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 2005-07-18

00:00:05amiconnYou'll need 3 (constant) mask works, a bit of ANDing shifting and XORing
00:00:16stripwaxmemmem - rom?! you'd store a fixed 256 byte table in the firmware image?!
00:00:18LearYou know, the archos players and recorders all bitswap the MP3 data before feeding them to the MAS...
00:00:36stripwaxLear - wow, really?
00:00:53amiconnLear: Yes, and I still have my 'geek' bitswap routine that is almost the same speed as table based swapping on SH1
00:01:19amiconnThe approach would very likely work faster than table-based on coldfire
00:01:44memmemAnyway, we have 3 bytes per pixel and the orientation of the display does not matter at all.
00:01:55memmem...for the X5.
00:01:59amiconn...and it's only slightly slower on SH1, because our table-based swap is heavily optimised too
00:02:46amiconnYes, with >= 1 byte/pixel it's not difficult to flip in software
00:03:10amiconn...yet still easier to do it in hardware if the hardware provides that function
00:03:23memmemWe just add -1, 1, -480, or 480 per pixel to get to the next one.
00:03:37stripwaxthat would be so slow, per pixel, no?
00:03:42memmems/1/3/
00:04:37memmemNo. The original firmware seems to have been compiled either by a bad compiler or without any optimization. Looks really bad.
00:05:55memmem(I haven't yet looked at m68k code generated by GCC with -O3, though.)
00:07:44 Quit muesli- (Read error: 60 (Operation timed out))
00:09:22 Quit matsl (Read error: 104 (Connection reset by peer))
00:10:50amiconnRockbox uses -O only, for several reasons
00:11:02amiconnThe codecs and some plugins use -O2
00:11:10LearLike Tremor. :)
00:13:14amiconnI would personally stay far away from -O3 with non-x86 gcc after I've seen what kind of buggy code that produces sometimes
00:13:25amiconnEven -O2 is sometimes problematic
00:13:59stripwaxamiconn - that sucks, are gcc devs aware?
00:14:09amiconnI dunno.
00:14:12stripwaxhrmm
00:14:51amiconnsh-elf-gcc produces really obviously wrong code with -O2 / -Os for special cases
00:15:12amiconnExample: There is an interrupt handler for the MAS interrupt.
00:16:15amiconn-O2 / -Os enables jump optimisation, meaning for a function that unconditionally calls another function at the end, gcc will replace the jsr (sub-function) - rts sequence with a jmp
00:16:38amiconn...which is obviously very wrong for an interrupt handler because of 2 reasons
00:17:13amiconn(1) A usual function only saves the non-scratch registers, while an interrupt handler has to save all regs
00:17:33amiconn(2) irq handler has to return with rte, not rts...
00:17:46amiconnYet sh-elf-gcc does this silly thing...
00:18:08memmemNo, that's not GCC's fault. You just cannot write interrupt handlers in C.
00:18:15stripwaxamiconn - blimey - so it will even replace jsr/rtE with jmp?!
00:18:20memmemUse a wrapper written in assembly.
00:18:34LearWhat about the interrupt attribute then?
00:18:57memmemIs that really documented as being supported for that architecture?
00:19:07amiconnmemmem: It *is* clearly gcc's fault
00:19:12LearIf you tell GCC that it is an interrupt handler, it should do the right thing, no matter the -O level.
00:19:30amiconnThe handler is properly marked __attribute__ ((interrupt_handler))
00:19:40 Quit spiralout (Read error: 104 (Connection reset by peer))
00:20:01stripwaxmemmem - if gcc is smart enough to generate the correct return instruction, it should be smart enough to not try and take it away again later!
00:20:08amiconnstripwax: It does this with -O2 -O3 and -Os, not with -O or -O0
00:20:28stripwaxamiconn yes, I'd gathered. it's dumb but workaroundable
00:20:50amiconnThis is about the only bug in gcc 3.3.x that prevents using -O2 or similar on archos
00:21:05amiconngcc 3.4.x does more silly things I was not yet able to find
00:21:05memmem"Use this attribute on the ARM, AVR, M32R/D and Xstormy16 ports to indicate that the specified function is an interrupt handler."
00:21:05memmem
00:21:51amiconn...apart from that using -O2 makes the code larger, and we're already short on space on archos...
00:22:07amiconn...and gcc 3.4.x generates larger binaries than 3.3.x
00:23:11memmem(That was for GCC 3.3.x.) For GCC 3.4.4 that section reads "Use this attribute on the ARM, AVR, C4x, M32R/D and Xstormy16 ports to indicate that the specified function is an interrupt handler." SH is not mentioned.
00:23:11memmem
00:23:14stripwaxamiconn - are daily builds 3.3.x or 3.4.x?
00:23:40amiconn3.3.x for archos, 3.4.x for iriver
00:23:44 Join ]RowaN[ [0] (a2b0y@82-43-212-52.cable.ubr10.newm.blueyonder.co.uk)
00:23:51stripwaxok
00:23:56amiconnmemmem: You read about interrupt, not interrupt_handler
00:24:08amiconninterrupt_handler
00:24:08amiconn Use this attribute on the m68k, H8/300, H8/300H, H8S, and SH to indicate that the specified function is an interrupt handler.
00:24:17amiconnIt clearly says SH...
00:24:17memmemTrue.
00:24:42memmemThen it's a bug indeed.
00:24:52amiconnThere is a workaround we could use... it works...
00:25:25amiconnJust introduce an empty volatile asm statement between the final function call and the closing brace of the function
00:25:29]RowaN[well, my 30gig iriver hd comes 2moro.. woo hoo!
00:25:34amiconnasm volatile ( "");
00:25:41*]RowaN[ preys for dhl to find me
00:25:52stripwaxamiconn - hehe, neat. and gcc doesn't optimise that out? haha
00:26:15amiconnIt can't... because it's marked volatile, and gcc doesn't look inside
00:26:39amiconn...because gcc doesn't know assembler mnemonics as input
00:26:39stripwaxamiconn oh, of course
00:26:52memmemIt might be better to write a wrapper in assembly and call the C code from there. Should also solve the problem about non-saved registers.
00:27:08amiconnI disagree
00:27:25stripwaxmemmem doesn't interrupt_handler also generate the code to save registers? (I'd be surprised..)
00:27:25amiconnGcc allows to write interrupt handlers in C, so we should use that
00:27:36memmemApparently, it doesn't.
00:27:43amiconnAssembler wrappers would only make it harder to read & follow the code
00:28:05amiconnIt does generate the code to save registers
00:28:14amiconnOtherwise it wouldn't make sense...
00:28:53amiconnIt's the jsr -> jmp optimisation enabled with -O2 and friends that has a bug
00:28:58memmemOK, I misunderstood you above.
00:29:33memmem-fno-optimize-sibling-calls
00:30:09amiconnYes, that'd be another workaround... but usually this optimisation is a good idea
00:30:41amiconnIt's just that it must not be done for interrupt handlers
00:31:15memmemI just coded the copy-RAM-to-LCD function in C and compiled it with GCC; even with -O the code is half the size compared to the code of the original iAUDIO firmware. So, speed doesn't seem to be an issue.
00:32:04amiconnPerhaps I should report that to the gcc team. The only thing stopping me from doing that so far is that they require a registration...
00:34:03 Join matsl [0] (~matsl@1-1-4-2a.mal.sth.bostream.se)
00:36:16stripwaxmemmem - heh, maybe the iAUDIO fw was very badly hand-optimised asm .. :-)
00:36:43*stripwax used to work at a company that had historically lots of "hand-optimised asm" that was, frankly, appalling.
00:36:57memmemI think they use hand-written assembly code for the codecs; lots of EMAC instructions.
00:37:09stripwaxthat's quite likely
00:38:10stripwaxhm, has anyone taken a look thru the H3xx fw to determine how to address the LCD? I know that's not really the rockbox way of doing things, but ..
00:38:31stripwaxactually do we have a bootloader yet?
00:40:25 Quit Lear ("Chatzilla 0.9.68.5 [Firefox 1.0.5/20050715]")
00:40:52stripwaxSlasheri - did you consider trying to increase the frequency of DMA0 interrupts so that size was always <= 512?
00:41:29stripwaxoutta here, gnight
00:41:37 Quit yngwi ("Chatzilla 0.9.68.5 [Firefox 1.0.4/20050511]")
00:41:48stripwax(malmsteen?)
00:41:51 Part stripwax
00:54:36memmemWhere's the iriver bootloader code in CVS?
01:00
01:00:23 Quit thegeek (Read error: 110 (Connection timed out))
01:05:59 Quit memmem ("ERC Version 5.0.3 $Revision: 1.726.2.17 $ (IRC client for Emacs)")
01:08:23 Join thegeek [0] (na@ti521110a080-2449.bb.online.no)
01:09:43***Saving seen data "./dancer.seen"
01:25:13HClbootloader
01:25:16HClyou have to check it out seperately
01:25:41Mooshe went out :)
01:26:46 Quit cYmen_ ("zZz")
01:27:43 Join memmem [0] (~user@p54A23260.dip0.t-ipconnect.de)
01:28:13memmemHCl: Thanks! I haven't found that fact documented anywhere.
01:46:20 Quit matsl (Remote closed the connection)
01:46:27 Quit memmem ("ERC Version 5.0.3 $Revision: 1.726.2.17 $ (IRC client for Emacs)")
01:52:38 Quit CheeseBurgerMan (Read error: 110 (Connection timed out))
01:56:05 Join CheeseBurgerMan [0] (~youshould@tc2-225-085.altelco.net)
02:00
02:06:12 Quit ]RowaN[ ()
02:30:27 Quit Sucka ("a bird in the bush is worth two in your house")
02:42:01 Quit Moos (" Want to be different? HydraIRC -> http://www.hydrairc.com <-")
02:46:14 Quit gromit` (Remote closed the connection)
03:00
03:09:45***Saving seen data "./dancer.seen"
03:43:22 Quit hicks_ (Remote closed the connection)
03:50:19 Join webguest65 [0] (~18c75b95@labb.contactor.se)
03:51:27 Part webguest65
03:51:33 Join webguest65 [0] (~18c75b95@labb.contactor.se)
03:57:07 Quit Asku (Remote closed the connection)
03:57:45 Quit webguest65 ("CGI:IRC (EOF)")
04:00
04:05:57 Join QT [0] (as@area51.users.madwifi)
04:17:23 Quit QT_ (Read error: 113 (No route to host))
05:00
05:06:34 Join amiconn_ [0] (~jens@p54BD66E3.dip.t-dialin.net)
05:09:46***Saving seen data "./dancer.seen"
05:15:06 Quit amiconn (Read error: 60 (Operation timed out))
05:15:07 Nick amiconn_ is now known as amiconn (~jens@p54BD66E3.dip.t-dialin.net)
05:44:25 Quit CheeseBurgerMan ()
05:44:49 Join CheeseBurgerMan [0] (~Dan@tc2-225-085.altelco.net)
05:56:13 Part CheeseBurgerMan
05:58:35 Join CBM [0] (~youshould@tc2-225-085.altelco.net)
05:59:06 Nick CBM is now known as CheeseBurgerMan (~youshould@tc2-225-085.altelco.net)
06:00
06:19:36 Join jwalk [0] (~jwalk_edm@67.71.173.144)
06:19:45 Quit jwalk (Client Quit)
06:30:32 Join Coldtoast [0] (~edan@ppp111-3.lns1.hba1.internode.on.net)
06:40:42 Nick CheeseBurgerMan is now known as CBM-away (~youshould@tc2-225-085.altelco.net)
06:42:39 Join Strath [0] (~mike@dgvlwinas01pool0-a218.wi.tds.net)
06:43:55 Quit Febs_away ("Chatzilla 0.9.68.5 [Firefox 1.0.5/20050711]")
07:00
07:09:29 Join oxygen77 [0] (~c1c28427@labb.contactor.se)
07:09:47***Saving seen data "./dancer.seen"
07:32:43 Join LinusN [0] (~linus@labb.contactor.se)
08:00
08:04:03amiconnMorning
08:06:45LinusNmorning
08:06:59LinusNi hooked up my la yesterday
08:07:39amiconnI've guessed that from your commits :)
08:08:30LinusN:-)
08:08:37amiconnDo we stay within specs for the LCD at 120 MHz?
08:09:04LinusNbtw, we are running the spi interface to the button adc too fast
08:09:11LinusNamiconn: yes
08:09:40amiconnOops. I think this should be corrected then.
08:10:08amiconnSpeaking of spi speed - did you also measure the remote LCD spi?
08:10:10LinusNbut we might consider adding one wait state anyway, so we can get rid of the extra NOP in the lcd loop
08:10:37LinusNthat way it will be faster in 48mhz
08:11:01LinusNno, i haven't yet measured the remote spi
08:11:37amiconnOkay.
08:11:48LinusNor have different loops for different frequencies
08:12:06amiconn1 ms per lcd_update() is nice - that means the grayscale lib will consume less than 10% cpu for fullscreen
08:12:21LinusNyup
08:12:24LinusNnice indeed
08:12:35amiconn(but still requires to boost for constant timer period :-/ )
08:13:03amiconnDid you read about our findings for the H3x0 (and X5) LCDs?
08:13:14LinusNnot in detail
08:13:18LinusNbrief me
08:13:42amiconnController not yet identified, but we're fairly sure about the module and panel manufacturer
08:13:55MrStaticVoidits a HD66789R in the H300 ;-)
08:14:10amiconnSomeone who works with a company dealing with embedded stuff could mail the module manuf.
08:14:12LinusNMrStaticVoid: and you know that for sure?
08:14:47amiconnModule manuf. should be Varitronix (Hong Kong), panel manufacturer IZ Display (Korea)
08:15:00MrStaticVoidhttp://www.m-display.com.tw/product/pdf/T200P00100.pdf
08:15:05amiconnBoth displays are very similar
08:15:15MrStaticVoidthere is one implementation of the varitronix lcd which notes the controller
08:17:00amiconnLinusN: IZ Display is derived from several facts: (1) The markings start with COG-IZ (2) There are web pages stating Varitronix and IZ took in KES part together (3) IZ is korean like iriver and cowon
08:17:15LinusNlooks promising
08:17:17amiconnThe smeared logo is the Varitronix logo almost for sure
08:17:36LinusNno wiki update?
08:18:02amiconnOops, should've done that....
08:19:02amiconnMrStaticVoid: I think it is likely that the X5 lcd also uses a HD... controller then
08:19:12MrStaticVoidprobably not
08:19:18amiconnNo?
08:19:21MrStaticVoidhold on
08:21:05amiconnLinusN: H3x0: COG-IZ12049-02 X5: COG-IZ12066-01
08:21:41amiconn..and the X5 lcd module shows a big clear 'VL' in addition to the smeared logo on the sticker
08:22:04MrStaticVoidwell it might be
08:22:12MrStaticVoidid go by this regarding the x5: http://www.varitronix.com/catalog/LCD/images/TFT3013-01.jpg
08:22:19MrStaticVoidits awfully hard to read
08:24:04MrStaticVoidor better yet javascript:void('0');
08:24:06MrStaticVoidgrr
08:24:36MrStaticVoidhttp://www.mouser.com/?handler=data.listcategory&Ne=400&terms=varitronix&Ntt=*varitronix*&crc=true&Ns=SField&N=421
08:24:53MrStaticVoidscroll down to the 160x128 module and click datasheet
08:25:05MrStaticVoidthere is a much clearer diagram of what i suspect is in the x5
08:26:05MrStaticVoidor maybe not, what do i know
08:35:31 Join XavierGr [0] (~XavierGr@ppp49-adsl-96.ath.forthnet.gr)
08:35:56XavierGrGood Morning!
08:40:26 Nick Strath is now known as StrathAFK (~mike@dgvlwinas01pool0-a218.wi.tds.net)
08:44:04 Quit oxygen77 ("CGI:IRC")
09:00
09:06:23 Join Harpy [0] (6hVBcU75FQ@dsl-hkigw7wbb.dial.inet.fi)
09:09:50***Saving seen data "./dancer.seen"
09:16:32 Nick Lynx_awy is now known as Lynx_ (~lynx@tina-10-4.genetik.uni-koeln.de)
09:17:38MrStaticVoidw00t
09:17:44MrStaticVoidi got an email from varitronix
09:22:09MrStaticVoidthey sent me a datasheet for their 2" 176x220 18bit module
09:22:12 Quit XavierGr ()
09:22:18MrStaticVoidanywhere i can upload it?
09:23:59 Join gromit` [0] (~gromit`@ras75-5-82-234-244-69.fbx.proxad.net)
09:31:35LinusNMrStaticVoid: the wiki
09:32:07MrStaticVoidah i see
09:33:56MrStaticVoidthat was painless, its up if you care to take a look
09:35:03 Join matsl [0] (~matsl@1-1-4-2a.mal.sth.bostream.se)
09:42:21 Join ashridah [0] (ashridah@220-253-121-116.VIC.netspace.net.au)
09:43:25 Join webguest32 [0] (~4509f7eb@labb.contactor.se)
09:50:36 Quit webguest32 ("CGI:IRC (EOF)")
09:50:36 Join web1 [0] (~4509f7eb@labb.contactor.se)
10:00
10:18:52*LinusN just got an email from micronas with updated data sheets for mas3539f
10:19:24LinusNnice of them
10:19:31Coldtoastvery
10:19:48Rickwhat's mas3539f?
10:19:56LinusNalong with a rockbox bug report :-)
10:20:10amiconnLinusN: Hmm?
10:20:11LinusNRick: the mp3 decoder chip in the archos ondio
10:20:26amiconnOndio SP to be precise
10:20:28LinusNyes
10:21:26Rickah
10:22:20 Quit Rick (Read error: 104 (Connection reset by peer))
10:22:32*amiconn is very curious what the updated datasheet says, and what the bug report is
10:22:40 Join pike| [0] (pike@c83-249-120-126.bredband.comhem.se)
10:22:54 Join Rick [0] (rick@pool-71-108-23-179.lsanca.dsl-w.verizon.net)
10:22:54LinusNthe updated data sheet is in the wiki
10:39:46 Quit pike (Read error: 110 (Connection timed out))
11:00
11:01:15 Join hicks [0] (~hicks@zeus.mups.co.uk)
11:04:02 Join cYmen [0] (~cymen@nat-ph3-wh.rz.uni-karlsruhe.de)
11:04:42 Quit matsl (Remote closed the connection)
11:07:45 Join preglow [0] (~c39fb93d@labb.contactor.se)
11:09:43 Quit preglow (Client Quit)
11:09:54***Saving seen data "./dancer.seen"
11:11:11 Join prethom [0] (~c39fb93d@labb.contactor.se)
11:12:42prethomwhat up
11:12:58LinusNnot much :-)
11:13:42prethomyou guys have been having code for breakfast, lunch and dinner the past week, it seems
11:13:59LinusNyup
11:14:30*LinusN is writing a biphase-mark decoder in perl
11:14:57LinusNmy brain hurts :-)
11:16:29prethomhaha, i dont even remember how it works
11:26:57prethomdoes rockbox work on h110 now, btw? saw you commited a bunch of changes
11:27:25LinusNyes it does
11:29:25prethomnice
11:29:49prethomreading 500 emails on a 28k8 is extremely boring, btw
11:36:40 Join memmem [0] (~user@p54A2213E.dip0.t-ipconnect.de)
11:38:08memmemHi MrStaticVoid, the HD66773R datasheet matches the firmware code! Thanks!
11:38:43memmemStrange device, though: "The hardware-dither circuit converts 18-bit one-pixel data to 16-bit data with hardware-dither conversion."
11:39:10MrStaticVoidgood news
11:39:41MrStaticVoidhopefully its not too strange
11:40:04memmemOn the interface side, it's not strange. But the graphics RAM has only 16 bits per pixel...
11:40:18MrStaticVoidoh
11:44:38 Part prethom
11:45:02 Join prethom [0] (~c39fb93d@labb.contactor.se)
11:45:12prethomyou're talking about h3x0 display here?
11:45:54MrStaticVoidwell memmem was talking about the x5
11:46:14MrStaticVoidbut it and the h300's display are closely related
11:46:26prethomahh
11:46:41MrStaticVoidthe manufacturer (varitronix) sent me the datasheets for both this morning
11:47:24prethomahh, so you've identified both now?
11:47:26memmemBad news: The firmware uses some registers not documented in the HD66773R datasheet, so we're close but not yet there.
11:47:34prethom;/
11:47:37MrStaticVoid:-(
11:47:42 Quit pill (Read error: 60 (Operation timed out))
11:50:41 Quit prethom ("CGI:IRC")
11:52:19MrStaticVoidwell i only requested information about the modules they listed on their site
11:52:49 Quit CoCoLUS (Read error: 110 (Connection timed out))
11:52:58MrStaticVoidmaybe it would be worth asking about the specific model found on the label
11:55:30 Join Zagor [0] (foobar@h63n1fls31o265.telia.com)
11:56:55 Join ]RowaN[ [0] (a2b0y@62.128.222.176)
11:57:10]RowaN[yay just installed new 30gig hd in my iriver h120 =]
11:58:47Coldtoastnice!
11:58:51]RowaN[is it normal for rockbox to report it as being 27.9gig?
11:58:55]RowaN[i didnt format it or anything
11:59:02Coldtoastit IS 27.9GB
11:59:12]RowaN[alrightey then
11:59:20]RowaN[i wonder if its worth ebaying my damaged 20gig drive
11:59:29Coldtoastthere's a couple of factors involved with GB ratings
11:59:39Coldtoastone of them being "What's teh definition of 1GB?"
11:59:40memmemMrStaticVoid: I think I found documentation of those missing registers. (Reading the TOC isn't enough ;-)
11:59:51Nibblerit depends on if "kilo" is 1000 or 1024
11:59:57MrStaticVoidheh
12:00
12:00:00MrStaticVoidthats good
12:00:30memmemI'll verify all the registers later.
12:00:31 Join CoCoLUS [0] (~coco@h081217139221.dyn.cm.kabsi.at)
12:00:49Coldtoastsee, 1KBcan be interpreted as 1000KB or 1024KB and then 1GB could be interpreted as 1000MB or 1024MB
12:00:49MrStaticVoidim curious, how are you verifing em?
12:01:02MrStaticVoida hex editor on the original firmware?
12:01:22memmemDisassembly of the original fw.
12:01:26Coldtoastmost companies go by 1000 tho these days it seems
12:01:46ZagorColdtoast: most harddisk manufacturers, yes. most other companies, no.
12:01:50MrStaticVoidoh, maybe ill have to look into that. always want to try something new
12:01:59ColdtoastI'm talkign abou tHDD manufactureres, yeah
12:02:10Coldtoastdamn! that was some bad typing
12:02:49Coldtoastbut anyway. you're not missing 2.1GB ]RowaN[
12:03:36ColdtoastI have supposedly 320GB of RAID that shows up as 300GB
12:06:19*memmem goes dining
12:06:24 Quit memmem ("ERC Version 5.0.3 $Revision: 1.726.2.17 $ (IRC client for Emacs)")
12:06:58ep0ch_k = 1000, K = 1024
12:07:17ep0ch_i thought
12:07:26Coldtoastdunno about that
12:07:45Slasheriplease, no Kelvins ;)
12:07:52ep0ch_heh
12:08:21ep0ch_so what's Slasheri working on today? :)
12:08:41Slasheriep0ch_: just now designing a pcb with pads at work.. =)
12:08:57 Join zeekoe [0] (~zeekoe@ip51cc69f6.adsl-surfen.hetnet.nl)
12:09:10zeekoeKi = 1024.. :)
12:09:16ep0ch_nce, what's it going to do?
12:09:26Coldtoastthink I'll delete every single 128Kb off of my machine
12:09:30zeekoehttp://en.wikipedia.org/wiki/Kibibyte
12:09:48Slasheriep0ch_: it's "a universal bluetooth controller unit".. :)
12:10:20Slasheri50 MHz fpga with nios and ~30 user i/o pins
12:10:22Coldtoasthttp://en.wikipedia.org/wiki/Kilobyte
12:10:26zeekoeSlasheri, can you write a driver for rockbox?
12:10:34zeekoethat would be 1337 :P
12:10:40Coldtoast"A kilobyte (derived from the SI prefix kilo-) is a unit of information or computer storage equal to either 1024 or 1000 bytes"
12:10:43Slasherihmm :D
12:11:06Slasheribut currently i don't have much time to do anything because i have to leave soon to the mil service :/
12:11:16ep0ch_which country?
12:11:19Slasherifinland..
12:11:44Coldtoastit's weird cos you really don't hear of Finland getting invilved in many wars
12:11:46ep0ch_they should do that over in the uk
12:11:54Coldtoastyet you still have compulsory mil service
12:12:10ep0ch_who knows what they're planning ;)
12:12:19crwlSlasheri, take a laptop with you and don't spend much time at Hotelli Hanhi ;)
12:12:42Slashericrwl: yes :(
12:12:46zeekoein holland it hasn't been compulsory since about 5 years i guess
12:12:47ep0ch_you looking forward to it Slasheri?
12:12:55Slashericrwl: hehe, nice idea .)
12:13:18zeekoehave2go
12:13:23 Nick zeekoe is now known as zeekoe|afk (~zeekoe@ip51cc69f6.adsl-surfen.hetnet.nl)
12:13:28 Nick ep0ch_ is now known as ep0ch (~ep0ch@81-6-218-35.gotadsl.co.uk)
12:14:08Slasheriep0ch: i don't have much choise.. but fortunately it's only about three weeks away from everything :)
12:14:22ColdtoastI hope they incorporate the "Rendering Engine" feature Netscape has into Firefox
12:14:28 Part Nibbler ("Leaving")
12:14:47ep0chColdtoast: what's that? the choice between IE or Gecko?
12:15:02Coldtoastyeah
12:15:06ep0chbluuuuurrrrrr
12:15:10crwlwhy would anyone want to use the IE rendering engine? :P
12:15:20Coldtoastso no need to fire up IE as it'll use th eIE rendering engine
12:15:24ep0chso we can have activex in firefox wahayy ;)
12:15:29Coldtoastcrwl: oh, let's see... Windows Update for one?
12:15:45Coldtoastyou HAVE to use IE for Windows Update
12:15:49crwlwhy would you want to update your windows via firefox?
12:15:53crwlyou can't get of rid IE anyway
12:16:01Coldtoastwhat?
12:16:14ep0chwell, if you dont want to use windows update, dont use windows
12:16:40ep0chmind you updating windows is easier than updating linux or bsd...
12:16:44Coldtoastyou can't see the value in going to webpages that NEED IE and flipping to the IE engine from your browser instead of firing up actual IE?
12:17:01ColdtoastI can
12:17:05crwlwell, in fact i can't
12:17:11Coldtoastell I can
12:17:35crwlwebpages that absolutely need IE (and i can't think of any right now, maybe i only go to the right pages) can do that
12:17:35ep0chthere is a plugin for firefox that opens up the current page in IE
12:17:37Coldtoastif you're fine with firing up IE tho, cool
12:17:49Coldtoastcrwl: I thought I just told you one?
12:17:56ColdtoastWindows Update NEEDS ie
12:18:06zeekoe|afkwindows update has this autoupdate thingy
12:18:08crwlColdtoast, what about that then? use automatic updates so you don't need to think about it :P
12:18:18Coldtoastautomatic updates? heh
12:18:22zeekoe|afk...which complains every half an hour that i need to reboot my computer :-<
12:18:32Coldtoastone of the first things I disable is Auto Update
12:18:59zeekoe|afkwhy?
12:19:04 Nick zeekoe|afk is now known as zeekoe (~zeekoe@ip51cc69f6.adsl-surfen.hetnet.nl)
12:19:13ColdtoastI'd rather run Windows update manually so I can shoose what not to install
12:19:15crwlbecause he likes to use IE for updating his windows, i think :)
12:19:17ep0chthe worst one is Security Center service, wtf is that all about!!?
12:19:25Coldtoastchoose
12:19:39zeekoewith autodownload&notify you can choose what to install, too
12:19:42Coldtoastcan you customise Automatic updates? No you can't
12:19:49Coldtoastok
12:19:51zeekoeyes you can
12:19:54zeekoe:
12:19:54ep0chyeah you can
12:19:56zeekoe)
12:20:04ep0chi have mine to just tell me when updates are available
12:20:12ColdtoastI disable it
12:20:32ep0chthen you need IE then lol
12:21:05ColdtoastI only want my machine talking to M$ when I tell it to. heh
12:21:22crwlmine never does that, what do i win? :P
12:21:25ep0chyou have the clock sync disabled too?
12:21:40Coldtoastnope. I use an Aussie ntp server tho
12:21:44Coldtoastyou can add your own servers
12:21:48ep0chyou can?
12:21:52Coldtoastyup
12:21:55zeekoehow do you know your machine does only talk to M$ when you tell it to?
12:22:00ep0chhidden somewhere in the registry i take it?
12:22:04Coldtoastfirewall
12:22:18zeekoeok
12:22:48zeekoeep0ch: doubleclick on clock, click internet time, type in timeserver
12:23:05zeekoeeasy as that
12:23:09ep0chheh
12:23:21ep0chthat's just too funny
12:23:26Coldtoastdo any of you Windows users actually use the Win inbuilt firewall?
12:23:33zeekoeyep
12:23:36ep0chyeah
12:23:39Coldtoastreally?
12:23:41zeekoenot that it blocks anything though
12:23:52zeekoei tried to install sygate, but it didn't work
12:23:52Coldtoastit's just a simple packet scheduler
12:23:59zeekoepacket scheduler?
12:24:01ep0chi'm behind a firewalled nat router so i dont really care too much
12:24:04Coldtoastwell, filter really
12:24:14zeekoeso what does it filter?
12:24:21Coldtoastyeah. I mostly don't bother with a firewall to be honest
12:24:39Coldtoastit just filters TCP packets I think
12:24:47ep0chbut i have the windows one on just incase someone sneaks onto my wireless
12:25:12Coldtoastyou don't MAC filter and use any encryption?
12:25:17ep0chi do
12:25:32Coldtoastwe don';t really have that sort of prob here
12:25:36ep0chbut still hackable by someone that knows what they're doing
12:25:50Coldtoastwireless is nowhere NEAR as widespread as in the US where I live
12:26:17ColdtoastI think I'm the only one running a wifi network on my street, for example
12:26:35Coldtoasttook my PDA out and had a bit of a look a couple of months ago
12:26:37ep0chi can access my neighbours from here...
12:26:46ColdtoastI want a Zaurus tho I think. So I can run Kismet
12:27:11Coldtoastthere doesn't seem to be a passive sniffer at all for Windows Mobile 2003
12:27:12zeekoeColdtoast, what country?
12:27:16ColdtoastAustralia
12:27:22Coldtoastbut I live in Tasmania
12:27:27ep0chi sat in the park last week, and picked up 3 access points in one spot
12:27:29Coldtoastthe teeny little island off the bottom right
12:27:36zeekoeok :)
12:28:03Coldtoastsomebody's running wifi next to the blockbuster just down from me tho. And it's open
12:28:23Coldtoastso I go to blockbuster, look for films and go check the film out on IMDB using my PDA
12:30:15Coldtoastwhat router are you using ep0ch?
12:30:39ep0chLinksys WAG54G, not my choice
12:30:48ColdtoastWRT54GS here
12:30:48zeekoeis that the linux running one?
12:31:04Coldtoastthe WRT54G/S are zeekoe
12:31:08Coldtoastthe one I have
12:31:12zeekoeok
12:31:14ep0chi dont think the WAG one is
12:31:15zeekoeneat
12:31:28Coldtoastwell, it is. sorta
12:31:33ep0chcan you SSH to the WRT?
12:31:40ep0chor telnet
12:31:41Coldtoastthese are NOTORIOUS for dyin gon large file transfers tho
12:31:45Coldtoastyou can
12:31:48Coldtoastboth
12:31:53ep0chits for me :)
12:32:11ColdtoastI can't transfer files between any machines using this WRT
12:32:33Coldtoastafter a short time it resets
12:32:34ep0chon wireless or hub?
12:32:39Coldtoastwireless
12:32:50Coldtoastnever used the wired ports
12:32:57ep0chi take it you've changed the channel?
12:33:02Coldtoastapparently it's down to the CPU overheating or something
12:33:22Coldtoastyeah. even dropped to 802.11b and it does it
12:33:30ep0chnice QA
12:33:33Coldtoastno encryption at all either. still does it
12:33:56Coldtoastppl have managed to ease the prob by installing heatsinks in them tho
12:34:02zeekoelol
12:34:03Coldtoastso it seems it's an overheating issue
12:34:42ColdtoastI have a DLink DI-624 here somewhere too. That's not a bad router either
12:34:54Coldtoasttho now it seems to have DNS issues intermittenylu
12:35:02Coldtoastintermittently
12:35:05ep0chnetgear ones are good
12:35:34ColdtoastNetgear's 108Mbps router uses the same circuitry as DLink's 108Mbps router
12:35:52Coldtoastthey're pretty much the same router
12:35:55 Nick zeekoe is now known as zeekoe|food (~zeekoe@ip51cc69f6.adsl-surfen.hetnet.nl)
12:36:22Coldtoastbbiab
12:38:25 Join Asku [0] (~aksu@adsl-39.180-DynIP.ssp.fi)
12:40:27ColdtoastI'd hate to not have a router
12:40:41Coldtoastused to just have everythign running thru a switch
12:42:37 Quit MrStaticVoid ("leaving")
12:53:20zeekoe|foodColdtoast, NAT routers are actually a bad solution
12:53:22 Quit ep0ch ("Trillian (http://www.ceruleanstudios.com")
12:53:25zeekoe|foodipv6 r00lz
12:54:06Coldtoasta guy I know on EFnet is running IPv6 stuff
12:54:18zeekoe|foodat my other address, i have internet from the univerity (via cable)
12:54:33zeekoe|foodone modem (Com21), switch + 8 computers
12:54:41zeekoe|foodall with a different ip address
12:54:44Coldtoastthe WRT54G/S are easy to get set up on IPv6
12:54:58zeekoe|foodthat is, ipv4
12:55:11zeekoe|foodthings like that are possible for everyone when ipv6 will be running
12:55:20zeekoe|foodno more portmapping
12:55:25ColdtoastI think the only diff betwen the G and GS versions of the WRT54 is the amount of RAM
12:55:48ColdtoastGS has double what the G has
12:55:58zeGS is the speedboost version or whatever
12:56:10Coldtoastoh yeah. that's right
12:56:21zeekoe|foodnice names...
12:56:22Coldtoastthat's the other diff then
12:57:00Coldtoastnot that I have speedbost enabled anyway. Too many notebooks and my PDA that don't see it when enabled
12:57:11zeheh
12:57:38zeit advertises that it remains compatible with the standard-g devices and i think it even claims to improve their performance still somehow
12:57:50Coldtoastmy PDA is B
12:57:57Coldtoastand my friends' laptops are B
12:58:26Coldtoastthey can't see the wifi at all with speedboost enabled
12:58:30Coldtoastnone of them can
12:58:52ColdtoastHadaka, I think, 4 laptops here so far and none of them even saw there was wifi there
12:59:12ColdtoastHadaka? damn nick completion
12:59:21ColdtoastHadaka=had. heh
13:00
13:00:49zeah, B.. still kinda weird
13:00:59zewtg linksys eh :p
13:01:06Coldtoast:)
13:01:17Coldtoasttoo much Linksys, not enough Cisco
13:01:18zemaybe wrt is we really tried
13:01:21Coldtoastthat's how I think of it
13:01:22 Join XavierGr [0] (~XavierGr@ppp13-adsl-50.ath.forthnet.gr)
13:01:31Coldtoasthah. yeah
13:01:34XavierGrHi all!
13:01:44XavierGrLinusN: Are you here?
13:01:51ColdtoastWRT is probably close to when Cisco took over
13:02:19XavierGroh amiconn is here too!
13:02:50zeekoe|foodto be or not to be, that's the question
13:03:13XavierGrI managed to understand the settings code and added the remote scrolling settings and menu.
13:03:20XavierGrAnyone interested for them?
13:03:41zeekoe|foodi have archos...
13:04:36 Join Febs_away [0] (~chatzilla@207-172-122-81.c3-0.rdl-ubr4.trpr-rdl.pa.cable.rcn.com)
13:04:43 Nick Febs_away is now known as Febs (~chatzilla@207-172-122-81.c3-0.rdl-ubr4.trpr-rdl.pa.cable.rcn.com)
13:08:25 Nick zeekoe|food is now known as zeekoe|afk (~zeekoe@ip51cc69f6.adsl-surfen.hetnet.nl)
13:09:55***Saving seen data "./dancer.seen"
13:12:41FebsLinus, if you're around, let me know if you find this more accurate than the prior version: http://www.misticriver.net/boards/showthread.php?p=276192
13:12:53LinusNjust came back from lunch
13:13:31LinusNFebs: looks ok
13:14:32 Quit edx (Read error: 110 (Connection timed out))
13:17:54XavierGrLinusN: I tried the thing we said last time for the remote. I hard copied the whole tree.c in a remote-tree.c (made the apropriate changes) and initialized it from main.c together with tree.c.Well guess what...it runs which is first. That way I will have display either for the remote or for the main unit not both...
13:17:55 Join austriancoder [0] (~5078751e@labb.contactor.se)
13:18:02XavierGrhi austriancoder!
13:18:25austriancoderhi
13:18:48LinusNXavierGr: you mean you have browse_root() and remote_browse_root()?
13:19:46XavierGryeah! Its working for the first one. The other is ignored.
13:19:58LinusNof course
13:20:06XavierGr:(
13:20:07LinusNbrowse_root() never returns
13:20:17XavierGrso much work for nothing...
13:20:42LinusNwe tried to explain it to you yesterday
13:20:48XavierGryeah my bad
13:20:52LinusNnah
13:21:11LinusNit's not exactly a trivial piece of code
13:21:37XavierGrit is just that I couldnt make the restructural work so I had for the easy way.
13:21:55LinusNthe good news is that you do have a remote browser
13:22:08LinusNall we have to do is merge them
13:22:11XavierGryeah but the quick hack is still more usable
13:22:22XavierGrcan we do that?
13:22:57LinusNwell, it's not all that easy
13:23:58XavierGr(keep in mind that its an exact copy of tree.c with different functions, but this simpletask took me ages to escape from multiple defines)
13:24:12XavierGrso not so much work is done.
13:24:22austriancoderah.. there are now datasheets for the lcds of h3xx and x5 ;)
13:24:51XavierGrthe only progress I made is with the remote settings. I managed to add remote scroll settings
13:26:13Coldtoastnow you have the datasheet for the h3x0, is the only real hurdle now the bootloader?
13:26:38LinusNColdtoast: i guess so
13:26:45Coldtoastgood news
13:26:46*LinusN feels the pressure
13:26:55XavierGrlol
13:26:57Coldtoastheh
13:27:11Coldtoastno pressure from me. I have a h140 :)
13:27:53Coldtoasthmm. so maybe when there's an official iriver release, it might even be for the h3x0 as well!
13:29:06austriancodermaybe
13:32:13XavierGrdo you think that we will need the same cursor,scrollbar and statusbar for the remote? Maybe I can implement this?
13:33:04LinusNi don't know
13:34:24XavierGrthe invert selector looks neat but I havent tried the statusbar or the cursor selector.
13:35:09austriancoderLinusN: Have you startede with the CPU_COLDFIRE thing?
13:35:22LinusNno
13:36:36austriancoderLinusN: ok.. then i will do it
13:53:58austriancoderLinusN: i would add this to config.h in firmware/export/: http://nopaste.php-q.net/147788
13:55:05LinusNyup
13:55:37austriancoderand the e.g. in codesc/libwavpack/SOURCES: http://nopaste.php-q.net/147789
13:57:05austriancodershould be ok so?!
13:58:05XavierGrlaunch time!
13:58:13XavierGrlater....
13:58:15 Quit XavierGr ()
13:58:51LinusNaustriancoder: yes
14:00
14:04:38austriancoderLinusN: Can i replace all CONFIG_CPU == MCF5249 in app.lds too?
14:06:24 Join RotAtoR [0] (~e@dhcp54-47.calvin.edu)
14:08:16 Join rooomish [0] (~Roman@195.47.96.56.adsl.nextra.cz)
14:08:27rooomishhi2all
14:09:12austriancoderhi
14:09:38LinusNaustriancoder: no
14:09:59LinusNmany of those checks should not check the cpu, but the platform instead
14:11:19LinusNaustriancoder: nah, i was wrong
14:11:36LinusNgo ahead and change them to CPU_COLDFIRE
14:12:12rooomishamiconn: hello, what is new about your WIKI planned chapter and pictures about UART boot mod? The links on Joergen soft are still outdated (rooomish - the flash killer of his Archos)
14:12:46austriancodercan i do the changes also for system.c thread.c firmware\common\mem*.S ata.c?
14:13:39austriancoderkernel.c
14:13:41austriancoder?
14:14:01LinusNata.c is problematic
14:14:23austriancoderthe rest is ok?
14:14:34LinusNseems so, yes
14:14:55austriancoderrolo.c - too?
14:16:43austriancoderpower.c?
14:17:57 Quit Febs ("Chatzilla 0.9.68.5 [Firefox 1.0.5/20050711]")
14:30:18LinusNrolo.c and power.c contains a lot of iriver specific code
14:35:59austriancoderLinusN: check if all looks ok, and i will commit it: http://nopaste.php-q.net/147800
14:38:12LinusNok to commit
14:38:50austriancoderfine ;)
14:39:04austriancoderso we are a step closer to iaudio port
14:43:08austriancodernow we need a mcf5250.h
14:44:19LinusNyep
14:44:38LinusNhow about this:
14:45:02LinusN1) create a new mcf52xx.h with all the common stuff
14:45:32LinusN2) the mcf5249.h and mcf5250.h includes this and just defines the unique stuff
14:45:40LinusNdo you think that's a good idea?
14:46:01austriancoderyep.. its a good idea
14:46:58austriancoderhave you time to do this?
14:47:12LinusNmaybe later this evening
14:47:59austriancoderfine
14:56:06Slasheriaustriancoder: Hmm, i think you have to do #if defined(CPU_COLDFIRE) && !defined(SIMULATOR) and not #ifdef CPU_COLDFIRE && !defined(SIMULATOR)
14:56:14Slasherithere are lot of warnings about that kind of things
14:56:23 Join edx [0] (edx@p54A8E308.dip.t-dialin.net)
15:00
15:00:27austriancoderwill fix it
15:00:33Slasherigood :)
15:01:28LinusNoops, i only checked the locations where the checks were made, i didn't bother to check the #ifdef's themselves :-)
15:02:23 Quit Coldtoast (Read error: 113 (No route to host))
15:09:30austriancodershoule be fixed now.. can't compile it.. i am not ony ma laptop
15:09:56***Saving seen data "./dancer.seen"
15:11:20LinusNrockbox does indeed send bad s/pdif frames
15:12:04LinusNbut i can't understand why
15:15:43austriancoderLinusN: should we add to the config-xx stuff like DRAMORIG used in *.lds files?
15:16:01LinusNyes, we should do that
15:16:32austriancoderok ;) I will do it
15:17:22 Join ripnetuk [0] (~george@82-70-100-230.dsl.in-addr.zen.co.uk)
15:18:15austriancoderfor what stands the extension lds?
15:18:30ripnetuklinker symbols <−−−− guess
15:21:30austriancoderi will do following now
15:21:37ripnetukXavierGr - did you get the file browser working on the iRiver remote?
15:21:40austriancoder(1) add the default values to config.h
15:21:54austriancoder(2) add spezific values to config-xx.h
15:22:09austriancoder(3) change *.lds files to use new config stuff
15:22:19austriancodersounds this ok for you guys?
15:22:47LinusNdefault values?
15:23:32austriancoderstuff in else tree: http://nopaste.php-q.net/147820
15:25:30LinusNaustriancoder: those are the archos values
15:25:52austriancoderah ok
15:26:22austriancoderso i will add the values for each platform in their own config-xx.h file
15:26:30LinusNyes
15:26:34 Quit ripnetuk ("arrrgggth")
15:29:09elinenbewho is the person working on the iAudio port?
15:29:44austriancoderme
15:30:00austriancoderLinusN: are the comments for each define ok? http://nopaste.php-q.net/147821
15:30:23 Nick Lynx_ is now known as Lynx_awy (~lynx@tina-10-4.genetik.uni-koeln.de)
15:30:45LinusNi must say that the comments are somewhat redundant :-)
15:31:07austriancoder*g* maybe
15:32:16austriancoderdoes this look better? http://nopaste.php-q.net/147822
15:32:35austriancoderelinenbe: i am on of the two guys working on iaudio port
15:35:03LinusNaustriancoder: remember that ROMSTART is an archos-only symbol
15:36:43austriancoderhmm.. ok
15:37:39austriancoderhmm... for the iaudio i dont know all values
15:38:35 Join slnna [0] (~3ba7452a@labb.contactor.se)
15:40:34slnnahi guys
15:41:08austriancoderhi
15:41:55austriancodershould i also update the gmini stuff?
15:42:03 Join Febs [0] (~chatzilla@64-190-36-240.client.cypresscom.net)
15:42:18slnnawonder if I could ask a quick question? Does anyone have some issues with mp3s occassionally stopping before the end of the track?
15:43:21LinusNslnna: how long before?
15:44:03slnnasay, it's a 5 minute track, and it stops at 1min42secs and doesn't go to the next song on the playlist. Everything just ... stops ...
15:44:21LinusNoh, interesting
15:44:52slnnabut it's not a hang, coz I can just go the the next song and everything ... just happened to me on FSOL's Papua New Guinea ...
15:45:24LinusNslnna: is it a vbr file?
15:45:50slnnaprobably, I have the iriver in the car, I can go down and get it if u want ;)
15:46:03LinusNtry this:
15:46:39LinusNselect the file in the browser, puch and hold the joystick, select "open with" and then "vbrfix"
15:46:59LinusNkeep a backup of the file just in case
15:46:59austriancoderis something like this valid?
15:47:01austriancoder#define DRAMORIG DRAMORIG STUBOFFSET
15:47:10austriancoderDRAMORIG is defines in config-xx.h
15:47:19slnnacool, I go down now and see if it works, back in a minute
15:47:19austriancoder#define DRAMORIG DRAMORIG STUBOFFSET
15:47:37austriancoder#define DRAMORIG DRAMORIG (plus operator) STUBOFFSET
15:47:40SlasheriHmm, if the player takes some vibration it may be unable to fill the file buffer. Did you drive the car while you experienced the stop?
15:47:41LinusN:-)
15:48:18LinusNaustriancoder: i suggest you don't do that
15:48:45austriancoderLinusN: Should i add a new define?
15:48:45LinusNbetter use a third #define
15:48:49austriancoderok
15:48:53LinusNDRAMBASE or something
15:49:08LinusN#define DRAMORIG (DRAMBASE+STUBOFFSET)
15:49:22LinusNfor example
15:49:35*LinusN thinks he found the s/pdif bug
15:49:48austriancoderare there different IRAMSIZE values used in *.lds files?
15:49:57LinusNstinkin' lyin' data sheet!
15:50:25LinusNaustriancoder: yes
15:50:58austriancoderhmm.. so we need some IRAMSIZE_XX defines?!
15:51:37austriancoderor how should i make it?
15:55:41LinusNi assume you mean the difference between app.lds and plugin.lds?
15:55:51austriancoderyep
15:56:00LinusNhmmm
15:56:13austriancoder#define IRAMSIZE_PLUGIN...
15:56:22austriancoder#define IRAMSIZE_APP...
15:56:25LinusNi guess IRAMSIZE should be the entire IRAM size
15:56:35LinusNaustriancoder: yes
15:56:51LinusNthat reminds me
15:57:08LinusNdoes the codecs and plugins have separate iram spaces?
15:57:20austriancodershould IRAMSIZE noe have the value of iram or should i make two defines
15:57:21LinusNi'm afraid they both use the same space
15:57:35austriancoderLinusN: I dont know
15:58:02LinusNmake it 3 defines, IRAMSIZE, IRAMSIZE_PLUGIN and IRAMSIZE_APP
15:58:13LinusNwhwre IRAMSIZE is the total
15:58:29LinusNgotta go, bbl
15:58:36austriancoderok
15:58:38austriancoderthanks
15:58:40 Part LinusN
15:59:26 Quit ashridah (Read error: 110 (Connection timed out))
16:00
16:02:02 Join ashridah [0] (ashridah@220-253-123-251.VIC.netspace.net.au)
16:06:45slnnaby the way, is there any support for double byte languages like japanese and chinese on rockbox yet?
16:10:30 Quit slnna ("CGI:IRC")
16:10:37 Join slnna [0] (~3ba7452a@labb.contactor.se)
16:11:17austriancoderslnna: i think that somebody is working on unicode support
16:11:56slnnaoh ... cool. Great work guys! :D
16:13:07slnnaI am seriously amazed at how much better a rockboxed iriver is compared to the original POS firmware even in the current state!
16:14:06austriancoderneed to go now.. see you soon
16:14:08 Quit austriancoder ("CGI:IRC 0.5.4 (2004/01/29)")
16:19:02 Join rooomish2 [0] (~Roman@195.47.96.56.adsl.nextra.cz)
16:24:16 Quit ashridah ("sleep.")
16:26:15 Join shx [0] (~a4810127@labb.contactor.se)
16:29:22 Join yngwi [0] (~chatzilla@chello080109107064.1.15.vie.surfer.at)
16:32:00 Join oxygen77 [0] (~c1c28427@labb.contactor.se)
16:32:18 Quit oxygen77 (Client Quit)
16:38:13 Quit rooomish (Read error: 110 (Connection timed out))
17:00
17:02:29 Join BBub_ [0] (belzebub16@dsl-084-059-232-236.arcor-ip.net)
17:09:01 Quit hicks (Connection timed out)
17:09:58***Saving seen data "./dancer.seen"
17:12:02 Quit slnna ("CGI:IRC")
17:19:13 Quit BBub (Read error: 110 (Connection timed out))
17:22:02 Quit Febs ("Chatzilla 0.9.68.5 [Firefox 1.0.5/20050711]")
17:25:35 Quit BBub_ (Read error: 110 (Connection timed out))
17:30:44 Quit yngwi ("Chatzilla 0.9.68a [Firefox 1.0.4/20050511]")
17:34:09 Join rooomish [0] (~Roman@195.47.96.56.adsl.nextra.cz)
17:41:19 Join LinusN [0] (~linus@labb.contactor.se)
17:53:23 Quit rooomish2 (Read error: 110 (Connection timed out))
17:58:01 Join webguest52 [0] (~51565c95@labb.contactor.se)
17:58:13webguest52hi
17:59:17 Quit webguest52 (Client Quit)
18:00
18:00:37LinusNwow, he was fast
18:00:45LinusN(and i wasn't)
18:03:44rooomishLinusN: Hi, Do you know if Joerg is on holidays?
18:04:41LinusNprobably, haven't seen him for a while
18:05:46rooomishLinusN: oki thanks - I am trying to contact him for updating the dead links to his side (http://joerg.hohensohn.bei.t-online.de/archos/uart_boot)
18:05:59rooomishsite
18:13:06 Quit ]RowaN[ ()
18:15:38 Join Aison [0] (~hans@zux166-181.adsl.green.ch)
18:21:17amiconnLinusN: wiki spam :(
18:22:35LinusNkilled, thx
18:37:00 Join belgarath [0] (~acc9cb0d@labb.contactor.se)
18:37:08belgarathhi
18:37:23belgarathmy h120 came back from america!
18:38:26belgarathso i have eagerly installed rockbox (love the peakmeter and greyscale and wps alignment btw), and have created a database file and enabled runtime database, and i would like to know how you rate songs
18:39:15belgarathwait sorry its in the playlist context menu doh
18:39:40 Join pregloq [0] (~c39fb9ba@labb.contactor.se)
18:39:48 Nick pregloq is now known as preblow (~c39fb9ba@labb.contactor.se)
18:41:51 Quit belgarath (Client Quit)
18:42:34 Join preglow [0] (~c39fb9ba@labb.contactor.se)
18:42:50 Quit preblow (Client Quit)
18:43:14preglowhmm
18:43:23preglowthe new CPU_COLDFIRE define is pretty misleading
18:44:10LinusNhow so?
18:44:13preglownot all coldfires, indeed, very few of them, actually have the emac unit, and CPU_COLDFIRE is now used to activiate the emac specific parts of the source
18:44:23LinusNah, yes
18:44:42LinusNmaybe a HAVE_COLDFIRE_EMAC?
18:45:03preglowsounds good, something like that is needed
18:45:34preglowi thought the iaudio units used the 5249 anyway
18:45:58LinusN5250
18:46:08preglowyeah, so i see
18:47:22preglowhas austriancoder gotten a bdm working?
18:48:13LinusNnot yet
18:48:15 Join Febs [0] (~chatzilla@64-190-36-240.client.cypresscom.net)
18:49:54preglowbloody woot
18:49:59 Join rubberglove [0] (~ceacad2d@labb.contactor.se)
18:50:01preglow128kb sram
18:51:20rubberglovehello all. is there a kind soul around who could send me or point me at the windows 98 usb drivers for the h100?
18:51:44LinusNrubberglove: 98 or 98se?
18:51:54preglow98 doesnt have usb support, does it?
18:52:04zeekoe|afkit does
18:52:06rubberglovegood question... it's not my computer ... one sec
18:52:07LinusNnope, only 98se
18:52:07 Nick zeekoe|afk is now known as zeekoe (~zeekoe@ip51cc69f6.adsl-surfen.hetnet.nl)
18:52:16preglowany more talk about an ipod port? :P
18:52:21zeekoemy dad's got 98 running with usb mass storage
18:52:30 Join webguest90 [0] (~3e4f4094@labb.contactor.se)
18:52:44LinusNzeekoe: which device?
18:53:01webguest90Windows 98 has usb support - just not Mass storage drivers built-in
18:53:02zeekoesome card readers worked
18:53:08zeekoeindeed
18:53:16amiconnLinusN: Plugins and codecs share the same IRAM space; that was the reason for the rockboy crash when music was playing.
18:53:33rubbergloveLinusN: yay. second edition. boy this is tough with dial-up.... especially the iriver flash madness site...
18:54:04amiconnIt's solved now as rockboy is the only plugin using IRAM and needs the audiobuffer anyway, so it claims the audiobuffer (stopping playback by that) before loading its own stuff into IRAM.
18:54:05LinusNrubberglove: if you have 98se, you don't need drivers
18:54:15LinusNamiconn: i see
18:54:36zeekoeok, so that's the difference :)
18:54:37amiconnWe could split more, but I'm not sure whether this would be wise
18:54:41rubbergloveLinusN: oh yeah? it seemed to ask for drivers... i'll try again...
18:54:51amiconn..giving 32 KB to each of core, plugins and codecs
18:55:01LinusNmaybe you just need to install them from the win98se cd?
18:55:10amiconnI think it would be better to increase available IRAM for codecs instead
18:55:25LinusNamiconn: agreed
18:55:26amiconnThe core doesn't need the 64 KB it reserves now
18:55:42rubbergloveLinusN: let's see if i can find the cd now....
18:56:02amiconnAfaik win98 (se or not) doesn't support usb mass storage devices without specific drivers
18:56:18LinusNhmmm
18:56:40rubbergloveamiconn: that's what i thought. the h100 came with a cd for win98, but that's a few hundred km from me right now.
18:56:53webguest90You should be able to install mass storage drivers from any device and have it work with any mass storage device though, shouldn't you?
18:56:59amiconnShould be the same driver as for H120/h140
18:57:03rubbergloveiriver.com has win98 drivers for the flash players....
18:57:20rubberglovemaybe i'll give that a go.
18:57:21webguest90rubberglove: the cd doesn't include any driver (mine doesn't, at least)
18:57:49amiconnrubberglove: I have a CD here: iriver iHP series install disc...
18:58:04amiconnIt says PC software & device driver
18:58:30rubbergloveamiconn: any hope of sending it over? hopefully it's tiny?
18:59:07amiconnHmm, there's only one file: iHP100Manager1.20.exe
18:59:14amiconn3 MB
18:59:44rubberglovethat'll only take me an hour or so up here on the mountain... could you throw it at rubberglove at gmail.com?
19:00
19:00:16amiconnI'll check whether this contains a win98 driver...
19:00:26*amiconn is firing up the win98 se vm...
19:02:44 Join BBub [0] (belzebub16@dsl-084-059-232-236.arcor-ip.net)
19:03:37zeekoethe driver cannot be too tiny, because the driver needs to do things like creating an icon in the taskbar for safe removal too
19:03:55preglowmore iram for the codecs would be lovely
19:03:56 Nick Lynx_awy is now known as Lynx_ (~lynx@tina-10-4.genetik.uni-koeln.de)
19:04:02amiconnrubberglove: It's indeed an all-in-one installer that unpacks a driver for win98 (se)
19:04:26preglowplugins requring iram will probably always be too demanding to run while the codecs run anyway
19:04:49rubberglovecool. if you send it to me, i'll give you a nice cold pint of beer.
19:05:13rubberglovewell, actually, since my matter transferrence beam is down, i'll drink it myself in your honour....
19:05:22amiconn.exe as attachment is okay?
19:05:54rubbergloveamiconn: oh yeah. gmail is picky about that. try martin.king at mail.mcgill.ca
19:08:07amiconnSending mail....
19:09:12preglowany more differences between 5249 and 5250 than the sram?
19:09:24amiconnyes
19:09:36amiconnOne audio SPI instead of 2
19:09:36LinusNmore dma capabilities
19:09:51amiconn3 IEC958 instead of 4
19:10:02LinusNhttp://www.rockbox.org/twiki/bin/view/Main/ColdFire
19:10:03***Saving seen data "./dancer.seen"
19:10:38preglowahh, thanks
19:12:42rubbergloveThe comment about iriver.com was right, but the drivers are very well hidden.
19:12:42rubbergloveThe IHP Manager software (executable) has the drivers hidden inside
19:12:48rubbergloveThe comment about iriver.com was right, but the drivers are very well hidden.
19:12:48rubbergloveThe IHP Manager software (executable) has the drivers hidden inside
19:13:20rubberglove^sorry. web client. and thanks for the help
19:15:01amiconnMail sent...
19:15:13rooomishamiconn: please help, I cannot go to the holidays without my Archos :-((( What is new about your planned cookbook, pictures and SW about UART boot mod in wikipages? I have now the serial converter but the links on Joergen soft are still outdated (rooomish - the flash killer of his Archos)
19:17:28preglowman, i gotta get myself a linux box again, i miss rockbox coding :/
19:17:38amiconnMeh, boring documentation :/
19:18:03amiconnrooomish: I'll see what I can do
19:18:15rooomishamiconn: thank you!!
19:19:56zeekoenow i have the time to implement my autoloadconfig thingy... but you deleted the resume screen :'-(
19:21:30zeekoei guess i have to think of another place for it
19:26:37 Join lodesi [0] (~moi@lns-vlq-36-gre-82-253-95-253.adsl.proxad.net)
19:30:57 Join asdsd____ [0] (~asdsd@h-67-100-24-142.miatflad.dynamic.covad.net)
19:31:56 Part asdsd____
19:35:14 Quit webguest90 ("CGI:IRC")
19:36:12 Join [IDC]Dragon [0] (~d90a3255@labb.contactor.se)
19:36:48[IDC]Dragonrooomish: what did you do?
19:38:48 Quit thegeek ("( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de )")
19:39:05*amiconn spots a rare guest
19:39:09amiconnhi Jörg
19:40:23[IDC]Dragonhi
19:40:36[IDC]DragonI'll be even more rare in future
19:40:42[IDC]Dragon:-(
19:41:14 Join thegeek [0] (na@ti521110a080-2449.bb.online.no)
19:44:01LinusN[IDC]Dragon: how so?
19:44:27*[IDC]Dragon found another subject for hackeing: home automation
19:45:18LinusNrockhouse
19:45:24[IDC]Dragonthe next weeks I'll be very busy developing little uC modules for my house
19:45:40[IDC]Dragonsortof, yes ;-)
19:46:04[IDC]Dragonbut I'll go for Atmel instead of SH
19:46:10amiconnHmm, home automation is another thing I wonder whether it makes sense
19:46:33amiconnI know it must make sense somehow, since so much is written about it
19:46:54[IDC]Dragonyes, spend weeks of work or money on saving you a few minutes later on
19:47:16amiconnHowever, I can't imagine a single application to would make sense to me...
19:47:56zeekoewell, you can say "i have a home automated house"
19:48:03[IDC]Dragonswitch of all the lights when you leave, automate the shades
19:48:03zeekoeand all your geek friends will say "wow"
19:48:43rooomish[IDC]Dragon Hi, I flashed my Archos - all was done cheksum was correct by after !rebboting my Archos is dead
19:49:09[IDC]Dragondid you use the proper file for your model?
19:49:19amiconnActually I would not want to automate most of the things that can be automated with home automation
19:49:37rooomish[IDC]Dragon I think yes
19:49:48[IDC]Dragonthere are some ridiculous marketing scenarios, yes
19:50:36[IDC]Dragonlike, checking your iron or stove from vacation
19:50:39 Nick StrathAFK is now known as Strath (~mike@dgvlwinas01pool0-a218.wi.tds.net)
19:51:13[IDC]Dragonrooomish: which model, did you back up the original flash content?
19:51:30rooomish[IDC]Dragon I suppose that you are Jorge - and the links to your pages on Wiki (Uart boot mode) are dead
19:51:41rooomish[IDC]Dragon ARCHOS 6000
19:51:47[IDC]Dragonmy webspace is gone, yes
19:51:47 Quit rubberglove ("CGI:IRC (EOF)")
19:52:15[IDC]Dragonamiconns responsibility then :-;
19:52:16rooomish[IDC]Dragon I have some backup (hope that correct)
19:52:24amiconn[IDC]Dragon: I actually mean most of the things. Like, why would I want to automate switching the light? If I ener a room, there's a switch to turn the light on, and if I leave again, I turn it off again. If I leave the light on, there is a reason no automation could know
19:52:48amiconnThat's what I wonder about...
19:53:12 Quit preglow ("CGI:IRC (Ping timeout)")
19:53:34[IDC]Dragonmaybe you want to switch it all off from bed
19:53:48 Join solex [0] (~jrschulz@c223092.adsl.hansenet.de)
19:53:52amiconnI'm not *that* lazy ;)
19:54:29amiconnNo, really, I take it that there might be useful applications, I just can't imagine *any* application that I might want to use
19:55:01[IDC]Dragonwe really need it for the shades
19:55:27[IDC]Dragonlight control was a nice to have
19:56:28rooomish[IDC]Dragon: Unfortunatelly the links on UART boot mode WIKI page
19:56:28rooomish to your site http://joerg.hohensohn.bei.t-online.de does not working.
19:56:28rooomish Have you any different site?
19:56:46[IDC]Dragonrooomish: I'll put my files into the wiki
19:56:54rooomishthank you
19:57:13[IDC]Dragonfor the Studio, amiconn made some pics, iirc
19:57:16amiconn[IDC]Dragon: You have pictures showing the recorder uart boot mod, correct?
19:57:32[IDC]Dragonrecorder, FM, ondio
19:57:47amiconnWe should put together a comprehensive uart boot wiki page
19:57:56amiconnI have pics for player/Studio
19:57:58rooomishyes!!!
19:57:58[IDC]Dragonyes
19:58:08rooomish:-)
19:58:26*rooomish will wait
19:58:32[IDC]Dragonthanks
19:59:17*[IDC]Dragon got quite some mail about the dead homepage
19:59:51[IDC]Dragonmostly about M-Bus, that was on #1 google
20:00
20:00:10amiconnI've put the flash images on my webspace, at least
20:00:50[IDC]DragonI have 100 new MB of webspace, but need a domain to use it
20:01:33[IDC]Dragon1&1 includes one, but I haven't found out if that's just the hosting, or the registration, too
20:03:25*amiconn has 200 MB and an own domain :)
20:04:35[IDC]Dragonhow's the domain handled? directly with denic?
20:04:51rooomishamiconn: what www address?
20:05:08amiconn[IDC]Dragon: T-Online
20:05:13amiconnwww.jens-arnold.net
20:05:18rooomishoki
20:05:25[IDC]Dragondid you have to pay extra?
20:05:29amiconnrooomish: Beware of German ;)
20:05:35amiconn[IDC]Dragon: Yup
20:05:49 Quit solex_ (Read error: 110 (Connection timed out))
20:05:52rooomishamiconn: aha :-(
20:06:05[IDC]Dragonhow much?
20:06:55amiconn'Homepage Starter' package. ftp://software.t-online.de/pub/service/pdf/lbhpstarter.pdf
20:08:19[IDC]Dragonaha
20:08:43amiconnDifferent topic - any news about bootbox?
20:08:57amiconnI made some bootbox related commits....
20:09:58[IDC]DragonI should commit mine
20:10:19amiconnAnything new to test on the various platforms?
20:10:20[IDC]Dragonbut don't expect any big changes
20:10:41[IDC]Dragonthey are covered now
20:11:08[IDC]Dragonbut most only by one person
20:11:25amiconnAny decision about the padding yet?
20:11:34[IDC]Dragonno
20:11:41amiconnThe flash plugins will need some work as well
20:11:52[IDC]DragonI'd like to use the same link file as normal
20:12:05amiconn?
20:12:44[IDC]Dragonbootbox uses the standard rockbox linking
20:13:20[IDC]Dragonthe only idea I have for padding would be a different linkfile
20:13:39[IDC]Dragonbut, we ucl pack it afterwards...
20:13:52amiconnHmm? You could pad just by defining the start address of image #2
20:14:04[IDC]Dragonnot nice
20:14:11amiconnWhy?
20:14:24[IDC]Dragonit breaks the normal procedure
20:14:33amiconnI think it would be the easiest solution to set the start address of image #2 to a fixed value.
20:14:38amiconn...for all platforms
20:14:57[IDC]Dragonyes, easy, quick and dirty
20:14:59amiconnThis allows for some growth of bootbox, and simplifies linking
20:15:11 Join austriancoder [0] (~5078751e@labb.contactor.se)
20:15:14[IDC]Dragonbut you're forced to use bootbox
20:15:31amiconnHow do you mean, forced?
20:15:41amiconnThe link address for rombox will change anyway
20:15:47austriancoderamiconn, LinusN: i will overwork the CPU_COLDFIRE thing to seperate it with emac's
20:15:52austriancoderwill do it later the day
20:15:55austriancoderneed to go now
20:15:58 Quit austriancoder (Client Quit)
20:16:04 Nick Lynx_ is now known as Lynx_awy (~lynx@tina-10-4.genetik.uni-koeln.de)
20:16:12[IDC]Dragonif you want to use Archos as image #1
20:16:14amiconnHow else would you reserve extra space allowing for bootbox growth without moving rombox start again later?
20:16:34[IDC]Dragonby padding the image
20:16:47amiconnHow do you pad an .ucl?
20:16:55amiconn...in a predictable way
20:17:10[IDC]Dragonthat's the question
20:17:26amiconnYes
20:17:28[IDC]Dragonanother option :-/
20:18:10amiconnAnd, I wonder why somebody would want archos as image #1, and how he would accomplish this without building everything himself
20:18:31amiconn(and having no safety net except when doing the uart boot mod too)
20:18:39 Join zeekoe_ [0] (~zeekoe@ip51cc69f6.adsl-surfen.hetnet.nl)
20:18:57[IDC]Dragoneasy: by using todays image file
20:19:06amiconnAfaik rockbox_flash does allow flashing image #2 only
20:19:19amiconnYes, but then he can't use later rombox versions
20:19:33amiconn...because they will be linked for the wrong address
20:19:55[IDC]DragonI could lower limit the image #2 start
20:20:24amiconnHow does the firmware tool determine the size of image #1?
20:20:43[IDC]Dragonwhich tool?
20:20:47amiconnHmm, btw, when padding, it would make sense to pad to a full flash sector
20:21:03amiconn...and that has to take bootloader size into account
20:21:12[IDC]Dragonyes
20:21:20amiconnmake_firmware
20:22:00[IDC]Dragonby file size
20:22:05[IDC]Dragoniirc
20:26:39amiconnLooks like the bootloader determines the position of image #2 from position and length of image #1
20:26:39 Quit gromit` (Read error: 104 (Connection reset by peer))
20:26:52[IDC]Dragonyes
20:26:56amiconnSo just 'pad' the size field in make_firmware...
20:26:57[IDC]Dragonsmart, eh?
20:26:57 Quit zeekoe (Read error: 110 (Connection timed out))
20:27:18 Join gromit` [0] (~gromit`@ras75-5-82-234-244-69.fbx.proxad.net)
20:28:37[IDC]Dragoncould work, an option to make_firmware
20:30:52[IDC]Dragonany perl hacker around?
20:36:38*LinusN looks away
20:37:30[IDC]Dragonsince an hour...
20:37:35[IDC]Dragonhi
20:38:56[IDC]Dragonhow do I trim trailing whitespace from a string, if it contains whitespace itself?
20:39:12[IDC]Dragoneg. "A B C "
20:39:21LinusNfrom the top of my head:
20:39:57LinusNs/\w+$//
20:43:38[IDC]Dragonseems to do nothing
20:44:08[IDC]Dragonah
20:44:18[IDC]Dragons/\s $//
20:44:38Harpys/\s+$//
20:44:41[IDC]Dragonmy plus gut lost
20:44:47[IDC]Dragongot
20:44:54[IDC]Dragonthanks
20:45:29LinusNsorry
20:46:01[IDC]Dragonnp, this was helpful
20:47:57 Join MrStaticVoid [0] (~jlee@69-175-94-207.frdrmd.adelphia.net)
20:51:46 Join Lear [0] (~chatzilla@h143n1c1o285.bredband.skanova.com)
21:00
21:05:03 Join belgarath [0] (~acc9cb0d@labb.contactor.se)
21:05:43 Join preglow [0] (~c39fbef3@labb.contactor.se)
21:10:04***Saving seen data "./dancer.seen"
21:18:16preglowanyone else have the problem where fragments of the old track reappear when changing to a new track?
21:20:43LinusNi had it, but i haven't tested it thoroughly since last friday
21:22:00preglowi get it all the time
21:22:04preglowconsistently
21:22:07preglowand have for a long time
21:24:47preglowi also get some weird skipping sometimes now
21:25:29preglowduring track changes
21:27:15 Join Stryke` [0] (~Chairman8@cpe-24-168-110-99.si.res.rr.com)
21:29:40preglowand while i remember it, i assume you havent looked at the imdct thing?
21:30:42LinusNnope
21:31:07preglownot important anyway, the current one does the job nicely even if it is ugly
21:31:12 Quit [IDC]Dragon ("CGI:IRC")
21:31:34preglowit's not the first ugly thing that has sprung from my hands
21:32:04 Quit preglow ("CGI:IRC")
21:32:17 Join preblow [0] (~c39fbef3@labb.contactor.se)
21:32:42 Quit lodesi ("Leaving")
21:32:49preblowi miss my linux box ://
21:33:54 Nick preblow is now known as preglow (~c39fbef3@labb.contactor.se)
21:35:52belgarathis song count and rating stored in the mp3 file, or the id3 database file?
21:36:50preglowa separate runtime db file, i believe
21:37:02preglowin any case, in a database
21:41:07preglowman, track changing is really having a bad day here
21:41:56 Quit zeekoe_ (Read error: 110 (Connection timed out))
21:44:02 Quit belgarath ("CGI:IRC (EOF)")
21:44:43 Join yngwi [0] (~chatzilla@chello080109107064.1.15.vie.surfer.at)
21:51:08 Join Coldtoast [0] (~edan@ppp111-3.lns1.hba1.internode.on.net)
21:56:26*amiconn is puzzled
21:56:45preglowwhat're you working on?
21:59:13amiconnWhy did the number of build errors for H3x0 target lower so much while all of today's changes don't touch lcd related stuff in an obvious way??
21:59:16Slasherii want to add optional dir caching to iriver :)
21:59:24amiconnNooooooooooooooo
21:59:29Slasherii just calculated it will not eat so much ram..
21:59:34Slasheri:D
21:59:37amiconnArrrrrrrrrrrrrrrrrrrrrrrggggggggggggggggghhhhhhhhhhhhhhhhhhhhhh
21:59:44yngwihi guys, im just trying to build my own crosscompiler :-)
21:59:47amiconnNo dir caching in rockbox pls pls pls
21:59:51ZagorSlasheri: only as much as you give it...
21:59:56Slasheriof course better caching than with iriver firmware has
22:00
22:00:00Slasheri:D
22:00:10SlasheriZagor: yes, the limits should be configurable..
22:00:20ZagorSlasheri: first question: why?
22:01:25SlasheriZagor: i would just like that directory browsing would be always instant, no matter if hdd is spinning or not.. no other reason for that :)
22:02:04Zagorwhy? "it feels good"?
22:02:21Slasheriyes, of course i could also increase the hd spinup time
22:02:48Slasheribut i still have to wait when the hd initially spins up
22:02:49amiconnI'd rather like to see some more existing rockbox features working on iriver, like voice UI
22:03:19Slasheriamiconn: that's true..
22:03:22ZagorSlasheri: you will *always* have delays, unless you cache the entire dir structure. and then you're not using "just a little" ram anymore
22:03:44amiconnI'd love to work on it, but (1) I don't understand that much of the codec magic and (2) I still have some LCD 'homework' pending
22:04:48Slasheriamiconn: i think the only big problem with the voice is to decide how to implement the codec loading: static codec or another dynamic memory area for that second codec?
22:04:52 Quit RotAtoR ()
22:05:45amiconnI would be glad to build the voice files once the voice UI is working, even per model using localisation v2 if I get an automatable upload access
22:05:49Slasheriafter that it should be trivial to feed the codec with the voice ui data
22:05:56Slasheripcmbuf has to implement some mixing also
22:06:00preglowthe codec will also have to be very fast if it is to not use iram
22:06:19preglowwhich it cant
22:06:27amiconnSlasheri: Voice UI should be able to deal with lower sample rate, in order not to waste space
22:06:28Slasheriah, true.. that's a problem too
22:06:51preglowyou seem very intent on using mp3, so this will be a great problem
22:06:53Slasheriamiconn: yep, but it can use dsp for that
22:07:41amiconnIf we want to be able to 'voice over' then it should be able to handle voice_sample_rate != music_sample_rate too
22:08:00LinusNamiconn: sample rate will always be 44.1kHz
22:08:01preglowwell, we have resampling code, and ill make some more as soon as im settled with a dev environment
22:08:03Slasherii think the best solution could be to have very simple codec with powerful compression that can be statically linked with rockbox
22:08:11amiconnpreglow: The mp3 focus will vanish as soon as Bagder's localisation v2 idea gets implemented
22:08:26Slasheriamiconn: dsp does that, no problem
22:08:57amiconn...because that will need different voice files per model anyway, without 'dead' snippets
22:09:15amiconn...and then we can tailor the voice files in quality
22:09:32preglowstill, if we are to allow mixing the voice ui over music, well have a speed problem
22:09:52amiconn..and use any codec we want; the only limit being that I'd need a command line encoder running under windows to be able to build the voice files
22:10:04preglowwell, all codecs should have that
22:10:06preglowall i can think of
22:10:44amiconnI don't think that we'll have much of a speed problem
22:10:49preglowno?
22:10:55preglowmp3 is dead slow without sram
22:10:56amiconn(1) The voice UI typically uses low sample rate
22:11:01preglowvorbis is not even realtime without i
22:11:23amiconn(2) The voice UI won't typically babble continuously for longer time intervals
22:11:44preglowwell, it will have to be very responsive anyhow
22:12:03amiconnYes, but I mean it shouldn't influence music playback much
22:12:04preglowif the codec we use is too slow, the ui wont be as responsive as it should
22:12:30amiconn...because short periods which are below realtime in total would be covered by the pcm buffer
22:13:22preglowhow long is the pcm buffer at the moment?
22:13:29preglowseems like its several seconds long
22:13:55amiconnI that if we start with mp3 because of the lack of a fast dedicated speech codec, I'd use my 'wanted' (but not possible due to space constraints) setting for encoding
22:14:07amiconnlame vbr 16 kHz mono
22:14:39amiconn...and I don't see why we couldn't use IRAM for the secondary codec
22:15:18amiconnWe have 96KB of IRAM of which the codecs get 32KB today
22:15:35preglowi think it would be a terrible waste to use 32kb of iram on an optional feature
22:15:39amiconnThe core has 64KB, and it doesn't even use the half of it
22:15:54preglowand i dont plan on allowing it staying that way :)
22:16:04amiconnYes I know
22:16:11Slasheripreglow: about 6 seconds
22:16:16Slasheri(1 MB)
22:16:25preglowholy crap
22:16:34preglowis that going to be shortened?
22:16:40LinusNi hope so
22:16:48amiconnpreglow: Btw, while the voice UI is optional, it is considered an important feature by a number of rockbox users
22:17:08amiconn(me among them - *very* useful while driving)
22:17:28preglowamiconn: indeed, but it is still an optional feature i dont think the majority will use, and there should be plenty of ways to implement it with no sram hogging
22:17:39amiconnIt is now about the only reason why I still prefer my archoses over the iriver
22:18:21preglowanyway
22:18:45preglowmight as well do that for now, i cant cook up something better at the moment myself
22:19:00 Join zeekoe_ [0] (~zeekoe@ip51cc69f6.adsl-surfen.hetnet.nl)
22:19:14amiconnAnyone knowing a small, efficient, open source voice codec available as integer implementation?
22:19:25preglowbut i would like the main codecs to be the main focus, they are the most crucial part of rockbox on platforms with no hw codec
22:19:48amiconnAnother option - make one of the codecs reentrant and static
22:19:49preglowamiconn: well, speex is by far the best, and it seemed to have very good potential for optimisation
22:20:10preglowamiconn: as for reentrancy, mp3 is very well suited to that
22:20:11amiconnThat way it could be used by both main audio thread and voice UI...
22:21:04preglowi believe libmad could function in the way you mean as it is, if you back its structs to ordinary ram when changing to voice ui and back
22:22:12amiconnHmm, that would still require to set some IRAM aside for the static codec, the only difference being that it wouldn't be used exclusively by the voice UI...
22:22:37preglowtrue enogh
22:22:38preglowenough
22:22:44 Join ep0ch [0] (~ep0ch@81-6-218-35.gotadsl.co.uk)
22:22:54preglowi'd love to experiment with speex
22:22:56amiconnI think we can get away with less IRAM than for the main codec
22:22:57amiconns
22:23:00ep0chhey, speakling of codecs for voice, would ADPC
22:23:03ep0ch M be suitable?
22:23:17amiconnToo little compression, I think
22:23:20ep0ch4:1
22:23:20preglowactually, it's the thing i want to do right now, but i have no dev env and my own pc has no internet
22:23:38ep0chuse 22khz then its 8:1
22:23:54preglowadpcm is computationally simple, at least
22:24:11preglow16khz would be fine for voice
22:24:29ep0chand it'll be mono
22:25:10amiconnep0ch: For voice, 16 kHz 16 bit mono is sufficient for voice, but I'd like higher compression
22:25:26ep0chit wont be 16bit, but 4 bit
22:25:33amiconnspeex might be the choice because of it being a dedicated speech codec, and opensource
22:25:33preglowbut anywho, i'll get working on speex as soon as i can
22:25:57preglowif no one gets there before me
22:26:05amiconn...but it will require some work, and I don't know how fast it will run on our coldfire
22:26:19preglowlike i said, lots of potential for tight asm code
22:26:43amiconnpreglow: Is the integer port finished now?
22:27:00preglowi think speex does function in fixed point mode now
22:27:11preglowsome files arent ported yet, but they're not essential either
22:27:18amiconns/integer/fixed point/
22:27:30Slasheripreglow: in the future that pcm buffer size could be configurable (with crossfade longer and without it short)
22:27:31amiconnHmm.
22:29:27yngwii just tried to build the binuils for the crosscompiler: a stupid question; how can i tell, it worked?? :-)
22:30:07preglowif you didnt get any compiler errors, chances are very good it worked
22:30:07amiconnpreglow: As you asked, I'm working on porting the grayscale lib to the new api, and then to H1x0
22:30:29preglowi like the flat shaded cube, btw, looks really nice ;)
22:30:30LinusNyngwi: you will notice
22:30:38yngwithanks :-)
22:30:57amiconnI already did some experiments on H1x0, and it *is* possible to use native 4-grey mode together with pixel flipping
22:31:00yngwii thougt i better start now learning c, else rockbox gets finished without me :-)
22:31:12LinusNyngwi: that would be terrible :-)
22:31:17yngwi:-)
22:31:28LinusNi mean, rockbox being finished
22:31:33preglowhahah
22:31:37LinusNwhat would i do???
22:31:37yngwiuh oh :-P
22:31:41amiconn...meaning the grayscale lib will support more shades than on archos (either 49 or 97, depends on some details I'll work out while implementing)
22:31:44LinusNi have to get a life
22:31:57preglowamiconn: you did a test, i assume?
22:32:02amiconnyup
22:32:18yngwino need to hurry, there are a lot poor mp3 players without decent firmware out there
22:32:27preglowlinusn: it's too boring, don't bother
22:32:29amiconnI have an experimental plugin showing 13 shades at one. That's without random pattern shifting, so it flickers a tiny bit
22:32:44ep0chi dont understand how you can get more than 4 shades? i know old demos used to do things like that, but how???
22:32:56preglowtemporal dithering
22:33:00LinusNep0ch: the same way as the demos do it
22:33:10amiconn...and it's hacky, so it requires to deactvate backlight fade in/out before running it, otherwise it will crash
22:33:25amiconnI need to write firmware/drivers/timer.c ....
22:33:45ep0chbasically you switch pixels on and off really quickly?
22:33:49amiconnyup
22:33:50LinusNep0ch: yes
22:34:09ep0chthat's kinda...
22:34:19ep0chermm... cheating :)
22:34:30ep0chbut i would love to see it :)
22:34:47preglowamiconn: any obvious artifacts from the slow lcd?
22:34:48amiconnep0ch: The H1x0 lcd does the very same thing internally to provide the 4 shades
22:34:59Learanyone using the win32 simulator? I've noticed some problems, and wonder if anyone else has seen them...
22:35:02ep0chreally? wow
22:35:04preglowamiconn: oh?
22:35:08amiconnpreglow: This is actually helpful; it helps to avoid flicker
22:35:50ep0chso the lcd is really mono
22:36:08amiconnpreglow: With my test plugin, I'm actually able to 'see' the internal patterns of the lcd, by superposition
22:36:23amiconn(the plugin allows to adjust the scan frequency in a wide range)
22:37:26 Quit web1 ("CGI:IRC (EOF)")
22:37:42amiconnLinusN: That's the nice thing: The lcd has an internal scan frequency of ~70 Hz in 4-grey mode (sounds familiar?)
22:38:04amiconnWith your improved lcd timing, this means 70 ms /second, or ~7% cpu load
22:38:16amiconn...plus some overhead, but that should be small
22:38:29ep0ch7% at 120 mhz?
22:38:33amiconnyup
22:38:33yngwiwhen the "cross compiler page" talks about the PATH (compiling gcc) is this the windows PATH, or something changed in cygwin??
22:38:47amiconnyngwi: It's the cygwin path
22:38:58 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
22:39:20preglowdefinitely looks like speex is usable in fixed point mode, yes
22:39:41amiconnep0ch: On archos we need 50% cpu for full screen, due to the lcd data line being serial
22:40:39ep0chi take it this grayscale patch wont be native to rockbox, but used just by plugins?
22:40:48amiconnOn archos, we need to feed 67 frames of 7168 pixels (896 bytes) per second
22:40:54amiconnep0ch: yup
22:40:55LinusNamiconn: nice
22:41:26amiconnOn iriver, we need to feed 70 frames of 20480 pixels (5120 bytes) per second
22:41:43ep0chbut faster CPU
22:42:00amiconnyup, and parallel LCD data lines
22:42:18ep0chsounds like you're having fun :)
22:42:43ep0chohhh, i have another idea for a codec....
22:42:57ep0chbut how close is the coldfile compared to a 68020?
22:43:19preglowcoldfire, and it should be quicker, i dont think the 68020 came with too high a clock
22:43:20ep0chcoldfire
22:43:43amiconnDue to the iriver version not flicking between (full) black & white but only between 2 adjacent of the 4 shades, and due to the slow lcd, it is much less susceptible to artifacts from not being able to synchronize internal and externa
22:43:44preglowwhat codec are you thinking of?
22:43:48amiconnl scan rate
22:43:49ep0cheagleplayer :) http://www-md.e-technik.uni-rostock.de/~bj74/software/eagleplayer/index.html
22:44:07ep0chto play all the old amiga mod like formats
22:44:23preglowhahaha
22:44:28preglowthat'll be "fun" to port
22:45:12ep0chi'm sure there's some nostalgic person out there would like to do it
22:45:56amiconnep0ch: That's what dumb is for. It's a mod player, but not yet functional as a codec
22:46:01*amiconn prods HCl
22:46:10ep0chnonono
22:46:38ep0chDUMB doesnt support all the formats
22:47:20ep0chthere are literally hunderds of mod formats that used the CPU to do some work on the Amiga that most mod players don't replicate
22:47:48preglowwhat, you mean the formats that contain code?
22:48:08preglowwe'll have to emulate those, the coldfire doesnt support all the instructions of the 68k family
22:48:19ep0chhmmm, not sure
22:49:36ep0chthink i'll email the authors, never know they might own an ihp ;)
22:51:37ep0choh the unix version is called UADE and uses WinUAE code to emulate the CPU if you're interested
22:52:58preglowwe're probably better off using a standalone emulator like starscream, if the code isnt dependent on having a full kickstart mapped to memory or something
22:53:17 Nick CBM-away is now known as CheeseBurgerMan (~youshould@tc2-225-085.altelco.net)
22:54:48ep0chheh well definitely fun then
22:55:23amiconnep0ch: deliplayer does the same thing.
22:55:33ep0chyes its based on eagleplayer
22:55:37ep0chi think
22:55:49amiconnNo, it's based on delitracker code
22:56:00amiconnIt's made by the same developers
22:56:40amiconndeliplayer uses a minimal m68k emulator based on uae techniques to play formats with embedded code, like Whittaker, FredMonitor etc
22:57:24 Quit preglow ("CGI:IRC")
22:57:46amiconn...but while I would appreciate mod support, I'm not that much interested in these exotic formats. >50% of all mods I have are ProTracker, and that's pretty generic
22:58:10ep0chi know its quite specialised..
22:58:26amiconnI can still play all those exotic formats on my real Amiga :)
23:00
23:02:38Zagoror in uae
23:03:20ep0chor uade piped to oggenc and play on the ihp
23:05:20amiconnWhy should I use UAE when I can use the real thing? ;)
23:05:35amiconnReminds me - I need to fix the audio out - again :-/
23:05:39Learin the simulator, would it be safe to make any assumptions about the size of the data returned by pcmbuf_callback?
23:06:05 Join muesli- [0] (muesli_tv@hmln-d9b8e198.pool.mediaWays.net)
23:07:09muesli-g'evening ladiez
23:07:48ep0ch:o
23:07:55ep0chwhere where?!
23:08:10muesli-just check my bra ^^
23:08:22HClo.o;;
23:08:29HClmuesli- is female? O.o;
23:08:36muesli-dunno
23:08:40muesli-maybe ^^
23:08:41ep0chjust today eh?
23:08:43HClthats even scarier o.o.
23:08:50muesli-hehe :D
23:09:03muesli-depends on my daily condition...
23:09:26HClo.o;
23:10:06***Saving seen data "./dancer.seen"
23:10:29muesli-well..its a surprise every morning
23:10:34muesli-even 4 me :D
23:10:50HClmhm o.o... *slowly backs away from muesli- while he nods*
23:11:09 Join [1]CheeseBurgerM [0] (~BurgerBoy@tc2-225-085.altelco.net)
23:11:15muesli-hehe
23:11:18muesli-well..
23:11:30muesli-just biught the new harry potter
23:11:39muesli-time to be relaxed...
23:11:40HClnice
23:11:54ep0chdoes anyone die in this one?
23:11:58HClshush.
23:11:59HCl :p
23:12:00 Quit CheeseBurgerMan ()
23:12:01 Nick [1]CheeseBurgerM is now known as CheeseBurgerMan (~BurgerBoy@tc2-225-085.altelco.net)
23:12:01HClno spoilers
23:12:26muesli-somebody is gonna be killed
23:12:43 Quit MrStaticVoid ("Lost terminal")
23:12:44muesli-guess its snape or hagrid
23:13:02ep0chhas dumbeldore died yet?
23:13:08thegeekhehehehehehehe
23:13:10ep0chhe's getting old...
23:13:10muesli-he will not
23:13:12thegeekmuahaha
23:13:18muesli-yes, he's old
23:13:30muesli-maybe its time for him to give up the spoon
23:13:36muesli-as we say in german
23:13:41ep0chyeah, and let harry take over
23:13:51muesli-yeah, harry 4 president ^^
23:14:04thegeekare your serious?`
23:14:07thegeekyou dont know it?
23:14:15ep0chshh
23:14:18ep0chdont spoil it
23:14:23muesli-shut up dude
23:14:24thegeekbut !
23:14:27thegeekit's so tempting
23:14:30ep0chshh :)
23:14:34muesli-shut up dude
23:14:36DBUGEnqueued KICK muesli-
23:14:36muesli-shut up dude
23:14:37muesli-shut up dude
23:14:44ep0chi wanna read it too now
23:14:50muesli-but wouldnt wonder either
23:14:57ep0choh i have like 5 books to read first
23:15:02thegeekhttp://content.ytmnd.com//94000/94316/image.jpg
23:15:03thegeekthere
23:15:05thegeekdont click it
23:15:05muesli-lol
23:15:07thegeekunless you want to know
23:15:26ep0chyou b*st*rd
23:15:27muesli-why dont i own op status...
23:15:34thegeekwtf?
23:15:34ep0ch:D
23:15:37thegeekI said dont click it
23:15:39muesli-LinusN kick him
23:15:41thegeekunless you want to know
23:15:58thegeek*feels scared*
23:16:00thegeek;)
23:16:10thegeekbut seriously
23:16:15thegeekthe book is good anyway
23:16:23thegeekknowing it wont spoil or ruin anything
23:16:39ep0chmuesli-: did you click?
23:16:43thegeekanother spoiler : http://www.filefarmer.com/Xizer/pottercrash.MOV
23:16:45muesli-nope, didnt
23:16:46thegeekit's extremely funny
23:16:48ep0chgood
23:16:51thegeekbut dont watch unless you want to know
23:16:58thegeekthe last 5 seconds are hilarious
23:17:15ep0chsorry can't do that
23:17:19ep0chi don't do MOV
23:17:31thegeekvlc
23:17:36muesli-the 5th was very scary
23:17:41thegeekor even quicktime alternative
23:17:48thegeekI like them in english
23:17:53 Quit Harpy (Read error: 110 (Connection timed out))
23:18:00thegeekwhen they are translated into norwegian they kinda loose the "edge"
23:18:14thegeekin english they are funny and interesting
23:18:15muesli-yepp..never read a translation
23:18:24thegeekin norwegain; mediocre
23:18:29muesli-oh i did..it was very boring
23:18:33thegeekit's not the translation is bad
23:18:37thegeekit's just not as good as the original
23:18:42muesli-yepp
23:18:45muesli-true.
23:19:20thegeekbut seriosly
23:19:23muesli-its just the way they let words stand as they are in english
23:19:25thegeeksomeone watch that movie
23:19:31muesli-doenst fit translated
23:19:32thegeekespecially the last seconds
23:19:35thegeekmhm
23:19:41thegeek*seriously
23:20:57ep0chwhat does ytmnd stand for?
23:21:00muesli-anyway, i pray for a fully supported iriver remote :D
23:21:30muesli-ep0ch dunno
23:21:35thegeekme neither
23:21:55muesli-thegeek greetz to mette-marit ;)
23:22:09muesli-hot babe :D
23:22:10thegeekwhat
23:22:11 Join f_x [0] (~54b891a9@labb.contactor.se)
23:22:13thegeekwhat now?
23:22:13thegeekhehe
23:22:16thegeekI've met her
23:22:23muesli-neither i did ;)
23:22:25thegeekwhen I was in the military
23:22:33thegeekI stood guard outside her "house" all the time
23:22:34 Quit f_x (Client Quit)
23:22:42thegeekopened the gates and stuff when she went running
23:22:48muesli-lame...didnt get into her bedroom ;)
23:22:55thegeekshe's not that hot.
23:23:01muesli-not`?
23:23:01thegeeknot bad, but kinda old
23:23:13muesli-doesnt matter..
23:23:13thegeekI've seen her in training clothes
23:23:30muesli-shes in the middle of 30?
23:23:33thegeekshe was nice though, thanked me and stuff;)
23:23:41muesli-:D
23:23:41thegeekyes
23:23:52thegeekin fact
23:23:53muesli-how old r u?
23:23:59thegeekI was standing there with 10 live shots
23:24:01ep0ch1973
23:24:11thegeekand an ag3 machinegun
23:24:20ep0chthis the crown princess?
23:24:22thegeekafter only 6 months in the military
23:24:23thegeekyes
23:24:33thegeekif someone really wanted to take them out
23:24:37thegeekit would be extremely easy
23:24:52 Join f_x [0] (~birk@p54B891A9.dip0.t-ipconnect.de)
23:25:07thegeekI would shure as hell not've engaged anyone if there was an attack
23:25:07muesli-sounds good...i would rather high-jack her :D
23:25:12 Join webguest20 [0] (~51429e1d@labb.contactor.se)
23:25:15muesli-err kidnapp
23:25:18muesli--p
23:25:21thegeekhehe
23:25:24thegeekthat too would be easy
23:25:29thegeekshe went running with a personal trainer
23:25:46thegeekher lifeguards where like 100 meters behind her at all times
23:25:46muesli-mmh..i hate jogging
23:25:54thegeekand there was only one of them
23:26:02thegeekbut then again
23:26:10 Join webguest43 [0] (~d567c275@labb.contactor.se)
23:26:16thegeekwho the hell would want to take out the crown-princess of _norway_
23:26:16muesli-you can see gerhard schröder in hannover jogging as well
23:26:16thegeek;)
23:26:25thegeekmhm
23:26:42muesli-didnt say that..rather pull her away ;)
23:26:55thegeekhumhum
23:26:56thegeekas I said
23:27:00thegeekshe's really not that hot
23:27:12muesli-hum
23:27:36 Quit webguest43 (Client Quit)
23:27:39ep0chyou've had better right?
23:27:40muesli-does norway offer anything hot?
23:28:08HClwhy has conversation dropped to this level..
23:28:17muesli-lol
23:28:18thegeekI'm not that one dragging it down
23:28:22thegeek*the
23:28:29muesli-err..ashes on my head
23:28:35muesli-i'm drunken :D
23:28:40thegeekah
23:28:47muesli-m in love with harry..
23:28:56muesli-cant deny it
23:29:02thegeekwhy?
23:29:04muesli-:o0
23:29:12thegeekexcept for the money
23:29:16thegeekI can't see any reason
23:29:27ep0chahhh
23:29:37ep0ch"you're the man now dog!"
23:29:37muesli-reason 4 wot?
23:29:43ep0chytmnd
23:29:48thegeekbtw
23:29:53thegeekdont go to ytmnd.com
23:30:00thegeekif you don't want to know the potter spoiler
23:30:02ep0chshhh
23:30:30muesli-thegeek! detentions!
23:31:00ep0chyeah you might just wanna skip past page 606
23:31:06thegeekindeed
23:31:08muesli-:D
23:31:27muesli-anyway...
23:31:41muesli-anything spectacular new todaY?#
23:31:57muesli-i'm yearning for remote support...
23:32:00muesli-:-//
23:32:28 Part f_x
23:34:04muesli-not :(
23:34:20muesli-will keep on reading harry
23:34:24muesli-cya mates :D
23:34:30ep0chcya
23:36:48 Join MrStaticVoid [0] (~jlee@69-175-94-207.frdrmd.adelphia.net)
23:52:33webguest20hi guys
23:52:42webguest20little question :)
23:53:03webguest20is the peak meter work good now?
23:54:03ep0chno commits since yesterday on ihp, so i doubt it
23:54:30 Quit muesli- (Read error: 113 (No route to host))
23:54:41webguest20do it can produce bug in playback?
23:54:49webguest20i'm afraid
23:55:09BBubno, it works
23:55:21BBubbut its not very qccurate yet
23:55:31BBub*accurate
23:55:35Coldtoastit DOES move tho. heh
23:55:42webguest20ok tanks guys
23:55:57webguest20i'll test it :)
23:56:10ep0chits really not accurate at all
23:57:22CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
23:57:22*HCl should check out a new build and give it some test runs
23:58:42ep0cham so tempted to get an X5...

Previous day | Next day