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 2005-10-03

00:00:25amiconnThe factors are 4 and 11, btw
00:07:10TiMiDI don't know if it can be usefull, but it's the cpu of the gmini
00:07:37XavierGr amiconn: any idea why cpu_idle_mode function can have a bad tuning effect on the FM radio for the iriver?
00:07:51XavierGrIs there any chance that it will be fixed with your work?
00:08:04amiconnIt shouldn't have any effect
00:08:31linuxstb__XavierGr: Have you tried testing with and without Linus's last change?
00:08:35XavierGrWell it has. I was doing test the other night and I made a quick option (on the radio) that toggle the function.
00:09:15XavierGrYou can clearly hear a slight mistuning and a slight "pop" on the toggle. Especially when the radio has a bad signal on the sation.
00:09:48XavierGrlinuxstb__: yes i removed thos 2 calls to the function and all is good now.
00:10:45XavierGrWhat battery cost will it have if I remove those 2 calls?
00:11:33 Join tvelocity [0] (
00:11:58XavierGrAlso, I thought that the lowest cpu frequency (currently) is 48mhz. Does the cpu_idle_mode gets this lower?
00:12:50amiconnYes, to 11 MHz
00:13:22XavierGrhmm this is not viewable by the debug menu I suppose.
00:13:31amiconnIt is
00:13:43XavierGrthen my bad.
00:13:51amiconnYou can even set it there, by pressing select
00:13:54XavierGrso any idea?
00:14:07XavierGrwhat causing this?
00:14:19XavierGryou can test it yourself.
00:14:32XavierGrtune the radio somewhere in the middle of 2 stations.
00:14:55XavierGrthen click to exit the FM (but keep playing) and then reenter.
00:15:15XavierGron the exit you will hear more static and on the enter a slight pop with the static gone.
00:15:24amiconnOkay, tuning between stations will make the radio pick up all sorts of electrical interference, that's normal
00:16:11XavierGrTry it and in normal station, but then it is a little. more difficult to notice
00:16:15amiconnIdle mode switches off the PLL, so it can't interfere any more...
00:16:27XavierGrnormal when the cpu frequency changes?
00:16:39amiconnHmm, in fact there might be a problem, because the VCO is running at a rather high frequency
00:17:12amiconnvoltage controlled oscillator
00:18:05amiconnIt generates a rather high clock (between 200 and 400 MHz) that is used as the base for generating the actual cpu clock
00:18:07 Quit Aison ()
00:18:44amiconnI just found out something - does the static have its maximum around 95.9..96 MHz?
00:19:40amiconnIf so, Linus chose a bad VCO frequency for 48 MHz mode. VCO frequency is 383.8644 MHz there. Divide that by 4...
00:20:03 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
00:20:15XavierGrI will make some test and get back to you.
00:25:41XavierGrOk here is a more clear situation. I tune manually to 96.0 and I hear nothing (a little pops and hiss) (all this at 11mhz) if I set it to 48 and 120mhz I can clearly here music, though in a very mistuned way.
00:27:10amiconnHmm, that's strange
00:29:29XavierGrAnd yes and in weel tuned stations you can hear a fain pop between transitions of 11mhz and the other states.
00:29:58amiconnThat's probably caused by the activation/deactivation of the PLL
00:29:59XavierGrI guess you will have to test it too, maybe I have a faulty unit.
00:30:19XavierGrBUT I cant hear it between 48 and 120mhz
00:30:37amiconnYes, there the PLL is just reprogrammed
00:30:57amiconn48 and 120 MHz use the PLL, 11 MHz does not
00:31:16XavierGrso, is there anything that can be done, or we just have to live with it as the LCD screen?
00:31:48amiconnI don't think there's anything we can do
00:31:58amiconn(except not using 11 MHz in the radio screen)
00:32:28XavierGrAnd I think that not using the 11 mhz maybe is a waste for the radio...
00:33:12XavierGrso I will leave those 2 calls and what the hell how often does someone will have to exit and reenter the FM
00:33:24 Quit paugh ("Leaving")
00:34:31XavierGrThough I solved that "BOOM" sound when reentering this small click will remain. (Linus had forgotten to make an if statment for some radio_set functions which cause the loud POP)
00:35:23XavierGrIs there any values for battery consumption between 11mhz and 48mhz?
00:40:11XavierGrSo amiconn do you time to inspect (or even commit) my patch for the FM preset?
00:43:32 Join webguest06 [0] (
00:47:40 Join Colddy [0] (
00:56:35 Join webguest23 [0] (
00:58:01webguest23Seems that IRC logs at go up to 2005/09/29 but no further... Manually editing the URL to point to 20051001.txt (for example) yields a 404 not found.
01:02:18 Join FireFly_ [0] (
01:03:26 Quit FireFly_ (Client Quit)
01:03:36 Join FireFly_ [0] (
01:03:37 Quit matsl (Remote closed the connection)
01:04:56 Join ep0ch [0] (n=ep0ch@
01:05:20ep0chhmm where are Octobers logs?
01:05:47webguest23Exactly! :)
01:08:01DBUGEnqueued KICK Colddy
01:08:01*Colddy Winamp 5.08 - Cocoa Tea - Who (playing) - 32kHz 128kbps stereo - 4m 9s long |>>>>>>>>>>> (20%)|
01:08:12 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
01:10:12 Part ep0ch ("Kopete 0.10.3 :")
01:15:11***Saving seen data "./dancer.seen"
01:18:49 Quit Moos ("Glory to Rockbox")
01:19:34 Quit _FireFly_ (Read error: 110 (Connection timed out))
01:24:12CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
01:24:12*Colddy Winamp 5.08 - Patrice - Murderer (playing) - 44kHz 128kbps stereo - 3m 11s long |>>>>>>>>>>> (13%)|
01:25:04 Join ansivirus [0] (
01:27:25*Colddy Winamp 5.08 - Gentleman - Intoxication (playing) - 44kHz 192kbps stereo - 3m 29s long |>>>>>>>>>>> (15%)|
01:28:11 Quit Colddy ("Lurker .gx.")
01:28:20 Nick FireFly_ is now known as _FireFly_ (
01:30:46 Quit ansivirus ("Leaving")
01:37:36 Quit XavierGr ()
01:49:49 Join linuxstb [0] (
01:50:16 Quit linuxstb__ (Read error: 104 (Connection reset by peer))
01:51:21amiconnWee, my method works for the timer tick, and it's rather precise
01:51:37amiconn(good for those who want to implement a clock on H1x0 ;) )
01:55:03TiMiDcongratulation !
01:55:18 Nick TiMiD is now known as TiMiD[ZZZzzz] (n=TiMiD[
01:55:18DBUGEnqueued KICK TiMiD[ZZZzzz]
01:55:26TiMiD[ZZZzzz]and good night !
01:56:10 Quit _FireFly_ ("Leaving")
02:04:27amiconnOooh, there's an ooooollld Bug in kernel.c - the timer count is off by one
02:06:50amiconn-> For *all* platforms (!!) <-
02:07:37 Quit mrelwood ("CGI:IRC")
02:14:59amiconnTimer error: 1/15000 for archos Ondio and player, 1/13824 for archos recorders, and 1/3582 for H1x0
02:16:19amiconnMeaning, a clock would drift ~6 seconds per day on archos, and ~24 seconds per day on H1x0
02:39:03 Quit actionshrimp ("a bird in the bush is worth two in your house")
02:46:39 Quit paugh ("Leaving")
03:15:12***Saving seen data "./dancer.seen"
03:25:43 Quit dpassen1 ()
03:58:50 Join grimreap [0] (
03:59:40grimreapis anyone familiar with bookmarking on the ondio?
04:01:54 Join Paul_The_Nerd [0] (
04:04:54 Join webguest82 [0] (
04:05:36 Join QT_ [0] (i=as@madwifi/users/area51)
04:06:55webguest82is it yet possible to display album art while playing music on IHP 120?
04:08:04 Quit webguest82 (Client Quit)
04:08:20 Quit tvelocity ("Leaving")
04:16:13 Quit QT (Read error: 113 (No route to host))
04:32:38 Join hd [0] (i=hd@
04:33:14 Quit goa (Read error: 104 (Connection reset by peer))
04:33:51 Nick hd is now known as goa (i=hd@
04:36:38 Join webguest78 [0] (
04:38:54 Quit Strath (Read error: 104 (Connection reset by peer))
04:40:24 Quit webguest78 (Client Quit)
05:15:13***Saving seen data "./dancer.seen"
05:19:42 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
05:24:20 Quit lostlogic (
05:24:20 Quit yosemite (
05:24:40NJoinlostlogic [0] (
05:24:40NJoinyosemite [0] (
05:40:48 Join random_man [0] (
05:42:23random_mani have a question about the remote patch that is about rwps files
05:53:38 Quit random_man ("CGI:IRC")
07:00:57 Join amiconn_ [0] (
07:14:12 Join matsl [0] (
07:15:17***Saving seen data "./dancer.seen"
07:15:49 Part webguest23
07:19:23 Quit amiconn (Read error: 110 (Connection timed out))
07:19:23 Nick amiconn_ is now known as amiconn (
07:43:05 Quit webguest06 ("CGI:IRC (EOF)")
07:52:28 Quit matsl (Remote closed the connection)
07:59:04 Join B4gder [0] (
08:07:49 Join ender` [0] (i=ychat@
08:45:03 Join _Vladoman [0] (
08:45:06 Join Bger [0] (n=Bager@
08:45:35Bgerhi all
08:45:44Bgerwhat's the problem with irc logs ?
08:55:02BgerB4gder ?
08:56:03B4gderthe server is evil on me atm
08:56:07B4gderworking on it
08:56:16Bgerokay, 10x
09:02:15 Quit Vlad0man (Read error: 110 (Connection timed out))
09:15:19***Saving seen data "./dancer.seen"
09:37:56 Join ashridah [0] (
09:41:56 Quit Slasheri (Read error: 145 (Connection timed out))
09:54:44 Join Slasheri [0] (
10:09:15amiconnHmm, no LinusN :(
10:15:35 Join LinusN [0] (
10:16:05amiconnhi LinusN :)
10:16:13LinusNhi, saw that you missed me
10:17:11amiconnI've implemented a new timer handling when changing cpu frequency according to my old idea
10:17:48amiconnI found a really old bug in kernel.c / tick_start() while doing that
10:17:50LinusNah, did it work well?
10:18:17amiconnNothing really serious, but the timer count was off by one
10:18:23 Join JoeBorn-having [0] (
10:18:48amiconnI double-checked the CPU datasheets for that
10:19:05amiconnThe error was rather small on archos, but somewhat larger on iriver
10:19:06LinusNwhich platform?
10:19:21LinusNoh again
10:20:10amiconnCheck sh7032.pdf page 321 (in adobe reader) (10.9.1 Note on Waveform Cycle Setting)
10:21:07amiconnand MCF5249UM.pdf page 177 (in adobe reader) (11.5.2 Timer Reference Registers)
10:21:27amiconnI'd like you to have a look at my code before commit
10:21:59amiconnThere's one thing I don't like that much myself, but I have no idea how to change it without major code shuffling:
10:22:35amiconnThe tick timer is set up in kernel.c, but the prescaler adjustment on clock change happens in timer.c (together with the user timer)
10:23:05LinusNah, yes
10:23:08amiconnI could separate this into 2 functions, but that would produce more code
10:23:17LinusNleave it like that
10:23:54amiconnI've also adjusted the refresh counters according to the new frequencies
10:24:11amiconn(and for all frequencies on h100, they were somewhat off as well)">
10:27:26B4gderis there an actual reason why configure doesn't enable sound for the windows sim?
10:27:41B4gderI mean apart from it just not being done
10:30:35LinusNi dunno
10:30:55LinusNamiconn: did you change the meaning of CPU_FREQ vs FREQ?
10:35:19 Quit JoeBorn (Read error: 110 (Connection timed out))
10:37:09amiconnLinusN: No, FREQ still is the current CPU frequency, CPU_FREQ is the base frequency
10:37:24amiconnHowever, I changed the method how timer setup uses them
10:37:54amiconnTimer cycles are always based on the base frequency, the multiplier is handled with the prescaler
10:38:42LinusNok, so timer_period in backlight.c is always based on the base freq
10:38:45amiconnThat's what my patch is about: it allows on-the-fly adjustment without resetting the counter on frequency changes
10:39:14amiconnIt also dictates the precondition that the higher frequencies must be integer multiples of the base
10:39:31amiconnHence 48->45 MHz and 120->124 MHz (x4 and x11)
10:40:08amiconnYes. backlight.c still needs to boost the CPU because the timer period is so small.
10:40:37amiconnOther uses of the user timer, like the grayscale lib, do no longer need that
10:41:27amiconnOn frequency change, the period is of course not really stable, and if an interrupt would occur during the PLL locking period, it is delayed
10:41:35LinusNi think the patch look good
10:41:45LinusNof course
10:41:53amiconnThis is such a minor disturbance that it doesn't hurt the appearance of the grayscale
10:42:51LinusNthere shouldn't be a need to surround cpu_boost() calls with #ifdefs
10:43:02amiconnI chose to delay the interrupt because allowing it during PLL relock would most likely mean that the next period might be stretched even more
10:43:16amiconnThe #ifdefs in plugins are necessary
10:43:38LinusNwe should fix that
10:44:00amiconn...because 'rb->cpu_boost()' would resolve to 'rb->' which is obviously invalid syntax
10:44:33amiconnThere are 2 things in system.c that I couldn't check:
10:45:03amiconn(1) I'm not sure whether the 2 waitstates for lcd and flash at 120 MHz are still sufficient for 124 MHz
10:45:45amiconn(2) Is it really correct that the IDE timings for 48 MHz and 120 MHz (now 45 MHz and 124 MHz) are the same?
10:46:04 Join _FireFly_ [0] (
10:46:25_FireFly_good morning
10:54:00LinusN1) iirc, the flash timing has always been wrong, i just haven't cared
10:54:18markunamiconn: Is 33 MHz for the lower frequency maybe enough? Some > real-time codecs could maybe benefit from it.
10:54:52LinusN2) i believe so, but i'll double check
10:57:10amiconnLinusN: Btw, I have the stopwatch plugin running on recorder and H1x0 for almost 9 hours now, the iriver playing music in the background (.ogg) for the first hour, and my watch's stopwatch for comparison
10:57:24amiconnError is <2 seconds on both :)
10:58:02 Quit JoeBorn-having (Read error: 110 (Connection timed out))
10:59:06 Join JoeBorn [0] (
11:05:35amiconnLinusN: Do you think I should commit the timer patch first? Maybe it's easier to fix the waitstates and IDE timing after that?
11:06:08SlasheriLinusN: hi :) have you tried the new dircache patch? (
11:06:20LinusNamiconn: yes
11:06:27LinusNSlasheri: will do
11:06:30SlasheriLinusN: if you find it ok, i could commit it soon (maybe today)
11:06:33Slasherinice :)
11:07:48amiconnOkay, buckle up...
11:09:41 Quit linuxstb (Remote closed the connection)
11:13:13LinusNSlasheri: when is the cache built for the first time?
11:14:14LinusNin dircache_init()?
11:14:23LinusNsorry init_dircache()
11:15:20***Saving seen data "./dancer.seen"
11:15:47LinusNi see that you use the fastboot file to determine the initial cache size
11:16:13LinusNi will never select fastboot, and thus i don't want the file
11:16:23amiconnI suggested to use the config sector for the cache size...
11:16:30LinusNand then it will always scan at boot
11:16:34amiconnLinusN: Why not fastboot?
11:16:49LinusNbecause it gives me a nice consistency problem
11:16:50amiconnIiuc, it will *always* scan on boot now
11:17:05amiconn..the difference being that fastboot scans in the background
11:17:42SlasheriLinusN: that's true, even the fastboot entry is nomore available on the menu.. I think i should move the cache size indicator to the config structure
11:18:22LinusNok, so the scan is always done in the background?
11:18:29Slasheriamiconn: it always scans on the background (except the first boot()
11:18:39Slasheriit will not affect the boot time
11:18:47LinusNexcept the first boot?
11:19:07Slasheriyep, then we must determine how large the cache will get and we cannot do that on background
11:19:34Slasheri(because we need to reserve the space for the cache)
11:20:07LinusNok, so what happens when i boot the second time, and the cache isn't built? can i still access the dir?
11:20:44 Quit JoeBorn (Read error: 110 (Connection timed out))
11:20:47Slasheriof course, if cache is not available, it's just not used and normal direct dir access is used instead
11:21:03Slasheriso the switching between cache and direct dir functions is automatic
11:21:16LinusNok, so the cache file is only used to keep the size info when fastboot is off?
11:21:28Slasherionly the headers are read
11:21:38LinusNbut the entire file is written?
11:21:43Slashericorrect :)
11:21:53Slasheribut we can fix that
11:21:58amiconnI'd call that a debug feature ;)
11:22:29LinusNwhat about the fastboot? is it loaded and used while the scan is performed?
11:23:06Slasherino, and currently user can't even enable that feature (it's not compiled at all, ifdeffed out)
11:23:38Slasherifastboot info is only saved on shutdown (or usb connection)
11:24:26LinusNhow long does a full scan take?
11:24:41Slasheriit's very fast because now we have direct fat_* functions
11:24:51Slasheria few seconds should be enough
11:25:01Slasheriso it shouldn't even eat the battery (much)
11:28:28 Join linuxstb [0] (
11:32:00amiconngtg, bbl
11:32:06 Part amiconn
11:32:56LinusNSlasheri: i'm not that worried about battery, i'm more worried about latency and safety (if you drop it while the disk is spinning)
11:33:43Slasheriah, dropping is never a good thing.. but what do you mean with latency?
11:34:04Slasheriuser should not notice any delays with bootup whether the cache is enabled or not
11:35:35Slasheribtw, i think many users have currently long disk spindown time because that's necessary to browse the files without caching. However, after i enabled the cache, i also lowered the spindown time to almost instantly spindown the disk
11:42:49LinusNthat's a good thing
11:43:04LinusNi'll try it myself asap, but atm i'm busy
11:43:34Slasheriok, good :)
11:47:12solexx_Slasheri: thanks a lot for your work!
11:47:25solexx_I got a build with dircache yesterday and i am really impressed
11:48:16Slasherisolexx_: hehe, that's nice to hear :)
11:51:27markunSlasheri: Does aminconn's latest commit make it possible to remove the cpu_boost from backlight fading?
11:52:54Slasherimarkun: Hmm, it looks promising. Might be, but we need to try that (i can't right now)
11:53:06LinusN10.40.08 # <amiconn> Yes. backlight.c still needs to boost the CPU because the timer period is so small.
11:53:13markunAh, too bad.
11:58:15CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
11:58:15*Bger greps /dev/hda1 for one lost (because of forgotten password with AES encryption) file :(
12:02:12solexx_um, how do you grep for encrypted content?
12:02:38Bgerno, for unencrypted
12:02:46Bgerwith very little hope to find it
12:02:59SlasheriBger: eh.. if you have lost a password for AES volume, it's then completely impossible to retrieve any data from there with the present machines we have
12:03:13BgerSlasheri : in fact the problem is with one RAR file ...
12:03:52Bgerbut newer RAR (ver >= 2.9) uses AES for encryption
12:04:07Slasheriah, hmm
12:08:09Bgerwhile on this topic, i think it's good to have some kind of plugin for encrypting/(decrypting for viewing and after that wiping the info on the disk)... for sensitive data
12:09:08Bgermy bad Engl. in action:)
12:09:22LinusNplugin for rockbox?
12:09:29LinusNget real
12:09:55LinusNwhy would you want that?
12:10:31Bgerbecause i have such info on my iriver and i often give it to my friends ... i don't want to have them looking at this
12:12:18Bgerat least nothing is stopping me from trying to make such thing for my own
12:12:43solexx_and then you spend five minutes entering your passphrase with the on-screen keyboard? ;-)
12:13:29Bgersolexx_ yes, that's a problem, but sometimes it's better to do it on the player itself...
12:15:40LinusNBger: things like PIN's, passwords and stuff?
12:15:58LinusNthen write a plugin for that
12:16:08Bgerthat's what i plan to do
12:16:15LinusNnot a general encryption and wiping plugin
12:17:00Bgerhow do u see it then ?
12:17:09Bgersee it like ...
12:20:03LinusNi see it as a plugin for keeping pin's, passwords etc in an encrypted database
12:22:44 Nick solexx_ is now known as solexx (
12:26:51LinusNthe passphrase doesn't have to be an alphanumeric string, ot could very well be a series of combinations of joystick movements, this has been suggested by many
12:28:12linuxstbI think choosing characters from a keypad is a more efficient way of entering a long enough keyphrase. But I haven't done the maths.
12:28:39Bgerthe joystick movements are only 4...
12:28:58linuxstbI guess joystick movements will be coded like 3up, 2left, 4right etc
12:29:15B4gdermovements are probably harder to remember
12:29:36Bgerfor sure
12:29:39linuxstbI agree, it seems a nice idea, but I think the virtual keyboard will do a better job.
12:30:01Bgerexcept u're a musician ... up=C, down=D, etc ... :)
12:36:34 Join ilikedirt [0] (
12:41:46Slasherijust use morse code to enter the passphare
12:41:58Slasheribtw, i would like to use morse also for entering filenames!
12:42:09Slasherithat could be a nice thing to implement
12:43:53 Join Bger_cgiirc [0] (
12:44:29Bger_cgiirchehe the AES in flash:
12:45:01 Quit Bger ("BitchX-1.1-final -- just do it.")
12:46:01Slasherianother great feature would be to allow volume control while hold is on
12:47:41LinusNnot all people agree on that
12:47:56Slasheritrue.. maybe an option?
12:47:59Bger_cgiircoption, option :)
12:48:28Slasheriwhen i have the player on my pocket, it's hard to change the volume without accidentally pressing the joystick or changing tracks
12:49:28 Part ilikedirt
12:50:47*LinusN whispers "remote... remote..."
12:51:22Slasherihehe, yes the remote is good but then you will get pretty much "wired" everywhere ;)
12:59:53 Join amiconn [0] (
13:07:45 Join JoeBorn [0] (
13:11:54NibblerBger_cgiirc: scareh
13:15:21***Saving seen data "./dancer.seen"
13:23:52Bger_cgiircNibbler: of what
13:24:11Nibbleri did not understand too much of rijndael...
13:24:28Nibblerfor me its security by obscurity :]
13:24:52 Join Vladoman [0] (
13:25:28Bger_cgiircheh, it's not
13:25:57Bger_cgiircwhy obscurity ?
13:27:02Bger_cgiircit's simple : you don't know the key - you can't reveal the original data
13:27:59Bger_cgiircthis is rude explanation, of course
13:28:35HClyou mean crude? :p
13:28:59Bger_cgiircthis too
13:30:54Bger_cgiircI'm in top 3 of rockbox people with worst english:)
13:31:10Nibbleri am sure it is NOT "security by obscurity", cause the algorithem is known. just for me its so obscure i cant understand much of it :-)
13:32:07*Bger_cgiirc too
13:32:22Bger_cgiirci don't think it's explained very well
13:33:33Nibblerah, ok then :)
13:36:04Bger_cgiircit's not said what state is (i suppose plaintext)...
13:36:42Bger_cgiircalso i don't understand the operation in MixColumns...
13:37:07Bger_cgiircalso what's S-box
13:38:49 Nick phaedrus961 is now known as phaedrus96 (
13:39:51 Nick phaedrus96 is now known as phaedrus961 (
13:41:39 Quit _Vladoman (Read error: 110 (Connection timed out))
13:48:53 Join XavierGr [0] (
13:52:36amiconnBger_cgiirc: A suggestion for entering a code: Instead of a real password using the vkeyboard, you could use sequences, like left-down-left-left-right-up-down-select
13:53:11amiconnUnsing left/right/up/down give means entering 2 bits per button press, not too bad
13:53:26amiconn*Using, -give
13:54:06amiconnIm not sure how well such sequences can be remembered though
13:54:46amiconnFor decent security we would need >64 bits, meaning >32-button-sequence
13:55:14 Join Paul_The_Nerd [0] (
13:55:29*amiconn should read the logs more thoroughly
13:58:07XavierGrLogs are gone for the October.
13:58:08_FireFly_i have a test-version with combined-bitmaps in wps(only main-lcd) ready.
13:58:48_FireFly_if someone want to test it can be grapped here:
13:58:50Paul_The_NerdWhy not just use some public/private key method. Your data can't be decrypted on-box like that, but you'd be able to encrypt using open with making at least "securing" data on your box very simple?
13:59:10_FireFly_a test wps is included
13:59:25B4gderwhat's "combined" bitmaps?
13:59:45_FireFly_e.g. 4 bitmaps are in one bitmap
13:59:54Paul_The_NerdOne .bmp file, that you use regions of, instead of the whole image.
14:00:23_FireFly_the iriver firmware uses combined-bitmap e.g. for the file-format (ogg, mp3)
14:00:31XavierGrWhat about integrating this bitmap on the WPS file?
14:00:55XavierGror is this an idiotic idea?
14:01:11_FireFly_i don't understand what do you mean
14:01:18XavierGrLike comments.
14:01:29XavierGryou have the notmal WPS syntax.
14:01:44XavierGrand then you can have the bitmap in side the wps.
14:02:07XavierGrThen the code will check if the code has a bitmap segment and draw the pictures.
14:02:26Paul_The_NerdXavierGr: That'd make it awfully hard to edit the wps using a simple text editor
14:02:33XavierGrThough someone will have to copy the actuall bitmap data inside the WPS
14:03:41XavierGrWell it can clearly be after the wps syntax, but yeah maybe.
14:03:48thegeek_that sounds like a horrible idea
14:04:26thegeek_I agree there should be some "standard" for distributing a wps
14:04:30XavierGrok I should stop talking then :p
14:04:34thegeek_with folders for the imagee and stuff
14:04:41Paul_The_NerdI think integrating it into the .wps itself may not be the best idea. But perhaps removing the ability for a wps to load separate bitmap files
14:04:52thegeek_but combining the actual bitmap data and the wps config itself... bah
14:04:58Paul_The_NerdThen, the .wps will automatically load a same-named .bmp (Bob.wps + Bob.bmp) and you only use parts.
14:05:01B4gderI think we should (try to) load images from the same dir as the WPS itself is located in
14:05:05Paul_The_NerdIf you wanted to go the way of standardizing it.
14:05:31thegeek_I dont like that either Paul_The_Nerd
14:05:36thegeek_why make it harder?
14:05:50thegeek_keeping the images separate is best, why would you want to combine them
14:06:08Paul_The_NerdWell for one thing, less files.
14:06:16thegeek_hardly worth it though
14:06:24Paul_The_NerdDepends on how you define worth.
14:06:38thegeek_how would you combine them into one image?
14:06:51thegeek_the different images would have different sizes
14:07:02thegeek_so no winamp-style standard
14:07:19thegeek_that means you have to define x-y coords or something in the cfg
14:07:34thegeek_and you cant just copy out one image or replace one image with another
14:07:38_FireFly_look at my example wps in the zip fiel ;)
14:07:42thegeek_you have to edit that single image
14:08:05thegeek_just load images from the same folder as the wps itself, if that fails, try to load from "wpsname"-folder relative to the wps-config itself
14:08:12thegeek_dont make it harder than it has to be
14:08:24Paul_The_NerdSome of us don't want to have a separate folder for each WPS though thegeek_
14:08:38thegeek_then make a third option
14:08:50thegeek_try to load from /.rockbox/wps/images
14:08:53Paul_The_NerdI suppose, you could have bob.wps and then a dir named bob, under which all the BMPs were
14:08:54Bger_cgiircit seems the operation in MixColumns is xor ...
14:08:58thegeek_just make it fallback
14:09:08Paul_The_NerdBut bob.wps doesn't go *in* the bob folder
14:09:23thegeek_that was my second option there
14:09:31Paul_The_NerdThat way you have .rockbox/wps/bob.wps then .rockbox/wps/bob/*.bmp as the bmps it autoloads.
14:09:31thegeek_and yeah
14:09:34thegeek_that would work well
14:09:47Paul_The_NerdBut again, if you autoload files, then you can't use them inside the WPS as a-z A-Z
14:09:53Paul_The_NerdUnless you know what order they're loaded in.
14:10:02Paul_The_NerdYou'd have to address them by filename, and that bloats the .wps size
14:10:18thegeek_bloats the wps size..
14:10:26thegeek_that's bullshit
14:10:30thegeek_just keep the filenames short then
14:10:44Paul_The_NerdEven then, they'd have to be one-character to match it.
14:10:45thegeek_it's not like the wps-config-format is optimized for bytesize
14:11:02Paul_The_NerdIt's hardly optimized, but it's still limited to a set number of bytes.
14:11:16thegeek_increase that limit then;)
14:11:40Paul_The_NerdThe current system works.
14:11:46thegeek_indeed it does
14:11:54thegeek_and you complain that it's limited?
14:11:59 Join webguest71 [0] (
14:12:05Paul_The_NerdI never complained that it was limited.
14:12:27thegeek_you said it was limited to a set number of bytes?
14:12:31Paul_The_NerdAll I said was "If you want to standardize, using parting you could do it this way."
14:12:47thegeek_autoloading images is not a good idea
14:12:59thegeek_having to reference each image is not that horrible
14:13:30XavierGrwell thegeek speak for yourself, maybe others will see that as a good option
14:13:39XavierGr(though I don't know what I prefer)
14:13:51thegeek_but really
14:13:58thegeek_if you "autoload" images
14:14:12thegeek_you could very quickly end up confusing the different images
14:14:21Paul_The_NerdYou could use single character filenames for the autoload, but the problem is that in a non case sensitive environment, you're down to a-z rather than a-z A-Z
14:14:24thegeek_I dont see how having to reference each image is bad
14:14:36XavierGrthat why I awas talking about integrating the bMP inside the wps
14:14:52Paul_The_NerdThe problem with that is still that it's far too complex for most people.
14:14:55thegeek_that is still a horrible idea
14:15:02webguest71if you autoloaded ONE image you could use Firefly's method and "cut" every icon or other image out of te main one
14:15:03Paul_The_NerdThe WPS is still on the "difficult" end anyway, for a lot of users.
14:15:04thegeek_we have to think ahead a bit here too
14:15:11thegeek_you cant just put the image inside the wps
14:15:18thegeek_especially when we get colour-screens
14:15:22thegeek_and the images increase in size
14:15:57XavierGrthat has nothing to do with the wps.
14:16:09XavierGrthe images will still be there if they are in a saperate folder.
14:16:09thegeek_yes it does?
14:16:20thegeek_you said
14:16:21XavierGronly difference it will be on a single file. (all of them)
14:16:22thegeek_inside the wps
14:16:31thegeek_as in inside the wps file itself?
14:16:41thegeek_the wps file is a textfile
14:16:57thegeek_what will you do when we get full-colour screens
14:17:00XavierGrbmp is ascii characters
14:17:08Paul_The_NerdThere's still another issue.
14:17:11thegeek_and the images start getting really big
14:17:22thegeek_putting that inside a textfile
14:17:26thegeek_is just a horrible horrible idea
14:17:31XavierGragain the bmps will be after the syntax
14:17:34Paul_The_NerdWhy really bother with more stuff for the current WPS format anyway, in the end?
14:17:42thegeek_still horrible
14:17:51thegeek_keep images out of textfiles
14:17:54Paul_The_NerdWhen ideally someone could actually take advantage of the bitmap screens and do per pixel positioning.
14:17:54thegeek_why do it?
14:18:01thegeek_why not keep them as separate files
14:18:11webguest71i don't see the need for putting it in the wps - one text wps file and one graphics bmp file (same root name) is not that hard to handle
14:18:17Bger_cgiircOdds of being killed by lightning (per day) 2^33 hehe
14:18:22XavierGrWell yes I don't think that we need to change the wps format, just talking and once you mentioned combined images...
14:18:25_FireFly_oh no what have i done ?? ;)
14:18:43Paul_The_Nerdwebguest71: The problem with that is, you still have the issue of "what happens when individual images within it are different shapes"
14:18:48Paul_The_NerdYou ahve a lot of wasted space.
14:18:55thegeek_as I said
14:19:00thegeek_keep them as separate files
14:19:17 Quit Febs (Connection timed out)
14:19:21thegeek_perhaps make a better folder-structure, but that's it
14:19:33webguest71is that wasted space that critical?
14:19:52Paul_The_NerdWell, for one thing, I don't know how the parting code works yet.
14:20:12Paul_The_NerdSo, I don't know if you define upper left and lower right corner within the bitmap (which would mean that it'd be easy for people to use"
14:20:26_FireFly_am 5 files with one image inside are bigger than one file with all 5 images inside
14:20:27Paul_The_NerdOr if you define width and height of the tiles, and it auto-parses the .bmp
14:20:36thegeek_I dont see _any_ benefit to putting all images inside one large image
14:20:40thegeek_less files, but so?
14:20:43amiconnThe space-waste isn't the problem, but loading heaps of small files takes longer than loading one bigger file
14:20:57amiconnIt should also help keeping the .wps file size down
14:21:00thegeek_why is that?
14:21:03_FireFly_i have compared my play-mode.bmp with the 5 seperate files from which i have created
14:21:06amiconn...and reduce the file clutter
14:21:31_FireFly_the 5 files have a size 458 bytes (all summed) and the one file only 210Bytes
14:21:32thegeek_I think just organizing the wps system would prevent clutter
14:21:35B4gderbut make it harder to do a wps and harder to share images between wpses
14:21:38_FireFly_the size is reduced half
14:21:40amiconnI didn't say we should drop multi-file support, but of course we could do that with clipping support
14:21:46webguest71_FireFly_ I like it what you've done - could lead to different skins for a wps file
14:21:49webguest71_firefly_ obviously different wpses with different cuts would need different graphics
14:22:32amiconnKeeping multi-file support and adding clipping support is most flexible, dropping multi-file support in favour of clipping support would probably save some more
14:22:57_FireFly_my currently implementation have both
14:23:10_FireFly_it supports combined bitmaps and seperate bitmaps
14:23:11Paul_The_NerdYeah, I'm definitely not arguing for dropping multi file
14:23:26XavierGrthen its settled
14:23:26_FireFly_my example wps has both in use
14:23:51Paul_The_NerdI was just arguing one side in an "If multi file was going to be dropped" case.
14:23:52XavierGrwhen this is going for commitment?
14:25:23*Paul_The_Nerd still wishes %s didn't make the whole line scroll.
14:25:46B4gderI think it should
14:25:54B4gderbut I have an idea on how to make things better
14:26:05XavierGrwell he means that he wants to choose.
14:26:21Paul_The_NerdWell, for example, one way would be to have everything to the right of it scroll
14:26:28Paul_The_NerdYou want the whole line to scroll, you put it first.
14:26:38webguest71what ever happened to the idea about moving away from a line based wps system?
14:26:55XavierGrfor some occasions that you want: 1. (scroll)Title
14:26:56B4gdermy idea "fixes" that
14:27:04B4gderbut still remains line-based
14:27:26Paul_The_NerdWhat's your idea B4gder?
14:27:27XavierGrthen implement B4gder!
14:27:27B4gderby allowing the WPS to create a set of "boxes"
14:27:36B4gdereach box being line-based like today
14:27:53amiconnImho line based design is too inflexibe as soon as we allow multiple fonts
14:27:54Paul_The_NerdBut the question is, are you basing on a number of characters for width?
14:28:03B4gderso the current wps would be a box for the whole screen
14:28:14amiconnAh yes, the box design being what I am considering fro quite some time
14:28:20B4gdereach box would be pixel specified
14:28:44amiconnThat's quite different from what I have in mind though, and imho more complex
14:28:56B4gderwe'd only need to a way to specify "the following data will be within a box with this size/coords"
14:29:11webguest71there was a post about a box system that was backwards compatible with current system
14:29:12B4gderI've tried to come up with a way that would use more or less the current syntax
14:29:26amiconnI'd rather base everything on boxes, defined by pixel coordinates
14:29:51B4gderwhy not line-based within the boxes?
14:30:13Paul_The_NerdSo, you're basically implementing an overly simplified CSS box model... Or perhaps arbitrary frames. :-P
14:30:14amiconnThen we have lines in boxes, awfully ugly to handle for scrolling
14:30:16B4gderI bet most people will think lines wthin them
14:30:31B4gderamiconn: no, you'd convert those to pixels for the scrolling
14:30:51amiconnImho a box should always contain one line of text only, possibly scrolling, and using the same font
14:31:00amiconnDifferent boxes could use different fonts
14:31:13B4gderthen you can't do the current wps syntax
14:31:19B4gderwhich my suggestion would
14:31:39amiconnYou want backwards compatibility at the cost of increased complexity?
14:31:42B4gdermy would even work with the exact wps files of today
14:31:49Paul_The_NerdYou could always write a new wps handler. .wp2 files, or something
14:31:53B4gderI don't see it as increased complexity
14:32:00Paul_The_NerdAs it is, my computer always wants to open my .wps in OpenOffice.
14:32:58 Quit webguest71 ("CGI:IRC")
14:33:27amiconnThe single-line box design would make things rather simple when it comes to things like multiple fonts, font attributes, foreground & background colours/shades, scrolling attribute etc
14:33:48B4gdernot really
14:33:50amiconnEach box would have its set of attributes
14:33:57B4gderonly if you'd allow different ones within the same box
14:34:10amiconnNo, that's the whole point of my suggestion
14:34:26B4gderbut then why rule out having more than one line using the same style?
14:34:26amiconnYOu would need to allow that with multi-line boxes
14:35:04amiconnWhy would I rule this out?
14:35:14B4gderI don't, you did
14:35:22B4gderyou said it would only be one line
14:35:27amiconnNo I didn't
14:35:32B4gderok, then I misunderstood
14:35:38B4gderthen our suggestions are basically the same
14:35:38amiconnYes. You would simply define a second box
14:36:09amiconnIn fact we can combine both suggestions with keeping the current wps format
14:36:12B4gder...and current WPS would simply be a full screen box
14:36:25B4gderwhy not?
14:36:41amiconn...combined with my idea of coordinate changes not applied until explicitly specified
14:36:49amiconnMy idea is now as follows:
14:37:25amiconnEach line in a wps implicitly starts a new box definition (unless it's a special line loading an image or such)
14:38:07amiconn...but as long as you don't specify coordinates, this box would be x=0, width=lcd_width, y=old_y+font_height, height=font_height
14:38:50amiconnThis way, every line of a 'classic' wps would be its own box, without a format change
14:39:09amiconn...and it's still possible to define boxes by special tags used at the start of the line
14:40:19amiconn...and in these tags, every coordinate value that's not explicitly specified would preserve its previous value (in a logical way, meaning the y coordinate will be incremented by font_height)
14:40:54Bger_cgiircamiconn: i see it this way too
14:41:22XavierGrall these sound great, but isn't a pain to implement all that?
14:41:26B4gderwell, the result would be the exact same as in my way of seeing it
14:41:44B4gderbut I would consider them all to be lines within a single box
14:41:50amiconn...the same way as my suggestion for partial images is to leave out coordinate values that didn't change from the previous declaration, in order to save space
14:42:13amiconnUnfortunately irc logging is broken :(
14:42:49_FireFly_in my current implementation the leave out of some values and insteed use the last one doesn't work very well
14:42:51amiconnB4gder: Internally there's a big difference, because seeing each line as a separate box will make things easier, imho
14:43:01 Join muesli- [0] (
14:43:02amiconnFrom a user's perspective, they are the same
14:43:03_FireFly_mybe i have make a failure
14:43:10B4gderwell I don't see the hard connection between the WPS and the internal representation
14:43:28B4gderbut it doesn't matter
14:44:54amiconnMy suggestion would e.g allow a rather elegant way of declaring a 2-column wps, even when both columns use fonts with different heights
14:45:26amiconnOnly 2 boxes would need to be declared explicitly, and not even fully
14:45:38B4gderso would it in my way too ;-)
14:45:50B4gderbut again, they're almost the same
14:46:06amiconnThe first line of the left column would set the width to lcd_width/2, then all lines for the left colum follow
14:46:25amiconnThe first line of the right column would need to set x and y (only)
14:46:27 Nick TiMiD[ZZZzzz] is now known as TiMiD (n=TiMiD[
14:46:27DBUGEnqueued KICK TiMiD
14:46:54_FireFly_hi TiMiD
14:46:57Paul_The_NerdHello TiMiD
14:46:59XavierGrhi TiMiD! any update on the remote?
14:47:14TiMiDslowly progressing
14:47:29TiMiDI'm documenting undocumented code parts
14:47:30XavierGrwhat's your status currently?
14:47:44muesli-any progress is progress ;)
14:48:07TiMiDI have the directory browsing working but far from complete
14:48:15XavierGrI mean what can you render on the remote right now?
14:48:21TiMiDsince a lot of things are unneeded in the code now
14:48:45XavierGrso you have filetree rendering on the remote?
14:48:59TiMiDbut it's not clean
14:49:21TiMiDI must change lots of things in apps using the tree infos
14:49:39TiMiDbecause the data structure is a little different
14:49:40Paul_The_NerdQuestion: Were you planning on using two separate wps files, or one file with two parts?
14:50:14TiMiDme ?
14:50:19amiconnB4gder: Perhaps I'm just thinking more about the internal representation...
14:50:23XavierGrI think that keeping the same wps (and dropping the lines that don't fit) is good.
14:50:41TiMiDI think 2 wps would be very good :)
14:51:01CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
14:51:01*Paul_The_Nerd prefers two files as well.
14:51:02amiconnTiMiD: Gets my vote
14:51:08_FireFly_a similar thing which i have made??
14:51:31 Join preglow [0] (
14:51:44 Quit muesli- (Read error: 104 (Connection reset by peer))
14:52:11TiMiD_FireFly_: well when I will be finished with tree, and if what I've done is accepted, maybe you could try to port wps to the remote the way I did it with filetree
14:52:26_FireFly_i could try :)
14:52:31TiMiDsince you may be more familiar with the code than me :)
14:53:16TiMiD(and there is a lot of work remaining, too much for me :P)
14:53:52TiMiDtree and wps are the most difficult parts I think
14:54:16TiMiDmainly because code is not very well documented and it takes a lot of time to understant how it works
14:54:46B4gderI'm not sure the wps should/could be done the same way as the tree code
14:56:18TiMiDI don't know since I didn't looked closely at the code
14:56:59XavierGrtree is the most difficult part I think.
14:57:17XavierGrI was lost when I attempted to make a remote patch.
14:57:35TiMiDbut the principle is simple : be able to render on any screen independantly and then call the rendering process nb_screens times :)
14:57:45TiMiDXavierGr: same thing here :p
14:57:53TiMiDit's like hell ^^
14:58:12XavierGrbut I think that menu.c will be easy to do.
14:58:22TiMiDyes since it's hard_coded
14:58:35TiMiDI wans thinking of something crazy
14:58:41TiMiDfor the menus
14:58:47TiMiDsince we have dircache
14:58:55TiMiDmenus could be just files and dirs
14:58:57XavierGrLinusN: I just sent you a mail with the latest radio.patch. I really hope that you will find some time to inspect it.
14:59:55TiMiDfiles would describe options and would be read by a viewer
15:00:16TiMiD(but it's just an idea)
15:00:56Paul_The_NerdWhat's the benefit in doing that?
15:01:28TiMiDPaul_The_Nerd: ultra-customizabel menus
15:01:34amiconnTiMiD: That's a really odd idea
15:01:36Paul_The_NerdHow so though?
15:01:51Paul_The_NerdI mean, even with the files, something still needs to interpret and handle them
15:01:57amiconn...and I can only find reasons against it
15:01:59Paul_The_NerdThat means that you can only have options relating to what you've planned for.
15:02:08Paul_The_NerdWhich is the same as having those options static and present.
15:02:47amiconn...the most important being that the dircache isn't mandatory (luckily), and it's not available on all platforms
15:03:05TiMiDamiconn: ok so it's impossible
15:03:44Paul_The_Nerdamiconn: What are the disadvantages of the dircache itself, do you know?
15:03:48TiMiD(btw, I wasn't planning on that, just an idea to remove most of the menus code
15:03:59amiconnPaul_The_Nerd: Yes, it needs RAM
15:04:14TiMiDbut it should be possible if the dircache only caches the menus rep all the time
15:04:16XavierGr<400kb though
15:04:32Paul_The_NerdWhere does it take it from, and how does this impact the average user?
15:04:47amiconnTiMiD: Currently the menu structures are compiled in, and they're const
15:04:59XavierGrwell smaller ram means a little bit smaller buffer.
15:05:07linuxstbPaul_The_Nerd: The audio buffer. So the disk will have to spin up more often during playback.
15:05:11amiconn...meaning that they can be read from ROM on archos when using rombox
15:05:11XavierGrso this means that the disk will spin up a little more.
15:05:21XavierGrBUT the disk will not spin up when browsing
15:05:23amiconn...leaving precious RAM free for buffering
15:05:30TiMiDamiconn: I know, I looked before starting working
15:05:41XavierGrand if you browse a lot... well you get the idea.
15:05:52TiMiDI browse a lot :)
15:06:01amiconnXavierGr: I know. It means that it depends on usage pattern
15:06:11TiMiDbut I don't think it will improve the battery lifetime
15:06:15XavierGramiconn: off course
15:06:30amiconnI don't browse a lot, most of the time I just resume after startup (possibly after loading a different .cfg - from the root)
15:06:38TiMiDbecause when you browse it's to play a file and the disk still needs to spin up ...
15:07:25XavierGryes that's true but thats not the case with the radio, or when you are searching for something to play.
15:07:57Paul_The_NerdSo, Dircache is a convenience feature at a slight impact to battery life (most likely)?
15:08:32amiconnDircache is a convenience feature that might increase or decrease battery runtime, depending on usage
15:08:53XavierGrwell said
15:09:03preglowooh, new patch
15:09:05Paul_The_NerdBut you'd have to do a lot of browsing relative to listening for it to have that sort of impact, no?
15:09:18preglownice, can't contact my dev box...
15:09:24XavierGryeah maybe.
15:09:30Paul_The_NerdThough I imagine it'd be a very marginal difference either way.
15:09:47linuxstbSlasheri (I think) also made the point that some people have a long spindown timeout to help increase browsing speed. Those users can now reduce the spindown time to almost nothing.
15:09:55XavierGrbut even then the battery wasted on the cache I think is very small.
15:10:19B4gderif you use a 1MB cache, that could be a minute of music
15:10:21XavierGrof course. That's what I did when I used the dircache!
15:10:29amiconnDepends on how large the cache is relation to the main RAM
15:10:44XavierGrbut the cache is lower than 400 kb
15:10:49amiconnThinking about H100 having only 16MB
15:10:56XavierGryeah on archos it might have major difference.
15:10:57TiMiDpeople who browse a lot (like me) don't use the full cache (I change music every 1 or 2 songs) :)
15:11:07linuxstbI think the other main argument against it was that it increases code complexity and therefore the chances of bugs.
15:11:11amiconnOn archos the full dircache is a no-no
15:11:29XavierGrso dircache is iriver specific?
15:11:37amiconnlinuxstb: yes, that too
15:11:39B4gder>8MB model specific
15:11:56Paul_The_NerdI don't browse much, but I hit next track *alot* at about halfway.
15:12:04XavierGrwhat were they thinking. 2MB ram?
15:12:12Paul_The_NerdI probably have the worst possible battery life use pattern.
15:12:19amiconnIt _might_ be feasible on archoses with 8MB mod, if you browse like hell
15:12:24B4gderXavierGr: they were the 2nd commercial mp3 player available
15:12:51*dwihno is also curious
15:13:26Paul_The_NerdMy boss actually got angry with me when I tried to explain to him that the iPod was not the first MP3 player, and in fact nowhere near it. He was so sure of himself.
15:13:41amiconnPaul_The_Nerd: Even when hitting next track, the dircache helps saving a tiny bit of battery, because the directory information is read from the cache instead of the disk
15:13:50XavierGrdont talk to ipod fan boys.
15:13:57B4gderHm, zagor wrote a page once about the very first players...
15:14:00dwihnoI got a parallell RIO 300 (I think it was) before the new millennium
15:14:09dwihno32 meg internal storage
15:14:18amiconnIirc archos players were around back in 2000...
15:14:24Paul_The_NerdSaehan's MPMan <−−- 1998
15:14:36dwihnoPaid roughly 1500 SEK for 32 megs of smartmedia ;)
15:14:45dwihno(divide by ~7 for USD)
15:14:48Paul_The_NerdApparently released in the US under a different brand, but same product name a few months before the Diamond Rio, commonly considered the first.
15:15:14Bger_cgiircLinusN: any news on h3x0 ?
15:15:17Paul_The_NerdXavierGr: I actually just asked him why he picked it over other players.
15:15:23***Saving seen data "./dancer.seen"
15:15:28Paul_The_NerdHis response "It's the first, so you know it has to be the best."
15:15:39Bger_cgiirc("THE question")
15:15:46Paul_The_NerdI know
15:15:52 Quit ashridah ("Leaving")
15:15:59Paul_The_NerdI *almost* asked him if he still watched Black and White television.
15:16:24Bger_cgiirchahaha :)
15:16:48Bger_cgiircthis is a HIT
15:17:25amiconnSlasheri: How much does dircaching add to binary size?
15:17:50_FireFly_ah i found the bugin my parsing code for creating parts of combined bitmaps;)
15:18:27B4gderah, _there_
15:18:43_FireFly_now it works also correct when some values are left off so that the last one will be used
15:19:33dwihnoZagor: That was cool!
15:19:48_FireFly_e.g. %xp|i|v|0|0|12|16|0|13|
15:19:56_FireFly_this works now correctly
15:20:24Slasheriamiconn: Hmm, i will measure that when i get to home (soon)
15:21:04Slasheriof course it increases the binary size but only for iriver (so that shouldn't be a problem)
15:21:30amiconnYes of course. I'm just interested in how much it is
15:21:51amiconn(related to my remark about 8MB-modded archoses)
15:22:02Slasheriah, yes
15:22:18Slasheribut now i will go, cu later ->
15:22:25_FireFly_amiconn with my remote-patchset+dircache+combined-bitmaps the size of the rockbox.iriver is 265632 bytes
15:23:34linuxstbThe last rockbox.iriver I built was 246472 bytes.
15:23:35XavierGramiconn: what is wrong with binary size, I thought you solved that?
15:24:23B4gderit cannot be solved
15:24:34B4gderits a limit in what files the Archoses load or not
15:24:42_FireFly_each new feature will increase the binary size
15:24:51amiconnWe now know that the limit fr fmr, v2 and Ondio is much higher than we thought
15:24:52B4gderreally? ;-) now the most problematic target is the recorder v1
15:25:50XavierGryes I know that a new feature will set the binary size to a greater value. I just thought that you found something to by pass it. At least temporary.
15:26:02Paul_The_NerdI thought there was some talk about having the firmware itself be stripped down, and have it load the rest... or something like that?
15:26:27B4gderuh, right, _that_ is actually the only fix
15:26:38XavierGrso how much binary size is left for the v1 recorder?
15:27:08amiconnEven if we don't hit the limit, or work around it, too large binaries are no good
15:27:26amiconnThey eat precious RAM (not all archoses can be flashed)
15:27:52XavierGr248.500 with the fm preset patch.
15:28:02amiconnXavierGr: Player and recorder v1 allow for 200KByte, fm recorder, recorder v2 and Ondios allow 400KByte
15:28:06XavierGr2 kb
15:28:39XavierGrdoes a plugin has to do with the binary size?
15:28:42TiMiDare you using gcc -Os ?
15:28:46Bger_cgiirchm, Rio is out of the DAP business
15:29:05XavierGrI use devkit which hasnt the latest gcc
15:29:28TiMiD-Os is available in old gcc :)
15:29:47XavierGrI thought you meant gcc version.
15:30:21TiMiDI don't think it's activated in the makefiles I saw (but there must be a reason for that because when I tried to use -O2 the firware had strange behaviours)
15:30:40amiconnTiMiD: No, we're using plain -O
15:30:59amiconn(and some specific flags, lie -fomit-frame-pointer)
15:31:21TiMiDwould it work with -Os ?
15:31:53amiconnWith -O2 and -Os gcc goofs on some optimisations, so rockbox crashes on archos
15:32:33amiconnI know a workaround which works with gcc 3.3.x, but gcc 3.4.0 and higher have more problems which I didn't find yet
15:32:56XavierGrso does a plug-in adds binary size?
15:33:05amiconnXavierGr: No
15:33:34XavierGrgreat. Then it shouldnt matter if my jpeg scrolling patch gets comitted.
15:34:09amiconn(as long as it doesn't require additional core functions to be exported, in which case the plugin api struct size increases a bit)
15:35:27amiconnPlugins have a size limit themselves
15:44:45Bger_cgiircany idea what does "static u8 ...[] __initdata" mean ?
15:44:55Bger_cgiircin linux kernel source
15:45:32linuxstbMy understanding is that it tells the linker to place it in the __initdata section.
15:45:58Bger_cgiircwhat's specific about this section ?
15:45:59B4gderyeps, used by the module loader iirc
15:46:30Bger_cgiircmemset-ed to 0? or?
15:50:22B4gderit is described in Documentation/Docbook/kernel-hacking.tmpl
15:51:51 Part amiconn
16:03:32 Quit B4gder ("time to say moo")
16:09:00Bger_cgiirc*that* is quit message
16:14:51XavierGrI agree.
16:52:06 Join muesli- [0] (
16:53:18Paul_The_NerdWelcome back.
16:53:27muesli-paul ;)
17:00:43 Join muesli__ [0] (
17:00:50muesli__re re ;)
17:07:41 Join JoeBorn-having [0] (
17:12:57 Quit JoeBorn-having ("")
17:13:19 Join Ismo [0] (
17:14:07muesli___FireFly_ didnt you like my request to increase rpws's file size :O ?
17:14:40TiMiDI'm more and more in favor of rewriting the whole tree.c . id3db.c, the code is insane :(
17:15:27***Saving seen data "./dancer.seen"
17:18:52XavierGrThat would be huge work!
17:19:09XavierGrwhen you say insane you mean bad written?
17:19:26 Quit muesli- (Read error: 104 (Connection reset by peer))
17:24:11 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
17:24:12TiMiDat the beginning, it should have been a pretty small code
17:24:20 Quit JoeBorn (Read error: 110 (Connection timed out))
17:24:21TiMiDbut now it's a 1500+ lines monster
17:24:31TiMiDand lots of programs interract with it
17:25:01TiMiDthe way the interractions are handled is not very well done (duplicated code and so on)
17:26:16TiMiDI have the feeling that rewriting it using the lists widgets natively would be a good thing (cleaner code)
17:26:34TiMiDalos, separate file tree code from id3db code
17:28:35_FireFly_muesli__: no i have currently no time to increase it ;) I'm playing with combined bitmaps
17:30:00 Join Vlad0man [0] (
17:30:02muesli__for those who care.. if your file is bigger than the limit there are mistakes on the screen
17:30:17_FireFly_muesli__: do you use my compiled version ??
17:33:06_FireFly_so size is to 1200 increased and it builds currently
17:33:14TiMiDwell ugly-hacked tree browser seems to work well
17:33:22muesli__cheers mate :D
17:33:25TiMiDtime to catch my train !
17:33:34TiMiDcu !
17:33:43muesli__bye thegeek_
17:34:04 Nick TiMiD is now known as TiMiD[ParisAgain (n=TiMiD[
17:34:04DBUGEnqueued KICK TiMiD[ParisAgain
17:34:25muesli___FireFly_ this one?
17:35:41XavierGrDoes the audio form the radio goes to the MAS?
17:36:35_FireFly_muesli__: yepp now updated
17:37:10_FireFly_it has also the latest cvs-changes included
17:37:22muesli__extra- :DDD
17:38:01 Join einhirn [0] (
17:38:26Paul_The_NerdWhat patches does that have on it?
17:39:56 Quit grimreap (Read error: 110 (Connection timed out))
17:41:57muesli__hm...just experienced a strange thing... can *.wps and *.rwps share bmps?
17:42:36_FireFly_muesli__: they are complete independend
17:43:14_FireFly_Paul_The_Nerd: this build has the latest cvs-changes + my remote-patchset + dircache_rev2
17:43:17muesli__because when i activate a rwps my main screen changes
17:43:40Paul_The_NerdThank you _Firefly_
17:44:48_FireFly_how does it change ??
17:45:50muesli__it affects the main lcd
17:46:02muesli__treats it like the remote
17:46:14muesli__but maybe theres an error in that file..have to check out
17:46:24Paul_The_NerdWait, so you can change the remote's WPS now?
17:47:01_FireFly_and have a seperate control over statusbar scrollbar and line-selector
17:47:50Paul_The_NerdI mean, I was happy with the remote's WPS
17:48:01Paul_The_NerdBut I was tired of seeing (No ID3) over, and over, and over, and over...
17:48:02Paul_The_Nerdad infinitum
17:49:06 Quit Vladoman (Read error: 110 (Connection timed out))
17:49:27markunDoes anyone know why there is no make on the tools dir during build?
17:49:53_FireFly_you have do it your selfe
17:49:57muesli___FireFly_ have to check out my rpws..guess it contains still some errors
17:50:52_FireFly_yepp i think it ,too because i have just tested the build
17:50:59_FireFly_and i had no problems
17:52:15muesli__i'Ve tried to transfer wps into my rwps...still quite buggy cause i have to understand what the code itself does ;)
17:58:31Paul_The_NerdHow do I enable dircache, by the way?
17:59:08_FireFly_it gaves a menu-point for that
17:59:21Paul_The_NerdYeah, but I can't find it in the menus... I must be blind or something.
17:59:47_FireFly_it's the same menu where you set up the spindown-time for the disk
18:00:02Paul_The_NerdThank you.
18:00:11Paul_The_NerdAre all wps tags valid in rwps?
18:00:28_FireFly_yepp it should be
18:00:43_FireFly_these tags, which i use in my rwps are working
18:00:53_FireFly_at least
18:02:51Paul_The_NerdThank you very much
18:03:38paughergh.. i enabled resume on start but now i'm stuck on a bad file. is there another way to unstick this other than removing that file via usb?
18:03:56paughit just locks as soon as i start it (iriver h140)
18:04:06XavierGrFirefly why dont you make some settings for the scrolling of the remote?
18:04:42XavierGrpaugh: Start original firmware and then delete that file.
18:04:48XavierGrthen start rockbox
18:05:10paughXavierGr, ah ok. cheers.
18:05:46XavierGrFirefly: It is is to add and it makes the remote much more usable. Currently the scrolling is very slow imho.
18:05:47Paul_The_NerdI haven't encountered a music file that's frozen me consistently ever.
18:05:50 Join arkascha [0] (
18:06:08XavierGrI can send you a dif with the changes I made in settings.c and settings.h which enables scrolling.
18:06:31XavierGrProbably you don't use p2p for music :P
18:06:43XavierGr(though me too, I have never encountered this)
18:07:25Paul_The_NerdI primarily use "Insert CD in drive." "Rip CD using dBPoweramp to lossless wavpack" "Enjoy"
18:08:02paughPaul_The_Nerd, i seem to have stacks of them. mostly oggs. probably encoded with an old version of the encoder.
18:08:46_FireFly_XavierGr: send me the diff :)
18:09:01Paul_The_NerdI hate downloading music anyway
18:09:03XavierGrwait a sec
18:09:15Paul_The_NerdI once found an album I really *really* wanted as FLACs and couldn't resist downloading it
18:09:37Paul_The_NerdUpon listening to it, it was abundantly clear it had been originally encoded in a lossy format at a relatively low bitrate, then transcoded to FLAC.
18:10:10_FireFly_there have someone missunderstood something ;)
18:10:13paughyeah. weird that people do that
18:10:51linuxstbIt probably went from CD -> MP3 -> CD -> FLAC.
18:10:54Bger_cgiircin fact, someone asked here how to convert ogg (or it was mp3) to flac ...
18:11:06Paul_The_NerdThat's actually a pretty likely case linuxstb
18:11:22Paul_The_NerdThat's a little more understandable.
18:11:35linuxstbI know. It's how a lot of people copy CDs - because it's the default settings in their software.
18:11:41XavierGrhmm I have a very old version of the files. I will try and send you the specific changes.
18:12:30linuxstbDoes anyone know how good (or bad) the "secure" mode in the Mac version of iTunes is?
18:16:19fuzzieif you mean the 'error correction', it's bad
18:17:09XavierGrPaul_The_Nerd: Why wavpack instead of FLAC?
18:18:25 Join webguest76 [0] (
18:19:24linuxstbI assume it's because wavpack decodes a lot more efficiently in Rockbox.
18:19:29Paul_The_NerdAt the time FLAC was still not decoding realtime consistently
18:19:35Bger_cgiircXavierGr: it seems that wavpack is better than FLAC
18:19:38Paul_The_NerdI'd get pops and gaps in the music
18:19:38Bger_cgiircand also makes smaller files
18:19:59Bger_cgiirc(in most cases)
18:21:10Paul_The_Nerd%s%?ia<%ia|%?d2<%d2|%?d1<Rock · box|>>>
18:21:11Paul_The_Nerd%s%?id<%id|%?d1<%d1|xob · kcoR>>%?iy< (%iy)|>
18:21:13Paul_The_NerdCan anyone explain to me why if I don't have an extra line between these, they display on the same line?
18:22:18XavierGrFireFly: I can't make an exact patch to apply only the remote_scroll changes. But it is easy to understand which changes are for the remote. So you can add them manually.
18:23:39webguest76How would I get a plugin to play a specific mp3
18:23:39webguest76from a directory?
18:25:09webguest76sorry that was meant to say: HI can anyone tell me how I would get a plugin to play a specific mp3 in a specifi directory?
18:25:22 Quit einhirn (Read error: 113 (No route to host))
18:26:56linuxstbwebguest76: I don't think you can. Look in plugin.h for the functions available to plugins.
18:29:33XavierGrFirefly: sorry for the PM you must be registered to answer me.
18:30:25XavierGrThere don't try to patch it as is. Just add manually the '+' statements to your code. The file states which file to change.
18:30:52XavierGrAlso some options maybe will be already there by you, so take those that you need.
18:34:46 Quit webguest76 ("CGI:IRC (EOF)")
18:36:52XavierGrIf you need help just ask.
18:40:27Bger_cgiircis there any routine for converting big endian to little endian (32bit) and reverse ?
18:41:54linuxstbBger_cgiirc: No - but there should be.
18:42:24Bger_cgiircthere are SWAB32 macroses..
18:42:27linuxstbThere are some functions (SWAB???) in the firmware directory somewhere, but they are not general purpose.
18:42:39linuxstbBut look at how they are defined.
18:43:19linuxstbIIRC, they convert little-endian to host-endian. Which isn't the posix definition of SWAB.
18:44:23CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
18:44:23*Bger_cgiirc must understand portions of linux/crypto/aes.c
18:44:35 Quit muesli__ (Read error: 104 (Connection reset by peer))
18:46:01 Join matsl [0] (
18:47:50linuxstbYes, but they are only defined on big-endian targets. For little-endian targets, they do nothing.
18:49:47linuxstbBut they may suit your needs - if you want to convert from little-endian to host endian.
18:50:03linuxstb(or vice-versa).
18:50:09Bger_cgiircok, i must write cpu_to_le32() and le32_to_cpu()
18:50:37Bger_cgiircwhich on little endian is just #define xxx(x) (x)
18:51:25Bger_cgiircand on big endian is what SWAB32() does... or i'm wrong ?
18:52:59linuxstbYes. I think SWAB should be defined on all targets, and simply swaps the byte order (using the existing asm-optimised routines). You can then define your set of byte-swap functions as macros that either use SWAB or do nothing.
18:53:23linuxstbBut that would mean changing any existing calls to SWAB in rockbox.
18:53:50linuxstbWhich is why I didn't do it when I found hte same problem a while ago - it touches too much code I'm not familiar with.
18:54:02Bger_cgiircaha :)
18:54:04Bger_cgiirclike fat.c :)
18:54:19linuxstbYes - I don't want to go there....
18:54:41Bger_cgiircfor sure
18:55:01Bger_cgiircso, it's ok to #define cpu_to_le32(x) SWAB32(x) ?
18:55:06linuxstbBut I will need to soon for the ipod - it's the first little-endian Rockbox target.
18:55:08Bger_cgiircand reverse
18:55:37linuxstbYes, that's how I understand those functions.
18:55:41preglowcan't you switch endiannes for arm?
18:55:54linuxstbI think it's less efficient running big-endian.
18:56:16linuxstbBut yes, there are commands to change the endianness.
18:57:10linuxstbBut I'm sure Rockbox is already quite safe in that respect, and I'll be looking out for problems as well.
18:57:16preglowyeah, me too
19:01:24Bger_cgiirchm, any standard way of getting 32bit int in rb ?
19:02:06linuxstbI know it's old news but I can't believe the "Zen patent":
19:03:10linuxstbBger_cgiirc: firmware/includes/inttypes.h
19:03:49XavierGripod is little endian?
19:03:54linuxstbXavierGr: Yes.
19:04:08XavierGrI thought that apple tends to make big endian.
19:04:15linuxstbNot any more.
19:04:37linuxstbTheir file formats are big-endian though.
19:04:39XavierGrIs there a reason behind this change?
19:04:50XavierGrand their macs
19:05:01linuxstbI don't think they decided to change - it's just the available processors.
19:05:02fuzziethe macs not for much longer, though
19:05:24 Quit Maxime (Read error: 104 (Connection reset by peer))
19:05:29Paul_The_NerdI can't believe they went x86.
19:05:36XavierGrthey changed their CPU company?
19:05:42fuzziethey're moving to Intel
19:05:49fuzziechanging archs
19:05:51XavierGrx86 no way I have missed some episodes?
19:05:58 Join Maxime [0] (
19:06:07XavierGrDidnt they had motorolla ones?
19:06:31fuzziethey're currently motorola/ibm powerpc users, yes
19:07:38_FireFly_the problem was/is that motorola/ibm couldn't increase the cpu power to the level apple wants
19:07:58_FireFly_somthing about doubleing the cpu power
19:08:14crwlthat power per watt thing
19:08:21linuxstbInteresting that Apple are now competing with Microsoft on a level playing field.
19:08:34_FireFly_not really
19:08:37fuzziethe problem is that no-one wants to make powerpc chips for laptops.
19:08:54_FireFly_MACOS will only run on mac
19:09:17Paul_The_NerdThat isn't true
19:09:26fuzziein theory
19:09:27Paul_The_NerdMacOS will only *officially* run on Mac hardware
19:09:38fuzziethe recent development builds certainly seem quite solidly locked
19:09:50_FireFly_they want to include a lock into there os and bios so that it can only run on MAC-pcs
19:09:52Paul_The_NerdI've read quite a few faqs on how to bypass that though.
19:10:03fuzzieyou haven't. :)
19:10:17fuzziethe original intel/mac builds didn't lock themselves to the TPM in any meaningful way
19:10:23fuzziethe recent ones do
19:10:30Paul_The_NerdI'll admit it's been a little while.
19:10:32 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
19:10:44fuzzieit might well end up workedaround, but we'll see.
19:10:50Paul_The_Nerd'eh. I'm quite certain someone will find a way anyway.
19:10:56linuxstbThat's not the issue - it's the cost of hardware to Apple that I'm referring to. They should now be able to build Macs with similar specs and prices as companies like Dell.
19:11:17Paul_The_NerdI don't think you were ever *really* paying for the hardware.
19:11:26Paul_The_NerdYou were paying for the design.
19:11:28fuzzieyes, their move to x86 seems somewhat suicide in that regard
19:11:39fuzziethey can't sustain 30% profit margins if their specs can be directly compared to Dell
19:12:13linuxstbPeople will be happy to pay a premium for Macs - is just that they will now be in the same ballpark performance wise.
19:13:46XavierGrWho buys Macs anyway???
19:13:57crwlwhy not?
19:14:02linuxstbNo-one - they're too expensive :)
19:14:07fuzzieI'm typing on an iBook right now. :)
19:14:20Paul_The_NerdI know many more iBook users than Mac desktop users.
19:14:26crwlimho you can't get better (12") laptop for 1000 EUR than ibook
19:14:35crwldon't know about the more expensive ones, though
19:15:00 Join webguest19 [0] (
19:15:00linuxstbcrwl: I agree. I'm now on my second iBook.
19:15:19fuzziei'm on my third iBook, and i've only paid for one
19:15:24fuzziethanks to Apple's wonderful logic board designs
19:15:28***Saving seen data "./dancer.seen"
19:15:32XavierGrwell I could understand aplle users but no way that I could understand MAC users.
19:15:49XavierGrapple = ipod
19:16:08linuxstbcrwl: Getting off-topic, is there any way I can set my uBook up so I can ALT+TAB between windows in the same application (Firefox, Terminal etc).
19:16:29crwllinuxstb, i don't know, i don't own an ibook
19:16:41crwli got my gf to buy one though :) i've been wondering about the same thing
19:16:45fuzzielinuxstb: i assume you know cmd-tilde does that?
19:16:55crwlcmd-tilde? i'll go try right away ->
19:16:57fuzziethere's a list, somewhere in the OS
19:17:00linuxstbfuzzie: No. No-one told me that :)
19:17:11Paul_The_NerdI keep thinking about getting an iBook
19:17:26Paul_The_NerdThe cheapest one I can find with WiFI.
19:17:30linuxstbThe only thing I know is that ALT+3 gives you a # symbol.
19:17:49linuxstbWhich is missing on at least the UK keyboard.
19:17:52fuzzieheh, the first thing I did to mine was switch to Dutch layout
19:18:19linuxstbThe first thing I did with my first iBook was install Debian. But I've stuck with Mac OS X on the new one so far...
19:18:53Paul_The_NerdI want to install linux on my current laptop, but as far as I can tell nobody's worked out the wireless on it quite yet.
19:19:04XavierGrif you want to upgrade a mac, what do you do... :p
19:19:13 Join Lear [0] (
19:19:15linuxstbYes, if someone knows the perfect Linux laptop, please tell me.
19:19:22crwlif you want to upgrade a pc laptop, what do you do...
19:19:29crwllinuxstb, i hear some ibm ones are pretty good
19:19:42crwlmight be expensive, though.
19:19:42Paul_The_Nerdlinuxstb: If they ever get the wireless working, I'd say emachines m6811 is a good shot for it.
19:19:46preglowdoes the peak meter interpret a max value as a clip?
19:19:47fuzzieXavierGr: you buy a processor upgrade, some RAM, another hard disk, etc? :)
19:20:03linuxstbpreglow: Is that a Rockbox question? :)
19:20:22Paul_The_NerdWhat, rockbox discussion in #rockbox? How DARE you!
19:20:24XavierGrI think yes.
19:22:02XavierGrI tried to make arrangements for the radio peak meter but no avail, I cant understand where the audio by the radio is sent.
19:22:25XavierGrit is not the pcm buffer for sure.
19:22:42_FireFly_mybe directly to the output
19:22:53 Quit linuxstb ("Client Exiting")
19:23:36XavierGrthen how can someone measure the peaks?
19:23:51preglowno, no, it's not in the pcm buffer
19:23:51LearRadio peak meter would require active recording (though not necessarily writing to disk) of the radio...
19:23:52preglowthat's for sure
19:24:02XavierGrGod I am so agnorant on this, I don't even know where to start
19:24:10preglowwhat lear says
19:24:39XavierGractive recording? then this will make the cpu go crazy.
19:25:06LearMight require more than 11 Mhz, but it shouldn't be that bad, I think...
19:25:17XavierGrthere is no chance with 48 or even 11mhz which is currently set.
19:25:28preglownot bad at all
19:25:31preglowof course
19:25:35XavierGroh little more than 11 you say eh.
19:25:36preglow48 would be much more than enough
19:25:53LearThe point is, the audio ADC must be active with proper input, and the data written to somewhere the CPU can read the peaks.
19:25:55preglow11mhz might just be enough
19:26:29LearWe're not talking MP3 encoding or anything here. :)
19:27:48XavierGrhmm thats why there is no s/pdif signal with the radio?
19:28:22LearRadio is analog...
19:28:23 Join muesli- [0] (
19:28:51XavierGropps my bad
19:29:22Bger_cgiircnite all ;)
19:29:22 Quit Bger_cgiirc ("CGI:IRC 0.5.4 (2004/01/29)")
19:30:29XavierGrso what is going to happen with the logs?
19:30:38Paul_The_NerdQuestion: Anyone know of a really good skin/case for an H120 that leaves the line out exposed? (And ideally the remote connection)
19:31:07XavierGrthe outro one
19:31:32XavierGrcheck misticaudio sthe site jeff runs on misticriver
19:31:50XavierGrI am very pleased with it.
19:32:07Paul_The_NerdI looked at the picture of the outro... something about it doesn't sit right with me.
19:32:34Paul_The_NerdI'm *thinking* of just puncturing a hole in my original iRiver case.
19:32:36XavierGrWell no problems for me so far and the screen protector is great.
19:32:51XavierGrThats what I did and screwed my first case.
19:33:10XavierGrI will give you a link for that.
19:34:49XavierGrcheck those 2 pictures, its my case mod.
19:35:07XavierGrthe first and then the seconds attempt.
19:37:37Paul_The_NerdYour case looks quite different from mine
19:38:08Paul_The_NerdMine, you insert the player from the top, and then close a small strap over it, that snaps on the back.
19:39:14XavierGrmine snaps on the front.
19:39:35XavierGryou have the old case.
19:40:06Paul_The_NerdI've had mine for a while.
19:40:24Paul_The_NerdI was thinking of basically doing what you did in your first mod, the line out only one.
19:42:53XavierGrit didnt't work for me so I moved to the second.
19:43:58XavierGrThe hole must be pretty big. In the picture while you think, "ok now I will plug the jack" when you try you will see that the plastic casing of the jack is very fat (especially those that arent made from iriver,
19:45:07Paul_The_NerdI though about that.
19:45:25Paul_The_NerdI'm wondering if I could just cut a horizontal slit
19:45:33Paul_The_NerdThat crosses both line out and line in.
19:45:38Paul_The_NerdDon't actually remove any material
19:45:46Paul_The_NerdSince it's sewn at the ends, the strip shouldn't extend.
19:46:10Paul_The_NerdAnd that way I should be able to plug, and when it's *not* plugged, it should still provide some protection from dust and especially larger particles.
19:46:21XavierGrSo I took a sharp cutter and cutted a rectangle to show both plugs.
19:46:54XavierGropps I think cut is cut in the past too...
19:47:31XavierGrwell I just extended the remote window to fit both line in/out
19:48:48Paul_The_NerdBut it didn't work well?
19:49:24XavierGrwell though the hole was doing its job...
19:49:53XavierGron my first attempt with the little hole, the drill went away and cut the screen.
19:50:15XavierGrSo I had a small tear on my irivers windshield!
19:50:55 Quit muesli- (Read error: 110 (Connection timed out))
19:51:01XavierGrI had that case for quite a long time then I bought the Outro which I think is very good case.
19:51:06 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
19:55:11 Join zeekoe_ [0] (
19:56:13Paul_The_NerdSomewhere I'd found a case that looked quite nice... they'd even imprint it for you, and the price wasn't too bad... but I think it's bookmarked elsewhere. =/
20:11:39XavierGrWhat an asshole I am I just submited a patch without first loging in. Now I can modify it.
20:12:34Paul_The_NerdThat sucks.
20:13:09BagderXavierGr: if you make an update, submit a new patch when logged in and mention what patch it replaces
20:16:39 Quit arkascha (Read error: 110 (Connection timed out))
20:19:09XavierGrOk though I think this version is bug free. Please test it yourself and commnet it. I would be pleased to see this comitted.
20:19:47 Quit webguest19 ("CGI:IRC (EOF)")
20:19:47 Quit webguest95 ("CGI:IRC (EOF)")
20:22:57LearShould I close #1312132 then?
20:27:54 Join linuxstb [0] (
20:28:23XavierGrClose it?
20:28:52XavierGrOnly if you want to commit it? :P (or you want me to re submit it logged in)
20:29:26LearAh, I thought you were about to submit a new, fixed patch. :)
20:29:51XavierGrNo it is the latest, until someone discovers a bug.
20:30:03XavierGrAgain test please.
20:32:02 Join solexx_ [0] (
20:32:26 Join DangerousDan [0] (
20:35:34 Quit t0mas (Read error: 104 (Connection reset by peer))
20:36:13 Quit paugh ("Leaving")
20:47:59 Quit _FireFly_ ("Leaving")
20:49:02 Quit solexx (Read error: 113 (No route to host))
20:49:08 Nick solexx_ is now known as solexx (
20:49:37 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
20:56:53 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
20:57:19 Join muesli- [0] (
20:58:51 Part LinusN
21:02:34 Quit Maxime (Read error: 104 (Connection reset by peer))
21:03:09XavierGrLinus left without saying a word all day!
21:03:21XavierGrLOL High-low
21:03:33 Join Maxime [0] (
21:04:06muesli-seems linus doenst like us anymore..
21:04:36 Quit Rick (Read error: 104 (Connection reset by peer))
21:04:50XavierGrI hope not...
21:06:02 Join Rick [0] (
21:13:33solexxwho cares? let's hope he still likes rockbox! ;)
21:15:02solexxdamn, having to learn b-trees sucks
21:15:29***Saving seen data "./dancer.seen"
21:16:19*solexx promises (again) not to start learning the day before the exam
21:17:56fuzzieyes, indeed, you should be doing it in the early morning of the exam day like a real student
21:34:39 Join webguest31 [0] (
21:34:51webguest31hello guys
21:35:11webguest31vpn-user here (from forums)
21:36:00webguest31any progress with the bugfixed charging algorithm for v1 recorders?
21:36:46solexxnow, who didn't see that coming!?
21:43:03 Quit matsl (Remote closed the connection)
21:54:28 Quit linuxstb (Read error: 104 (Connection reset by peer))
21:54:42 Join linuxstb [0] (
22:01:21 Join ebone_ [0] (
22:01:32ebone_hey fellas.
22:01:55ebone_you know the one thing i will miss from the iRiver firmware ?
22:02:05ebone_and i know there is nothing that can be done about it, sadly.
22:02:19ebone_WOW SRS
22:02:30ebone_that shit really does sound good through a car stereo system.
22:02:41ebone_over headphones, i don't use it at all.
22:02:58ebone_but on a stereo system ... WOW really is, err wow !
22:04:26preglowlike you said, nothing we can do about it
22:04:52fuzzienot a known algorithm?
22:05:59preglowpatented as well
22:06:19preglowwe can do some processing of our own, though, just need someone to invent something first ;)
22:06:42fuzzieit's not as if the mp3 algorithms aren't patented
22:08:20XavierGrwhat if someone knows the wow algorithms and say that it is his work.
22:08:44XavierGrWho is going to know if it used wow or something similar.
22:09:02solexxsomeone reading the source
22:09:29XavierGrand if it isn't open source?
22:09:47solexxand whether it is his own work (from *before* the patent application) has to be proved in court
22:09:54solexxif someone sues him...
22:10:10XavierGrno no you got it wrong.., I mean:
22:10:17solexxif it isn't OS, it cannot be in rockbox...
22:10:36preglowfuzzie: i'm willing to bet the srs people are more aggressive with defending their patents, the mp3 people have more or less said they wont go after decoders anymore
22:10:57preglowanyway, i sure as hell don't know how srs wow works anyway
22:11:04fuzziei was more thinking the encoding
22:11:24XavierGrsomeone gets the WOW algorithms, and makes a program of his own. And the calls that algorithm "owo" (example) who can be sure that he stealed the WOW algorithms?
22:11:37solexxa judge
22:11:40fuzzieXavierGr: that doesn't matter in the context of patents
22:11:57fuzzieif it's patented, even if you invent it independently you're libable.
22:12:00fuzzieliable, even
22:12:08fuzziebut, yes, I have no idea how it works either, just curious :)
22:12:27 Join linuxstb_ [0] (
22:12:43XavierGrbut if noone has checked his code then how SRS can sue him.
22:13:07XavierGrHe can claim that he used hiw own algorithm
22:13:52solexx...and prove it in court, if SRS becomes suspicious
22:14:11XavierGryeah that what I am talking.
22:14:42solexxok, but he will only win if it is really not covered by the patent
22:14:50XavierGrBut I guess that SRS algorithms aren't easy to fond...
22:15:06solexx(and even if he wins, he will have to pay a lot before that decision)
22:15:28XavierGrWhat you mean covered byt the patent?
22:15:47fuzziethe SRS stuff is patented.
22:15:58solexxa patent doesn't cover an exact algorithm, but an abstract idea to handle a problem
22:16:12solexxas far as i understand
22:16:32solexxat least, the "new kind" of patents ("software patents")
22:16:41XavierGrso noone can make a program that alters the sound.
22:16:51solexxthink of the amazon "one click" patent
22:17:03XavierGrBecause there have been many different algorithms for it. EAX as an example
22:17:17XavierGrI bet they do not pay SRS a penny
22:17:25fuzzieright, so if you really really were to use your own algorithm then it'd be fine
22:17:30solexxit all depends on the patent application
22:17:39fuzziebut it'd have to be different enough that it wasn't covered by the patent
22:18:19XavierGrI make the program and hide the code. We are talking about sounds which can be sure if I used the same algorithm?
22:18:44 Join tucoz [0] (
22:18:48XavierGrEAX made her own algorithm so did SRS so can an "x" company or user.
22:19:09XavierGrthe result is not the same,
22:19:15XavierGrbut who can be sure?
22:19:16tucozhi, sorry for bumping into the discussion. Would OpenAL be possible to use?
22:19:58solexxXavierGr: a lawyer. ;-)
22:20:21tucozNot that I would want it, but if someone wants 3dsound, that might be an option.
22:20:40XavierGrEven then the court would have to innocent the guy.
22:20:55XavierGr^claim innocnet
22:21:00fuzzietucoz: OpenAL doesn't do anything which would be useful..
22:21:11tucozfuzzie: ok
22:21:33fuzzieXavierGr: if it weren't open source and weren't easily reverse-engineerable, you might get away with it, yes
22:21:40XavierGrAnd as far as I know software patetns aren't comply with the European laws.
22:21:53fuzziethat doesn't stop EU countries from happily issuing them, though
22:22:39fuzzieand it'd mean no-one from the US could be involved with the code
22:25:16XavierGrWell just an "if" question on a non open source program.
22:25:46solexxdamn, i even cannot find the patent nr.
22:26:12preglowi don't think the srs stuff is too complex
22:26:23preglowbut still enough to not be apparent from just listening at the sound
22:26:54 Quit linuxstb (Read error: 110 (Connection timed out))
22:27:13XavierGrAs for the mp3 encoding. We could program a patented encoder but don't commit it to cvs. Host it on another site and the claim from the Rockbox site that this is not supported by the main developers, though this is unethic.
22:29:11tucozI would suggest using a crappy mp3 encoder for not-that-important recordings, like lectures and stuff. And have a lossless encoder for hq recordings
22:29:29tucozhmm, that is what we have right now, isn't it?
22:29:59tucozor speex
22:30:12 Join amiconn [0] (
22:30:19XavierGrhi amiconn
22:32:39 Nick TiMiD[ParisAgain is now known as TiMiD (n=TiMiD[
22:32:39DBUGEnqueued KICK TiMiD
22:32:57TiMiDI have a strange compilation problem
22:33:10TiMiDtree.c:317: error: structure has no member named `dirstart'
22:33:12TiMiD //int start = tc.dirstart;
22:33:20TiMiDthis is line 317 ...
22:34:41BirdFishDoes anyone have any news on the iaudio X5 port?
22:34:52Bagderthere is no news
22:35:02TiMiDoops sorry to bother you, seems that I wasn't compiling the right code ^^
22:35:16BagderBirdFish: that port is stalling since a few months
22:35:37Bagderhopefully it'll get back up to speed again
22:36:14BirdFishBagder: I figured that is what you meant. But some posts dissappeared on Iaudiophile's forum and I wanted to check so that I could respost and make others aware of the current situation again.
22:36:51CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
22:36:51*amiconn spotted a problem in a forum post that sounds familiar
22:37:19amiconnScreen updates for the H1x0 remote seem to disturb radio reception - seems we need screen freeze
22:37:27 Join arkascha [0] (
22:38:02 Join Philip_0729 [0] (
22:39:24tucozamiconn: hehe, I had completely forgotten about that one. Hmm, maybe that was something else. some low beep when using the remote in general.
22:40:34amiconnI'm still pondering the implementation of a special low-emi lcd driver mode for fm
22:40:47*tucoz is completely rockboxed by now and never uses the remote
22:41:06amiconn(thinking archos here, but it may also helpful for iriver remote)
22:42:27tucozamiconn: I do not remember if I have asked you this before, but.. Do some shades of gray flicker on your iriver?
22:42:49amiconnDo you mean the native shades, or the grayscale lib?
22:42:56tucozgrayscale lib
22:43:08amiconnThat's normal, due to how the grayscale lib works
22:43:56tucozIt doesn't matter. Mandelbrot is cool no matter what. Especially on a dap that do not support that amount of shades anyway :)
22:44:22amiconnYes, mandelbrot uses 9 shades (both on H1x0 and archos)
22:44:50Learamiconn: just curios regarding the timer changes... fading uses "shorter" timer periods then?
22:45:27tucozhehe, I really like that. Off topic: have I dreamt this, or did some c64'ers use the cpu in the floppy drive for extra computation power?
22:45:29amiconnMy timer changes avoid the timer restart at frequency changes, but of course the period isn't 100% stable
22:46:01amiconn...when the frequency changes, and an interrupt is delayed during pll relock
22:46:15amiconn...which may take up to 10 ms, typically 3 ms
22:46:50amiconnEverything that uses a timer period of 10ms or greater and can stand an occasional glitch will work without locking the frequency by boosting,
22:47:21amiconnbut backlight fading uses timer periods in the microseconds range
22:48:02amiconn(50 s...5 ms)
22:49:10tucozseems like I didn't dream "Old Commodore 64 floppy drives had a CPU and some RAM in them that you could actually upload and execute code with"
22:49:42Bagdertucoz: right, that's how they work
22:50:08Bagderiirc, there was a 6502 in the floppy drive too
22:50:14 Join Strath [0] (
22:50:52tucozBagder: same as for the 64?
22:51:50tucozhehe, those were the days. It is really amazing to think of what people could do with their 64s.
22:52:21Bagderour stuff ;-)
22:52:23tucozhave read that :)
22:52:38tucozI have to power up my old one soon
22:52:47LearHrm... Loaded a config file that changed bass and treble only. Playback stopped... :/
22:53:51XavierGramiconn: remote updates interefer with the remote? But thats not the issue with the iriver firmware.
22:54:17*XavierGr gets to test the FM radio with the remote and default firmware.
22:54:28tucozjust for the record. I came to think of that when I thought of amiconns work on the grayscale lib. Doing stuff that isn't "possible"
22:54:46*amiconn would need to find out how to enter fm with the stock firmware first :/
22:55:06BagderI also feel that lost when I used it
22:55:14BagderI once managed to activate it by mistkae
22:55:23Bagderand couldn't figure out how to get out of it ;-)
22:55:39XavierGrWell I think that it is not that difficult to operate. I am used to it.
22:55:40amiconnHaha, I also managed to enter it once, but didn't find out how to tune
22:55:48tucozamiconn: hold play
22:55:54tucoz..for a while
22:56:11tucozand hold play to exit
22:57:19amiconntucoz: Nah, it's not really impossible. The crucial point was the idea how to display shades of grey on a b&w display, and that idea wasn't even mine
22:57:57amiconnThe rest was testing, thinking, retesting, optimising, and lots of implementation work
22:58:09amiconn(involving assembler for speed)
22:58:21tucozamiconn: I still think that it is cool
22:59:46XavierGrHmm actually they are right, there is somekind of mistuning when I change the volume. (and the lcd updates)
23:00:12tucozI mean it is possible, obviously. But, the idea might seem impossible.
23:00:38amiconnI'd rather expect that for the remote, considered how close the data lines are to the headphone lines which are used as the antenna
23:01:28 Join ep0ch [0] (n=ep0ch@
23:01:29tucozif one is thinking of hw limitations i.e.
23:01:32amiconntucoz: I agree that it is a cool idea...
23:01:37 Quit Lear ("Chatzilla [Firefox 1.4.1/undefined]")
23:01:45XavierGrBut why we need the freeze screen this appears only when the lcd changes.
23:01:50amiconnThe idea is borrowed from the video plugin, which is [IDC]Dragon's work
23:02:13amiconnXavierGr: Rockbox does update the lcd periodically
23:02:18tucozoh, ok.
23:02:32tucozthat does also sound like fun.
23:02:52tucozgot to go. Bye
23:02:55XavierGrLucky those that have the batch without the ticking issue.
23:02:55 Part tucoz
23:02:58ep0chXavierGr: i;ve just applied your radio patch
23:03:07ep0chit didn't break anything :)
23:03:10amiconnOf course I had to 'invent' some techniques, like the pseudo-random pattern shifting that minimises flicker
23:03:10XavierGrand :)
23:03:42webguest31how is bugfixing going for the charging algorythm of the recorder v1?
23:03:43ep0chit doesnt pick up the existing default fm preset file if you go via the settings menu?
23:03:44amiconnWithout it, the grayscale lib would flicker like hell, esp. on archos
23:04:48XavierGrep0ch: What exactly do you mean?
23:05:07XavierGrIt will remember presets in .rockbox/presets/ directory.
23:05:28webguest31amiconn: You were about to send me a test-binary for the charging algo, nor?
23:05:49*amiconn completely forgot that
23:05:52ep0chthe old radio had a file called .rockbox/fm-presets-default.fmr
23:05:56webguest31I see :)
23:06:13ep0chi need to move this to .rockbox/presets then...
23:06:31XavierGrthis will load the presets and pop the radio, but it will not remember it after a shutdown.
23:06:54XavierGrBut if you had presets that you want you save it to default directory presets.
23:06:54amiconnwebguest31: Perhaps I should do some tests myself. I don't know whether the logging code in current powermgmt.c will tell us much
23:07:14webguest31amiconn: You have a recorder v1?
23:07:24XavierGrJust push A-B button and then save to the prompted directory with a filename under 20 characters
23:07:44amiconnwebguest31: Yes, that's one of my rockboxes
23:07:55ep0chi've loaded the default fmr file up and saved it in the radio. that works.
23:08:03webguest31amiconn: And your box works correctly?
23:08:12linuxstb_Anyone familiar with ARM assembler, linker scripts, map files, crt0.S etc ?
23:08:48amiconnwebguest31: Not really... I don't have as big problems as you, but I do get much less runtime from my 2500mAh cells than I would expect
23:09:06amiconnI get 6..8 hours, I'd rather expect twice as much
23:09:08Bagderlinuxstb_: somewhat, yes, yes, somewhat, but I'm heading to bed in minutes
23:09:14webguest31amiconn: Ok so what will we do?
23:09:37ep0chXavierGr: but i think the radio should look at .rockbox/fm-presets-default it .rockbox/presets does not exist or contain any .fmr files
23:09:55webguest31amiconn: I only have very little C knowledge and more little jukebox knowledge :)
23:10:12Bagderlinuxstb_: fire off a mail to the list
23:10:29 Join tvelocity [0] (
23:10:43XavierGrep0ch once this is commited and people know about it it will not have any use to search in .rockbox directory.
23:11:08linuxstb_The problem is that my ipod bootloader is not running reliably. I think the problem is that the code gets loaded to one address, and then my relocation of the code (and data) to a different address isn't working properly.
23:11:31ep0chXavierGr: thats true
23:11:33linuxstb_But I'll keep thinking about it tonight, and try and diagnose what's going on.
23:11:38amiconnwebguest31: These problems definitely need to be fixed, however, there's always little time
23:11:41XavierGrbootlaoder? You amde a rockbox bootlaoder that quick? Are you joking?
23:11:46Bagderbtw, there's a .ss-generating site coming up soonish
23:12:20webguest31miconn: I understand. no stress, I can still rolo the archos firmware
23:12:50linuxstb_XavierGr: It's in progress, it's not working. I've got the framework working, but there's problems. I then need to write the ATA driver before the bootloader can actually load anything...
23:12:52ep0chXavierGr: one last thing, while you're working on the radio, can you make the buttons consistent with the rest of rockbox, as there is not way to get to the rockbox settings menu in radio mode :(
23:13:04Bagderlinuxstb_: sure, just mail away when you think we/I can offer a hand in the debug job
23:13:30Bagdernow: bedtime
23:13:33amiconnwebguest31: As soon as I have some testing code, I'll post about it on the ml. I think I'll need logs taken with various boxes, various cells etc to tune the algorithm
23:13:38ep0chXavierGr: but me likes preset mode
23:14:13amiconnMy idea is to measure voltage with and without charging current. Perhaps that will give us some more hints about the real status of the cells
23:14:22webguest31miconn: Ok - I see I have to subscribe to the list :)
23:14:33linuxstb_Bagder. Thanks.
23:14:36XavierGrep0ch: Didn't cought that. How can you get to rockbox settings when you are inside the radio? When you set the radio on (either by selecting an fmr file or choose FM radio) A-B button will open the FM menu.
23:14:52webguest31Damn - my "a" needs more power to get
23:15:27amiconnwebguest31: You save me a 'pling' and a red line everytime ;)
23:15:31***Saving seen data "./dancer.seen"
23:15:53TiMiDwho thinks a complete rewrite from swratch of the file browser would be a good thing ?
23:16:22 Quit Philip_0729 ()
23:16:44ep0chXavierGr: oh just realised that pressing the joystick now takes you to rockbox settings. But this isn't consistent. Can you swap the A-B button with joystick press?
23:16:54webguest31BTW: I got 17 hours of playtime with my brand new 2600mAh batteries!
23:17:07webguest31I think thats massive
23:17:21amiconnXavierGr: Allowing to call the main menu from the radio screen isn't a good idea imho
23:17:33XavierGrep0ch:I never touched the button hnadling this is the attitude by default.
23:17:38ep0chamiconn: why not?
23:17:45XavierGryeah why?
23:17:49amiconnIt allows cycles like menu->radio->menu->radio...
23:17:56XavierGrno it dont
23:18:03XavierGrI coded that way that will not.
23:18:24XavierGrI check the fm status before everything.
23:18:25amiconnIt will eat up all menu slots, or overflow the stack, whatever comes first
23:18:38ep0chwell something has changed in the radio, before you couldn't go to the settings menu at all. now you can with joystick button
23:18:54amiconnXavierGr: So, how do I enter the radio screen when the radio is already playing
23:19:03ep0chif i'm not mistaken
23:19:03amiconn(after leaving it with the radio on)
23:19:09XavierGrchoosing radio again by the menu
23:19:29XavierGrthen it will check if the radio is already on and it will only draw the screen.
23:19:52XavierGrthat's the same as before though I just check it with an if.
23:19:55amiconnYes, but that doesn't prevent the cycle I was talking about
23:20:52XavierGrI cant really got what you are saying. Can you be more specific and which is the change before after my patch?
23:21:31amiconnBefore the patch you couldn't enter the menu from the radio screen
23:21:45amiconn...for a reason, because the radio screen is entered from the menu
23:22:07XavierGrso before my patch you couldnt load a new setting?
23:22:19XavierGror change the settings for the display?
23:22:28amiconnYes, by leaving the radio screen
23:22:34ep0chyou would have to have to filetree mode
23:22:45ep0chs/to/go to
23:22:46amiconnHmm, perhaps I didn't understand what you mean
23:23:12XavierGrplease try the patch and see for youself that there is no point of confusion.
23:23:22XavierGrI think that it is well implemented that way.
23:23:27 Quit muesli- (Read error: 110 (Connection timed out))
23:23:32amiconn..but then I was irritated by ep0ch saying that you couldn't use the menu before.
23:23:58ep0chwell you couldn't
23:24:04XavierGrI think that I could use it even without my patch.
23:24:06amiconnIn fact that has always been possible
23:24:11XavierGryes leaving the radio screen with Play instead of Stop (on iriver)
23:24:32XavierGrwhy not go to the settings. How could one get to the FM again?
23:24:55XavierGrthat's right. So there is no difference right?
23:24:57amiconnXavierGr: No, that was a misunderstanding.
23:25:24XavierGrThank god! I had been doing a lot of work into this.
23:25:31amiconnI thought you mean entering the menu by actually calling it (again), but you mean by leaving the radio screen
23:25:52XavierGrep0ch: Keep testing. and tell me if you like the cahnges. it has always been possible. Just the buttons have changed
23:26:40amiconnI think this function could be put on AB (consistent with other screens) and the fm menu put on long Select (as the 'context menu' of the radio screen)
23:26:56XavierGramiconn: I didn't changed any buttons. The button handling is the smae.
23:26:57ep0chwell i'm convinced before you could not go directly from radio screen to rockbox settings before. but now you can go from radio screen to rockbox settings by pressing the joystick button
23:27:05ep0chand i just wanted this swapping with A-B :)
23:27:46amiconnep0ch: You could do so before, definitely
23:27:57XavierGras amiconn said.
23:28:05 Quit ender` (Connection timed out)
23:28:14XavierGrBecause I will say this again. I didn't change the button handling at all.
23:28:25amiconnAh, it even was select
23:28:39XavierGrep0ch; we can change the button handling if the devs want to.
23:28:40amiconnSeems I'm a little confused...
23:29:05ep0chme too now :)
23:29:27XavierGrI will test a build without my patch why dont you test a build with my patch?
23:29:48amiconnPlay was the fm menu, and A-B was the context menu, allowing to add presets etc
23:29:48ep0chRockbox just switched off by itself when in radio...
23:30:35XavierGrdo you have autoturn down on?
23:31:09XavierGramiconn: play (in the side) will render the preset list.
23:31:15ep0chlet me wait a few minutes and see...
23:31:22XavierGrA-B will get the fm radio menu.
23:31:37XavierGrand select will exit the radio but keep playing.
23:31:46XavierGrthat was the button handling before and after my patch.
23:31:49ep0chXavierGr: there was a thread somewhere about making the buttons consistant in fm mode
23:32:51XavierGrYes just tested it is how I say.
23:33:11 Quit webguest31 ("CGI:IRC")
23:33:23XavierGrEven in the old builds Select (joystick click) will bring up the main menu of rockbox.
23:33:28linuxstb_A joystick press takes you from the FM screen back to the main menu, leaving the radio on. This is consistent with a joystick press from the WPS, which takes you back to the browser.
23:34:21amiconnlinuxstb_: It's consistent and inconsistent at the same time, depending on the view
23:34:56linuxstb_If you understand that Rockbox goes browser -> menu -> FM and browser -> WPS, then it makes sense.
23:35:05amiconnSeeing the presets list as the fm mode's browser, it's inconsistent
23:35:22ep0chA/B = go to main menu (currently A/B goes to radio context menu)
23:35:23ep0chSTOP = quit radio (this is the only one that's correct)
23:35:23ep0chPush in joystick = select from presets (this currently goes to the main menu)
23:35:23DBUGEnqueued KICK ep0ch
23:35:23ep0chPush+hold joystick = context menu (currently does nothing)
23:35:23ep0chPlay = toggle mute/unmute (currently selects from radio presets)
23:36:21 Join lImbus [0] (
23:36:24amiconnHmm, I just had an idea...
23:36:29lImbushi all
23:37:02amiconnWhat's currently missing on iriver and Ondio is a way to delete presets, due to how the initial radio implementation works (on archos fm recorder)
23:37:21amiconnIf the presets list would have a context menu...
23:37:23XavierGrI add tha to my patch.
23:37:42amiconn...but the whole radio menu stuff needs some rewriting
23:38:10XavierGrwhat you mean exactly?
23:38:24amiconnrasher tried to implement the button assignment from that post. It didn't work, compilation failed due to duplicate case: values...
23:38:32XavierGrdelete filepresets or preset entries ( I suppose the 2nd)
23:39:11XavierGrI made an option clear preset list.
23:39:21 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
23:39:59XavierGrand currently the preset list has already a context menu
23:40:06XavierGrEdit preset and delete preset.
23:41:14ep0chhmm I have no idea why rockbox switch off by itself....
23:41:29ep0chthe battery is low, but not too low
23:41:51 Join webguest95 [0] (
23:42:03XavierGryou were on the radio screen or outside of it?
23:42:08ep0chin radio
23:42:25XavierGrI had this one too but I was outside.
23:42:29ep0chi;m in radio now, haven;t touched anything, and it's fine for the last 10 minutes
23:42:40ep0chwas your battery lowish?
23:42:50XavierGrI think yes.
23:42:56XavierGrBut I am not sure.
23:43:07ep0chmaybe a dodgy battery reading shutting rockbox down?
23:43:27ep0chi assume rockbox shuts down on low battery?
23:43:33XavierGrFiddle with it a little. Load save, add exit (without stopping) reenter and then let it for another 10 minutes.
23:43:59 Quit arkascha (Remote closed the connection)
23:45:47ep0chXavierGr: if there's room maybe it would look better (to me) if there was a space between preset number and station name.
23:47:53XavierGrOk I can add that. what others think?
23:48:23ep0chFM radio is not running the CPU @11 Mhz ?
23:48:38XavierGryou mean like "1. Station name" instead of "1.Station name" ?
23:48:43XavierGrit is.
23:49:05ep0chXavierGr: yes, but i'm not THAT bothered about the space really :)
23:49:48ep0chwell, i've gone to the debug screen while the radio is on the the cpu frequency is at 45 mhz
23:49:58XavierGrDid you played with the preset mode? Did you load the FM from a preset file?
23:50:21XavierGryes becuase you need normal cpu power outside of the radio.
23:50:39linuxstb_I'm listening to the radio ATM (unpatched rockbox from a few days ago), and it turned itself off a few minutes ago. The battery is fully charged.
23:50:44XavierGrCan you test the low frequency mistuning?
23:50:50ep0chyou do? oh... i always found menus fine at 11mhz
23:51:03ep0chlow frequency? how low?#
23:51:16XavierGrwell do this:
23:51:30XavierGrEnter the menu and tune the radio between 2 stations.
23:51:47XavierGrThen press click and then renter the radio screen.
23:52:05XavierGrListen carefully to the signal do you hear any changes?
23:52:16XavierGrReapeat that and tell me if you can hear that.
23:52:42XavierGralso this can be done from the debug menu.
23:53:21ep0chi've heard the same when listening to other audio files
23:53:27ep0chlike a low pass filter effect
23:53:44ep0chthats probably at around 30 khz
23:53:59Maxime30khz? you wouldn't hear it..
23:54:09ep0choh i mean 15 then :p
23:54:27Maximehuman ears are from 20Hz to 20Khz :x
23:54:42Maxime(for a very good ear)
23:54:46ep0chyeah i know, i'm used to sampling rates
23:55:25Maximei've found a website with multiple samples like 15Khz, 16khz..
23:55:34Maximeto compare this may be useful no?
23:56:04ep0chXavierGr: would you say it sounds like a low pass filter effect what you hear?
23:56:19ep0chhigh frequencies just dont seem to be there
23:56:38XavierGrsomething like that.
23:56:53XavierGrand on a well tuned station I can hear a faint pop.
23:56:53ep0chtry it with an mp3 or vorbis
23:57:07ep0choh i think we;re hearing different things
23:57:18XavierGrtry what?
23:57:26XavierGrchange the frequency while playing?
23:57:30ep0chplay on mp3 of vorbis then change the frequency to 11
23:57:35XavierGrbut this may crash the player or the playback.
23:57:56ep0chwell don't do it for too long
23:58:05 Join ashridah [0] (
23:58:06ep0chor the buffer will empty
23:58:37ep0chbut i can definitly hear a change in the frequencies
23:58:44XavierGramazing you are right I can clearly hear a pop.
23:58:58ep0chwow i dont get a pop! :)
23:58:59XavierGrI hear a pop when doing this.
23:59:04ep0chjust a dulled sound
23:59:31XavierGrwell I didn't checked for long to hear the dull sound but I can clearly hear the pop.
23:59:35ep0chwhat player have you got?
23:59:54XavierGrand 1900mah battery.

Previous day | Next day