:57:05 Join DerDome [0] (
00:57:26 Quit DerDome (Client Quit)
00:57:26 Quit Zagor ("Client exiting")
00:57:48 Join DerDome [0] (
00:57:52 Quit DerDome (Remote closed the connection)
01:00:17 Quit stripwax__ ("Miranda IM! Smaller, Faster, Easier.")
01:00:38 Quit Nico_P (Remote closed the connection)
01:00:46 Quit l403 ("Leaving")
01:04:01 Quit eflavoie (Remote closed the connection)
01:05:31 Join DerDome [0] (
01:05:37 Quit DerDome (Remote closed the connection)
01:08:27 Join Strife89 [0] (n=michael@
01:10:02 Join Thundercloud [0] (
01:11:19 Join mofux [0] (
01:15:02 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
01:16:53 Quit lasser (Read error: 110 (Connection timed out))
01:18:24 Quit Thundercloud (Remote closed the connection)
01:23:31 Quit MethoS-- (Read error: 104 (Connection reset by peer))
01:25:32 Quit skipper (Remote closed the connection)
01:30:07 Quit moos ("Rockbox rules the DAP world")
01:38:34 Quit Strife89 ("Bye all. :)")
01:43:49***Saving seen data "./dancer.seen"
01:48:03 Quit mofux (Remote closed the connection)
01:49:24 Part pixelma
01:49:41 Join pixelma2 [0] (n=marianne@rockbox/staff/pixelma)
02:03:22 Quit gevaerts (Nick collision from services.)
02:03:31 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
02:07:59 Join DerDome [0] (
02:12:10 Quit saratoga (" ajax IRC Client")
02:20:33 Quit |AhIoRoS| ("Abandonando, see you")
02:28:35 Quit jhulst (Remote closed the connection)
02:33:16 Join Notlistening [0] (
02:37:14 Quit knnylftn (Remote closed the connection)
02:37:48 Quit obo (Read error: 110 (Connection timed out))
02:38:26NotlisteningHi can any tell me if the rockbox libraries support printf and before you all do duh you never want to use printf there is a valid reason
02:38:30*jhMikeS screams at this PMIC charging contraption
02:52:23 Quit tessarakt (Read error: 110 (Connection timed out))
02:58:55LloreanNotlistening: I'm curious about your valid reason
03:00:24jhMikeSThere's one in the bootloaders that's hacked together
03:01:56jhMikeSbootloader/common.c starting at line 60
03:01:58Notlisteningok I have got a arm based development board that i am working on getting rockbox working on and it has a serial output for debugging and that uses a printf to output debug info
03:04:05jhMikeSmessing with logf could get you somewhere as well (it can do serial)
03:04:18 Nick fxb is now known as fxb__ (
03:04:51NotlisteningjhMikeS: Thanks will give that a go and see what we get :)
03:05:37 Join Darksair [0] (n=user@
03:05:56jhMikeSfirmware/logf.c, probably need serial_tx implemented for your hardware
03:08:04NotlisteningI will try the printf code first and then look further into different ways if we get nothing cheers
03:09:10 Quit Notlistening ()
03:10:32 Quit DerDome ("Leaving.")
03:14:29 Join super_ [0] (
03:15:28 Quit super (Read error: 104 (Connection reset by peer))
03:32:52 Join AndyIL [0] (i=AndyI@
03:43:53***Saving seen data "./dancer.seen"
03:45:11 Quit AndyI (Read error: 110 (Connection timed out))
03:49:09 Join cg [0] (
03:53:08 Quit LambdaCalculus37 ("Ka-chunka")
03:55:03 Join nuonguy [0] (
04:02:27 Join metal_falsetto [0] (n=48e68c64@gateway/web/cgi-irc/
04:07:32metal_falsettoHey all −− I'm trying to install Rockbox on my Sansa e260 (using Ubuntu Hardy 64) and am experiencing some crashing when trying to install via the Rockbox Utility.
04:07:34 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
04:17:13 Quit Bensawsome ("The awsome is gone :(")
04:17:16 Quit metal_falsetto ("CGI:IRC (EOF)")
04:30:45 Join miepchen^schlaf_ [0] (
04:35:14 Quit ap0 (Read error: 104 (Connection reset by peer))
04:37:34 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:40:50 Quit fdinel (Read error: 110 (Connection timed out))
04:49:32 Join blkhawk- [0] (
04:57:49 Join kkurbjun [0] (
04:58:46kkurbjunhow do I revert the last commit I made?
04:59:52 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
05:06:40 Quit blkhawk (Read error: 110 (Connection timed out))
05:07:29 Nick blkhawk- is now known as blkhawk (
05:14:43 Quit _lifeless (Remote closed the connection)
05:15:04 Join _lifeless [0] (n=lifeless@
05:19:17 Quit XavierGr ()
05:22:11JdGordonkkurbjun: get a svn diff of the commit, apply the patch (it will say its applied already and do you want to revert the patch) then commit again
05:22:28JdGordonwhich apparently you worked out :D
05:23:02 Quit Seed (Read error: 60 (Operation timed out))
05:24:41 Join midkay [0] (n=midkay@rockbox/developer/midkay)
05:24:43 Join midkay_ [0] (
05:24:44 Join midkay__ [0] (
05:25:12kkurbjun:), yep, thanks for that
05:25:25kkurbjunI just added it to the wiki for future reference :D
05:25:34JdGordon:) hows things?
05:25:54JdGordoncant get you to have a crack at the mr500 can i? :D
05:26:44kkurbjunit's going well. :), not yet, it's on the todo list eventually. Good news is that I think this flashwriter is pretty much done
05:27:05kkurbjunso I can move on to something else eventually (like the mr500)
05:27:08JdGordoncool, so work can start on a real bootloader?
05:27:13kkurbjunhow's it on your end
05:27:22JdGordonall good.. finshed uni so have plenty of time
05:27:41 Quit midkay (Client Quit)
05:27:41 Quit midkay__ (Client Quit)
05:27:42 Quit midkay_ (Client Quit)
05:28:06kkurbjunthe bootloader is pretty much done for the gigabeat, I run my player with the bootloader flashed now. Nice, it's got to feel good to be done with classes
05:28:22 Join midkay [0] (n=midkay@rockbox/developer/midkay)
05:28:57kkurbjunAlthough with the economic mess everyone is in school looks appealing :)
05:28:58JdGordonyep, better than that though... finshed forever... :D
05:30:06kkurbjunyeah, school was good, but it is nice to not have any more homework
05:30:36JdGordonhaha yeah
05:36:30 Quit JdGordon ("Konversation terminated!")
05:43:55***Saving seen data "./dancer.seen"
05:45:34 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
05:48:59 Quit Horscht ("electromagnetic radiation from satellite debris")
06:40:40 Join __lifeless [0] (n=lifeless@
06:45:49 Quit _lifeless (Read error: 145 (Connection timed out))
06:57:30 Part kkurbjun
07:04:35 Join _lifeless [0] (n=lifeless@
07:20:03 Quit __lifeless (Read error: 110 (Connection timed out))
07:22:41 Quit reacocard (".")
07:23:05 Join reacocard [0] (n=reacocar@
07:39:58 Nick fxb__ is now known as fxb (
07:41:16 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:43:56***Saving seen data "./dancer.seen"
07:45:49 Nick fxb is now known as fxb__ (
07:46:36 Quit jhulst (Read error: 60 (Operation timed out))
08:06:12 Join kugel [0] (n=chatzill@unaffiliated/kugel)
08:13:10 Nick super_ is now known as super (
08:20:13 Join tessarakt [0] (
08:20:17 Quit tessarakt (Read error: 104 (Connection reset by peer))
08:21:42 Join gromit`` [0] (
08:27:03 Quit BigBambi (Read error: 104 (Connection reset by peer))
08:28:16 Join draft [0] (
08:30:02 Quit miepchen^schlaf_ ()
08:30:45 Quit gromit` (Read error: 110 (Connection timed out))
08:33:07 Quit draft (Client Quit)
08:33:22 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
08:37:40 Quit GodEater (" ajax IRC Client")
08:39:24 Join GodEater [0] (i=c2cbc962@gateway/web/ajax/
08:47:23 Quit nuonguy ("This computer has gone to sleep")
08:51:22 Join ender` [0] (
08:53:08 Quit Llorean (Read error: 113 (No route to host))
08:59:14 Join Bagderr [0] (n=daniel@gateway/web/cgi-irc/
09:00:00 Join Rob2222 [0] (
09:08:23 Quit amiconn (Nick collision from services.)
09:08:29 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
09:10:40 Join petur [50] (n=petur@rockbox/developer/petur)
09:12:52 Join culture [0] (
09:16:05 Quit _Andrew_ (Read error: 110 (Connection timed out))
09:17:00 Quit Rob2223 (Read error: 110 (Connection timed out))
09:20:36 Join CaptainSquid [0] (
09:24:28 Quit pixelma2 (Nick collision from services.)
09:24:35 Join pixelma2_ [0] (n=marianne@rockbox/staff/pixelma)
09:24:39 Nick pixelma2_ is now known as pixelma2 (n=marianne@rockbox/staff/pixelma)
09:27:03 Quit Chronon (Remote closed the connection)
09:27:32 Quit Acky (Read error: 104 (Connection reset by peer))
09:27:53 Join Acksaw [0] (
09:28:41 Nick Bagderr is now known as B4gder (n=daniel@gateway/web/cgi-irc/
09:33:15 Quit kugel (Remote closed the connection)
09:34:30 Quit culture (Read error: 110 (Connection timed out))
09:35:27amiconnHmpf. Codec dependencies are still borked
09:36:17*amiconn would also like to know what's up with the Player build
09:40:51 Join einhirn [0] (
09:43:58***Saving seen data "./dancer.seen"
09:47:08 Join ap0 [0] (
09:52:13 Join Nibbler [0] (
09:53:38GodEaterwow - now there's a bit of Rockbox history I didn't know
09:54:12*JdGordon wonders what history GodEater found in the past hours of silence :p
09:55:09 Quit MrDuck (Read error: 113 (No route to host))
09:55:37GodEaterabout the Neo car player that ran rockbox
09:55:44GodEaterLinusN's post in the forums
09:56:48B4gder'openneo' even has a wiki page in our wiki
09:57:08B4gderthere's a "beta" for download from sourceforge dated 2003
09:57:48LinusNand i refrained from telling the real story of the fork
09:58:44LinusNthey took the code and stripped all the copyright headers
09:59:03B4gderand removed "rockbox" from it everywhere
09:59:03GodEaterthe swine
09:59:24LinusNthey eventually did the right thing when we told them about the gpl
09:59:38GodEateryou didn't have to threaten legal action ?
10:00:12LinusNi like to think it was classic gpl ignorance
10:00:34GodEaterthat's a very generous thing to think ;)
10:01:28Zagorthey also claimed they wrote it themselves
10:01:28amiconnlinuxstb: around?
10:01:53*amiconn found something suspicious in libdemac's filter.c
10:02:36LinusNin this case, it seems SSI hired some guys to write the software, and it loks like they didn't tell SSI that they stole it
10:03:16LinusNbut that's history
10:05:33 Join Acky [0] (
10:10:30B4gderand they claimed they "had" to fork it since they had so different requirements
10:11:06B4gderas in the hard drive always spinning, lba48, low mem situation
10:11:10scorchelike requiring their own branded version without mention of rockbox? ;)
10:11:20B4gderit was more like that yes
10:11:36B4gderbut they did a whole lot of weird changes
10:11:45B4gderI subscribed to their commit lists while it existed
10:12:04scorchethey had a public commit list?
10:12:10B4gderyes, but a delayed one
10:12:23B4gderthey had some internal cvs and synced with the external every once in a while
10:14:00 Join kugel [0] (n=chatzill@unaffiliated/kugel)
10:14:30B4gderfor example, they ripped out our playlist handling
10:14:42JdGordondid they put something better in?
10:14:49B4gderbetter in their view I guess
10:14:59B4gdernot in mine
10:15:04Zagorand, as I recall it, our painstakingly perfected shuffle, instead replacing it with a shoddy random
10:15:15B4gderyeah, that was an effect of that
10:15:31JdGordonsome people actually prefer random over shuffle...
10:15:54B4gderI think it was mostly because they didn't understand our way
10:16:12scorchewhy?...just for the ability to play songs twice?...besides that, i dont see any sort of reason why random might be better
10:17:18JdGordonbecause it keeps the playlist in order, so if you decide you'd like to listen to a song which is "close" to the current track, you dont have to go searching the playlist for it
10:17:32JdGordonyou know where it should be in relattion to the current track
10:18:43ZagorJdGordon: we can do that with shuffle too. but you can't do shuffle with random.
10:19:06Zagorwonderful sentence...
10:19:58JdGordonno we cant... our shuffle mangles the playlist
10:20:06JdGordonwe have to find the file in the browser to do it
10:20:33Zagorjust unshuffle, move to wherever, and reshuffle. our playlist is not mangled, just reordered in a controlled fashion
10:20:49JdGordonnot as easy to do
10:21:04Zagorno it's three button presses instead of one
10:21:37JdGordonnot if you removed shuffle from the quickscreen
10:22:35 Quit Acksaw (Connection timed out)
10:22:57GodEaterwho'd be silly enough to do that ? ;)
10:23:02*scorche wonders if JdGordon ever tried the random implementation in the archos's OF
10:23:06Zagormy point is we can do whatever random-lovers want, but the opposite is not true
10:23:17Zagorwe have not chosen to do it, but we have the option
10:23:54JdGordonscorche: I've only ever touched a archos during our tower building.. never used rockbox or its of on it
10:24:11B4gderits a fine brick for towers!
10:24:29JdGordonalthough, if its like the iriver of its pretty bad, randomness wasnt so good iirc
10:24:39ZagorB4gder: how does buildzip collect files? will it be terribly confused if, for example, some .rocks are in their subdirs?
10:25:18Zagoror rather, how do I adapt it so it doesn't get confused?
10:25:36JdGordonplease dont leave plugins in subdirs... buildzip might handle it but sometimes its easier to manually copy the rocks and trying to remember which go in a subdir and which dont...
10:25:45linuxstbamiconn: I'm around now
10:25:48 Quit BHSPitMonkey (Remote closed the connection)
10:25:52B4gderit scans for them in the subfolders and put them in .rockbox/rocks so I figure it should be rather easy to make work
10:27:22amiconnlinuxstb: I probably misread something, need to double check
10:28:01linuxstbamiconn: OK, I'll be around (more or less) all day, so just ping me...
10:29:21JdGordonso, what was the final verdict on the wps view mode patch? OK to commit without the silly setting to allow it?
10:29:53 Quit voltagex (Remote closed the connection)
10:37:22JdGordondoes anyone have a good name for the define the ams sansas need to skip the OF on the disk? HAVE_SUPERFLOPPYMODE_OFFSET is a bit longwinded
10:39:14 Join tvelocity [0] (n=tony@
10:39:21GodEaterHV_SPRFLPYMD_OFFST ?
10:39:31linuxstbWhy do you need the HAVE_* define at all? You could do "#ifndef XXX ; #define XXX 0 ; #endif" in config.h, and then check for #if XXX > 0 (if needed at all)
10:39:44*GodEater follows the unix tradition of regarding vowels as something other people need
10:40:10*linuxstb still wants to know why the unmount command is called umount, but is straying off-topic
10:40:14JdGordonthats what I had to start with, gevaerts rekons having a function return the value is better, and I changed my mind because a #defined number wont work for multivolume targets
10:41:35linuxstbBut isn't it only needed for the first disk? That method doesn't stop a function being used on some targets.
10:42:50gevaertsJdGordon: I would drop the superfloppy bit in the name. If a player appears that has a partition table at that offset, this still applies
10:48:02 Join what [0] (n=79b45a2d@gateway/web/cgi-irc/
10:52:41JdGordongevaerts: sort of thing?
10:53:55JdGordonIs having the function prototype at the top of disk.c be worse than having to add a new header for the one line?
10:54:22 Join robin0800 [0] (
10:55:11gevaertsJdGordon: something like that, yes. Why not just add the prototype to storage.h?
10:56:11linuxstbWouldn't it be better as a macro, and not use HAVE_STORAGE_OFFSET at all ? i.e. just have the line "const int offset = STORAGE_OFFSET(IF_MV(drive));", which would evaluate to 0 for most targets.
10:56:16robin0800Commit 19097 Power off bug could this effect other arm targets?
10:57:14linuxstbrobin0800: No - the only file that was changed was one only used by the Gigabeat F/X
10:57:38JdGordonlinuxstb: but if it is a function wont that cause warnings with function not being eclared?
10:57:55 Join lasser [0] (
10:58:09linuxstbYou declare it in something like storage-target.h - but you don't have to worry about that yet.
10:59:22robin0800linuxstb: is the bug present in other targets?
11:00:21linuxstbrobin0800: The fix was in a target-specific file, so I would guess it was a target-specific bug. But of course that doesn't mean other targets don't have similar issues - I've no idea.
11:00:39JdGordonIs there any difference really between the two?
11:01:49linuxstbIt removes a #if from disk.c, and allows things to be done at compile-time when the offset is a known constant.
11:03:25amiconnlinuxstb: Seems I was wrong regarding my suspicion, but I found something else instead:
11:03:41amiconnHISTORY_SIZE is 512, independent of the filter order
11:04:12amiconnThat means excessive memmoving for the longer-order filters, and unnecessary memmoving if the buffer would allow for a longer history
11:04:34amiconnThere are 5 different filter orders, but only 3 buffers, used as necessary
11:05:19JdGordon"#define STORAGE_OFFSET(IF_MV(drive)) IF_MV(drive==1:0?)61440" ?
11:05:28amiconnI think the filters should use the *longest* available buffer instead of the shortest, and use the extra space for less history moving
11:06:18amiconnThe only exception to this rule is if the longest available buffer is not in IRAM, then it shouldn't be used unless required
11:06:22linuxstbamiconn: Yes, that sounds sensible.
11:07:44ZagorJdGordon: a macro looks sensible, imho
11:07:45*amiconn thinks it would be best to pass the buffer size to init_filter_*(), then it can figure out the history size
11:08:14 Join Seed [0] (
11:08:51JdGordonlame, apparently you cant put IF_MV(drive) insiode the macro definition
11:10:17kugelJdGordon: why don't we just do something like : if (0==drive) "use offset"; else "dont use offset"
11:10:35JdGordonin the sd driver you mean?
11:10:53JdGordonbecause thats baaaadd mmkay
11:11:05JdGordonthe SD driver should not be fiddling with the sector being ascked for
11:11:14kugelin the card initialization. That worked for me when I tried to load from the microsd and internal depending on the microsd inserted
11:11:20kugelhm ok
11:12:52JdGordonlol @ the comment in mv.h about getting rid of the IF_MV macros "once decided on how to proceed with multivolume" which is now how many years old? :D
11:12:56kugelthough I think it's fine in the target-specific driver
11:13:32kugelJdGordon: I think that should be done similar to multiscreen (e.g. FOR_NB_VOLUMES, etc)
11:13:43amiconnJdGordon: mv.h is less than a month old afaik
11:13:50JdGordoncopied from old code
11:13:57JdGordonkugel: serves different purposes
11:14:29 Quit kugel ("ChatZilla 0.9.83 [Firefox 3.0.3/2008101315]")
11:14:30amiconnThat comment is certainly not copied
11:14:43JdGordonoh really.. /me shuts up then :p
11:17:37_lifelesshello, how i can get write acces for twiki?
11:19:02 Quit what ("CGI:IRC (EOF)")
11:20:35gevaerts_lifeless: what's your wiki username? And out of curiosity, what sort of changes are you planning?
11:21:09 Quit larstobi (Read error: 60 (Operation timed out))
11:21:27 Join larstobi [0] (n=larstobi@
11:21:50_lifelessit's AntonVeretenenko, i want to upload pcb pictures for gogear hdd1620/30
11:22:05_lifelessan then commit some code
11:22:27JdGordonlinuxstb: ?
11:23:06JdGordonthe macros arn't exactly pretty...
11:23:43gevaerts_lifeless: done. Welcome :)
11:23:55 Join MrDuck [0] (
11:24:04GodEater_lifeless: any code you've written should go in the patch tracker
11:24:16 Quit n17ikh|Lappy ()
11:25:03 Join n1s [0] (n=nils@rockbox/developer/n1s)
11:25:16_lifelessgreat, thx
11:25:29_lifelessGodEater, ok got it
11:34:22 Nick Darksair is now known as Darksair{away} (n=user@
11:38:22*JdGordon wonders where funman disappeared to
11:41:11 Join mrkiko [0] (n=IRCExplo@
11:44:00 Join bodymind [0] (n=bodymind@
11:44:01***Saving seen data "./dancer.seen"
11:44:30 Join dany_21a_ [0] (
11:45:51 Nick fxb__ is now known as fxb (
11:47:10JdGordonarg, ok maybe this would be easier just leaving it in the sd driver :(
11:48:43mrkikoHi all guys! W rockbox...
11:48:52mrkikoI would like remembering you I hate gprs :D bye!
11:48:53 Part mrkiko
11:50:45 Join kugel [0] (n=chatzill@unaffiliated/kugel)
11:50:51linuxstbJdGordon: I wouldn't put the macros in config-target.h, but in a target-specific .h file, but I'm not sure what possibilities there are with the new storage code. "Someone" also needs to clean up the target/arm/ tree to create a PP subdir (or two)... I would also write the macro using #ifdef HAVE_MULTIVOLUME, rather than the less obvious IF_MV...
11:51:10linuxstbBut my attempt is here - (not a real patch, I just edited your patch to show what I'm thinking)
11:52:14linuxstbI think this is just something that can be done many different ways, so just pick one...
11:53:21 Quit reacocard (".")
11:53:25kugelJdGordon: the fuze's offset should be the same as e200 offset (not clip,... offset)
11:55:04amiconnCombining the filter buffers would be even better. Then single filter compression levels could make even better use of the available ram
11:57:46linuxstbJdGordon: Thinking about the offset, I think we're probably trying to think too far ahead. So maybe just add a #define to config-$target.h and keep it in the driver as funman has it now. There's no real reason to change it at the moment.
11:57:56 Join reacocard [0] (n=reacocar@WL-112.CINE.HMC.Edu)
11:58:12kugelwhich is what I proposed :)
11:58:52linuxstbkugel: Yes, but it's always worth considering other methods... I still think we may want to change it in the future, but it can wait until we actually need to.
11:58:56amiconnThe offset is only needed in disk.c iiuc
12:06:33JdGordonyes, its only needed there
12:06:38 Nick Darksair{away} is now known as Darksair (n=user@
12:06:57JdGordonseems its not actually getting mounted as a superfloppy so my change breaks it :(
12:07:08 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
12:18:44linuxstbWon't it also be need for a usb_storage driver?
12:20:22JdGordondepends how usb access the disk?
12:21:01 Join moos [0] (
12:21:14linuxstbusb storage just exposes the raw sectors
12:25:04 Join Zambezi [0] (
12:33:16JdGordonlinuxstb: how does the needing a pp folder fit into our message before?
12:33:17 Quit kugel (Remote closed the connection)
12:35:30linuxstbIt doesn't specifically, it's just that the more I look at that tree, the more PP-specific things I notice in the top-level generic target/arm/ directory.
12:37:34JdGordonhow should it be done though? we dont want to make pp at the same level as arm... and putting ipod/sansa under it would go one too many levels
12:38:37 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
12:44:59 Quit tvelocity (Read error: 110 (Connection timed out))
12:47:38linuxstbJdGordon: I'm not sure. The problem is that some things are target-specific based on CPU, and others are target-specific based on manufacturer/model. It just happens that the vast majority of targets are now ARM-based, so lots of things are thrown in there when arm isn't the real common denominator (e.g. lcd-ssd1815.c, usb-drv-arc.c)
12:51:46JdGordonmaybe the really target specific bits in them go where they should, and the rest moves up to drivers?
13:06:43 Join Linton [0] (
13:09:14LintonHi, I have a WPS design question: is there a way to make part of the background transparent, so the users bg colour is used?
13:09:33 Join Thundercloud [0] (
13:11:49LintonI tried using always-dispayed images for the non-transparent part of the background, but are shown in front of text and progress-bar elements - can this be avoided?
13:12:04linuxstbLinton: No, either the backdrop is used, or the background colour, not a combination of the two.
13:13:25LintonAm I likely to find a work-around of any kind, or should I just give up on the idea?
13:14:02 Quit Thundercloud (Remote closed the connection)
13:14:25linuxstbGive up on the idea. You (the theme designer), not the user gets to choose the colours...
13:15:48LintonOK, thanks for the info.
13:19:01robin0800Linton: Can you not use a background where some of it is transparent?
13:20:31 Quit Nibbler (Read error: 110 (Connection timed out))
13:21:53Zagorwith full builds, the speed difference is more reasonable: 2:50 vs 3:14 for -j1 and nearly identical 1:25 vs 1:26 for -j4
13:23:20Lintonrobin0800: Unfortunately magenta (FF00FF - magic transparent colour most bitmaps) is just shown as magenta when used on the background
13:24:06robin0800Linton: even on the player?
13:24:46 Join fredddy [0] (
13:24:51LintonAh, interesting question - I may be putting too much faith in testing on the sim. You think there is likely to be a difference?
13:27:01robin0800Linton: I know you can see background of icons in paint programs but this disapears on the unit don't know about the sim try an icon
13:27:54n1sif that differs between sim and target it's a bug and AFAIK transparency doesn't work for backdrops, like linuxstbalready said...
13:29:26robin0800n1s: is icons the only place transparency works?
13:29:48LintonIt seems the sim is accurate - just done a little test on my gigabeat and am looking at a nice pink rectange.
13:29:51linuxstbrobin0800: Rockbox has diifferent functions for dealing with transparent bitmaps and normal bitmaps. Only the specific transparent bitmap functions look for magenta and replace it. The backdrop code doesn't.
13:30:08moosn1s: congrates for your FM works
13:30:31n1smoos: thanks :) now to integrate it nicely :)
13:30:48mooshehe ;)
13:31:20 Quit robin0800 (Remote closed the connection)
13:31:55 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
13:35:12 Join Nibbler [0] (
13:39:50 Join kugel [0] (n=chatzill@unaffiliated/kugel)
13:41:49 Join _Andrew_ [0] (
13:42:04 Part Linton
13:44:03***Saving seen data "./dancer.seen"
13:45:19Nico_PZagor: still, it's quite a speedup. also there seem to be plenty of other advantages
13:45:39Nico_PI haven't tried, but I assume recompiling a single plugin is much faster now?
13:45:47Zagoryes. speed was never the goal
13:46:06Zagormake `pwd`/apps/plugins/jpeg/jpeg.rock
13:46:28Nico_Pnice :)
13:46:47Nico_Pwhat about "make" when only one plugin has been modified?
13:47:25Zagorsince this is a single huge dependency tree (the dep file is > 1MB) you just run 'make' and it will build what needs building
13:47:35Zagorand nothing else
13:52:29Nico_Pawesome :)
13:54:28 Quit Nibbler ("Ex-Chat")
13:57:48 Quit kugel (Remote closed the connection)
13:58:36 Join kugel [0] (n=chatzill@unaffiliated/kugel)
14:12:26Zagor"124 files changed"
14:12:40Zagorit's growing into a whopper
14:20:10dany_21a_hi all... is the current SVN trunk supposed to work (iie. load the rb-fw from the flash) on a fuze?
14:20:35dany_21a_I get a *PANIC* disk_init failed!
14:21:04dany_21a_(its a 8GiB fuze)
14:21:54dany_21a_ah - okay... than its "okay" :) i thought i read something on the irc
14:21:56linuxstbZagor: Why do you need a new "BMP2RB_FORMAT" variable, and why is there only one? There are potentially 4 different bmp formats (mono, native, remote_mono, remote_native).
14:22:28kugeldany_21a_: it's possible though with a few modifications, I've done it
14:22:29Zagorpotentially, yes. only one target uses more than one.
14:23:09linuxstbI thought all (or at least, most) of the remote targets have different displays to the main LCD?
14:23:11Zagorbut yes, I need to revise that portion since it obviously does not work on that target
14:23:13dany_21a_kugel: any change that this will be in svn soon?
14:23:23Zagorlinuxstb: yes but the same display format
14:24:01kugeldany_21a_: I'd think so. it's just that someone needs to create a patch and someone who committs it
14:24:14kugelmaybe I'll put up the patch later today
14:24:32fredddyhow can I disassemble the Fuze firmware ?
14:24:59linuxstbZagor: What do you mean? e.g. H300 has 16-bit colour main LCD, and mono remote. X5 has 16-bit main, 2-bit remote, H100 has 2-bit main, 1-bit remote...
14:25:32dany_21a_kugel: large changes? if not, could you shortly describe it/pastbin it?
14:25:34GodEaterfredddy: objdump is a freebie (but not fantastic) dissassembler ?
14:26:21 Join fyre^OS [0] (
14:26:32Zagorlinuxstb: ah. looks like I hastily misinterpreted the options in tools/configure.
14:27:02fredddyGodEater, if have it installed but it always says "File format not recognized"
14:27:04kugeldany_21a_: should be a one liner, wait a second
14:28:02linuxstbfredddy: You need something like "arm-elf-objdump -b binary -m arm -D file.bin > file.txt" - where file.bin is about 128KB starting from 1024 bvtes (I think) into the firmware image file.
14:28:40linuxstbfredddy: But it's a mixture of ARM and thumb code, so you also need to do a thumb version (add something like -M force-thumb to the options)
14:29:08fredddycould [−−start-address=address] from the man page help ?
14:29:46linuxstbI never use it. But I think that defaults to zero anyway, which is where the firmware is executed from.
14:30:20JdGordonkugel: can you test that patch i put in the forum thread?
14:31:03amiconnZagor: There'll be more than the 4 current formats...
14:33:32kugeldany_21a_: try and build a new bootloader
14:33:38linuxstbfredddy: Sorry, I misunderstood your question. I've never tried that option, so I'm not sure what it's doing.
14:35:58kugelJdGordon: sure, later
14:37:02kugelJdGordon: looking at it you check for multivolume twice? first time at the creation of the STORAGE_OFFSET macro, the second time when using it
14:37:27fredddythx for your help linuxstb I´ll try my best in looking for the Fuze buttons
14:38:41 Join Strife89 [0] (i=a810eba1@gateway/web/ajax/
14:39:25kugelJdGordon: STORAGE_OFFSET is already dependant on multivolume so the additional IF_MV_NONVOID looks reduntant to me
14:39:58 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
14:40:16kugelfredddy: would be awesome if you find them
14:40:40JdGordonits not dependant
14:42:13 Quit fyrestorm (Read error: 110 (Connection timed out))
14:42:29kugelJdGordon: that part looks very dependent to me
14:42:31linuxstbJdGordon: I was also wondering why IF_MV_NONVOID is needed - I didn't use it in my patch.
14:42:49*JdGordon double checks and realises it was a copy/paste error
14:42:54JdGordonits not needed but does no harm
14:43:37linuxstbI also wouldn't use a temporary variable, even though gcc would hopefully optimise it away
14:44:02Strife89gevaerts: Hmmm, "Maxtor drives are known for their firmware problems." That I was unaware of.......
14:44:03*linuxstb still thinks it's best to just leave it as it is for now though...
14:44:04kugelthe const int offset would also not be needed
14:44:14kugelargh, linuxstb beat me :S
14:44:19 Quit moos (Read error: 104 (Connection reset by peer))
14:44:55 Join moos [0] (
14:45:15 Join faemir [0] (
14:45:39 Part B4gder
14:45:39JdGordonyes its probably not needed just yet, but that means it wont be done until later when someone tries to figure out why the hell the sd driver is returning nonesense sectores
14:46:53kugelI thought SanDisk did some fat hacks and put their firmware "partition" into that part
14:47:35JdGordonwhats to say someone wont want to try reading/writing to that OF "partition"?
14:48:00kugelafair that was exactly the way to recover the e200v2
14:48:23kugelwriting a dump of that part into that part if the OF wasn't functional anymore
14:50:34ZagorBMP2RB_FORMAT removed
14:50:36 Join J-23 [0] (
14:51:44kugelso, writing to that part would render the player into a brick (as long as rockbox isn't installed)
14:52:41 Quit fyre^OS (Read error: 110 (Connection timed out))
14:55:39 Join nplus [0] (n=nplus@141.25.Globcom.Net)
14:56:36Zagorhmm jpeg.rock becomes 0x114 bytes too large :-(
14:56:47Zagoron archos recorder, that is
14:57:27n1sis it using -Os like the core does on archos?
14:58:05J-23hmm, strange - after booting current Rockbox SVN code on my e280v2 (patched with JdGordon's patch) I see grey screen and grey (little darker than background) "Connected" text
14:58:16ZagorI'll build it in svn and compare the maps
14:58:31JdGordonthe connected is probably from the OF after usb?
14:58:57J-23it's in same place as on screen when player's connected to computer
15:00:50J-23hm, it disappeared after some time
15:02:29Zagorahh, pluginlib is added wrong, so it ends up completely in jpeg.rock
15:03:39 Join LambdaCalculus37 [0] (i=44a04303@gateway/web/ajax/
15:04:19 Join TheSphinX^ [0] (
15:06:15 Join robin0800 [0] (
15:07:42bodymindwhere is that patch for e200v2?
15:09:34linuxstbJdGordon: Your patch looks wrong if the disk has a partition table...
15:10:12JdGordonthats what I thought... but then if the disk does it wouldnt be using this?
15:10:29JdGordonthe disk has a mbr apparently starting at that offset
15:10:49 Quit Strife89 (" Ay, more Trig......")
15:12:24linuxstbThe point is that the first N sectors of the disk aren't exposed via USB, so aren't part of the "real" disk. There is no MBR at the real sector 0 on the disk - it's just the firmware image. In effect, the NAND is split into two logical disks.
15:12:38*linuxstb doesn't know how the OF would react to a partition table though...
15:13:41JdGordonwhich part of the patch looks wrong?
15:13:47JdGordonunless i upped the wrong version
15:14:16JdGordonno, thats the right one
15:15:04linuxstbThe patch simply doesn't deal with the potential situation of a partition table being there.
15:16:31 Quit kugel (Read error: 113 (No route to host))
15:16:45 Quit TheSphinX^ ("XChat@Linux")
15:17:35bodymindah tkx 8)
15:23:06 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
15:26:00Zagorah, I was missing -ffunction-sections in the pluginlib build
15:28:06linuxstbZagor: I assume your patch also gets rid of all the (almost identical) Makefiles in the individual SUBDIR plugin directories? Unifying those was something I've been meaning to do...
15:28:47Zagorwell it replaces them with minimal <dir>.make files, with typically 10 lines each
15:29:13Zagorcompile and link rules are inherited from upstream files
15:29:35linuxstbThat sounds like what I had in mind.
15:34:18linuxstbZagor: I don't want to distract you from the main build system, but if you have a chance, could you look at the Makefile in rbutil/mkamsboot/ and see if you can suggest any improvements? It currently has lots of copy/pasted rules (building multiple binaries from the same dualboot.S source file, using different #defines for different builds) and I'm sure there must be a cleaner way...
15:35:42Zagorsimply using $< and $@ would clean it up a little bit
15:36:23Zagoralso that will allow you to use a single link rule, since those are the same
15:36:52Zagori.e: dualboot-clip.elf: dualboot-clip.o
15:37:02Zagorarm-elf-ld -e 0 -Ttext=0 -o $@ $<
15:37:30ZagorI'll mock something up and paste it for you
15:37:53 Part LinusN
15:39:40n1sshould tracker tasks marked as "Due in version 3.0" that are still open be marked as "Due in version 3.1"?
15:43:58 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
15:44:04***Saving seen data "./dancer.seen"
15:44:30linuxstbn1s: Sounds sensible - assuming we can ever agree on when 3.1 should be...
15:45:47n1sI'll do it then, I think we should decide as soon as possible so people know a while before the freeze happens.
15:46:37linuxstbOK, let's decide then. When do you want to release 3.1? ;)
15:47:15n1sI still like your suggestion of 23rd Dec
15:48:32 Join m0f0x [0] (
15:48:33linuxstbIt seems a few people would prefer a 4-month cycle, and amiconn has said he would prefer a January freeze/release, rather than releasing before Christmas, so that would suit him...
15:49:00n1sperhaps it's time to try out the RSB for this one?
15:49:03linuxstbI think 3 months is enough, but there's not much difference between 3 and 4, so am happy to go with 4 for the sake of getting an agreement.
15:49:44n1sI don't feel strongly either but the difference between 4 and 5 is even smaller and so on ;)
15:54:20n1sbtw, any h10 users who can comment on FS #7672 around?
15:54:27preglowi think i'd prefer 4 months
15:54:46preglowthree releases a year sounds much better than four
15:57:01 Quit ap0 (Read error: 104 (Connection reset by peer))
15:58:00LambdaCalculus37As long as it's three stable releases a year. ;)
15:58:11ZagorI'm getting assembler errors in libffmpegFLAC on m68k
15:58:20Zagor"Error: symbol `l1_shift' is already defined"
15:58:30Zagordoes this ring a bell with anyone?
15:59:07n1slinuxstb: is FS #8391 still useful to keep open?
15:59:15preglowi'll check it out
15:59:22 Quit MrDuck (Read error: 110 (Connection timed out))
16:00:04linuxstbZagor: Thanks. There's still too much duplication for my liking, but it shows me the way forward.
16:00:32preglowsymbol used properly as far as i can see
16:00:33Zagorlinuxstb: it's difficult to remove more duplication and still keep it semi-readable
16:00:48Zagorpreglow: it goes away when compiled with -O2...
16:01:27Zagor-O3 gives the error, -O2 doesn't
16:01:58preglowthat turns on inlining
16:02:08Zagorthis is one of the trickier things with single make: every directory wants different -O flags
16:02:09preglowthat function is used several times in the same place
16:02:18preglowbut gcc can't be THAT stupid
16:02:57linuxstbZagor: I was thinking of using something like ".arm.o" for arm binaries, and ".o" for native binaries. And maybe renaming the bootimg files to have the same name as the dualboot - i.e. having one set of rules to go from "dualboot-target.elf" to "dualboot-target.o" (the file currently called bootimg_target.o)
16:03:29preglowZagor: hack shndec.c, find that label, make it local and see if it vanishes
16:03:35Zagorlinuxstb: ok, sounds fine
16:03:47preglowjust prefix it with . or something
16:03:52 Join tvelocity [0] (n=tony@
16:04:04n1sZagor: i got that same failure when testing gcc 4.3 for coldfire, removing 'volatile' from that asm block "fixed" it but that seems very fishy
16:04:34Zagorpreglow: nope. Error: symbol `.l1_shift' is already defined
16:04:50preglowZagor: well, numberic labels? :>
16:05:02preglowthose should never yield that error
16:05:32preglown1s: if removing the volatile fixed it, it sounds like gcc just ditched that block of code altogether
16:05:54preglowthought it really shouldn't, given those constraints
16:06:06preglowamiconn: what were the rules gcc uses to see if it can ditch an asm block again?
16:06:11Zagornumeric labels fixed it. is that a proper solution? my asm is ... rusty.
16:06:28preglowZagor: nothing wrong with it, the labels don't seem descriptive to me anyway
16:06:39n1sI think i tested shn decodign after that and it worked but i have no idea why that change should make sense
16:06:56preglowi'd rather see functions like that implemented naked to not rely on gcc like this
16:07:01preglowit breaks so badly between versions
16:07:45preglowi had problems with libdemac when trying gcc 4.3 for arm
16:07:52preglowthe only goddamn gcc i got to work on os x
16:07:54n1syes, inline asm is the biggest source of breakage when "upgrading"
16:08:04linuxstbn1s: I'm not sure - that comment from October 2008 seems like a different issue. There was an old problem with some ipod bootloaders not dual-booting, but that mysteriously went away (IIRC).
16:08:16 Quit CaptainSquid ("Miranda IM!")
16:09:50Zagorshndec.c:(.icode+0x1a): relocation truncated to fit: R_68K_PC16 against `*UND*'
16:11:00preglowZagor: what are you trying to do, btw?
16:11:10Zagorbuild :-)
16:11:14preglowZagor: that looks like another inline asm bug
16:11:25amiconnpreglow: Gcc tries to optimise asm blocks similar to function calls. If none of the block's output values are ever used, it'll drop the block, *unless* the block is marked volatile
16:11:30ZagorI'm going through the different targets to make sure they build with my new makefiles
16:11:46 Join tyfoo [0] (
16:11:47amiconnObviously it won't drop a block with no outputs (works the same way as a volatile block)
16:11:49preglowZagor: oh, and changing the makefiles alone provoke these errors?
16:12:37amiconnZagor: Do you have an explanation for the weird Player build behaviour? It still happens, even with after your rombox fix
16:12:38Zagorpreglow: yes. since I'm using the same flags for all files (unless I add exception rules) some files get other -O flags than in the "old" build
16:13:15preglowZagor: you're going to put them back, i assume? if a codec has different O flags, it's usually for good reasons
16:13:39amiconnZagor: The -O options in the current build system are fine tuned in several cases. If you change them, you'll lose performance in several cases
16:14:04 Quit jhulst (Read error: 148 (No route to host))
16:14:09*linuxstb deletes the same sentence...
16:14:26Zagorof course
16:14:40preglowany reason to care about these small build errors then? :P
16:15:22Zagoramiconn: I haven't looked into that problem.
16:15:37Zagorpreglow: I guess not
16:15:48*Zagor crawls back under his rock :)
16:16:33 Join Llorean [0] (
16:16:48*preglow is thankful he doesn't have to look at makefiles
16:19:38 Join Thundercloud [0] (
16:19:57 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
16:23:10preglowanyone know if there's a reason not to go with 2.19 binutils and 4.0.3 gcc? i'm starting to think binutils is my os x problem
16:24:53LambdaCalculus37preglow: I should check my binutils as well to see what version I have.
16:25:41n1spreglow: will not work for coldfire, but afaik other targets should be no problem
16:26:08n1s(unless there's a way to override ld's compatible ISA check)
16:27:01LambdaCalculus37preglow: Do you have sh-elf-ld in your /usr/local/sh-elf/bin folder?
16:27:20LambdaCalculus37For some reason, I don't have that file, and I can't build any Archos builds because of that.
16:28:57preglown1s: arm is my immediate concern
16:29:08preglowLambdaCalculus37: haven't even tried building for sh
16:29:13preglowlike i said, arm is my immediate concern
16:29:52LambdaCalculus37Can you try to set up for sh and see if you can the sh-elf-ld binary?
16:32:10preglowwell, i can't even get the arm-elf-ld binary
16:32:23preglowso it would not surprise me if i coudn't get the sh one either
16:37:42preglowconfigure: error: installation or configuration problem: C compiler cannot create executables.
16:37:48preglowisn't that just the cutest error?
16:38:52linuxstbWhich native gcc version do you have? Does OS X have multiple versions installed that you can choose from?
16:39:06preglowoh, great
16:39:09preglowan autotools bug
16:39:26preglowld: can't open output file for writing: conftest.dSYM, errno=21
16:39:36preglowthis is why it fails, gcc gets passed some shit from autotools
16:39:56preglowhave i mentioned i hate autotools?
16:40:04linuxstbNot this afternoon.
16:40:26 Join TheSphinX^ [0] (
16:40:37preglowwell, let's try new binutils, then
16:41:13preglowsince fighting autotools sounds about as appealing as suicide by drytumbling right now
16:42:49preglowperhaps i could just regen configure
16:42:59 Join kugel [0] (n=chatzill@unaffiliated/kugel)
16:43:48 Quit Bensawsome (Read error: 54 (Connection reset by peer))
16:45:15preglow2.19 did build
16:45:21preglowbut i knew that already
16:45:31 Quit faemir (Read error: 110 (Connection timed out))
16:45:35 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
16:46:27preglowanyone know if there's a stock routine to regen autotools in gnu packages?
16:46:41preglowautoreconf didn't seem to do all that was needed
16:46:43 Join kkurbjun [0] (
16:47:14 Quit kkurbjun (Client Quit)
16:48:56 Join Schmogel [0] (
16:49:17*amiconn figured out the optimisation formula for the filter buffer sizes
16:50:18n1sooh, even faster ape?
16:50:43amiconnI hope so (for -c2000 and up). It's building right now
16:51:33n1sbtw, re the midi plugin, is the gain from that early termination in the arm multiplier significant?
16:51:46amiconnWhether the effect is visible or not depends on how much overhead the memmove()s actually add to the filters
16:52:20preglown1s: depends how much it multiplies and how the operands look
16:52:30amiconnI optimised the sizes so the amount of copied bytes per sample becomes minimal
16:52:34preglownever did look much at midiplayer
16:52:47 Quit lasser ("ChatZilla 0.9.83 [Iceweasel 3.0.3/2008092816]")
16:53:45amiconnThe single-filter cases are simple - simply use the whole (IRAM) buffer. The 2-filter and 3-filter cases are a bit more complex
16:53:49n1spreglow: one of the multiplies that is done once for each voice for each sample has one operand that is always between 0-127 the other is signed 16 bit iirc
16:54:31amiconnThe amount of copying becomes minimal if the history size's ratio is that of the square root of the orders
16:54:41preglown1s: then yes, significant
16:54:48preglown1s: make sure that operand is always the last one
16:55:25 Join jgarvey [0] (
16:55:30n1spreglow: and there is no way to do this in c right? (or will gcc do it if one has char type)?
16:55:30amiconnE.g. -c4000 filter orders are 256 and 32 (1:8), hence the history sizes should be as close as possible to 1:2.828..
16:55:38preglown1s: no way, no
16:55:43preglowbloody overkill ape format...
16:56:03preglown1s: no idea if gcc will be clever enough to do it properly with char, doubt it
16:56:51n1spreglow: haha, yes doubtful. I'll try to make some asm macro or something when i get LambdaCalculus37's sansa then
16:57:35preglowoh sweet lord, it installed the bins, but called them ar, not arm-elf-ar, etc....
16:57:38preglowhow i hate autotools
16:58:14LambdaCalculus37Getting a complete, proper environment set up under OS X is proving to be a bitch, indeed. :/
16:58:58preglowhow hard can it bloody well be?
16:59:11preglowthese guys are used to old rotten vax and crays and whatnot, and autotools works there
16:59:46preglowgreat! gcc build gives same error
16:59:52 Quit Acky (Connection timed out)
16:59:57preglownow, if i wasn't so goddamn tired of having to use vmware to develop, i'd stop right here
17:00:07n1spreglow: my toolchains has them prefixed in for example /usr/local/m68k-elf/bin/ but with no prefixes in /usr/local/m68k-elf/m68k-elf/bin/
17:00:14 Quit J-23 ("Flying cow pressed ^D on my keyboard.")
17:00:36 Quit XavierGr ()
17:00:39preglown1s: other way around here :-)
17:01:09preglown1s: what's the arm-elf dir for?
17:01:15n1scan't you just "fix" your PATH then? :)
17:01:41preglowwell, it has fewer tools, and why the hell same tools in several dirs?
17:01:53 Join J-23 [0] (
17:02:55n1si don't know why they do that.
17:03:12 Join MrDuck [0] (
17:06:20preglowthey're the same files.........
17:06:40 Join ap0 [0] (
17:06:43 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
17:07:15 Quit Thundercloud (Remote closed the connection)
17:07:17preglowwell, it's either back to vmware with me
17:07:23preglowor i learn to live with the 4.3.2 install i have that works
17:07:32preglowand i don't relish the thought of using another compiler than other devs do
17:07:55 Quit Zagor ("Client exiting")
17:20:01 Join PaulJam [0] (
17:26:27bodymindJdGordon|zzz: i get a ATA Error: -1 on my e260v2 how can i debug this?
17:28:06 Join miepchen^schlaf [0] (
17:29:01*amiconn didn't expect the result he got
17:29:11linuxstbamiconn: APE?
17:29:57amiconnAdding an extra element to struct filter_t (for storing the history size) costs more performance than the bigger buffers gain
17:30:29linuxstb:( I was expecting good news...
17:30:51amiconnI have an idea how to fix this though (needing no extra element), and perhaps even get rid of another one
17:32:00 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
17:32:59amiconnBasically, struct filter_t has some redundant pointers
17:34:10amiconnhistorybuffer and coeffs are redundant (their difference is always == ORDER)
17:35:35amiconnSo I could replace historybuffer with an end-of-filter pointer. Checking that in the if() is even better, since it requires no added offset
17:38:52 Join Necromas [0] (n=chatzill@
17:39:40preglowamiconn: i use that optimization to free a register in the flac routines
17:40:19preglowwhich i guess are the same as the ape ones, only with sane filter orders
17:41:55 Join {phoenix} [0] (
17:42:11 Quit Schmogel ("Miranda IM! Smaller, Faster, Easier.")
17:42:17*linuxstb wonders if he'll live to see the TAK source code published
17:42:25Necromascan anyone confirm if the 120gb ipod classic is rockboxable? what "gen" is it?
17:42:44linuxstbNecromas: It's a "Classic" - 6th gen - not rockboxable
17:44:08***Saving seen data "./dancer.seen"
17:45:16 Quit TheSphinX^ ("XChat@Linux")
17:46:05Necromasoh and kudos to anyone working on rockbox for the sansa view
17:46:45linuxstbNo-one is (afaik)
17:48:00kugelLambdaCalculus37: Would you be available for a backlight fading test?
17:49:02Necromaslies, there is a forum thread for it
17:49:15Necromasthat's not dead
17:49:24LambdaCalculus37That doesn't mean that that someone is working on it right now.
17:49:34LambdaCalculus37That forum thread is just for research right now.
17:49:42Necromasresearch = work
17:49:48LambdaCalculus37kugel: A little busy right now, but I'll try my best.
17:50:59kugelI think I got a nicer way now, let's see if it works
17:52:53 Quit reacocard (".")
17:55:29 Quit robin0800 (Remote closed the connection)
17:58:32 Quit kugel ("ChatZilla 0.9.83 [Firefox 3.0.3/2008101315]")
18:00:11 Join Thundercloud [0] (
18:00:21 Join kugel [0] (
18:04:01 Quit J-23 (Read error: 104 (Connection reset by peer))
18:04:07kugelLambdaCalculus37: no, it doesn't work :( I don't even get behind the bootloader
18:05:12 Quit ap0 ("Ba")
18:05:30 Join J-23 [0] (
18:08:51 Join Sjano [0] (
18:09:55 Join ap0 [0] (
18:10:03SjanoI'm having trouble with my ipod video, it doesn't seem to recharge when i hook it up with a computer unless i boot up in the original firmware.
18:10:32Sjanoand i also can't enter the ipod unless i boot it up in the original firmware
18:10:55linuxstbWhat do you mean by "enter the ipod" ?
18:11:16Sjanolike go in to F: <- ipod
18:11:24Sjanoin explorer (on windows)
18:11:59linuxstbRockbox doesn't have usb support on the ipods - it should reboot into the original firmware's emergency disk mode when you attach it to a computer (but that may be unreliable...). Same with charging - I don't know the exact status, but am not surprised that it doesn't charge in Rockbox.
18:12:47 Quit bmbl ("Woah!")
18:13:13Sjanomaybe because i use the current build or something then.... (used to atleast charge on build 18xxx
18:13:24 Quit ap0 (Client Quit)
18:13:43 Quit J-23 (Read error: 104 (Connection reset by peer))
18:14:39 Quit petur ("work->...")
18:14:58kugeliirc charging should work, but don't quote me on that
18:15:12Sjanoalso the rockbox utility can't innstall a new bootloader and autodetect detects ipod video 30gb insted of 60gb (are there any diffrence on that?)
18:15:54Sjano(installing the firmware on it works, though)
18:16:13 Join MethoS [0] (
18:16:48kugelSjano: rbutil can't distinguish between the 30 and 60 gb version
18:17:18 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
18:18:14Sjanokugel: yeah, i just choosed the 60gb. guessed it would be something like that.
18:18:40 Quit kugel ("ChatZilla 0.9.83 [Firefox 3.0.3/2008101315]")
18:19:27linuxstbSjano: The real difference is RAM - 30GB players should have 32MB RAM, and 60GB/80GB should have 64MB RAM. So it's safe to use the "30GB" version of Rockbox on both, but you can't use the "60/80GB" version on a 30GB - so rbutil defaults to the safe option.
18:19:40domonokyrbutil can not distinguish the 30 and 60gb versions, but it is save to install a 30gb version on a 60 (not the other way round) ...
18:19:55*domonoky is too slow...
18:20:34 Join J-23 [0] (
18:21:27Sjanoif i manually install the firmware, is the bootloader included?
18:22:22linuxstbdomonoky: rbutil should be able to if Mr Someone implements this - (there's a RAM field)
18:22:57linuxstbSjano: What does rbutil say when you try to install the bootloader? The common problem is that you're not running rbutil as the Adminstrator/root user.
18:23:34domonokylinuxstb: mr. Someone seem to have too much todo.. :-)
18:24:00n1slinuxstb: it could also be solved by runtime ram size detection
18:24:18linuxstbn1s: Indeed - something else on Mr Somone's list.
18:25:04 Join kugel [0] (n=chatzill@unaffiliated/kugel)
18:25:38kugelso can anyone tell me how to debug rockbox build where I cannot even get to the main menu?
18:26:39n1sprinting stuff as early as possible and continuing further into the satrtup to see where it fails
18:26:58amiconnlinuxstb: Speedup for -c2000 and -c3000 is ~0.1% realtime on PP5002, -c4000 speedup is immeasurable
18:27:55domonokykugel: put some outputs/panicfs into apps/main.c ..
18:28:08LambdaCalculus37amiconn: Still very slow on PP5002.
18:28:10amiconnSo it looks like the memmove() is neglectible, but the filter wrap condition simplification could probably be committed (it's basically independent)
18:29:27 Quit linuxstb (Remote closed the connection)
18:31:43 Join Schmogel [0] (
18:32:00 Join bertrik [0] (
18:32:43*kugel still waits for 1 of the 2 ape user to thank amiconn for his recent optimizations :p
18:33:20gevaertsI believe one of them also has backlight fading enabled ;)
18:33:24 Join ap0 [0] (
18:33:36kugelprobably :p
18:33:37 Quit bodymind ("Ex-Chat")
18:34:14kugelgiven that I expect over 50% of the people to want this backlight fading :)
18:39:36 Quit Darksair ("Everything that has a beginning has an end.")
18:40:00 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
18:42:24kugeldomonoky: uh
18:42:35 Join aarcane [0] (
18:42:39kugeldomonoky: the panicf didn't happen.
18:42:47kugelBut I'm in the main menu now :?
18:43:02domonokykugel: and where did you put it ? before the init() ?
18:43:26kugelbefore backlight_init() in init()
18:43:36domonokybe carefull if you put it into the init function itself.. there are two init functions..
18:43:46domonokyone for the sim, and one for target...
18:44:22domonokyand for me the init just takes very long, so be sure to wait a good bit after bootup...
18:44:28kugelah right
18:44:34kugelI put it in the wrong one
18:44:42amiconnlinuxstb: Am I right in thinking that the 'adaptcoeffs' are always a power of 2 ?
18:44:48 Part Sjano
18:44:50*domonoky did the same yersterday.. :-)
18:45:11amiconn(either positive or negative of that)
18:45:35kugeldomonoky: haha :p
18:50:06 Quit PaulJam (".")
18:52:00kugelevil typos :S
18:52:26kugelbtw: can codec building be disabled like plugins?
18:52:52 Join kronflux [0] (
18:55:16amiconnThat wouldn't make sense....
18:55:43amiconnRockbox without codecs is worth nothing on an swcodec target
18:55:57kugelfor testing only of course
18:56:03amiconnTesting what?
18:56:38kugelhm, make bin should do the job
18:56:52amiconnFor plugins, it makes sense, because adapting plugins to a new target needs considerable work (different buttons, screen resolution...)
18:57:03 Quit pixelma2 ("-")
18:57:27amiconnCodecs don't need that. Optimisation is another matter, of course
18:57:32 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
18:58:13 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
18:59:34 Quit neddy ("Leaving.")
19:01:10 Join mofux [0] (
19:05:12kugeldomonoky: do I need to include something for panicf?
19:06:04domonokykugel: no, panicf should work as is in main.c, maybe make a test with a panicf before the init()
19:06:30kugeldomonoky: I'm not in main.c anymore. The issue must be in backlight code.
19:07:40*domonoky thinks you need to include panic.h :-)
19:07:57kugelI did that and I don't get any panic
19:09:00domonokyand you are sure that this code is really called ?
19:09:16kugelnot anymore
19:10:16 Quit Shiny (Read error: 104 (Connection reset by peer))
19:10:22 Join Shiny [0] (
19:16:52 Quit Thundercloud (Remote closed the connection)
19:18:31kugelI'm so stupid
19:20:10kugeldomonoky: yep the code wasn't called :S
19:22:27 Quit m0f0x (Read error: 110 (Connection timed out))
19:25:35 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:26:24kugeldomonoky: believe it or not, I've never been that happy about a panic
19:26:47 Quit tvelocity ("Αποχώρησε")
19:31:13 Quit J-23 (Remote closed the connection)
19:36:06 Join J-23 [0] (
19:37:28kugelnice to see that the code that I wrote while not having access to my target actually worked, although a missing #define in my target config prevented it from actually workin
19:37:38 Join Horschti [0] (
19:38:24 Quit Horscht (Nick collision from services.)
19:44:14***Saving seen data "./dancer.seen"
19:50:53 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
20:02:10 Quit ap0 ("Ba")
20:02:19 Join ap0 [0] (
20:06:51preglowanyone know if sd cards should switch to data state at once when receiving a read command?
20:16:09 Part fredddy ("Verlassend")
20:16:55gevaertsHow should multi-driver storage handle errors in *_init()? Continue with the next driver or give up? If the former, what should it return?
20:17:24 Join Strife89 [0] (n=michael@
20:18:05*linuxstb waits for someone to ask an easy question
20:18:25gevaertsWhat is your favourite colour?
20:18:56*linuxstb doesn't even know that
20:20:51*LambdaCalculus37 throws linuxstb into the Gorge of Eternal Peril :)
20:21:17 Join TheSphinX^ [0] (
20:23:27nanokis charging of the x5 target supposed to work while running rockbox (that is: work reliably, and to an actual full charge)?
20:25:08rashernanok: TargetStatus lists charging as "Partial" for the X5
20:25:28LambdaCalculus37nanok: Try charging with the device off instead.
20:25:30rasherI don't know what that means though
20:25:40rasherMaybe someone with a clue should add a footnote
20:25:53nanokLambdaCalculus37: thanks, i thought so..
20:26:18nanokso the 5h or so i got today should not scare me too much. maybe there is still hope for this old battery :)
20:27:17nanokLambdaCalculus37: (i am currently charging with the player off, btw, i was asking to know if i should still hope or just order a new battery :) )
20:30:46 Quit Horschti ("I got raided by the FBI and all i got is this lousy quit message")
20:31:42LambdaCalculus37nanok: If you wish to, order a new battery for it. It never hurts to have a spare, anyway. :)
20:32:20preglowgevaerts: well, what makes sense? is an _init failure something we can recover from, or even something to be expected from time to time?
20:33:01gevaertsI assume we can't recover, but failing to init one storage device doesn't mean anything about the others
20:33:26preglowwe might want to let the individual driver decide, really. perhaps let it return several values, one meaning "failed, but go on", one meaning "failed, panic"
20:33:37pixelmaLambdaCalculus37: but no sense in keeping a battery in the drawer and let it deep discharge...
20:33:37preglowbut again, an ata driver might not always be primary storage
20:33:56preglowso that decision should probably not be in the driver
20:34:28LambdaCalculus37pixelma: True...
20:34:33gevaertsI'm going to do "first init failure stops init" now. We can always change it later
20:35:01 Join Acksaw [0] (
20:35:29*preglow wonder why his card refuses to switch to data state :/
20:37:11 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
20:41:53 Quit J-23 (Read error: 104 (Connection reset by peer))
20:43:10 Join Zagor [242] (
20:43:11 Quit MethoS (Remote closed the connection)
20:43:31gevaertsSo what should storage_spinup_time() return? I guess the most sensible is the maximum of all drive(r)s
20:43:50 Join J-23 [0] (
20:43:55gevaertsAnd what about storage_last_disk_activity()?
20:44:08 Quit AndyIL (Read error: 60 (Operation timed out))
20:44:29*gevaerts wants to have a CF+Ramdisk ipod running today
20:47:24gevaertsstorage_last_disk_activity() is only used by powermgmt.c to decide on idle poweroff, so I guess it needs "most recent of all"
20:47:38 Join AndyI [0] (i=AndyI@
20:50:24kugelcan anyone give me a short introduction to logf? I made a logf("..") in my file and built a logf build
20:50:49preglowgevaerts: shouldn't that be renamed to storage_last_activity, then? :/
20:50:49kugelthe logf screen is empty though, and I can't find any file that seems to be created by logfdump
20:50:53preglow:/ -> :>
20:51:16gevaertspreglow: good point. I think that's for a separate commit though :)
20:51:48gevaertskugel: you also need to #define LOGF_ENABLE in the files where you want it
20:52:21kugelok, lets try that
20:52:22 Quit _Andrew_ (Read error: 104 (Connection reset by peer))
20:54:58kugelgevaerts: still no luck
20:55:18linuxstbkugel: Which target?
20:56:00linuxstbYou need to define LOGF_ENABLE before the #include "logf.h" I think.
20:57:20kugellinuxstb: thanks
20:57:33kugelalthough, the messages are cut. is there limited space?
20:57:34 Join MethoS [0] (
20:58:25 Join `Zee [0] (n=chatzill@
20:58:35`Zeehi guys
20:58:47gevaertskugel: yes. 32 character IIRC
20:59:03`Zeei just wanted to say you guy rock
20:59:17`Zeethe developers of rockbox that is
20:59:46gevaertsThanks :)
21:01:39 Join Chronon [0] (
21:02:44 Quit kronflux (Read error: 104 (Connection reset by peer))
21:03:05 Join kronflux [0] (
21:03:07ChrononDoes anyone know if line out works properly on the c200s? There's a person on the forums using a dock that works with OF and apparently with Rockboxed e200, but not with a Rockboxed c200.
21:03:21 Quit `Zee ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
21:03:36 Join `Zee [0] (n=chatzill@
21:06:41ChrononIf you can shed any light on that, the topic is here:
21:07:23ZagorI don't know if any dev has a dock
21:07:30kugelI think I found out why one should always make { } after an if ()
21:07:50rasherMy father has one, but no c200
21:08:10ChrononZagor: the set of people with both seems very small. . .
21:08:11rasherAnd I don't know if I qualify as a dev, in that respect
21:08:29Zagorrasher: but does your father? :)
21:08:35gevaertsI once saw a dock in a shop. I should have bought it
21:09:08ZagorChronon: I see no reason to doubt the person saying it works on e200 but not on c200
21:09:44ChrononI don't doubt them.
21:09:57gevaertsCan I assume that any storage driver knows at boot time how many drives it handles?
21:10:10 Part dany_21a_
21:10:29rashergevaerts: not usb host :)
21:10:50gevaertsrasher: that one can have a hardcoded maximum instead :)
21:12:18 Join tessarakt [0] (
21:13:03gevaertsAll I need is that it can tell me a number after _init()
21:13:21nanokis the pinout of the plug known already? maybe a plug from a standard c/e200 could be "hacked" to test the functionality of available connections?..
21:13:55nanokwouldn't be extremelly fun for a diy project, i must admit, but it should be a start, for testing
21:13:56bertrika printf in the clip bootloader should show immediately (without any calls to lcd_update), right?
21:14:51linuxstbbertrik: It depends on the value of "verbose"
21:16:06bertrikah thanks
21:16:59*linuxstb wonders why the AMS bootloaders set verbose to false
21:18:12preglowand now i discover i can't input curly braces into vmware
21:18:16preglowboy, do i need a drink
21:18:24*LambdaCalculus37 hands preglow a beer
21:19:12preglowalready got one, but two at a time sounds good now :P
21:19:47 Quit TMM (Read error: 104 (Connection reset by peer))
21:20:11preglowhaha, and gnome pukes when i try to change the layout
21:20:22preglowi so goddamn hate computers sometimes
21:20:36preglowwith so many fine hobbies in the world, i'm doing this
21:21:28 Join Thundercloud [0] (
21:24:59 Join kachna|lappy [0] (
21:25:50kugellinuxstb: JdGordon made a commit to the bootloader recently
21:26:04 Quit MrDuck (Read error: 104 (Connection reset by peer))
21:26:13kugellinuxstb: r19085
21:27:57 Join miepchen^schlaf_ [0] (
21:28:35 Join TMM [0] (
21:31:42 Quit J-23 (Remote closed the connection)
21:32:06kugelLambdaCalculus37: I think I got it nicely done now
21:32:20 Quit Schmogel (Read error: 104 (Connection reset by peer))
21:33:32LambdaCalculus37kugel: Post a patch at Flyspray, and I'll try it out later today.
21:33:52*gevaerts spots a mess
21:33:53 Join faemir [0] (
21:35:09 Join petur [50] (n=petur@rockbox/developer/petur)
21:36:50 Quit kronflux (Read error: 104 (Connection reset by peer))
21:37:26 Join kronflux [0] (
21:37:36bertrikgevaerts, where?
21:38:30 Quit tyfoo (Read error: 54 (Connection reset by peer))
21:38:40gevaertsBy some accident, all HAVE_MULTIVOLUME targets up to now have had exactly two volumes, so (a) the number isn't set in config-*.h, and (b) some places may assume 2
21:39:44 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
21:40:10 Join sarixe [0] (n=sarixe@
21:40:11 Quit Bensawsome ("The awsome is gone :(")
21:40:46gevaertsThis is in addition to NUM_VOLUMES being taken to mean number of drives in some parts of the code and number of FAT filesystems in other parts
21:41:31 Join tyfoo [0] (
21:43:40`Zeewhats the best way to go about installing rockbox on my e260 got it today
21:43:48`Zeei think i did it right
21:43:52`Zeeused the installer app
21:44:16***Saving seen data "./dancer.seen"
21:45:27kugellcd_enable is buggy (on e200 at least)
21:45:44gevaerts`Zee: that should be the best way
21:45:50 Join _Andrew_ [0] (
21:46:01`Zeeyeah only problem is the themes
21:46:11`Zeebecause it rockbox 3
21:47:19LambdaCalculus37`Zee: Rockbox Utility current can't install themes. You'll have to do that yourself.
21:47:35`Zeeyeah tried a feew but they never worked
21:47:37 Join tvelocity [0] (
21:47:54linuxstb`Zee: Themes from where?
21:48:47`Zeei grabbed a few from this site
21:48:48kugellinuxstb: in lcd-e200.c, shouldn't the display_on variable set to true (false) just before lcd_enable is finished and not within?
21:48:58`Zeei grabbed a few from this site
21:49:19kugelthe way it's now, there's a chance that lcd_enabled returns true even if lcd_enable isn't done yet
21:49:25 Quit Shiny (Read error: 104 (Connection reset by peer))
21:49:38 Join Shiny [0] (
21:49:42linuxstb`Zee: Try these -
21:50:39kugelI assume this is the reason my while(!(lcd_enabled)); loop can create a infinite loop under certain conditions
21:51:01`Zeethe rockbox font pack on that site is it the same as the one in the installer
21:51:22kugelcould've another reason though too
21:52:47kugelalso, I spot many udelays and sleeps in the driver, is that realy required?
21:52:59 Join culture [0] (
21:53:08gevaerts`Zee: which one? or
21:53:26gevaerts`Zee: site has never been official, and it looks horribly outdated. You're better of ignoring it
21:54:11linuxstb`Zee: Yes, that's the standard fonts that the installer downloads.
21:54:11gevaertsThat's the same font pack as the one from the installer, yes
21:54:15kugelalso I would assume e200 doesn't need lcd_sleep since it uses lcd enable all the time
21:54:38*bertrik checks if there's something more on the i2c buses in his clip
21:54:40`Zeeals is the firware the latest in the installer or do i have to update it after with the other install?
21:54:45 Quit parafin (Nick collision from services.)
21:54:50 Join parafin|away [0] (
21:55:05*gevaerts decides that he can't do multi-driver properly without cleaning up HAVE_MULTIVOLUME
21:55:34kugelgevaerts: you know it, /someone/ has to do it ;)
21:55:57gevaertsHow do we handle this? Max. partitions per drive, and max. nr of drives, and the NUM_VOLUMES for fat.c being the product of both?
21:56:24`Zeealso how much space should i have available before i install the full version?
21:58:13LambdaCalculus37Mr. Someone has a *lot* of work to do now. :)
21:58:36kugelgevaerts: sounds good. max nr. of drives can be two
21:58:38 Quit sarixe ("Ex-Chat")
21:59:11 Nick parafin|away is now known as parafin (
21:59:33 Quit LambdaCalculus37 (" ajax IRC Client")
22:00:15gevaertskugel: no assumptions of that sort this time :)
22:00:49kugelwell, I doubt you will ever have the possiblity to have more than your microsd+internal memory at the same time
22:01:00kugelexcept in rockbox as an app maybe
22:01:09linuxstbusb host?
22:01:27kugeluhm, yea, forgot about that
22:01:33gevaertsPeople swapping in a two-CF card ATA adapter?
22:01:57kugelso, really, what's the need to have (a configurable) lcd sleep if lcd_enable is always done before?
22:01:58gevaertsrockbox as an app won't use this code anyway
22:02:35 Quit jhulst (Read error: 110 (Connection timed out))
22:04:02`Zeealso how much space should i have available before i install the full version?
22:06:52gevaerts`Zee: hard to say exactly. How much do you have?
22:07:07`Zeeabout 1.1gb
22:07:18gevaertsThat's plenty
22:07:49gevaertsDepending on what exactly you install you need somewhere between 5 and 50MB I guess
22:08:40bertrikno luck finding any other devices on i2c so far, I was hoping to find the fm chip
22:09:17 Quit Shiny (Read error: 54 (Connection reset by peer))
22:09:30kugelbertrik: also no trace of buttons?
22:09:38kugelar wrong target
22:09:43kugelnevermind :p
22:11:53 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
22:12:30kugelbertrik: remember that issue with backlight fading when you went from off to on in the settings (the screen would turn white)
22:12:42 Quit {phoenix} (Read error: 104 (Connection reset by peer))
22:13:15bertrikyes, I remember
22:13:20*gevaerts wonders if rockbox will ever support things like that sansa slotmusic player, i.e. a single drive with hotswap
22:14:04`Zeegevaerts i installed a full install from the installer thats all
22:14:07kugelI can't find a solution for that. It's not reliably reproducible at all, and it only seems to happen in this particular setting
22:14:26 Join lasser [0] (
22:14:47kugelbertrik: i.e. it doesn't happen in lists etc no matter at which point you let it fade up again
22:17:26 Quit Zagor ("Client exiting")
22:17:54kugelbertrik: any ideas?
22:18:11bertriklast patch I tried didn't seem to have the problem anymore
22:20:49 Quit Strife89 ("Ah, work to do.....")
22:21:57kugelI have my doubts, we just might have tried it often enough
22:23:57kugelmy while loop seems to fix this particular is for just about every screen except this option screen
22:24:58 Join HBK1 [0] (
22:25:05kugelbertrik: it only happens if you change from on to off and on again very quickly
22:25:20kugeland then not even always
22:25:37 Quit GodEater (" ajax IRC Client")
22:27:09 Quit ap0 (
22:27:09 Quit bertrik (
22:27:09 Quit bmbl (
22:27:09 Quit ender` (
22:27:09 Quit _lifeless (
22:27:09 Quit denes_ (
22:27:09 Quit tchan (
22:27:09 Quit nanok (
22:27:09 Quit puzzles (
22:27:09 Quit blithe (
22:27:09 Quit Slack (
22:27:09 Quit thegeek (
22:27:09 Quit Gareth (
22:27:09 Quit tim__b (
22:27:09 Quit Galois (
22:27:09 Quit crashd (
22:27:09 Quit Bensawsome (
22:27:09 Quit tvelocity (
22:27:09 Quit TMM (
22:27:09 Quit pixelma (
22:27:09 Quit Nico_P (
22:27:09 Quit synergist (
22:27:09 Quit Tuplanolla (
22:27:09 Quit stu8ball (
22:27:09 Quit Kohlrabi_ (
22:27:09 Quit plus_M (
22:27:09 Quit havien (
22:27:09 Quit Bjoern-Erik (
22:27:09 Quit basti (
22:27:09 Quit ChanServ (
22:28:48NJoinblithe [0] (
22:30:14 Join PaulJam [0] (
22:36:50 Quit kachna|lappy (No route to host)
22:42:10 Quit HBK- (Read error: 110 (Connection timed out))
22:43:10kugellcd_enabled isn't reliable at all
22:43:53 Join GodEater [0] (i=c2cbc962@gateway/web/ajax/
22:43:57gevaertspreglow: (and others) You may want to have a look at FS #9545. It's not entirely done yet, but it compiles at least for some targets. No other tests have been done
22:45:33NJoinap0 [0] (
22:45:33NJoinbertrik [0] (
22:45:33NJoinbmbl [0] (n=Miranda@unaffiliated/bmbl)
22:45:33NJoinender` [0] (
22:45:33NJoin_lifeless [0] (n=lifeless@
22:45:33NJoindenes_ [0] (
22:45:33NJoinnanok [0] (n=nanok@
22:45:33 Join puzzles [0] (n=dan@xmms2/developer/puzzles)
22:45:33NJoinSlack [0] (
22:45:33NJointhegeek [0] (
22:45:33NJoinGareth [0] (
22:45:33NJointim__b [0] (
22:45:33NJoinGalois [0] (
22:45:33NJoincrashd [0] (
22:46:00NJoinChanServ [0] (ChanServ@services.)
22:46:00Mode"#rockbox +o ChanServ " by
22:46:39gevaertsI'll concentrate on testing the IF_MV/IF_MD split during the next few days, as that touches all targets. I think multi-driver is a bit less critical right now
22:47:50NJointchan [0] (
22:47:50NJoinBensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
22:47:50NJoinTMM [0] (
22:47:50NJoinpixelma [50] (i=pixelma@rockbox/staff/pixelma)
22:47:50NJoinNico_P [50] (n=nicolas@rockbox/developer/NicoP)
22:47:50NJoinsynergist [0] (
22:47:50NJoinstu8ball [0] (
22:47:50NJoinTuplanolla [0] (n=jani@unaffiliated/tuplanolla)
22:47:50NJoinKohlrabi_ [0] (
22:47:50NJoinplus_M [0] (
22:47:50NJoinhavien [0] (
22:47:50NJoinBjoern-Erik [0] (
22:47:50NJoinbasti [0] (n=basti@
22:48:45 Quit bmbl ("Woah!")
22:49:12 Quit TMM (Connection reset by peer)
22:49:18gevaertspreglow: I've used the original switch approach for now. I'm not convinced that function pointers are better, but feel free to propose better ideas
22:49:53 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
22:49:54 Join TMM [0] (
22:49:55preglowgevaerts: can't think of anything better than just having an array of structs (of function pointers), where the array will contain structs per ifdef
22:50:12preglowgevaerts: that solution means hacking just one ifdef instead of several, but i guess it's not something you do often
22:50:13 Join LambdaCalculus37 [0] (
22:50:16preglowso anything goes
22:50:54preglowspeedwise both approaches mean jump table, so i guess they'll be about equally fast
22:51:07preglowone less function call in pointer approach :>
22:51:40 Join n17ikh|Lappy [0] (
22:53:00`Zeedoes rockbox have an alarm clock
22:53:03gevaertsThe problem I see with function pointers is that I don't think they can be used for all functions, so you'll end up with a mix
22:53:08`Zeesay i plug this into its dock
22:53:09 Quit nplus (Remote closed the connection)
22:53:17`Zeewill it wake me up?
22:53:31`Zeefor the sansa e260
22:53:40preglowgevaerts: nah, they won't work if not all functions have an equivalent in each driver
22:54:47gevaertsThey do have an equivalent, but for some functions you need to call the one corresponding to the drive while for others you need to call them all once (i.e. just looping won't work, as there may be drivers that handle more than one drive)
22:55:25preglowyeah, true
22:55:26preglowgo switch
22:55:56preglowwell, woot
22:56:03preglowat least not sd controller claims it's getting data
22:56:09preglowi wonder why the fifo never claims to fill, then
22:56:21gevaerts`Zee: have you looked in the manual?
22:56:41`Zeeyeah i see a clock
22:56:48`Zeenot an alarm
22:57:00`Zeethinking i missed it
22:57:31`Zeedamn earphones for 3k
22:57:39`Zeesmall compact ones
22:57:44kugelbertrik: Ok, I could now reproduce this with the current build. It's not my code which causes this
23:00:29LambdaCalculus37kugel: Got a patch posted for me to try?
23:04:22 Quit `Zee ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
23:07:29 Join m0f0x [0] (i=m0f0x@gateway/tor/x-d12175fbe3c191e9)
23:12:57 Quit domonoky (Read error: 104 (Connection reset by peer))
23:15:07preglowreceiving data, but fifo never gets anything at all
23:17:52 Nick HBK1 is now known as HBK (
23:18:25 Quit ap0 (Read error: 104 (Connection reset by peer))
23:25:22 Quit TheSphinX^ ("XChat@Linux")
23:30:37kugelLambdaCalculus37: buh
23:30:59kugelLambdaCalculus37: one second, need to messure bin size
23:32:21 Quit petur ("plop")
23:35:09 Quit tyfoo ("Carpe diem")
23:35:22 Quit n1s ()
23:35:25kugelLambdaCalculus37: if I saw this correcly my new patch adds more than 2K binsize
23:35:33kugelwhich can't be right
23:36:51LambdaCalculus37kugel: That's not good. Better see what's wrong.
23:37:03LambdaCalculus37jhMikeS: Ping
23:37:51kugelLambdaCalculus37: the previous patch only added very few 100 bytes
23:39:00kugelLambdaCalculus37: ok, calm down. my "svn build" alone is already 1500bytes bigger than the one shown in the build table
23:39:45*LambdaCalculus37 is calm... he only had 5 cups of coffee :P
23:40:35LambdaCalculus37And no beer. ;)
23:40:57kugelbut I wasn't ;)
23:41:44 Quit kronflux ("Leaving")
23:44:19***Saving seen data "./dancer.seen"
23:47:20 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
23:56:30 Join pie [0] (
23:57:44 Quit m0f0x ()
23:58:57 Nick pie is now known as razym (

Previous day | Next day