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 2006-02-06

00:01:51saratogahow useful is the 3G port likely to be?
00:02:01saratogafrom what I hear, the 5002 is pretty slow
00:02:08 Quit S3TUP ("CGI:IRC (EOF)")
00:02:34linuxstbWe won't know until we test it.
00:03:51saratogawhat was the issue with the 5002 anyway?
00:03:53linuxstbThe slow part of the 5002 is the cache, but Rockbox is optimised to put time-critical code and data in fast internal RAM - which isn't affected by the cache.
00:03:59saratogaI read something about it having very slow access to main memory
00:04:08saratogahaha
00:04:11saratogathanks
00:04:34saratogaoh you mean the Icache is slow, not the DRAM?
00:05:18linuxstbThis is what the ipodlinux Wiki says: "The 5002 has a "broken" cache (1 wait state per access for program or data, meaning you effectively have half the effective clock rate when running code from external memory). This means that running code that doesn't fit in the internal 96kbyte SRAM of the player is very inefficient, both in terms of CPU cycles and power."
00:06:18saratogaok, that sounds like the i-cache then since there is no d-cache
00:06:38 Join Paul_The_Nerd [0] (n=Paul_The@cpe-66-68-93-2.austin.res.rr.com)
00:06:53linuxstbThe ipods have a unified cache.
00:07:03saratogawhat?
00:07:13saratogai thought there was no caching at all of data
00:07:15imphasingThere is a small 96k cache directly on the processor though, that is very fast.
00:07:28saratogathats the i-ram, not cache (I think?)
00:07:35imphasingEh..it's possible.
00:07:36imphasing:D
00:07:40linuxstbsaratoga: Yes, each of the two ARM cores has its own 8KB unified (code and data) cache.
00:07:46saratogaoh ok
00:08:13linuxstbPlus there is 96KB of fast IRAM which isn't cached (it doesn't need to be).
00:08:19 Quit Therx ("May be back later")
00:08:21imphasingAh..
00:08:26linuxstbBoth cores can access that IRAM.
00:09:43saratogais there any documentation at all for the ARM cores on the 5002?
00:09:47amiconnAny idea about the cache organisation?
00:09:52saratogaI've never done ARM and need to brush up on it
00:09:59imphasingYeah, there's quite a bit of info out there.
00:10:07linuxstbYes - there is lots of info - they are standard arm7tdmi cores
00:10:07amiconn(direct mapped/set associative/fully associative ?)
00:10:13imphasingThere's some really freat pamphlets about it, that have everything you need to know.
00:10:14Bagdersaratoga: its an arm7tdmi core, and there's docs for that on arm.com
00:10:17imphasings/freat/great/
00:10:18saratogais tha core standized?
00:10:25linuxstbamiconn: I have no idea.
00:10:36imphasingstandardized?
00:10:43saratogaIE: all arm7tdmi cores have the same latencies, etc
00:10:49imphasingAh..
00:10:52imphasingI'm not sure..
00:11:29josh_http://www.get-linux.org/~oremanj/t/arm7tdmi.pdf <−− for your reading pleasure
00:11:35imphasingI've got a ~250 page doc on the arm7tdmi that explains everything in great depth.
00:11:37imphasingThat's the one.
00:11:38imphasing:D
00:11:42imphasingjosh_++
00:11:44saratogasweet
00:15:25josh_btw, TDMI stands for Thumb/Debugger/Multiplier/ICE
00:15:50imphasing(!)
00:15:51josh_or maybe /Interrupt
00:15:53imphasingIt's ICE?
00:15:58imphasingAh..
00:16:05josh_depending on who you ask
00:16:20imphasingHeh..
00:18:31 Join TCK [0] (n=tckocr@81-178-216-103.dsl.pipex.com)
00:18:47 Quit Thus0 ("Leaving")
00:21:07linuxstbsaratoga: I assume you have a 3G?
00:21:42saratogayeah
00:22:23linuxstbPlanning to help with Rockbox? :)
00:22:40saratogaI'd like to
00:22:57saratogajust limited on time until I graduate in a couple months
00:23:06saratogaso I'm just reading for now
00:23:40imphasinggraduate == college?
00:23:55saratogayeah
00:24:04imphasingCS?
00:24:04imphasing:D
00:24:19saratogaI'm doubleing in CS and CE
00:24:29imphasingAh, cool.
00:24:44saratogarockbox is the kind of thing that interests me :)
00:24:56imphasingYeah, me too.
00:25:12imphasingMostly because it's a proprietary OS
00:25:17imphasingWhich is beyond cool.
00:25:20Bagderand I'm only here to look cool ;-)
00:25:40josh_imphasing: proprietary != unique
00:25:56josh_Rockbox is open-source, definitely *NOT* proprietary.
00:25:57imphasing=./
00:26:04imphasingNo?
00:26:04josh_And, in general, proprietary is *NOT* "beyond cool'.
00:26:15josh_Proprietary is something like Windows.
00:26:24saratogai think its interesting just because its simple enough that you can understand what its doing (well aside from a lot of the audio stuff)
00:26:28imphasingIt doesn't have to mean "commercial" though
00:26:36josh_imphasing: actually, it does.
00:26:39imphasingOh man..
00:26:41imphasingYou're right.
00:26:48imphasingJust googled it.
00:26:49imphasing:(
00:26:50josh_proprietary means one entity owns it and controls it
00:26:53saratogahaha
00:26:53imphasingyeah
00:26:55josh_which is not the case with open-source software
00:27:21hyarionjosh_ 1, imphasing 0
00:27:49imphasingmore like, josh_ 2385 x 10 ^ 3785, imphasing 0
00:27:56hyarion:)
00:28:09josh_haha
00:29:20saratogafor now i'll keep an eye on the 3G progress
00:29:28saratogahopefully I'll be able to start playing with it soon
00:29:29josh_(10 ^ 3785) is 0b111011001001 xor 0b000000001010, which is 0b111011000011, which is 3779 :P
00:29:45imphasing=/
00:29:50hyarionbtw, I looked at the data sheets for HD66789 is it really compatible with HD66735?
00:29:50imphasingYou suck.
00:30:19imphasingIn this case, '^' denotes the exponential operator.
00:30:19imphasing:P
00:30:20hyarionI read it was on either ipl or rockbox, but it looks like there is differeces
00:30:24josh_ah :P
00:31:00amiconn.
00:31:11 Quit Rondom ("I'm leaving on a jetplane, don't know when I'll be back again...")
00:31:17hydrahead3G started working? Is there gonna be a greyscale 4g soon for testing?
00:31:29imphasingThe bootloader is mostly working.
00:31:48saratogatheres a binary posted an hour or two ago for the 4G, but it crashes
00:31:59saratogaso probably soon if someone with a 4G starts testing
00:32:08imphasingThis binary is just the bootloader though, right?
00:32:23saratogai assumed it was all of rockbox, since I think it finally compiles
00:32:30saratogabut I didn't check since I don't have a 4G
00:32:34imphasingAh..
00:32:38imphasingWas it from cvs?
00:32:53hydraheadwell i can test but I dont know the first thing about debugging software :(
00:33:07linuxstbhydrahead: Can you read C?
00:33:39hydraheadnot really, only messed with basic and pascal in highschool
00:34:06imphasinglinuxstb: Did something largish happen while I was sleeping? As in, getting the "normal" build to compile for 3G/4G?
00:34:12linuxstbThe 4G greyscale port needs someone to compare the code in Rockbox with the code in ipodlinux and fix any differences. At the moment, the 4G greyscale build is just an iPod Photo build using the 3G's LCD driver.
00:34:39linuxstbBut I would guess it's just the LCD driver that has any differences.
00:35:20linuxstbimphasing: Yes. slimx is currently debugging the 3G build (it crashes after displaying the Rockbox splash logo). The 4G build crashes even earlier than that. But they both compile.
00:35:25hydraheadsince the processors are the same, thats a good guess
00:35:26imphasingjosh_: Do you know what file the LCD driver is in the ipl tree?
00:35:32imphasinglinuxstb: Woo!
00:35:41imphasingIs it in CVS?
00:35:43josh_imphasing: linux/arch/armnommu/mach-ipod/fb.c I believe
00:35:48imphasingAh, ok.
00:35:51josh_linuxstb: yeah, 4g crashes before doing *anything*
00:35:53linuxstbimphasing: It's in various places - I use the one josh_ just mentioned.
00:35:56josh_well, freezes
00:36:08josh_linuxstb: there's also some LCD initing you need to do
00:36:14imphasingI'll poke it with my ipod and see what happens.
00:36:17hydraheadlinuxstb: and what do you need to view C ?
00:36:28saratogatext editor
00:36:34josh_linuxstb: check in ipodloader2/fb.c
00:36:38saratogaon Windows, Textpad is pretty good
00:36:39linuxstbjosh_: Yes, I noticed some differences - but I'm just going to leave it to a 4g owner to fix.
00:36:40josh_at the very end (fb_init())
00:37:57hydraheadaha, well I guess that I could try to see any syntax differences but i wouldnt know how to fix them since i propapbly dont understand them
00:38:29imphasingHeh, leave it to me to bork my whole tree, and forget to make a backup before I start editing things..
00:38:32linuxstbI'll have a quick look at the lcd code now then, and fix any obvious differences.
00:39:32imphasingAugh, lots of hunk failed's on slimx's latest patch..
00:39:39hydraheadace! anyone working on the variable cpu speed?
00:39:42slimxi've corrected the lcd driver
00:39:53slimxnow rockbox logo is nice to see
00:39:58slimx:)
00:40:00imphasing:)
00:40:13imphasingYour patch doesn't apply cleanly though
00:40:14imphasing:(
00:40:30slimximphasing, there is no patch at the moment
00:40:45slimxnow consider the ipod 3g patch as closed
00:40:49slimxall is in cvs
00:41:19imphasingYay!
00:41:44Bagderslimx: nice work on the 3g port btw
00:43:13slimx:) thx
00:43:15 Quit Strath (Read error: 104 (Connection reset by peer))
00:43:56slimxthx also to linuxstb nothing would had been possible without his help
00:43:58slimx:)
00:43:59saratogawill it actually boot, or does it still crash?
00:44:05Moosslimx: oui félicitations
00:44:32slimxh h un autre frenchi dans les rangs
00:45:04MoosOui ;)
00:45:13Bagderje sui un baguette
00:45:20Moosohoh :)
00:45:33MoosHi Bagder btw
00:45:36Bagder:-)
00:45:39Bagderevening
00:46:18amiconnBagder: I hope the X5 button assignments for X5 are usable.
00:46:35 Quit actionshrimp (Read error: 104 (Connection reset by peer))
00:46:35amiconnI don't know anything about the constraints
00:46:52Bagderwhat are you saying?
00:46:54amiconnhrm, for mandelbrot I mean :)
00:46:58Bagderah
00:47:08Bagderwell, I don't think anyone knows the constraints yet
00:47:13saratogaam I updating from CVS right? cvs update -dP
00:47:14Bagderor perhaps Linus knows something
00:47:20saratogait seems to grab too many files
00:47:43 Join Strath [0] (n=mike@dpc67143207026.direcpc.com)
00:47:45Bagderyou do get quite a bunch
00:47:57amiconnMost targets have constraints for pressing multiple buttons
00:48:03saratogawell yeah, but i just updated it today, and it still seemed to pull a lot of stuff through
00:48:05amiconn(both mechanical and electrical)
00:48:10saratogai mean it should only grab a handful of things right
00:48:48Bagdersaratoga: no, you get the whole lot
00:48:49saratogaand I get a lot of "it is in the way" messages as I go
00:48:53Bagderoh
00:48:57Bagderthat is weird
00:48:59*amiconn is doing fancy asm stuff
00:49:16amiconnLots of similar code, have to take care to not mix up things
00:49:17saratogafor instance: "cvs update: move away uisimulator/sdl/uisdl.c; it is in the way"
00:49:25CassandraAha! Victory!
00:49:40linuxstbjosh_: Can you try this build? The LCD driver was completely wrong before - I'm not surprised nothing happened. http://www.davechapman.f2s.com/rockbox/rockbox-4g.zip
00:49:43*amiconn wants faaaast scrolling :)
00:49:44CassandraNow does anyone know where I can find a high res version of the Rockbox logo?
00:49:59Bagderyes
00:50:01Bagderhang on
00:50:09linuxstbI only know this one: http://www.rockbox.org/rockbox400.png
00:50:36*Cassandra is physically sickened by the gross hack used to shoehorn the manual into the build system.
00:50:42CassandraNever mind though, it works.
00:50:42Bagderhttp://www.rockbox.org/viewcvs.cgi/*checkout*/www/tshirt-contest/Attic/rockbox3540.jpg
00:50:56linuxstbAlso, rasher made an svg version: rasher.dk/rockbox/rockbox.svg">http://www.rasher.dk/rockbox/rockbox.svg (I've no idea how good it is)
00:50:57Bagder3540x1096
00:51:08CassandraProbably doesn't help that I've never really understood make in detail.
00:51:14CassandraThanks Bagder
00:51:36CassandraDon't we have one in a lossless form?
00:51:52Bagderwe might somewhere
00:52:18CassandraThat would be better, if it's locateable.
00:52:28linuxstbhydrahead: You could try this bootloader - http://www.davechapman.g2s.com/rockbox/bootloader-4g.bin
00:52:42CassandraI can easily convert that jpeg to png, but it kind of defeats the point of using png.
00:53:06hydraheadumm ok, just point me to some sort of an istruction for installing it
00:53:14hydrahead*instruction
00:53:32linuxstbhttp://www.rockbox.org/twiki/bin/view/Main/IpodInstallation
00:53:54hydraheadThanks! I'm on it
00:54:20 Join JdGordon [0] (n=Miranda@c220-239-137-122.smelb1.vic.optusnet.com.au)
00:54:40linuxstbMake sure you keep a copy of the "bootpartition.bin" you make during the installation. You can then simply restore that file back to your ipod if the bootloader doesn't work (or if you ever want to uninstall Rockbox).
00:54:56JdGordonmorning all
00:55:40 Quit hydrahead ("CGI:IRC (EOF)")
00:56:11 Join toology [0] (n=5774a9c7@labb.contactor.se)
00:56:28 Nick toology is now known as hydrahead (n=5774a9c7@labb.contactor.se)
00:57:29slimxprefetch abort ...
00:57:29 Quit saratoga ("CGI:IRC (EOF)")
00:57:31linuxstbhydrahead: Sorry, I mistyped the URL: http://www.davechapman.f2s.com/rockbox/bootloader-4g.bin
00:57:36 Join LinusN [0] (n=linus@labb.contactor.se)
00:57:41slimxat C4000002
00:57:49hydraheadcould you just repeat what file i should back up?
00:58:10slimxsomething is really strange :p
00:58:20 Join jmg [0] (n=Jernau@shinobi.thoughtcrime.org.nz)
00:58:23linuxstbIt's bootpartition.bin - no need to back it up, just don't delete it.
00:58:25jmghi all
00:58:35jmgi just got an archos pma400
00:58:52jmg(not currently supported by rockbox)
00:59:00 Join DMJC [0] (n=James@220-245-174-43-sa-pppoe.tpgi.com.au)
00:59:08DMJCanyone else build rockbox on linux?
00:59:18linuxstbslimx: Have you made any changes to firmware/crt0.S ?
00:59:19BagderDMJC: lots of us do
00:59:24 Quit mikearthur (Remote closed the connection)
00:59:27DMJCcool
00:59:44slimxthe bootloader loop infinitely till i press a button :p
00:59:48DMJCneed to setup my build environment again
00:59:54DMJCbeen a long time since I last did it
01:00
01:00:04 Join mikearthur [0] (i=mike@82-41-205-190.cable.ubr11.edin.blueyonder.co.uk)
01:00:14*Cassandra sighs. Google tells me there used to be an image at http://daniel.haxx.se/logo-contest/rockbox3540.png
01:00:18slimxbut print loading rockbox each time :p
01:00:24CassandraSadly it seems to have disappeared.
01:00:31slimxi don't think so linuxstb
01:00:39hydraheadamn, i forgot how to make it run in disc mode
01:00:39jmgI was just wondering why the pma400 is too braindead to play ogg
01:00:49 Join mirak [0] (n=mirak@AAubervilliers-152-1-9-93.w82-121.abo.wanadoo.fr)
01:01:03*amiconn is curious if and when someone will come up with asm optimised mem* routines for arm
01:01:28 Join NicoFR [0] (n=nico404@rob92-6-82-231-243-63.fbx.proxad.net)
01:01:40Bagderjmg: because it doesn't run rockbox? ;-)
01:02:21slimxpress menu and play button hydrahead
01:02:24jmgBagder: i think the problem is deeper than that ;)
01:03:05Bagderjmg: well, it certainly must have cpu enough
01:03:15linuxstbAsk Archos to implement it. Or port Rockbox....
01:03:55jmg150mhz enough?
01:04:06Bagderyes indeed
01:04:08jmglinuxstb: was thinking of porting opie
01:04:28Bagderisn't that PMA ARM based?
01:04:37jmgyes
01:04:49Bagderso 150 MHz is way more than you need
01:04:51amiconnOh no!
01:04:53jmga lot of zaurus stuff works for it
01:04:58linuxstbI've just read that it's a TI chip - OMAP5910
01:05:01amiconnARM taking over ;)
01:05:02jmgyeah
01:05:07jmgthat's the problem
01:05:11jmgthe compiler is nonfree
01:05:13Bagderhaha
01:05:16BagderTI trap
01:06:07Bagderhey
01:06:15Bagderwe decided to make some devcon2006 shirts
01:07:05linuxstbAnother competition? Or a dictatorship?
01:07:25***Saving seen data "./dancer.seen"
01:07:33Bagderdictatorship
01:07:43linuxstb:) We trust you...
01:08:09CassandraI don't.
01:08:51 Quit ender` (" "I believe the use of noise to make music will increase until we reach a music produced through the aid of electrical instru)
01:09:03JdGordonsoo... hows it going on the 3g?
01:09:06DMJChow do you setup the build environment on linux?
01:09:27linuxstbhttp://www.rockbox.org/twiki/bin/view/Main/CrossCompiler
01:09:28DMJCtrying to build the main firmware + games plugins for h100
01:09:58LinusNDMJC: there is this link called "documentation" on www.rockbox.org
01:10:20imphasingJdGordon: I think slimx got a "normal" build to boot, but it freezes
01:10:23imphasinglast I heard
01:10:29DMJCI found the cygwin stuff but that's for windows
01:10:30JdGordonok
01:10:34 Join kramerica [0] (n=lkd@216-197-247-182.sktn.hsdb.sasknet.sk.ca)
01:10:51DMJCafaik last time someone actually sent me a copy of the compiler binaries
01:10:53LinusNDMJC: "Building the cross compiler"
01:11:17jmghmm
01:11:27*jmg downloads archos sdk
01:11:54jmgthe omap dev kit is 900 USD
01:12:00DMJCbeing shaped SUCKS
01:12:44slimxthat's it JdGordon :)
01:12:49 Quit goa (Read error: 104 (Connection reset by peer))
01:13:06slimxit freezes after displaying the logo
01:13:15JdGordonok, well good luck with it
01:13:29jmglinuxstb: archos provide a version of gcc for building it
01:13:31imphasingI haven't had any luck getting it to compile though
01:13:34imphasingbesides the bootloader
01:13:35linuxstbslimx: Have you discovered which of the ???_init() functions causes the freeze?
01:13:53linuxstbimphasing: Not all the necessary changes are in CVS yet - slimx has been working on it.
01:13:54 Join goa [0] (i=hd@gate-hannes-tdsl.imos.net)
01:14:28imphasingAh, ok.
01:14:28 Quit Moos (Read error: 104 (Connection reset by peer))
01:14:31 Join Moos [0] (i=DrMoos@m53.net81-66-159.noos.fr)
01:14:41DMJCso what kind of stuff is planned for devcon?
01:15:04Bagdergetting devs together the primary objective
01:15:07Bagderis
01:15:09hydraheadhey im looking at theese installation instructions, but i dont exactly understand c)
01:15:29Bagderthen if we manage to get a few, we can start thinking about what we want to accomplish
01:15:32hydraheaddo i need the ipod wizard or not?
01:15:38linuxstbNo.
01:15:43 Join ashridah [0] (i=ashridah@220-253-120-146.VIC.netspace.net.au)
01:15:55linuxstbIgnore the references to ipodwizard.
01:16:07jmgDO YOU ACCEPT THESE TERMS AND CONDITIONS?
01:16:07jmgYES
01:16:07jmgARE YOU SURE THAT YOU WISH TO ACCEPT THESE TERMS AND CONDITIONS?
01:16:09jmgYES
01:16:12jmgthanks archos
01:16:48DMJCGCC 4 works on coldfire?
01:16:57DMJCis that the default GCC 4 code?
01:17:07linuxstbYes, but it produces (mostly) slower code. The recommended version is 3.4.5
01:17:07jmglol i need to give them my serial number to get the sdk :/ :(
01:17:11DMJCah k
01:18:26hydraheadhmm on comand ipodpatcher 3 it says:error locking disk, acces is denied
01:19:00hydraheadperhaps i should reboot, this is the first time ive plugged it using usb, i used firewire until now
01:19:27JdGordonwhats the line to undo a patch? patch -r file does nothing...?
01:19:39LinusNpatch -R < file
01:20:09DMJCsigh
01:20:11DMJC17mb
01:20:14JdGordonthanx
01:20:37*Cassandra ponders whether to have "recorderv2" and "recorderfm" be seperate targets in the manual build system
01:20:57Cassandra(Since some V2s have the radio installed anyway.)
01:21:50LinusNCassandra: http://www.rockbox.org/tshirt.old/rockboxHires.tif
01:22:05CassandraIs there any difference between the Ondio SP and the Ondio FM other than the lack of a radio in the SP?
01:22:16CassandraThanks, Linus.
01:22:16LinusNsp can't record
01:22:31LinusNiirc
01:22:36Moosyep
01:23:15CassandraHmmm. I'm trying to work out whether to combine those as a target too.
01:23:32CassandraSince FM is just a superset of SP, that might be workable.
01:25:25linuxstbhydrahead: I've read reports that those installation instructions work fine with firewire as well.
01:25:30 Join webguest33 [0] (n=18e406cd@labb.contactor.se)
01:25:47hydraheadok ill try it that way
01:26:18CassandraLinusN: Object not found
01:26:18 Quit webguest33 (Client Quit)
01:26:21 Join webguest33 [0] (n=18e406cd@labb.contactor.se)
01:26:25 Join einhirn [0] (n=Miranda@p54862B45.dip0.t-ipconnect.de)
01:26:37*Cassandra decides that it's easier to move from less platforms to more than the other way around.
01:27:04CassandraFor now I'll keep both pairs as single targets. We can change later if that seems sensible.
01:27:05 Quit webguest33 (Client Quit)
01:27:11amiconnYes, Ondio SP can't record and has no radio. Other than that, it's the same as Ondio FM from a user's view
01:27:23LinusNCassandra: oops - http://www.rockbox.org/tshirt.old/rockboxHighRes.tif
01:27:31amiconn(only visible difference from the outside is the missing line in socket)
01:28:18CassandraRight. Which you wouldn't expect if you can't record anyway.
01:29:33amiconnThe fun thing about the MAS3539F is that it actually *can* record, just the MP3 encoder is missing in the ROM.
01:29:48amiconn...and archos even connected the parallel port buffer to the MAS.
01:30:00DMJCheh this should build in like 3 seconds
01:30:22amiconnSo if someone mounts the missing microphone and/or line in socket, it would be possible to record with the PCM codec
01:30:28CassandraBasically I think we need a new manual platform if there's either different keymappings, unique features in all models, or different screen shots.
01:31:02hydraheadlinuxstb: found it using firewire!
01:31:34linuxstbhydrahead: OK, good.
01:31:43slimxipod 3g bootloader had a big bug against rockbox loading
01:32:15slimxbad ram address
01:32:16 Quit muesli__ (Read error: 110 (Connection timed out))
01:32:17linuxstbAh yes, the asm instruction...
01:32:18CassandraI think that implies the need for different manuals for player, recorder, recorder2fm, ondio, h1xx, h3xx, ipodnano, ipodcolor (and potentially ipodvideo, ipodbw)
01:32:23slimxyes
01:32:27hydraheadso this original bootpartition bin it made is gonna stay on the ipod?
01:32:34josh_Cassandra: you can use \if and stuff
01:33:00josh_something like
01:33:12Cassandrajosh_: Yeah, I know, thanks. It's more a question of working out which platforms need to be different at the moment.
01:33:14slimxnow i can debug the rockbox apps/main.c (init )
01:33:34linuxstbhydrahead: "bootpartition.bin" is a backup of the boot partition on your iPod. The Rockbox bootloader is installed to that partition, which is why it's a good idea to keep that file.
01:33:36josh_\def\ifdefined#1{\ifx\expandafter\csname#1\endcsname\relax\else}
01:34:02CassandraIck! What on earth does that do?
01:34:07josh_\ifdefined{modelplayer} text for player \fi \ifdefined{modelcolor} text for color \fi
01:34:16hydraheadok, understood that, checking jut in case :)
01:34:17josh_and then LaTeX it like
01:34:17linuxstbjosh_: Is that raw TeX?
01:34:23josh_linuxstb: yep
01:34:30*linuxstb looks away
01:34:48josh_prompt$ latex '\def\modelcolor{color} \input yourfile.tex'
01:34:56 Quit einhirn (Read error: 104 (Connection reset by peer))
01:34:58 Part LinusN
01:35:02josh_my point is, there are ways to keep it all in one file :P
01:35:32linuxstbjosh_: I think we're going to be using the "optional" package for that stuff. But that's not the problem - the problem is deciding how to use those conditionals to structure the manuals.
01:35:42josh_ah, ok
01:35:51josh_didn't even realize there *was* a package for that
01:35:57hydraheadi now ran this ipod_fw command, is it ment for it not to show any output like: done or something?
01:36:12jmgNewFeature: Support for the FM Remote Control was added to the Music
01:36:13jmg and AudioCorder application
01:36:20linuxstbhydrahead: No, it should at least say something.
01:36:38josh_hydrahead: it should say, "Generating firmware compatible with [insert ipod version here]..."
01:37:22linuxstbjosh_: I've uploaded a replacement rockbox-4g.zip (same URL). Are you able to test it?
01:37:29josh_sure, 1 sec.
01:37:37linuxstbThe LCD code was completely wrong before...
01:38:13imphasing3G != 4G(b&w)?
01:38:20slimxlol linuxstb
01:38:28linuxstbimphasing: true.
01:38:32slimxi've sent a new patch
01:38:38slimx:p
01:38:41hydraheadipod_fw -o apple_os.bin -e 0 bootpartition.bin - i typed this and its not showing anything :/
01:38:44Cassandraimphasing, in what way? Are the controls or screen res different?
01:39:10midkaycontrols, for one..
01:39:14linuxstbhydrahead: Ah, yes. At that stage, it shouldn't display anything. But it should have created a file called apple_os.bin. Can you see it?
01:39:20CassandraAh, OK.
01:39:24slimxnormal buil corrections
01:39:25josh_Cassandra: they use completely different CPUs
01:39:28josh_PP5002 vs PP5020
01:39:30slimx*build
01:39:33josh_with all the I/Os in different places
01:39:42Cassandrajosh_, irrelevent from the perspective of the manual.
01:39:53linuxstbThe buttons are in different places
01:39:59josh_ah yeah, sorry - forgot we were talking about the manual ther
01:40:03CassandraThat's more relevant.
01:40:04slimxnow times to sleep
01:40:06josh_s/$/e/
01:40:07hydraheadIm looking at the root and the ipod control but i dont see anything new
01:40:10linuxstbSee http://ipodlinux.org/Generations for pictures
01:40:13slimxsee ya tomorrow
01:40:28slimxand have a look at the new patch
01:40:34slimx;-)
01:40:35linuxstbgoodnight.
01:40:41CassandraWell, I'm only setting things up for the current ports that are part of the build system anyway.
01:40:51 Quit slimx (Remote closed the connection)
01:40:52CassandraWe'll add the other iPods when they come on line, of course.
01:41:43josh_ok
01:41:51josh_linuxstb: backlight came on for a few seconds, Rockbox logo was displayed briefly
01:41:55josh_now the screen is blank
01:42:08linuxstbThat's not bad....
01:42:21linuxstbDoes anything happen if you press MENU ?
01:42:28josh_no
01:42:34josh_but I have a suspicion it's a contrast problem
01:42:47JdGordonhmm... so.. ive been waiting for the wrong versino to work :p my friend has the b&w 4g, not 3g.. hows that 1 going?
01:42:47josh_because you can barely see the edge of the screen
01:43:04linuxstbThat sounds possible. Any idea what the default contrast should be set to?
01:43:09josh_and the distinction there depends on the contrast, and if contrast is too low you won't be able to see anything
01:43:12hydraheadwhere is the bin file supposed to be?
01:43:21hydraheadin the root?
01:43:28josh_linuxstb: I don't know what the range is raw... the iPL interface is 0-128 and 96 gives good results.
01:43:34linuxstbhydrahead: In the directory you typed the ipod_fw command
01:43:39hydraheadahaa
01:43:40hydraheadok
01:44:03hydraheades its there, sorry :)
01:44:17linuxstbjosh_: OK, thanks.
01:45:01hydraheadnow under e) should i replace color with 4g?
01:45:01imphasingslimx: New patch?
01:45:01imphasing:D
01:45:27josh_linuxstb: lcd_cmd_and_data(0x4, 0x4, contrast);
01:45:42linuxstbjosh_: Maybe it isn't a contrast problem though - if you can see the Rockbox logo, then the contrast should be OK. Unless Rockbox then changes it....
01:45:52josh_I think it might be that RB changes it.
01:46:03josh_during LCD init or something
01:46:17linuxstbThe logo is displayed after the lcd init....
01:46:21josh_hrm
01:46:23josh_idk
01:46:27linuxstbBut let me check something.
01:46:31josh_ok, cool
01:47:09imphasingHrm.
01:47:14imphasingpatch doesn't apply cleanly..
01:47:15imphasing=/
01:47:22hydraheadlinuxstb: sorry to bug you again: how should i type the command for this bootloader? just change it to 4g instead of nano or color?
01:47:24linuxstbYes, Rockbox itself has a contrast setting, and that is wired-up to the ipod's lcd_set_contrast() function. And I think the default value is 40....
01:47:33josh_ack.
01:47:36josh_that might be a problem.
01:47:39josh_40 is almost invisible.
01:47:53linuxstbhydrahead: Yes.
01:48:21linuxstbIn fact, the default contrast is 28....
01:48:24josh_yes, the arg to that 0x4 command is indeed 0 (low) to 127 (high)
01:48:33linuxstbI'll change it to 96.
01:48:41josh_cool, thanks
01:48:49 Quit youngcereal (Connection timed out)
01:51:04 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
01:51:28linuxstbjosh_: OK, new rockbox-4g.zip in the same place. I've hard-coded the contrast to 96.
01:51:42josh_ok, thanks
01:51:43linuxstbSorry, wait 10 seconds....
01:51:55linuxstbOK, it's there.
01:52:35hydraheadallright, got the bootloader going
01:52:44linuxstbIs it displaying messages/
01:52:45linuxstb?
01:52:57 Quit mirak (Remote closed the connection)
01:53:06hydraheadi showed the blue screen with some text and then it loaded the original firmware
01:53:23hydraheadarrrgh
01:53:30hydraheadit showing everything
01:53:37hydraheadlike in a mirror
01:53:44midkayhydrahead, because you didn't install the rockbox.zip?
01:54:07linuxstbhyarion: Do you mean the original firmware is displaying things wrongly?
01:54:18hydraheadyrs
01:54:20hydrahead*yes
01:54:31josh_linuxstb: watch your LCD inits
01:54:51BHSPitLappy( I think hyarion's asleep)
01:54:53josh_see fb.c in ipodloader2 - esp the part about #ifdef YOU_WANT_TO_SCREW_UP_THE_COLORS_IN_RETAILOS
01:55:06 Join youngcereal [0] (n=youngcer@p54B2212E.dip0.t-ipconnect.de)
01:55:06 Quit XavierGr ("Trillian (http://www.ceruleanstudios.com")
01:55:06linuxstbYes, I noticed that (and removed that code) :)
01:55:15josh_and.. rockbox works!
01:55:20linuxstbHurray :)
01:55:23josh_well, sort of.
01:55:29linuxstbBoo....
01:55:31hydraheadits displaying everything like in a miror, also has some letters wich are descending vertically
01:55:32josh_is input supposed to do anything?
01:55:33BHSPitLappy18:54 rockbox sort of works
01:55:34hydraheadwhat now? :/
01:55:55josh_linuxstb: input isn't doing anything
01:56:04josh_btw, the 4g's input is exactly like the latter versions' (photo etc.)
01:56:09josh_it's different from the mini's though.
01:56:38josh_it's showing a directory listing of my FAT32 partition
01:56:51BHSPitLappythat would be the default view, yes
01:56:55linuxstbOK, let me try something. There's one thing we changed with your button drivers.
01:57:10josh_ok
01:57:18josh_does the scroll wheel do anything?
01:57:21hydraheadshould i try to reset the original firmware or try installing the rockbox zip?
01:57:26josh_(is it supposed to, that is)
01:57:56linuxstbhydrahead: Wait a moment, I'm uploaded a new rockbox.zip which I hope will work.
01:58:18linuxstbOK, it's here: http://www.davechapman.f2s.com/rockbox/rockbox-4g.zip
01:58:27josh_linuxstb: the clock works :-)
01:58:39linuxstbOf course :) You can even set the time.
01:58:44josh_cool
01:58:48josh_does it sync to the hwclock if you do?
01:59:19linuxstbYes
01:59:30josh_cool
01:59:38imphasinglinuxstb: Are you going to commit those changes to CVS?
01:59:51imphasingAnd if you feel like doing something with the 3G, I can test for you.
01:59:55linuxstbThe code is trivial - it's just the exact opposite of reading the RTC registers. In the kernel it's in pcf50605.c (I think).
02:00
02:00:42linuxstbimphasing: No, I'm leaving the 3G for slimx. I'll commit my changes when it's either working or I've exhausted my ideas.
02:00:51linuxstb(my 4G changes)
02:01:21josh_linuxstb: input still does nothing
02:01:25imphasingAlrighty.
02:01:32imphasingLCD driver is the same for 4G and 3G?
02:01:41 Quit midkay ("Leaving")
02:01:48linuxstbThe addresses are different, and the initialisation is slightly different.
02:01:55josh_imphasing: yes, except for different base addresses
02:01:55amiconnI can have a look at the 2bpp horiz driver when doing optimisation work for the 16bpp driver
02:02:18linuxstbThat would be good. We're going to need a sim as well.
02:02:25amiconnI plan to tackle this after memmove() and memset16() (which in turn is needed for 16bit opt)
02:02:37hydraheadnow it wont go into disc mode...
02:02:37 Quit NicoFR ()
02:02:56hydraheadoe do i get the original firmware back, ipodupdate?
02:02:57amiconnPerhaps I should also finish my .lang work before...
02:03:01hydrahead*how
02:03:15amiconnMy tree is somewhat messed up atm
02:03:43linuxstbhydrahead: It will go into disk mode.... You can then use ipodpatcher to restore your original bootpartition.bin back to your ipod.
02:04:29linuxstbi.e. just type "ipodpatcher -w N bootpartition.bin" (replacing N with the diskno of your ipod).
02:04:30amiconnick
02:04:54linuxstbamiconn: You need more trees...
02:06:32 Join midkay [0] (n=midkay@24.143.70.99)
02:06:37hydraheadhmm ive copied the rockbox files, ill try to reboot it now
02:08:38hydraheadwe got liftoff! :)))
02:08:40linuxstbjosh_: I've run out of ideas now then. As you said, the 4G button driver should be the same as the Photo (and Nano/Video).
02:08:44hydraheadi can see the menu
02:08:58linuxstbhydrahead: So do the buttons work for you?
02:09:16linuxstbOr can you just see the files on your ipod?
02:09:22 Join Zoide [0] (i=Zoide@aodio.Stanford.EDU)
02:09:47hydraheadno, im in the rockbox menu, it seem the buttons arent working
02:09:50 Nick Zoide is now known as Zoide7777 (i=Zoide@aodio.Stanford.EDU)
02:09:53hydraheadalso some letters are funny
02:09:56 Quit imphasing (Remote closed the connection)
02:10:00hydraheadlike they have the top missing
02:10:12linuxstbI expect there are still some bugs in the lcd code.
02:10:24linuxstbBut I don't know why the buttons don't work...
02:10:48 Join imphasing [0] (n=imphasin@c-69-250-93-218.hsd1.dc.comcast.net)
02:10:51Zoide7777hi, I tried installing the latest CVS Rockbox on my greyscale 4g, and all I get is the Apple symbol. Is anyone having the same problem?
02:11:45hydraheadbuttons definetly not working, the battery is blinking(expected) and the clock display is working
02:11:54lostlogiclinuxstb: you got a minute to check this on the ipod? fixed the issue with overflowing the buffer. http://lostlogicx.com/transfer/rockbox/pcmbuf_rework2.patch
02:11:59linuxstbZoide7777: Yes - it's not finished yet...
02:12:24lostlogic(yay superbowl halftime coding :))
02:12:42hydraheadand in the menu i got: contacts, notes, calendar and rockbox
02:13:04CassandraHmm. Why is CVS complaining that I can't add to /cvsroot/rockbox/CVSROOT/EmptyDir when I try to 'cvs add manual"?
02:13:05imphasingIs the superbowl today??
02:13:07imphasing=/
02:13:12Zoide7777linuxstb: oh, so at this stage it doesn't go past the Apple symbol?
02:13:30linuxstbhydrahead: That's not the menu - that's the file browser.
02:13:39lostlogicimphasing: steelers 7, seahawks 3 at the half (in case you care)
02:13:41CassandraCVS - I hates it I does, precioussss.
02:13:43linuxstbZoide7777: Yes, the version in CVS has bugs in the LCD driver.
02:13:54hydraheadaha, sorry, never saw a rockbox display before
02:13:57imphasinglostlogic: I'm from seattle, so go hawks.
02:13:59imphasing:D
02:14:29imphasingCassandra: (Darcs || SVN) == nice!
02:14:33imphasing:)
02:14:33Zoide7777linuxstb: ok, thanks for the info
02:14:36 Quit tvelocity ("Αναχωρεί")
02:14:58 Join XavierGr [0] (n=XavierGr@ppp84-adsl-230.ath.forthnet.gr)
02:16:02linuxstbCassandra: If you want to create a new top-level directory, you can't use cvs add.
02:16:55linuxstbBut I forget how you do do it.... Maybe "cvs import"
02:17:32 Part kramerica
02:17:32CassandraHmm. If I use import, it'll drag in all the files, and I need to be able to specify which ones are binary.
02:17:59linuxstbJust create import to import an empty "manual" directory. Then check out that empty directory and carry on.
02:18:09linuxstbs/create/use/
02:18:14hydraheadlinuxstb: ive now loaded the original firmware by holding menu during booting, and the display is still not ok
02:18:45linuxstbhydrahead: OK, thanks. I'll have to look at that - but not now (it's 1.20am)
02:19:04linuxstbBut we seem to be making progress.
02:19:17hydraheadok just tell me how to fix the original firmware
02:19:22hydraheadby that commandline?
02:19:45linuxstbYes - that will remove Rockbox. Next time you boot, the original firmware should be OK (the Rockbox bootloader won't mess up the LCD)
02:23:35 Quit midkay ("laterz!")
02:24:16*Cassandra swears at CVS. Now it's claiming I've not set CVSROOT.
02:24:22hydraheadok everything is back to normal, thanks linuxstb
02:24:28CassandraHow come all the other CVS commands know what it is then?
02:26:15linuxstbThe other cvs commands will know because you have a "CVS" directory in your current directory.
02:27:27linuxstblostlogic: Yes, the overflow problem has been fixed. Normal playback seems to work well.
02:27:42hydraheadbye ppl, now that i got the hang of it ill help with further versions :)
02:27:42linuxstbBut I try to stop playback, or skip to the next track, it just freezes.
02:28:45 Quit imphasing (Remote closed the connection)
02:28:58lostlogiclinuxstb: bagh, that's definitely not happening here, must have broken something ipod specific :(
02:29:14linuxstbOr you've just exposed the bugs in my pcm_playback.c code...
02:29:14 Quit thegeek (Read error: 104 (Connection reset by peer))
02:29:14 Quit hydrahead ("CGI:IRC (EOF)")
02:29:43CassandraAh, crap - what do I want to set the vendor and release tags to?
02:29:45linuxstbAlso, the audio buffer doesn't seem to get fully filled. It stops filling at about 24MB (out of about 29MB).
02:30:07 Part Zoide7777
02:30:13linuxstbCassandra: Yes, that's the question I can never answer....
02:30:14lostlogiclinuxstb: hmm, I haven't laid finger 1 on the audiobuffer code
02:30:16 Join imphasing [0] (n=imphasin@c-69-250-93-218.hsd1.dc.comcast.net)
02:30:30linuxstblostlogic: That makes sense - I noticed the same behaviour with the existing code.
02:30:36 Join thegeek [0] (n=thegeek@s026b.studby.ntnu.no)
02:30:43linuxstb(but wasn't 100% sure)
02:31:10lostlogicok... I'll try to hunt in the ipod specific parts, but may need you or preglow to help me find bugs tomorrow.
02:31:17linuxstbShould your code be faster than the existing code?
02:32:00linuxstbMy ALAC test tracks seem to skip less - some tracks that skipped once or twice now don't skip, and other tracks skip less.
02:32:25Paul_The_NerdI thought a goal of it *was* performance improvement?
02:32:34lostlogicIt's very close, my code uses memory more effectively...
02:32:41linuxstbI thought the goal was simplification...
02:32:48lostlogicmain goal was improving structure
02:32:55lostlogicperformance is a hopeful side effect of that :)
02:33:30lostlogicPaul_The_Nerd: you interested in testing the latest patch I just posted with voice?
02:33:57Paul_The_Nerdlostlogic: Sure, why not.
02:34:04Paul_The_NerdH120, iPod Nano, or both?
02:34:29lostlogicPaul_The_Nerd: whichever you feel like
02:34:43lostlogiclinuxstb says it's broken anytime playback needs to stop on ipod
02:35:21Paul_The_NerdThe link you posted is the patch I need?
02:36:04lostlogicyeah
02:40:38Paul_The_NerdOkay, I'll let you know how it goes.
02:42:20linuxstblostlogic: gapless playback is still flawless....
02:42:28linuxstb(with FLAC)
02:42:53CassandraHardeep posted a gapless fix for MP3. Wonder if it worked.
02:46:11BHSPitLappydo the mp3's have to have a special flag
02:46:20Paul_The_NerdYes
02:46:38BHSPitLappywell the only music I have where gapless is necessary is AAC actually
02:46:44Paul_The_NerdBHSPitLappy: Basically, I think LAME is the only encoder that does it, as far as I know. For MP3s.
02:46:54BHSPitLappylame's what I use.
02:47:02Paul_The_NerdAnd of course, the player has to read the tags.
02:48:13BHSPitLappyany slackware users here?
02:48:24lostlogiclinuxstb: I'm wondering about putting a call to dma_stop() as the last line in fiq()
02:48:30Paul_The_Nerdlostlogic: Same freezes it had before for voice on h120
02:48:56lostlogicPaul_The_Nerd: gah! so if you are using voice menus, after a few menus it frezes the player, or stops playing them?
02:49:02Paul_The_NerdFreezes the player
02:49:22lostlogicPaul_The_Nerd: hmmmmmmmmmmmm
02:49:50Paul_The_NerdSoft freeze though. If you hold stop long enough, you can still shut off
02:50:44Paul_The_NerdAnd, I guess voice menus aren't enabled on iPod yet.
02:51:08BHSPitLappywhat a weird hummer commercial
02:51:15lostlogicBHSPitLappy: yes!
02:51:30linuxstbPaul_The_Nerd: I haven't disabled voice menus afaik. But I haven't tested them.
02:51:44BHSPitLappylostlogic: eh?
02:51:54lostlogicBHSPitLappy: it was weird... my gf and I were just saying so
02:52:02BHSPitLappylol
02:53:42Paul_The_Nerdlinuxstb: Well, they certainly aren't reading with lostlogic's build, at least. At all. No freezes on them, at least. ;-). Lemme try with a normal build
02:54:48linuxstblostlogic: I tried adding that dma_stop(), but it doesn't change anything.
02:55:21lostlogiclinuxstb: :(
02:55:41BHSPitLappyhahaa
02:55:43BHSPitLappyjackasses
02:55:51linuxstblostlogic: Is ipod playback the only thing stopping you from committing it?
02:55:59lostlogiclinuxstb: well that and voice menus now
02:56:15lostlogicI should download a set of voice menus so I can try them
02:56:32Paul_The_NerdWhat ever happened to the AT&T thing?
02:57:33linuxstbMmm. I see that I haven't implemented pcm_play_pause()...
02:58:02 Join BHSPitMonkey [0] (n=billybob@67.64.118.242)
02:59:59 Join midkay [0] (n=midkay@24.143.70.99)
03:00
03:00:51*BHSPitLappy punts midkay
03:01:14*midkay hits BHSPitLappy in the face, affectionately.
03:01:17lostlogicwhere does the voice file go on the player?
03:01:36BHSPitLappyroot, I thought
03:01:41BHSPitLappydon't quote me though
03:01:45Mooslanguage dir
03:02:11lostlogicty
03:06:15 Quit BHSPitMonkey (Remote closed the connection)
03:07:26***Saving seen data "./dancer.seen"
03:07:41lostlogicPaul_The_Nerd: consistently the 4th voice read freezes?
03:08:12Paul_The_Nerdlostlogic: Actually, 2nd.
03:08:32lostlogicPaul_The_Nerd: ok, so it's not always the same
03:08:37Paul_The_NerdBut usually, I would just hit down, listen, hit down again, listen, freeze, after enabling voice.
03:12:03lostlogicPaul_The_Nerd: interestingly it works during playack, but freezes otherwise
03:12:31Paul_The_NerdHrm. Well, that seems like behaviour chock full 'o clues.
03:12:52Paul_The_NerdLinuxstb: Yes, voices seem to definitely *no* play at all on iPod Nano. Just tried it with a clean CVS build
03:13:27lostlogicPaul_The_Nerd: yep, I at least know where to look... _then_ I'll probably commit it and let linuxstb and preglow fight with the ipod bug (or bitch at me till I fix it :))
03:13:36Paul_The_NerdHahaha
03:16:50JdGordondoes any1 know if you can open files (on windows) like \\\\.\\PhysicalDrive0\folder\file.exe
03:17:17linuxstbJdGordon: Yes, I think you can.
03:17:27JdGordongreat :D
03:17:36 Quit midkay (Remote closed the connection)
03:19:09 Join midkay [0] (n=midkay@24.143.70.99)
03:21:16 Quit midkay (Client Quit)
03:24:58 Quit Moos ("Glory to Rockbox!")
03:27:58yeahxit reads the track names?
03:29:38Paul_The_Nerdyeahx: Huh?
03:30:40yeahxI've read something that sounds like rockbox has speach on the site and then someone just meantioned it
03:30:58lostlogicin preparation for making voice work when pcm is paused, anyone mind if I allocate an extra 8k dedicated voice buffer?
03:33:11Paul_The_Nerdyeahx: Well, the menu system can be read aloud. If you run a script available on the site (and have Text to Speech support enabled in windows) or if you prepare speech files for each file, you can have it read out filenames while browsing. Otherwise you can command it to spell them, or just tell you the number (3 for third file in the folder, etc)
03:33:37Paul_The_NerdThe voice files available on the website right now for the menus are a little out of date though, so not all menu entries will be supported.
03:34:34yeahxoh ok, so this would mainly be for blind navigation then, thats interesting
03:34:39 Join kkurbjun [0] (n=Jim@c-24-8-222-177.hsd1.co.comcast.net)
03:35:06Paul_The_NerdThat's its intended purpose, yet.
03:35:11lostlogicnevermind.
03:36:18kkurbjundid someone get doom running as a viewer?
03:43:24ashridahkkurbjun: people were looking into it, but we were hindered by not having an actual copy that can be compile :)
03:43:53JdGordonbny hindered... u mean thrown into a brick wall...
03:44:05kkurbjunashridah: what was wrong with compiling what is online?
03:44:27kkurbjunI don't think that making doom a viewer is the proper way to approach it
03:44:30ashridahkkurbjun: it failed :)
03:44:50ashridahkkurbjun: it is if you want to use a PWAD or IWAD :)
03:44:56kkurbjunashridah: where, I would like to correct the problem
03:45:25kkurbjunif you would like to load individually made levels that depend on doom1 or 2 it would make a mess though
03:45:28ashridahof course, since you should be able to add multiple PWAD's, it'd be nice if it could be done within the plugin, by the sounds of it
03:45:30 Join pabs_ [0] (n=pabs@xor.pablotron.org)
03:45:34ashridahnot that i have a player that can play it
03:45:36Paul_The_Nerdashridah: How would you handle PWADs? Is there some way to determine if they're produced against Doom 1 or 2, or can you arbitrarily load them all against a Doom 2 one, if you've modify the executable?
03:45:42 Quit BHSPitLappy (Read error: 104 (Connection reset by peer))
03:45:48 Quit pabs (Nick collision from services.)
03:45:52ashridahPaul_The_Nerd: hell if i know :)
03:45:53 Nick pabs_ is now known as pabs (n=pabs@xor.pablotron.org)
03:46:26kkurbjunthe other problem is that loading multiple pwad's won't work right now
03:46:39kkurbjunthe realloc code has amnesia
03:47:41kkurbjunashridah: either way though I would like to know where compiling fails so that people can work on it.
03:48:05Paul_The_NerdI mean, if I were designing the UI, I'd make it so you can launch it on an IWAD as a viewer, then it shows a list of PWADs (presumably all stored in a subfolder with a name matching that of the IWAD) for you to select from (checkboxes next to them, or something)
03:48:20ashridahkkurbjun: JdGordon was working on it after i had a quick look. bug him :)
03:49:17JdGordonkkurbjun: where is the latest sources?
03:49:47kkurbjunPaul_The_Nerd: that was one option that I thought of.. the other is having a full interface to pick which base wad you would like without the viewer.. I guess that rockboy set the expectation of running things as a viewer though
03:49:59kkurbjunJdGordon: they are on sourceforge
03:50:20kkurbjunyou have to apply the diff to a cvs copy
03:50:44kkurbjunand then put the files in doom.zip to apps/plugins
03:50:59kkurbjunshould leave you with apps/plugins/doom
03:51:02Paul_The_Nerdkkurbjun: I think viewers just feel simpler than launching it, and then browsing to the file, because there's less need to put the files in a specific location (or include a browser that lets you go up/down levels)
03:52:02Paul_The_NerdI wouldn't object to a nice full interface, mind you.
03:53:00kkurbjunPaul_The_Nerd: True, I'll have to look into it more and see how individual levels use doom1/2. If someone wants to make it a viewer for now though feel free too.
03:53:26JdGordoncompiling...
03:53:49JdGordonu can load user wads through the command line right?
03:54:43kkurbjunyes, but like I said the realloc code does not work correct so sending the -files or -file (whichever it is) won't work properly right now
03:55:11Paul_The_NerdJdGordon: A user wad (PWAD) is essentially a Patch against the main wad (IWAD), so a plugin launching a user wad would need to know which IWAD it was against somehow.
03:55:39Paul_The_NerdThere's actually one "Open Source" IWAD out there.
03:55:40kkurbjunselecting the base iwad would be fine to do though
03:56:00Paul_The_NerdA group went through and replaced the *entirety* of Doom 2's resources, etc, with from-scratch ones.
03:56:31kkurbjunPaul: the shareware wad would be ok to distribute though wouldn't it?
03:57:08 Join BHSPitLappy [0] (i=Steve-O@67.64.118.242)
03:57:26yeahxblah its all outdated and should be fine
03:57:29Paul_The_NerdRedistributing the shareware Wad is *probably* okay, depending on the license they included
03:57:43DMJChas anyone tried building the compiler stuff on athlon 64?
03:57:45Paul_The_NerdYeahx: "should be fine" is different from "legally 100% safe" which is kinda the ideal.
03:57:52DMJCI'm getting malloc errors on binutils
03:57:55DMJCand gcc won't compile
03:57:59Paul_The_Nerdkkurbjun: http://freedoom.sourceforge.net/
03:57:59BHSPitLappyiDoom distributes the shareware wad
03:58:12kkurbjunJdGordon: did you get it to compile
03:58:21yeahxyeh, people can easily get the ones they want
03:58:24JdGordoncompiling
03:58:29kkurbjunI'm looking for a copy of the license for the shareware wad
03:58:30Paul_The_NerdBHSPitLappy: I'm pretty sure the Freedoom IWAD actually works with all Doom 2 PWADs.
03:59:06Paul_The_Nerdkkurbjun: I'd still suggest using the iwad from Freedoom instead. It contains replacement resources for everything in Doom 2.
03:59:41JdGordonMAKE in doom
03:59:42JdGordon....
04:00
04:00:00kkurbjunPaul: that's an option, honestly though I really prefer doom1 over doom2, and if it's legal to distribute doom1 shareware that'd be ideal for me
04:00:02JdGordonm_fixed.c:62: error: can't find a register in class `DREG' while reloading `asm'
04:00:15BHSPitLappyPaul_The_Nerd: http://ipodlinux.org/forums/viewtopic.php?t=4984
04:00:25BHSPitLappyJdGordon: how'd the color picker turn out
04:00:28kkurbjunJdGordon: what platform are you compiling for?
04:00:50JdGordonBHSPitLappy: the patch is on the tracker... just waiting for higher powers to add it
04:00:55JdGordonkkurbjun: h300 sim
04:01:13kkurbjunI havn't tested the sim or tried to get it working
04:01:29kkurbjunI can't compile one on my system
04:01:34JdGordongah, ok.. only compiles on the target?
04:01:47kkurbjunyes, but to correct that error...
04:01:50Paul_The_NerdBHSPitLappy: I miss the old AliensTC for doom. BEST Doom Total Conversion I'd played, until Fox killed it
04:01:51kkurbjunohe second
04:02:25DMJCsailor moon doom
04:02:30kkurbjunyou need to ifdef the asm code
04:02:33DMJCworst conversion ever
04:02:50kkurbjunto make it use the #if 0 code
04:03:51ashridahhahah. what i found hysterical were the porn doom packs. replacing nearly every texture in the game with an animated 4-frame sex scene :)
04:04:12kkurbjunJdGordon: I'll uplaod a new version that should fix it
04:04:14ashridahno actual relation to the game, they've pretty much just gone down the line replacing as they go
04:04:15BHSPitLappyI want chex quest on my nano!
04:04:56Paul_The_NerdChex Quest didn't even change the weapon balance. Hehehe
04:05:13JdGordonbbs
04:07:51DMJCI was unable to build gcc or binutils
04:08:01DMJCI'm assuming the problem is Athlon 64 architecture
04:09:18kkurbjunJdGordon: I've put new code up that should fix the problem
04:09:53 Join yeahx_ [0] (n=aarond@c-67-160-124-13.hsd1.wa.comcast.net)
04:20:28 Quit yeahx (Read error: 110 (Connection timed out))
04:22:03imphasinglinuxstb: Do you know what I need to comment out to get the bootloader to load linux, or rockbox?
04:22:08imphasingon the 3G
04:33:04 Join ehntoo [0] (n=ehntoo@24-177-166-0.dhcp.mrqt.mi.charter.com)
04:34:16linuxstbimphasing: Have you seen slimx's last patch?
04:35:06imphasing"cleaned lcd-ipod.c"?
04:35:10imphasingI couldn't get that one to apply
04:35:14imphasinglots of hunk errors
04:38:23 Quit kkurbjun (Read error: 104 (Connection reset by peer))
04:38:35 Quit belze ("Be Not Nobody...")
04:39:12linuxstbimphasing: Yes, that patch isn't good :(
04:39:43 Join kkurbjun [0] (n=Jim@c-24-8-222-177.hsd1.co.comcast.net)
04:40:04imphasing:(
04:40:30imphasingThe bootloader will build straight from CVS, but I'm not sure what I need to do to get it to boot anyhting besides retailos
04:41:46 Join ModernExecutive [0] (n=bob@pool-141-152-166-148.norf.east.verizon.net)
04:47:03ModernExecutiveanyone know if linus had any time to work on a new h300 us fw + bl?
04:47:19DMJCcrap
04:47:28DMJCI cant' get the crosscompiler to build
04:47:49imphasing64 bit?
04:48:33DMJCyeah
04:49:53 Join yeahx [0] (n=aarond@c-67-160-124-13.hsd1.wa.comcast.net)
04:50:10DMJC64-bit native linux
04:50:36DMJC.././opcodes/disassemble.c:392: undefined reference to `print_insn_sh'
04:50:36DMJCcollect2: ld returned 1 exit status
04:50:52DMJCthat's when I force host to i686 and select m68k-elf
04:52:54 Quit ModernExecutive ()
04:52:54 Quit webguest77 ("CGI:IRC (EOF)")
04:53:28 Part yeahx
04:54:46linuxstbDMJC: Even if you manage to get the cross-compiler compiled, some of the Rockbox build tools are not 64-bit clean yet....
04:55:04DMJCUm I don't think you understand
04:55:18DMJCI was trying to force the cross compiler to build in 32-bit mode
04:55:32DMJCthen I would've executed it as a 32-bit app
04:55:37DMJCathlon 64 can do that
04:56:01imphasingyerayeah
04:56:16imphasingEr..
04:56:25imphasingIgnore the "yera" part..
04:56:33BHSPitLappydo we want to know what it was?
04:56:50DMJCheh
04:58:04JdGordonPaul_The_Nerd: was it you who was gonna merge the colour settings and my colour chooser patches?
05:00
05:01:16Paul_The_NerdJdGordon: I was gonna try to look at it, but I honestly haven't had a chance, at the moment. It's been a very busy weekend, with far far too little sleep.
05:01:26 Join Rob2222 [0] (n=Miranda@ACB08D1F.ipt.aol.com)
05:01:28JdGordonok
05:01:32Paul_The_NerdSorry
05:02:16JdGordonno prob
05:02:23JdGordonit looks like the damn patch doesnt work anyway
05:02:28Paul_The_NerdWell, that sucks.
05:02:41 Quit yeahx_ (Read error: 110 (Connection timed out))
05:04:12kkurbjunJdGordon: did you try compiling the sim source?
05:04:20kkurbjunor new source?
05:04:22JdGordonnot yet...
05:04:54JdGordonif we cant use the -file switch then i wouldnt know where to start to get it to load the wad... so cant really help
05:06:48kkurbjunJdGordon: you can add a flag for rockbox to fix the identify version code in d_main.c. I would be interested to know if the sim compiles though.
05:07:28***Saving seen data "./dancer.seen"
05:08:14kkurbjunor you could add some flag like -iwad and then check it in identify version
05:08:39JdGordonok, busy atm.. but ill have a look'
05:09:09 Nick aegray_ is now known as aegray (n=aegray@12-210-86-210.client.insightBB.com)
05:09:13kkurbjunno worries, I'm just supplying idea's. I plan on working on it once I have sound working.
05:09:22Paul_The_NerdOne thing you *could* do is force users to rename the files to .pwad and .iwad.
05:09:31Paul_The_NerdThen, if they launch a .pwad, it simply asks them to select another file.
05:09:35Paul_The_NerdThis repeats until they pick an iwad.
05:09:44Paul_The_NerdAs an idea for viewer-like use
05:13:22 Quit DreamTactix291 ()
05:13:53 Join DreamTactix291 [0] (n=DreamTac@adsl-149-149-154.bna.bellsouth.net)
05:14:01linuxstbAm I right in saying that you have one iwad, which is then (optionally) modified by a pwad?
05:14:07Paul_The_NerdYes
05:14:19BHSPitLappywikipedia has a decent explanation of the *wads
05:14:27Paul_The_NerdBHSPitLappy: Link?
05:14:31BHSPitLappyumm
05:14:44JdGordonkkurbjun: kkurbjun/doom.zip">http://www.mines.edu/students/k/kkurbjun/doom.zip is latest source?
05:14:57BHSPitLappyhttp://en.wikipedia.org/wiki/Doom_WAD
05:15:08linuxstbSo one option would be for doom to look in the same directory for an iwad if a user selects a pwad.
05:15:40kkurbjunJdGordon: yes
05:16:32BHSPitLappyor have a plugin that's a simple interface where you pick enough wads to "complete a puzzle" and then launch doom
05:17:01Paul_The_Nerdlinuxstb: You can use an arbitrary number of pwads though. Not just one.
05:17:10 Quit mikearthur (Remote closed the connection)
05:17:25JdGordonuh oh.. i fucked my tree..... how do i get cvs to redownload any file that has been modified?
05:18:14BHSPitLappydamn arborphiles
05:18:25Paul_The_NerdDendrophiles
05:18:48linuxstbPaul_The_Nerd: Sounds like a nightmare then.... :) Unless you either write a complicated new UI "widget", or create new ".doom" files that are plain text files listing wad files to be used.
05:19:05Paul_The_Nerdlinuxstb: I was actually thinking of that .doom thing. :) Seems a good idea
05:19:12linuxstbOr just load every wad in the same directory as the one you choose.
05:19:13ashridahJdGordon: cvs update -C iirc
05:19:15Paul_The_Nerd'cuz then you could just replicate the command line parameters.
05:19:32 Quit Rob2222_ (Read error: 113 (No route to host))
05:19:34JdGordonyup, ta
05:19:38kkurbjunthe .doom idea would be a good idea
05:19:40ashridahit replaces locally modified files wth the original, and saves the modified one in .#filename.revision iirc
05:19:47kkurbjunand pretty easy
05:20:03kkurbjunrelatively that is to creating a full interface
05:20:21Paul_The_NerdThat way all the wads can be hidden somewhere nice, like under .rockbox/doom, and then you can put shortcuts for each of your mods wherever you want them readily accessible.
05:20:50Paul_The_Nerdkkurbjun: I know some source ports have allowed loading of deHacked patches via the command line too. Do you know if this source code is based off of one of those?
05:20:54linuxstbPaul_The_Nerd: Is Doom running on the ipod?
05:22:03kkurbjunPaul_The_Nerd: no, I don't have the dHacked patches code in there. I started backporting some of the prBoom features, but the dehacked suppport requires rewiting all of the file code.
05:22:09Paul_The_Nerdlinuxstb: haven't tried.
05:22:24kkurbjunI've never used dehacked so I didn't really see a benefit
05:22:32Paul_The_Nerdkkurbjun: Alright. Not exactly a feature of the original anyway. ;-)
05:22:48kkurbjunwhat types of things did people use dehacked for?
05:23:01ashridahhahah. dehacked modified the doom.exe
05:23:09ashridahyou could use it to modify the animations of various creatures
05:23:12Paul_The_NerdIt was used in some of the more complex TCs. It lets you change values on weapon damage, or which AI an enemy uses... projectile behaviour
05:23:22ashridahsince they were compiled into the executable.
05:23:22kkurbjunhmm
05:23:28ashridahlike making a rapidfire BFG
05:23:29Paul_The_NerdYou could even set one entity to launch another as a projectile. It was used to have an egg spawn a facehugger enemy in the Aliens TC
05:23:38ashridahor turning the chaingun you have into a chain-shotgun :)
05:23:51Paul_The_Nerdashridah: I tended to enjoy the non-balance breaking ones. :-P
05:23:56ashridahPaul_The_Nerd: or have ghosts come out of dead creatures :)
05:24:05kkurbjunwere there any "good" conversions that used it?
05:24:48ashridahkkurbjun: as paul said, the aliens TC used it to reasonable effect.
05:25:03Paul_The_Nerdkkurbjun: AliensTC was most excellent. Other than that though, it seemed pretty rarely used, though I did see a few alternate gametypes created with it, like various sports, or hockey, or other such things.
05:25:19Paul_The_NerdBut, in the end, I wouldn't qualify it as an "essential" feature in any way
05:25:28kkurbjunI may look into it later, but it's not really a priority now
05:25:43Paul_The_Nerdkkurbjun: are there keymappings for ipod?
05:26:24ashridahman, i can remember some good times having jousting matches with doom.
05:26:30ashridahthose were the days :)
05:26:56kkurbjunno, I don't have an ipod to test with or a way to build a sim. I think there are some additional flags that are required to get doom to run correct on the ipod
05:26:56Paul_The_Nerdashridah: Maps with lots of berserk packs, and no guns?
05:27:07kkurbjunPaul: do you still have the changes you made?
05:27:16ashridahPaul_The_Nerd: nah, basically chainsaw battles
05:27:42Paul_The_Nerdkkurbjun: Are the keys still hard coded, or can I just add another set of #defines in the right file?
05:27:44ashridahyou'd run from either end of a barrier you can't pass, cut each other up a bit, then go back for another pass :)
05:27:50Paul_The_NerdHehehe
05:28:33kkurbjunlinuxstb: I also tried writing directly to the memory for doom. It seemed to give a bit of a performance improvement, but I can't measure it with my old fps code. Paul you can modify the defines I have in i_video.c.
05:28:48kkurbjundirectly to the lcd memory that is
05:28:58Paul_The_Nerdkkurbjun, linuxstb: Okay, gonna test out doom on iPod then, see how far I get.
05:29:22kkurbjunPaul, the new compiler flags you need are on sourceforge I believe
05:34:51Paul_The_NerdWhat would you use as a conditional in a makefile for ipod targets?
05:35:32kkurbjunPaul: I'm not sure, I would just add the flags to the CFLAGS section at the top for now to see if it works
05:35:41Paul_The_NerdI did that
05:35:55Paul_The_NerdBut I'd much rather have ifdef something cflags += etc,
05:36:21Paul_The_NerdThat way, if I do get it working, I can give you a nice little patch and not break any of your stuff. ;-)
05:36:23ashridahPaul_The_Nerd: conditional stuff in makefiles is really non-standard iirc
05:37:18ashridahthat's one of the reasons they use cpp on files like SOURCES iirc
05:38:24linuxstbPaul_The_Nerd: Look in apps/plugins/Makefile - for the section that adds rockboy (and probably Doom) to the build. This uses "ifeq".
05:38:54ashridahdoesn't that only work in gnu make?
05:39:08linuxstbProbably - but Rockbox demands gnu make
05:39:33*ashridah remembers the pain of trying to use a conditional make target using solaris make
05:39:35*ashridah shudders
05:39:59Paul_The_Nerdlinuxstb: I saw that. But do I just use IPOD_NANO instead of what it uses?
05:41:14kkurbjunPaul_The_Nerd: the cflags should be the same for all the IPOD/ARM targets
05:41:55linuxstbThe problem is that Makefiles don't have access to the #defines in config-*.h - all you really have is the target name.
05:42:49Paul_The_Nerdlinuxstb, kkurbjun: Okay, I think I've got the makefile thing worked out. Thanks.
05:43:18Paul_The_Nerd*think* being the operative word
05:45:26Paul_The_NerdWarnings and Errors, wee! So, at least I have doom properly wrapped in a conditional.
05:45:51 Quit PaulJ (".")
05:45:57kkurbjunPaul_The_Nerd: what was the error?
05:46:02Paul_The_NerdOne sec
05:46:38Paul_The_Nerdm_fixed.c:47: error: impossible constraint in 'asm'
05:46:57kkurbjunPaul: did you get the latest source on sourceforge?
05:47:16Paul_The_NerdI followed the link in your comment
05:48:00kkurbjunhmm
05:48:08kkurbjuntonight that is?
05:48:18Paul_The_NerdYes
05:48:22Paul_The_NerdJust a few minutes ago
05:48:38kkurbjunok, one second
05:49:01Paul_The_NerdIt's in a "#if !defined(CPU_COLDFIRE) && !defined(SIMULATOR)" block
05:49:20kkurbjunoops, first should not have the !
05:49:23kkurbjun#if defined(CPU_COLDFIRE) && !defined(SIMULATOR)
05:49:42Paul_The_NerdI was thinking "Hrm, that seems really strange for wrapping a block of assembly" ;-)
05:50:10kkurbjun:)
05:50:50Paul_The_NerdMaaaaan.... so many warnings. :)
05:51:10Paul_The_NerdOkay, fixing the button definitions now
05:51:33kkurbjunjust updated cvs link with that ! fixed
05:53:34Paul_The_Nerdifndef BASICLAYOUT?
05:54:03kkurbjunyeah, that was a temporary for now while I was messing with the layout
05:54:26 Join midkay [0] (n=midkay@24.143.70.99)
05:54:55kkurbjunhad another weird ifdef on the video initialization that I fixed too
05:55:01Paul_The_NerdOk
05:55:13kkurbjunit shouldn't affect your work though
05:56:22 Quit RotAtoR ("zzzzzzzz")
05:56:47Paul_The_NerdI just commented out the first section, and put in #if CONFIG_KEYPAD == IPOD_4G_PAD and put my definitions there.
05:57:21kkurbjunsounds good
05:58:18Paul_The_NerdDOOMBUTTON_ENTER, is it only used during menus? Can it safely be the same as DOOMBUTTON_SHOOT?
05:59:03Paul_The_Nerdlinuxstb: Can all buttons be modified with Select?
05:59:06kkurbjunI havn't tried it, but probably.
05:59:38Paul_The_Nerdkkurbjun: Okay, I'm trying Y, ENTER and SHOOT as "Select" for the moment
06:00
06:00:08kkurbjunPaul that won't work
06:00:21kkurbjunI need to rename that Y one to change weapon
06:00:33Paul_The_NerdAaah
06:01:25Paul_The_NerdOh, man, I just forgot all about Left/Right anyway, since I was using wheel rotation for turning
06:01:31Paul_The_NerdI can just use those, and not worry 'bout it
06:01:51Paul_The_Nerdi_video.c:146: error: 'struct plugin_api' has no member named 'button_hold'
06:02:43Paul_The_Nerdkkurbjun: So, new error −−^
06:02:49kkurbjunoy, one second, I'll fix that one too
06:04:02kkurbjunneeds #if CONFIG_KEYPAD == IRIVER_H300_PAD
06:04:11kkurbjunthat before if (rb->button_hold()&~holdbutton)
06:04:16kkurbjunand and #endif
06:04:24kkurbjunafter holdbutton=rb->button_hold();
06:04:50kkurbjunthat's in the getkey function
06:05:08Paul_The_NerdYeah, got it. :)
06:07:40Paul_The_Nerdkkurbjun: Also, as a note: cpu_boost doesn't work yet. That line also provides a small problem at 125 and 143 in rockdoom.c
06:08:15Paul_The_NerdA couple #ifndef CPU_ARMs fixed those though
06:08:21kkurbjunok, I'll add an ifdef for that one too
06:08:45Paul_The_NerdAlso, where do I change the plugin ram size?
06:09:28kkurbjunfirmware/export/config-ipodnano.h
06:09:37kkurbjun#define PLUGIN_BUFFER_SIZE 0x80000
06:09:45kkurbjunto #define PLUGIN_BUFFER_SIZE 0xF0000
06:10:11Paul_The_NerdGotcha
06:10:16kkurbjunyou have to re-run configure after changing that in a clean make directory
06:10:33Paul_The_NerdIndeed.
06:16:49kkurbjunPaul: is it running?
06:17:00Paul_The_NerdHad a compile problem
06:17:08kkurbjunwhat was it?
06:17:37Paul_The_NerdWell, I *think* I just forgot to hit save in my editor window with the buffer size change
06:18:15Paul_The_NerdNo, it still doesn't build
06:18:35kkurbjunwhat's the error?
06:18:45kkurbjunhmm, probably iram now..
06:18:48Paul_The_Nerd /opt/arm/lib/gcc/arm-elf/4.0.2/../../../../arm-elf/bin/ld: address 0x2002144 of /home/DarkkOne/rockbox-devel/ipod/apps/plugins/doom/doom.elf section .bss is not
06:18:48Paul_The_Nerd within region PLUGIN_RAM
06:19:00kkurbjunnope
06:19:02kkurbjunhmm
06:20:34kkurbjuntry making it #define PLUGIN_BUFFER_SIZE 0x100000
06:21:04kkurbjunor to be safe 0x200000
06:21:27Paul_The_NerdYeah, can scale it back later
06:23:54kkurbjunPaul: there may be another error about iram
06:24:10Paul_The_NerdOk
06:24:20 Quit JdGordon (Read error: 104 (Connection reset by peer))
06:24:49Paul_The_NerdNope
06:24:49kkurbjuna quick fix for it would to be to remove ICONST_ATTR at line 585 in tables.c
06:24:52kkurbjungood
06:24:53Paul_The_NerdIt built fine with that
06:27:57Paul_The_NerdOkay, Doom2 gets "Buffer Error"
06:28:04kkurbjunhmm
06:28:11kkurbjunwhat cflags are you using?
06:28:16Paul_The_NerdDoom1 SW gets "Buffer Error" and then "Couldn't realloc lumpinfo"
06:29:33Paul_The_NerdI just added -fpack-struct -mstructure-size-boundary=8
06:30:19Paul_The_NerdAnd the -fpack-struct thing seems to have worked, since there were all these warning messages: doomdata.h:59: warning: #pragma pack has no effect with -fpack-struct - ignored
06:30:27Papricakkurbjun, you have compiled version for the new version?
06:31:48kkurbjunPaprica: it's not a significant change right now from what's available. We're working on getting it functioning on the IPOD
06:32:11kkurbjunPaul, let me look into that
06:32:14Paul_The_NerdOkay
06:32:17Papricaok
06:32:23Paul_The_NerdIs it part of the realloc stupidness?
06:32:33Paul_The_NerdYou mentioned it was forgetful
06:32:38 Join Sinbios [0] (n=Sinbios@HSE-Toronto-ppp301652.sympatico.ca)
06:32:50kkurbjunno, it's a problem with reading the structures from the wad file I believe
06:32:53Paul_The_NerdAh
06:33:25linuxstbPaul_The_Nerd: If you are using the BUTTON_SCROLL_* events, then I don't hink you can test for them using button_status() - you have to use button_get()
06:33:39linuxstbs/hink/think/ :)
06:34:33Paul_The_Nerdlinuxstb: I'll remember that point once input is actually happening. :)
06:34:53kkurbjunpaul: are those errors just in doomdata.h?
06:35:05kkurbjunor warnings?
06:35:14Paul_The_NerdNo, they're *all* over
06:35:38Paul_The_NerdI think
06:35:43Paul_The_NerdLemme double check that. Buffer's too small
06:37:13kkurbjunlinuxstb: is the arm big or little indian?
06:38:11imphasingIs there a shortage of people with 5G ipods, for testing?
06:38:24imphasingI'm buying a new ipod, and I wonder which would be best for development..
06:38:48Paul_The_Nerdkkurbjun: Sorry, yeah, looks like it is just doomdata.h. It's just showing up bunches because of all the includes, I suppose
06:39:04Paul_The_NerdLine 59 and 213 and that's it
06:40:10kkurbjunPaul: that shouldn't be the problem.. you tried once to remove -D__BIG_ENDIAN__ from the cflags?
06:40:29Paul_The_NerdOh man
06:40:35Paul_The_NerdI forgot to remove it this time
06:41:33 Join DarthLappy [0] (n=DarthShr@202-161-21-45.dyn.iinet.net.au)
06:42:34Paul_The_NerdShould I replace BIG with LITTLE or should that be unnecessary?
06:43:42kkurbjunmake sure the whole doom directory gets rebuilt
06:43:43kkurbjunyou don't need to do that
06:43:49Paul_The_NerdOkay
06:43:53Paul_The_NerdDoing a clean build ayway
06:48:44Paul_The_Nerdkkurbjun: Without -D__BIG_ENDIAN__ gets a "Data abort at 01E5FBF8"
06:49:18kkurbjundoes it do anything else?
06:49:22kkurbjundoom that is
06:49:29Paul_The_NerdWell, it looks like it runs through most of the startup process first
06:49:34Paul_The_NerdIt goes by very quickly
06:49:34 Join damaki_ [0] (n=Chocolat@ALille-153-1-24-202.w83-198.abo.wanadoo.fr)
06:50:42 Quit DMJC ("Leaving")
06:51:22linuxstbPaul_The_Nerd: A Data abort means an attempted unaligned memory access.
06:51:33linuxstbAnd yes, the ARM is little-endian.
06:51:50kkurbjunPaul: it's very close to working. just need to get past the wad loading. I'm not sure if it's endianness problem or structure packing right now. I can write some test code later, but I need to sleep soon. Can you send me a diff of what you have so far.. linuxstb: is that a rockbox error then?
06:52:18linuxstbIs what a rockbox error? The "data abort"
06:52:20linuxstb?
06:52:51kkurbjunyes the "data abort" I mean is it somethign that rockbox is monitoring?
06:53:20linuxstbYes, the ARM CPU throws an exception, which Rockbox then catches and gives you that error.
06:53:49kkurbjunhmm, I wonder how idoom got around that
06:54:10kkurbjunit might be my structure packing attributes in doomdata.h
06:54:36linuxstbDo you understand the alignment rules for ARM?
06:54:36kkurbjunhow does the ipod read bmps then?
06:54:44kkurbjunno I don't
06:55:14Paul_The_NerdHow do I make diff compare two directories, as I seem to be horrible when I can't just use CVS diff
06:55:41linuxstb32-bit reads have to be aligned to 4 bytes - so reading a 32-bit value from address 0x10003 is illegal (=data abort), but 0x10004 is legal.
06:55:42kkurbjundiff -ubB doom/ doomold/
06:55:58linuxstbSimilarly, 16-bit reads have to be aligned to 2 bytes - 0x10001 is bad, 0x10002 is good.
06:56:04linuxstbBytes can be read from anywhere.
06:57:03linuxstbThe Archos CPUs (SH) will also fail if you attempt an unaligned memory access. The Coldfire doesn't complain, but the read/write to memory will be slower.
06:58:18kkurbjunlinuxstb: thanks, that's very helpful, I didn't realize that.
06:58:23Paul_The_Nerdhttp://66.68.93.2/doom.diff
06:58:44Paul_The_Nerdlinuxstb has alerted me that the key defines may not work, since I used the wheel ones.
07:00
07:00:24kkurbjunPaul: somethign that you may try is removing all the __attribute__((packed)) from doomdata.h. They are what is changed from idoom. I don't think that they are needed anyway. that was just when I was trying to figure out the graphics errors
07:00:27Paul_The_Nerdkkurbjun: The makefile changes *should* have it do big endian on all the irivers 'cept the 790, and do the ipod-necessary option on the ipods.
07:01:31Paul_The_Nerdlinuxstb: Can I just replace button_status with button_get?
07:03:27 Quit damaki__ (Read error: 110 (Connection timed out))
07:03:48Paul_The_Nerdkkurbjun: I'll try it without the packed then
07:04:35kkurbjunPaul: yeah, I just checked on the H300 and they are not needed
07:04:49Paul_The_NerdWell, we'll see how that affects the ipod. :)
07:05:54kkurbjunthey are also in rdata.c, p_spec.c p_spec.h and d_textur.h
07:06:45 Join damaki__ [0] (n=Chocolat@ALille-153-1-18-68.w83-198.abo.wanadoo.fr)
07:06:53kkurbjunscratch p_spec.h and d_textur.h
07:07:04kkurbjunnevermind
07:07:06Paul_The_NerdOkay
07:07:17Paul_The_NerdSo, just rdata.c p_spec.c and doomdata.h?
07:07:29***Saving seen data "./dancer.seen"
07:07:52kkurbjunno, my find wasn't working correct, they are in p_spec.h and d_textur.h, I spoke too soon
07:07:58kkurbjunas well that is
07:12:33 Quit thegeek (Read error: 104 (Connection reset by peer))
07:13:58kkurbjunpaul: did that make a difference?
07:14:33Paul_The_NerdYes, to an extent
07:14:41Paul_The_NerdGImme a sec
07:14:47Paul_The_NerdNeed to set the Backligh to always on
07:16:02Paul_The_NerdOkay, got past the loading, past the refresh daemon, paused briefly at init_flats, then did a *lot* of stuff very quickly, that was at least as long as the screen (vertically) then another Data Abort, new address
07:16:59Paul_The_NerdoOPS
07:17:07Paul_The_Nerdkkurbjun: But... I somehow missed some of the packeds
07:17:26kkurbjunwhich ones?
07:17:55Paul_The_NerdIn d_textur.h, I just used a find and replace, but these ones didn't have a trailing space.
07:17:58Paul_The_NerdI was being lazy
07:19:55kkurbjunpaul search for "packed" in your doom dir to make sure they're all cleared out
07:20:09kkurbjunI don't think that the ones in d_textur.h will fix it
07:20:09Paul_The_Nerdkkurbjun: That's how I found those two. grep -r "packed" *
07:20:14kkurbjunoh ok
07:20:24Paul_The_NerdThose were the only three left
07:20:54 Quit damaki_ (Read error: 110 (Connection timed out))
07:22:15 Join B4gder [0] (n=daniel@static-213-115-255-230.sme.bredbandsbolaget.se)
07:22:26Paul_The_Nerdkkurbjun: Still gets a data abort though.
07:22:42Paul_The_Nerdkkurbjun: It's different depending on which .wad I load.
07:23:34kkurbjunPaul: in rockdoom.c at the end of the printf command add rb->sleep(HZ*2);
07:23:46kkurbjunso that you can see where the loading is failing
07:23:59kkurbjunbefore the return
07:24:14Paul_The_NerdOk
07:28:39Paul_The_NerdWell, with the shareware, it was in the Refresh Daemon step.
07:30:06Paul_The_NerdKkurbjun: With Doom2.wad present, it gets *past* R_init_planes I believe, but since the text doesn't scroll...
07:30:30Paul_The_NerdThat one shows up partially offscreen.
07:32:10Paul_The_Nerdkkurbjun: Changing the clear screen bit of your printf to be easier to scale for different LCD sizes, and to not draw text halfway off screen. ;-)
07:33:06kkurbjun: ), yeah, it was real basic, just enough that I had a good idea what it was doing
07:33:36kkurbjunPaul: the problem is with r_data.c
07:33:58kkurbjunidoom has some different packing attributes in it that I don't have right now.
07:35:03Paul_The_NerdP_Init: Init playloop state
07:35:05Paul_The_NerdThat's how far it gets
07:35:29kkurbjunoh, maybe not
07:35:46Paul_The_NerdHehehe
07:36:43Paul_The_NerdWell, with Doom1.wad it doesn't get past R_Init for the Refresh Daemon
07:36:52Paul_The_NerdThe P_Init is with Doom2.wad
07:36:58Paul_The_NerdDoom1.wad == Shareware 1.9
07:37:49kkurbjunhttp://pastebin.com/541274
07:38:30kkurbjunPaul: that's the diff between idoom and rockdoom in r_data.c
07:38:43kkurbjunthe + is what idoom has
07:38:57kkurbjunif you fix the stuff up with that it may work
07:39:04kkurbjunI'm stopping for real now
07:39:10kkurbjunit's late here
07:39:30kkurbjunlemme know if you get anywhere with it though
07:39:47Paul_The_NerdBut... those are packed
07:40:24kkurbjunyeah, I can't think right now. It may have something to do with the alignment as linuxstb was saying
07:40:34kkurbjunthey arn't a full packed structure though
07:42:16Paul_The_NerdI'll see what it does, I guess
07:43:17kkurbjunlike I said, it's really close to running. once you get past the alignment stuff it should work fine as long as the button mappings are ok
07:43:36kkurbjunPaul: thanks for trying all this too
07:43:47Paul_The_NerdNo problem
07:43:49Paul_The_NerdI like to help where I can
07:48:45Paul_The_Nerdkkurbjun: Well, that change didn't fix Doom1 SW
07:49:00Paul_The_NerdBut, you should sleep, so you can think.
07:56:25Paul_The_NerdOr doom2, either.:(
07:59:54 Join BHSPitMonkey [0] (n=billybob@67.64.118.242)
08:00
08:05:06 Join hardeep [0] (i=hardeeps@SDF.LONESTAR.ORG)
08:06:42 Join ender` [0] (i=ychat@84.52.165.220)
08:11:41CassandraBagder: You here?
08:11:45B4gderyeps
08:12:39 Quit midkay (Remote closed the connection)
08:13:17CassandraI'm having problems with creating a manual module in CVS. Apparently I need a vendor tag and version, but I have no idea what I should set those to.
08:13:38B4gder"cassandra" and "initial" or similar
08:13:42CassandraSorry, release tag.
08:13:50CassandraOK
08:13:58B4gderthey're rather pointless
08:14:13B4gderI usually import an empty dir
08:14:21B4gderand then cvs add entries like normal
08:15:03CassandraYeah, I thought as much.
08:17:22 Join midkay [0] (n=midkay@24.143.70.99)
08:18:00 Join needleboy [0] (n=needlebo@HFA62-0-184-248.bb.netvision.net.il)
08:24:07Paul_The_Nerdkkurbjun: For when you wake up, the problem seems to be in line 206 of r_things.c, if you see this
08:25:36Cassandrabagder: Can you add manual to the cvs updates list?
08:26:13B4gderwill do
08:27:14CassandraThanks.
08:27:17 Join darkless [0] (n=darkless@cpe.atm2-0-1131059.0x50a648ba.banxx1.customer.tele.dk)
08:27:41CassandraGods messing with the build system internals makes me feel dirty,.
08:28:46CassandraDamn - am2 boards unlikely to be out before summer according to Anandtech. :P
08:37:57Paul_The_Nerdkkurbjun: Oops, previous statement was in error.
08:38:38linuxstbPaul_The_Nerd: Did you ever try Doom in ipodlinux?
08:39:13 Join thegeek [0] (n=thegeek@s026b.studby.ntnu.no)
08:39:15 Join thegeek_ [0] (n=thegeek@s026b.studby.ntnu.no)
08:39:48Paul_The_Nerdlinuxstb: Nope.
08:42:24Paul_The_Nerdlinuxstb: Why?
08:43:27linuxstbI was just curious how it compared with kkurbjun's version.
08:43:47Paul_The_NerdAh
08:43:52Paul_The_NerdWell, I haven't played that yet either.
08:43:53Paul_The_Nerd;-)
08:44:04linuxstbI know :) Did you find enough buttons?
08:44:12Paul_The_NerdThere are plenty
08:45:09Paul_The_NerdOh, linuxstb, can I substituted button_hold for button_status, or do they pass/return different things? (Still working on tracing bugs, so I haven't even looked at that yet.)
08:45:28 Quit thegeek_ ("( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de )")
08:46:27B4gderBryant strikes again
08:47:50linuxstbI was expecting some ARM optimisations.... :)
08:51:52 Join zhilik [0] (n=_if_you_@ppp83-237-251-211.pppoe.mtu-net.ru)
08:53:30linuxstbPaul_The_Nerd: Sorry, I don't know how the hold works.
08:54:06Paul_The_Nerdlinuxstb: No worries. I'll figure it out when it actually matters. :)
08:54:34 Join LinusN [0] (n=linus@labb.contactor.se)
08:57:17*Cassandra idly wonders what the point the score on the builds was.
08:57:31CassandraB4gder, if you mean breaking the builds, that was me, I'm afraid.
08:58:10CassandraImportant lesson: Never commit while sleep deprived. You're bound to miss something.
08:58:41CassandraWhat amazes me is that the manual builds completed. I'd have expected the typo to affect them too. Oh well.
08:59:34CassandraRight - I *must* sleep.
09:00
09:01:36 Quit LinusN (Remote closed the connection)
09:01:46 Join LinusN [0] (n=linus@labb.contactor.se)
09:07:31***Saving seen data "./dancer.seen"
09:10:36B4gderand now we need to autobuild nice manuals...
09:11:08 Join mozetti_ [0] (n=mozetti@p54A8530B.dip.t-dialin.net)
09:11:40LinusNB4gder: go! go! go!
09:12:32 Quit needleboy (Read error: 104 (Connection reset by peer))
09:12:32 Join _FireFly_ [0] (n=icechat5@pd95b7c08.dip0.t-ipconnect.de)
09:13:13 Join needleboy [0] (n=needlebo@HFA62-0-184-248.bb.netvision.net.il)
09:14:23B4gderhttp://daniel.haxx.se/rockbox/rockbox-build.pdf
09:14:29Paul_The_NerdHey, what would *(int *)variablename do?
09:14:33B4gderan example of the h1xx manual I just built
09:14:55LinusNdoes anybody know if the next/prev-dir combos (right + long-right etc) are disabled on the archos pn purpose
09:15:01B4gderPaul_The_Nerd: it typecasts the variable to an int pointer and then read the contents from that address
09:15:02midkayhm. so much unfinished work i could (should) finish.. breakout, rockblox update..
09:15:13Paul_The_NerdB4gder: Okay, that's what I thought. Hrm
09:15:37LinusNmidkay: don't bother about breakout, brickmania is enough
09:15:50 Quit goa (Read error: 104 (Connection reset by peer))
09:15:56midkayoh, right.. well, it isn't running on the recorders, is it? :)
09:16:07 Join goa [0] (i=hd@gate-hannes-tdsl.imos.net)
09:16:12LinusNmidkay: ah, no
09:16:18B4gderno images...
09:16:19LinusNmidkay: port it
09:16:39LinusNquite a challenge on that small lcd :-)
09:16:44midkayLinusN, when i've got a 90% finished and 100% compatible version of my own? ;)
09:16:52LinusNoooh
09:17:12midkayoh. do you know if something's changed with the scrollbar function or something?
09:17:30midkayin the clock plugin credits roll it locks up when the progress bar should be drawn...
09:20:29LinusNthe scrollbar and the progressbar turned into one quite a while ago
09:20:48midkayi remember that but i don't believe that's the problem..
09:21:16Paul_The_Nerdlinuxstb: Well, I've figured out exactly what line is killing Doom, at least at the moment. But I'm not really sure why.
09:22:15midkayoh, and.. i know i can do "case BUTTON_SELECT | BUTTON_RIGHT" but what about defining that somehow? (ie #define BTN_F3 BUTTON_SELECT|BUTTON_RIGHT, but that doesn't work)
09:22:18 Join DarthL4ppy [0] (n=DarthShr@202-161-21-45.dyn.iinet.net.au)
09:23:12Paul_The_Nerdmidkay: You mean like #define DOOMBUTTON_OPEN (BUTTON_SELECT|BUTTON_MENU)
09:23:18 Join Kohlrabi [0] (n=Kohlrabi@dslb-082-083-145-236.pools.arcor-ip.net)
09:24:38midkaythat works? i tried that, iirc..
09:24:49LinusNmidkay: we do that everywhere in rockbox
09:25:09 Join fairway [0] (i=fairway@217-162-176-206.dclient.hispeed.ch)
09:25:11fairwayhi
09:25:15fairwayhow do I use backdrops on ipod?
09:25:16midkaywtf then... would #define BTN_F3 (BUTTON_SELECT | BUTTON_RIGHT) work?
09:25:23LinusNmidkay: yes
09:25:34midkay_wtf_.. well, i guess i need to try it again :) thanks
09:25:35Paul_The_Nerdfairway: Browse to a bmp file, hold select on it, and choose "Set as backdrop." The BMP must be 172*132 pixel.
09:25:41fairwaythanks
09:25:42midkaythat makes things so much easier..
09:26:09midkaya lot easier to port to other devices, i was going nuts with tons of #ifdef CONFIG_KEYPAD == IPOD_4G_PAD :)
09:26:12Paul_The_Nerdfairway if the BMP is in /.rockbox/backdrops/ it'll be saved.
09:26:21BHSPitLappygah sure is late
09:26:26LinusNmidkay: check how it is done in the other plugins
09:26:47 Quit Kohlrabi (Client Quit)
09:26:51midkayLinusN, was going to, i guess i forgot :)
09:27:01midkayor.. i think i checked brickmania..
09:27:09midkaybut didn't see anything there relevant
09:27:35LinusNbrickmania might not be the best example
09:28:02fairwayis the skipping bug fixed?
09:28:03midkayyeah, i was using that as reference a few times though, considering it's the only plugin that runs fullscreen full-color on the ipod videos..
09:28:05fairway;)
09:28:19 Quit TCK (Read error: 104 (Connection reset by peer))
09:28:29BHSPitLappywhat about bejeweled?
09:28:42LinusNwhat about it?
09:28:46 Join Jungti1234 [0] (n=jungti12@124.60.15.86)
09:29:00BHSPitLappydoes it not fall under that category
09:29:25Jungti1234needleboy
09:29:39midkayBHSPitLappy, talking to me?
09:29:46BHSPitLappymmmaybe.
09:29:55midkayi don't think i ever played that, maybe a missing level file or something.
09:30:06BHSPitLappy:/
09:30:29midkayoh, no, it does work..
09:30:30midkayweird
09:30:48 Quit DarthLappy (Read error: 110 (Connection timed out))
09:30:52midkaybut cool.. well, i just chose one, i didn't know bejeweled worked :) or maybe it didn't when i was testing them all
09:31:11 Join HaKuNa [0] (n=c31ce021@labb.contactor.se)
09:31:51Paul_The_Nerdmidkay: Bejeweled's worked on the iPod since before sound did. :-P
09:32:41midkayhaha. i can't tell if you're being sarcastic or not ;) (if not: too bad sound doesn't work on the video ipod!! !!!!!!!!)
09:32:49midkay:)
09:33:20Paul_The_NerdWell, for Nano and Color, it was working before sound.
09:33:45 Quit hardeep ("[BX] Occifer, I'm not as think as you stoned I am!")
09:33:53 Join JdGordon [0] (n=Miranda@c220-239-137-122.smelb1.vic.optusnet.com.au)
09:34:14midkayi don't see video in your short list there ;) nah, i really don't remember.. but i really think i would have noticed if it were working, i went down the list of plugins a few weeks ago, and just updated all of them a few days ago, so.. either it was fixed recently or i accidentally skipped over it..
09:34:35Jungti1234needleboy!!
09:35:24midkay"new target - ipod video" in the cvs history for bejeweled.c on dev 18 05, so i must have skipped over it :)
09:35:48 Quit fairway ()
09:36:15 Join Siku [0] (n=Siku@f303b.w3.ton.tut.fi)
09:36:30Paul_The_NerdHehehe
09:36:35 Quit _FireFly_ ("Clap on! , Clap off! clap@#&$NO CARRIER")
09:36:48Paul_The_NerdOkay, break from beating my head against Doom sources. Gonna set up 'tex
09:37:11midkayfun stuff! and i'll take a nice break from being awake now.. laters all
09:38:09JdGordonlinusN: have u done anything with the colour changer thingy? i tried merging it with the coloer settings patch but that patch was buggered...
09:39:11LinusNJdGordon: no i haven't
09:50:51 Join safetydan [0] (n=dan@195.27.52.9)
09:54:56 Join gtkspert_laptop [0] (n=gtkspert@203-166-235-199.dyn.iinet.net.au)
09:57:46*safetydan sees the EQ UI is popular
09:58:28B4gderyes, seems lots of peple jumped at it immediately
09:58:51ashridahhm. i didn't see it get committed...
09:58:56ashridahpatch someplace?
09:59:10Paul_The_NerdOn the tracker
09:59:29Paul_The_Nerdhttp://sourceforge.net/tracker/index.php?func=detail&aid=1423974&group_id=44306&atid=439120
10:00
10:00:31ashridahtah
10:02:10B4gdersomeone is keeping the wikipedia page fairly up-to-date
10:02:46 Quit Jungti1234 ()
10:03:11B4gderif that somone who edits it reads this: ifp7x0 is _not_ the first flash player to run rockbox, Archos Ondio and iPod Nano are flash players too
10:05:17 Join Zagor [0] (n=bjst@194-237-150-170.customer.telia.com)
10:05:38 Join damaki [0] (n=Chocolat@ALille-153-1-17-105.w83-198.abo.wanadoo.fr)
10:07:25 Join DangerousDan [0] (n=Miranda@newtpulsifer.campus.luth.se)
10:09:55*amiconn appears
10:11:38*amiconn stares at two rows of equally distributed build errors
10:12:11ashridahheh
10:12:26 Join slimx [0] (n=slimx@81.255.106.202)
10:12:46B4gderChristi's attempt to get high score
10:13:10ashridahheh
10:13:18ashridahneed to add rows for the manuals
10:13:26ashridah>:)
10:13:29B4gderhah
10:13:41B4gder*then* she'd get high score ;-)
10:13:50slimxlo
10:13:59B4gdermy test build gave like a bazillion warnings
10:14:10 Part mozetti_
10:14:14*ashridah tosses safetydan and preglow each a cookie of pwnage
10:14:32 Quit mafj2 (Remote closed the connection)
10:14:54amiconnB4gder: What's necessary to build the manual?
10:15:00 Join petur [0] (n=d4efd6a6@labb.contactor.se)
10:15:04amiconnIs it possible on cygwin?
10:15:08B4gderI'm not sure, 'cvs co manual' to start with
10:15:13ashridahamiconn: latex
10:15:20ashridahwhich will be largish
10:15:24ashridahbut probably included in cygwin
10:15:26B4gderI just tried on the build server and it seemed to build
10:15:36Paul_The_NerdWhat targets can the manual be made for?
10:15:40ashridahyou'll need both TeX and LaTeX iirc
10:15:44ashridahPaul_The_Nerd: all of them
10:15:48 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
10:15:50ashridahalthough some are basically identical
10:16:12Paul_The_NerdIpod Nano doesn't work
10:17:27Paul_The_NerdH120 did though, so it's not my setup. Also, is there a step after "make" or is that it?
10:18:48 Quit damaki__ (Read error: 110 (Connection timed out))
10:19:02ashridahdamnit, what's "Q" stand for again?
10:20:32Paul_The_Nerdamiconn: The Cygwin package it Tetex
10:20:51Paul_The_Nerdamiconn: And yes, I've succesfully built the H120 manual in cygwin.
10:21:05amiconnDo I need tetex-extra as well?
10:21:09Paul_The_NerdI dunno
10:21:11Paul_The_NerdI got it
10:21:22Paul_The_NerdDidn't try it without it though, just figure I may as well
10:21:37amiconnI'll try without it first
10:21:40Paul_The_NerdOkay
10:23:10 Join NicoFR [0] (n=nico404@rob92-6-82-231-243-63.fbx.proxad.net)
10:23:36Paul_The_NerdYeah, iPod Nano doesn't work
10:23:39Paul_The_NerdiPod 4G does
10:25:37ashridahamiconn: i don't know. it contains a bunch of fonts and a few other bits
10:28:16safetydanashridah, Q is related to the "width" of the filter
10:28:22 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
10:28:55ashridahsafetydan: yeah, gotchya
10:33:03 Quit BHSPitLappy (Read error: 110 (Connection timed out))
10:37:57 Nick Paul_The_Nerd is now known as Paul_The_Away (n=Paul_The@cpe-66-68-93-2.austin.res.rr.com)
10:42:37 Join muesli__ [0] (n=muesli_t@88.134.20.76)
10:43:11amiconntete-extra *is* needed on cygwin. It contains some necessary fonts
10:43:20amiconn*tetex-extra
10:43:37amiconnManual build works, but throws tons of warnings
10:43:46B4gderyeps, for me too
10:43:55B4gderand I get no pics
10:44:30peturLinusN: if you think of committing the recording gain patch, make sure you take the latest version of my entry in the tracker as I've fixed a warning when building for Archos
10:44:46amiconn'Overfull \box (...pt too wide) in paragraph at lines <blah>'
10:46:04amiconnNo images here as well, except the logo on the front page
10:47:27LinusNpetur: sure
10:47:40 Quit gtkspert_laptop (Read error: 104 (Connection reset by peer))
10:48:21peturbbl
10:48:24 Part petur
10:54:01 Quit muesli__ ("ich will Khe!!!")
10:55:36Paul_The_AwayHrm
10:55:42Paul_The_AwayIs there anyone with an iPod in here, preferably a nano?
10:59:01Paul_The_AwayNevermind. It was a local thing of some sort
10:59:33 Join muesli__ [0] (n=muesli_t@88.134.20.76)
11:00
11:00:27 Join amiconn_ [0] (n=jens@p54BD5D54.dip.t-dialin.net)
11:00:47 Quit amiconn (Nick collision from services.)
11:00:47 Nick amiconn_ is now known as amiconn (n=jens@p54BD5D54.dip.t-dialin.net)
11:07:35***Saving seen data "./dancer.seen"
11:07:59 Quit zhilik (Read error: 110 (Connection timed out))
11:17:30 Join hshah [0] (n=hshah@hirenshah.plus.com)
11:23:44 Join tucoz [0] (n=81b17b04@labb.contactor.se)
11:24:07*tucoz notices the LaTeX docs has been comitted :)
11:24:12 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
11:24:16muesli__preglow r u arround?
11:26:03 Nick paugh is now known as AliasCoffee (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
11:26:55tucozregarding the output when building the manual. LaTeX is a bit more verbose in its output compared to GCC and other compilers
11:27:55tucozan overfull \hbox means that the line, figure, table etc that it refers to is outside of the page margins
11:28:31tucozand I think the only image included now is the frontpage's rockbox logo
11:31:01tucozJust a note, probably the most common and hardest error to track (in my opinion) is when you write an underscore _ . That has to be preceeded by a backslash. I.e. \_
11:34:53 Join gtkspert_laptop [0] (n=gtkspert@203-166-235-199.dyn.iinet.net.au)
11:40:49tucozI just noticed something. If I don't have the m68k-elf-gcc in my path when I run ../tools/configure and select h120 for instance. The make script complains that it can not find that, although I selected manual
11:41:12tucoz*configure script
11:54:15 Join MarcoPolo [0] (n=MarcoPol@che78-2-82-227-240-106.fbx.proxad.net)
11:54:37 Quit MarcoPolo (Remote closed the connection)
11:57:01 Join einhirn [0] (i=Miranda@bsod.rz.tu-clausthal.de)
12:00
12:03:13 Join mikearthur [0] (i=mike@82-41-205-190.cable.ubr11.edin.blueyonder.co.uk)
12:03:56Bg3ris the h3x0 the only player that can charge via USB for now ?
12:04:05 Nick DarthL4ppy is now known as DarthLappy (n=DarthShr@202-161-21-45.dyn.iinet.net.au)
12:05:09amiconnBg3r: Afaik the iPods can charge from USB as well
12:05:19amiconnBg3r: Any news concerning battery_bench ?
12:05:39Bg3ramiconn working on it
12:05:41Bg3r:)
12:05:47Bg3rif i can say "working" ... :)
12:06:08Bg3ranyone against adding this
12:06:09Bg3rbool acdc_charger_inserted(void)
12:06:10Bg3r{
12:06:10DBUGEnqueued KICK Bg3r
12:06:10Bg3r#if defined(IRIVER_H300_SERIES)
12:06:10Bg3r return (GPIO1_READ & 0x00400000);
12:06:10***Alert Mode level 1
12:06:10Bg3r#endif
12:06:10***Alert Mode level 2
12:06:10Bg3r}
12:06:20Bg3rto drivers/power.c ?
12:06:23amiconn?
12:06:40amiconnThere already is a function to check for inserted charger
12:07:09Bg3ramiconn yep, but on h300 it returns
12:07:15Bg3r#elif defined(IRIVER_H300_SERIES)
12:07:16Bg3r return (GPIO1_READ & 0x00400000) || usb_powered();
12:07:25amiconnYes, that's wrong
12:07:34Bg3rah ?
12:07:45amiconnIt's the reasoon why inserting USB power shows the standard plug
12:07:58Bg3rwhy wrong ?
12:08:02amiconnIt should show the USB plug, like on Ondio/FM recorder
12:08:07 Nick Paul_The_Away is now known as Paul_The_Nerd (n=Paul_The@cpe-66-68-93-2.austin.res.rr.com)
12:08:15Bg3rhm ... didn't know this
12:08:46Bg3rshould i remove it ?
12:09:00Bg3rthe || usb_powered() part
12:10:40amiconnI think so. The charging detection will perhaps need to be adapted
12:10:52amiconnIirc, lostlogic added this
12:11:14Bg3ryep
12:11:18Bg3riirc too
12:12:00CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
12:12:00*amiconn is stress-testing & timing memmove()
12:12:15Bg3rgood to hear :)
12:12:22 Join DarthL4ppy [0] (n=DarthShr@202-161-21-45.dyn.iinet.net.au)
12:13:12Bg3rbtw this "usb_powered" isn't equal to "usb_charging" or i'm wrong ?
12:13:17 Quit DarthLappy (Nick collision from services.)
12:14:28Bg3rlike powered != charging
12:14:33 Join Membrillo [0] (n=sam_kill@CPE-60-229-179-140.nsw.bigpond.net.au)
12:15:14amiconnusb_powered is equivalent to charger_inserted, power-wise
12:15:19 Nick DarthL4ppy is now known as DarthLappy (n=DarthShr@202-161-21-45.dyn.iinet.net.au)
12:15:34amiconnWhether it's actually charging depends on the charging circuit of course
12:15:47Bg3rokay
12:16:11***Alert Mode OFF
12:16:21 Quit markun ("leaving")
12:16:25Bg3rso New/FM/V2 recorders can charge via usb ?
12:16:37Bg3r-New
12:16:44amiconnyes
12:17:05amiconnand the Ondio can be powered from USB (it has no charging at all)
12:17:15Bg3rah, just to ask:)
12:18:05amiconnThe Ondio is designed to run from alkalines
12:18:10amiconn3x AAA
12:18:42Bg3rafk...
12:20:22 Join webguest44 [0] (n=5687a08f@labb.contactor.se)
12:24:52 Quit webguest44 (Client Quit)
12:25:05 Join ripnetuk [0] (n=george@host-87-74-125-82.bulldogdsl.com)
12:25:35 Join webguest54 [0] (n=86e20188@labb.contactor.se)
12:25:36ripnetukhi... is anything known about the Archos AV500? (as in is a Rockbox port a therotical possibility)?
12:25:42 Join markun [0] (n=karl@bastards.student.ipv6.utwente.nl)
12:26:17 Quit webguest54 (Client Quit)
12:28:06 Quit gtkspert_laptop (Read error: 104 (Connection reset by peer))
12:28:48tucozripnetuk: doesn't that run linux?
12:28:59ripnetukdoes it? i know the iStation V43 does
12:29:21 Quit Membrillo ()
12:29:23ripnetukif it does /me buys now :)
12:29:43tucozI think I read that somewhere. Don't know for sure though
12:29:58ripnetuki guess we could run the Rockbox SIM on linux based devices like the V43, but I thought the Archos was non-linux... i will google
12:30:33tucozcheck this as well: http://www.archopen.org/tiki-index.php
12:32:15 Join Moos [0] (i=DrMoos@m53.net81-66-159.noos.fr)
12:32:28ripnetukthanks tucoz - looks interesting
12:33:02tucozat least that project could possibly kick alive the gmini port
12:33:21 Quit AliasCoffee (Read error: 104 (Connection reset by peer))
12:36:32 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
12:37:02 Quit paugh (Remote closed the connection)
12:41:51 Join Matze [0] (i=Miranda@p5484E996.dip.t-dialin.net)
12:44:39 Part ripnetuk
12:46:37 Quit darkless (Client Quit)
12:47:08 Quit markun ("leaving")
12:48:18JdGordonim boooored
12:49:59amiconnLinusN: Imho it would be way easier to use a 16 bit framebuffer for X5, and convert on the fly in lcd_write_data()
12:52:32 Join markun [0] (n=karl@bastards.student.ipv6.utwente.nl)
12:52:35 Part tucoz
12:53:42*safetydan wonders if he can just do dsp->replaygain += get_replaygain_int(eq_max_gain * 10) >> 23
12:55:25linuxstbslimx: Are you around?
12:56:43CassandraOf course you can - the question is will it do what you wanted?
12:58:56Paul_The_NerdCassandra: Manual doesn't work for Nano
12:59:29linuxstbCassandra: Nice work on committing the manual. What's the status on the content? Is it mostly your 2.4 manual at the moment?
13:00
13:01:22Paul_The_NerdScratch that, it *does* work, kinda. It gives a message, and waits for user input, then if you just hit enter it still creates a pdf.
13:02:17linuxstbSounds like it's checking for an undefined variable - which is why it's prompting the user.
13:03:12Paul_The_NerdRunaway argument?
13:03:29Paul_The_Nerd./preamble.tex:12: File ended while scanning use of \@argdef.
13:03:57Paul_The_NerdAlso, are these manuals supposed to be... target specific already, or are they all pretty much identical in text right now?
13:07:31amiconnThere are currently lots of old references in the manual. It's the first step towards a modular manual
13:07:37***Saving seen data "./dancer.seen"
13:07:57amiconnE.g. if you build the manual for recorder, there is also a table describing the player buttons
13:08:00Paul_The_NerdOkay
13:08:15Paul_The_NerdI just saw the Rombox section.
13:08:23amiconnThere are also references to the flash packages at [IDC]Dragon's homepage
13:08:38amiconnThey are long gone, and we have bootbox for quite a while
13:09:02Paul_The_NerdFair enough.
13:09:22CassandraPaul_The_Nerd, what happens?
13:09:42Cassandralinuxstb, yes - it's converted from the 2.4 manual
13:10:09Paul_The_Nerd(./platform/ipodnano.tex)
13:10:10Paul_The_NerdRunaway argument?
13:10:10Paul_The_Nerd{\playertype ]
13:10:10DBUGEnqueued KICK Paul_The_Nerd
13:10:10Paul_The_Nerd./preamble.tex:12: File ended while scanning use of \@argdef.
13:10:10Paul_The_Nerd<inserted text>
13:10:10***Alert Mode level 1
13:10:10Paul_The_Nerd \par
13:10:12Paul_The_Nerdl.12 \input{platform/\platform.tex}
13:10:18Paul_The_NerdThen a ? prompt
13:10:29CassandraPaul_The_Nerd, no, the target specific stuff is barely started.
13:10:29Bg3rPaul_The_Nerd : enter
13:10:32Bg3rjust hit enter
13:10:39Paul_The_NerdI did hit enter
13:10:51CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
13:10:51*Cassandra looks confused.
13:11:05CassandraYou've gotta love LaTeX error messages.
13:11:22Paul_The_NerdBg3r: I hit enter, and it finished normally (I think) but other targets don't have that
13:13:38LinusNamiconn: i don't know about easier, the conversion is not straightforward, unless we use a lookup table
13:15:19CassandraThanks Paul, fixed.
13:15:34Paul_The_Nerd:)
13:15:57CassandraThe manual is over a year old, it would seem.
13:16:05CassandraIt doesn't seem like that long.
13:20:11***Alert Mode OFF
13:20:36linuxstbCassandra: Is there a reason you're trying to keep the number of different manuals small? I would have thought that now you've automated it, it would be simpler to produce one manual for each target - using the same set of targets that the main Rockbox build system uses.
13:30:26 Join zhilik [0] (n=_if_you_@ppp83-237-251-112.pppoe.mtu-net.ru)
13:31:32JdGordonlinuxstb: anything happening with the yesno screen?
13:35:04linuxstbThere was a discussion about it yesterday - check the logs from 14.32.56 onwards. (about 2 minutes after you left...)
13:35:13JdGordonok
13:36:41ashridahcursed timezones
13:36:49*ashridah takes out a hammer and starts whacking on the ground
13:36:57ashridahyou'll be flat once i'm done! i'll show you!
13:39:28 Join mafj2 [0] (i=debian-t@tor/session/x-f66937ec95b04165)
13:40:13JdGordonhmm... so undecided...?
13:41:33 Join actionshrimp [0] (i=nn@dhcp-163-1-214-233.seh.ox.ac.uk)
13:44:33 Join damaki_ [0] (n=Chocolat@ALille-153-1-46-98.w83-198.abo.wanadoo.fr)
13:45:38 Join petur [0] (n=d4efd6a6@labb.contactor.se)
13:48:22 Quit ashridah ("sleep")
13:56:06 Quit damaki (Read error: 110 (Connection timed out))
13:56:50 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
14:00
14:03:15Papricalinuxstb, have you try the statusbar patch?
14:06:25Papricamm there is a way to stop the compile?
14:06:35JdGordonctrl-c ?
14:06:53Papricaok, i will try
14:06:56Paprica10x =]
14:09:10 Quit JdGordon ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
14:09:21 Join mirak [0] (n=mirak@AAubervilliers-152-1-9-93.w82-121.abo.wanadoo.fr)
14:12:20 Quit Matze ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
14:12:54needleboyhey ben
14:13:51 Join perpleXa [0] (i=perpleXa@62.216.185.132)
14:14:02perpleXahi
14:14:34perpleXawill there ever be support for ipod minis?
14:17:08Paul_The_NerdperpleXa: Ideally, anything that iPod Linux runs on could see Rockbox so long as anyone takes the effort.
14:18:14perpleXahmm nice, i'm currently running ipod linux on it, only problem is that it spews a malloc error when i try to play mp3's of ~70mb
14:18:28LinusNhehe
14:18:34amiconnhaha
14:18:41amiconnmalloc() ...
14:19:19*Paul_The_Nerd grins.
14:19:20perpleXabtw, is rockbox able to read apev2 tags?
14:23:11Moosnot with mp3's files yet
14:23:37 Quit NicoFR ()
14:25:00perpleXa:(
14:25:28Paul_The_NerdI don't think anyone's made a convincing argument for them other than "I like them better" or "My files are already tagged with them."
14:26:25 Join ripnetuk [0] (n=george@host-87-74-125-82.bulldogdsl.com)
14:26:37MoosPaul_The_Nerd: full tag support is suficent no? :P
14:27:23Papricaneedleboy hi =]
14:27:23perpleXawell i dont like idev2 tags, since they're written at the beginning of files
14:27:32perpleXathats fucking slow with foobar
14:28:11Paul_The_NerdWell, the beginning of the file is kinda better for reading them (on machines with limited RAM) I imagine.
14:28:14amiconnTags at the beginning are way better on a mobile device
14:28:29MoosperpleXa: just convert them to id3 v2
14:28:35amiconnIt saves the seek-to-end and back when reading tags
14:28:43Paul_The_NerdBesides, it's not like you should have to retag your files very often.
14:28:55amiconn-> needs less battery
14:29:05perpleXathat takes like 3 days to convert all my mp3 tags to id3v2
14:29:18MoosOuch :-(
14:29:20LinusNyes
14:29:27 Join Kohlrabi [0] (n=Kohlrabi@dslb-082-083-142-087.pools.arcor-ip.net)
14:29:47Paul_The_NerdperpleXa: Break it up into 6 groups of files, and run the operations while you sleep.
14:29:54MoosperpleXa: there is a patch in the tracker for ape V2 tags for mp3's
14:30:20perpleXai'm using foobar plugins like play counter and such shit, if i would use id3v2 tags it would freeze ~1min after each track
14:30:40perpleXai hope they fix that, so i can use id3v2 tags =)
14:30:59Moosuse other tagger instead ;-)
14:30:59perpleXaMoos: could you gimme the link, please?
14:31:03Paul_The_NerdEvery time, or just the first time it rights the tag?
14:31:13perpleXaevery time
14:31:14Paul_The_Nerd*writes
14:32:30 Join dufuS [0] (n=hohoho@DSL217-132-191-102.bb.netvision.net.il)
14:32:53 Nick dufuS is now known as san_benedetto (n=hohoho@DSL217-132-191-102.bb.netvision.net.il)
14:33:08san_benedettohi
14:33:12san_benedettoneedleboy ?
14:33:53 Join damaki [0] (n=Chocolat@ALille-153-1-62-174.w86-196.abo.wanadoo.fr)
14:33:53perpleXais it currently running on ipod minis too?
14:34:20Paul_The_NerdIs what?
14:34:41san_benedettohe means rockbox i think
14:35:18 Nick san_benedetto is now known as San_Benedetto (n=hohoho@DSL217-132-191-102.bb.netvision.net.il)
14:35:27perpleXayeah ofc
14:35:28perpleXa:)
14:36:22Paul_The_NerdWell, you already asked if Minis would be supported.
14:36:39Paul_The_NerdWhen I said they could potentially be supported, I think that kinda implied they aren't yet, so I was confused as to why you asked again.
14:36:43perpleXai asked if they will be supported, didnt know that they already
14:36:54perpleXaah
14:36:55perpleXak :P
14:37:01MoosperpleXa:http://sourceforge.net/tracker/index.php?func=detail&aid=1245274&group_id=44306&atid=439120
14:37:02Paul_The_NerdRight now it's Nano and 4G Color
14:37:13perpleXayeah moos, already found it, thanks
14:37:19Moosnp
14:37:33perpleXahmm, so i gonna buy a nano :X
14:39:15muesli__buy a nano buy scratches ;)
14:39:20Paul_The_NerdYeah
14:39:28Paul_The_NerdIf you buy a nano, figure out some way to never let ANYTHING touch it
14:39:29Paul_The_NerdEVER
14:39:39LinusNi'm about to introduce a new directory structure in firmware/
14:39:48ripnetukhold it in a electromagnetic field like they do anti-matter
14:40:04Paul_The_Nerdemf + electronics = fun? ;-)
14:40:05perpleXayeah i've already seen that issues
14:40:05LinusNi'm thinking about adding a directory tree like this:
14:40:22LinusNfirmware/target/coldfire/iaudio/x5
14:40:24perpleXaa friend has a 3 months old nano, that looks worse than my 2 years old mini
14:40:31ripnetukim thinking of buying a archos av500 - anyone know how hackable it is?
14:40:52Paul_The_NerdLinusN: It'd be nice, I think.
14:40:53LinusNall coldfire-specific code, like asm versions of memset() etc goes to target/coldfire
14:40:55Papricagui/statusbar.c: In function `gui_statusbar_draw':
14:40:55Papricagui/statusbar.c:181: warning: statement with no effect
14:41:00Papricawtf it wants?
14:41:16LinusNall iaudio common functions go to target/coldfire/iaudio
14:41:27SlasheriPaprica: you have something like "variable;"
14:41:32LinusNand the x5 specific stuff to target/coldfire/iaudio/x5
14:41:38SlasheriPaprica: and that does nothing
14:41:50Papricaoh shit
14:41:55Papricaforgot to do
14:42:02Papricafunction() with the ()
14:42:04Paprica=\
14:42:06Paprica10x
14:42:07Slasherihehe
14:42:07MoosLinusN: wee !
14:42:08LinusNfor example, the ata_enable(), ata_reset(9 etc
14:42:34LinusNthat may clear some of the #ifdef hell we're currently experiencing
14:43:15peturas long as you don't end up with a load of directories with few files in each...
14:43:30LinusNripnetuk: av500 has a texas instruments dsp i think
14:43:36amiconnSounds good, but may be problematic when considering the sims as well
14:43:57LinusNso you can forget it unless you can cough up loadza dollars for the compiler
14:44:11LinusNamiconn: why so?
14:45:11amiconnWell, some target specific code (which would e.g. be buried in firmware/target/coldfire/h300) is needed for the sims as well,
14:45:13 Join __peer__ [0] (n=hrm@bzq-218-213-161.red.bezeqint.net)
14:45:20__peer__hello, anyone had exprience with PX-100 ?
14:45:23LinusNamiconn: give me an example
14:45:32LinusN__peer__: what is that?
14:45:41__peer__Portable Headphones
14:45:49ripnetukis the texas instruments the one missing an open compiler?
14:46:01LinusNripnetuk: as well as data sheets
14:46:13ripnetukok... that makes it easy to knock off my list - thanks :)
14:46:25amiconnLinusN: Things that are simulated, e.g. backlight
14:46:29LinusNav500 is a cool machine nevertheless
14:46:42LinusNamiconn: but that's still simulated
14:46:51Zagor__peer__: i've tested them. sounds pretty good. a lot better than the px-200!
14:46:59LinusNthis directory structure is only for target builds
14:47:00 Quit damaki_ (Read error: 110 (Connection timed out))
14:47:02amiconnWe could split those things into the common part and the actual platform specific code,
14:47:27amiconnbut I'm afraid that will just replace ifdef hell with file hell
14:47:28__peer__Zagor: Will I feel much diffrence between them and cheap 5$ headphones?
14:47:35__peer__They mostly for home use.
14:47:39LinusNamiconn: the backlight code in targets/coldfire/iriver/h300 will only be the lowest level code
14:47:40ripnetukyes, but as soon as Archos moves onto another model, we can forget about software updates... also I am addicted to Rockbox features (multiple independent bookmarks mostly)
14:47:49Zagor__peer__: yes you will
14:48:16__peer__they worth 70$?
14:48:28ripnetuki was thinking of the istation v43 - that runs Linux, so i should be able to port the Rockbox sim to it hopefully
14:48:28amiconnLinusN: Another disadvantage would be that the lowest level acces will have to be done by functions -> more overhead
14:48:34peturLinusN: again, will you not end up with many directories with very little files in?
14:48:41LinusNamiconn: could be macros as well
14:48:50LinusNpetur: yes
14:48:55amiconn(unless we use .h files with loadsa macros, or inline functions)
14:49:15amiconnMacros aren't exactly easy to read
14:49:54LinusNthe advantages would be that porting will be a more straightforward process, you can start with empty stubs for the lowlevel functions
14:50:20LinusNand you (almost) always know where to look for the hardware specific functions
14:50:28LinusNfor a particular target
14:50:49amiconnThat's correct. Well, it seems this is something that has to be tried
14:51:12amiconnIt also means that someone will have to cleanup stone-age leftovers
14:51:16LinusNi'm trying it for the x5 port
14:51:27LinusNamiconn: yes
14:51:57slimxlinuxstb, are you around ?
14:52:47 Join tucoz [0] (n=81b17b04@labb.contactor.se)
14:53:07tucozHi, what would a suitable name for Latex help be (in the wiki)?
14:53:22tucozLatexHowto?
14:53:31LinusNdefine "help"
14:53:32B4gderHowToBuildManual ?
14:53:58tucozI mean, to help people working with the latex docs
14:54:03LinusNperhaps ManualHowto, to follow the howto convention
14:54:24tucozLike tips and tricks, pointers to latex docs etc
14:54:46B4gderbtw, what do we need to do to get images in the manual?
14:54:52amiconnLinusN: I went for a slightly different approach for integrating memmove() / memcpy()
14:55:07LinusNamiconn: yes?
14:55:11tucozB4gder, hold on. I'll make paste for you
14:55:31amiconnI made the asm versions two files, with memmove() jumping into the forward branch in memcpy_a.S when forward copying has to be done
14:55:50amiconnA comment at the top will point this out
14:56:28amiconnThis has the advantage that memmove_a.o doesn't need to be linked when memmove() isn't used,
14:56:29LinusNthose two would end up in target/coldfire in the new structure
14:56:49amiconnNo, those 2 will need to be splitted in two each,
14:56:59amiconnand end up in target/coldfire and target/sh1
14:57:05LinusNyes, of course
14:57:23*B4gder is in favour of splitting ifdefs into files
14:57:40amiconnI added memmove.c from newlib, as it seems memcpy.c is also from newlib
14:57:48amiconn(for arm targets and sims)
14:58:50 Quit HaKuNa ("CGI:IRC (EOF)")
14:59:42tucozhttp://pastebin.com/541555
15:00
15:00:03tucozB4gder, have a look at that and ask me if there is something you don't understand
15:00:14imphasingYum, postscript..
15:00:21B4gderwell, my problem is not really the latex code
15:00:28B4gderI don't get any pics in the manual
15:00:31B4gderwhen I build it
15:00:35LinusNtucoz: well, i think he's asking how to make the already existing pics appear in the manual
15:00:54B4gderthe only pics that appears is the front page logo
15:01:03tucozaha, are those present?
15:01:03B4gders/pics/pic
15:01:38tucozthe pictures I mean. When I set up the proposal, I only included the frontpage image. The others were gifs or jpegs iirc
15:01:40B4gderhm, maybe they're just missing...
15:02:06B4gderso how is the manual supposed to know how to get the gifs and jpegs?
15:02:36tucozI don't think those formats should be used. Instead, png was suggested as the standard format
15:02:50B4gderso this is still left todo?
15:02:54tucozyes
15:03:02B4gderokiedoki
15:03:03tucozlot's of screendumps to be made
15:03:14tucozrather _loads_
15:03:18B4gderI figure a bunch of the old ones could just be converted
15:03:59tucozyes, maybe Cassandra have the originals. I think we could find something in the writer-source as well
15:04:19LinusNi guess the old dumps are obsolete anyway
15:04:26B4gderthey might be so, yes
15:04:37LinusNgotta go, cu folx
15:04:38B4gdergetting screendumps should be quickly done though
15:04:39tucozBut, pdflatex should handle png and jpg's fine
15:04:44 Part LinusN
15:05:09B4gderI guess I'm using tetex
15:05:52B4gderah, I have a pdflatex in my path
15:06:03tucozthat is ok. I think the buildscript runs pdflatex to generate the .pdf
15:06:44B4gderI get a ~1MB pdf so I guess most things work
15:07:11tucozyes, that sounds ok. The tables need to be taken care of, as they look like ... now.
15:07:39***Saving seen data "./dancer.seen"
15:07:41B4gderI'm preparing for automated manual builds
15:07:48tucozlatex keyword in google generally comes up with something nice.
15:08:21 Join bandgeekndb [0] (n=d19e6629@labb.contactor.se)
15:08:34amiconnBagder: What? automated manual sounds a bit undecided ;)
15:08:48Paul_The_NerdYeah, that phrase concerned me.
15:09:22amiconnWow! This 600-lines asm monster is actually working :-)
15:09:26B4gderright, we should instead just write all the docs now to describe how rockbox should work in the end and then just implemented it so ;-)
15:09:32B4gderjust like we use to
15:10:18 Nick paugh is now known as AliasCoffee (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
15:11:55bandgeekndbquick question if someone has a minute - i finally realized why one of my music folders wasn't showing up under "file-tree" navigation. The album name (also the folder name) started with a period, which i've now learned hides folders in Rockbox (took me long enough, I know)
15:12:11bandgeekndbwhen i try to rename another folder in windows with a period in front, it rejects the new folder name
15:12:24B4gderuse the command line prompt
15:12:36bandgeekndbah
15:12:37bandgeekndbthank you
15:12:42B4gderthat's the "windows explorer" deciding you may not do that
15:14:06bandgeekndbany way to do that w/o the cmd line prompt - access to the cmd line prompt has been disabled on my school computer
15:14:23bandgeekndbstupid IT department thinks we're a bunch of kiddy hackers and doesn't trust us
15:14:26B4gderI don't know
15:14:38B4gderI prefer Linux ;-)
15:14:52tucozwe could probably snatch a good deal of screenshots from the wiki as a start
15:15:01bandgeekndbi can do that when I get home i guess
15:15:11tucozthose that are up-to-date and in png.
15:15:12bandgeekndbalthough there, i have cmd-line access as well
15:15:24bandgeekndboh well. thanks anyway B4gder
15:15:28B4gdertucoz: yes, or just go screenshot mode and make a whole bunch at once playing around on the player
15:15:36B4gdernp
15:15:40tucozand that, yes
15:16:03 Part bandgeekndb
15:16:52safetydanor... he could change the view setting to "All" and rename it in rockbox
15:16:54safetydanbut he's gone now
15:17:13Paul_The_Nerdbandgeekndb: Most school computers let you get to the command prompt by typing cmd at start->run or hitting CTRL-ALT-DEL and choosing Run from the file menu
15:17:20Paul_The_NerdUrk, he left
15:17:22Paul_The_NerdI am slow
15:17:48Paul_The_NerdI've found that school "tech" departments tend to take the lazy way out, and in most cases just delete the icon.
15:17:49tucozisn't it also possible to use " " around a filename in windows
15:18:16tucozto sort of force it to be that name.
15:19:15peturnope, just to pass names with spaces on a commandline (amongst others)
15:20:27tucozok
15:22:31 Part tucoz
15:28:36San_Benedettohey
15:28:41San_Benedettowhere do i put
15:28:46San_Benedettoprogressbar.patch
15:28:47San_Benedetto?
15:29:02 Join TCK [0] (n=tckocr@81-178-216-103.dsl.pipex.com)
15:29:36Paul_The_NerdDo you have a development environment set up?
15:29:42safetydanHave you checked out the source?
15:30:11San_Benedettowhat source?
15:30:20San_Benedettoi don't understand
15:30:24San_Benedettoi just want to install the
15:30:40Paul_The_NerdTo use patches, you have to have the source code and a means of compiling it.
15:30:42San_Benedettoipodcolor wps
15:30:44B4gderhttp://www.rockbox.org/twiki/bin/view/Main/WorkingWithPatches
15:31:02San_Benedettodo i need to do anything to use the ipodcolor wps?
15:31:14B4gderread that URL
15:31:31 Join tvelocity [0] (n=tony@ipa33.0.tellas.gr)
15:31:39B4gderI have no idea what you'd need to do or not
15:32:07Paul_The_NerdDoes the WPS creator say you need anything specific?
15:32:08B4gderthere's no ipodcolor wps in the patch tracker afaik
15:32:25 Join darkless [0] (n=darkless@62.79.44.48.adsl.vby.tiscali.dk)
15:32:28San_Benedettoin the misticriver forum there is
15:32:36B4gderthen ask there
15:32:49B4gderif they don't bother to show us, then how can we know?
15:33:11 Quit darkless (Read error: 104 (Connection reset by peer))
15:33:34 Join darkless [0] (n=darkless@62.79.44.48.adsl.vby.tiscali.dk)
15:39:54 Quit DarthLappy (Success)
15:40:31 Quit TCK (Read error: 104 (Connection reset by peer))
15:40:47 Join Matze [0] (i=Miranda@p5484E996.dip.t-dialin.net)
15:42:00PapricaSan_Benedetto, no
15:43:54XavierGrYay! :) I just received an email that iriver got my defective player and they are working on it!
15:44:04Bg3rwow :)
15:44:07Bg3rvery good, XavierGr
15:44:12Papricahaha =]
15:44:33novimonare there any rss readers / "podcasting software" to iriver
15:45:15 Join TCK- [0] (n=tckocr@81-178-216-103.dsl.pipex.com)
15:47:13Paul_The_Nerdnovimon: Any of them that work with UMS devices.
15:52:58*preglow returns
15:53:07Paul_The_NerdPreglow!!
15:53:53amiconnhi preglow
15:54:11preglowhello
15:54:16novimonfound one, its called happyfish, looks pretty good
15:54:29preglowwow, the eq gui has really shaped up
15:54:36 Quit B4gder ("time to say moo")
15:54:39Paul_The_NerdYeah
15:54:40amiconnpreglow: You returned just at the right time. Do you know whether the codecs use memmove() a lot?
15:54:48XavierGrpreglow any commits comming up?
15:54:54Paul_The_NerdBut word is there's bugs relating to eq + resampling 22100 files.
15:54:55*amiconn is about to commit freaky asm memmove() for coldfire & sh1
15:55:08needleboyPaprica, what's up with the scrolling margins patch?
15:55:12XavierGramiconn any optimizations?
15:55:24needleboypeople are close to cut my head off because of it :)
15:56:06amiconnThe fun thing is that backwards copy ended up even a bit faster than forward copy
15:56:11Papricaok, i will work on it now..
15:56:42needleboycool man, you're a god
15:57:17preglowamiconn: ooooh
15:57:23preglowamiconn: some of them doi
15:57:27preglowamiconn: musepack, among others
15:57:53 Part petur
15:59:26amiconnpreglow: I have no musepack files at all. Could/would you do a speed comparison with & without asm memmove() on iriver?
16:00
16:00:21amiconnI'll commit in pieces. memmove() itself first, then wiring it to codecs and plugins
16:00:35preglowamiconn: i'll try and see
16:00:44amiconnSome parts in the core can also be changed to use memmove()
16:00:57preglowamiconn: good reason to test out lostlogics new codec null output driver as well
16:00:59amiconnOne place I know is the font cache
16:03:39safetydanyay, preglow is back
16:03:45safetydannow you can tell me the things I got wrong
16:04:54preglowamiconn: just had a check, only faad, libmad and musepack use memmove, and of those i think musepack does it the most
16:05:05preglowsafetydan: i'll try the gui out right now
16:05:17preglowstrange, when i tested my eq right before i left, i could only get noise out of it
16:05:20amiconnfadd is no longer used
16:05:20preglownoise and sines
16:05:27preglowamiconn: what?
16:05:37amiconnBla, noncsense
16:05:37safetydanI think that was because I had gain and q swapped at one point
16:05:47preglowsafetydan: that might be it...
16:05:55*amiconn should think before typing
16:06:20amiconnI confused libfaad and libflac
16:07:27San_Benedettohey
16:07:36San_Benedettocan someone tell me what are .rwps files?
16:08:03Papricamm
16:08:16Papricasomeone can help me with the marginds patch
16:08:21Paprica?
16:08:38safetydanSan_Benedetto: Remote WPS files
16:09:26Paprica.....????/
16:09:57 Join damaki_ [0] (n=Chocolat@ALille-153-1-78-160.w86-207.abo.wanadoo.fr)
16:10:10safetydanPaprica: I'd take a look but I'm at work and can't really do rockbox stuff
16:10:33Paprica=\
16:11:07 Quit TCK- (Read error: 110 (Connection timed out))
16:14:55Papricawooopiii =]
16:17:07 Quit mirak (Remote closed the connection)
16:18:23 Quit needleboy ()
16:20:10San_BenedettoPaprica
16:20:16Paprica?
16:20:23San_Benedettomake a new game
16:20:24San_Benedetto:P
16:21:16Papricalol
16:21:17Paprica=]
16:21:27Papricaim working on the gui now..
16:21:49 Quit damaki (Read error: 110 (Connection timed out))
16:21:54preglowsafetydan: i think this is very good, but it's very tedious to use, thanks to the small step size
16:22:04preglowsafetydan: perhaps some step acceleration is in order?
16:23:41 Quit DreamTactix291 (Read error: 104 (Connection reset by peer))
16:24:12 Join DreamTactix291 [0] (n=DreamTac@adsl-149-149-154.bna.bellsouth.net)
16:24:40preglowand anything on how the eq actually sounds? i haven't used it much myself
16:25:17safetydanapparently it sounds fine
16:25:28safetydanneeds a pre-amp to stop clipping though
16:25:32preglowyes, indeed
16:25:38preglowi think the gain sounds too weak
16:25:39safetydanand step acceleration is on the TODO list
16:25:41preglowbut i can measure that
16:26:05preglowrockbox really needs a generic function or something for doing accelartion
16:26:08preglowif at all possible
16:26:13safetydanhttp://misticriver.net/showthread.php?t=36519 should give you some idea of what people think
16:26:13preglowi'd also love to use it in the pitch screen
16:26:20preglowok, will read
16:26:33amiconnThe button driver already does repeat acceleration
16:26:49safetydanI should probably make more of a distinction between shelf and peak filters
16:27:55preglowahh, yes
16:28:13preglowamiconn: a two-step thing?
16:28:16safetydanOh, there was one guy saying that the EQ doesn't work on 22Khz files
16:28:29safetydanhttp://forums.rockbox.org/index.php?topic=1887.msg17801#msg17801
16:28:34preglowsafetydan: i know, that's weird, sample rate shouldn't matter at all for the core eq code
16:28:50preglowthe only place you need the sample rate, is when calculating the cutoff value to send to eq_*_coefs
16:29:14preglowhowever, i think i know what the bug is
16:29:22preglowyou don't cap frequency based on sample rate
16:29:32pregloweq freq can be maximum samplerate/2
16:29:45preglowif this is not true, the cutoff calculation will overflow
16:30:05amiconnEQ is applied before resampling?
16:30:06safetydanah, of course
16:30:13preglowamiconn: no, after
16:30:27preglowwait...
16:30:35safetydanQ should also be limited to 1.0 for the shelf filters shouldn't it?
16:30:46preglowyes, after
16:31:04preglowsafetydan: it doesn't make sense to use anything but 0.707 for shelving filters
16:31:24preglowsafetydan: what would happen if you used a higher value, is that the shelving filters would get some peaking tendencies around the cutoff freq
16:33:12 Join ep0ch [0] (n=ep0ch@84.12.178.113)
16:33:24preglowbut yeah, amiconn has a good point, resampling is done before the eq
16:33:53ep0chi cant compile with the eq_patch anymore :(
16:34:34ep0chi get errors like "eq_menu.c:2201: error: redefinition of 'eq_menu'"
16:34:46safetydanodd
16:34:49safetydanWhich patch release?
16:35:04ep0ch11
16:35:04 Join tucoz [0] (n=81b17b04@labb.contactor.se)
16:35:22ep0chi had the same with version 10 before
16:35:30Paul_The_Nerdsafetydan: That is a minor problem if you -R one of the earlier version, then apply a later one, funkyness can ensue, as it doesn't remove those files.
16:35:38tucozAny wiki gurus around? I try to write LaTeX in the wiki, but I keep getting a LaTeX? text when I preview
16:35:54Paul_The_Nerdep0ch: patch -R to get rid of it, delete eq_menu.c and .h, and then apply the patch again, and see if that fixes it.
16:36:02ep0chk
16:36:18amiconntucoz: That's because LaTeX is recognised as a wiki word
16:36:41amiconnEither you add a description for it afterwards, then it will turn into an ordinary wiki link
16:36:45tucozamiconn: Ok, is it possible to tell it that it isn't a wiki word?
16:36:58josh_tucoz: try !LaTeX
16:36:59amiconnIf you don't want that, you'll have to write <nop>LaTeX
16:37:05safetydanhttp://www.rockbox.org/twiki/bin/view/TWiki/WikiSyntax
16:37:10tucozthanks :)
16:37:57safetydanpreglow: someone also asked for a lower, lower bound on Q, more around 0.3 area
16:38:06ep0chPaul_The_Nerd: thanks, that seems to have fixed it
16:38:40preglowsafetydan: i would not have thought that was of any use, but sure, i can fix that
16:39:07preglowamiconn: but yeah, the button acceleration just does one step of acceleration, yes?
16:39:17amiconnNo, multiple
16:39:24ep0chi think i got some really unbearable resonance when i used a really high Q, and i think it broke the eq until i rebooted
16:40:51ep0chpreglow: ipod question, can ipods charge from usb?
16:41:00josh_ep0ch: yes
16:41:02preglowep0ch: sure
16:41:02slimxno
16:41:07Paul_The_Nerdep0ch: Hold Menu while plugging in the USB cable
16:41:10slimx:p
16:41:19preglowep0ch: high q == high resonance
16:41:23preglowep0ch: so small wonder
16:41:35ep0chpreglow: but it broke the eq too
16:41:47preglowep0ch: it shouldn't 6.4 really isn't much
16:41:48slimxdoesn't it charge on firewire only preglow ?
16:41:56preglowslimx: newer ipods doesn't have firewire
16:42:01preglowslimx: i don't know about older ones
16:42:06ep0chpreglow: i.e. changing it back to a reasonable level kept the resonance going
16:42:14preglowep0ch: is this with a new patch?
16:42:16Paul_The_NerdI know the newer ones don't charge on 1.1, I've heard. Only 2.0
16:42:16 Quit San_Benedetto ("soon ;x")
16:42:25ep0chit was quite an old one, i'll try the new one in a few mins
16:43:08preglowep0ch: i've used q == infinity, and it still doesn't crash the eq
16:43:12preglowit just starts making a sine wave
16:43:44ep0chwierd, i used it at a high frequency... let me retest...
16:43:49t0maspreglow: is your eq patch uploaded somewhere already?
16:43:53safetydanpreglow: if you could, could you write up the limits for parameters since I'm just guessing and reading comments in the eq.c file
16:44:00preglowsafetydan: sure
16:44:23preglowt0mas: https://sourceforge.net/tracker/?func=detail&atid=439120&aid=1423974&group_id=44306
16:44:27t0mask
16:44:57 Join ModernExecutive [0] (n=bob@pool-141-152-166-148.norf.east.verizon.net)
16:45:16 Quit ModernExecutive (Client Quit)
16:46:36 Join ste__ [0] (n=steve@slawson.plus.com)
16:46:38preglowep0ch: tested at all freqs, 24db gain, 6.4q, works just like it should
16:47:48preglowi think you used a buggy patch
16:48:04ep0chyeah i think so, working good now :)
16:48:32 Quit __peer__ (Read error: 110 (Connection timed out))
16:49:10ep0chlostlogic: you here?
16:51:28preglowi'm trying to figure out how to deal with the pregain
16:51:57preglowa gain can be integrated into one of the eq filters at no additional cost
16:52:12preglowbut it'd be kind of hard to implement, since any or none of the filters might be active at any time
16:52:42 Join __peer__ [0] (n=hrm@bzq-82-81-118-121.red.bezeqint.net)
16:55:09Paul_The_Nerdpreglow: Why not just put a 6th slider in there, for simply "pregain"
16:56:28preglowPaul_The_Nerd: i was more thinking how it should work internally, for max efficiency
16:56:43preglowdoing a separate gaining pass on the audio is a waste when one of the filters can do it
16:57:10preglowsafetydan: which parameters are you wondering about?
16:57:46Paul_The_NerdAaah
17:00
17:00:17preglowsafetydan: i think this merits a commit now
17:01:50safetydanThe maximum for Q
17:01:57safetydanwhether it's sensible for cutoff to go as high as 22050
17:02:02ep0chiceing on the cake would be a button combo to accelerate step size
17:02:03safetydanthings like that
17:02:46tucozok, first step towards a help on the documentation effort. http://www.rockbox.org/twiki/bin/view/Main/ManualHowto
17:04:48lostlogicep0ch: yes
17:04:49lostlogicsorta
17:05:54Paul_The_Nerdtucoz: You might want to mention that TeTeX is available as a Cygwin package. If you get TeTex and TeTeX-extras, you have everything you need under windows, without needing Miktex
17:06:31preglowsafetydan: whether that is sensible is a matter of psychoacoustics, my code allows all frequncies up to samplerate/2
17:06:43preglowsafetydan: q can go as loud as a 16 bit number allows
17:06:53preglowso something around 65536.9999
17:07:01ep0chlostlogic: i was trying out your pcmbuf_rework.patch and had a few comments, but i've just noticed there is now a version 3 so i'll just test that one to see if i get the same issues...
17:07:02tucozPaul_The_Nerd: ok
17:07:23preglow65535, that would be
17:07:33preglowbut no, values much higher than 20 or so isn't needed in an eq
17:07:41***Saving seen data "./dancer.seen"
17:07:41preglowand you might actually be able to go lower than 0.5 now
17:07:42preglowhmm
17:07:46lostlogicep0ch: ok, let me know −− the only issue I'm aware of is with voice menus when playback is stopped
17:08:08preglowbtw, i just did a test now, and the eq works just like it should, gain is correct and frequency is correct
17:09:16ep0chlostlogic: i've had playback glitches, pitch control < 100% stopped audio, and for low bitrate vorbis files i'm convinced it boosts more than cvs
17:09:29tucozPaul_The_Nerd: I'll just remove the MiKTeX link then
17:09:31 Quit ste__ (Read error: 110 (Connection timed out))
17:09:52Paul_The_Nerdtucoz: Might be the simplest idea. :)
17:09:58lostlogicep0ch: the playback glitches should be good... what do you mean "<100% stopped audio"?
17:10:12 Join ste__ [0] (n=steve@slawson.plus.com)
17:10:33ep0chlostlogic: you know the pitch control? if you slow the speed down audio stops, however speeding up is fine
17:10:41 Part tucoz
17:11:19ep0chlostlogic: but let me check with the new patch
17:13:04 Quit Zagor ("Client exiting")
17:13:38lostlogicep0ch: kk
17:14:35ep0chlostlogic: yeah latest patch breaks pitch
17:15:53 Part ste__
17:15:59lostlogichmm, that's weird, I've never used the pitch feature, will look at it sometime today
17:16:41ep0chyeah i dont use it, i just wanted to see how i could break your pcm buffer :D
17:17:33preglowhmm
17:17:38preglowpitch feature doesn't do anything special
17:17:41preglowit just enabled the resampler
17:18:52ep0chhmmm so maybe files < 44.1 khz are broken too
17:19:04preglowamiconn: that's some hardcore assembler you've got there
17:19:05lostlogicif the dsp_input_size or dsp_output size results change or get wrong it could cause problems.
17:19:19preglowlostlogic: oh, they'll change, yes
17:19:43amiconnNo more hardcore than memcpy(). In fact it's highly symmetric to that
17:19:50 Join PaulJ [0] (n=PaulJ@vpn-3138.gwdg.de)
17:20:29lostlogicpreglow: Well this explains where the problem is, and it is a problem that the old buffer tolerated wrongly.
17:20:48preglowwhat would it be?
17:21:11preglowamiconn: memcpy is hardcore too
17:21:20preglowso yeah
17:21:25lostlogicpreglow: that the result of dsp_input_size(dsp_output_size(size)) is not size most likely
17:21:33lostlogic(but that's just a guess at this point)
17:21:52preglowit might differ by one sample
17:21:54preglowi guess
17:22:52 Quit einhirn ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
17:26:39amiconnpreglow: Different points of view :) I don't think this kind of asm is hard, it's just lengthy. Then I don't really understand how jpegs are decoded...
17:28:09preglowthe lengthy part is makes it hardcore, hehe
17:28:24preglowasm gets kind of hard when it gets long
17:28:30amiconnWell, a lot of different code paths depending on alignment
17:28:31preglowunless you've thought things well out
17:29:00amiconnEach of memcpy() and memmove() for coldfire is almost 1KB
17:29:39safetydanpreglow: I'll work on some additional tweaks tonight (disabling Q setting for shelf filters, better distinguishing between shelf and peak, step acceleration)
17:29:51lostlogicep0ch: is this pitch thing in CVS??
17:29:55safetydanmention anything else on IRC you wan't fixed and I'll look for it in the logs
17:29:58safetydanfor now
17:30:11 Part safetydan ("home time")
17:30:17ep0chlostlogic: oh yeah, on ihp-1x0, you play a song, in wps hold down play and use up/down
17:30:27ep0chlostlogic: not sure about h3x0
17:30:47preglowit should work for h3x0
17:31:00preglowamiconn: seems to have gained a couple of percent boost for musepack
17:31:16preglownot bad for a single function called only once per frame
17:32:14lostlogicep0ch: wow, it does soemthing weird fo-sho
17:32:16 Quit ripnetuk ("Leaving")
17:32:36ep0chlostlogic: but its fine with higher pitch
17:33:03lostlogicep0ch: yep. I'll look into it.
17:33:36lostlogicwhat's weird is that the codec buffer continues to be consumed even when nothing is being played back on the lower pitches... :-\
17:34:11 Join ghode|afk [0] (i=testing@host-84-9-105-221.bulldogdsl.com)
17:34:23preglowamiconn: yeah, memmove shaved off 2-3% for musepack at 190 kbps
17:35:22 Quit tvelocity ("Αναχωρεί")
17:37:32 Quit muesli__ (Read error: 110 (Connection timed out))
17:47:55 Join ]RowaN[ [0] (n=5690c514@labb.contactor.se)
17:49:00]RowaN[can someone remind me what format bmps have to be in for the WPS? i cant seem to get any to display anymore
17:49:22 Join TCK [0] (n=tckocr@81-178-216-103.dsl.pipex.com)
17:49:36]RowaN[ahah, 8bit
17:50:01Paul_The_Nerd1, 8, or 24
17:50:11Paul_The_NerdI believe
17:51:25 Join chilin_dude [0] (n=chatzill@spc1-leed6-3-0-cust194.seac.broadband.ntl.com)
17:52:14chilin_dudehey
17:54:16Paul_The_NerdHello
17:55:18chilin_dudehi
17:57:01]RowaN[guys i know greyscale bmps doesnt have transparency, but i thought i made a little greyscale icon and placed it somewhere that would be ok.. but its making all the text disappear.. is that normal for now?
17:58:18Paul_The_Nerd]Rowan[: Yes.
17:58:27Paul_The_NerdWait, ALL the text?
17:58:34Paul_The_NerdOr just, text that crosses over it?
17:59:18]RowaN[all the text
17:59:22]RowaN[apart from when the text scrolls
17:59:34]RowaN[my wps is just like this:
17:59:35]RowaN[%x|b|rowan2.bmp|106|56|
17:59:35]RowaN[%x|a|rowan.bmp|0|0|
18:00
18:00:02Paul_The_Nerd]RowaN[: Is rowan.bmp fullscreen?
18:00:08]RowaN[the one positioned at 106,56 is the greyscale little image, and that causes all the text to vanish
18:00:22]RowaN[rowan.bmp is not greyscale.. this image covers the entire screen
18:00:35chilin_dudehows progress with gigabeat going?
18:00:40Paul_The_NerdIt's saved at a Monochrome bmp (1 bit per pixel) or is it 8/24, but only using B&W?
18:01:15]RowaN[not sure.. both images are displaying, its just the text thats vanished
18:01:39Paul_The_NerdYes, but the problem is that if rowan.bmp isn't specifically 1 bit per pixel, the white from it is drawing over the text
18:01:41Paul_The_NerdIt's the real problem
18:02:05]RowaN[hmm photoshop only lets me save as 4 or 8 + bpp
18:02:09]RowaN[with this b&w image
18:02:14Paul_The_NerdThen use Paint
18:02:28]RowaN[dont seem to have it
18:02:38]RowaN[ahah
18:03:05Paul_The_NerdFigure it out?
18:03:13]RowaN[just found paint =]
18:03:17Paul_The_NerdGotcha
18:03:39]RowaN[yey it works =]
18:04:56Paul_The_NerdOh, just remembered something. Linuxstb, preglow: On my ipod, if I have repeat mode set to shuffle, it freezes at the end of a playlist.
18:06:06preglowit doesn't like you!
18:06:56Paul_The_NerdI'd assumed that much. I was just curious if this was a "Doesn't like anyone" situation, or a "Specific animosity toward Paul and/or iPods" situation. ;-)
18:09:16 Join ModernExecutive [0] (n=bob@pool-141-152-166-148.norf.east.verizon.net)
18:09:30 Quit ModernExecutive (Client Quit)
18:09:43 Join foo_ [0] (n=mafj2@tor/session/x-ccb6c11ebb97ac2b)
18:12:19preglowi have no idea, i've never used shuffle with rockbox
18:12:26preglowi'll see if i can check it out
18:12:30Paul_The_NerdRegular shuffle works fine.
18:12:42Paul_The_NerdIt's just when it hits the end of the playlist, with the shuffle repeat mode.
18:12:46lostlogicep0ch: pitch down definitely works with CVS?
18:12:52Paul_The_NerdAnd, if it's regular "repeat all" also, no problems.
18:14:23 Join NightCat [0] (n=pink__ki@garant.metacom.ru)
18:15:03ep0chlostlogic: defintely
18:15:41 Quit NightCat (Remote closed the connection)
18:15:45 Quit preglow ("leaving")
18:15:45 Join muesli__ [0] (n=muesli_t@88.134.20.76)
18:16:17 Join preglow [0] (n=thomjoha@hekta.edt.aft.hist.no)
18:16:59lostlogicI officially see no way in which my changes would impact pitch.
18:17:35Papricamm
18:17:44ep0chodd
18:17:49Papricatesters for the new scroll margins patch?
18:19:01lostlogicPaul_The_Nerd: I've made improvements to voice −− it no longer freezes, but it does stop playing menus after the 3.5thish one
18:19:34Paul_The_Nerdlostlogic: So, progress. :) I personally don't use voice, especially since Audrey won't be back, but it's a nice feature.
18:20:14lostlogicPaul_The_Nerd: hehe, yeah, and blind users _need it_ so I obviously can't commit with it partially broken
18:20:22 Quit ]RowaN[ ("CGI:IRC (EOF)")
18:20:28 Join Febs [0] (n=40be24f0@labb.contactor.se)
18:23:09 Quit Slasheri (Read error: 145 (Connection timed out))
18:23:29Papricahttps://sourceforge.net/tracker/?func=detail&atid=439120&aid=1417090&group_id=44306
18:23:59 Join Slasheri [0] (i=miipekk@ihme.org)
18:25:28 Quit TCK (Read error: 110 (Connection timed out))
18:25:53FebsPreglow, just tried the EQ. Great work!
18:26:58ep0chanyone know how to resample in audacity?
18:27:15preglowFebs: thanks
18:27:46 Quit Febs ("CGI:IRC (EOF)")
18:28:27 Join Febs [0] (n=40be24f0@labb.contactor.se)
18:28:37FebsDamn, I hate the web-based IRC client.
18:28:46preglowas you should
18:29:04chilin_dudefirefox :)
18:29:09FebsAnyway, on the EQ, frequency is currently adjusted in 10 Hz steps.
18:29:26FebsAny plans to make that a logarithmic adjustment?
18:29:55Febs10 Hz is a significant change at 50 Hz. Less so at 10kHz!
18:32:20preglowFebs: yeah, i know, i'd more like a logarithmic acceleration, though
18:32:29preglowFebs: you might still want to finetune at higher frequencies
18:32:51amiconnWhat about basic the step size on the number of decimals?
18:33:12amiconn(that's roughly logarithmic)
18:33:28 Join DangerousDan [0] (n=Miranda@newtpulsifer.campus.luth.se)
18:34:23lostlogicpreglow: Seems like dsp_input_size is rounding toward zero excessively maybe? When the pcmbuffer needs wrapping (has 1 or 2 samples of space left), those last 1-2 samples don't get written when pitching down, because dsp_input_size says it can't fit anything in them.
18:34:54preglowoh
18:34:57lostlogicep0ch: interestingly, this also explains worse performance on _low_ bitrate files
18:34:59preglowwell, it does round down
18:35:13lostlogicI'll work around it, because it will make performance on low bitrate better any way.
18:36:05ep0chinteresting, what's the issue?
18:36:30lostlogicep0ch: too small of chunks are ending up getting fed to the DMA
18:36:46ep0chahhhh
18:36:47lostlogicat least I think that's it
18:36:52lostlogicwe'll find out soon
18:37:29*amiconn would like a better resampler :/
18:37:45preglowyeah, me too, but it's not a priority
18:38:21 Quit chilin_dude ("Chatzilla 0.9.70 [Firefox 1.5.0.1/2006011112]")
18:38:48 Join imphasing_ [0] (n=a766a239@yossman.net)
18:38:53amiconnHmm. For me the resampler has higher priority than EQ
18:38:59 Join damaki__ [0] (n=Chocolat@ALille-153-1-4-214.w83-198.abo.wanadoo.fr)
18:39:23amiconnWith a bad resampler, non-44.1kHz tracks sound horrible, but EQ is optional stuff...
18:40:13preglow48khz tracks sound completely adequate
18:40:18preglowand those are more or less what i did it for
18:40:21preglowit isn't meant for 22khz files
18:40:30amiconnThe voices are 12 kHz...
18:40:40preglowyes, and it wasn't designed for them at all
18:40:55muesli__preglow theres a load option for eq settings..but hwo to save settings?
18:41:05preglowmuesli__: with the save settings item?
18:42:25muesli__didnt see any
18:42:29preglowit's there
18:42:53lostlogicpitch down works
18:43:31muesli__sorry, i seem to be dumb :o
18:43:35ep0chcool
18:44:57 Join einhirn [0] (n=Miranda@p548606D4.dip0.t-ipconnect.de)
18:49:16preglowmaking a new resampler wont be easy either
18:49:18ep0chhehe i think resampler should go before eq in the dsp chain
18:49:29preglowep0ch: that depends on whether it samples up or down
18:50:30ep0chi get lots of noise on a 22khz file when eq is just over 11khz
18:50:41preglowyes, that'll be fixed
18:50:46ep0choh cool
18:51:27preglowit has to be fast, reasonably accurate, offer seamless resampling ratio changes
18:53:16 Quit damaki_ (Read error: 110 (Connection timed out))
18:57:36slimxlet's leave my office see ya all at home :)
18:57:40 Quit slimx ("Leaving")
18:59:47PaulJPaprica: the scrolling margins patch doesn't work correctly for me (i used %m|11|108| ): http://www.stud.uni-goettingen.de/~s291827/test/dump%20350206-185407.gif
19:00
19:00:15Papricaauch.....
19:00:15 Quit imphasing_ ("yossman.net freenode.net/wikipedia webchat CGI:IRC 0.5.4 (2005) (EOF)")
19:00:23Papricacan you send me the wps?
19:02:37PapricaPaulJ?
19:03:08novimondoes rockbox have format a disk ability?
19:03:30novimoncause windows wont recognize my iriver anymore
19:04:09novimoneven the usb mode wont work
19:04:39PaulJPaprica: http://www.stud.uni-goettingen.de/~s291827/test/PJ_Ice_R.zip
19:04:49Paprica10
19:04:51Papricax
19:06:14PaulJPaprica: but i haven't used a clean source. Maybe this is the reason for the problems.
19:06:21Papricanono
19:06:25Papricawait a sec
19:07:17novimonwhen i start the normal firmware it says "chek the hdd connection"
19:07:23novimonand the "check the hdd"
19:07:43***Saving seen data "./dancer.seen"
19:09:38 Join ep0ch| [0] (n=ep0ch@84.12.178.113)
19:09:42Papricaok i know way it happens...
19:09:50Papricawait a few minutes..
19:11:21novimonhmm
19:11:23novimonfixed it
19:11:32novimonsnapped the player agains my hand
19:11:39novimoni think the hdd connection is loose
19:11:45ghode|afkhehe
19:12:20novimonbut i wonder if I should send it for fixing
19:12:27novimonor continue using
19:12:42novimonhow long is the warrante for ihp120?
19:13:09ghode|afkerrr not long, a year at most
19:13:12novimonok
19:13:13Paul_The_NerdIf it says iHP on it, it's probably expired. :-P. That being said, it's a year.
19:13:18novimonim past that so..
19:13:40Paul_The_NerdYou could always open 'er up, and make sure the HD is snug.
19:13:48novimoncan I do that
19:13:53novimonis that hard to do? :)
19:14:15ghode|afknot too hard, there are loads of guides out there
19:14:46novimonyou wouldnt happen to have URL for one handy, would you?
19:15:03Paul_The_NerdRemove screws, remove end bits, remove back, look, make sure snug. It's really not that terribly complicated as long as you're careful.
19:15:13ghode|afklook at the h1x0 forums on misticriver.com
19:15:26novimonok
19:15:43novimonI'll try it
19:15:59novimonI'll report back when I've messed things up :)
19:17:08 Quit Paul_The_Nerd ("Leaving.")
19:17:13amiconnHmm.
19:17:36amiconnWhere to put the memset16() declaration? It's not a standard string function...
19:17:51preglowindeed not
19:18:46PapricaPaulJ, http://sourceforge.net/tracker/index.php?func=detail&aid=1417090&group_id=44306&atid=439120
19:18:51Papricatry it please
19:19:56 Quit ep0ch (Read error: 110 (Connection timed out))
19:20:05PaulJPaprica: there is no patch attached
19:20:28ep0ch|/name ep0ch
19:20:30 Nick ep0ch| is now known as ep0ch (n=ep0ch@84.12.178.113)
19:20:31Papricaok
19:20:31ep0choops
19:20:33Papricatry now
19:21:12PaulJPaprica: ok. i'll try it
19:21:18Paprica10x
19:27:50 Join TCK [0] (n=tckocr@81-178-216-103.dsl.pipex.com)
19:29:18 Quit einhirn (Read error: 104 (Connection reset by peer))
19:31:24 Join safetydan [0] (n=dan@81-178-235-210.dsl.pipex.com)
19:34:39 Join NicoFR [0] (n=nico404@rob92-6-82-231-243-63.fbx.proxad.net)
19:34:51 Join matsl [0] (n=matsl@1-1-4-2a.mal.sth.bostream.se)
19:36:40 Join damaki_ [0] (n=Chocolat@ALille-153-1-76-147.w86-207.abo.wanadoo.fr)
19:36:54PaulJPaprica: If the statusbar is enabled all text is pushed down. http://www.stud.uni-goettingen.de/~s291827/test/dump%20350206-193319.gif
19:37:32Papricasec
19:37:37Papricacheck it.
19:38:26Papricalol
19:38:31Papricaits not happen to me
19:38:32Paprica=\
19:38:33Papricamm
19:38:33 Quit kkurbjun ("Leaving")
19:38:44Papricawhen it happen?
19:38:52PaulJi will try it with a clean cvs
19:39:08Papricaok
19:40:55 Join perl|bbq [0] (n=say@cpe-66-65-89-236.nyc.res.rr.com)
19:43:50*preglow might have eq plotting working faster than expected...
19:44:26safetydannow that will be something to see
19:45:46preglowi've got some prototype code working right now
19:46:17 Join hd [0] (i=hd@gate-hannes-tdsl.imos.net)
19:50:29preglownow, how would such a thing be integrated into, do you think?
19:50:38preglowit seems this is very realistic after all
19:50:47safetydanWell, there's not really room on the UI screen as it stands
19:50:55 Quit damaki__ (Read error: 110 (Connection timed out))
19:51:02preglowno, i'm thinking it might be bound to a button
19:51:05preglowso you can switch screens fast
19:51:37preglowi don't think we can afford to have this thing be realtime updated anyway, it's pretty cpu intensive
19:54:49safetydanSounds like the best way to do it
19:55:08 Join _FireFly_ [0] (n=FireFly@p54A444B0.dip.t-dialin.net)
19:55:44Papricahttp://misticriver.net/showthread.php?p=387888#post387888
19:55:49Papricaveryyyyyyyy nice
19:59:09 Quit goa (Connection timed out)
19:59:10 Nick hd is now known as goa (i=hd@gate-hannes-tdsl.imos.net)
20:00
20:00:04PaulJPaprica: i tried now with a clean cvs version, but there are still problems: 1.If the statusbar is enabled in the menu, the text is pushed down, even though i use the %wd tag in the wps
20:01:09 Quit damaki_ (Read error: 110 (Connection timed out))
20:01:11 Quit XavierGr ("Trillian (http://www.ceruleanstudios.com")
20:01:18Papricacan you send me the wps again?
20:02:45PaulJ2. the %m tag (without x1 and x2 values) doesnt work, the text in the lines below the %m stilll have the margins
20:03:23 Join Lear [0] (n=chatzill@h194n1c1o285.bredband.skanova.com)
20:04:01Papricaok
20:07:27PaulJPaprica: here is the wps (it is a different one, because the other one doesnt work correctly with cvs): http://www.stud.uni-goettingen.de/~s291827/test/br22.zip
20:07:33preglowlooks like phase problems is going to make this non-trivial
20:09:23 Join Thus0 [0] (n=Thus0@248.111.102-84.rev.gaoland.net)
20:10:03Papricamm
20:10:07Papricawhat is this tag?
20:10:08Paprica->>
20:10:58 Quit t0mas ("brb")
20:11:32PaulJthe > are the closing brackets of the conditional
20:14:17Papricamm dont know why but i could'nt do what you say in 1..
20:14:30preglowsafetydan: btw, since the eq is now placed after resampling, you should use NATIVE_FREQUENCY and not dsp->frequency for calculating the cutoff
20:19:35 Join slimx [0] (n=slimx@vau75-7-82-234-251-56.fbx.proxad.net)
20:20:33Learpreglow: for the Q parameter in the equalizer, does a high value make the band wide or narrow?
20:21:51preglownarrower
20:22:02preglowi
20:22:14preglowi'm pondering on doing a bandwidth parameter too
20:22:31preglowjust need to figure out how simple it is to do a sinh() in fixed point
20:22:51PaulJPaprica: Statusbar enabled looks like this for me: http://www.stud.uni-goettingen.de/~s291827/test/dump%20350206-201639.gif
20:23:03 Quit goa ("Client suicide")
20:23:10PaulJ(it happens only if the first lines scroll)
20:23:47PaulJIf the statusbar is disaled it looks like this: http://www.stud.uni-goettingen.de/~s291827/test/dump%20350206-201700.gif
20:24:58PaulJ(in both cases the text flickers)
20:25:56preglowseems eq graphing is realistic after all
20:27:17 Quit TCK ("well, if you say so.")
20:28:56preglowdoes any other daps allow you to see the freq response of the eq? :>
20:29:02preglowi'm hoping for another rockbox first, heh
20:29:08safetydanGo Rockbox!
20:29:15safetydanI'll fix that NATIVE_FREQUENCY in the next release
20:29:19preglowyep
20:29:32Bagdereveryone just mentions rio karm when eq and dap is brought up
20:29:36Bagderkarma
20:29:42preglowi'd say our eq is just as good
20:29:50preglowspecwise it should be
20:33:42Learbut is it as fast? ;)
20:36:50 Join petur [0] (i=petur@d54C1AB96.access.telenet.be)
20:43:40 Join ModernExecutive [0] (n=bob@pool-141-152-166-148.norf.east.verizon.net)
20:43:45 Quit AliasCoffee ("Leaving")
20:43:47 Quit ModernExecutive (Remote closed the connection)
20:49:38 Quit Febs ("CGI:IRC (EOF)")
20:57:30 Join ModernExecutive [0] (n=bob@pool-141-152-166-148.norf.east.verizon.net)
21:00
21:00:13amiconnBagder: Any suggestion where to put the memset16() declaration? string.h doesn't sound appropriate...
21:00:22NicoFRlinuxstb: are you there ?
21:00:39Bagderamiconn: we should probably have a memory.h
21:07:31 Join linuxstb_ [0] (n=linuxstb@i-83-67-212-170.freedom2surf.net)
21:07:46***Saving seen data "./dancer.seen"
21:07:50linuxstb_NicoFR: ?
21:07:54NicoFRhi
21:08:01 Quit linuxstb (Nick collision from services.)
21:08:03 Nick linuxstb_ is now known as linuxstb (n=linuxstb@i-83-67-212-170.freedom2surf.net)
21:08:06safetydanpreglow: should I disable Q setting for the shelf filters? and would it be worth upping the precision of Q so we can use 0.707 instead of 0.7?
21:08:11NicoFRit's about the backdrops in wps
21:08:30linuxstbWhat about them?
21:08:48NicoFRis it normal that a wps that doesn't have an %X tag uses the backrop of the previously loaded wps ?
21:09:04 Quit Thus0 ("Leaving")
21:09:17linuxstbYes, I've read about that - it's a bug I think.
21:09:25NicoFRi think so too
21:09:40NicoFRisn't you who worked on backdrops ?
21:09:59linuxstbYes.
21:10:14linuxstbPlus Linus did some work on them as well.
21:10:54NicoFRi was looking at the code so i wanted to ask you if it was worth me trying to fix it
21:11:21linuxstbIt's OK, I'm having a look now.
21:11:21ep0chsafetydan: Q setting for bass is quite interesting imho
21:11:34NicoFRok
21:12:12*petur wonders what he did wrong to the gods of cvs commits
21:12:28peturkeeping a patch up to date is a serious pita :(
21:13:03Bagderpetur: I don't think you've done anything wrong, just that no cvs committer has yet taken the time to merge your work
21:13:23NicoFRlinuxstb: and what about having another tag that would fill the BG with a solid color ?
21:14:12linuxstbYes, we need that.
21:14:18lostlogiclinuxstb: you OK with me committing my pcmbuf stuff with it still probably breaking ipod skip/stop?
21:14:38safetydanep0ch: interesting in that it does something useful or interesting in that it makes the sound go funny in interesting ways?
21:14:56lostlogicep0ch: want to torture test it again before I do that? :)
21:16:01ep0chsafetydan: interesting that it adds an extra dimension to the bass
21:16:09ep0chlostlogic: i'd love too
21:16:28ep0chlostlogic: patch in the patch tracker?
21:16:34lostlogicep0ch: not yet, about to upload
21:17:08peturwell, shameless patch advertisment follows... latest comment sums up what it does:
21:17:10peturhttps://sourceforge.net/tracker/index.php?func=detail&aid=1403437&group_id=44306&atid=439120
21:17:43linuxstblostlogic: I'll have another go at fixing it later tonight. The current ipod playback is buggy, but at least it doesn't freeze, forcing a reset.
21:18:30linuxstbI'm going to try and rewrite the pcm_playback code for ipod, but try and more closely follow the logic of the uda1380 driver. I'm now understanding things more.
21:18:53peturBagder: I'm sure I did something wrong to Linus (some pretty bad first patches as I was learning all these new things: cygwin, cvs, dos2unix,...)
21:20:02lostlogiclinuxstb: hmm... might better for me to commit first any way, because my stuff does modify the pcm_playback.c a bit, at least the initialization, and not holding on to extra data...
21:21:58lostlogicep0ch: new version posted on sf
21:22:11*ep0ch wonders if linuxstb wants to update menu_files.patch to work with cvs one day
21:22:40preglowLear: dunno, if anyone can give me perforance specs from the karma, i'd be happy to compare :-)
21:22:42ep0chlostlogic: k
21:22:47 Join nounours [0] (n=lalmeras@gs.team.openwide.fr)
21:25:20preglowat least i can't make my current eq code much faster
21:25:57 Quit perl|bbq (Read error: 110 (Connection timed out))
21:26:09 Join webguest67 [0] (n=5087cc2c@labb.contactor.se)
21:26:54webguest67preglow, does the cutting of freq work with your EQ ?
21:27:04preglowwebguest67: of course
21:27:11preglowjust use negative gains
21:27:41webguest67eargerly waiting your committ, :-)
21:28:32 Part webguest67
21:28:54safetydanpreglow, any thoughts on Q removal for shelf filters?
21:29:06 Join goa [0] (i=hd@gate-hannes-tdsl.imos.net)
21:29:11preglowsafetydan: well, i don't think we should _rmove
21:29:13preglowargh
21:29:26preglowsafetydan: well, i don't think we should _remove_ them, but then again, people shouldn't tweak it an expect good results :)
21:29:30safetydanah okay
21:29:33 Join hardeep [0] (i=hardeeps@SDF.LONESTAR.ORG)
21:29:34safetydandocumentation issue then :)
21:29:40 Quit _FireFly_ ("Leaving")
21:29:50preglowwe should really use another parameter for the shelving filters
21:29:53lostlogicgah, did something recently take up 2.5 megs of codec buffer space, or is it me somehow? Does enabling voice do that or changing language?
21:30:06preglowthe audio eq cookbook has a S parameter, where S = 1 means maximally flat frequency response
21:30:06 Quit Sinbios (Read error: 104 (Connection reset by peer))
21:30:16preglowwhich is exactly equal to q = 0.707, but a bit clearer
21:30:27preglowlostlogic: enabling voice does that, sure
21:30:29 Join Sinbios [0] (n=Sinbios@HSE-Toronto-ppp301652.sympatico.ca)
21:30:36preglowlostlogic: not only enabling it, just keeping a voice file around will allocate it
21:30:54lostlogicahhhhh, ok, that makes sense, I was about to ask "why didn't it go away when I rebooted?"
21:31:04safetydanpreglow: that sounds like a good idea actually
21:31:18safetydanI see some confusion on the forums about why there's a Q for shelving filters
21:31:28preglowsafetydan: problem is explaining to ordinary people what a higher S does
21:31:42preglowalso, i need to code it first :-)
21:31:50safetydanthere is that
21:32:04safetydanfour things left on my todo list
21:32:29preglowand those are?
21:32:29*safetydan starts digging in to the WPS code to add a "EQ enabled" setting
21:32:33preglowi'm quite eager to commit this thing
21:32:38 Quit foo_ ("Leaving")
21:32:50safetydanEQ reset option
21:32:53safetydanWPS tag
21:33:00safetydanStep acceleration
21:33:02 Join optimistik [0] (n=optimist@d216-232-100-172.bchsia.telus.net)
21:33:12safetydanGenre based EQ settings
21:33:20safetydannone of them are really critical
21:33:22preglowwouldn't the config file system cover that?
21:33:59safetydanI would have thought so, but some people want it to switch to .rockbox/eqs/<genre>.cfg based on ID3 tags
21:34:04preglowbah
21:34:13safetydanAs I said, not critical
21:34:20preglowso it's basically cool to go now?
21:34:32optimistikcan someone help me my ipod has rockbox on it and won't boot into the apple os or go into diskmode if i try i hear the hd starting and stopping a lot and then i get the sad ipod thing
21:34:38safetydanI'll upload a new patch in a second
21:34:46safetydanfixes the NATIVE_FREQUENCY thing at least
21:34:52preglowyeah, good
21:34:56preglowhmm
21:35:20preglowi'd also like a pregain ready, but that could wait
21:35:34preglowoptimistik: sounds like you need to restore it
21:36:03preglowoptimistik: what ipod model?
21:37:10safetydanoh yeah, pregain, I'll add that to the list
21:37:12optimistik4th gen 60gb photo
21:37:28preglowsafetydan: i think we'll just have to be content with the replaygain method
21:37:44preglowit'll slow down the eq yet another notch, but yeah
21:37:44 Quit optimistik ()
21:38:05 Join linuxstb_ [0] (n=linuxstb@i-83-67-212-170.freedom2surf.net)
21:38:40ep0chlostlogic: patch seems good to me, i havent tested voice though
21:38:47 Quit linuxstb (Read error: 110 (Connection timed out))
21:39:36 Join Henrico [0] (n=henrico@cn-mss-cb01-1900.dial.kabelfoon.nl)
21:39:37*ep0ch remembers when 128 kbps mp3 used to play without boost :s
21:40:00lostlogicep0ch: you sayin' it's slower still?
21:40:10ep0chno
21:40:17ep0chits on par with latest cvs
21:40:29*safetydan waits for sourceforge
21:40:53ep0chi think either my builds are slow or cvs build got slower over the past months
21:40:58lostlogicah yes, we need to figure out some way other than mutexes to protect pcmbuf_play_stop and pcmbuf_flush_fillpos from each other
21:41:03preglowep0ch: they still do...
21:41:30ep0chthen something is most definelty wrong with my setup
21:41:35safetydanpreglow: latest patch should be up with change summary attatched
21:41:35preglowh1x0?
21:41:40ep0chyeah...
21:41:43preglowwell, works fine here
21:41:53pregloweven 160kbps is just _almost_ non-boost
21:41:53safetydanattached even
21:42:07 Join perldiver [0] (n=say@cpe-66-65-89-236.nyc.res.rr.com)
21:42:08lostlogicep0ch: what compiler you usin'?
21:42:15ep0ch3.4.4
21:42:58ep0chahh i had replaygain on and crossfeed lol
21:43:22*lostlogic always plays with replaygain on
21:43:27lostlogicnot crossfeed though
21:43:35ep0chhmm still boosts slightly
21:43:46safetydanpreglow, one thing that might be important, not all strings are localised
21:43:49safetydanI got lazy
21:45:16preglowoh, well that needs fixing
21:46:23 Quit Sinbios ("If the definition of a klutz is someone who doesn't have eyes on their ass, then yes, I suppose I am a klutz.")
21:46:28ep0chwhat compiler does rockbox.org use for builds?
21:46:39lostlogicep0ch: 3.4.4 or 3.4.5 iirc
21:47:01 Join Mmmm [0] (n=mscarrat@cpc1-hem13-0-0-cust973.lutn.cable.ntl.com)
21:47:26Bagderm68k-elf-gcc (GCC) 3.4.1
21:47:29peturhey Mmmm
21:47:44ep0chty
21:47:44Bagdereach build log shows the version
21:47:45safetydanpreglow, okay, I'll go through and do the i18n thing
21:47:47MmmmSo...Petur...anyone committed our patch yet? :D
21:47:59peturnope... :(
21:48:10preglowwhat patch? recording?
21:48:24MmmmIs there something weong with it? does it smell or something?
21:48:31ep0chhehe
21:48:32preglowahh, _that_ one
21:48:33peturI've been keeping it up to date with cvs and fixed a warning with Archos...
21:48:38preglowyeah, it smells like a bag of shit
21:48:39 Quit solexx ("leaving")
21:48:40preglowso i wont touch it
21:48:45BagderMmmm: yes, that is the _only_ patch that isn't committed... :-)
21:48:52Bagderhahaha
21:48:54Mmmmbut its lovely!
21:49:20preglowyeah, i kn0ow
21:49:22preglowi should try it one day
21:49:29Mmmm:D
21:49:30peturI guess most devs with commit access aren't into recording...
21:49:45Mmmmcertainly seems that way eh? :(
21:50:15 Quit mafj2 (Connection reset by peer)
21:50:15preglowsafetydan: hmm, i just remembered something
21:50:32preglowsafetydan: when a band is enabled after being disabled for a while, the filter history for that filter should be zeroed
21:50:34 Join webguest16 [0] (n=5087cc2c@labb.contactor.se)
21:50:56preglowthe filter history for all filters should be zeroed whenever audio stops
21:51:16preglowor you'll get a little resonant pop whenever you press play again
21:51:17peturMmmm: anyway, my last changes are important for Archos: one mistake and one cosmetic issue (ran it in the sim to test)
21:52:04Mmmmreally? for some reason I though recording didnt work in the sim! I've been plugging and unplugging!
21:52:24peturI also saw that abs() is redefined in a zillion places...
21:52:41peturMmmm: just the UI...
21:52:57Mmmmabs()...hmmm..was that me or you ;)
21:53:03webguest16using the EQ gui interface, the changes made are only audible after 5 plus seconds, and then only if another mode or band is selected, anyway of making the EQ changes audible in realtime
21:53:30peturMmmm: no, I mean in other files we didn't touch ;)
21:54:10Mmmmwell of course..as if we would go and do something silly like that eh? :D
21:54:56lostlogiclinuxstb_: let me know if you figure anything out about the ipod playback brokenness −− I'm itchin' to commit this stuff, been workin' on / sittin' on it for too long.
21:55:03Learwebguest16: crossfade user perhaps? :)
21:55:14webguest16not at LL
21:55:24webguest16LL = all
21:55:39 Quit linuxstb_ (Read error: 110 (Connection timed out))
21:56:01preglowsafetydan: even seems to work on ipod
21:56:03LearWell, the non-crossfade audio buffer isn't that long, so the change should happend a bit quicker, I would've thought..
21:56:17 Join muesli- [0] (n=muesli_t@88.134.20.76)
21:56:20 Join linuxstb_ [0] (n=linuxstb@i-83-67-212-170.freedom2surf.net)
21:56:28safetydanI'll fix the history filter thing as well while I'm i18n'ing
21:57:03webguest16its no biggie, just thought it might be a bug :)
21:57:03ghode|afkeq changes take 20 seconds for me :/
21:57:22safetydanpreglow: could I move the eq_*_coef calls inside the while loop in dsp_process? would that make the EQ more responsive?
21:57:43preglowno, not really
21:57:52preglowthe problem is the length of the pcm buffer
21:57:57preglowit's several seconds long
21:58:15 Quit DangerousDan (Read error: 104 (Connection reset by peer))
21:58:18 Join DrMoos [0] (i=DrMoos@m53.net81-66-159.noos.fr)
21:58:26ep0chlostlogic: and for your next trick please add a playback destination option i.e. out or null or disk with and without dsp :)
21:58:29 Quit Moos (Read error: 104 (Connection reset by peer))
21:59:57 Quit muesli__ (Read error: 110 (Connection timed out))
22:00
22:00:09preglowsafetydan: please test and see that the eq sounds good when switching a filter on and off after fixing the history thing
22:00:26preglowit _might_ be better to always let the filters have some history, even though it's technically speaking invalid
22:00:38 Quit imphasing (Remote closed the connection)
22:00:43 Join DangerousDan [0] (n=Miranda@newtpulsifer.campus.luth.se)
22:01:25 Part webguest16
22:01:29Learpreglow: without crossfade, the pcm buffer isn't that long...
22:01:40preglowLear: it's still several seconds long
22:01:49preglowLear: more than long enough for a really noticable delay
22:01:50 Join imphasing [0] (n=imphasin@c-69-250-93-218.hsd1.dc.comcast.net)
22:01:51Learsafetydan: you might want to call it from time to time in eq_menu.
22:01:52 Join Thus0 [0] (n=Thus0@84.102.111.248)
22:01:52 Join Thus0_ [0] (n=Thus0@84.102.111.248)
22:01:58ep0chpreglow: which gcc are you using to get realtime playback of 128kbps mp3?
22:02:01safetydanLear, good point
22:02:01lostlogicep0ch: it's on my brain −− haven't figured out exactly how to implement it without costing performance is the main problem at this point, probably have to make a "pluggable" interface to the pcmbuffer.
22:02:03preglowep0ch: 3.4.5
22:02:12preglowLear: why would he want that?
22:02:12 Quit Thus0 (Read error: 104 (Connection reset by peer))
22:02:12 Quit Thus0_ (Read error: 104 (Connection reset by peer))
22:02:21 Join Thus0 [0] (n=Thus0@248.111.102-84.rev.gaoland.net)
22:02:43preglowif we're still talking about calling eq_*_coefs
22:03:06Learwell, I was thinking of eq_update_data... :)
22:03:14safetydanwait... I was thinking that was related to the update lag
22:03:25Learah, more or less the same.
22:03:27safetydaneq_update_data gets called at the beginning of each PCM block
22:03:42Learwhich is more often than needed, imho...
22:03:53preglowoh yes
22:04:01preglowif all it does is update coefs
22:04:21Learyep.
22:04:27preglowcoefs should be updated when the slider/value is changed, everything else would mean a noticable performance impact
22:04:36Learjust call it on load, and whenever the settings change.
22:05:09 Join bazz [0] (n=nick@fw.marklogic.com)
22:05:15safetydanah so many things to fix
22:05:23 Quit Thus0 (Client Quit)
22:05:24*safetydan should have just preglow commit it early and stayed quiet
22:05:30safetydanlet preglow even
22:05:43bazzis there a way to do a scrolling line that doesn't scroll the width of the whole screen in a wps (so like, display the folder name, scrolling, on only half the screen)?
22:05:44preglowhahah
22:05:53Learouch, like six div64, bunch of fracmuls...
22:06:19peturbazz: I think Paprica is working on that
22:06:31bazzpetur: okay, but not yet then right?
22:06:47NicoFRbazz:http://sourceforge.net/tracker/index.php?func=detail&aid=1417090&group_id=44306&atid=439120
22:06:47peturthere's probably a patch already
22:06:55Papricabazz, there is a patch on the SF but it little buggy
22:07:09bazzthanks guys :)
22:07:49bazzPaprica: buggy how?
22:07:51preglowsafetydan: nothing that doesn't impact the way it sounds is critical
22:07:55preglowsafetydan: everything else can be done later
22:08:05Papricait have some problems..
22:08:35preglowsafetydan: hey, think of the fame and glory
22:08:54preglowsafetydan: but yeah, it even works on ipod, grand work
22:08:56bazzPaprica: okay, well, i'll play with it a bit thanks :)
22:09:03Paprica=]
22:09:38preglowsafetydan: btw, you've got a lacking indent around the if (dsp->eq_enabled) is dsp_process() :>
22:09:47preglow_in_ dsp_process
22:10:02 Quit imphasing (Remote closed the connection)
22:11:23safetydanodd, looks fine on my screen
22:11:32preglowsafetydan: tab?
22:11:37safetydanspaces
22:11:41preglowthe eq_process call isn't indented here
22:12:01safetydanI'll check the next patch
22:12:09preglowwtf
22:12:16preglowthe semicolon is lacking too here :>
22:12:21preglowsomething's on
22:12:27safetydanyeah okay that's just getting weird
22:12:45preglowjust ignore me
22:12:49safetydanDo the iPod key settings make sense? I sort of just made them up
22:12:49preglowi'm sure all will be well
22:13:17Learsafetydan: move that to dsp_set_eq (and clear eq_state too), and add some to calls to eq_menu while you're at it. :)
22:13:22LearI'm just about to try that here...
22:13:47preglowsafetydan: we'll see, for now, they work
22:15:43safetydanI've just redone the filter update stuff. Now it only updates on change from the user not in dsp_process
22:16:08preglowexcellent
22:17:36 Quit DangerousDan (Read error: 104 (Connection reset by peer))
22:19:32*safetydan notices the oddity of having DangerousDan and safetydan on the same channel
22:19:36preglowhaha
22:19:45 Join DangerousDan [0] (n=Miranda@newtpulsifer.campus.luth.se)
22:20:31safetydanStep back DangerousDan! safetydan is here!
22:20:47muesli-schizzo? :d
22:20:54 Quit ghode|afk (Read error: 104 (Connection reset by peer))
22:21:09 Join Thus0 [0] (n=Thus0@154.111.102-84.rev.gaoland.net)
22:21:25safetydangah, now I've broken the compile
22:22:21preglowcause for celebration!
22:23:00lostlogicwait wait... DangerousDan comes in... then safetydan breaks things!?
22:23:37preglowyou're in #rockbox now
22:23:50preglowreality itself has to withstand some bending
22:26:11 Join imphasing [0] (n=imphasin@c-69-250-93-218.hsd1.dc.comcast.net)
22:27:40 Quit imphasing (Client Quit)
22:30:49safetydanWhat's the best place to reset the filter history? DSP_RESET?
22:31:26amiconngah
22:31:43amiconnBackdrop seriously complicates optimisation
22:32:00preglowsafetydan: i don't know, it'd need to be in a place that gets called whenever playback stops
22:32:11preglowsafetydan: as in it should not happen at an ordinary track transition
22:32:23preglowsafetydan: so it'll probably be a bit tricky :>
22:32:43preglowamiconn: well, no shit
22:33:03preglowamiconn: but the gui code is built so that you can optimise everything else without having to optimise backdrop code as well, yes?
22:33:23lostlogicep0ch: for your wav out needs, would the raw PCM data work, or would that be a major pain?
22:34:10preglowraw pcm data = signed 32 bit data of arbitrary precision :>
22:35:49amiconnpreglow: Backdrop is low level stuff, the gui layer has nothing to do with it
22:35:57lostlogicok, so how does one output wav?
22:36:00preglowahh, yeah, i meant lcd drawing code
22:37:26amiconnAs long as the primitives only use the fastpixelfuncs it's no problem, but that's the area I want to optimise
22:37:41 Join JdGordon [0] (n=Miranda@c220-239-137-122.smelb1.vic.optusnet.com.au)
22:37:42 Quit mikearthur (Read error: 104 (Connection reset by peer))
22:37:58amiconnCurrently all primitives do one function call *per pixel*, which is quite some overhead
22:38:05lostlogicSlasheri: any chance that your audio system early initialization changes make voice not work right on bootup?
22:39:13amiconnPrimitives that fill adjacent pixels (fillrect, hline) can use memset16() instead, but have to use memcpy() for background if backdrop is in effect
22:39:18 Join w0rds [0] (n=w0rds@sp-200-155-78-175.datacenter1.com.br)
22:40:27safetydanwould it seem reasonable to have stop_codec_flush call something like "dsp_reset"?
22:40:32amiconnlostlogic: Voice for swcodec is buggy as-is in cvs
22:40:35safetydansorry, eq_reset
22:40:41preglowone function call per pixel isn't _that_ much
22:40:47preglowthey should be able to use just scratch regs
22:41:12preglowsafetydan: ask lostlogic :>
22:41:21preglowi have no clue about that part of rockbox
22:41:28amiconnpreglow: Parameter passing happens on the stack, plus rts is a slow operation
22:41:31safetydanI'm getting lost in there too
22:41:33 Quit Matze ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
22:41:38safetydanI think I'll leave it on the todo list for the moment
22:41:42perldiveris there any development of rockbox for toshiba's gigabeat?
22:41:44lostlogicsafetydan: I want to get rid of that function... what are you trying to do?
22:42:10safetydanreset eq filter history every time playback stops
22:42:46 Nick DrMoos is now known as Moos (i=DrMoos@m53.net81-66-159.noos.fr)
22:43:10preglowlostlogic: as in only them, not on continous track changes, etc
22:43:14lostlogicsafetydan: define every time the playback stops −− should it reset on skips due to buffer underflow, on track changes, both, neither?
22:43:15preglows/them/then/
22:43:18amiconn...and on coldfire, filling adjacent pixels with memset16()/memcpy() will give significant speedups, since these functions use burst mode if possible
22:43:31preglowlostlogic: only when there's a significant gap in the playbacj
22:44:03preglowat least the normal conditions, on buffer underflow and stuff like that, we'll survive not handling it optimally
22:44:04lostlogicI assume a pause would not require a skip? This is a system timer based operation?
22:44:17lostlogics/skip/reset/
22:44:31 Join BHSPitLappy [0] (i=Steve-O@67.64.118.242)
22:44:33preglowhmm, lemme think here
22:44:58 Join imphasing [0] (n=imphasin@c-69-250-93-218.hsd1.dc.comcast.net)
22:45:09preglownow that i think about it, the only case that needs to zero the filter history, is the case when the output is interrupted by new audio
22:45:26preglowlike skipping tracks
22:45:30imphasinghrm
22:45:42imphasinginteresting..
22:45:54safetydanwhat about crossfade?
22:47:00preglownope
22:47:04preglowonly rough transitions
22:47:12lostlogicthen it should be reset on play initialization and in codec_request_next_track if new_track and not crossfade, I think.
22:47:38 Quit zhilik ("http://www.zhukovsky.net")
22:47:46amiconnIiuc this may break gapless playback...
22:48:03lostlogicit won't impact gapless
22:48:06 Quit perpleXa (Remote closed the connection)
22:48:11preglowit shouldn't
22:48:19lostlogicbecause preglow says only on rough (manual) track changes
22:48:33amiconnah ok
22:48:35preglowproblem is, if there is a sudden change in output amplitude, the filter will make a resonant pop
22:48:48 Join solexx [0] (n=jrschulz@d020063.adsl.hansenet.de)
22:48:53preglowif audio is gapless, the audio should not change suddenly, and so the playback engine should not reset filter history
22:48:58lostlogicwhich could happen at track change even if it is an automatic change
22:49:21preglowindeed, but trigging on manual changes is the best we can do
22:49:27lostlogicyep
22:49:38amiconnI don't think this pop is a problem
22:49:46preglowit's annoying
22:49:49preglowand very noticable
22:49:51amiconn(on automatic track changes that don't 'match')
22:50:20preglowbut yeah
22:50:23preglowit's far from critical
22:50:28preglowso might as well let it wait
22:50:38amiconnI'd think the UDA treble& bass filters do the same
22:50:47preglowthey're not resonant
22:50:48safetydanI've borken something serious somewhere, EQ just does white noise now
22:51:02BHSPitLappywhat pop are we talking about?
22:51:03preglowsafetydan: what've you changed?
22:51:15BHSPitLappythe volume jump after a fadeout?
22:51:24safetydanmoved eq_update_data calls out of dsp_process
22:51:43preglowBHSPitLappy: no, just the sudden jump you get from sample a to sample b during a track transition that's not continuous
22:51:56BHSPitLappyhmm
22:52:17BHSPitLappywell, why does my problem happen :P
22:52:32Learsafetydan: me too, but I changed a few more things...
22:52:50amiconnThe volume jump after fadeout happens because the fadeout doesn't check whether pcm playback finally finished
22:53:08preglowi think software codec platforms should use a software based fadout
22:53:15preglowthe hardware based one clicks like mad
22:53:33amiconnWe had the same problem on archos back then for low bitrate tracks, and I added a wait-for-MAS-running-out-of-data check
22:53:49preglowyeah, i remember
22:55:41 Quit w0rds ()
22:55:44amiconnSimilar effect - the playback engine stops delivering data, but the MAS keeps playing until it runs out of buffered data
22:55:56amiconn...which takes longer with lower bitrates
22:56:33safetydanI'm going to have to leave this for the moment
22:56:47safetydanseems something weird happens when you update the eq coef's outside of dsp_process
22:56:49 Join LinusN [0] (n=linus@labb.contactor.se)
22:56:55safetydanlike a burst of white noise, then it settles down
22:57:35preglowoh?
22:57:40safetydanI can post a patch with what I have (mostly i18n fixes and moving eq_update_data calls), but it's not usable
22:57:46preglowhow fast does it?
22:57:51preglowsettle down, that is
22:57:59safetydana second, maybe two
22:58:03preglowhmm
22:58:10preglowand it settles down to ordinary eqed audio?
22:58:12preglowor silence or what?
22:58:19safetydanordinary audio
22:58:26safetydanactually, should I reset the history on coef change?
22:58:30preglowno
22:58:32preglowby no means
22:58:37preglowbut this i strange
22:58:48preglowthe coefs shouldn't be updated differently now than before
22:59:08preglowafter all, if you did no control changes before, the coefs would always be the same anyway
22:59:16Learpreglow: I made similar changes, but it doesn't settle down for me...
22:59:19safetydanIt's roughly once every HZ/10 for changes from the UI
22:59:23preglowso nothing much should change...
22:59:48pregloware you completely sure the macsr is set correctly?
22:59:52safetydanoh
22:59:53safetydanwait
22:59:54safetydanno
22:59:55DBUGEnqueued KICK safetydan
22:59:55safetydandarn
22:59:55preglow:-)
22:59:56safetydanhang on
22:59:57preglowi knew it
23:00
23:00:13preglowyou need to be careful with that one
23:00:18preglowmacsr is saved per thread
23:00:21amiconnmacsr is saved in the thread context...
23:00:29amiconn:)
23:00:48preglowthere should be aboslutely no changes in sound with this change
23:00:53safetydanno
23:00:53preglowand a ton of cpu savings
23:01:09safetydanI forgot I'd changed that in eq.c
23:01:18preglowoh?
23:01:20safetydanwell it was safe when I was calling from dsp_process only
23:01:40preglowyeah, it was
23:02:21safetydanIs it that much faster that I should be setting up macsr in eq_update_data()?
23:03:24preglowsafetydan: btw, eq_update_data should, for max efficiency, only update the bands that have changed, but it's not _that_ important
23:03:33preglowsafetydan: and yeah, eq_update_data would be a nice place to do the macsr set
23:03:34Learsafetydan: if you only call it when settings actually change, it won't matter.
23:04:00safetydanokay, last fix before sleeping
23:04:45preglowthink i might do a little go at the eq graphing before i call it a night
23:05:05preglowno one would be interested in the phase response of the eq, no?
23:06:02lostlogicthinking of ways to provide pluggable output. calling functions via function pointer instead of directly costs 1 instruction per call, right?
23:07:31amiconnDepends how often the function is called
23:07:50***Saving seen data "./dancer.seen"
23:08:04amiconnIf it's called more than once from within a loop, calling via pointer may be equally fast
23:08:12lostlogicamiconn: yah
23:10:24 Quit Mmmm ()
23:10:31safetydanlast patch for tonight is posted, only lightly tested though
23:10:37safetydanlucky v13
23:10:55safetydanany problems post them to the patch tracker or here and I'll take a look tomorrow
23:11:21lostlogiconce I get my pcmbuf refactoring in, could make the destination of the pcm data null, output or file just by changing the pcmbuf_add_chunk function (which shouldn't balled much faster than 5hz.
23:12:49preglowsounds wonderful
23:12:53lostlogicthen make a plugin that wraps that plays to a destination one file at a time, and we can tiem codecs or write wavs for precision comparisons or whatever
23:13:12preglowagain
23:13:14preglowsounds like bliss
23:13:18CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
23:13:18*preglow huggles lostlogic
23:13:25*lostlogic won't make the plugin though :-P
23:13:31preglowthat poor playback system really needs this
23:14:04lostlogicit's not a bad system... it just suffers from waht I call "Just in case coding"
23:14:19preglowhaha, which is?
23:14:43 Join ashridah [0] (i=ashridah@220-253-123-152.VIC.netspace.net.au)
23:14:59Learhm... boost seems to be down a fair bit (but I moved eq_data to iram too)...
23:15:23lostlogicinstead of writing tight code that prevents weird cases, handle the weird cases with if catches... "Just in case" they come up
23:17:40preglowLear: i designed it so that moving eq_data to iram should not matter much unless the block size is really small
23:18:04 Part safetydan ("night all")
23:18:07Learwell, the processing block is 256 samples...
23:18:22preglowisn't it whatever the codec passes along?
23:18:28preglowi thought 256 samples was a resampler thing
23:19:05 Quit ashridah (Client Quit)
23:19:08lostlogicdsp_process will get 1 codec packet, up-to 1024 samples iirc
23:20:22Learwell, yes, but look at convert_to_internal, it never returns more than 256, thus breaking up the outer loop in dsp_process in 256 sample blocks.
23:20:49Learhe, 128 samples even.
23:21:11preglowthat needs to be changed
23:21:16preglowunless there is good reason
23:22:35lostlogicI wonder what kind of impact changing the PCM chunk size would have on performance...
23:22:44 Join ashridah [0] (i=ashridah@220-253-121-139.VIC.netspace.net.au)
23:22:53preglowwell, in most cases now, the codecs pass their data on in iram
23:22:55Learthis way, all filters knows the sample_buf and resample_buf is large enough, if they are needed.
23:23:01preglowand doing it all in one pass can be nothing but faster
23:24:08Leartrue, but this does keep things simpler... :)
23:24:27lostlogicpreglow: the codecs have an 8kb buffer allocated somewhere to pass back, or where does it come from?
23:25:04preglow8kb?
23:25:10preglowisn't the size dependent on the codec?
23:25:24preglowit is, unless we're talking about different thing
23:25:24preglows
23:26:19lostlogicThe codec creates an output buffer and passes it to the insert_split_callback which calls dsp_process on it...
23:27:05Learbut the sample_buf could easily be larger, now that resample_buf doesn't "have to" be 4x sample_buf. Would help the non-resample case...
23:27:25Learlostlogic: yes, and that buffer is codec-specific...
23:27:29preglowyes
23:27:33NicoFRhow hard would it be to fix the annoying progressbar behaviour when seeking (ie it goes back to the original position for a split second) ?
23:27:35preglowallocated statically on iram for most codecs
23:27:38preglowall i've touched, at least
23:31:00 Quit Lear ("Chatzilla 0.9.70 [Firefox 1.5.0.1/2006011112]")
23:32:15NicoFRanyone ?
23:33:39preglowshould be
23:33:41preglowbut no idea
23:36:06NicoFRok
23:36:16NicoFRmaybe i'll look into it
23:37:32preglowbut oh well
23:37:33preglowi require sleep
23:37:37preglowlater
23:39:56 Quit bazz ("Leaving")
23:45:09 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
23:51:05 Quit matsl (Remote closed the connection)

Previous day | Next day