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

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

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

#rockbox log for 2006-01-03

00:03:49 Join Mongey [0] (
00:19:14 Quit BHSPitLappy (Connection timed out)
00:23:08muesli-just built (hope i did) the simulator. how do i start it?
00:23:33Bagder'make install' then './rockboxui'
00:25:36muesli-mmh /home/guest/tools/convbdf: not found | no wps module present, cant do the wps magic
00:26:15muesli-but i am still in ~/build>
00:26:18Bagderh300 sim?
00:26:34 Quit ender` (Connection timed out)
00:27:09Bagderand when you run 'make', doesn't it mention convbdf among the first few lines?
00:27:42Bagderthe missing wps stuff is probably because you need to checkout the wps module
00:27:49muesli-gcc -g convbdf.c ...
00:28:52Bagderyou have a fonts dir?
00:29:31Bagdercheck out the fonts and wps modules
00:29:32muesli-this ine
00:29:57Bagderyes, that one
00:31:30 Join ashridah [0] (
00:31:58muesli-its still "making"
00:37:43 Join aliask [0] (
00:42:43muesli-how do i get the wps module?
00:42:54Bagdercvs -d blablabla co wps
00:45:33muesli-ok...i see that h3xx pic in a seperate window..when i press the numbers nothin happens
00:47:36BagderI've never used the win32 sim so I can't offer much help there
00:48:03muesli-ok ;)
00:50:00muesli-but thank you anyway
01:01:02 Join LinusN [0] (
01:05:37MoosBonsoir LinusN
01:06:18muesli-n8 m8s
01:06:19 Quit muesli- ("ich will Kühe!!!")
01:07:53 Quit Mongey ()
01:10:00 Quit Strath (Read error: 104 (Connection reset by peer))
01:13:44 Join Strath [0] (
01:21:24 Join cbrian [0] (
01:21:27 Part cbrian
01:23:07 Quit Moos ("Happy new year @ all")
01:28:06 Quit aliask ("Chatzilla 0.9.69 [Firefox 1.5/2005111116]")
01:32:49 Part Paul_The_Nerd
01:36:07 Join matsl [0] (
01:39:44 Quit matsl (Client Quit)
01:40:58nico_i have a question about rockbox architecture :
01:41:48nico_is code in firmware/ only use for bootloader or also for application rockbox ?
01:42:40nico_i do not really understand the organisation between apps/ and firmware/
01:43:08markunnico_: apps should containt the hardware independant code, firmware the hardware dependant.
01:44:05 Join mymomthelush [0] (
01:44:10 Part mymomthelush
01:44:36nico_so you means low level code with firmware
01:46:54markunyes, I believe low level code should be in firmware
01:49:49***Saving seen data "./dancer.seen"
01:56:02LinusNthe "firmware" and "apps" directories come from the beginning of rockbox, when we had the idea that rockbox would support different "applications" that ran on top of the base firmware
01:56:36nico_ok ok
01:56:46LinusNwhich actually is almost the case now, where the bootloader uses the firmware lib, like the application does
01:57:30LinusNthe names "firmware" and "apps" are kind of misleading nowadays, but cvs doesn't really like renaming fines and directories
01:57:43nico_firmware is like a kernel which give us hardware access, scheduler of threads ...
01:59:09LinusNsort of, yes
02:08:36nico_i am happy to have been able to speak a-dady-of-rockboy
02:08:59nico_rockbox ...
02:09:27nico_i am a new user of it and it seems to be very god
02:10:26nico_PS : i would write 'a rockboy-daddy'
02:10:31nico_good night
02:10:44HCla what now?
02:11:24nico_sorry : rockboX-daddy
02:11:31*HCl sleeps.
02:12:12 Quit nico_ ("Leaving")
02:20:08 Join Matze [0] (
02:22:52 Quit ashridah ("Leaving")
02:28:04 Quit Strath (Read error: 104 (Connection reset by peer))
02:28:21 Join Jungti1234 [0] (n=jungti12@
02:29:10 Join Strath [0] (
02:44:49 Join dretothew [0] (
02:45:43 Quit dretothew (Client Quit)
02:47:52 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
02:53:53 Join chris_hotwire [0] (
02:54:52 Quit chris_hotwire (Client Quit)
03:00:50 Quit DreamTactix291 (Read error: 110 (Connection timed out))
03:01:09 Join ataxy [0] (
03:13:36 Join DJDD_ [0] (
03:16:25 Part YouCeyE ("Leaving")
03:16:40 Join DrumRBoy320 [0] (
03:29:00 Quit actionshrimp ("a bird in the bush is worth two in your house")
03:49:53***Saving seen data "./dancer.seen"
03:50:13Jungti1234iriver G10
03:54:42 Part LinusN
03:57:08 Quit Maxime` (Read error: 104 (Connection reset by peer))
03:57:35 Join Maxime [0] (
04:04:43saa[b_r]idergood morning jungti
04:04:53Jungti1234hi :)
04:05:37saa[b_r]iderI just saw it! is it real, or fake?
04:06:20saa[b_r]iderI wonder how the controls will be
04:06:39Jungti1234I don't know
04:08:07saa[b_r]iderbefore I got my H340, I was thinking of getting the PMP
04:14:15 Nick DrumRBoy320 is now known as DrumRBoy320|Away (
04:23:07 Join livesNbox [0] (
04:24:23livesNboxhey guys... I have an h320 −− when I plug in the power adapter it boots into oem firmware. how can I make it automatically boot into rockbox when power is plugged in ?
04:27:41DrumRBoy320|Awaythat would require a change in the bootloader
04:27:59DrumRBoy320|Awayyou would most likelly have to "mod" it yourself
04:28:06DrumRBoy320|Awaywhich i wouldnt suggest
04:28:13DrumRBoy320|Awaysorry :(
04:28:38livesNboxheh.. boo..
04:28:52livesNboxkinda stinks because i leave it in my car with the car setting..
04:28:57livesNboxit powers down great.. perfect..
04:29:24livesNboxbut when I start the car back up.. it boots into iriver firmware.. then I have to pull the power cord.. turn the unit off.. then turn it back on with power out... then plug the power back in..
04:30:08saa[b_r]iderhow exactly do you use the car settings?
04:30:20livesNboxwhat do you mean?
04:30:30livesNboxi don't think it's called the car settings.. it has another name
04:31:07livesNbox"car adapter mode" it's in the general options
04:32:32saa[b_r]ideryeah.. so how do you benefit from this feature?
04:32:45saa[b_r]ideror what does it actually do?
04:32:49livesNboxwell... it pauses the unit when the power turns off to it
04:33:05livesNboxthen your idle settings will kick in (however long you have your idle set for) and turn the unit off
04:33:21livesNboxor if you get back in before the idle turns it off, it will resume playback 5 seconds after your car starts.
04:33:42livesNboxthe last missing link is... to boot back into rockbox and resume playback after the unit has idled off.
04:33:50livesNboxand the power comes back on
04:35:11saa[b_r]iderthat sounds cool... booting into RB when pluged into USB or an AC adapter is planned for the next bootloader, so don't worry about it
04:35:23livesNboxok cool.. where do you find out about that stuff ?
04:36:32saa[b_r]iderit was mentioned somewhere between the RB forum, MR forum, or this channel :D development for the H300 has been crazy for the past month, I forgot where exactly :)
04:36:49livesNboxheh.. ok...
04:37:00saa[b_r]iderbut I'm almost sure Linus wants this for the upcoming bootloader 4, if I'm not mistaking
04:37:06*livesNbox ignores the channel title...
04:37:08livesNboxany idea when?
04:37:34saa[b_r]iderkeep your fingers crossed!
04:37:35livesNboxby the way.. is this the same linus that's behind linux kernel ?
04:37:44livesNboxit's not is it.
04:38:19saa[b_r]iderso which charger do you use? is it a specific iRiver charger, a PSP charger, or a generic one?
04:38:25saa[b_r]iderin the car that is
04:39:42livesNboxgeneric one.
04:39:58livesNboxthen I wired it to the fuse panel.
04:40:15 Quit mikearthur (Remote closed the connection)
04:40:16 Join DreamTactix291 [0] (
04:40:28livesNboxi wanted one with a right angle plug on it.. so I had to get a generic one.
04:40:46saa[b_r]iderI was always affraid of frying my player with a cheap generic charger... and most of them came in 5V not 4.5
04:41:10livesNboxyeah. I found one with the right voltage.. I think it's actually one with variable voltage
04:41:59livesNboxno problems.. been using it for years...
04:42:05livesNboxI think I got it like at k-mart or something
04:42:13livesNboxwell.. maybe not years.. maybe like "a year"
04:42:44livesNboxanyway.. I guess I'll just keep my eye out for the next bootloader.
04:43:21livesNboxeven though nobody has any idea of where it is in development :)
04:43:28livesNboxat least, nobody I know.
04:45:44saa[b_r]iderI'll ask around
04:46:02livesNboxok.. if you find anything out.. I'd appreciate a memoserv or an email..
04:46:10 Join mikearthur [0] (
04:46:44livesNboxthanks man.. see ya later!
04:46:54saa[b_r]idertake care
04:47:01 Quit livesNbox ()
04:51:05 Quit Rick (Read error: 104 (Connection reset by peer))
04:51:52 Join Rick [0] (
04:55:04 Quit bagawk ("leaving")
05:00:33 Quit DreamTactix291 (Read error: 110 (Connection timed out))
05:02:46 Quit Matze ("Miranda IM! Smaller, Faster, Easier.")
05:03:30 Join Rob2222_ [0] (
05:04:17 Join ashridah [0] (
05:18:51 Quit Rob2222 (Read error: 110 (Connection timed out))
05:34:42 Quit darkless (Read error: 104 (Connection reset by peer))
05:40:56 Nick lostlogi1 is now known as lostlogic (
05:49:56***Saving seen data "./dancer.seen"
05:52:50 Join darkless [0] (
06:19:08 Join grimreap [0] (
06:19:39grimreapI have a mmc card that used to work fine with my ondio fm128, but I had to format the card and now the ondio doesn't see it, any ideas what I could do?
06:19:49grimreapwindows computers seem to be able to use it just fine
06:21:49Maximemaybe another FS ?
06:22:08grimreapI tried formatting fat16 and fat32
06:22:37 Join mymomthelush [0] (
06:26:52 Quit mymomthelush (Client Quit)
06:33:06 Quit grimreap (Read error: 104 (Connection reset by peer))
06:36:40 Quit Jungti1234 ()
06:41:50 Join DreamTactix291 [0] (
06:43:23 Quit DJDD_ ("Trillian (")
06:44:59 Quit ataxy (Excess Flood)
06:50:47 Join Bluechip [0] (
06:59:33 Join aliask [0] (
07:14:01 Join RotAtoR_ [0] (
07:15:46 Quit RotAtoR (Read error: 110 (Connection timed out))
07:19:42 Quit aliask ("CGI:IRC (EOF)")
07:28:46 Quit RotAtoR_ ("Leaving")
07:30:16 Quit DreamTactix291 (Read error: 110 (Connection timed out))
07:49:57***Saving seen data "./dancer.seen"
07:51:31 Quit saa[b_r]ider (Read error: 110 (Connection timed out))
07:54:34 Join JazzBone [0] (
08:06:04 Join Paul_The_Nerd [0] (
08:09:41 Join Bedlore [0] (
08:09:55Bedlorehi ppl :)
08:10:45Bedlorejust dropped in to ask about iriver mp3 encoding, I read it is unlikely to ever happen. is that still the case?
08:16:01ashridahread where?
08:17:03ashridahi don't think anyone's working on optimising an encoder enough to run at >realtime (which would be necessary to record on the fly), but i wouldn't say it's impossible to do (don't know about licensing issues tho)
08:17:17BedloreFAQ, I cant find it again if you like
08:17:47Bedlorehave you used ogg live encoding?
08:17:56ashridahbesides, i believe someone's got a lossless encoder working, there's just really no record+encode framework yet
08:18:02ashridahme? no
08:20:39 Join Jungti1234 [0] (n=jungti12@
08:22:39Paul_The_NerdYeah, there's currently a lossless wavpack encoder included as a "viewer" type plugin
08:23:14Bedloreyeh I saw that, I really need small files though
08:26:44Paul_The_NerdBedlore: Well, you could use the wavpack encoder viewer as a starting point, and port over a lossy format
08:34:05 Part Bedlore
08:46:31Jungti1234good morning
08:47:26*amiconn notices new target categories in the forum
08:50:45Jungti1234It's soap.
08:51:51 Join _FireFly_ [0] (
08:53:29 Join Kohlrabi [0] (
08:59:17 Join ataxy [0] (
09:06:02 Join ender` [0] (i=ychat@
09:08:35 Join Zagor [0] (
09:09:39 Join DreamTactix291 [0] (
09:26:29 Quit Jungti1234 ()
09:30:28 Join saa[b_r]ider [0] (n=saab_rid@
09:34:43 Join mymomthelush [0] (
09:35:47 Part mymomthelush
09:41:12 Part Paul_The_Nerd
09:46:31 Part Bluechip
09:50:01***Saving seen data "./dancer.seen"
10:01:04 Quit DreamTactix291 (Read error: 110 (Connection timed out))
10:08:47 Join amiconn_ [0] (
10:23:34 Quit Kohlrabi ("Leaving")
10:27:24 Quit amiconn (Read error: 110 (Connection timed out))
10:27:24 Nick amiconn_ is now known as amiconn (
10:42:25 Join aliask [0] (
10:42:42aliaskSlasheri: Are you around?
10:44:10 Join Jungti1234 [0] (n=jungti12@
10:45:06aliaskOr in fact, anyone who knows. Is there a way to pull the modified and created time for a specific file out of the datacache?
10:46:57aliaskI thought I heard slasheri say that it was in there.
10:47:55Slasherialiask: yes, it is in the same structure you get with standard readdir
10:48:29Jungti1234 - Can you see it?
10:48:32preglowSlasheri: so, how's tagcache doing?
10:48:55Slasherialiask: wrt_time and wrt_date.. in fact there is no created time
10:49:13Slasheripreglow: i haven't had time for it in the last few days but i will look it soon again
10:49:26preglowgot a patch lying around?
10:49:35aliaskThanks Slasheri.
10:50:11Slasheripreglow: Hmm, not a really working patch at the moment. But if you wish to test, please ask amiconn for a patch with small fix for archoses
10:56:29 Join Bger [0] (n=Bager@
10:57:00Bgermorning and happy new year 2 all :)
10:57:27markunalso to you Bger
10:57:51Bger10x ;)
10:58:20 Join DangerousDan [0] (
10:58:46 Join LinusN [0] (
10:59:24Bgerhm, question to all h100/h300 owners : what do you think about moving the long "joy"(navi) push in WPS to the REC key, and to map the long joy/navi push to "view current playlist" ?
11:00:27aliaskI'm in favour, but because our player is fairly useless without the backlight, how can we just turn the backlight on?
11:00:27Jungti1234hi Bger
11:00:54Bgeraliask that's another question :)
11:00:58Bgerhi, Jungti1234
11:01:11BgerLinusN ?
11:01:13Bgeramiconn ?
11:01:17BgerBagder ?
11:02:29Bgeraliask regarding this, i think about a new setting (and maybe a specific button only, that is only this button will "eat" the first press for *only* turning the backlight on)
11:02:48LinusNlong navi push means "context menu" in all screens, i don't want to change that
11:03:06aliaskA good idea. The hold button does put the backlight on though.
11:03:17aliaskMaybe just have the record button bring up the playlist.
11:03:27aliaskIt would cause less confusion.
11:03:44BgerLinusN but in wps imho short navi -> tree ; long navi -> current playlist
11:03:51Bgeris more intuitive
11:04:31BagderI'm with LinusN here
11:04:34preglowme too
11:04:48Bgerthan what about mapping the rec to "view current playlist" ?
11:04:50LinusNthe long navi push is about the only thing that is consistent in rockbox
11:04:59preglowexactly, heh
11:05:05Bgerhehe, okay
11:05:15Bagderbut consistency might be seen as inconsistency? ;-)
11:05:21preglowif we're going to change our consistency, i don't think "less consistent" is the way to go
11:06:25Slasherii think it would be better to have just "actions" and user could create the menus as he/she wish
11:07:42Bgerthe reason i suggested this (the long navi push) is because the "view current playlist" is too far as it is now
11:07:47Bgerthat is, in button presses
11:08:06Bgerwell, at least imho
11:08:47LinusNBger: every user has his own opinion about which function should be on a quick-access button
11:09:40BgerLinusN yeah, you're right, and maybe the Slasheri's suggestion is the way to solve this ...
11:09:42LinusNcheck the patch tracker for instance, several "<my favourite function> on <my favourite button>" patches
11:10:14LinusNi don't think full configurability is always a good think
11:10:17BgerLinusN while on this topic, what do you think about mapping the h100's remote buttons 1:1 to the buttons on the main unit?
11:10:37LinusNBger: can it be done?
11:10:56LinusNfull configurability means two bad things:
11:10:59Bgerwhy? something like WPS - "good" built-in default and a way to change
11:11:09LinusN1) much more complicated code
11:11:26LinusN2) a nightmare for the support
11:11:37Bgerheh ...
11:11:49LinusN(support meaning helping people when things don't work as they should)
11:12:09Bgeri don't mean *full* configurability
11:12:21Bgerbut a way for user to edit one action menu
11:12:33Jungti1234See it.
11:13:16Bgerlike making the REC button on H[13]00 "action menu"
11:14:08Bgerwas this "<LinusN> Bger: can it be done?
11:14:14Bger" about the h100's buttons ?
11:17:49 Quit Jungti1234 ()
11:17:52Bgerat least to the level h300's LCD remote maps to the h300's main unit ...
11:21:13Bgeri mean if we want to support iriver remotes runtime, we have 2 possibilities: 1) making h100 and h300 remotes' buttons function equal (i.e. play on h100 does the *same* thing as on h300) 2) using a "mapping" button function
11:21:22Bgeror at least i see this 2
11:21:32 Quit ataxy ("Chatzilla [Firefox 1.5/2005111116]")
11:22:55Bgers/support iriver remotes runtime/ support iriver remotes detection/changing runtime
11:23:55BgerLinusN ?
11:24:40LinusNyes, it would be nice if they were as similar as possible
11:24:48Bgeranother q: do we need to initialize the LCD screen in the rockbox itself (isn't the bootloader init enough)
11:25:55BgerLinusN the problem is that 1:1 button mappings on h100 remote are not so good as for h300 ...
11:26:29LinusNthen it shouldn't be done on the h100
11:26:53Bgerthen ...
11:26:57 Join ataxy [0] (
11:27:02Bgerthen what ?
11:27:41Bgergoing on 2nd possibility ?
11:31:37saa[b_r]iderHi Bger :)
11:31:45Bgerhi, saa[b_r]ider
11:32:13saa[b_r]iderbtw, I noticed something about the H300 remote patch
11:32:28Bgeryes ?
11:32:36saa[b_r]ideryou know how RB now goes directly into the record screen?
11:32:45saa[b_r]iderat startup that is
11:33:34saa[b_r]iderwhen this screen is displayed on the main LCD, the remote LCD simply displays a "Rockbox" logo
11:33:35Bgerhm, reset your settings
11:33:54Bgerthat's not this patch's problem...
11:34:46Bgerhuh, see, it should be doing the same without this patch too
11:35:07 Quit ataxy ("Chatzilla [Firefox 1.5/2005111116]")
11:35:44saa[b_r]iderbasically, when you enter the recording screen, the screen isn't updated on the remote, and the remote buttons do nothing, untill you exit the recording screen
11:37:00_FireFly_yeah because the recording screen isn't yet ported to display also on the remote-lcd
11:37:20saa[b_r]iderso the devs are aware of this already?
11:37:32Bgersully me, i started to look at the apps/main.c :)
11:38:19saa[b_r]ider:) Bger: what would happen if I reste my settings?
11:38:28amiconnsaa[b_r]ider: The recording screen is not yet adapted to the multi-screen gui api
11:38:46Bgerif _FireFly_ is right (and i'm pretty sure he is) than reseting won't help you with anything
11:38:47*amiconn wonders whether TiMiD will return one day
11:39:12Bgeramiconn iirc he is in japan currently
11:39:59saa[b_r]iderBger: yeah I understand that it hasn't been ported to display on the remote LCD. I thought that there was a way for the player not to directly load into the recording screen
11:40:11Bgersaa[b_r]ider there is
11:40:32Bgeryou can change it in Settings->recording->recording settings or something similar
11:40:41preglowamiconn: more duplicated gui code left?
11:41:01 Join muesli- [0] (n=muesli_t@
11:41:33Bgersaa[b_r]ider did you find it ?
11:41:52amiconnpreglow: Yes, there are some more screens that are not yet adapted
11:41:53 Join DreamTactix291 [0] (
11:43:28saa[b_r]iderBger: I'll check it out
11:45:01saa[b_r]iderok, much better :)
11:46:24saa[b_r]idernow I don't have to take out the player everytime I turn it on using the remote!
11:47:16LinusNi wonder why so many people have this problem
11:47:35saa[b_r]iderLinus: what problem?
11:47:46Bgerwith recording screen on startup
11:48:00LinusNthat setting defaults to off
11:48:17LinusNsaa[b_r]ider: are you running a patched rockbox?
11:48:19Bgersome patch maybe ...
11:48:29saa[b_r]iderI compile my own build..
11:48:30Bgeryes, he is running my remote patch for sure ...
11:48:37Bgermaybe also other ones
11:48:40saa[b_r]ideryeah, no other patches
11:48:47Bgerno other ? strange ...
11:48:55LinusNBger: better bump the settings version then
11:49:21BgerLinusN but this setting is at the end ...
11:49:33 Join matsl [0] (
11:49:55Bgersaa[b_r]ider other patches recently ?
11:50:04***Saving seen data "./dancer.seen"
11:50:04saa[b_r]iderBger: I still don't know why I had to add the tetrox source to compile after using your patch......
11:50:16Bgeri'll take a look at it
11:50:34saa[b_r]iderI use sources, I don't use CVS
11:50:46Bgersaa[b_r]ider ah, yes, my error
11:51:15saa[b_r]iderhehehe... at least we spotted it :)
11:51:24saa[b_r]idertetrox is cool though huh?
11:52:06Bagderthat one of Eli's things?
11:52:09Bgerbtw, the rockbox's calendar plugin has some buffer overflows (or i'm really dumb)
11:52:11Bgeryes :)
11:52:27Bagderanyone know if he's gonna submit them to us?
11:53:03BgerBagder do we need his blessing to do it ?
11:53:09saa[b_r]idereli and paprica say they're constantly updating, and they'll submit when they're happy with their plug-ins
11:53:25BagderBger: not if they are GPL, other than plain courtesy no
11:53:39Bagdersaa[b_r]ider: ok, thanks
11:53:45saa[b_r]idernot that they don't want to submit
11:54:35saa[b_r]idereli is currently working on a new viewer for bmp, GIF and jpeg.. he wants people to help too
11:55:24Bagderthat's a reason for submitting...
11:55:53Bgeryes :)
11:56:20Bagderhe can't expect people to help before the code is available in CVS or tarballs imho
11:56:58saa[b_r]idermaybe you guys can submit this, it sounds promosing :)
11:57:30saa[b_r]iderhe calls it "Viewbox"
11:57:59Bgerregarding calculator.c : static bool wday_has_memo[6];
11:58:13Bgerfor (k = 0; k < 7; k++)
11:58:14Bger wday_has_memo[k] = false;
11:58:28Bgerand similar
11:59:55 Quit _FireFly_ ("Leaving")
12:00:23LinusNeli doesn't seem to be the most cooperative person
12:02:01LinusNhe says he'll submit the patch "when it's ready"
12:02:20Bagderabout the same time Rockbox is ready then
12:02:26Bgerprobably ...
12:02:45LinusNso all of the plugins will probably remain on their plugin web page until we all grow old
12:02:47saa[b_r]iderhe has been asked before if he was going to submit, and he has said that eventually he would. He has added made quite a few contributions... and his most recent work shows that he's all for help. plus he has said he barely has time since he's in the army and all
12:03:12Bagderthen he should submit them instead
12:03:15LinusNall he has to do is submit the patches
12:03:47 Join DJDD_ [0] (
12:03:54saa[b_r]iderlet's try to convince him then :) he hasn't been active recently, I'll keep an eye for him
12:04:31LinusNdo so, i won't bother
12:04:35BgerBagder what happened with the bmp2rb thing in the build system
12:04:56Bagderask linuxstb, he works on it
12:05:00LinusNnothing happened, to my knowledge
12:05:29saa[b_r]iderLinus: have you spoken to him before? I have suggested to him before to join this channel to get in touch with you guys
12:05:38LinusNhe has been here
12:07:05saa[b_r]ideryeah I saw him once or twice, but has he said anything that indicated he didn't want to cooperate?
12:07:43Bgeri don't remember such thing ...
12:08:28saa[b_r]idergood. so I'll just try to make him understand how it would help to submit
12:08:47Bgerdo so, please
12:08:56saa[b_r]iderBger: did you get my query?
12:09:00LinusNhe hasn't said he didn't want to cooperate
12:09:29LinusNhe has said that he won't submit the patches until the plugins are "finished"
12:09:52saa[b_r]iderthat pretty much sounds like him :)
12:09:55LinusNand that isn't the open source way imho
12:10:03Bgersaa[b_r]ider if you mean privmsg, i didn't
12:13:54saa[b_r]iderbtw, what's the status of RB loading when an AC adapter or USB is plugged?
12:14:08Bgeri.e. ?
12:14:25LinusNif he really wants the plugins included in rockbox, he should work *with* us, not do his own race and then dump a shitload of "finished" sourcecode on us to clean up
12:15:36saa[b_r]idercalm down, calm down :) why can't all the developers just get a long :D
12:16:54LinusNstill, he's entitled to develop his plugins like he's doing
12:17:13LinusNthat's after all one of the reasons for the plugin interface design
12:17:25LinusNto allow for 3rd party plugins
12:17:40 Quit muesli- ("ich will Kühe!!!")
12:18:38 Quit matsl ("ERC Version 5.0 (CVS) $Revision: 1.776 $ (IRC client for Emacs)")
12:20:38saa[b_r]iderLinus: so what's the status on loading RB on AC adapter or USB being plugged?
12:21:06 Join muesli- [0] (n=muesli_t@
12:26:28saa[b_r]iderbtw, this pong compiles with no problem...
12:28:23saa[b_r]ideropps wrong window
12:29:43 Quit DreamTactix291 (Read error: 110 (Connection timed out))
12:37:15LinusNsaa[b_r]ider: loading rb on AC adapter shouldn't be hard at all
12:37:45*Bger whispers "bootloader usb...."
12:38:30saa[b_r]iderit's planned for bootloader 4 isn't it?
12:38:54BgerETA ?
12:38:59*Bger shuts up
12:40:52BgerLinusN i asked you before ... what about not (re)initializing the LCD in the rockbox, and only doing so in the bootloader
12:41:41Bgerbecause ... it's visible...
12:42:10Bgerand is about .2 sec ? ;)
12:42:14LinusNtha philosophy behind this is that rockbox shouldn't depend on the bootloader
12:42:37Bgerbut on iriver targets it depends ...
12:42:52LinusNunfortunately, yes (if you mean the uda1380 reset)
12:43:18Bgerah, no, i just wanted to say that we can't boot rockbox without bootloader
12:43:31 Join BalanceST [0] (
12:44:02LinusNof course, i just mean that rockbox shouldn't depend on what the bootloader does with the hardware
12:44:25LinusNbut it still does in several ways...
12:44:34LinusNnothing is perfect
12:45:03Bgeralso ... will an output to the remote in the bootloader add much code size ?
12:46:15 Quit Rob2222_ (Read error: 104 (Connection reset by peer))
12:46:33Bgerbecause as it is now ... the display is blank up to rockbox logo ...
12:47:00LinusNno, it won't increase code size much
12:47:24Bgerare you against this then ?
12:47:51LinusNwell, it's against the do-as-little-as-possible approach, but it's pretty annoying when starting from the remote
12:48:16Bgeryes, the annoying part is the reason i'm asking for it ...
12:48:28LinusNso i guess it should be added
12:48:46 Part BalanceST
12:49:00*Bger hugs LinusN and hands him a beer
12:49:29 Join ST [0] (
12:51:27preglowdo-as-little-as-possible is a good strategy, but i think remote support is a must
12:52:40 Join Rob2222 [0] (
12:53:08LinusNi want to fix the remote autodetection
12:53:33 Join An [0] (n=vdgraya@
12:53:35STis there goin to be non-lcd remote support for the h3x0
12:53:38preglowany idea how to approach it?
12:54:07Bgerafaik there is a way to distinguish h300 non-lcd from any lcd ...
12:54:29markunwell, the iriver firmware can, can't it?
12:54:34 Join Moos [0] (
12:54:46Bgeryes, it can for sure
12:54:49Bgerso there is a way
12:55:29LinusNyes, of course
12:55:50BgerLinusN while on this topic ... to ask again about the remotes :))
12:56:10LinusNabout remapping the buttons?
12:56:10Bgerhow do you suggest to work around button issues ?
12:56:24LinusNi haven't really thought about it that much
12:58:04Bgerwhat about a function that has a param like (bool raw) ... if (raw) returns raw button else returns something like UP; DOWN; LEVEL_UP (EXIT ONE LEVEL); LEVEL_DOWN (ENTER ONE LEVEL DEEPER) (virtual) and this func remaps the remotes ?
12:58:11 Quit ST ()
12:58:28 Join Voker57 [0] (n=Voker57@
12:58:34Bgerthis should add button consistency ...
12:58:35 Join ST [0] (
12:58:51Bgerin the menus, tree ...
12:58:55saa[b_r]iderhi An
12:59:54preglowarm code size will become a problem for sure
13:00:01preglowwe really should start looking at thumb code
13:00:04preglowi wonder why apple doesn't use it
13:00:05Bgerbad ... :(
13:00:06 Quit edx (Read error: 110 (Connection timed out))
13:00:16LinusNBger: that's almost what we are trying to do with all the #defines
13:00:20 Quit Rob2222 (Read error: 104 (Connection reset by peer))
13:00:33 Part Voker57 ("Virca 1.1.20")
13:00:40Bgerexcept that it's not runtime ...
13:01:00Bgerand if we want remote switching runtime ...
13:01:45STvirtual HDD LED for WPS isnt in the daily builds right?
13:01:55LinusNfor wps?
13:01:57BgerST no, it isn't
13:02:44 Quit ST (Client Quit)
13:02:50 Join ST [0] (
13:03:39 Part An ("Virca 1.1.20")
13:03:48STok thx
13:05:40STwhose tried out the Color Changer patch by Sanitarium ( ?
13:07:00saa[b_r]iderST: I read it's buggy... but it works
13:07:59STye i like, just needs some more work
13:08:47LinusNsaa[b_r]ider: how is it buggy?
13:09:10STthe settings arent saved on reboot of the player
13:09:17LinusNST: i just committed the virtual led patch
13:09:18STsettings for the colours that is
13:09:22LinusNST: how lame
13:09:32saa[b_r]iderI haven't tried it my self, but I read it doesn't work with all WPS screens
13:10:26STLinusN: kewl, thx
13:10:47saa[b_r]iderLinus: all the problems were reported here:
13:10:53STsaa[b_r]ider: it works with my wps, but ye those bugs are annoying,
13:11:20aliaskOn the topic of patches, I made one and submitted it to the tracker. It's been sitting there for a few days, and it's a bugfix. If not committed I'd like some kind of feedback on it if possible.
13:11:40saa[b_r]iderI compile my own modified lcd.h with my own RGB values... and I'm fine with that for now :)
13:12:20LinusNaliask: which one?
13:13:06 Join Jungti1234 [0] (n=jungti12@
13:13:06aliaskFor the Yesno screen, on iRiver (H300 at least), it said that "ON" was to reply yes, but in fact it was Navi/Joy press.
13:13:32LinusNaliask: i can't decide if i want "yes" to be Play on the iriver or not...
13:13:58STaliask: hehe yes, i thought the delete wouldnt work for a while
13:14:04Bgeraliask me too
13:14:10aliaskWell, the other option is to change the lang file, but then it becomes platform specific, which the lang file doesn't support.
13:14:25aliaskAha! :)
13:14:26BgerBagder ... :)
13:15:20BgerLinusN so back to my favourite topic ... :) what do you suggest? i really want some ideas/guidelines to start working on it
13:15:43LinusNthe translate() function sounds like a good idea
13:16:17Bgeronly for iriver targets or ?
13:16:49 Join Rob2222 [0] (
13:19:29LinusNthat's the only target with multiple remotes, but i guess it's cleaner to do it on all platforms
13:19:50LinusNmaybe it could be a no-op macro on the other platforms
13:20:26aliaskIs it possible to get a .rock to build as a viewer without having an association in viewers.config?
13:20:49Bgerso what do you suggest ? button=translate(get_button_w_tmo()) ?
13:22:38*ST np: Usher - Think Of You [03:51m/128Kbps/44KHz]
13:22:48 Quit AciD (Read error: 104 (Connection reset by peer))
13:23:53*ST np: Usher Ft. Fabolous - Caught Up (Remix) [04:34m/192Kbps/44KHz]
13:24:28 Join _FireFly_ [0] (
13:25:04 Join PaulJ [0] (
13:25:14LinusNBger: something like that, yes
13:26:05LinusNBger: or maybe do the translation in the button driver
13:27:25Bgerin the button driver will be better imho
13:28:12Bgerbut as another func or as current one but with one additional param ?
13:29:18LinusNno, do the translation in button_read()
13:31:43Bgerhuh.. but this will affect *everything*
13:32:14Bgeri already do translation there but only in aspect of adc_read->real_button
13:33:09Bgerbut this is the low_level mapping
13:33:32Bgeri'm talking more about real_button->action mapping
13:34:25Bgermaybe action=button_translate(button_returning_function) is the best in this case
13:34:37LinusNwhat is an "action"?
13:34:57Bgerlike "mark the item above" go up one level (in tree, in menu)
13:35:28Bgerthat is the action rockbox perform when you press the button
13:35:36LinusNok, that would of course have to be done locally
13:35:56LinusNthere's no way to centralize that, is it?
13:36:16Bgerbut i think the mappings in the menu and in the tree should be the same, or i'm wrong
13:37:13Bgerwe were talking about consistency ... :)
13:38:16 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
13:38:30 Quit DJDD_ ("Trillian (")
13:41:27 Join DJDD_ [0] (
13:43:16 Join Matze [0] (
13:50:05***Saving seen data "./dancer.seen"
13:50:54 Quit _FireFly_ ("Do fish get thirsty?")
13:55:58 Join Pi [0] (
13:57:31 Join webguest86 [0] (
13:57:50 Quit webguest86 (Client Quit)
13:59:15*ST np: Usher ft. Fabulous - Caught Up (Remix) [00:00m/0Kbps/0KHz]
14:01:05 Join linuxstb [0] (
14:05:44*ST np: Kazaa Mp3z,Pix Etc# - Green Day - Boulevard Of Broken Dreams [04:20m/245Kbps/44KHz]
14:05:46Bgerhi, linuxstb ;)
14:05:49preglowamiconn: you did get a thumb build working, yeah?
14:06:52linuxstbHello from Boston, MA.
14:06:53 Join _FireFly_ [0] (
14:07:17Jungti1234np = now playing?
14:07:50 Quit Febs (Read error: 110 (Connection timed out))
14:07:53aliaskI've had my fill of unpredictable output for tonight. G'night.
14:07:59 Quit aliask ("Chatzilla 0.9.69 [Firefox 1.5/2005111116]")
14:08:08preglowlinuxstb: hello from hellhole, norway
14:08:37 Quit ST ()
14:09:14linuxstbHow's the weather? They talk in strange F numbers over here. But I know it's cold.
14:09:54linuxstb(relative to London - it's probably hot compared to trondheim)
14:10:18Bagder-8C here atm
14:10:21Bagderlots of snow
14:10:23preglownah, it's pretty hot up here
14:10:25pregloweven rained yesterday
14:10:29preglowwhich explains my discontent
14:10:41preglowlots of snow at home, but i'm not at home
14:10:43 Join DreamTactix291 [0] (
14:10:48Jungti1234very cold
14:12:15PaulJhow can you login annonymous in the wiki? the login TWikiGuest/guest does not work.
14:12:21preglowwhy not register?
14:12:29preglowi think you have to anyway
14:13:24PaulJi don't want to register everywhere, and i don't want to use my realname as username
14:13:33LinusNtoo bad then
14:13:42preglowthen you can't use the wiki
14:13:50PaulJok, thank you
14:14:07 Join edx [0] (
14:14:54linuxstbYou can read the wiki without logging in, you just can't edit it anonymously.
14:15:12preglowwell, yeah, s/use/edit/
14:15:14PaulJmaybe someone should cange the errorpage when you cancel the login. it shows: Note: You can also login anonymously as user TWikiGuest with password guest.
14:15:30preglowgood idea
14:16:01preglowbut no, we have enough wikispam as it is, anonymous logins wouldn't help
14:16:24PaulJyes, i can understand that
14:18:20linuxstbpreglow: Are you working on anything ipod-wise at the moment?
14:19:59amiconnpreglow: I didn't try; I just tried to build binutils + gcc with thumb-interwork suppoort
14:20:11amiconnThis *seems* tp have worked
14:20:52amiconnI have no idea how to try a thumb target build, and of course I have no target
14:21:25preglowlinuxstb: no, i'm pretty much bogged down with work for some time to come
14:21:55preglowamiconn: so, all you did was pass some interwork flags to configure?
14:22:38amiconnThe strange thing is that my /opt/arm folder is exactly the same size as before
14:22:50preglownow really
14:22:57preglowif you did a thumb build of rockbox, something apparently did work
14:23:08preglowi'll try it myself when i have time
14:23:14preglowthumb code might be very beneficial for us
14:23:25preglowgives better cache usage, and afaik, ipod only has a 16 bit data bus
14:23:52amiconnI did *not* try a target build with thumb code, as I don't know how
14:24:25preglowahh, right, i thought you said you did
14:24:41preglowyou just need to add -mthumb and -menable-interwork, i think
14:24:46preglowpluss add some asm directives here and there
14:26:06Bagder-menable-interwork should only be needed if we use parts that are not thumb, right?
14:26:40preglowand we will
14:26:55preglowfor one, the 64 bit mac instructions are not available in thumb code
14:27:05preglowand ordinary arm code is faster if you code by hand
14:28:03Papricawhy the MAX_MENUS define to 5?
14:28:12Bagderwhy not?
14:28:29Papricamm if i add a sub menu to sub menu
14:28:35Papricait give me an error
14:28:46Paprica"Out of menus!"
14:28:52linuxstbThen increase it.
14:28:52preglowBagder: actually, none of the fancy math functions are in thumb
14:29:03Papricai did that
14:29:20linuxstbDid that fix the problem?
14:29:24Papricabut i want to know what is the reason
14:29:29BagderPaprica: it would be silly to have arrays bigger than we need
14:29:34linuxstbBecause everything needs a maximum
14:29:46*preglow can see a possible malloc debate coming
14:29:54Papricais i set it to 10
14:29:58Papricait takes more memory?
14:30:05amiconnpreglow: The "pluss add some asm directives here and there" is the interesting point
14:30:12Papricaohh =\
14:30:25preglowamiconn: how? it's dirt simple, you just add .code 32 or .arm at the top of ordinary arm code
14:31:14amiconnYes, just that I don't know in what places the arm asm hides (other than crt0.S)
14:31:48preglowcrt0.S you wont have to change, all exception handlers run in arm asm
14:31:51preglowarm mode, i mean
14:31:55preglowyou need to change thread.c
14:32:01preglowand that's about it
14:32:37amiconnWhy not crt0.S, then?
14:32:55preglowbecause that is exception handler code
14:33:25preglowplus stuff that gets called by bootloader
14:33:31amiconnYes, and that means they _have_ to be marked as code32 iiuc
14:33:55preglowcode32 is default, so the entire file will be that
14:34:25 Quit ashridah ("Leaving")
14:34:38amiconnAre you sure that this is still valid when gcc is called with -mthumb?
14:34:40preglowthe bootloader will need to know what type of code the entry point is
14:34:45preglowpretty much
14:35:47 Join ST [0] (
14:36:01preglowto switch to thumb mode, you need to use a special branch instruction that switches to thumb mode if the lower address bit is 1
14:36:19preglowso the image bugger in the bootloader needs to be properly aligned
14:36:48preglowso we don't get any accidental thumb mode switches, but i bet gcc already takes care of this
14:36:59preglowforget it, non-aligned accesses are legal on arm
14:37:13preglowaren't, god damn, i type worse than usual today
14:38:15 Quit DJDD_ ("Trillian (")
14:43:24 Quit linuxstb ("CGI:IRC (Ping timeout)")
14:44:45 Quit JazzBone ("Leaving")
14:58:24 Quit PaulJ (Read error: 110 (Connection timed out))
15:01:09 Quit DreamTactix291 (Read error: 110 (Connection timed out))
15:01:47 Join Kohlrabi [0] (
15:11:54 Quit Kohlrabi ("Leaving")
15:12:11Jungti1234hi Paprica
15:12:37Papricawhats up?
15:13:11Jungti1234I want that change tetrox key.
15:13:25Papricaits not a big problem
15:13:39Papricayou cant do it yourself?
15:14:00Jungti1234I can't program.
15:14:04Papricayes what? you cant o r you can
15:14:18Jungti1234I can't
15:14:30Papricawhat u want to change?
15:14:55Jungti1234Tetrox is no hold function.
15:15:57BgerPaprica i think he means something like pause
15:16:46Jungti1234I'm going to find picture.
15:17:03PapricaBTW, your ipod wps is nice
15:17:22Papricawith my status bar it is very nice ;]
15:21:28muesli-hi Jungti1234
15:21:35Jungti1234hi muesli- :)
15:22:32PapricaJungti1234, what do you mean
15:22:36muesli-what i always wanted to do you pronounce seoul? like soul oder se-oul?
15:22:37Papricain this game
15:23:31muesli-your capital?
15:24:11muesli-how do you pronounce it
15:25:17muesli-woh..never heard that version in the news
15:25:37PapricaJungti1234 do you mean to the sadows?
15:26:26 Join georgeblunt [0] (n=georgebl@
15:26:26muesli-they say it like soul or se-ul
15:27:09Jungti1234It stores bricks for a moment.
15:28:08Jungti1234muesli-: Many people pronounce it that is se-ul.
15:28:55Jungti1234For a moment, I go to other place.
15:32:24 Quit ST ()
15:32:34Jungti1234I came back.
15:32:58Jungti1234It stores bricks for a moment.
15:33:38Papricaim dont understand
15:33:51Papricawhere it store them
15:38:36Jungti1234Paprica: You can know though you play it.
15:43:35preglowthe code gcc generates for arm truly doesn't look very good
15:50:07***Saving seen data "./dancer.seen"
15:50:16Bagderyou use any arch option?
15:50:40Bagderiirc, by default it makes some very-old-arm-arch choices
15:51:13Bagderlike doing 2 8bit stores instead of a single half-word store when writing 16bits
15:51:36Bagderat least that one hit me at work
15:54:29preglownothing out of the usual
15:54:36preglowjust looking at disassembled rockbox code
15:57:52 Quit georgeblunt (" HydraIRC -> <-")
15:57:53 Nick Paprica is now known as Paprica[bbl] (
15:58:05preglowhaving to always use separate instructions for loads and stores seems to increase code size quite a bit
15:58:47 Join hshah [0] (
15:59:13Bagder-march=armv4t shouldn't hurt at least
15:59:41Bagdernot that I know if it adds anything
15:59:50preglowwell, -march=arm7tdmi should be even more specific, no?
15:59:50Bagderto the already used -mcpu=arm7tdmi
16:00:06preglowi'll try
16:01:16Rob2222I think I found the difference to recognize the H3xx remote types. :)
16:01:50Bagderthe linux Makefile for ARM uses -mapcs -mno-sched-prolog too
16:03:12preglowBagder: -mapcs might be nice
16:05:20preglow-march=armv4t did nothing at all
16:06:30preglowalmost all the bloody code is ldr and str
16:06:49preglowsounds like something asm optimising might fix...
16:10:03 Join DocterD [0] (
16:11:04 Join Kohlrabi [0] (
16:12:44 Quit Jungti1234 ()
16:13:57 Quit Kohlrabi (Read error: 104 (Connection reset by peer))
16:14:01preglowRob2222: lets hope it works on other players as well, then
16:17:52 Join Kohlrabi [0] (
16:19:33 Quit _FireFly_ ("Leaving")
16:20:08 Quit Kohlrabi (Client Quit)
16:22:17 Quit DocterD ("Leaving")
16:33:06Lynx_how fast should file transfers to the h340 be with usb 2? i get about 4-5 mb/sec
16:39:08Bgeri get about 8-10mbytes/s
16:41:32 Join DreamTactix291 [0] (
16:42:44Lynx_hmmm, ok
16:42:53 Nick Lynx_ is now known as Lynx_awy (
16:44:37 Quit saa[b_r]ider (Read error: 110 (Connection timed out))
16:44:38 Quit joshn_454 ("KVIrc 3.2.0 'Realia'")
16:53:26 Join _FireFly_ [0] (
16:55:23 Join matsl [0] (
17:05:34 Join saa[b_r]ider [0] (n=saab_rid@
17:07:45 Join DangerousDan [0] (
17:13:25saa[b_r]iderhello everyone
17:14:12saa[b_r]iderhow's the remote coding going Bger?
17:14:39Bgernothing new...
17:15:42saa[b_r]iderwas the wiki link above made by Rob2222 actually helpful to your project?
17:17:29 Quit Zagor ("Client exiting")
17:20:53 Quit Nibbler (Read error: 113 (No route to host))
17:21:03 Join Nibbler [0] (
17:21:53Bgerhm, i have some ideas (they are perlplexity's ideas in fact)
17:22:05Bgerbut we'll see next days
17:22:14amiconnGPIO1 bit 20 is the lock sitch bit for the H1x0 remote, so this isn't the final solution
17:22:42markunpreglow: what happened to the crossfeed improvements?
17:22:55amiconnIdeally, we should be able to handle all 3 remotes (H1x0, H300 LCD and H300 non-LCD) on both targets (H1x0 and H300)
17:23:51Bgeramiconn no reason of not being capable to do it on h100 once we do it for h300
17:25:25preglowmarkun: jlo gave me a blind test, then vanished
17:26:25markunSo you don't even know which of the implementations came out best?
17:27:40preglowi don't even remember what results i had, it's been a couple of weeks
17:32:12 Join actionshrimp [0] (
17:34:47Rob2222Oh damn, right @ remote ... GPIO1_READ changes cause the HOLD button...
17:35:29Rob2222Well, I reverb the WIKI
17:36:31Rob2222Well, cant we make the remote selectable by the config area until we got it finally automatically?
17:36:43 Join NicoFR [0] (
17:40:42 Join Kohlrabi [0] (
17:45:41 Quit Kohlrabi ("Leaving")
17:50:09***Saving seen data "./dancer.seen"
17:53:43 Join Krazyjim [0] (
17:58:35amiconnPerhaps it would be interesting to check how iriver distinguishes the remotes
18:03:00 Join YouCeyE [0] (
18:08:14mirakpreglow: was it you who wanted to buy a H300 ?
18:08:32preglowhell no
18:08:40saa[b_r]iderhe's in love with his nano :)
18:08:47preglowi've already got the super sweet h120
18:09:23mirakpreglow: does the nano have an equivalent cpu than other ipods ?
18:09:29mirakor is it lighter ?
18:09:33preglowsame thing
18:09:39preglowit's even more powerful than most of the others
18:09:44preglowsame cpu as in the video
18:09:46*saa[b_r]ider stands corrected...
18:09:56mirakI am interested in adapting libmpeg2 like I did for xvid
18:10:04preglowsaa[b_r]ider: oh, i do love my nano
18:10:11mirakI got to be motivated
18:10:19preglowis that the ffmpeg one?
18:10:19mirakan external motivation would be a plus :p
18:10:24saa[b_r]iderpreglow: what color is your H120?
18:10:29preglowsaa[b_r]ider: silver
18:10:40mirakpreglow: I think it's an external project but maybe ffmpeg uses it
18:10:42mirakI don't know
18:10:44preglowmirak: well, don't look to me for that, i don't much care about video
18:11:14preglowdone any more work with the xvidcore thing, then?
18:11:19mirakunfortunely you are good at optimising too :(
18:11:39mirakthe poor performances didn't motivated me
18:11:51preglowthat's the part that would have motivated me, hehe
18:12:03preglowunfortunately, i don't even have time to code on my nano at the moment
18:12:08mirakI tried to check what could be but in iram, but I think you said nothing much could be done
18:12:16preglownah, i haven't said that
18:12:23preglowi haven't even looked at it
18:12:38NicoFRand if you clean the code ?
18:12:38mirakok so I looked at it and,
18:12:54mirakNicoFR: the code fits in plugin_ram
18:13:03mirakthat's not the issue anymore
18:13:13mirakissue is about decoding speed now
18:13:37mirakso one thing would be to have an imdct in assembly
18:13:52NicoFRand would a JPEG decoder motivate you by any chance ?
18:13:55mirakarm have some imdct in ffmpeg
18:14:04NicoFRor is someone already working on that ?
18:14:05mirakNicoFR: I think there is already one
18:14:19NicoFRthere is one but it uses the grayscale lib IIRC
18:14:41mirakI would put effort in xvid or mpeg2
18:15:03mirakanyway any video codec uses a bidimentional imdct
18:16:18mirakwell my goal was to have video at more than 10fps like the iriver firmware
18:16:58mirakxvid and mpeg4 can't do that imho on this device.
18:17:01NicoFRmaybe it would be possible with a less cpu-intensive format ?
18:17:15markunanimated gif! :)
18:17:18mirakmpeg2 is less intensive they said
18:17:54lostlogicdo we have enough RAM to buffer a less compressed format?
18:17:57mirakeven with mpeg2 we would still need an imdct in assembly. and that's the most difficult thing probably to do
18:18:02 Join [1]Krazyjim [0] (
18:18:31miraklostlogic: yes, we would need to load more data from hard drive
18:18:52mirakbut I can watch 6 hours of xvid movie with normal firmware
18:19:19markunWould be nice if we could make the optimized imdct work for all the codecs. Everyone has their own fixed point format, that makes it a bit difficult.
18:19:24miraksomething like that. even 3 hours would be ok for me with a good framerate
18:19:37mirakmarkun: yes I agree :D
18:20:17preglowso, you don't use iram and you don't use emac?
18:20:18mirakmarkun: if you look at the imdct used by mpeg or jpeg, it's just two imdct
18:20:25 Join Kohlrabi [0] (
18:20:35 Quit Kohlrabi (Client Quit)
18:20:50 Join Kohlrabi [0] (
18:20:50mirakmarkun: run one after the other, one on the line the other on columns. most of the code could be reused probably
18:20:53merbananmirak: idct not imdct if you talk about video
18:21:01preglowhe does
18:21:09preglowmirak: link me to this thing so i can have a quick look
18:21:48mirakmerbanan: what's the m for ?
18:21:49markunpreglow: ?
18:22:06markunOr were you talking about the xvid player?
18:22:09preglowi mean the xvidcore thing
18:22:25mirakpreglow: what do you want ?
18:22:32merbananmirak: it is used in audio (aac,vorbis,ac3)
18:23:09mirakmerbanan: yeah but the m is for modified right ?
18:23:15mirakmerbanan: yeah but the m is for modified, right ?
18:23:43preglowmodified because it's a lapped transforms as opposed to the ordinary dcts, which are ordinary block based
18:23:59merbananmy point is that it's best to keep them apart
18:24:03preglowanyway, the only mdct i have seen is one dimensional
18:24:05mirakpreglow: what means lapped ?
18:24:14preglowmirak: the transform blocks overlap
18:24:42mirakpreglow: then an interpolation is done ?
18:24:54preglowthe overlapping is a feature of the transform
18:25:04preglowit eliminates boundary artifacts
18:25:20preglowboundary artifacts are very noticable for audio
18:25:30 Quit Matze ("Miranda IM! Smaller, Faster, Easier.")
18:25:48mirakfor image as well, there is the blocks
18:25:58preglowyes, but you really only see them when you zoom in
18:25:58mirakthat's less disturbing probably
18:26:08preglowfor audio, it would have been completely useless
18:26:27markunwould imdct not be nice for video?
18:26:47preglowi don't even think i've ever seen it formulated for 2d
18:27:01markunHm.. no, probably not
18:27:02preglowanywho, the problem seems to be dealt with differently in video
18:27:07preglowh264 has deblocking, for instance
18:27:15preglowbut done in a completely different way than overlapping the transform
18:28:33mirakI am not sure why there is blocking on image, because I have done in univ dct and idct on full image in grayscales
18:28:50preglowthere always is blocking
18:28:54mirakit works as well, and you can cut off frequencies
18:29:00preglowit's just unnoticable, since the blocks are so small
18:29:08preglowbut yeah
18:29:16preglowthis idct can be pretty easily optimised
18:29:18mirakpreglow: you see what I mean ?
18:29:46NicoFRmirak: what kind of studies are you in ?
18:29:48merbananmirak: cut more coeffs and you'll see blocks in the picture
18:30:10mirakmerbanan: ok, I should reinstall matlab and do some tests ^^
18:30:35mirakNicoFR: I did a master in datamining and artificial intelligence
18:30:41mirakwell it was wide
18:31:01mirakwe did image analysis
18:31:02NicoFRwhat is 'datamining' in french ?
18:31:15mirakthat's datamining also in french, I am french
18:31:24mirakit's "fouille de données"
18:31:50NicoFRoh, never heard of that, but it's what I understood
18:32:39mirakI tryed to do a master in Bordeau of sound and image processing comrpession etecetera. I got rejected
18:34:08mirakmerbanan: you practiced or studied that ?
18:34:38 Quit Krazyjim (Read error: 110 (Connection timed out))
18:34:38 Nick [1]Krazyjim is now known as Krazyjim (
18:35:22merbananmirak: yes
18:35:56merbananmirak: I studied signal processing
18:37:04mirakmerbanan: my speciality is more pure informatic than maths, and signal process is really mostly maths. That was odd, but very intersting.
18:37:11 Quit DreamTactix291 (Read error: 104 (Connection reset by peer))
18:37:49mirakmerbanan: I don't know what's your background
18:38:29merbananalmost finished University
18:38:55merbananin the northen part of sweden
18:39:25 Join DreamTactix291 [0] (
18:41:05mirakpreglow: I think linuxstb tried xvidcore plugin on is ipod but it was crashing after decoding the first frame
18:41:13preglowyes, that's right
18:47:39 Quit muesli- ("ich will Kühe!!!")
18:48:07 Join bandgeekndb [0] (
18:49:04 Quit bandgeekndb (Client Quit)
18:50:35 Join Benacool [0] (
18:51:31BenacoolHi guys! I have a little question about playlists on my H320..
18:52:55NicoFRgo ahead
18:53:11BenacoolIs it normal that when I use a playlist with accentuated filenames in it that rockbox skip these accentuated files ? (It show (ERR) in the playlist viewer in front of the filename)
18:54:11Benacoolbut when he create is own playlist (on the fly) it recognize them
18:55:50Benacool(this bug is only with m3u playlists created with Winamp for exemple)
18:57:05_FireFly_maybe a m3u-create failure in winamp
18:57:10Pisilly question, do the playlists work with the origonal fw?
18:57:18Benacool2 sec<
18:57:31_FireFly_Benacool: check the two playlists
18:57:40_FireFly_in which way they are differ
18:57:47Benacool2 min ;-)
18:58:51Benacoolyes the playlist works in the original firmware
18:59:16Benacoolthe dynamic one is not converted so didn't work but its normal
18:59:19preglowi'm out, later
19:01:26Benacoolall playlists with accentuated filenames in it don't work well, it always skip these files
19:02:25 Quit Krazyjim (" HydraIRC -> <- State of the art IRC")
19:02:27Benacool Du Soleil - Alegría.m3u
19:02:55amiconnHow did you build these playlists? On the unit itself, or in a PC program
19:03:11amiconnIf you did the latter, you need to save the playlist as UTF-8
19:03:37Benacoolyea but how i say to EAC to save m3u files in UTF-8 lol ?!
19:03:56*amiconn didn't even know that EAC can create playlists
19:04:14Benacoolwhen you rip your CD he can make a playlist ...
19:04:16amiconnI don't use playlists often
19:04:34Benacooland each time a RIP a CD he create a playlist
19:04:52_FireFly_Benacool: load the playlist-files in an editor which can save in UTF-8 and save the files as UTF-8 with it
19:04:55Benacooli'll test with winamp
19:05:11Benacoolyea ... but i must do that with every playlists -_-
19:05:35_FireFly_hmm under linux there would be a simple way to do it with an script ;)
19:06:01BenacoolLinux don't like my NVRaid ;-)
19:06:16_FireFly_NVRaid is an faik raid
19:06:21Benacooli know i know
19:06:34_FireFly_convmv is the program
19:06:54_FireFly_i have run an raid on such an fiak-raid with software-raid under linux
19:07:08_FireFly_maybe you can run the convmv under cygwin
19:07:21_FireFly_or ther is maybe an win32 build of it
19:07:38Benacoolyea they are apps to recognize these fake raid but don't try to boot directly on it ... it's to difficult for me lol
19:08:10_FireFly_no problem here
19:08:20_FireFly_is also an NV-Sata-faikRaid
19:08:30_FireFly_i ran evms
19:08:33_FireFly_under gentoo
19:08:51_FireFly_and also the boot is on the raid as Raid-1
19:09:48Benacooli read that you can create some kind of ram-drive where the app can recognize the fake raid array and be able to boot on it after
19:10:02_FireFly_evms has an initrd for it
19:10:21_FireFly_to setup the raid-array
19:10:24_FireFly_on boot
19:10:32Benacool(i don't know a lot on Linux ...)
19:10:45Benacoollike ... what's initrd lol
19:11:06_FireFly_initrd is an initramdisk ;=
19:11:20_FireFly_an initrd is loaded and mounted when the kernel-boots
19:11:36_FireFly_most binary-distributions uses an initrd
19:11:58_FireFly_to load all necessary hardware drivers for the maschine
19:12:04_FireFly_on boot
19:12:51Benacoolso you must have it on an another hard-drive ?
19:13:17Benacool(i have two drives in raid 0 (on the nvraid) + another sata drive)
19:14:11_FireFly_i don*t know if it is for raid0 also but at least for Raid1 (mirror) you can access each partition seperatly becuase they have the same files on it
19:14:22Benacooli could simply install linux on this SATA drive ....
19:15:15_FireFly_Benacool: you can install it on the seperate sata-drive and mount the raid0 array elswhere in the filetree
19:15:30_FireFly_but afaik you can't use the current raid0 array under linux
19:15:38_FireFly_you must setup it again
19:16:07_FireFly_becuase your current one is created under "windows"
19:16:40Benacooli found a little app that permit to recognize raid created under windows but ... damn i don't it in my favorites ^^'
19:16:59Benacooli search it
19:19:06Benacooli think it's that
19:19:22_FireFly_you can handle dmraid-array with evms
19:19:56_FireFly_it has several modules for the differents software-raid solutions under linux
19:20:01_FireFly_dmrad lvms
19:21:12_FireFly_but you can also only use dmraid
19:21:21_FireFly_it's your choice ;)
19:22:17Benacoolon top of that all my partitions are in NTFS ... lol
19:22:59_FireFly_then use convmv with cygwin
19:23:41_FireFly_if cygwin has this program included
19:25:14Benacoolconvmv ... this app convert something ? or it allows to rocognize my partitions in ntfs ?
19:25:19markunconvmv is not what he needs
19:25:30markunThat would only rename his filename
19:25:41_FireFly_ah right
19:25:42Benacoolok ... not interesting lol
19:26:36 Join ratpack92 [0] (
19:26:45_FireFly_then you must do it manual with an utf-8 capabel editor
19:26:52 Part ratpack92
19:26:53_FireFly_or create the playlists on rb
19:26:59markunOr find another program that can change multiple files
19:28:12Benacoolyea i will save my playlists again with the notepad in utf-8 ... iwas thinking that you wanted to help me to install linux ... loll
19:29:28Benacoolthanks ^^
19:30:13Benacoolhello !
19:30:14mirakI am trying to use IDATA_ATTR to put an array into IRAM but I fail
19:30:26 Join Thus0 [0] (
19:30:28mirakI am not sure what to do
19:31:53mirakok seems to work now
19:32:52Benacoolha ha !
19:33:19Benacooli'll try that
19:37:41 Quit Nibbler (Read error: 110 (Connection timed out))
19:37:57 Join Nibbler [0] (
19:40:46Benacoolvoila !
19:46:25 Join leftright [0] (
19:50:10***Saving seen data "./dancer.seen"
19:52:10leftrightLinusN; found a bug, using file tree tracks that dont have leading zeros, aren't listed sequentialy in numerical order in Rockbox, but ok in Explorer or suchlikes, without leading zeros up to number 10 the files are listed randomly in albums/directories
19:52:55LinusNthat's not a bug
19:54:14markunwell, if they are listed randomly that doesn't sound good
19:55:06markunit should still be 1, 10, 11, 2, 3 ...
19:55:29leftright1,2,3 etc without leading zeros methinks
19:56:17_FireFly_it shoudl 0,10,11,...,2,20,...5,..,55
19:56:59_FireFly_when no leading 0 is used
19:57:12leftrightwhy does exporer list them sequentially then ?, unless MS is wrong ofcourse
19:57:12_FireFly_in 2 digits numbering
19:57:23_FireFly_no it doesn't
19:57:30markunleftright: is there a right or wrong here? It's just different
19:58:07_FireFly_leftright: the ms-explorer list such files also "random"
19:58:09leftrightok its just odd to have files listed non sequentially unless they have leading zeros
19:58:46leftrightmy explorer lists them sequentially, thats why i noticed it
19:58:51_FireFly_how should know the compare
19:59:01_FireFly_leftright: afaik not
19:59:16leftrightMine does
19:59:29_FireFly_or you have only files which begins with numbers 0-9
19:59:52KohlrabiAt least XP does sort sequentially wihtout leading zeroes
19:59:56LinusNthe files are lexically sorted
20:00:01Kohlrabi9x doesn't afaik
20:00:15leftrightnope I have more than 10 files in albums and all are listed sequentially in Explorer
20:00:15_FireFly_leftright create following files 1files, 2file,10file
20:00:22amiconnWindows tries to be 'clever'. Whether this is really clever is debatable
20:00:42_FireFly_they will be listet 1file,10file,2file
20:00:49leftrightI dont have to create it, I have it on my player right now
20:01:09_FireFly_but this is no bug
20:01:12_FireFly_at all
20:01:20leftrightthank you,
20:01:21Kohlrabiit sorts: 1file, 2file, 10file
20:01:37Kohlrabiexplorer, that is
20:01:43markun_FireFly_: maybe the german windows sorts differently?
20:01:47_FireFly_then windows makes a seperate work to sort such file
20:01:48leftrightI will add leading zeros to my entire 40 gig music collection
20:01:53Kohlrabiyes, it does
20:02:01Kohlrabiit sorts sequentially nonetheless ;)
20:02:13markunleftright: or add the feature to your rockbox to sort the numbers
20:02:14Kohlrabileftright: yeah, you should do that
20:02:16_FireFly_markun: the last time i have looked at the ordering of such files under windows is long long ago ;)
20:02:57Kohlrabileftright: should be quick with foobar2k-masstagger
20:03:20leftrightdear oh dear, if I could code i wouldn't be bothering you with trivia
20:04:05leftrightI hate leading zeros in track numbers though, just seems wrong
20:04:13Benacoolleftright: you can use some software to rename all your file form the ID3Tags if they are well made
20:05:21Benacooland it can put 01, 02 , etc in front of your filename with the track tag
20:05:47markunleftright: I also can't say that I love them, but most programs don't sort correctly without them, not only rockbox
20:05:57leftrightdont want leading zeros in ID3, only in file name
20:06:20Benacooli take the tag 1 (track) and put 01 in the filename
20:06:47markunor use foobar like kohlrabi says
20:06:53Benacooloh ok ^^'
20:07:02amiconn...or The Godfather
20:07:32 Join Lear [0] (
20:07:36leftrightthanks I have a masstager which can do that, the point is that I dont want leading zeros in ID3, only in the file name
20:08:00Benacoolwith audiograil you do that
20:08:03_FireFly_most tag-programms can also change the file-names
20:08:13Benacoolyou can*
20:09:23 Join xmixahlx [0] (n=xmixahlx@
20:09:31leftrightyou can change file name to tag name, but not rename file names with leading zeros without adding leading zeros to ID3 as well
20:09:33Kohlrabii dunno about other masstaggers
20:09:46KohlrabiBut f2k-masstagger can handle filename and tags seperately
20:10:09 Quit DangerousDan (Read error: 110 (Connection timed out))
20:10:33 Quit NicoFR ()
20:10:44_FireFly_easytag (linux) also ;)
20:10:56leftrighthmm, I'll look into f2k
20:11:13Kohlrabiyeah, just ask if you need halp with the masstagger
20:11:54leftrightthank you for the offer
20:12:25 Quit mirak (Connection timed out)
20:12:47BenacoolAudiograil renames your files from a tag track with 1, 2, 3 and create filenames with 01, 02, 03 ... (if i well remember)
20:13:01Benacooleven if the tag is 1, 2 ,3 ...
20:13:09Kohlrabisounds good
20:13:32Kohlrabibesides...did anyone ever encounter a 100+ tracks album?
20:13:56Benacoolthe problem is that audiograil is not free :/
20:14:25amiconnKohlrabi: No. That isn't possible because of the CD specification
20:14:28Maxime(mp3 tag&rename, great tool!)
20:14:36amiconnBenacool: The GodFather is free
20:14:39 Join mirak [0] (
20:14:59Benacoolok... it's an app lol ...
20:15:21_FireFly_as amiconn has already posted
20:16:15*amiconn is installing ubuntu in a VM for a test
20:16:45_FireFly_for using dd from ubuntu for the ipod i guess?
20:16:56Kohlrabithat's qhy I chose "album", not CD
20:17:11Kohlrabi(i didn't know about CD specs, though ;))
20:17:18amiconnI'm looking for a usable distro, as I'm planning to get an AMD64 PC
20:17:26_FireFly_ah ok :)
20:17:49amiconnKohlrabi: All of today's albums are published on CD, so no album will exceed the CD specs
20:18:07_FireFly_i guess you don't want a source-distri like gentoo
20:19:06markunFreeBSD! :)
20:19:18markunWell, not a distro
20:19:24_FireFly_afaik ubuntu is based on debian
20:19:48amiconnI alreday have a debian VM
20:20:18 Part leftright
20:20:40amiconnI don't want an rpm-based distro
20:21:05_FireFly_but an binary distro? :)
20:21:39amiconndebian is nice, but either you install the ages-old 'stable' variant, or you run into various version conflicts
20:21:51amiconn(happened to me with testing)
20:21:56_FireFly_yeah that is a big prolem with debian
20:22:04amiconnEven testing packages are rather old sometimes
20:22:08_FireFly_if you want stay up to date
20:22:35merbananamiconn: gentoo is fairly usable, if you have the time compiling it
20:22:36Benacoolhow godfather works lol ... when a click on apply in rename tab, nothng happen ?!
20:22:54_FireFly_merbanan: ;)
20:24:00lostlogicI use gentoo stable for my server and unstable for my workstations... aside from crossdev's failure to actually build a working cross compiler for rockbox I have few problems with it.
20:25:09amiconnI don't want to compile my distro either. It's already more than enough hand-work to get a linux box up & running compared to windows
20:25:29_FireFly_but then it rans forever ;)
20:26:05lostlogica perspective I completely disagree with, but this is not the time nor place. recommend SuSE I do.
20:26:06_FireFly_windows gets slow if you install/unistall much programms
20:26:30_FireFly_SUSE uses rpm and he wants no rpm
20:26:49lostlogic::blink blink:: slackware.
20:27:01 Join webguest06 [0] (
20:27:02amiconnlostlogic: I have rather bad experiences with SuSE, that's why I don't want rpm
20:27:21lostlogicdo you want dependency management?
20:27:35lostlogicif so, debian, if not, slackware
20:27:59saa[b_r]iderdebian is cool
20:28:15_FireFly_ah no flamewar which distri is the best please :)
20:28:16 Quit webguest06 (Client Quit)
20:28:39*saa[b_r]ider uses debian-based Mepis. (no flaming intended)
20:29:22amiconnlostlogic: Yes I want dependency management, and I already had more than just a quick look at debian
20:29:33amiconnThat's why I'm trying ubuntu now
20:29:56amiconnThe good thing with VMware is that you can test this & that...
20:30:34_FireFly_amiconn: if you don*t liek gnome the try kubuntu it has kde as default
20:31:02amiconnI like gnome more than KDE
20:31:16Lear_firefly_: Currently, the apt-get UI for Kubuntu isn't as good as the Ubuntu one...
20:31:39_FireFly_i didn*t hat kubuntu yet seen running myself
20:32:14amiconnlostlogic: As for the installation times & handwork - I usually get a windows machine up & running (OS & all built-in hardware) within 2 hours. With linux it's more like 2 days
20:32:33amiconnThat's not an inherent flaw of Linux, rather it's a flaw of the HW manufacturers
20:32:59_FireFly_taht is is the biggest prolbem for the newest hardware
20:34:56 Part Benacool
20:34:59 Join Benacool [0] (
20:36:11 Join Benacool{} [0] (
20:36:22 Quit Benacool{} (Client Quit)
20:38:22 Join Benacool{} [0] (
20:38:26 Quit Benacool{} (Remote closed the connection)
20:38:58 Join NicoFR [0] (
20:40:18Learlostlogic: using the profiler, have you made any tests with Tremor using no iram at all? I was thinking it might be useful to re-do the iram usage based on profiling results...
20:50:13lostlogicLear: I had done some a while back, but not on the latest profiling or tremor code.
20:50:53LearNot completely wasted if I should start looking at it then, I guess...
20:52:13lostlogicyeah −− the thing to be wary of is that inlining changes makes significant change in how profiling looks. which is why I created and used that NO_PROF_ATTR to not profile inlined calls.
20:53:54mirakamiconn: depends of the hardware
20:54:30mirakamiconn: 2 days seems a bit long for linux. What are you doing with it ? :p
20:54:32LearSo inline calls are not included at all? (In call count, that is.)
20:55:06_FireFly_maybe 2days are only with gentoo or LFS
20:55:27amiconnNo, that was with binary distros
20:55:35amiconn(SuSE especially, but also debian)
20:55:57amiconnThe problem is that no distro provides drivers for all components
20:56:01amiconn(it can't)
20:56:27lostlogicLear −− if they are marked with that attribute, they aren't profiled at all. you can leave them profiled, but it seems to skew results somewhat due to the extra jsrs and instructions on each call, and if course profiling a song takes a lot longer when nothing that's going on.
20:56:41_FireFly_gentoo has many drivers as ebuild
20:57:15amiconnThe last problem I had was getting wlan to work (with wpa)
20:57:35amiconnThat even was with a card where the chipset manufacturer supplied open source drivers
20:57:36_FireFly_which card ??
20:57:39lostlogicthat's a rough edge for sure still.
20:59:08amiconn_FireFly_: LevelOne WNC-0301 (Ralink RL2500 chipset)
20:59:22lostlogic(that said, I use exclusively wireless networking at home with wpa)
20:59:22amiconnlostlogic: Not the only one
20:59:58amiconnAnother linux problem is that there is sometimes more than one system for the same purpose, but the systems are incompatible
21:00:06amiconn(e.g. oss <-> alsa)
21:00:19lostlogicamiconn: depends on what you need and a lot of things. I'm a big linux geek, so I completely don't notice a lot of things that really bother windows users, and I don't use a full gui environment to see the issues with those.
21:00:43 Quit DreamTactix291 (Read error: 110 (Connection timed out))
21:01:13amiconnMy main requirement is that I need a working system, so that I can work *with* the PC instead of working *for* the PC
21:02:01amiconnIt's of little importance how the system is called.
21:03:05lostlogicthat;s the point of a distribution... if you use packages provided by the distribution or explicitly for that distribution then you shouldn't run into those compatibility issues. Ubuntu does seemlike a good idea for you.
21:15:43 Quit NicoFR ()
21:16:47 Join muesli__ [0] (i=muesli_t@
21:23:57amiconnWtf is this ubuntu doing??
21:24:55amiconnI installed gcc, but: gcc -> bash: gcc: command not found
21:24:58mirakyeah ubuntu is fine
21:25:29mirakamiconn: ls -l /usr/bin/gcc*
21:25:55muesli__amiconn interesting?
21:26:10mirakamiconn: shouldn't happen
21:26:29amiconnYeah, 'shouldn't' is the word
21:26:48amiconnAh, now it works...
21:26:54amiconnLinux is strange...
21:27:15amiconnWhy is there both a gcc4 and a gcc package (which is also gcc4)???
21:27:31ender`use gentoo, use gentoo :)
21:27:49mirakamiconn: probably a metapackage
21:28:01mirakamiconn: you can install concurently gcc3 and gcc4
21:28:36mirakamiconn: you probably have gcc4 now. if you want to set the good one you can use update-alternatives
21:29:01mirakyou shouldn't need that anyay since you need a crosscompiler
21:29:24amiconnI need gcc for installing vmware tools
21:29:35amiconn...and of course for compiling rockbox sims
21:30:22 Join oxygen77 [0] (
21:30:30amiconnI can't install the complete VMware tools because it complains that the kernel was built with gcc 3.4.5, and the installed gcc is 4.0.2
21:32:32mirakamiconn: you can give a shot at qemu, that's a good free alternative to vmware on linux
21:33:05mirakamiconn: #qemu . You will need kqemu kernel module or qvm86
21:33:15amiconnThat would require a host PC with linux, which I don't have
21:33:32ender`qemu works on windows, too
21:34:01ender`(though the accelerator module is considered alpha on windows, and sometimes causes qemu to unkillably hang)
21:34:36mirakender`: I have seen that. I never tried it.
21:34:51mirakender`: there is colinux also on windows wich can be interesting
21:34:58mirakand linux too by the way
21:35:21 Part oxygen77 ("Leaving")
21:35:23 Join JoeBorn [0] (
21:35:25*ender` personally prefers SFU on windows, though he wasn't able to build a cross-compiler for rockbox on it yet
21:35:38miraksfu ?
21:35:47ender`services for unix
21:35:53mirakdon't know that
21:35:56ender`posix layer for nt kernel
21:36:22mirakcolinux is nice because you run things without emulation, it runs at native speed
21:36:37amiconnender`: Some time ago I tried the same. The SFU 'make' doesn't even recognise the -C option...
21:36:56ender`get gnumake there
21:37:03ender`(and a ton of other useful utils)
21:37:04mirakyou should give a shot to colinux probably
21:37:46LearNice, a profiling build seems to be working...
21:37:48ender`SFU mostly ships with BSD utilities
21:38:09ender`i built myself gnu coreutils (most of them anyway), since i'm more used to them
21:39:11ender`i'll have to try colinux sometime
21:39:53ender`good thing about interix (SFU) is that you still can run native win32 programs directly, so it's a bit easier to cross-compile stuff that's intended to run in win32
21:40:06ender`(i use interix to compile gimp)
21:43:32 Join saratoga [0] (
21:43:42Learlostlogic: I get no symbols in the parsed profiler output (and a null pointer exception during run of the parser). Any suggestions?
21:44:08LearA lib file missing from grouped map
21:44:34 Quit akaidiot (Read error: 110 (Connection timed out))
21:44:45lostlogicLear: had you recompiled anything between making the version you profiled and parsing?
21:45:01lostlogicand what was your ProfileReader commandline
21:45:27LearExcept java-specific stuff, profile.out apps/codecs/ libTremor.a
21:46:03lostlogicI wonder if the relative paths are breaking it or something... can try with absolutes?
21:47:19lostlogicI kept expecting preglow to perlize my java there, so I didn't really test it outside of how I was calling it
21:47:28LearI can, even if it is a bit of a pain. :) Btw, I get this message too: "iError parsing LaO: [, .bss, 0x32ecc990, 0x4," followed by a path...
21:48:51LearDoens't find the map file with absolute paths (in Unix style...). Remember I'm running Cygwin, maybe the map file layout is different enough?
21:48:53lostlogichaven't seen that one, I'll have to revisit the parser and maybe turn it into perl at some point
21:50:00lostlogicLear: it's possible that the objdump output is different enough :(
21:50:12***Saving seen data "./dancer.seen"
21:50:17amiconnWhy should it?
21:51:36 Join Sacro [0] (i=Sacro@adsl-83-100-210-85.karoo.KCOM.COM)
21:51:58LearIt fails on the first "real" row of the .bss section, it seems...
21:57:12LearPerhaps because the line is completely devoid of parentheses...
21:57:56 Quit xmixahlx ("blah blah blah")
21:59:15mirakis there a way to know how full the iram stack is ?
21:59:25Learmap files
21:59:49LearIn the memory configuration section, you can see how large it is.
21:59:50mirakLear: in the case of a recursive calling wich is dynamical how can you know ?
22:00:25Learah, you mean the stack; I looked more at the iram part. :)
22:00:35mirakby the way is the stack in iram of rockbox kernel or in plugin or codec ram ?
22:00:36LearOnly some stacks are in iram...
22:00:54mirakiram I mean for the last
22:00:59LearAfaik, the main thread stack and hte codec thread stack...
22:01:52 Join DangerousDan [0] (
22:02:11Learlostlogic: I get an SymbolMap$BadMapFormatException at SymbolMap.parseLibraryAndObject(, btw...
22:04:21LearOh well, manual matching works for now... :) (And I do get the feeling that a perl script to do the matching would be quite easy to write, for one experienced in perl, at least...)
22:10:41Benacool_FireFly_: you are there ?
22:11:21BenacoolWinamp don't like UF-8 m3u lol ...
22:12:01Benacooljust for info
22:12:06_FireFly_carp program ;)
22:13:09Benacoolhe read "02 - Les Étoiles Filantes.mp3" instead of "02 - Les étoiles Filantes.mp3"
22:14:14ender`winamp doesn't support unicode
22:14:27 Join solexx_ [0] (
22:16:07Benacoolso i have unicode m3u for rockbox and normal one for winamp ... :/
22:17:02 Quit mirak (Remote closed the connection)
22:17:04 Join akaidiot [0] (
22:18:06_FireFly_Benacool: have you the latest version ??
22:18:10_FireFly_of winamp
22:19:23_FireFly_Benacool: can winamp play the files in the m3u ??
22:21:05Benacoolyea i have version 5.12 and he can't play the files in the palylist because it don't found them
22:21:48_FireFly_ maybe the lanuage-pack helps
22:22:06_FireFly_or any other language-packs
22:22:23Benacooli have the french one ...
22:22:48 Join mirak [0] (
22:22:58_FireFly_ok then winamp is really crap ;)
22:24:19_FireFly_ try this
22:24:45Benacoolany suggestion then ?
22:25:06Benacooloh i tested this something like a year ago ...
22:25:16Benacooli'll dl it
22:26:38_FireFly_QCD has amazing support for Unicode (filenames, ID3 tags, etc),
22:26:44_FireFly_from an forum post
22:26:56 Quit solexx (Read error: 110 (Connection timed out))
22:27:06_FireFly_and the best it is free ;)
22:28:30amiconnbah, crappy linux
22:28:45amiconnWhy oh why doesn't it just work??
22:28:50_FireFly_or crappy vmware ;)
22:29:18amiconnrhythmbox doesn't do what it should
22:29:29amiconnThe strange thing is that it works in debian
22:30:58mirakamiconn: rythmbox is crap
22:31:14miraktry quodlibet instead for exemple
22:31:44amiconnThat's another such point... if rhythmbox is crap, why is it the default??
22:32:05mirakahah good question
22:32:16mirakreally I am wondering that since a long time
22:32:28amiconnYes. This question has to be asked a whole lot more
22:32:33mirakyou might need to install gstreamer* plugins
22:32:39amiconn...regarding operating systems
22:32:44mirakfor rythmbox
22:32:55mirakit rely on that, but I don't like it much anyway
22:33:39miraktotem is also the default, and I don't like it as well
22:33:51mirakI prefer gxine
22:33:52amiconnThey are installed...
22:34:08amiconnJust found that the multimedia settings were set to ESD instead of ALSA...
22:34:29Benacool_FireFly_: thanks again... QCD recognize them perfectly ;-)
22:34:49mirakprobably esd uses alsa
22:35:21_FireFly_esd is only an sound-deamon it uses either oss or alse
22:35:23amiconnHmm, still doesn't werk
22:35:31_FireFly_werk? ;)
22:36:13amiconnyup ;)
22:36:14markunI like the interface of Muine, but gapless playback doesn't work
22:36:43mirakwhat's muine ?
22:36:52_FireFly_mirak: a gnome media-player
22:37:07_FireFly_bmp is a fork of xmms using gtk2
22:37:10lostlogicLear: sorry. got pulled away by my job... I'll see if I can get to perlizing the parse code and making it smarter in the next few days
22:37:41mirakI am going to try it
22:37:52mirakamiconn: if you have an ipod there is also gtkpod
22:37:52LearNo worry, have something very basic in Python that might work...
22:38:05mirakand banshee
22:38:16amiconnmirak: No I don't, and I don't plan to have one
22:38:24Benacoolsomebody know the utility of the hex "EE BB BF" at the beginning of a UTF-8 file ? (something to do with the byte-order mark ??)
22:38:37LearNot all functions can be found, it seems...
22:38:43Learbenacool: correct.
22:38:55Benacoolwhat is it ?
22:39:00markunbyte odering in UTF-8?
22:39:05markundon't think so
22:39:05Benacoolyea ...
22:39:28lostlogicLear: map files get non-statics, but for statics you need objdump combined with the map file start addesses
22:39:32Benacoolbecause rockbox recognize them as caracters
22:39:44Benacoolin playlists
22:40:04_FireFly_The BOM for UTF-8 is 0xef 0xbb 0xbf
22:40:08Benacool(crappy converter)
22:40:25lostlogicthe parseLibraryAndObject method is supposed to get an offset for a given object file within a section of ram
22:40:31Learlostlogic: ah, yes, that must be it. Does your parser handle non-statics (I guess so, since it seems to parse the lib file too)?
22:40:48amiconnmarkun: It's the unicode BOM encoded as UTF-8. Not necessary for byte order, but as a marker for an UTF-8 file
22:40:56lostlogicyeah, it does... when it works
22:40:57 Join DreamTactix291 [0] (
22:41:14Learbenacool: yes, that isn't recognized at the moment...
22:41:30Benacoolok ;-) so i remove it
22:41:43Benacoolin my playlists
22:42:15Learunicode/utf-8 support isn't yet complete for files, afaik...
22:42:29_FireFly_an BOM for UTF-8 isn't needed at all
22:42:48_FireFly_as the faq on say
22:43:01markunAlso the viewer needs to be updated. phaedrus was working on it, but he doesn't have time for it anymore.
22:43:16LearCould be useful for Rockbox to decide if a file should be treated as UTF-8 or as the codepage setting says...
22:43:16amiconn[22:42:51] <amiconn> .... Not necessary for byte order, but as a marker for an UTF-8 file
22:47:38 Quit _FireFly_ ("Leaving")
22:48:34lostlogicare wps updates stopped when the LCD is off?
22:49:49mirakI have not found yet a music player that synchronise files on a UMS digital audio playr
22:50:24mirakBagder: unison
22:50:49mirakBagder: but the problem remains the same, you can most of the time only sync a full folder
22:51:13Bagdernot with rsync
22:52:06lostlogicrsync seems to have problems with fat32. it is fine if you use chksums, but then it's not a whole ton faster than just copying everything...
22:52:40Bagderit most certainly is
22:52:55mirakBagder: rsync you need to have a kind of image
22:52:56Bagderit syncs the files that are missing
22:53:10Bagderrsync syncs files
22:53:13miraka local image of what is on the dap
22:53:22mirakyes but it syncs all files
22:53:30Bagderall given files, yes
22:53:46mirakI mean, if you have a folder music/ if you rsync to the dap root it will stuff all files in it
22:54:36Bagderif you tell it to sync the whole folder, yes
22:54:57lostlogicmirak: it is designed to only update, but it has a hard time telling what to update (at least in my experience) between reiserfs and fat32
22:55:15BagderI disagree
22:55:25BagderI sync fine between ext3 and fat32
22:55:35lostlogicI must just suck.
22:56:45mirakwell maybe you want to try unison it works, but the problem is that it default to syncing all files, and doesn't remember what you didn't wanted to sync.
22:56:59mirakso you must say to do nothing on most of the files :-/
23:01:24LearHe he, profiling is interesting... :)
23:02:06lostlogicLear: Yay, makes me happy that someone else is finally playing with it! :)
23:04:06LearReally gives you an idea how much iram speeds things up, on a function per function basis.
23:05:01LearE.g., compared to nothing in iram, moving the pcm array(s) in mdct_backward to iram about halves execution time...
23:05:04lostlogicyeah, is interesting to strategically try to balance between iram and code cache...
23:06:22Moosis this profiling thing could help for others codecs too?
23:06:41LearSure, as it tells you where to focus optimization work...
23:07:34lostlogicwas why it was worth writing a real profiling patch instead of just modifying tremor functions to count themselves manually
23:08:04Moosgood thought :X
23:08:12mirakLear: on xvid I have put an array used to mask the idct result, and the boost is visible. I think it decodes 3-4 images per seconds now
23:08:52 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
23:08:54markunMaybe 10fps are not that far away then
23:09:32mirakyep, there is many arrays that can be put into iram
23:10:05miraksome of them are into a O(2) loop
23:10:08saratogaWhat the latency on the IRAM like?
23:10:43lostlogic1 bus cycle?
23:10:44saratogawhats I mean
23:11:03saratogaso its basically programmable L1 then?
23:11:14lostlogic1 bus cycle is not 1 cpu cycle
23:11:20lostlogicl1 is 1 cpu
23:11:36saratogawhy a bus cycle? i thought it was ondie?
23:12:15lostlogicstill has to go through the ram controller iirc. someone may have to correct me.
23:12:52 Quit Pi (Read error: 110 (Connection timed out))
23:13:06saratogahow fast is the DRAM then?
23:13:56LinusNiram should be 1 cpu cycle
23:14:36lostlogicah, the core speed is 2x clock, as is the L1
23:15:49LinusNthe bus speed is 1/2 the cpu clock
23:16:27MoosLinusN: hi, did you see this :D :
23:17:26Moosit still working after this ;)
23:17:26LinusNi hope the poor dog didn't cut himself
23:17:32Mooshahaha :D
23:18:31lostlogicsaratoga: whoops, sorry
23:24:38 Quit Lear ("Chatzilla 0.9.69 [Firefox 1.5/0000000000]")
23:25:59muesli__how safe is 128bit encryption? can it be decoded fast/easy?
23:26:14Bagderit depends
23:26:18saratogawhich type
23:26:34muesli__i am talking about trillian icq encryption, if that helps
23:26:42muesli__i reckon they are using blowfish
23:28:53muesli__SecureIM with Blowfish 128 bit encryption
23:29:32 Join ataxy [0] (
23:29:35muesli__SecureIM with Blowfish 128 bit encryption and Diffie-Hellman key exchange.
23:29:54lostlogicLinusN: is the L1 faster in practical terms than the SRAM? I can't tell from reading the manual.
23:31:19 Join jborn_ [0] (
23:31:26saratogaI thought their was only an L1 I cache?
23:31:56lostlogicsaratoga: correct.
23:32:22saratogaso then the relative speed is irrelevent
23:32:31saratogasince all you can do is fetch ops
23:32:49saratogaor am i missing something
23:33:27muesli__Bagder did that information help you to answer my question?
23:33:37lostlogicyou can put code in IRAM, but tyat prevents it from being cached in the L1, so if IRAM is slower than L1, sometimes moving things to iram could slow us down.
23:34:39saratogathats interesting
23:34:46saratogabut makes sense
23:35:14Bagdermuesli__: that is secure enough ;-)
23:36:03miraksomeone was talking about an utf8 converter ?
23:38:03amiconnIram and icache are both 1-cycle so they are equally fast
23:38:59amiconnPutting code in IRAM gives a slight speedup compared to icache
23:39:15amiconn...because it's already there and doesn't need to be fetched from SDRAM
23:39:38 Quit ataxy ("Chatzilla [Firefox 1.5/2005111116]")
23:39:39amiconnUsing IRAM for data is more efficient most of the time though, because there is no dcache
23:41:42lostlogicso iram is only worthwhile for code when the oft called code won't all be icached. (exists in more than 512 16 byte areas)
23:42:44amiconnEither that, or if it suffers from cache address aliasing
23:42:52amiconn(the icache is direct mapped)
23:46:54 Quit Kohlrabi (Read error: 104 (Connection reset by peer))
23:47:33 Quit JoeBorn (Read error: 110 (Connection timed out))
23:48:18 Join Kohlrabi [0] (
23:48:28 Join Sucka [0] (
23:50:13***Saving seen data "./dancer.seen"
23:50:16 Part LinusN
23:50:37mirakI will reenable the timers on xvidcore and see how much it speed up to put lookup tables an arrays into iram
23:52:25 Join akaidiota [0] (
23:55:29 Nick jborn_ is now known as JoeBorn (
23:57:04 Nick Sucka is now known as actionshrimo (
23:57:12 Quit Benacool (Read error: 110 (Connection timed out))
23:57:48 Quit actionshrimp (Nick collision from services.)
23:57:51 Nick actionshrimo is now known as actionshrimp (
23:58:28 Quit Strath (Read error: 104 (Connection reset by peer))

Previous day | Next day