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).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2005-05-12

00:22:43 Join cool2bdave [0] (
00:23:47 Join ashridah [0] (
00:24:13 Quit lodesi ("Leaving")
00:24:35 Quit Sucka ("a bird in the bush is worth two in your house")
00:30:38HClhow much inch was it again?
00:30:48*HCl might buy an 26" soon..
00:31:07cool2bdaveHi guys, I am suck again trying to compile with cygwin cygwin/rockbox-daily-20050511/build/apps/lang.o: file not recognised : File format not recognised collect2: ld returned 1 exit status {Any Ideas?]
00:31:49***Saving seen data "./dancer.seen"
00:31:49cool2bdaveI just ran make for the fm build
00:34:14HCltry deleting lang.o
00:37:18cool2bdaveit just has the same problem with the next *.o
00:37:29HClthen delete the next .o
00:44:31cool2bdaveok I started again with a new tarball, i get /tools/convbdf: command not found
00:44:49HCldunno :/ sorry
00:44:53HCland i gotta go to sleep :/
00:46:32cool2bdavethanks for trying
00:51:24cool2bdaveI ran make in the tools directory then tried again, I was able to build something
00:53:27 Quit matsl ("Leaving")
01:09:39 Quit cool2bdave ("Leaving")
01:23:04 Join Aditya [0] (
01:34:10 Join XShocK [0] (
01:34:12 Part XShocK
01:40:20 Quit DMJC ("Leaving")
01:54:07 Quit Aditya (Read error: 54 (Connection reset by peer))
01:54:14 Join Aditya [0] (
02:25:43preglowgot my new monitor working in linux
02:26:05ashridahi'm not sure that's as big an acheivement as it used to be ;)
02:26:34preglowit didn't work at first
02:26:37preglowi had to hack the x config file
02:26:51preglowto make it actually use the dvi output
02:27:03ashridahby hack, i presume you mean 'enter sane values for the vert and horizontal refreshes' ?
02:27:12preglowa nvidia driver option
02:27:18ashridahyeah, that can be a little more fun.
02:27:25preglowbesides, it's an lcd panel
02:27:28preglowno refresh values ;)
02:27:40ashridahanyway, i've got to run.
02:27:44preglowme too
02:27:46 Quit ashridah ("uni")
02:27:50 Quit preglow ("lol")
02:31:50***Saving seen data "./dancer.seen"
03:28:41 Quit cYmen (Read error: 113 (No route to host))
03:50:08 Join XShocK_away [0] (
04:02:29 Quit XShocK_away (" HydraIRC -> <- IRC with a difference")
04:05:41 Join QT [0] (as@area51.users.madwifi)
04:16:36 Quit xam ("Chatzilla 0.9.67 [Firefox 1.0.3/20050421]")
04:19:26 Quit QT_ (Read error: 110 (Connection timed out))
04:29:47 Join courtc [0] (
04:31:53***Saving seen data "./dancer.seen"
05:45:27 Join lostlogic [0] (
05:47:44 Join Strath [0] (
06:02:57 Quit Aditya (Read error: 110 (Connection timed out))
06:06:37 Quit StrathAFK (Read error: 110 (Connection timed out))
06:31:54***Saving seen data "./dancer.seen"
06:47:05 Join LinusN [0] (
07:18:09 Quit NibbIer (Read error: 110 (Connection timed out))
07:18:21 Join NibbIer [0] (
08:23:13 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
08:24:04 Join ashridah [0] (
08:31:49 Quit gromit` (Read error: 104 (Connection reset by peer))
08:31:50 Join gromit`` [0] (~gromit`
08:31:55***Saving seen data "./dancer.seen"
08:42:50 Quit gromit`` (Read error: 104 (Connection reset by peer))
08:42:54 Join gromit` [0] (~gromit`
08:48:55 Join Aison [0] (
09:50:24 Join Harpy [0] (
09:52:11 Nick Lynx_awy is now known as Lynx_ (HydraIRC@
09:53:09 Join [IpL]hatto [0] (
09:53:23 Part [IpL]hatto
10:31:56***Saving seen data "./dancer.seen"
10:47:10*rasher debugs fwpatcher.exe
10:49:42rashersomething's wrong..
10:49:51rasherit rejects a properly patched file..
10:52:23rasheruh.. the code seems wrong.. it checks the md5sums one by one, and aborts as soon as it sees one that doesn't match, or am I reading it wrong?
10:57:50Bagderit should rather do the opposite ;-)
10:58:08rasherthat's what I thought
10:58:18rasherstill, one of my tests should have turned out successesful
10:58:47rasheroh, no
10:59:00rasherit would just try the next, and error there
10:59:14rasherI think I may be able to handle this..
10:59:26rasherunless someone else feels like it
10:59:51Bagdernah, you go ahead ;-)
11:00:24 Quit ashridah (Read error: 148 (No route to host))
11:11:43 Join ashridah [0] (
11:14:01amiconnLinusN: Did you read about Jörg's proposal for mutexes in the logs? (A mutex should not block if it is already owned by the same thread)
11:14:20amiconnWhat do you think? Would this be a good idea, or promote sloppy coding?
11:14:54amiconnImplementation should be simple, provided that the mutex functions are moved to thread.c (where they belong anyway imho)
11:18:33rasherfwpatcher worked!
11:19:05rasherexcept not
11:19:51rasheroh, it does
11:20:17rasherdidn't expect patching an already firmware would be succesful, but I can see why
11:21:49t0masamiconn: imo making that mutex non blocking for the thread that locked it would promote very sloppy thread coding...
11:22:19t0masbut maybe there is a good use for it... that I missed
11:22:31rashera compromise would be to detect the condition, and yell at the developer doing it :)
11:23:24*rasher commits fwpatcher bugfix
11:23:57rashertested it on all 1.65/1.63 firmwares, and it produced sane md5sums
11:25:19rasherI should probably test if it detects errors as well.. :-S
11:26:09LinusNamiconn: i'm not sure i want that
11:26:22LinusNlet me think about it
11:26:44rasherexcellent, it detected a corrupt file
11:27:20 Join amiconn_ [0] (
11:28:05rasherheh, viewcvs' diff for my fix looks crazy
11:29:23rasherBut it looks like we have a working fwpatcher
11:30:34rasherworks in wine, even
11:30:54rashernot that there's any reason it shouldn't
11:34:40 Quit amiconn (Read error: 145 (Connection timed out))
11:34:40 Nick amiconn_ is now known as amiconn (
11:34:43 Join [IDC]Dragon [0] (
11:34:56LinusNhow do i build fwpatcher?
11:35:07*[IDC]Dragon read in the log about mutexes
11:35:09rasheryou need mingw
11:35:20rasherand then it's just a case of running make in the dir
11:35:46LinusNafter copying bootloader.bin there...
11:35:52rasherright :)
11:36:16rasherthe 2005-02-11 bootloader.bin that is
11:36:21amiconnLinusN, t0mas: This is the way ordinary windows mutexes work, and pthread mutexes do also know this operating mode (recursive)
11:36:37[IDC]Dragonin order to not promote sloppy coding, we could instead panic if the thread already owns the mutex
11:36:39amiconnCurrently, rockbox mutexes work like 'fast' mutexes
11:37:23*LinusN goes to lunch
11:37:27amiconn[IDC]Dragon: The pthread_mutex_lock manpage is very interesting
11:39:26[IDC]DragonI've seen more interesting reading... ;-)
11:43:35[IDC]Dragonamiconn: on a different subject, I never reviewed you 2GB MMC changes, but it looked like less than I suspected
11:43:59[IDC]Dragondid something work in favour of you?
11:44:17[IDC]Dragonlike, partial block access or so
11:46:18*[IDC]Dragon wonders if there is a MMC to CF adapter
11:46:48[IDC]Dragonso I could use me MMC in my cameras' CF slot
11:49:02amiconnThere are quite some changes, but I streamlined the driver as well
11:49:16amiconnThere is a block cache function now
11:49:54amiconnMost MMCs can do partial block access on read. If they can, it's also enabled by default.
11:50:27amiconnHowever, there are not many cards that can do partial block access on write.
11:50:52amiconnMy guess is that the 2 GB card can, but it's not enabled by default
11:51:20amiconnEnabling it would require programming the CSD, which in turn requires a CRC7 routine
11:52:49amiconnSo my solution to cache whole blocks when dealing with partial block data is not necessarily the fastest, but it's simpler and more compatible
11:53:38 Join Shagnar [0] (
11:54:34Shagnarhiho guys
12:16:23Shagnaris it possible to copy files on the H1x0 with rockbox ?
12:16:48rashernot afaik
12:17:04rasherWhy would you?
12:17:17rasher(not opposing the idea, just wondering)
12:18:30Shagnarin holidays... i could fill my ifp795 (512mb) with my H without a pc (i don't have a pc for the two weeks)
12:18:47Shagnaralso i could organize the songs on my H
12:19:11rasherhow would that work?
12:23:01ashridahShagnar: the H1xx don't have usb-host mode capabilities.
12:23:07ashridahso unless the ifp-795 does...
12:23:11Shagnarit works
12:23:20Shagnarwith a so known "usb-on-the-go-copy-box"
12:23:37Shagnare.g. i can transfer files from my cam on to my H
12:23:44Shagnarthat works, that is not the problem
12:23:44ashridahexcept that requires extra hardware
12:23:44ashridahthe H3xx has it
12:23:46ashridahH1xx's do not
12:24:10Shagnarbut the host function of the H3xx is not as compataible as a copybox
12:24:25Shagnarmy cam isn't supportet by the h3xx, but by the copybox
12:25:08 Join preglow [0] (
12:25:28ashridahwell, usb<->harddrive support is pretty much completely controlled by a hardware chip inside the iriver hardware. rockbox just flips a few bits to enable it, and goes to sleep.
12:25:33ashridahso rockbox can't really do anything about it
12:25:57preglowthat's one of my nicer bugs, no doubt
12:26:03rasherIndeed :)
12:26:23rasherI guess it worked when you had only one md5sum in the list?
12:26:28preglowBagder: i got dead pixels on my 2001fp :/
12:26:41Shagnarthe copybox manages the usb<>hdd support
12:26:46Shagnarthat all works
12:27:05Shagnarmy problem is that i can't copy files on my H
12:27:42rasherwhy do you need to copy?
12:27:52rasherisn't rename(/move) what you want?
12:29:54ashridahpreglow: already? bugger
12:30:19Shagnarrasher its like that:
12:30:48Shagnarby creating a dir named "copybox" in the /root i can also copy vom H140 to any other usb device
12:31:09Shagnarso i oculd copy the songs i want to have on the target in this directory and then copy these files
12:31:26Shagnari hope you understand, my englisch isn't the best ;o)
12:31:50rasherAh, I understand now
12:32:00***Saving seen data "./dancer.seen"
12:32:03rasherdoes copybox remove the contents of this dir+
12:36:17Shagnarjust copies
12:36:48rasherin theory you could move the files there, then back again :)
12:36:48Shagnarwhy? is rockbox able to move files?
12:37:04rasherI forget..
12:37:07Shagnarbut how?
12:37:10rasherhang on.. I'm unsure :)
12:37:54rasherit seems not :)
12:38:47rasheroh, it is
12:38:59rasherrename to '/root/copybox/filename'
12:39:07rasherit's not exactly convenient
12:39:09 Part LinusN
12:39:55amiconnrasher: Someone should write a norton commander plugin...
12:40:34Shagnarrasher what do u mean? didn't undertand really... how can i move files?
12:40:55rasherif you rename a file to something including a path, it will be moved as well
12:42:01Shagnarnow i did understand
12:42:09rasherBut as I said, it's not very practical
12:42:25rasherespecially not when you have to move a lot of files
12:42:40Shagnaryes it isn't. but the files don't get lost or something?
12:43:39rasherno, but I'm not sure how it handles if the path doesn't exist
12:44:35rasherAh, 'rename failed'.. that's fine
12:45:04rasherwhy :(?
12:45:26Shagnartried it but didn't work... hmm encore une fois
12:45:58rasheroh, you shouldn't include /root/
12:46:00rasherjust /
12:46:30rashersorry about that :)
12:46:51Shagnarah fine
12:46:55Shagnarnow it works
12:47:12Shagnarwell, not very practical but it works :) can i also move directories like that?
12:47:19rasherno idea
12:48:13amiconnShagnar: Should work too, just try it
12:48:24Shagnarkk :)
12:48:57rasherdoesn't look like that's worknig
12:49:33Shagnarwell, the most important thing is moving files, and that's working altough
12:51:00Shagnarthanks a lot
13:07:47 Join LinusN [0] (
13:13:27preglowLinusN: would you happen to know how the emac unit does pipelining? i can't find it mentioned in any of the docs
13:13:50LinusNin what respect?
13:15:31preglowi know it pipelines the partial multiplies if i do several mac instructions on the same accumulator
13:15:43preglowbut does it work the same if i work with different accumulators in each instruction?
13:16:36preglowright now i do the latter in windowing, so i can better avoid the accumulator fetch stalling, BUT on the assumption that pipelining works the way i hope
13:17:34LinusNi'm not sure how it works
13:17:52preglowno, and motorola has not seen the need to mention it
13:19:48preglowread it, and it's on the mac arch, not emac
13:20:02preglowso no multiple accumulators
13:22:24preglowbut i see no reason for the pipelining not working with different accumulators, from how the emac unit internals are described
13:22:59LinusNme neither
13:25:09preglowthought i'm starting to think the long multiplies doesn't pipeline the way i thought at all
13:25:34preglow32 bit multiplies reuse the 16 bit multipliers iteratively to produce a 64 bit result
13:25:53preglowi don't see how that can pipeline at all to get a 1 cycle/mac throughput
13:26:05preglowbut that's what motorola claims is possible
13:32:39LinusNso interleaving accumulator usage is pointless then
13:33:31preglowi should do some timing tests
13:36:18preglowgod damn, this monitor makes viewing pdfs even more of a pleasuer
13:36:20preglowpleasure, even
13:37:20LinusN2001fp with dvi...aaaaahh
13:37:42preglowutilized the dvi immediately, yes
13:38:05t0masI still have to buy a DVI cable
13:38:13rasherhow does one create an .ico?
13:38:58LinusNi once heard about this:
13:39:31ashridahrasher: grab kiconedit or himp
13:39:32ashridahgimp even
13:39:47rasherah, gimp
13:39:52rasherthat works
13:40:11ashridahhmm. actually, kiconedit doesn't actually do .ico
13:40:17ashridahuseless thing
13:40:25preglowno, their emac propaganda brochure seems to say the 32x32 bit mul is fully pipelined
13:40:56rasherblack power :D
13:51:20rasherfun, now my rockbox isn't shutting down
13:54:03preglowrasher: mine doesn't either
13:54:09preglowi had it on for eight hours without knowing it
13:54:23rasherit does some times.. I think
13:54:55preglowThe EMAC has a single-clock issue for 32x32-bit multiplication instructions and implements a four-stage execution pipeline.
13:56:22preglowanother source again says the issue rate is 3 clocks for 32x32 bit muls, but that doc is for the mac unit, i suppose it is possible they also did these kinds of enhancements when designing emac
14:00:54rasherLinusN: doesn't support multiple-file icons it seems
14:05:49 Join MoosCamaro [0] (
14:06:03MoosCamaroHi all
14:15:08preglowi'm starting to wonder what the odds of motorola giving away their mp3 decoder source code are
14:15:12preglowi'd really love to it, heh
14:15:29HClwe can always reverse engineer it out
14:15:30 Join DMJC [0] (
14:15:41t0masHCl: mt/archieven/004859.html">
14:16:22preglowHCl: that is such a major bother, and i'd very much rather have it in its commented form anyway
14:16:42HClt0mas: nice.
14:16:48HCl*goes to print and post*
14:16:52preglowi don't see why they have to be so incredibly anal retentive about distributing it
14:16:59preglowbut then again, they are corporate
14:21:29Lynx_preglow: maybe the quality is really bad ;)
14:23:16t0masHCl: forward it to some downloaders around your apartment... the more letters the more fun :P
14:23:59HClt0mas: i have :P
14:24:05HCli sent it to everyone i know XD
14:24:11t0masyeah, I just did :P
14:24:18HClincluding the main irc channel of the university :p
14:24:41*t0mas just emailed it around the office :)
14:25:08Lynx_t0mas: what is it about exaclty? i don't really speak dutch. where to the letters go?
14:25:17HClt0mas: add the date onto the letter
14:25:22t0masto the dutch copyright organiation
14:25:26t0masHCl: yes
14:25:55t0masLynx_: and there is a dutch law, all organisations collecting personal info (like name, email, ip etc..) must send you all they have about you on request
14:25:58t0mas(within 4 weeks)
14:26:07Lynx_ah, ok
14:26:25t0masso everybody sends a letter to the dutch equivalent of the RIAA, and they have to respond in 4 weeks
14:26:27Lynx_maybe we have that in germany too
14:26:38t0mascheck it ;)
14:26:40HCl :P
14:26:51Lynx_what info do you assume they have?
14:26:52HClthis'll be fun.
14:26:56HClthats not the point
14:26:57HClthe point is
14:27:07HClthey have to process 10000's of letters
14:27:09HClin 4 weeks
14:27:14HClor they'll get sued
14:27:14HCl :P
14:27:20Lynx_by whom?
14:27:27HClby us, who're sending the letters :)
14:27:34t0masHCl: no, dutch: "Kanton rechter procedure, met een dwangsom van 1000 euro per week telaat."
14:27:41Lynx_HCl: would you really go throught the trouble of sueing them?">
14:27:44HClt0mas: nice :)
14:27:59HClLynx_: yes, actually, since it would pay me 1000 euro per week, and my dads a lawyer :)
14:28:17Lynx_HCl: hehe. Can germans send a letter, too?
14:28:21HClno :P
14:28:24preglowrasher: where is the source material you use?
14:28:30t0masHCl: if they don't respond... post your dads e-mail address somewhere
14:28:34Lynx_it's like DDOSing by mail...
14:28:38HClit is :P
14:28:39rasherpreglow: the large jpeg in that dir
14:28:46preglowthere isn't a vector source format?
14:28:51t0masHCl: hoofddorp... is that close to you?
14:29:01rasherNot that I know of
14:29:02*t0mas checks train times...
14:29:12t0masdon't want to miss the post truck stopping by there :P
14:29:34rasherpreglow: Maybe if we could find the original author
14:29:40rasherwhoever that might be..
14:31:07rashera vektor version would be very nice
14:32:04***Saving seen data "./dancer.seen"
14:36:17LinusNoriginal author of the rockbox logo?
14:37:22LinusNthomas saeys
14:37:29LinusNa friend of ours
14:37:48rasherof course, assuming that he kept a vector version of it for a couple of years
14:37:48LinusNand i think the original is bitmapped
14:37:56rasherAh, *sniff*
14:38:03LinusNi'll ask him
14:38:50rasherwell I guess it's likely, since the "box" includes bitmap data
14:39:16rasherI have that :)
14:39:30rasherIt's good enough really
14:39:35rasherbut vector would still be nice
14:41:43LinusNshould be enough to make a silly icon :-)
14:41:47preglowtiff doesn't hurt
14:42:06*preglow views rasher-pig.mpg
14:42:34*HCl and markun just bought a battery replacement for the iriver
14:42:42preglowionity one?
14:43:03preglowlet me know how hard it is to install
14:43:16HClLook for item 5186210735 on but as to what brand I can't remember and the ebay description doesn't tell you
14:43:19HClI have an H140 and it was a "snug" fit to say the least but it eventually fitted.
14:43:22HClIf I get 29+ hrs usage then it'll be the best £11 ever spent
14:43:27HCl :)
14:44:10HClconfirmed to last 29hrs :)
14:44:19ashridahHCl: a li-poly or li-ion battery?
14:44:24HCllet me check
14:44:36dwihnohey, what happened to that t-shirt competition? :)
14:44:45ashridahHCl: nice.
14:45:07ShagnarHCl isn't this battery too thick?
14:45:11HCl :)
14:45:24HClthere have been 3 people who've confirmed that it fits (snugly) into the h140
14:45:25HCl :)
14:45:59ashridahmm. fortunately the li-polys tend to be relatively flexible, since they don't have to worry about going up in flames if they're cut (iirc)
14:46:17 Part LinusN
14:46:22Shagnar this one?
14:47:24t0masdoes it fit h120 too?
14:47:32HClShagnar: yea, that one.
14:47:39HClt0mas: dunno, but markun took the risk, at least.
14:47:42HClhe has an 120
14:47:50t0mashm... can you tell me if his works?
14:48:04HClit'll take a while to get to him.
14:48:23ashridahi'm never away from home long enough to need a higher capacity battery, myself.
14:50:29rasherpreglow: any objections to adding one of my icons to fwpatcher (they are 19kb though :-S)
14:52:32 Quit courtc (Read error: 110 (Connection timed out))
14:53:54t0masOk, time to get to the postoffice :P
14:54:40 Quit [IDC]Dragon ("CGI:IRC (EOF)")
14:54:53 Quit lostlogic ("Going to the moon")
14:56:26 Join infernum [0] (
14:57:27preglowrasher: do what you like, i don't claim to own the code, heh
14:57:51preglowit's put in cvs so you can enhance it
14:57:53infernumAnyone knows of a firmware for AV420 ? :/
14:58:30infernumthey rlsed one ?
14:58:39 Quit ashridah ("Leaving")
14:58:42BagderI have no idea
14:58:46BagderI use rockbox
14:59:13infernumfor what client ?
14:59:32t0masLynx_: german people can send letters :D
14:59:42Bagderinfernum: Archos Jukebox 5000, 6000, Studio, Recorder, FM Recorder, Recorder V2 and Ondio MP3 players
14:59:49Bagderand upcoming on iRiver
14:59:58infernumgod damnit
14:59:58Lynx_t0mas: hehe
15:00:02infernumim dying for smthing new for my archos
15:00:06infernummake it work faster and some new features :(
15:00:10t0masas the organisation is recording ip info from all p2p networks, I'm sure they have some german ip's
15:04:54 Join LinusN [0] (
15:06:10 Part infernum
15:06:11 Join infernum [0] (
15:06:48Bagderthey (linav, avos) are just another project that should've ported rockbox instead
15:07:07infernumwhats avos ?
15:07:21Bagderyou're the av owner, not me
15:07:49infernumi dont know about other oses :(
15:08:18LinusNgoogle for "linav" and "avos"
15:08:18Bagderbut you have heard of google, have you?
15:08:23infernumi had
15:08:26infernumcouldent find anything
15:08:32infernumLinav has only AV320
15:08:37preglowavos just links to linav
15:08:37infernumnothing for av420
15:08:58infernumboth devices are so similar why is it such a problem :(
15:08:58Bagderavos was first, I believe they joined forces later on or similar
15:09:31preglowmake them come to the light side!
15:09:42rasherSo which icon is nicer on"> ?
15:10:05preglowi like the Rb ones
15:10:11preglowor do you mean size?
15:10:19rasherthe size is the same
15:10:30infernumwhats that :D
15:10:45infernumin archos site..
15:10:46infernumFOr av420
15:10:46rashersome Microsoft DRM crap
15:10:48infernumwtf thats nice
15:10:54infernumi see windows.. in it..
15:11:22Bagderrasher: I like the Rb ones better too
15:11:31LinusNme too
15:11:31rasherAlso in 16x16 size?
15:11:43Bagderperhaps not
15:11:54preglowthat's too small
15:11:57 Quit gromit` (Remote closed the connection)
15:12:07rasherGuess I could make one that uses the other icon in 16x16 size
15:12:09LinusNwhat are the icons for?
15:12:23rasherand potentially other things, I guess
15:12:34preglowwhich now finally works, it seems
15:12:57preglowafter having pruned away one of my more entertaining bugs
15:12:58rasherSo one using rb for 32x32 and 48x48 and the thingy (wossname?) for 16x16?
15:13:03 Join bobTHC [0] (
15:13:05*preglow applaudes in rasher's direction
15:13:31bobTHChi !
15:13:33*Bagder slaps the "hero of the day" sticker on rasher's forehead
15:13:43infernumah this play for sure is bullshit
15:14:10 Join Nukeleus [0] (
15:14:16Nukeleushello everyone :)
15:15:10BagderI mean ¡hola!
15:15:38*Bagder makes sure quelsaruk is not around
15:16:08preglowi'm always in spaniard detection mode
15:16:35rasheryay, this works
15:16:53rasherUsing rb for 32x32 and 48x48 and the simpler one for 16x16 .. that works nicely
15:16:59preglowi can hear "hola", "pablo" or "señor" being uttered in a five mile radius
15:20:19 Quit infernum (Read error: 131 (Connection reset by peer))
15:21:17Nukeleusrasher what are you doing?
15:22:02rasherCreating a rockbox icon
15:22:13Nukeleusoh i c
15:29:07 Quit Nukeleus ()
15:31:09 Quit bobTHC ("CGI:IRC (Ping timeout)")
15:31:24Bagderall uses of perror() give a warning in the main.c in fwpatcher
15:31:47preglowi know
15:31:57preglowthe entire iriver.c inclusion is a bloody hack
15:32:04preglowsomeone might want to fix that
15:32:34Bagderah, right
15:32:44preglowthe code never was intended to be used like it is
15:32:51preglowit's intended to be used by a command line tool
15:33:43preglowand seeing as how i seem to be the only codec guy, i think i'll use my time on that instead
15:33:59rasheryou do that
15:34:29preglowright now i'm battling with windows again
15:36:10rasherInteresting.. when rockbox wasn't shutting down, touching a button was enough
15:36:18rasherto make it shutdown 10 minutes later
15:36:33rasherI think..
15:36:51 Join bobTHC [0] (
15:36:52preglowthe last thing i do before i don't touch it anymore, is loading a plugin
15:52:47 Join ripnetuk [0] (
15:54:05rasherafternoon, ripnetuk
15:56:49ripnetukhows life in rockbox land?
15:56:57rasherfun fun fun
15:57:00LinusNsad and boring
15:57:08ripnetuk:( cheer up
15:57:13ripnetuki have a cold
15:57:29ripnetukhey - is the bdm port on the iRiver a jtag port? or somethign else?
15:57:38LinusNit's a bdm port
15:57:46ripnetukso thats a different standard to jtag?
15:57:59t0masnever expected a BDM port to be a BDM port :|
15:58:00LinusNit's a motorola specific thing
15:58:12LinusNmost cpu's have similar things
15:58:16ripnetuktOmas - you live and learn :)
15:58:48ripnetuku ever worked with jtag?
15:59:19Bagdernot really "worked with", just used
15:59:22preglowi've used it a couple of times
15:59:59ripnetukim guessing you only know which pins are the jtag interface on a 'unknown' device by either them being marked on the PCB or documentation
16:00:53preglowi sure as hell wouldn't know any other way
16:01:17ripnetukhmmm... i was hoping for something like 'it kicks out a xxMHz square wave when not being used" - wishful thinking
16:01:25ripnetuki guess it wouldnt :)
16:01:54ripnetukthe reason i ask is because I have a shiny new sony PSP that is dying to run some homebrew apps, but its locked (code signing etc)
16:02:12rasherdidn't the psp get cracked recently=
16:02:23LinusNthe japanese version iirc
16:02:33ripnetukthey found an explit on the version 1.0 firmware, but the newer firmware is locked down
16:03:23ripnetukdammit - if they allowed (small) unsigned binaries to run, it would be much less likely that people would try and break the protection
16:05:04preglowit would also make it far easier to break the protection
16:06:18 Join Aditya [0] (
16:06:40 Quit Rick (Read error: 131 (Connection reset by peer))
16:06:50 Join Gibbed [0] (
16:10:43*HCl posted letter :)
16:12:29 Quit bobTHC ("CGI:IRC (Ping timeout)")
16:13:51*ripnetuk wishes i had time to port Rockbox to the pocketPC platform so it ran on my SPV smartphone... MEdia Player is a POS
16:15:37HClthere are plenty of good mp3 players for pocketpc O.o.
16:15:47ripnetukyeah, but me want rockbox
16:16:40*HCl fails due to a logic error :/
16:18:56*preglow puts HCl on his feet again
16:19:31HClthanks :x
16:22:33 Part LinusN
16:22:52Adityamedia player is not a piece of shit
16:23:06AdityaIMHO WMP is quite well written
16:24:44preglowno gapless playback :/
16:26:29Adityaif a song ends it ends.. if the artist put in extra space at the end. I want it there
16:26:33Adityaits usually for a purpose
16:26:37preglowwell, yes...
16:26:40preglowbut i don't want extra gaps
16:26:47AdityaI dont have extra gaps..
16:26:54preglowwmp makes it for you
16:27:03preglowlast time i checked, wmp inserted gaps between all tracks
16:27:19AdityaI dont know what mutated version you are using..
16:27:24preglowait, i can check again
16:29:06preglowbloody media guide...
16:29:42preglowputs a nice little gap between two songs that should not have a gap between them
16:29:48preglowgranted, it's not as big as some other players
16:30:01Adityaeh.. its nice
16:30:15preglownothing that inserts gaps between my tracks is nice
16:30:18Adityawinamp is better in some things.. plugins generally, but WMP is not a POS player
16:30:22HCli remember winamp
16:30:26preglowi hate winamp
16:30:31HCli actually couldn't get it to insert gaps properly.
16:30:38preglow_insert_ gaps? :P
16:30:42Adityafoobar is best if you are just listening to music
16:30:49HClwhen i got a metal song and a soft song.
16:30:54HCli don't want them bumped right on top of eachother
16:30:57HClbut i want a gap between them.
16:31:09preglowAditya: what other things do you want to do?
16:31:20HClwork with my remote
16:32:08***Saving seen data "./dancer.seen"
16:37:58HCland do the dishes, ofcourse.
16:50:09AdityaI still keep wmp and winamp around for videos
16:50:12Adityaand DVDs
16:50:35preglowi use media player classic for that
16:50:59HClmplayer for windows! :d
16:51:15preglowmplayer is ok
16:51:23preglowbut i prefer something a bit less command line for stuff like this
16:51:43HClyou can drag files on top of a shortcut to it
16:51:44HClworks fine
16:52:25AdityaI don't have anything against any software.. they are merely tools
16:52:31Adityayou use what you use
16:55:44ripnetukgotta go... talk to ya later
16:56:17 Quit ripnetuk ("Leaving")
16:56:39 Quit thegeek (Read error: 131 (Connection reset by peer))
17:01:16 Join xam [0] (
17:08:05 Nick Gibbed is now known as Rick (
17:12:12 Join TCK- [0] (
17:20:35 Quit TCK (Read error: 145 (Connection timed out))
18:00:58 Join TCK [0] (
18:01:51 Quit MoosCamaro ()
18:07:50 Quit xam ("Chatzilla 0.9.67 [Firefox 1.0.3/20050421]")
18:19:29 Quit Shagnar (Read error: 110 (Connection timed out))
18:19:47 Quit TCK- (Read error: 110 (Connection timed out))
18:32:10***Saving seen data "./dancer.seen"
18:55:10 Join courtc [0] (
18:56:10 Nick Lynx_ is now known as Lynx_awy (HydraIRC@
18:58:53 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
19:31:13 Nick tvelocity[away] is now known as tvelocity (
19:33:24 Join Sucka [0] (
19:58:58HClguess what :3
19:59:08*HCl has 4 days of free time and is bored :x
20:01:31*preglow coughs 'dumb! dumb!'
20:02:24HCl :p
20:02:31HClexactly :p
20:02:35HClbut first watching some tv
20:11:45 Join courtc_ [0] (
20:12:39 Quit courtc (Read error: 110 (Connection timed out))
20:12:50 Nick courtc_ is now known as courtc (
20:17:24 Nick tvelocity is now known as tvelocity[away] (
20:32:12***Saving seen data "./dancer.seen"
20:37:09 Join Stryke` [0] (~Chairman8@resnet-241-86.resnet.UMBC.EDU)
20:38:08HClwhere have my dumb warnings gone...
20:39:39*HCl starts by making an mod2wav
20:45:11preglowyes, as long as you know it won't be usable
20:45:21preglowunless you make some math routines, that is
20:48:07HClyea yea.. :/
20:48:16HClat least the other day i came across a way to do sqrt()
20:48:35HCli'll start by converting the floats.
20:50:02 Join MoosCamaro [0] (
20:52:45HClno typedefs in rockbox, right?
20:58:58preglowoh, there are plenty of ways
20:59:23preglowdon't know about typedefs
21:00:27preglowthe stupid way about doing the float->int conversion first, is that you don't know if the code worked in the first place
21:01:48amiconnIt should be possible to test that. Before doing the conversion, write the test plugin and build on the simulator (with math support)
21:02:36HClpreglow: i tested the code with a linux compile, it worked.
21:02:53HCli'm planning to make a fixed point linux version first.
21:03:02HClit should be fairly trivial to port that to rockbox then.
21:10:02 Quit Aditya (Read error: 131 (Connection reset by peer))
21:19:20HClit shouldn't have compiled already, why did it >.<
21:23:36preglowfixed point linux port first might be a good idea
21:23:53preglowif you're particularily clever, you'll write a set of macros that closely mimic the emac unit, and try to work with those
21:26:27*HCl spots pi
21:26:32HCli don't know the emac unit, sorry
21:27:34HClour approximation of pi is gonna be horrid though.
21:28:00HCldunno how much quality loss that'll cause.
21:28:11HClnot many bits for behind the fixed point dot.
21:28:30HClit'll be something like 3.141 :/
21:28:31preglowwell, that depends on the range you'll use
21:28:48preglowyou don't HAVE to use the same fixed point format everywhere
21:28:54preglowas long as you scale stuff correctly
21:29:07HCldoubling the amount of bits would still give 3.14159 :/
21:29:14HClwe might get away with it anyways.
21:30:36preglowyou should review the code completely to see where the largest scale is
21:30:52preglowand use that for deciding how many bits to use
21:31:23preglowbut yeah, i sincerely doubt you'll end up using pi in fixed point code anyway
21:31:40preglowfunctions like log, exp should be implemented as lookup tables
21:31:51preglowpossibly with interpolation
21:35:30preglowforget what i said about scaling, i suck
21:36:03*amiconn spotted an easy-to-fix bug in the tracker
21:36:07preglowbut anyway, emac can do both 32x32 bit muls giving the lower and the higher 32 bits
21:36:10amiconnStill a quite old one :/
21:36:45preglowit actually gives you the higher bits of x*y << 1
21:36:54preglowand operands are signed
21:47:52*HCl screams at the horrible floats
21:48:46 Join einhirn [0] (
21:58:56preglowwhat's up with the comments in video.c? i thought rockbox used /* */
21:59:09*t0mas is going to bed...
21:59:10t0masgood night :)
21:59:15 Part t0mas
21:59:27amiconnpreglow: I know. Blame [IDC]Dragon ;)
21:59:47preglowhaha, i don't care, i like using single line comments
22:00:36amiconnI don't like them that much
22:01:07amiconn...and in fact, I like it that rockbox style is to not use them
22:01:10HCli use them for what they are meant for, - single lines :/
22:01:12HClbut yea
22:01:16HCli need to convert mine too.
22:01:18amiconnThat way I can use them as debug comments
22:02:02amiconnThey look so C++ish to me....
22:02:27preglowis it custom to hate c++ around here, or something?
22:02:34 Join lodesi [0] (~Unknown@
22:02:48amiconnI don't know much about C++
22:04:04preglowi use it for most larger projects
22:04:58preglowi've gotten pretty used to it, to the point that i sincerely hate having to deallocate stuff by hand ;)
22:05:09HClfwah. i want a way to tell gcc that type fixed is an int but type conversions aren't allowed from ints :/
22:05:12amiconnI'm not that much in oo
22:05:46*HCl doesn't know how one would code flexibly without oo :/
22:05:54HClanyways, *goes back to dumb*
22:05:58preglowi'm not that much of an oo person myself
22:06:08preglowbut i do use it
22:10:35 Join TCK- [0] (
22:18:32 Nick tvelocity[away] is now known as tvelocity (
22:19:44 Join thegeek [0] (
22:20:34HCli use oo when speed isn't very important.
22:21:12preglowthere's nothing that says oo is slow
22:22:17HCli know from experience that it often is :)
22:22:37HClsimply due to more functions getting called cause of the oo structure
22:24:05preglowif you inline the accessors and stuff, you shouldn't fare too bad
22:24:19HClyea, i guess you're right about that
22:24:22HClinlining would help
22:24:33preglowoo depends heavily on inlinging
22:24:44preglowfor stuff like overloaded operators and accessors
22:24:58preglowgcc 3.4 bugged on me and stopped inlining operator[] for vectors once
22:25:09preglowi believe you can imagine what a performance hit that gave
22:25:47HClconverting dumb to fixed point is making me want to scream.
22:25:55preglowjust ask if there's some difficulties
22:26:00HCli think there are gonna be soooooooooooooooooo many bugs due to places i've overlooked
22:26:05preglowof course
22:26:07HCli'm going off
22:26:09preglowdo it in pieces
22:26:13preglowand verify each stage
22:26:24preglowas long as that is possible
22:26:28HClnot very :/
22:26:39preglowif i don't remember incorrectly, the only things that are float are volume and delteas
22:26:47preglowvolume is a no-effort port
22:26:48HCli don't know whether functions return an int or a float, a float needs conversion where an int doesn't if its just multiplication..
22:26:51preglowso you're stuck with delta
22:27:08HClvolume is a no effort port?
22:27:13HClthen can you do that for me first? heh.
22:27:15preglowvolume is just a multiply
22:27:21HClmost of what i've done so far is volume and delta, yea.
22:27:39preglowwhat range is volume intended to have?
22:28:03 Quit TCK (Read error: 110 (Connection timed out))
22:29:24HClno idea.
22:29:26HCloh wait
22:29:27HClno i know.
22:29:34HCl0.0 - 1.0
22:29:46preglowthen simple
22:30:01preglowthe emac units native format is +/-0.31 fixed point
22:30:03preglowwhich is what that is
22:30:54preglowso whenever you need to scale by volume, you just do a simple mac.l data, volume, acc0
22:31:03preglowand the answer will be ready to be used in acc0
22:32:13***Saving seen data "./dancer.seen"
22:32:22HCli'll leave that to you :X :X :X
22:32:41preglowbut to make that job simple on me, you really need to consider it now
22:33:04preglowit's really quite simple, you just need to know what operations the emac can do, and stick to that
22:36:49HCli think i finished.
22:37:22preglowand all the emac unit can do is accx += a*b, if it's in integer mode
22:37:31preglowor accx += a*b << 1, if it's in fractional mode
22:37:54preglowin integer mode you can only fetch the lower 32 bit of the accumulator
22:38:01preglowand in fractional mode you can only get the upper 32 bits
22:39:01preglowyou'll probably mostly be interested in the upper bits, so you'll probably need fractional mode
22:39:21preglowoh, it can also do accx -= a*b (<< 1) in both modes
22:40:40*HCl smiles and nods, having no clue what preglow is trying to say
22:41:29preglowwhat, what's incomprehensible?
22:42:33HCldon't know how to apply that to the existing code :/
22:44:04preglowthen you're going to have a hard time :P
22:44:22HCli'll leave that to you :(
22:44:35preglowbut no, it's of course possible to port it fixed point knowing nothing of that
22:44:37HCli'll be happy if it still plays recognisable sounds after the conversion to fixed point
22:44:57preglowbut then parts of the job will have to be done again when trying to use the emac unit
22:45:12preglowbut sure
22:45:22preglowif it's a first attempt at fixed point porting, don't worry about anything
22:45:26 Nick tvelocity is now known as tvelocity[away] (
22:46:41 Join muesli- [0] (
22:46:47preglowwill probably take you a wee bit time to get intimate with it
22:47:24preglowi really look forward to playing xms on my h120
22:47:50muesli-whats xms?
22:48:17preglow.xm files
22:48:42muesli-what are they good for
22:48:55preglowthey're a tracked music format
22:49:11preglowthe file contains a bunch of sounds, and instructions on how to paste them together to a piece of music
22:49:31muesli-never heard of them
22:49:34HCli give up x.x;
22:49:35preglowthen you're missing out
22:49:41preglowHCl: that was quick
22:49:42HCl"warning: division by 0"
22:49:57HClyea, well. i don't have enough fixed point stuff to prevent divisions by 0
22:50:02HCland i don't want to have to deal with that u.u
22:50:12HCl :(
22:50:14preglowyou do of course have to have an if around divs
22:50:19preglowunless you know they'll never be zero
22:50:30HCli KNOW this will ALWAYS be 0
22:50:36HClbecause its simply too small
22:50:38HClto represent
22:50:39preglowthen you're doing something wwrong
22:50:44HCl :(
22:50:49preglowuse more frac bits
22:50:58preglowwhere is this, btw?
22:51:03HClthen i'll get trouble when multiplying them..
22:51:12HCl playing->delta = DIVFX(FTOFX(1.0/65536.0),playing->delta);
22:51:17preglowyou need to shift before multiplying
22:51:27HCli can't, then i lose precision.
22:51:31preglowof course
22:51:32preglowbut that's ok
22:51:37HClugh :(
22:51:39HCli give up :(
22:51:41preglowthere's nothing that says you can never lose precision
22:51:51HClyou can look at what i got so far if you want :/
22:51:54HClshall i put it online? :(
22:51:55preglowlibmad in it's current incarnation looses about 16 bits of precision everywhere
22:52:05preglownah, if i'll do it i'll do it from scratch
22:52:14HCl :(
22:52:19preglowthe problem being i can't do this until late in the summer
22:52:26preglowi need to concentrate on my master's stuff right now
22:52:42preglowwhich i am already having a hard time in doing
22:53:03HClyea, i don't blame you.
22:53:08HCli'm just not very good at math :/
22:53:23preglowdoing fixed point is really more like being good with bits
22:53:44preglowthe math is the same, you just need to make sure there is always enough precision to represent what you want
22:54:02HClwell, this delta is gonna become 0, there's no way i'll have enough precision for it..
22:54:43preglowwhat's going on in that code, btw? you're going to divide playing->delta by 1.0/65536 ?
22:54:57HClno, we divide 1.0/65536 by delta
22:55:00HCland store it in delta
22:55:30preglowahh, yes, i meant that :-)
22:55:48preglowwhy is that being done?
22:57:05HCli don't know :(
22:57:11HClbut its screwing my fixed point totally over.
22:57:19crashdhey guys
22:57:30preglowfixed point divides should be avoided at all costs anyway
22:57:31crashdsomeone told me a few days ago, but does anyone have an idea on which platform the h10 runs on ?
22:57:41preglowcrashd: portalplayer
22:57:47crashdthat's the trick :]
22:57:50crashdany indy dev for it ?
22:58:01preglowipodlinux is probably the closest at the moment
22:58:12*HCl goes to clean the deepyfrier...
22:58:12preglowsince that too uses a portalplayer chipset
22:58:19crashdyeah, same/similar hardware
22:59:18 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
23:03:02crashdpreglow: but there's no actual work on the h10 yet then ? :\
23:04:17Bagdercrashd: nopes
23:04:39preglowwe're more than busy enough with h1x0
23:04:45crashd;) hehe
23:04:45preglowand h3x0, really
23:04:49crashdso i can see
23:04:50preglowdon't think too much code will differ
23:05:12crashdi'd love to contribute, but im not really au faix with the hardware system and ive got a shitload on at work
23:05:12preglowonce we get codecs ported, and a playback system up, porting to new archs will be easier still
23:05:27preglowsure, no worries
23:05:38crashdmaybe one day ;)
23:05:44preglowyou're welcome anyday
23:05:45Bagderwith the codecs in place, we have a solid ground for basically all players
23:06:01Bagderthe only tricky part is the dealing with various button setups
23:06:08preglowahh, yes
23:07:12preglowit's a pity there aren't more people interested in codec work :/
23:07:16preglowthere's a bunch of work to be done
23:07:26preglowbut that might happen once we have playback
23:07:42preglowsince there's a good reason to get stuff done, then
23:07:57Bagderonce it starts playing music...
23:08:58preglowoh, how these green pixels are mocking me
23:11:07muesli-l8er girlz
23:39:22 Quit lodesi ("Leaving")
23:41:19 Quit Harpy (Read error: 145 (Connection timed out))
23:43:31preglowHCl: the dumb author really does seem to enjoy the float dynamic range, yes...

Previous day | Next day