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 2005-11-14

00:00:11amiconnpreglow: Rockbox doesn't handle thread functions that simply exit. The proper way to terminate a thread is to call remove_thread()
00:00:41preglowamiconn: does that work from within a thread as well?
00:00:41linuxstb_preglow: If it doesn't break the bootloader, then commit.
00:01:00amiconnYes, threads should be able to terminate themselves
00:02:02 Join XavierGr [0] (
00:02:12amiconnJust make sure that at least one thread stays... I don't know how the scheduler would react when num_threads ever gets zero
00:07:33]RowaN[i wish foobar had a "playing next" feature like rockbox =]
00:10:31preglowarm asm is sweet
00:11:28 Part len0x
00:16:53linuxstb_The IPL people have been busy:
00:16:54preglowlinuxstb_: only functions i can see scrolling use, is lcd_fillrect and lcd_putsxyofs, both of which are implemented
00:17:09 Quit gromit` (Read error: 104 (Connection reset by peer))
00:17:27linuxstb_But are they implemented correctly? They may still contain the h100 code :(
00:18:02preglowpreglow's variable plural english experiment
00:18:10preglowi have no idea, heh
00:18:16linuxstb_fillrect is wrong.
00:18:31preglowfillrect only deletes old data, i believe
00:18:31 Quit MacDancer ("Leaving")
00:18:42preglowi should be able to see some moving garbage even if that didn't work
00:19:08linuxstb_I would have thought so.
00:19:59 Join gromit` [0] (n=gromit`
00:20:31linuxstb_lcd_putsxyofs should be working though - it's used by the lcd_puts function.
00:22:51preglowthe thread IS running
00:22:52preglowi just verified that
00:23:26preglowi don't think we're running at too high a clock rate yet
00:23:31preglowany idea which?
00:23:54 Quit gromit` (Read error: 104 (Connection reset by peer))
00:24:15 Join edx [0] (
00:24:28linuxstb_I think I read somewhere that it's about 16MHz. But I'm not sure.
00:25:27 Join gromit` [0] (n=gromit`
00:25:35amiconnDoes ipodlinux adjust the cpu clock on the fly like rockbox on iriver?
00:25:58linuxstb_I'm not sure.
00:26:26linuxstb_I would be surprised if it's as tightly controlled as Rockbox though.
00:26:43linuxstb_We'll find out when we get to that stage.
00:26:52amiconnThat will be a fun part I think, making the timer work, and work correctly with varying cpu clock
00:27:02amiconnAll thins without hw docs...
00:27:15preglowamiconn: rtc...
00:27:29preglowahh, those timers :P
00:27:42preglowyes, that'll be fun
00:27:53preglowsomeone should start pressuring portalplayer!
00:27:55 Quit ender` (Read error: 104 (Connection reset by peer))
00:28:01preglowso we might have docs some time in twenty years
00:34:42 Quit Moos ("Glory to Rockbox")
00:36:39Bagdercan a MIPS write 16 bits to an odd address?
00:38:32preglownever used mips
00:39:25*Bagder debugs a SIGSEGV on IRIX 6.5
00:40:05BagderI'm quite sure this is the problem, it cannot access unaligned addresses like this
00:40:05 Quit chopped_pork (Read error: 104 (Connection reset by peer))
00:44:13preglowscrolling works
00:44:16preglowi'll commit threading
00:44:59linuxstb_What was the problem?
00:45:04preglowme being a complete fool
00:45:20preglowthere's some corruption of the screen
00:45:26preglowbut that's not a threading problem ;)
00:45:39linuxstb_That will obviously be the lcd_fillrect.
00:45:50preglowlinuxstb_: suffice it to say i just now learned of the lcd_puts_scroll funtion
00:46:10preglowlooking for bugs in all the wrong places, as usual
00:46:25preglowshall i commit the stack fix as well?
00:46:33preglowthat is, initing sp to stackend
00:46:44linuxstb_I was just about to ask you about that.
00:47:25amiconnlcd_fillrect should be incredibly simple on 8bit+ displays...
00:47:39linuxstb_Feel free to implement it :)
00:47:56amiconnIt' basically only memset()ting, except for the DRMODE_COMPLEMENT
00:48:25amiconnNah, for >8 bit it's not exactly memset()ting, but very similar
00:48:46linuxstb_A 16-bit memset...
00:49:12amiconnlinuxstb_: Perhaps I'll work a little on the 16 bit driver, but that requires getting the sim to work first
00:49:23preglowor you getting an ipod!
00:49:43amiconnThat's rather unlikely
00:49:45preglowamiconn: think of all the sexy arm assembler you're missing out on
00:49:46linuxstb_Or a H300...
00:50:12amiconnH300 is more likely than any ipod, although I dislike colour lcds on portable devices in general
00:50:24linuxstb_Get a b/w ipod....
00:50:29preglowi was a bit surprised at the quality of the lcd, actually
00:51:05preglowof course, it's still a bit unusable without backlight
00:51:17amiconnIt's not about quality (recent colour lcds are rather good; I have a Samsung mobile phone with colour)
00:51:19preglowbut i also think bw displays don't really shine without a backlight either
00:51:31preglowand that's not meant literally, heh
00:51:38amiconnIt's just that it requires backlight to be readable, and that eats battery
00:51:51amiconn...and I don't really need colour
00:51:58preglow'course you don't need it
00:52:04preglowbut it allows for a ton of bells and whistles
00:52:08preglowwhich are fun, occasionally
00:53:43*tucoz whispers mandelbrot.rock in color to amiconn
00:54:44preglownow there's a nice example if i ever heard one
00:55:15preglowand on arm you've got a shiny, full 64 bit accumulator!
00:55:49amiconn64bit math, multiple shades, all readily available
00:56:18preglowi'm really looking forward to working on codecs for this little thing
00:56:25amiconnThe grayscale lib was much more of a technical challenge...
00:56:30preglowwhich reminds me, i've got some musepack changes pending
00:56:45amiconnAnd aac iirc?
00:56:58preglowi never was able to do much for aac
00:57:00tucoznothing of a challenge. That is the spirit. With your recent memcpy I see what you mean.
00:57:08preglowi've got some stuff pending, but it doesn't seem to help at all...
00:57:27preglowhaha, yes
00:57:36amiconntucoz: Yes, and soon I'll extend it into memmove(). The archos implementation is already done
00:57:41preglowi can see what you meant by calling it a monster
00:58:13*preglow laments that we'll never have optimised arm mem* functions :/
00:58:20***Saving seen data "./dancer.seen"
00:58:54preglownow there's a challende for you
00:58:57preglowchallenge, even
00:59:04amiconnDo I hear "grayscale lib for b&w ipods!"?
00:59:09tucozamiconn, cool. I really learn something new everyday by just following the discussions, and looking into the commits on Rockbox.
00:59:09amiconnShow off ipodlinux...
00:59:11preglowwriting nice mem* functions in thumb arm
00:59:16preglowamiconn: hehe
01:00:10tucozhehe, that would be something.
01:00:22amiconnpreglow: Regarding memcpy() - I was even able to shave off some more bytes
01:00:25 Part tucoz ("Leaving")
01:00:30amiconnNow it's exactly 950 bytes
01:00:54preglowis that too in your commit?
01:02:43 Quit gromit` (Read error: 110 (Connection timed out))
01:02:57 Join DJDD [0] (
01:08:11preglownow, what next
01:12:12linuxstb_crt0.S ?
01:13:22preglowsounds good
01:13:30linuxstb_Rockbox could possibly run after that.
01:13:34preglowjust need to get an overview over what it should do
01:14:24amiconntimer(s)? button driver?
01:14:36linuxstb_I said "run", not "work" :)
01:15:21linuxstb_What are the timers used for?
01:15:28preglowbutton driver shouldn't be too much work either, but no, i think i'll do crt0.s
01:16:39amiconnlinuxstb_: One is needed for the timer tick. One more can be used freely; on iriver it is shared between backlight fading and plugins, e.g all grayscale plugins
01:17:09amiconnUnfortunately the coldfire has only 2 timers, SH1 has 5 ...
01:17:58preglowand we'll probably never find out what ipod has :)
01:18:39preglowas many timers as 5 is rather rare, though
01:20:55linuxstb_The ipod has at least 2 - they are defined in the ipl source.
01:21:16preglowso, currently there is only bootloader code in crt0.S
01:21:29linuxstb_preglow: Yes.
01:21:45linuxstb_Most of it is inside the #ifdef BOOTLOADER define
01:22:04preglowsome of it can probably be reused
01:22:59linuxstb_Looking at the IPL source, I think it just has two timers.
01:23:58preglowtwo timers are by far the most common number i've encountered
01:24:15linuxstb_But Rockbox should run without them - we emulate the current_tick.
01:24:41XavierGranyone knows hot to block all internet activity from windows XP firewall?
01:24:48XavierGrDon't laugh at me please.
01:25:00linuxstb_Unplug the ethernet cable?
01:25:30XavierGrI want this done through XP firewall to make someone believe that there is no internet...
01:25:46*amiconn is looking at this alien thing named arm assembler
01:25:57XavierGrWell obviously I want some specific apps to work on the internet without the user knowing it.
01:25:58preglowamiconn: it's pretty different, but i think it's cool
01:26:35amiconnFirst, the source->destination logic seems to be switched compared to our other targets
01:26:51preglowamiconn: yup, but it's pretty logical
01:27:06preglowamiconn: like for add, which is "add r0, r1, r2", which means r0 = r1 + r2
01:27:33preglowamiconn: and sub, which is "sub r0, r1, r2", and means r0 = r1 - r2, so it follows algebraic order
01:28:20amiconnI was about to ask about the 3 parameters
01:28:24preglowhaving a separate destination register is really nice
01:28:49amiconnSH seems to be somewah in-between coldfire and arm then
01:29:35amiconnWhile it doesn't know insns with 3 parameters (at least not for SH1), insns taking only one parameter on coldfire often take 2 on SH1
01:29:59amiconne.g. neg %d0 −−> neg r1,r2
01:30:04preglowyes, the one param instrs are pretty annoying
01:30:13preglowdoes sh use fixed length instructions?
01:30:19preglow32 bits?
01:30:28preglowarm uses 32
01:30:34preglowbut there's also a 16 bit length mode
01:30:39preglowcalled thumb arm
01:30:50preglowso it's actually got two instruction sets
01:30:53amiconnYes, read about it. I wonder why...
01:31:07preglow32 bit mode is tons and tons more flexible
01:31:13preglow16 bit mode is faster for 16 bit buses
01:31:33amiconnyes, but supporting two complete instruction sets in a cpu core sounds like overkill to me
01:31:44preglowthey've implemented it very efficiently
01:31:52amiconncan you mix modes?
01:31:56preglowthumb instructions are "uncompressed" to ordinary 32 bit instrctions internally
01:32:01preglowamiconn: yes, via a branch instruction
01:32:12preglowamiconn: if the lsb of the branch target is 1, then thumb mode is set
01:32:33preglowarm forces alignment of branch targets, i believe
01:32:36amiconnIf I read some gcc sources correctly, the higher SH cpus also know 3-parameter instructions
01:33:07amiconnSH1 is just the 'smallest' member...
01:33:15preglowamiconn: arm actually has some four parameter instructions as well
01:33:33preglowamiconn: where to of the parameters describe a 64 bit combined register
01:34:19amiconnhow many registers?
01:35:00preglowr15 is ip, r14 is link register, r13 is (usually) sp
01:35:00amiconnsounds familiar...
01:35:19preglowall of them can be manipulated as you want
01:35:20amiconnAh, so ip (aka pc) is part of the general set?
01:35:28preglowso a jmp on arm is actually a mov pc, rxx
01:36:10amiconnOn SH, r15 is sp, pc is separate
01:36:40preglowso there are fewer regs on arm, yes
01:36:58linuxstb_amiconn: Do you think FLAC is feasible on the SH1?
01:37:14preglowi don't
01:37:36linuxstb_My impression was the SH1 at 20MHz is faster than the coldfire at 20MHz.
01:37:51amiconnThe RAM is definitely faster
01:37:58linuxstb_That's what I was thinking.
01:38:09linuxstb_But of course there is less IRAM.
01:38:15preglowhow fast are muls on sh?
01:38:20amiconn...but the SH1 is running at 11.0592 MHz (recorders) resp. 12 MHz (player, Ondio)
01:38:34linuxstb_Did I just dream the 20MHz part?
01:38:47preglowlet's hope not
01:39:00amiconnpreglow: muls are 1 cycle provided you don't read the result too early
01:39:07preglowright, like emac
01:39:18amiconnThe multiplier works in the background and needs 3 cycles
01:39:39amiconnThe destination of a mul is *always* the macl register
01:39:40preglowso shifting by n is actually slower than a multiply? ;)
01:39:52amiconnFor variable n yes
01:40:00preglowi still think that sounds weird
01:40:03preglowi love barrel shifters
01:40:06amiconnThe sh can shift, just not by variable n
01:40:16 Part LinusN
01:40:20amiconnFixed shifts are 1, 2, 8 and 16
01:40:41amiconn...and there's even an advantage compared to cf - it can rotate
01:41:54amiconnWhen comparing shift and multiplication, note that the SH1 multiplier only does 16*16->32
01:42:02preglowahh, right
01:42:07amiconnSH2+ is full 32 bit
01:42:09preglowmight as well just forget flac, then
01:42:23amiconnThe mac accumulator is 40 bits on sh1
01:42:45preglowshifting on arm is nice
01:42:59preglowmost arithmetic instructions can have a variable size shift or rotate
01:43:10preglowadd r0, r1, r2 lsl 5, for example
01:43:49amiconnI thing the only sofware compression feasible on SH1 are the simple wav compression schemes
01:44:29preglowyou might be able to cope with adpcm
01:44:41amiconnThe SH1 will be quite busy just with feeding the data
01:45:04amiconnThanks to the silly spi we have to bitswap...
01:49:04amiconnstmia and ldmia look like the equivalent of movem
01:50:09preglowthey are
01:50:12preglowjust more flexible
01:50:35preglowstm is the base instruction, with postfixes ia, ib, da and db
01:50:51preglowstands for increment before, increment after, decrement after and decrement before
01:51:00preglowwhich decides how the address gen works
01:51:14preglowyou can also update the address register after the instruction by adding an !
01:51:41preglowwhich is quite handy for function prologues
01:52:35 Quit XavierGr ("Trillian (")
01:52:41preglowthis possibility used to be in 68k as well, but not coldfire
01:58:48 Quit _DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
02:01:53 Join Cassandra [0] (
02:02:35CassandraI've updated my WPS in preparation for commision to CVS. The thing is, I need a more interesting name for it than 3d.
02:03:44CassandraThe last one (the files themselves aren't uploaded yet).
02:03:57CassandraSuggestions welcome.
02:07:57Paul_The_NerdFrom the looks of it, doesn't it show basically all the information you *can* show, and demonstrate all the functionality available. You seem to have used almost everything you could...
02:08:49Paul_The_NerdI'm terrible at naming things, but it seems something involving either 'complete' or 'example' would be logical.
02:08:59CassandraYep. I think it has everything currently in CVS (+ room to add replaygain and whatever when it gets commite)
02:09:28Cassandra(although I'll have to make more images available first. I'm actually using all 52 now.
02:11:44CassandraI like the idea, I'll have a think along those lines.
02:12:42Paul_The_NerdI mean, it's a good place to get a practical look at how to use all the WPS features so far.
02:13:06 Quit actionshrimp ("a bird in the bush is worth two in your house")
02:13:14CassandraSure there must be something I've not used.
02:13:27CassandraIt is lovely and runic though.
02:14:20Cassandra(although back under 2k now that images are stored in a subfolder by default)
02:15:31Paul_The_NerdThere may be, but it's got most of it.
02:17:05 Quit ]RowaN[ (Read error: 110 (Connection timed out))
02:17:06CassandraOh yeah, I don't display the album year.
02:17:16CassandraNever seemed like that useful a piece of info to me.
02:17:39Cassandra(although I suppose I could stick it in brackets after the album name.)
02:21:31XShocKCassandra, it looks nice.
02:25:06 Quit yngwi ("Chatzilla 0.9.68a [Firefox 1.0.7/20050919]")
02:25:23CassandraThanks. I'm quite pleased with it myself.
02:28:27XShocKbut i can't make it work on my player..
02:28:46XShocKfirst of i put files into .rockbox/wps and not to .rockbox/
02:29:03XShocKonly then i could "browse .wps"
02:29:33XShocKand when i play, no "3d look is seen" only name of the file, and next files name...
02:33:05CassandraSorry, I did say the new version isn't up yet, just the images.
02:33:19CassandraLet me fix that.
02:35:38XShocKaaah. ok
02:41:53CassandraOK - it's up there now.
02:42:00CassandraExtract to /.rockbox/wps/
02:42:15CassandraYou need a bleeding edge build, possibly.
02:42:55CassandraNo, the last daily should be sufficient.
02:54:58XShocKso. what whall i see?
02:58:21***Saving seen data "./dancer.seen"
03:00:00 Quit edx (Read error: 110 (Connection timed out))
03:14:50 Join solexx_ [0] (
03:20:03 Nick linuxstb_ is now known as linuxstb (
03:24:42 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
03:27:39 Quit solexx (Read error: 110 (Connection timed out))
04:58:25***Saving seen data "./dancer.seen"
05:47:57 Join DJDD_ [0] (
06:06:30 Quit DJDD (Read error: 110 (Connection timed out))
06:13:32 Quit DJDD_ (Read error: 110 (Connection timed out))
06:27:44 Quit RotAtoR ()
06:40:16 Join BirdFish [0] (n=bradbox8@
06:58:26***Saving seen data "./dancer.seen"
06:59:35 Quit ehntoo (Remote closed the connection)
07:04:48 Join pengo [0] (
07:06:09pengo "Auto change directory" is dumb and less descriptive than "move to next folder". It should be "Fall-thru to next directory" or something
07:33:55 Join Jungti1234 [0] (n=jungti12@
07:44:13 Quit pengo ("Download Gaim:")
07:51:30Slasherihmm, am i correct that there are color lcd remote controls for h300 and they can connected to h100 also?
07:51:49Slasheriif so, i would like to buy one and write the driver to support h100 with color lcd too :)
07:54:22 Join LinusN [0] (
07:54:42LinusNSlasheri: no, the h300 remote is not color
07:54:50Slasheriah :/
07:55:02LinusNsame shit, different buttons
07:55:05Slasherii thought i saw somewhere a picture where it was color but i cuold be wrong
07:57:14 Join ender` [0] (i=ychat@
08:00:08 Join _FireFly_ [0] (
08:14:07 Quit _FireFly_ ("Leaving")
08:24:37 Join edx [0] (
08:26:54 Join B4gder [0] (
08:28:13 Join einhirn [0] (
08:29:50 Join DMJC [0] (
08:30:01DMJCmy broken iriver is getting replaced..
08:30:05DMJCH340 series
08:30:20DMJCwas hoping to get a H140 back
08:31:02B4gderask for a BDM to go with it! ;-)
08:34:21LinusNdamn! i can't update the lcd display data on the h300 unless i write *really* slow
08:35:02LinusNone pixel/tick!
08:35:10*B4gder faints
08:35:45LinusNthere's obviously something i did wrong, but what?
08:36:00LinusNstill, the lcd is displaying my data
08:36:11B4gderthat's good at least
08:36:23B4gderplaying any videos yet?
08:36:26*B4gder ducks
08:38:00LinusN1 frame / minute :-)
08:42:52 Join DJDD [0] (
08:50:50SlasheriLinusN: oh =)
08:51:19Slasheridid it even boot compiled code?
08:52:16Jungti1234It is much developments.
08:54:09DMJCis it in color?
08:55:27DMJCso will rockboy play GBA games?
08:55:45B4gdera _little_ premature question
08:58:08DMJCthis sucks
08:58:13DMJCI was enjoying rockbox a lot
08:58:21DMJCnow I'm stuck with crappy iriver firmware again
08:58:27***Saving seen data "./dancer.seen"
08:58:47LinusNDMJC: hang on, i'm working hard on the h300 bootloader
08:59:07LinusNah, full speed lcd access
08:59:38LinusNit turns out that it's way better to be smart than to be silly
09:00:16B4gdernow that's a first
09:00:25*B4gder writes this down
09:00:29LinusNtook me 36 years to find out
09:02:05 Join _FireFly_ [0] (
09:02:42DMJChow fast is fullscreen lcd access?
09:02:49LinusNdon't know yet
09:02:56DMJCheh any chance of video working properly (ROFL)
09:06:17 Join webguest72 [0] (
09:08:02 Quit webguest72 (Client Quit)
09:10:12 Part LinusN
09:10:21 Join LinusN [0] (
09:17:18 Quit Kohlriba (Read error: 110 (Connection timed out))
09:26:57 Join edx_ [0] (
09:28:21 Quit edx (Read error: 104 (Connection reset by peer))
09:28:35 Join Kohlrabi [0] (
10:18:22 Join webguest64 [0] (
10:20:53 Join webguest09 [0] (
10:20:53 Quit webguest64 (Client Quit)
10:23:03Jungti1234No one exist here?
10:23:20B4gderno, we are all figments of your imagination
10:23:58Jungti1234Can you answer on my question?
10:24:12B4gderdon't ask if you can ask, just ask
10:24:59LinusNwee! i can display text on the h300 lcd!
10:25:18LinusNand a seriously distorted rockbox logo :-)
10:26:15LinusNi can't understand why lcd-16bit.c has an 8-bit framebuffer
10:26:21Jungti1234 <- Do you know why break so?
10:26:49Maxou`hum, do we have a "readme' about the new WPS disposition ?
10:26:52B4gderJungti1234: because the WPS is wrong
10:27:04B4gderMaxou`: the mailing list post I made
10:27:12LinusNB4gder: that's the recording screen
10:27:27*B4gder is clueless
10:27:32B4gderI've never recorded
10:27:44Maxou`ok B4gder
10:27:48LinusNJungti1234: the recording screen isn't unicode compatible
10:27:59Jungti1234You informed 'WPS' meaning, but I forgot. Can you respeak?
10:28:15B4gderJungti1234: if you run odd patches, you should discuss bugs with the patch authors first
10:29:16Jungti1234I used unicode rockbox.
10:29:33B4gderJungti1234: yes, and that is a patched version that is not in the core code
10:30:11B4gderJungti1234: I'm just saying that you shouldn't expect all of us to know the quirks and details of all patches that are flying around
10:30:53LinusNthe recording screen uses the internal system font, which is not unicode
10:33:18Jungti1234yes, and that is a patched version that is not in the core code
10:33:18Jungti1234I'm just saying that you shouldn't expect all of us to know the quirks and details of all patches that are flying around
10:33:18Jungti1234the recording screen uses the internal system font, which is not unicode
10:33:25Jungti1234Is record part developing?
10:34:42Maxou`hum, the Snap WPS won't show the "Normal" or "SFL" bmps oO
10:34:48Jungti1234Is unicode supported to recording screen sometime?
10:35:26markunI guess it will
10:35:31LinusNsome day
10:35:42Jungti1234hi markun
10:35:47markunmorning :)
10:35:56Jungti1234I waited you.
10:36:10markunare you using the simulator now?
10:36:20Maxou`the good idea would be to do a "WPS" Selector in config menu, instead of browsing the .rockbox folder (which would allow you not to show this folder), the menu would list the WPS present.. no?
10:37:41B4gderMaxou`: there is one already
10:37:49Maxou`escuse me lol
10:37:58B4gderand it browses .rockbox/wps where they are located these days
10:38:06Maxou`sry ^^
10:38:16Maxou`but.. don't understand this WPS
10:38:22Maxou`don't shows two BMPs
10:41:15Maxou`(what's the rockbox default font ? :x)
10:41:24Jungti1234Is there simulator apart?
10:42:31Maxou`"rockbox_default" ... /me shuts up
10:58:31***Saving seen data "./dancer.seen"
10:59:02Jungti1234It is examination tomorrow. :(
10:59:15Jungti1234Must study. Bye.
10:59:37 Quit Jungti1234 ("Bye Bye~")
11:01:24linuxstbpreglow: You around?
11:02:29 Join _FireFly__ [0] (
11:02:29 Quit _FireFly_ (Read error: 104 (Connection reset by peer))
11:02:53 Nick _FireFly__ is now known as _FireFly_ (
11:03:06 Quit _FireFly_ (Client Quit)
11:03:39 Join _FireFly_ [0] (
11:12:34 Join amiconn_ [0] (
11:12:35 Quit _FireFly_ (Read error: 104 (Connection reset by peer))
11:13:57 Join _FireFly_ [0] (
11:14:43 Quit markun (Read error: 54 (Connection reset by peer))
11:14:48LinusNlinuxstb: u there?
11:16:58 Join linuxstb_ [0] (
11:17:09LinusNlinuxstb_: u there?
11:19:28 Join markun [0] (
11:19:57markunproper joined arabic in rockbox (from a bbc news article)
11:21:12linuxstb_LinusN: Yes.
11:21:50LinusNi wonder why the framebuffer is 8-bit in lcd-16bit.c
11:23:12 Quit linuxstb_ (Read error: 104 (Connection reset by peer))
11:23:16 Join linuxstb_ [0] (
11:25:33linuxstb_It's simply to make it a type that is consistent with the other framebuffers. I did originally have it as an array of shorts, but then changed it.
11:26:36LinusNwhy should it be consistent?
11:27:15linuxstb_I'm not arguing in favour of it. I just wanted to make the least changes to Rockbox at the time. I would be happy to change back to shorts if you think that's better.
11:28:42LinusNhehe, i just did, for the h300 port, but i haven't committed it
11:29:01linuxstb_No problem then :)
11:29:24linuxstb_Did you change the screen_dump function in apps/misc.c ?
11:30:44 Quit amiconn (Read error: 110 (Connection timed out))
11:30:45 Nick amiconn_ is now known as amiconn (
11:30:59LinusNno, not yet
11:33:24linuxstb_How are your 16-bit values stored? I assume you noticed that they are byte-swapped on the ipod.
11:33:46 Quit linuxstb (Read error: 110 (Connection timed out))
11:35:43LinusNthey are stored big-endian
11:37:20linuxstb_So we should use the htobe16 and betoh16 macros instead of swap16
11:38:05LinusNor something like that
11:46:00preglowLinusN: sounds like you're progressing nicely
11:46:02preglowlinuxstb_: i am now
11:46:03linuxstb_Do you recognise any of the LCD commands used by ipodlinux? i.e. do they match any datasheets you've seen.
11:46:47LinusNno, it isn't even close to the h300 lcd
11:46:53linuxstb_preglow: I added a yield() to the usleep command in the bootloader, and tried a lcd_puts_scroll() but it didn't work. Is there anything else I needed to do?
11:47:34preglowlinuxstb_: yes, you need to do a couple of yields before you lcd_puts_scroll as well, so that the scrolling thread has time to get going
11:47:46preglowlinuxstb_: the first thing the scrolling thread does is reset the scrolling_lines variable
11:48:09linuxstb_OK, I'll try that.
11:48:49LinusNdid any of you convert a color rockbox logo?
11:49:21linuxstb_Yes - it's in the bootloader directory. But it's byte-swapped for the ipod.
11:49:34LinusNin the bootloader dir???
11:49:49preglowbootloader displays a couple of logos
11:50:03linuxstb_It's a very small logo - we probably want a bigger one for the main splash screen.
11:50:23LinusNhow did you convert it?
11:51:09 Join ender1 [0] (i=ychat@
11:51:59linuxstb_A poor quality perl script that I hacked from one in the ipodlinux source and some manual editing.
11:52:21linuxstb_I mean my changes were poor quality.
11:52:36linuxstb_It takes an xpm file as input.
11:53:19linuxstb_I could send it to you if you wanted to clean it up. Or you could start from something else.
11:54:24LinusNi was thinking of updating bmp2rb
11:54:41linuxstb_That's a better plan.
11:54:57LinusNthat's what i thought you did
11:55:52linuxstb_No, I didn't spot that program.
11:56:03LinusNi see
11:56:19linuxstb_Looking at it now, it would have been a lot easier.
11:57:06 Join ashridah [0] (
12:04:49 Join webguest81 [0] (
12:04:56preglowthis setting up of interrupts is going to be tricky
12:05:02webguest81Hello all
12:06:00webguest81I'm wondering if you guys thought to add the patch for delete the syteme volume info dir of windows ?
12:06:56webguest81it is big dir for nothing, with the directory cach, that make few ko for nothing
12:07:47SlasheriHmm, that's interesting. iriver display just got a full of straight lines and playback stopped. There must be some bugs in the code :)
12:07:58webguest81the "master" LinusN, any thought?
12:08:08 Quit edx_ ()
12:08:23 Quit ender` (Read error: 110 (Connection timed out))
12:08:26webguest81in your code Slasheri you mean :D
12:08:29markunSlasheri: I have found another bug, but I don't know how to reproduce it
12:09:03markunSometimes my iriver shuts down after it played the last file in the playlist.
12:09:11Slasheriwebguest81: i believe the problem is in cvs
12:09:14preglowSlasheri: did your iriver get well again?
12:09:20Slasheripreglow: yep, working fine
12:09:59Slasheripreglow: i am now writing a safe flash plugin for iriver and specifying a firmware file format, that has bootloader and rombox splitted into a single file
12:10:35 Join lamed [0] (
12:10:36webguest81it will be commited for users wanted it?
12:10:48webguest81it could be fun :)
12:10:55lamedolah everyone.
12:10:55Slasheriwebguest81: yes, commit will come later
12:11:19Slasherithen it should be possible for anyone who wants, to erase the iriver firmware and replace it with rockbox fw
12:12:10webguest81now recordings is in CVS not a lot of things missing
12:12:28webguest81why we need of iriver fw now? :D
12:12:40webguest81just a full support of the remote
12:12:46lamedquestion: say you have changed a file, will invoking cvs update -dP delete your change? (-i think it tries to diff the file to the original by version and patch it back to the new file!)
12:12:48Slasherithe code will also introduce a little more advanced boot loader for rombox, that has a simple failsafe menu and could recall some settings from eeprom
12:12:55webguest81and definitively bye bye to the original fw
12:13:13_FireFly_if my wps-widget gets into cvs then the main-parts are working also on remote
12:13:13LinusNlame, no it will not remove your changes
12:13:14webguest81Slasheri: rombox like archoses ones?
12:13:35webguest81yes _FireFly
12:13:51webguest81I hope WPS in remote will come soon
12:13:53Slasheriwebguest81: sorry but i don't know how rombox works on archos.. this will be probably a little different
12:14:07lamedlinusn: how not? it _does_ updates the file, correct? so am i right with my assumption?
12:14:27webguest81Slasheri: no problems it was just for understand the concept
12:14:42Slasheriwebguest81: iiuc, archoses have two rockboxes in flash. Iriver will have only one and a bootloader, that doesn't need to be touched when upgrading rombox
12:14:44_FireFly_webguest81: if you can compile yourself then you can get my patch in the tracker
12:14:48LinusNlamed: it merges your changes with the updated file
12:15:10_FireFly_if there where problems it will display a error message about it
12:15:24webguest81_FireFly: you know why isn't commited yet?
12:15:41lamedlinusn: so if the file in the repository is changed too much, what happends?
12:15:50_FireFly_no dev with cvs access had yet enought time to review it
12:16:05LinusNlamed: then you have conflicts, which are marked in the file
12:16:22webguest81_FireFly_: isn't the part of TiMiD the guys made the begin of mutidiplay support?
12:16:30Slasherihehe, i found an ogg file that crashes rockbox
12:16:41Slasheriwill try on simulator soon
12:16:46_FireFly_webguest81 he had created the base for it yes
12:17:09webguest81and you the WPS part, yes?
12:17:27preglowSlasheri: i'm willing to bet it's one of the ones with a ton of faulty granuleposes
12:17:29lamedlinusn: it won't product other .reg .org files? just mark them in file? -that's interesting, i should add a commit to the wiki
12:17:38preglowSlasheri: try using ogginfo on it
12:17:43lamedlinus: sorry, .rej
12:17:46webguest81_FireFly_thanks for this, I hope it will commit
12:18:09_FireFly_as i said it already if you can compile rb yourself you can use it already
12:18:47webguest81I can't :(
12:20:05webguest81LinusN, Bagder: is the official release for irivers will come soon (or wait for h300), now is it quasi full
12:20:16Slasheripreglow: Hmm, ogginfo doesn't display anything weird. But i will check more when i get home
12:20:27webguest81one bug freeze and bugs hunting could be fine for irivers
12:20:29SlasheriI could then send the file to you if you want
12:22:06markunSlasheri: Do you think auto shutdown should happen when playback stops and the user has not touched his unit for a while? I don't think it should.
12:22:44_FireFly_markun when should then auto shutdown work ??
12:22:51LinusNmarkun: of course it should
12:23:00lamedlinusn:thanks! btw..
12:23:34markunWell, after playback stops it should wait another timeout period before shutting down. Now it sometimes shuts down imediately
12:23:52SlasheriHmm.. immediately?
12:24:05Maxou`the auto shutdown shouldn't happen when the playback is "paused", only when it's stopped.. like the original fw.. no?
12:24:08Slasherithat is probably then a bug
12:24:19markunOr am I the only one getting this behaviour?
12:24:25LinusNit should not shut down immediately
12:24:34 Quit ashridah ("Leaving")
12:24:35webguest81bye all
12:24:44 Quit webguest81 ("CGI:IRC")
12:25:15markunlooking at handle_auto_poweroff I can see how this can happen.
12:25:19LinusNhmm, maybe that's how the system works atm
12:25:26preglowSlasheri: sure
12:25:37LinusNit is calculated from the last disk access iirc
12:25:37markunMaybe last_event_tick should be reset when playback status changes
12:25:51LinusNmarkun: yes
12:26:10crwlbtw, has someone thought of implementing the stereo width setting for iriver?
12:26:28preglowcrwl: yes, i have, but it needs some internal changes
12:26:31crwlit's difficult to listen to a lot of stuff from sixties right now :)
12:26:37markunBecause I listen to albums a lot and when they finish I often look at my iriver to discover it turned off.
12:26:42 Join webguest53 [0] (
12:26:48preglowcrwl: markun's got something you'll like, then
12:26:56webguest53how is the h3xx port going ?
12:27:25crwlpreglow, hum?
12:27:40preglowcrwl: he's made a crossfeed plugin
12:27:47preglowcrwl: that makes even vintage beatles sound pleasant
12:28:18LinusNwebguest53: forward i guess
12:28:48markunThere are a few problems with it. It takes up quite some CPU, makes the volume a lot lower and cannot be turned off right now.
12:29:50markuncrwl: Do you want a patch?
12:30:13 Quit DMJC (Read error: 104 (Connection reset by peer))
12:30:30preglowthe cpu usage worries me
12:30:38crwlmarkun, why not
12:30:38preglow'cause it's quite a simple plugin, really
12:31:13markunpreglow: I changed the code a bit, maybe you can take a look at it again.
12:33:30markunpreglow: I was trying to get rid of all the move instructions (but didn't disassemble to find out)
12:33:41markunWell, not ALL the move instructions :)
12:35:04webguest53long before realese?
12:35:50markuncrwl: you have to apply the patch in the apps dir
12:37:04crwlmarkun, okay, i'll try it when i get home again
12:37:35preglowthink it's about time i made one tree for each platform
12:37:35 Quit webguest53 ("CGI:IRC (EOF)")
12:37:46preglowi'm working on, that is
12:38:00crwlright now i need to go slightly closer to slasheri to increase my iriver's l33t power
12:38:18B4gderto Finland? ;-)
12:38:44crwlwell, here i am already
12:38:59crwlprobably in the same town in fact
12:39:05*B4gder didn't check
12:39:19crwlin fact if i'm not mistaken, i'm going to the adjacent building
12:39:47preglowleetness level must be increasing rapidly, then
12:39:52preglowgive him a pat on the back for me
12:40:30crwlanyways, the bus is coming... ->
12:40:35lamedwhy everyone saying the nokia chagger polarity is inverted then iriver's charger? the signs are the same (hadn't had a multimeter yet to check)
12:41:09preglowmarkun: it's looking a _lot_ tighter now, but still there's something to be done
12:42:26markunlamed: my nokia charger doesn't say what the polarity is..
12:43:05markunpreglow: would have been nice of gcc to do the optimisation for me
12:44:06preglowmarkun: you still can't count on the compilers completely, i'm afraid
12:46:38lamedmarkum: i dont have one here to see if i'm wrong :/ i'll have to check it at the weekend.
12:47:10preglowmarkun: i think it should be possible to convert that it into quite a nice little asm loop
12:47:20B4gderlamed: several people have claimed they fried their units that way
12:47:25preglowlooks like there is enough room for all constants and arrays in the registers
12:47:32preglowso you don't need to do the constant reloading gcc does
12:47:36B4gderand LinusN once investigated the results of a nokia charged irivier, iirc
12:47:39preglowdsp is all about assembler, i'm afraid
12:48:23markunpreglow: would you be so kind to take a shot a rewriting it?
12:48:43preglowyep, just need to have some lunch
12:48:44 Join Vladoman [0] (
12:48:50_FireFly_the problem with the nokia chargers could be that these charger can only deliver 500mA
12:50:03LinusNmy nokia charger doesn't fit the h100 connector
12:50:23linuxstb_I once plugged my iriver charger into my Nokia, and nothing bad happened - but the phone didn't charge either.
12:50:24LinusNthere isn't just one nokia charger
12:53:28 Join ender` [0] (i=ychat@
12:53:28 Quit ender1 (Read error: 104 (Connection reset by peer))
12:58:34***Saving seen data "./dancer.seen"
12:59:22HClhow many volts is the iriver h120/h140 battery supposed to be?
13:01:53LinusN4.2V when fully charged
13:04:38linuxstb_preglow: Did you make any progress with crt0.S ?
13:08:06markunHCl: problems with our battery?
13:08:49preglowlinuxstb_: none at all
13:09:14preglowlinuxstb_: searched for the relevant parts in ipl, found something, and started comparing with some docs
13:12:48linuxstb_I think the best approach is just to do something similar to the iriver code.
13:13:49linuxstb_For now, I think we can ignore any hardware initialisation (assuming the Apple flash firmware has done enough) and just set up the memory contents correctly - i.e. zero out the bss, copy the iram section etc
13:14:03preglowwell, yes, and we need to set up interrupts
13:14:34preglowwhich is a bit of work
13:14:44linuxstb_Yes, but can that be done later? i.e. get a simple crt0.S first, and then add the rest.
13:15:25linuxstb_But interrupts is probably the most important thing to do next anyway.
13:15:30preglowit is
13:15:36preglowwe need it for buttons
13:15:46Slashericrwl: hmm, where are you? =)
13:17:24preglowmarkun: dsp.c:513: warning: operation on `index' may be undefined
13:17:27preglowmarkun: do you get that?
13:30:03 Quit Rick ("I don't need to be here.")
13:31:58 Join Rick [0] (
13:32:41preglowlinuxstb_: did you get the scrolling to work, btw?
13:36:39linuxstb_preglow: No, but I only had a quick attempt. I added about 4 or 5 yields between lcd_init and my lcd_puts_scroll and it doesn't help.
13:36:59preglowi just added an usleep there
13:37:03preglowworked fine then
13:37:05linuxstb_I'e added a yield inside the usleep() while loop as well.
13:37:28linuxstb_OK, I'll try that. How long did you usleep for?
13:38:30preglowbut that shouldn't matter
13:40:20preglowi'll have a quick diff to see if i did anything else i forgot
13:41:55linuxstb_No, that doesn't work either.
13:44:04linuxstb_preglow: Have you built the arm compiler for cygwin?
13:45:01SlasheriHmm, i got the internal eeprom working and a dump of it. Now it should be possible to use it on bootloader :)
13:45:19preglowlinuxstb_: i can't stand cygwin
13:46:13linuxstb_I was just curious if it compiled OK there.
13:46:22linuxstb_I forgot you use your linux box for compiling.
13:46:34preglownot my box, at the moment
13:46:37preglowi need to get that going again
13:48:08pregloweither that or get cross compilers to work on amd64
13:51:48*LinusN has a nice coloured rockbox logo on his h300 lcd :-)
13:52:02B4gderpictures pictures!!!
13:52:56LinusNwill take pics in an hour or so
13:53:01LinusNgotta go
13:53:40preglowlinuxstb_: dev box is lagging like mad right now, so can't be completely sure, but doesn't look like i've done anything that's not commited
13:53:59linuxstb_Can you send me your ipod.c file?
13:54:25linuxstb_But if it's in a mess, don't worry - it's not important.
13:54:33preglowit's not
13:55:32linuxstb_Are you still getting problems with corrupted fonts?
13:56:19preglowlinuxstb_: no
13:58:07markunpreglow: yes, I get that as well.
13:59:11preglowlinuxstb_: does it work?
13:59:16linuxstb_preglow: It must be your thread tests that cause the scrolling to work. That seems to be the only difference.
13:59:21linuxstb_(yes, it does work)
13:59:43linuxstb_t2 is never incremented though.
13:59:47preglowahh, i know
14:00:09preglowi used bbb to test the scrolling thread
14:07:43HClmarkun: nah
14:09:13preglowmarkun: my dev box is lagged to hell curently, and i need to go, will look at optimising later
14:09:24linuxstb_HCl: Did you write databox?
14:10:13markunok, bye
14:11:04 Join DJDD_ [0] (
14:17:35wubblaharr harr... I can't wait for the H300 port :-D
14:18:56wubblahow can LinusN execute code on his h320?
14:19:11wubblai thought that there is no bootloader, yet
14:19:28B4gderbecause he as a BDM
14:19:42B4gderwhich can load code and execute in the cpu
14:20:10 Join Sandking [0] (
14:20:14wubblai thought that was just some kind of debugging-device...
14:20:28B4gderit is
14:20:38 Quit Kohlrabi (Read error: 110 (Connection timed out))
14:20:40wubblaso that you can read specific registers or something else like that...
14:21:11B4gderthat it can
14:21:15Slasheriwubbla: it allows something like full control to the cpu and system bus
14:21:30wubblai see :)
14:21:35Slasherithat is what hardware debugging devices are for
14:22:23wubblado you need to reverse-engineer everything or is ther some kind of documentation?
14:22:50Slasheriwe have documentation for the cpu, display and almost for all peripherals
14:23:30Slasheribut at least electrical reverse-engineering is needed to create the schematics etc.
14:24:05Slasherithere is no need to reverse engineer the firmware itself
14:24:10 Join arkascha [0] (
14:24:30wubblaby the way, why are some header files in the CVS "marked" as executable?
14:24:57wubblafor example: firmware/export/audio.h
14:24:58B4gderI blame the windows people
14:24:59Slasherihmm, probably they are committed from windows
14:25:03CassandraProbably because someone commited them that way.
14:25:18CassandraSillly Windows users. No cookie.
14:25:30wubblahehe :)
14:25:58CassandraBTW - someone think of a better name than "3d" for my WPS.
14:27:14wubblaare mpeg-files decoded using a software codec or utilizing a hardware device?
14:27:37 Quit DJDD (Read error: 110 (Connection timed out))
14:27:42CassandraHmm - that link doesn't work. It's at the bottom of the page.
14:27:44B4gderwubbla: on iriver it is all done in sw
14:27:54wubblathis is related to the h300
14:28:15wubblaisn't there a hardware decoder on board?
14:28:30CassandraNot unless you count the CPU. :)
14:28:36 Join ipnetukr [0] (
14:28:44B4gderand the ADC/DAC
14:28:57SlasheriCassandra: oh, looks like a nice wps :)
14:29:16CassandraI'll be commiting it into CVS before too long, I hope.
14:29:46B4gderand you'll save some 250 bytes with the new loading system! ;-)
14:29:55Cassandra*nods* Yep.
14:30:07*B4gder used that wps when modifying to this new system
14:30:34ipnetukr:) my first weekend of exclusive rockbox usage on the iRiver went very well indeed... i didnt miss the irRiver firmware once :)
14:30:57ipnetukrits got a lot better since the Archos days
14:30:58Cassandra*grin* The only reason I didn't use WPS/3d/ in the first place was the extra space it would require in the cfg file.
14:31:19*Cassandra hasn't used the iRiver firmware in over half a year.
14:31:45ipnetukri needed firefly's wps patch to push it into the usable catagory
14:31:59ipnetukrbtw - is the wps on that hard coded atm?
14:32:01markunI turned on the iriver firmware by accident when I was playing with the remote
14:32:21ipnetukr6 lines on the remote :)
14:32:21CassandraWPS is just eye candy, really.
14:32:31CassandraBut I think a lot of people consider that important.
14:32:50CassandraThat's why I've been trying to design a comprehensive, clean looking RPS.
14:33:32CassandraI'm hoping we'll consider using a graphical WPS as default on iRiver. I think the original spartan Rockbox look puts a lot of people off.
14:33:59CassandraWhich is a shame, because Rockbox beats the iRiver firmware hands down in terms of performance.
14:34:06Cassandraand features.
14:35:01CassandraIs firefly's patch for remote WPS support? That's not in the main CVS yet, is it?
14:36:23CassandraWish I could use greyscale on the WPS rather than resorting to optical illusions.
14:36:56 Quit lamed ("CGI:IRC (EOF)")
14:37:01ZagorCassandra: I like it, it gets that C64 look... :-)
14:37:09CassandraI'm hoping that maybe we'll finally get 4 grey WPS as a side effect of the work on the H3xx and iPod colour LCDs.
14:37:42Cassandrazagor: Ah - retro chic. :)
14:39:12CassandraMust get around to adding an "in WPS only" option for the status bar.
14:39:48solexx_Cassandra: I'd appreciate the opposite :)
14:39:55 Join Kohlrabi [0] (
14:40:06linuxstb_I thought the plan was to add a tag to a WPS to explicitly enable or disable the status bar.
14:40:32CassandraErnm in browser I mean.
14:40:53_FireFly_i had a patch for this
14:41:02B4gderI'll make a system that'll generate a .cfg file for each WPS we bundle
14:41:12CassandraNot sure. badger has some plans for bundling config files .... oh, there we go.
14:41:31B4gderto easier get the proper font set etc
14:41:49Zagorbut I'd still want status bar to toggle off in wps and on in browser
14:42:02B4gderyes, me too
14:42:12CassandraIt was a design feature of 3d that it works with a wide range of fonts.
14:42:14_FireFly_Cassandra: but i have to adapt it to the new version if my widget gets into cvs
14:42:34CassandraWhat's the widget?
14:43:01_FireFly_which adds multi-screen support to wps and loading of remote-wps(*.rwps) files
14:43:12CassandraOh, right.
14:44:16_FireFly_maybe i find some time today to make a patch which readds the possibility to disable/enable explizitly the statusbar in wps upon my wps-widget patch
14:44:19CassandraYeah - that would be good, but others than I have a better idea of the eventual design goals of wps support. I know that there are plans for the two screens to function independently.
14:44:45_FireFly_yepp with few modification my code is ready for that
14:44:58_FireFly_to handle wps independently
14:45:05ZagorI still very much oppose that, though.
14:45:23CassandraI think it may add unnecessary complexity.
14:45:45CassandraBut it may be possible to have it as a (non-default) option.
14:46:07ZagorThat would still make the code as complex. and the bugs.
14:46:32CassandraI'm with you Zagor. I don't really see the point.
14:46:39_FireFly_i don*t think it make sense to have idependent handling of remote and main-display on iriver
14:47:11CassandraFirefly: Fix up the two patches, and I'll try to have a look at them with a view to merging them into CVS.
14:47:28CassandraRemote WPS support is something we could certainly do with.
14:48:03CassandraGotta make sure it doesn't break Archos builds though.
14:48:21_FireFly_it shouldn't but i can't test it
14:48:39 Join Jungti1234 [0] (n=jungti12@
14:48:56_FireFly_my widget uses the old wps-code with some modification to the new interface which i had put upon these function
14:49:03 Join webguest78 [0] (
14:50:46B4gderok, I'll show you how I imagine the bundled-wps system
14:53:37ipnetukrfirefly - sorry, had to go and do work for a bit - did i hear you say your wps-widget patch supports .rwps files?
14:56:20 Join muesli- [0] (n=muesli_t@
14:56:30ipnetukrcool... :)
14:57:08ipnetukri love being able to see track length on the remote (radio shows)
14:58:12ZagorB4gder: wouldn't it be a better idea to create a new cvs module for wps files?
14:58:25ZagorI imagine it can grow pretty extensive
14:58:36linuxstb_Zagor: Good idea.
14:58:39***Saving seen data "./dancer.seen"
14:58:57B4gderI wouldn't mind that, I just thought users might
14:59:07linuxstb_Why should users care?
14:59:22B4gderusers who checkout code
14:59:43B4gderbut sure, I'll create one
15:01:35linuxstb_Have you thought about target specific WPS files in your build scripts?
15:02:53linuxstb_OK :)
15:03:22B4gderI'll have them depend on the LCD
15:03:35B4gdersize and depth I guess
15:04:09 Quit webguest78 ("CGI:IRC (Ping timeout)")
15:06:26 Quit Sandking ()
15:07:47 Join Febs [0] (
15:08:33*Cassandra wonders if it's worth giving the config style stubs a .thm extension, so we can then allow users to select a theme.
15:08:48CassandraOr is this all getting ridiculously complex.
15:09:38linuxstb_I'm not writing any code, so I like the idea :)
15:10:44B4gderyou mean so that we could remember the theme file for next boot?
15:11:03B4gderthese options will be remembered too
15:11:14linuxstb_It also makes it more future-proof. When other parts of the rockbox UI inevitably become customisable.
15:11:39B4gdersure, I just don't see the difference between a .thm and a .cfg
15:11:57linuxstb_The syntax is the same, the purpose is different.
15:12:43linuxstb_But maybe a ".thm" is just a ".cfg" inside a special directory.
15:13:01B4gderI think we should at least start that way
15:13:08linuxstb_I'm just thinking that "browse WPS" could be changed to "browse themes".
15:13:12*Zagor votes, as usually, for not using silly three-letter acronyms
15:13:18CassandraAlthough I'm not entirely sure that moving all the skinning options into the WPS file itself isn't a better approach.
15:13:38CassandraThen people could have a different font for their WPS and their browser.
15:13:49CassandraAnd installing a WPS becomes an atomic operation.
15:14:06Cassandra(As opposed to "install WPS, set status bar, load font".
15:14:12linuxstb_Zagor: I agree. ".theme" is definitely better.
15:14:35B4gderselecting one of those .cfg files in the wps/ dir will be a "set theme" operation
15:15:25B4gder(the site is currently being slow as a toad due to bandwidth hogging)
15:15:41linuxstb_Is the plan for these .cfg files to live in the .rockbox/wps/ directory?
15:15:51B4gderyes, I think so
15:15:52CassandraWho's hogging the bandwidth? And how fast are toads?
15:16:01CassandraThat would make sense, yes.
15:16:05linuxstb_Then I don't think it really matters if they are called .cfg or .theme
15:16:12B4gderCassandra: I believe the people in the office are downloading too much crap
15:17:01CassandraBut I am beginning to wonder if rather than build on all this extra infrastructure for themes, we shouldn't just make font and statusbar configurable from the WPS.
15:17:29B4gderfor that to be good, we first need multi-font support
15:18:01B4gderas I don't want to be _forced_ to switch font in the dir browser just because I don't to load a new wps
15:18:10B4gder"becase I load"
15:18:29CassandraI can see that. Theorectically there's no reason why the current infrastructure can't reload fonts on swapping between WPS and browser is there?
15:18:58CassandraThat might actually be the way forward for now then.
15:19:16Zagorit's rather slow though
15:19:30CassandraIs it? Hmmm.
15:19:42Zagorif you need to spin up the disk to load the font
15:19:51CassandraTechnically what are the downsides of keeping two fonts in memory?
15:20:03CassandraNot good for Archos, is it?
15:21:25crwlSlasheri, i was at MaA for a while
15:22:05Zagorusing extra memory is never good, of course. but there is no other downside as far as I remember.
15:22:24B4gderno, it just hasn't beed done
15:24:06CassandraRight. I may have a poke around with that this afternoon then.
15:24:23thegeek_beeing able to use more than one font in the wps/.theme would be kinda nice;)
15:24:26markunIs there a more elegant way to write: return &(unsigned short){0};
15:25:16markunit used to be return ""; before switching to unsigned shorts from chars..
15:26:13B4gderreturn NULL ?
15:28:19Zagorimho NULL should be used instead of empty strings
15:28:40CassandraWell I don't think I'd be adding support for multiple fonts in the WPS.
15:28:52CassandraJust for a different font in the WPS to the browser.
15:29:37markunZagor, B4gder: ok, I'll change it.
15:30:07ZagorCassandra: if you fix double-font loading I bet you can't keep your fingers away from double-display in wps... :-)
15:30:10thegeek_is it a lot of work to enable support for two concurrent fonts?
15:30:30Zagornot really. the code was written for multiple fonts.
15:32:33 Quit _FireFly_ ("Did anyone see my lost carrier?")
15:39:14 Nick Lynx_awy is now known as Lynx_ (
15:49:18DBUGEnqueued KICK Jungti1234
15:49:18*Jungti1234 ̽ʴϴ.(I go to study.)
15:50:32LinusNlinuxstb_: i updated bmp2rb
15:50:49LinusNis the ipod lcd also 5-6-5 RGB?
15:51:53B4gderI'm pretty sure it is
15:52:06B4gderI recall them saying so
16:03:41linuxstb_LinusN: Thanks and Yes.
16:03:50B4gderI have a first shot at WPS-bundling
16:04:10B4gderI'll start with one single WPS to work out the system with
16:04:10LinusNlinuxstb_: the best part is that you don't need to care about endianness
16:05:03linuxstb_LinusN: Where ? The low-level driver needs to write the data as a big-endian 32-bit word containing two pixels.
16:05:28LinusNok, so the ARM is little endian?
16:05:45linuxstb_Which is why I byte-swap the framebuffer.
16:06:03LinusNok, then we should probably add another format to bmp2rb
16:06:41linuxstb_Yes, I was thinking the same thing this morning.
16:10:18 Quit DJDD_ ("Trillian (")
16:11:34linuxstb_LinusN: Are you planning on committing your changes to the make the framebuffer to 16-bit?
16:11:44LinusNyes, soon
16:12:06LinusNwill probably break the ipod lcd code though
16:12:20linuxstb_I was just about to say don't worry about breaking the ipod code - I'll look at it this evening.
16:13:13CassandraGods I hate the Linux automounter. It's so damn flaky.
16:13:30linuxstb_I never enable it.
16:14:30CassandraIt's useful if you have Windows shares you need to access.
16:14:37dwihnoA small question for all of you... I need a new and fast, quiet sata-150 disk... any recommendations?
16:14:54dwihnoSome WD-10krpm or seagate? (no, I hate maxtor)
16:15:16linuxstb_LinusN: Does that mean that you have a functional H300 LCD driver?
16:15:23B4gderok, starting now all automatic builds _should_ get wpses bundled
16:15:39LinusNlinuxstb_: partly
16:15:49muesli-dwihno samsung drives are damn quiet, nearly no audible
16:16:31dwihnomuesli-: how are they performance-wise?
16:16:53muesli-you wont notice 1 or 2mb more or less
16:17:11dwihnoaccess times? how many rpm?
16:17:36muesli-that count on those useless synthetic benchmarks..they are all in one comparable
16:17:44B4gderdoes anyone bother about such specs these days?
16:18:00dwihnoof course everybody does ;D
16:18:13B4gderthey're all bloody fast now
16:19:23Slashericrwl: oh.. i would have been very close :/
16:19:24dwihnohow about disk rotation speed?
16:19:33muesli-7200 i guess
16:19:45B4gderI like sf
16:19:54B4gder"The connection to the server was reset while the page was loading." ;)
16:20:06B4gderwhat the heck is that for an error message
16:20:08dwihnomuesli-: you aren't a samsung agent, are you? ;)
16:20:10Slashericrwl: btw, 'why did you be there? :)
16:20:32muesli-i had samsung's before..they rocked, no problems ;)
16:20:51muesli-but the new maxtor generation is very quite too
16:20:58crwlSlasheri, i had an unbelieveably boring lecture :P
16:21:16Slashericrwl: hehe, nice :) so you are now studying here? cool :)
16:22:01dwihnomuesli-: a really quiet disk would be nice as well ;)
16:22:24 Quit muesli- ("ich will Khe!!!")
16:24:17 Join _FireFly_ [0] (
16:24:39CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
16:24:39*amiconn appears
16:24:59amiconnToday I had another occasion of non-reacting STOP button in wps :-(
16:25:27crwlSlasheri, yes, i've been doing so for the autumn now... though i've had kind of a slow start :)
16:25:45B4gderamiconn: I've had it too
16:25:48_FireFly_amiconn: have you tried my patch maybe that it solve this bug
16:26:09amiconnSeems I need to go back to old trusty archos for a while
16:26:24LinusNamiconn: no, you need to find and fix the bug
16:27:26LinusNgoing back to your trusty archos isn't a good way to get a bug fixed
16:32:07markunWhere is Moos when you need him.. I want to show off with my arabic progress :)
16:32:52markunlinuxstb_: Is it better to have the ipod in little-endian mode, or is it impossible to switch to big-endian?
16:34:19linuxstb_My understanding is that it is more efficient in little-endian mode.
16:34:26amiconnLinusN: Perhaps you're right, but (1) there are already many areas I'm working on and (2) the iriver playback code isn't exactly transparent
16:35:13_FireFly_amiconn: try my wps-widget maybe it solve this bug, if is a bug in wps
16:35:32Slashericrwl: Hmm, please join #jyu or #ynna on ircnet :)
16:37:36LinusNamiconn: i understand completely
16:38:06Cassandrabadger: Commited my WPS, so you have something else to test your builder against.
16:39:29B4gdermake zip seems to do good
16:39:32Slasheriamiconn: Hmm, what do you mean with non-reacting stop-button?
16:39:37SlasheriDid you press it too long?
16:39:56amiconnNo, I measn that shortt STOP doesn't stop playback, it simply does nothing
16:39:58B4gderSlasheri: sometimes you just can't stop the music with the stop button
16:40:11Slasherii have never encountered that
16:40:22amiconnSomethimes pressing 2 seconds will initiate soft shutdown, but even that doesn't always work
16:40:23CassandraI've had it a couple of times.
16:40:38_FireFly_why does everybody ignore my suggestion ??
16:40:38Slasheriamiconn: but the music was still playing and wps working?
16:40:44amiconnThe strange thing is that the button driver itself works, I can adjust the volume and all
16:40:54CassandraAlso USB insertion while playing is still crashing Rockbox regularly.
16:41:02amiconnGot that 3 times now, within 2 days
16:41:03B4gder_FireFly_: because you say "maybe it solves"... it isn't clearly an attempt to fix the bug, is it?
16:41:23amiconnOne time I could only use the 8-seconds shutdown
16:41:33_FireFly_we don*t know it until someone test it or??
16:41:50B4gderso therefore we should try all sorts of weird patches to see if they solve bugs we see?
16:42:03B4gderseems like a weird way to fix bugs
16:42:13 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
16:42:16LinusNin fact, i think we should take a serious look at _FireFly_'s patch
16:42:20linuxstb_I'm assuming that _FireFly_'s code is a rewrite of the WPS screen. So it may not have the same bug.
16:42:22B4gderthat we should
16:42:30_FireFly_because i can't remember to have this bug
16:42:30linuxstb_That's assuming that the bug is in the WPS code.
16:42:32t0masintresting things happened in the time I was away
16:42:37t0masipod support? :)
16:42:43B4gderhi t0mas
16:42:48linuxstb_t0mas: Progressing nicely.
16:42:55CassandraThe WPS code is very spaghetti currently.
16:44:16_FireFly_is there any reproduceable way to trigger this bug ??
16:44:29B4gderI don't think so
16:44:44B4gderat least I haven't found any pattern
16:44:45*t0mas hasn't touched a C compiler for > 6 weeks
16:46:00_FireFly_i can't also reproduce the bug which crash rb when connecting usb while playing
16:46:26CassandraIt's not an every time thing.
16:46:27t0mascould be a racecondition?
16:46:27_FireFly_the only thing is, that i will get back to filetree after disconnecting from usb
16:46:31t0masthat's hard to reproduce
16:46:34amiconnB4gder: It *might* have to do with config loading
16:46:55CassandraYou could try it with my WPS. I get it quite frequently.
16:47:00amiconnAll occasions I had were in the car, where I loaded a different .cfg at the start, then resumed playback
16:47:08B4gderamiconn: I doubt it, I never load configs
16:47:28amiconnWhen I wanted to stop at my destination, it didn't react
16:47:40SlasheriHmm, there is some problem with the cfg loader at the moment, because some of my older configs crashes it
16:47:43Cassandra(the USB crash)
16:48:02Cassandraslasheri: Confirmed. I've seen that.
16:48:49_FireFly_Cassandra: where can i get your wps ??
16:48:54CassandraOf course I'm also using a logf build.
16:48:59Cassandrafirefly: cvs
16:49:56CassandraWow - there was some really wacky stuff going on with the build system this mornign.
16:50:14Cassandra_FireFly_, check out badger's new WPS module. Should be in there.
16:50:28B4gder_FireFly_: checkout the new wps module
16:50:35B4gderthen make zip as usual
16:51:45 Join DangerousDan [0] (
16:52:05_FireFly_Cassandra: ipodVol is that yours ??
16:52:52B4gderno, boxes is hers
16:53:07B4gderWPSLIST says so too
16:54:36Cassandra(why would we want to ape the iPod. Let's have our own branding. :) )
16:56:30_FireFly_it doesn't crash when i connec/disconnec the usb-cable
16:56:41 Join mashalla [0] (
16:57:25 Join tucoz [0] (
16:57:38_FireFly_while playing a file
16:57:51tucozHi, I get compiler errors when building the X11 sim for h120
16:58:07tucozCC drivers/lcd.S
16:58:18tucozdrivers/lcd.S: Assembler messages:
16:58:27tucozdrivers/lcd.S:316: Error: invalid character '.' in mnemonic
16:58:30tucoz...and so on
16:58:37 Join simonjw [0] (
16:58:40***Saving seen data "./dancer.seen"
16:59:41 Quit simonjw (Client Quit)
17:00:33linuxstb_tucoz: Yes, I get the same. Looks like LinusN's last change.
17:00:55tucozok, then it is probably easy to fix
17:01:57CassandraHmm - can't reproduce at the moment either, but I just got a freeze at "Loading ..." after pressing resume after removing USB.
17:02:40 Join lamed [0] (
17:03:01lamedmarkum: here?
17:03:17amiconnIriver rockbox almost reproducably freezes when plugging USB during startup
17:04:09_FireFly_for me not
17:04:09 Quit Febs ("CGI:IRC (EOF)")
17:04:18 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
17:04:27ipnetukrhow can you have usb plugged in on startup?
17:04:28amiconnLinusN: It seems you broke the H1x0 sims...
17:04:34ipnetukrdoesnt bootloader do usb in place of rockbox?
17:04:51_FireFly_when rb itslef boots you can plug in usb
17:04:52markunlamed: I'm here
17:05:00_FireFly_e.g. when the logo is shown
17:05:11_FireFly_then the bootloader isn't there anymore
17:06:04 Quit B4gder ("time to say moo")
17:07:10CassandraAnd now I'm getting perfect function every time.
17:07:17CassandraI hate intermittent bugs.
17:08:00 Join Febs [0] (
17:08:25_FireFly_yepp these bugs are hard to solv
17:08:25*Cassandra postulates that the problem may be rundb related.
17:08:32 Quit Zagor ("Client exiting")
17:08:52CassandraNo particularl evidence. I just have that enabled and "rundb_init" was one of the last things on the logf
17:10:07CassandraAh interesting.
17:10:14CassandraGot it to hang on the USB screen
17:10:28CassandraLast two things on the logf are both "rundb write"
17:10:53CassandraThen it's hit an ATA error -71 eventually
17:11:06_FireFly_maybe we have the culprits
17:11:38CassandraBut there haven't been any recent rundb changes, I don't think.
17:13:11 Quit lamed ("CGI:IRC (EOF)")
17:14:32amiconnCassandra: Strange thing is that I never observed this usb hang on archos - but the rundb code itself is the same on all platforms
17:15:14 Join muesli- [0] (n=muesli_t@
17:15:24_FireFly_maybe it's something which is used from rundb code
17:15:44_FireFly_a function which is called from the rundb-code
17:16:11CassandraWell, rundb is just the last thing to have written something to logf - that doesn't necessarily mean it's the culprit.
17:16:45 Quit goa (
17:16:45 Quit Slasheri (
17:18:03 Quit linuxstb_ (Read error: 104 (Connection reset by peer))
17:18:14CassandraAnyway, I have to go lie down for a bit. Hope people like the new boxes wps.
17:18:15tucozI have been trying to find out why wmv files are (tried to be) played, although no codec is mapped to load these files.
17:18:47tucozAny ideas where to look?
17:19:47 Join goa [0] (
17:19:50***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
17:20:21 Join lamed [0] (
17:20:47lamedmarkum: what does your crossfeed do..?
17:20:48markunlamed: you were looking for me?
17:21:19 Quit Cassandra (" HydraIRC -> <- IRC has never been so good")
17:21:30markunIt's a filter to make some music sound better through headphones. It's a bit like stereo-width.
17:21:38 Join linuxstb [0] (
17:22:32markunlamed: here is some info:
17:22:46tucozthat should be wma-files, not wmv
17:24:04linuxstbtucoz: It is so users of the iriver firmware can add them to playlists, and then use those playlists under the iriver firmware.
17:24:17lamedthanks markum, very interesting... but i gtg now
17:24:35markunIt's markuN btw :)
17:25:14linuxstbBut if Rockbox doesn't react well to unsupported files in the playlist, then that's a bug. "supported" files that for some reason are unplayable probably cause the same problems, so it's not a problem specific to WMA.
17:25:22tucozlinuxstb, yes I know. But looking at playback.c and trying to find out why rockbox attempts to play them, gives me no hints, as the default (in the switch-case) should give a splash saying that the file doesn't have a codec.
17:25:47linuxstbYes, you're right. WMA is different because there is no codec.
17:26:18tucozBut, now wps is started (and iirc, crashes)
17:27:26linuxstbSlasheri is the person to ask.
17:27:40tucozyes, seems like he is gone
17:29:57NJoinSlasheri [0] (
17:30:44_FireFly_tucoz: he is back :)
17:31:04tucozSpeak of the devil ;) Slasheri, do you know why Rockbox tries to play WMA files?
17:32:04tucozIt seems like you have taken care of the different codecs in playback.c, but it seems like it is not really working as it should.
17:43:36 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
17:48:53 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
17:57:47 Join XavierGr [0] (
17:58:24XavierGrHi all!
17:58:37muesli-high XavierGrakles :D
17:58:39XavierGrWow a lot of activity lately... :)
17:59:21XavierGr_FireFly_: I am impatient for your WPS widget.
17:59:33_FireFly_XavierGr: thanks
18:00:10_FireFly_it's on tracker if you want to use it now :)
18:00:31XavierGrI will definitelly try it.
18:00:36Maxou`wps widget? what's that? :x
18:00:54XavierGrIt is just that I dislike patches... too many fronts with them
18:01:37XavierGrNow I will have to update my radio patch for fmr preset handling, and then again maybe my work will be ignored due to the serious ongoing changes.
18:01:42_FireFly_Maxou`: it habnles multiscreen support ;) and loading of remote-wps-files is also possible as in my unofficial remote-patch
18:02:34XavierGr_FireFly_: Well if you wps widget gets commited then this will be the end for your unofficial remote build!
18:02:58_FireFly_yepp only minor changes are then left :)
18:03:06XavierGrNow almost everything is rendered (except plug-ins and some details)
18:03:41_FireFly_quick-screen, some splash-messages, rolo and virtual keyboard
18:04:12_FireFly_these are left when my wiget get's commited
18:05:08_FireFly_and maybe some additional settings for the remote-display
18:05:20_FireFly_which i had also in my remote-patch
18:14:47 Quit muesli- ("ich will Khe!!!")
18:17:11XavierGrIf I bump the plugin version the old plugins are rendered useless?
18:17:14preglowi just ate enough sushi for two men
18:17:22LinusNXavierGr: yes
18:17:47LinusNXavierGr: they need to be recompiled
18:18:00 Join chopped_pork [0] (
18:18:03XavierGrbummer. then all "3rd party plugins" are uselless if cvs changes the plugin version.
18:18:12LinusNXavierGr: yes
18:18:34LinusNwhy bump it?
18:19:19XavierGrwell I am making a battery benchmark tsr plugin that will write down all sort of info about the battery when the disk is accessed.
18:19:25chopped_porki just would like to note that the boxes.wps in the bleeding edge build hang rockbox here (ihp-120), it just shows a blank screen and the backlight is on forever - needs to be reseted afterwards
18:19:46_FireFly_for me not
18:19:47XavierGrThat way the user can make graphs (e.g. excel) with these values to see vlotage runtime, estimated runtime behaviour e.t.c
18:19:49LinusNnote that they will be incompatible only if you bump PLUGIN_MIN_API_VERSION
18:21:04XavierGrso the "MIN" will make them incompatble? What about the other def? I added 2 new functions to be able to get the data needed?
18:21:16XavierGrWhich def must be bumped and in which case?
18:21:22chopped_porkhm ipodVOL.wps works fine
18:21:30LinusNread the comments in plugin.h
18:22:12chopped_porkregardless of if the player is playing or not...
18:22:25chopped_porkit reads something from the disk for a while and then it blanks
18:22:44XavierGrso if I understood correctly I must insert new functions in the end of the struct and bump only the plugin_api_version?
18:22:46chopped_porkno music aswell
18:23:42XavierGrso if I introduce those functions at the start of the struct, or let's say the middle, the MIN version must be bumped too right?
18:24:14XavierGrah got it
18:24:21XavierGrit is clear on the commnets.
18:24:48amiconnLinusN: ping...
18:24:51XavierGrno only text rendering to let the user know wht he is doing and proper tsr plugin exit is needed.
18:26:27LinusNamiconn: pong
18:26:42amiconn[17:02:30] <amiconn> LinusN: It seems you broke the H1x0 sims...
18:27:16amiconnI wonder why the build system tries to assemble the .S files for the sims in the first place
18:27:26*Jungti1234 ߿ ƿ̽ϴ.(I go to study.)
18:27:28linuxstbBecause SOURCES tells it to.
18:27:45linuxstbOr do you mean we should have a special case to exclude .S files?
18:27:48Jungti1234I'm very hungry. T.T
18:27:51 Quit mashalla (Read error: 110 (Connection timed out))
18:27:54amiconnYes, but w-h-y? Doesn't make sense at all
18:28:16Jungti1234Korea time is 2:28.
18:28:19amiconnAll .S should be #ifndef SIMULATOR in SOURCES
18:28:30LinusNyup, fixing that now
18:28:31linuxstbamiconn: I was thinking the same.
18:29:21 Quit ipnetukr ("Main screen killed!")
18:30:56chopped_porkoh and it doesn't even show the contents of the wps for a brief while like all other do
18:30:57amiconnLinusN: What are you currently working on? I could do some work on the sims (greyscale and colour support, win32 first), then work on completing the 16bit lcd driver
18:31:57LinusNamiconn: sims would be great
18:32:34LinusNi'm currently thinking about how to handle the lcd_bitmap function prototypes
18:32:37chopped_porkhm when i run it via .cfg it's fine
18:32:49amiconnLinusN: What's the problem with these?
18:33:04LinusNthe color bitmaps are 16-bit, so the prototypes are different (char vs short)
18:33:09markunLinusN: will you take a picture of your H3xx with the color logo?
18:33:13amiconnI would keep char *
18:33:17*linuxstb remembers why he made it char*
18:33:23LinusNamiconn: why?
18:33:30amiconnThis way it would be easier to add 24(18) bit support
18:33:34LinusNmarkun: soon
18:33:52amiconnUsing long for 18 bit would be a waste
18:33:54 Join ep0ch| [0] (n=ep0ch|@
18:34:11XavierGrLinusN" So you have a bootloader for your own use now? Or you just load code?
18:34:17 Nick paugh is now known as AliasCoffee (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
18:34:19LinusNXavierGr: i use the bdm
18:34:23tucozhmm, shouldn't a dummy recording screen show up in the sim?
18:34:32amiconnIt shouldn't matter at all internally; we can always cast between char* and short*
18:34:34LinusNusing char means a lot of nasty typecasting
18:34:35tucozin the iriver sim
18:34:56LinusNtucoz: it does in the archos sim, but not yet in the iriver sim
18:35:06tucozLinusN, ok, so I noticed :)
18:35:10amiconnWe need to cast occasionally anyway, e.g. when memset()ting the framebuffer
18:35:30linuxstbBut we don't memset a > 8-bit framebuffer.
18:35:36amiconnWhy not?
18:35:59tucozamiconn, #ifndef SIMULATOR lcd.S worked fine. Got the sim running.
18:36:01LinusNwhen would we memset?
18:36:04linuxstbWhy? The background/foreground colours are not black/white
18:36:10amiconnAt least clearing it to black is possible
18:36:20linuxstbYou clear it to bg_pattern
18:36:27LinusNwhich is 16-bit
18:37:19amiconnLinusN: how about a typedef char pixel; / typedef short pixel; depending on the lcd?
18:37:32LinusNthat's what i had in mind
18:38:41linuxstbI agree, but it's not accurate to call 8-bits of the display a pixel for the 1bpp and 2bpp devices.
18:39:07LinusNor something
18:39:15amiconnLinusN: In fact this sounds like a good idea; 24(18)bit support would then typedef pixel as a struct (or array) of 3 chars...
18:39:38preglowno, bitfield
18:39:46preglowanyone know how efficient those are?
18:39:56Jungti1234bye good night
18:39:59 Quit Jungti1234 ("Bye Bye~")
18:40:02preglowseems most coders don't use them
18:40:11ep0ch|i have a dumb cvs question (in cygwin)
18:40:15amiconnBitfields are probably not a good idea. The layout is compiler and machine dependent
18:40:25LinusNthat's because bitfields aren't very portable
18:41:28preglownot portable?
18:41:48XavierGrdo we know how many characters in default font the screens can fit?
18:41:59XavierGrwidth is the same for all targets no?
18:42:22XavierGrI could test it on the SIM if no one is sure
18:42:41ep0ch|in order to cvs up, i have to "cd rockbox\apps" then cvup, then "cd rockbox\firmware" then cvsup etc. I want to just "cd \rockbox" then run cvs up to update the whole lot, but i get a CVSROOT not specified error, how can i fix it?
18:42:52_FireFly_XavierGr: maybe scree_access has a member in his struct for that
18:43:33linuxstbXavierGr: No. My understanding is that the fonts are a fixed size in pixels, and the LCDs are all different sizes.
18:43:44preglowlinuxstb: any idea why my threading tests would suddenly make threading work?
18:43:50*amiconn votes for fb_type
18:44:02linuxstbpreglow: No, I haven't thought about it.
18:44:24XavierGraslo I think rockbox should have a function that will enable to print large text, without the need to check if it fits. E.g to change line when the text moves beyond LCD width.
18:44:49amiconnsplash() des that
18:45:03XavierGrset to false?
18:45:20LinusNep0ch|: how did you check out the files in the first place?
18:45:45LinusNyou should check out the entire repository in one go
18:45:52LinusNcvs co rockbox-devel
18:46:08ep0ch|LinusN: cvs login, then cvs -z3 co rockbox
18:46:08linuxstbamiconn: I agree with the principle, but am not sure the name "fb_type" is intuitive. To me, that implies the entire type of the framebuffer structure.
18:46:09*amiconn usually doesn't do that.
18:46:38 Join Lear [0] (
18:46:43LinusNep0ch|: then it should work to cvs update in the rockbox dir
18:46:55ep0ch|:( not for me
18:46:56amiconnUsually I do checkout each module individually, because I don't want cvs to create an additional dir level
18:47:06ep0ch|i'll start from clean
18:47:18amiconnI still can cvs up from my cvs root
18:47:19 Join Moos [0] (
18:47:50amiconn(and afaik rockbox-devel doesn't contain all modules)
18:48:30Learlinusn: seems like your little lcd.s fix broke the h120 sim...
18:48:30linuxstbIs there any reason the scrolling logic is duplicated in the different lcd drivers? Is there any difference for all the bitmap LCDs?
18:49:11Learlooks simple to fix though. :)
18:50:11ep0ch|ok my bad, starting from a clean directory seems to have sorted it. ty.
18:52:18MoosHi markun :)
18:53:16ep0ch|meh now i can't build tools, ipod_fw.c errors
18:55:13linuxstbep0ch|: Ignore the error. ipod_fw.c is the last file to be compiled, so everything else would have worked fine.
18:55:14ep0ch|i think the ipod target should be removed until it works
18:55:21linuxstbAre you using the DevKit?
18:55:53LinusNep0ch|: no, the "devkit" should be updated
18:56:08preglowbut ok
18:56:12linuxstbBluechip has fixed the problem, but I've promised him an arm compiler before he makes a new release.
18:56:18preglowanyone have any other ideas on how i can write raw to a partition?=
18:56:47preglowLinusN: oh, i forgot to mention i'm in windows
18:57:06LinusNamiconn: flippixel is supposed to toggle background/foreground, right?
18:57:18preglowLinusN: cygwin dd doesn't do partitions marked 0, and the time i tried writing to a partition it did know, it overwrote everything of importance anyway
18:57:23linuxstbpreglow: Write your own (cygwin) C program to do it that seeks in /dev/sda and writes to the correct position.
18:57:33 Join Eraser|85 [0] (n=eraser85@
18:57:41preglowlinuxstb: i don't really believe the error is in cygwin
18:57:48preglowlinuxstb: that is, not in dd
18:58:02linuxstbCan you access the entire device via /dev/sda ?
18:58:07amiconnLinusN: No, flippixel() is supposed to complement the pixel
18:58:12preglowbah, i keep forgetting the other problem i had
18:58:33amiconntoggling fg/bg isn't possible
18:58:44***Saving seen data "./dancer.seen"
18:59:16linuxstbpreglow: It should be OK once we start running Rockbox itself. Which hopefully will be quite soon. You can just update rockbox.ipod in the same was as your rockbox.iriver.
18:59:47preglowlinuxstb: yep, that'll be grand
18:59:56linuxstbOr you could hack the existing bootloader to load another bootloader from the FAT32 partition.....
18:59:57preglowi think i'll boot linux
19:00:21linuxstbBut that could get confusing :)
19:00:59Learlinusn: should I fix the h1x0 sim?
19:01:13LinusNLear: no need, i will commit soon
19:01:40ep0ch|LinusN: I changed the default rockbox font yesterday to Nimbus-12. The status bar wasn't too happy displaying my numeric volume though. I take it the statusbar doesnt take variable font height into consideration or something?
19:02:20amiconnLinusN, linuxstb: What about fb_data ?
19:02:39 Quit webguest09 ("CGI:IRC")
19:03:15LinusNfb_data is fine
19:03:32LinusNep0ch|: exactly
19:03:49LinusNep0ch|: would be great if you fixed that
19:06:09ep0ch|well no promises as my C knowledge is somewhat lacking
19:11:15 Quit linuxstb ("Leaving")
19:18:13 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
19:19:00amiconnLinusN: Hmm, in fact the .S files were hidden for the simulator builds as long as they did use the CONFIG_CPU macros in the conditionals, since these are never defined for the sims. Obviously the *_SERIES macros are.
19:19:12Eraser|85H10 is not supported by rockbox i suppose.. :D
19:19:16Eraser|85am I right?
19:19:47Eraser|85I just sent an email at iRiver asking for mpc support in H10 series
19:19:58Eraser|85I already know that it'll be useless
19:20:11 Join NicoFR [0] (
19:20:22Eraser|85but why not trying :)
19:20:38 Quit _FireFly_ (Read error: 113 (No route to host))
19:21:02 Quit Moos ("Glory to Rockbox")
19:21:58 Join _FireFly_ [0] (
19:23:31XavierGramiconn: Does the splash function has a character limit?
19:25:01amiconnBagder: It's an impressive build table now. But on my screen it would be good for another 24 configurations and still fit on screen :)
19:25:29 Join webguest37 [0] (
19:27:00_FireFly_XavierGr: yepp
19:27:02 Join linuxstb [0] (
19:27:33 Quit webguest37 (Client Quit)
19:27:37_FireFly_XavierGr: search for MAXLETTERS in gui/splash.c
19:27:42linuxstbamiconn: If we supported all the ipods, we could probably get another 15 configurations there.
19:28:02 Join al [0] (
19:28:05linuxstbSo don't wish for something you don't want :)
19:28:09 Part al
19:28:30 Nick _FireFly_ is now known as Fire|away (
19:28:58 Join alThor [0] (
19:29:46 Nick alThor is now known as rand_althor (
19:31:48 Quit AliasCoffee ("Leaving")
19:42:34XavierGrI will repeat my question, because I can't understand why I get this behaviour: Does splash has a character limit?
19:42:48Maxou`19:27] <_FireFly_> XavierGr: yepp
19:43:20XavierGrops my bad then
19:43:24 Nick Fire|away is now known as _FireFly_ (
19:43:36_FireFly_and XavierGr: search for MAXLETTERS in gui/splash.c
19:43:42XavierGr_FireFly_: Do you know which is the limit?
19:44:09_FireFly_look the message above your last one ;)
19:44:11linuxstbClever - top-posting in IRC....
19:44:27XavierGrhmm 128 characters
19:45:21XavierGrAnyway to call to splashes without the text overwritten?
19:45:31XavierGrThough I don't think there is...
19:45:39XavierGr(to call 2 splashes)
19:45:43 Quit Eraser|85 ("If I give you one Euro and you give me one Euro too, we both have one Euro in the end. But if I give you one idea, and you gi)
19:46:24_FireFly_there is no way because the splash is displayed on the center of screen
19:46:37XavierGryeah i thought so
19:46:58XavierGrSo which is the best way to render a big paragraph?
19:47:44XavierGrlcd_puts for each line of text?
19:47:57_FireFly_currently through several puts-calls afaik
19:48:33XavierGrboring.... to call so many puts
19:49:08_FireFly_you could create a kind of widget for that
19:49:15XavierGranother way could be to let the user change page
19:49:25XavierGrand render 128 characters every page turn
19:56:19markunMy original unicode testfile displays without problems as of today:
19:56:49 Quit linuxstb ("CGI:IRC")
19:57:18XavierGrNice work
19:57:20chopped_porkwoo unicode
19:57:39markunarabic was the last challenge
19:57:45LinusNdamn cool
19:57:47tucozwow, that is cool
19:59:33lamedway the go! it acctually displays hebrew correctly from right to left as well!
20:00:21markunyes, hebrew and arabic are RTL
20:01:05markunBut we should also have a right-allign if the string starts with hebrew or arabic.
20:01:15markunAnd scroll the other direction
20:04:12amiconnHow would you scroll a line that mixes ltr and rtl languages?
20:04:25Maxou`don't scroll.. lol :x
20:04:46*Maxou` gets out −−> [- ]
20:04:55lamedyou'll probebly choose by it's first word.
20:05:08lamedi'm acctually thinking about it a lot lately
20:06:24markunI also think the first word
20:07:41lamedI feel that any directory should be displayed only in one alignment, left or right, by the more items that there are. but usually that isn't a problem because you will have every language sorted one after the other alphabeticly
20:07:42 Join linuxstb [0] (
20:08:04 Join mashalla [0] (
20:08:52linuxstbThe current bleeding edge source download is missing the contents of the codec directories.
20:09:50linuxstbAnd if anyone is interested to know, gcc-3.4.4 and binutils-2.16 both compile cleanly for ARM under cygwin.
20:10:55 Join Sandking [0] (
20:11:21_FireFly_linuxstb: and some more afaik
20:11:40_FireFly_is missing in the archiv of the bleeding edge sources
20:12:02linuxstbIt's possible that it's where FILES uses a "*" by itself - i.e. not *.[ch]
20:19:31 Quit rand_althor ()
20:25:56 Part tucoz ("Leaving")
20:28:18 Quit dpassen1 (Read error: 110 (Connection timed out))
20:29:31*amiconn wonders why 'make install' for the sim doesn't package the wps'es
20:32:35Learamiconn: they were installed for me earlier today.
20:34:42 Quit Sandking (Read error: 110 (Connection timed out))
20:38:33amiconnHmm, it does work for the h1x0 sim, but not for the recorder sim.
20:38:40amiconnEven the wps dir isn't there
20:39:29linuxstbMaybe no WPS files have been deemed suitable for the recorder's smaller LCD.
20:40:42XavierGrwhich was the page on rockbox wiki that showed all players with their features?
20:40:48XavierGrI cant find it...
20:41:26linuxstbamiconn: Yes, both WPS files are defined as needing a minimum of 160x128 LCD.
20:42:14linuxstbLinusN: Nice. That will make lots of people very excited.
20:42:27XavierGrpsss Linus: that indeed rocks!
20:43:49thegeek_LinusN : hehe
20:44:01XavierGrlinuxstb: thanks
20:44:10thegeek_I can imagine the excitement that will result from that
20:44:17lamednice to know: rb still works and record when iriver fw reports low battery. (i need mp3 encoding pronto)
20:44:17thegeek_LinusN pastes one image
20:44:45thegeek_and h3xx owners around the world rejoice;)
20:45:09markunLinusN: great work!
20:45:38Maxou`hum, does WPS now can change teh font?
20:50:51_FireFly_Maxou`: no
20:51:04_FireFly_the cfg file change the font globaly
20:51:08_FireFly_on loading
20:52:32LearGreat, SourceForge just went down...
20:53:03LearAh, back again, still slooow though.
20:53:59XavierGrwhy is sourceforge so slow anyway
20:55:05*Febs finds that the Mistic River thread on H300 progress is already updated with the picture.
20:55:25LinusNthat was fast
20:56:59 Join Jungti1234 [0] (n=jungti12@
20:58:26Maxou`what is the use of "irivum explorer" oO
20:58:48***Saving seen data "./dancer.seen"
20:59:30 Quit lamed ("CGI:IRC (Ping timeout)")
21:01:48 Join Rincewind [0] (
21:03:57*Maxou` will try
21:07:43ep0ch|is the radio working properly these days? (ihp)
21:08:17LinusNep0ch|: define "properly"
21:08:21ep0ch|i'm stuck in 'preset mode' no idea how to change back
21:08:31LinusNhold ON
21:08:44XavierGrI think Anton changed the configuration.
21:08:48LinusNsame key :-)
21:09:13XavierGrNow if fmr preset handling is commted...
21:10:05Rincewindare there plans to make the option-menus configurable? I would like to change the positions of entries I use often
21:11:08LinusNRincewind: no plans
21:11:49Rincewindthen I would like to send this in as a feature-request ;)
21:12:24LinusNor compile your own version?
21:13:02RincewindI would have to learn c first. I only program with delphi al little bit
21:13:13ep0ch|i miss the old radio preset patch. so atm radio preset is only half implemented?
21:13:44XavierGrthe old radio preset patch?
21:14:11XavierGrI thought I was the first to introduce multiple fmr file handling
21:14:31LinusNi believe you are
21:14:32ep0ch|thats probably the patch i'm talking about :)
21:14:43XavierGrthen :)
21:14:51XavierGrbut now the patch is broken :(
21:15:05XavierGrI will try and make again the parts that handle diffrent fmr files
21:16:05 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
21:16:12ep0ch|how does one save a preset atm?
21:16:46XavierGrthere is ony 1 default preset list
21:16:47ep0ch|pressing select just makes the screen go blank
21:16:51linuxstbRincewind: The kind of change you want to do (moving menu options around) might be a good way to start learning C... It should be a relatively easy change.
21:17:00XavierGryou can add 32 presets to this list
21:17:17XavierGrthen if you want another list you must rename the default file and make another.
21:17:37XavierGrand every time you want a different preset list you must rename the files.
21:18:01ep0ch|XavierGr: are you talking about your patch or current CVS build?
21:18:15RincewindI'm too scared of breaking my Iriver, even if it is relatively easy to do
21:18:21XavierGrcurrent CVS
21:19:04XavierGrtrey holding the select button
21:19:14ep0ch|heh just figured :)
21:19:32linuxstbRincewind: It's very very difficult to break your iriver by changing the parts of Rockbox you will be looking at. The worse that will happen is that the firmware crashes, so you just reset, enter USB mode and then restore a known working version of Rockbox.
21:19:46linuxstbYou can also practice on the simulators that run under either Windows or X11.
21:19:47ep0ch|tapping select confused me cos the screen went blank
21:19:54XavierGrhmm ep0ch is right
21:20:04XavierGrthis is a bug
21:20:20XavierGrIt lets the user see the preset list even if there is no presets.
21:20:57XavierGrIMHO it is very confusing
21:20:58ep0ch|whats the default filename for presets?
21:21:07XavierGralso the frequency changes to 0.0
21:21:28Rincewindso I have to download the source, get an editor and a compiler and then I have to find the right place in millions of lines of code...
21:21:33ep0ch|in presets directory
21:21:58ep0ch|hmmm it not picking up my old presets then
21:22:02XavierGrin /.rockbox
21:22:10 Join Moos [0] (
21:22:11XavierGrthat was my patch
21:22:24XavierGrthe preset directory is with my patch.
21:22:26ep0ch|excellent that explains it all
21:23:37Rincewindthe daily build for today will be out in 5 minutes, right?
21:24:18ep0ch|bleeding edge will
21:24:23Rincewindsorry, i didn't read the whole page
21:24:58ep0ch|XavierGr: thanks for sorting my little radio problem out
21:25:28 Join Kaggen [0] (
21:25:55XavierGrep0ch|: np :0
21:26:00XavierGrI mean :)
21:26:04 Join DangerousDan [0] (
21:27:47 Quit mashalla (Read error: 110 (Connection timed out))
21:38:25RincewindI'm afraid building my own version of rockbox is to difficult for me because i have no idea how to work with cvs, I'm working with windows (don't hit me) and english isn't my mother tounge.
21:39:19LinusNRincewind: what did you have in mind regarding the options?
21:40:04RincewindI have a config file in mind, where you can assign positions and names to every item in the options menu
21:40:25Rincewindmaybe a xml-file
21:41:39Rincewindso that you can edit the structure of the menus on your computer
21:43:27thegeek_actually, I think it would be nice to have some kind of universal .theme that would skin the wps(both remote and normal), setup fonts, and could also to some extent skin the menu's
21:44:41Rincewindthere are to things which annoy (is this the right spelling) me: 1. I don't use bookmarks, so I don't need bookmarks to be the first menu item. 2. I use the sleep timer very often, I would like to put it in the first sreen of the options menu
21:45:08Rincewindthe first "to" is actually a "two"
21:45:48thegeek_it would indeed be nice to have some personalization of the menu's, I tend to use the sleep timer a lot too;)
21:46:14Rincewindthemes would be great too, so that you can set your font together with the wps you are using
21:50:17LinusNpersonally, i think the bookmark item is totally misplaced
21:50:30LinusNand i want a new scheme for the sleep timer
21:51:18LinusNRincewind: suggest a new menu layout and we'll consider it
21:51:58XavierGrwhch way can I get a bool value (or 0,1) to see if a file already exists?
21:52:16Rincewindok, i think about it
21:52:23linuxstbI think just doing: fd=open("filename",O_RDONLY);
21:52:38linuxstbIf fd >=0 then the file exists, and you should close it.
21:52:52XavierGrhmm yes that's convinient thanks
21:53:03XavierGrI am still not used to all the handlers of open.
21:53:44Rincewindbut it is very likely that my menu layout represents my personal usage, so it may be not good for everybody
21:54:15linuxstbXavierGr: You may find reading the source code to open (in firmware/common/file.c) educational. It's very clearly written and tells you how all the flags are interpreted.
21:55:42RincewindI have to go now. gn8 everybody!
21:55:45LinusNXavierGr: why do you need to know?
21:56:11XavierGrlinuxstb: Indeed, I used to read only the API document so far
21:56:44XavierGrLinusN: The plugin must give some instructions to the user as to how it must be used.
21:56:54 Quit Rincewind ("bye")
21:57:26XavierGrCause I want to write all sort of stuff, I choosed to let the user know about it from within the batter_testt.txt which will hold the benchmark
21:58:05XavierGrso the plugin will test if there is already a file like that. If it isn't it will just rewrite all the info.
21:58:38XavierGrAnd with a splash in plugin initialization I will ask the user to check battery_test.txt for more information.
21:59:10XavierGrafter the infomration (how to make the benchmark and stuff) there will be all the values considering the battery in sequence.
21:59:35XavierGrthen with a graph program you can make quite handy graphs of the performance of the battery.
22:00:55 Quit Kohlrabi (Nick collision from services.)
22:01:00 Join Kohlrabi [0] (
22:06:39 Join linuxstb_ [0] (
22:15:05 Quit Lear ("Chatzilla [Firefox 1.5/undefined]")
22:15:06 Quit NicoFR ()
22:15:08 Join ehntoo [0] (
22:16:46_FireFly_*g* i have just re-created my wps-sb-tag patch upon my wps-widget :)
22:21:56*amiconn finally went ahead and got a gcc bugzilla account
22:23:28 Quit linuxstb (Read error: 110 (Connection timed out))
22:23:31markunWhere can I bump the settings version?
22:24:24_FireFly_it's #define CONFIG_BLOCK_VERSION 31
22:24:50markun_FireFly_: thanks, was looking in settings.h
22:25:09_FireFly_this define is only needed in settings.c
22:25:24 Join LedFloyd [0] (
22:30:08preglowmarkun: nice work
22:30:39markunpreglow: the unicode?
22:31:05 Join davidd [0] (
22:31:09markunI just made a crossfeed setting. Should I commit it?
22:32:06Moosplease do :)
22:32:07 Quit ehntoo ("Leaving")
22:32:09LinusNdo it
22:32:14markunok ok :)
22:32:34markunI had to bump the settings version.. sorry.
22:34:44markunI wanted to put it in the same #ifdef SWCODEC with the rest of the dsp settings
22:34:53markunbut I could place it at the end of course.
22:35:02Jungti1234kekeke... :P
22:35:31markunJungti1234: riiiight.. :)
22:35:37LinusNmarkun: it doesn't really matter
22:36:28markunMaybe I should get rid of the warning I get: dsp.c:499: warning: operation on `index' may be undefined
22:36:37LinusNof course
22:36:40Jungti1234That is imitation.
22:37:07markunthe operation is ++index, where index is a long
22:37:23markundon't know why it complains about it.
22:37:26XavierGrJungti1234: Make a USB OTG screens or a tree view of rockbox
22:37:31XavierGrmaybe some will fell for it
22:37:41LinusNmarkun: it depends on the context
22:37:53preglowindex = ++index % 13
22:37:56linuxstb_LinusN: I'm not about the best solution for the first ipod error - caused by RGB_PACK involving an (inlined) function call to swap16.
22:38:12LinusNpreglow: that *is* undefined
22:38:35preglowhow come?
22:38:47Bagderthat's just how ANSI C is defined
22:38:58preglowyeah, but why is it?
22:39:19LinusNit is undefined when the index is incremented
22:39:52Jungti1234XavierGr: Good :D
22:40:04markunso I should change it to index++; index %= 13; or something?
22:40:10preglowwhy isn't index just incremented, the moded?
22:40:14LinusNyou have to updates of the "index" variable in the same expression
22:40:33_FireFly_markun or index = (index+1) % 13
22:40:34LinusNindex = (index +1) % 13;
22:40:50*preglow just realised he doesn't know c at all
22:41:02markunyes, a the ++ was a bit useless :)
22:41:12LinusNlinuxstb_: error: initializer element is not constant ???
22:41:51linuxstb_Yes - RGB_PACK resolves to a function call.
22:41:53LinusNjust make it an ifdef then
22:42:13LinusNor rather, put the assignment in lcd_init()
22:42:26linuxstb_That's what I was thinking - the second option.
22:42:40linuxstb_The other errors are minor. I'll commit the fixes now.
22:42:58LinusNhave you tried bmp2rb?
22:43:44LinusNconvert it to bmp and then use bmp2rb
22:44:06LinusNthen you can create a new file: apps/recorder/logo-ipod.c
22:44:12LinusNor something
22:44:28linuxstb_No, not yet - I'll probably try it later.
22:45:04*preglow just remebered arm instructions can be conditional
22:45:31Bagdernow that is kind of essential in arm asm
22:46:11preglowespecially when you just need to skip branch over two instructions anyway
22:46:22preglowyou save one instruction, and probably one cycle anyway
22:46:52linuxstb_LinusN: Have you changed the shade of "rockbox blue" ?
22:47:12*Bagder detects big money on his account just arrived... sweet funding
22:47:57Jungti1234Very cold..
22:48:07preglowi'd like some funding too, please!
22:48:43LinusNlinuxstb_: i took it from the web page
22:49:09preglowBagder: god bloody damn
22:49:18preglowthat is a nice sum
22:49:19LinusNthe lighter shade
22:49:45linuxstb_I think I must have taken the darker shade. I prefer your version, but it means one of my logos has the wrong background.
22:50:00linuxstb_Which brings us to the issue of bitmaps with transparent bits.
22:50:12preglowuse some absurd shade of pink
22:50:15LinusNyeah, that will be nice
22:50:20*amiconn just filed a gcc bug report
22:50:32LinusNamiconn: weak aliases?
22:50:33preglowamiconn: file your mul routines while you're at it
22:50:41amiconnLinusN: yeps
22:50:57amiconnOr rather, all aliases to inline asm symbols
22:51:19linuxstb_Anyway, I've got to go now, so I'll just commit the ipod fixes and worry about shades of blue some other time.
22:53:11Jungti1234I go to the school. bye.
22:53:28 Quit Jungti1234 ("Bye Bye~")
22:54:34 Join Sandking [0] (
22:54:44 Quit davidd ()
22:57:23markuncrossfeed comitted. Now I just have to wait for the red and yellow :(
22:58:07_FireFly_you have to wait a bit longer a build is currently in progress
22:58:26Bagderwe need a faster build server! ;-)
22:58:52***Saving seen data "./dancer.seen"
22:59:45preglowit's a dual cpu 900mhz something, yes?
23:00:06amiconnmarkun: Imho you should check for build problems before commit. I do...
23:00:08Bagderdual amd 1600mhz
23:00:46LinusNBagder: i believe it's the disk, not the cpu
23:00:49markunamiconn: I did, but using FreeBSD I always have some patches applied to get rockbox to compile
23:00:57BagderLinusN: I think you're right
23:01:20LinusNstill crawler DOS'ing?
23:01:46BagderI don't think so
23:02:21chopped_porkwoo crossfade
23:02:41DBUGEnqueued KICK chopped_pork
23:03:08ep0ch|markun: is crossfeed more effecient than the previous crossfeed patch?
23:03:28ep0ch|excellent :)
23:03:53preglowand i'll see if it's easy to optimise approximately now
23:03:56Bagderit sure is nice to see the h300 build gone yellow
23:05:10LedFloydum, i don't know if I'm allowed to speak in here, but what exactly does "yellow" mean? some sort of status?
23:05:48chopped_porkmarkun: any estimate on when unicode will be in cvs?
23:05:59markunchopped_pork: no, don't ask me :)
23:06:00BagderLedFloyd: check the cvs build table and you'll see
23:06:03chopped_porklooks lovely as already noted :)
23:06:37markunThe problem is that one of the archos rockbox versions gets too big with unicode
23:07:18markunOne sollution is the optimize till it fits, the other is to compress the rockbox image and extract it in memory
23:09:16amiconnIt probably (hopefully?) will fit when TiMiD finishes multiscreen support and some old functions can be removed
23:09:36amiconnCurrently a number of functions are included twice, like splash()
23:09:53chopped_porkis your work available as a patch somewhere? it's so cool it might finally make me setup a crosscompiling enviroment so i can have my own rockbox compile :>
23:10:10amiconnAfair unicode did fit before the multiscreen work
23:10:23markunyes, it did.
23:10:36_FireFly_amiconn: a quick solution for splash would be to replace each left splash call with gui_syncsplash
23:11:04chopped_porkthanks markun
23:11:06_FireFly_and i think the same could be done for the two functions in widget.c
23:11:07markunThe summary is outdated, but cannot be changed
23:11:12amiconnIt's not only splash(), it's splash(), statusbar, button bar and probably others
23:11:32chopped_porkis there any guide available to making a build environment in linux? (running gentoo here)
23:14:42chopped_porkis it possible to build rockbox (coldfire flavor) with gcc4 or is that a bad idea?
23:14:52Bagderbad idea
23:14:54_FireFly_amiconn: yepp at least in 10 files in which status_draw is called
23:15:08chopped_porkalrighty, thanks again
23:15:36preglowchopped_pork: it works just fine
23:17:25chopped_porkgiven i'll have the time, would it be of any use if i did a .ebuild for gentoo?
23:17:45Bagderchopped_pork: it could make a few people happy, I'd say
23:17:59 Part Kaggen
23:21:12 Join ashridah [0] (
23:21:51_FireFly_markun: it's me or does crossfeed lower the volume a bit??
23:22:30markun_FireFly_: yes, quite a bit
23:23:07_FireFly_and i think it is a also little dull
23:23:19_FireFly_the sound
23:23:20markunYes, I also think so :)
23:23:36markunBut it works very well for some recordings with extreme stereo.
23:23:41 Join actionshrimp [0] (
23:23:45markunbeatles, jimi hendrix..
23:23:56preglowmarkun: you told me about some crossfeed plugin author you were going to mail, did that happen?
23:23:56_FireFly_so there is space for improvements :)
23:24:08markun_FireFly_: feel free to improve it
23:24:25Maxou`what's "crossfeed"?
23:25:21 Join Midgey34 [0] (
23:25:27 Quit Midgey34 (Client Quit)
23:25:58 Join Midgey34 [0] (
23:26:55_FireFly_markun i think the prvious one had an better sound then this one
23:27:31preglowthe previous one didn't differ in sound
23:27:34markun_FireFly_: I didn't change any of the values I think
23:27:53preglowno, everything of importance is the same
23:29:24 Join bitmastro [0] (
23:31:15markunbona sera
23:31:26bitmastroAt first, thanks to linuxn for the screenshot.. I'm still crying
23:32:13bitmastrosecond.. I was thinking... the last popular codec missing is the wma, right?
23:32:31XavierGryes and I don't think that we will see this thing soon
23:32:39preglowwhat screenshot?
23:32:44amiconnpreglow: Regarding the 64 bit mul routines - I have no idea how to contribute them. They obviously don't fix a bug, and I have no idea where they have to be placed in the gcc source, so I can't file a patch either
23:32:54preglowXavierGr: we just might, someone's done a bit of the work for us
23:32:58preglowamiconn: mailinglist, perhaps?
23:32:59bitmastropreglow: the rockbox logo on the h300
23:33:10preglowbitmastro: url?
23:33:13markunhey, the build should be finished by now..
23:33:48_FireFly_markun: ok it was the song on which crossfeed has a bad drawback to the sound
23:34:02preglowmarkun: should, but it's not
23:34:23amiconnLinusN: The H300 baclight is hooked to some kind of pwm output, so it's dimmable in hardware?
23:34:43bitmastroah.. I was saying.. about the wma... anyone looked at the neuros code?
23:34:49Moosmarkun: the prediction isn't very accurate like somwhere ;)
23:34:51markunpreglow: maybe because I changed some .h files that caused a lot to recompile
23:35:03preglowmarkun: doubtless, that'd fool ccache
23:35:23preglowbitmastro: neuros doesn't distribute the wma code
23:35:43bitmastroit seems to me like it's a fixed point implementation
23:35:51preglowor does it...
23:36:03 Join elinenbe [0] (
23:36:16preglowbitmastro: that looks like a dsp interface to me
23:36:42elinenbeLinusN: quick question... will rockbox boot on the H300 now?
23:36:50XavierGrpreglow: who did the work? Is there a major chance for WMA then?
23:37:01amiconnmarkun: Crossfeed wastes a .lang string, imho.
23:37:11preglowXavierGr: someone on the forums posted a partially fixed point converted version of the wma decoder in libffmpeg
23:37:33amiconnIn stead of "Enabled crossfeed" = yes/no in the setting itself it would be better to use "Crossfeed" = On/Off
23:37:48elinenbepreglow: that's awesome!
23:38:02amiconnWasted .lang strings are bad because they increase binary size for all targets...
23:38:05CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
23:38:05*amiconn prods Bagder
23:38:15*Bagder appears
23:38:24Bagderat your service!
23:38:28preglowelinenbe: awesome, yes, but someone still needs to work on it
23:38:33preglowi wont be
23:38:36amiconnBagder: [23:38:26] <amiconn> Wasted .lang strings are bad because they increase binary size for all targets...
23:39:12*Bagder blinks and talks about the wheather
23:39:23MoosBagder: what's about your V2 localisation ? XD
23:39:55Bagderit was just a bit too much work for my little time I spend on rockbox these days
23:40:11Bagderso its still pending
23:40:26Mooshehe :)
23:41:48preglowmarkun: do remember that the highpass filter isn't really good, though
23:41:58Bagdersomeone else is free to take the ball and run for it
23:43:06preglowmarkun: and btw, the comment for apply_crossfeed is on crack
23:43:29ashridahhey, did that multiscreen WPS widget stuff get merged?
23:43:41Bagderred red red red red
23:44:01Bagderand red
23:44:30Midgey34oh man, newest build is crazy
23:44:53thegeek_please elaborate
23:45:00Midgey34the redness I mean
23:45:21Moosisn't critical markun will fix this ;)
23:45:23 Part ep0ch|
23:45:32Midgey34I know :)
23:45:52Midgey34I still excited by linus' progress on the h300
23:45:54 Join ]RowaN[ [0] (
23:46:02]RowaN[guys what is this "crossfeed dsp effect" thing
23:46:17XavierGrok I know that it is a rockbox policy to hold columns in source below 80. But I use a very big text paragraph? How can I break this in multiple lines?
23:46:37Bagdervery big text paragraph?
23:46:41Bagderyou mean comment?
23:46:48XavierGrNo fdprint
23:46:55Bagderyou can write "string"
23:47:01_FireFly_ah got the volume drawback a bit lower :)
23:47:13Bagderand they are concatenated by the compiler
23:48:12XavierGrso you recommend breaking 507 characters in 7 strings?
23:48:20XavierGrit seems very tiresome to me
23:48:21Bagderat least, yes
23:48:54 Join mashalla [0] (
23:49:13markunDon't think this is cause by my commit, is it?
23:49:54Bagderit looks like LinusN's commit is involved
23:50:28Bagderseeing the backlight references
23:52:20Febs]RowaN[, crossfeed feeds a little of the right channel into the left and vice versa.
23:52:44FebsIt simulates the experience you get when you are listening to speakers instead of headphones ...
23:53:01FebsWith speakers, the sound bounces off walls, floors, etc before it gets to you.
23:53:02]RowaN[ah, i'll have to give it a go
23:53:07FebsWith headphones, it doesn't.
23:53:23FebsSo many feel that crossfeed makes headphones sound more natural.
23:53:58ashridah]RowaN[: most music is mixed assuming you're listening with speakers. some music has some instruments completely on one side only. it also has the effect of making what you're listening to sound like it's inside your head. crossfeed should make it sound like it's infront of you too.
23:54:22*Febs catches up on some of the recent discussion.
23:54:41FebsAs someone noted above, it's good for older recordings like early Beatles albums.
23:54:50ashridahokay. why do i get 'No wps module present, can't do the WPS magic!' when i run make zip?
23:55:04ashridahstill seems to create the file
23:55:16ashridahFebs: and jazz music often comes that way too.
23:55:33Bagderashridah: you need to checkout the 'wps' module to get the wps stuff added to the zip
23:55:45Bagderits brand new
23:56:02ashridahBagder: just cvs co wps in the rockbox directory containing all of rockbox's source?
23:56:05Bagderused for the wps bundling script
23:56:10XavierGrBagder: Are you sure there is no other way to cut the text? It is very troublesome and unintuitive, plus if someone wants to change the text it should be very difficult.
23:56:11Bagderashridah: yeps
23:56:51BagderXavierGr: having 500 bytes text in source code is _always_ cumbersome no matter how you do it
23:57:16 Quit mashalla ()
23:57:29Bagderhaving a 500 bytes long line is not exactly comfortable either
23:57:45XavierGrwell turn wordwrap on :p
23:57:57XavierGrok if there is no other way I will split it.
23:58:10markunpreglow: ACC etc are not declared for non coldfire platforms (ipod)
23:58:53XavierGrBagder: Do you think a 2 dimensional array is better or stay with string1[80], string2[80] e.t.c?

Previous day | Next day