tuxnes-devel — Development discussion for the TuxNES project

Re: [Tuxnes-devel] Weird configure thinghy

[Mike M. <mel...@pc...>]

On Mon, 26 Feb 2001, Jeroen Ruigrok/Asmodai wrote:

> If anyone has the experience to fix this please set it to -O by default.
> Otherwise I will if not objected against and will document how people
> can tweak TuxNES to squeak out the last remaining few cycles.

	I can't figure this one out. I know it's not always the default
for these scripts since I've written some simple programs and scripts that
don't use any optimizer level. If you can figure out how to fix it, have
at it.

-- 
	-Mike Melanson

[Tuxnes-devel] Sound and GTuxNES

[Mike M. <mel...@pc...>]

Hi team,
	Sorry to be sending so many messages tonight, but I really, truly
want to get to a release soon. It's been way too long since the last
release and TuxNES has advanced significantly since then.

	First question: Since sound is so good now, should it be on by
default? I'm thinking that the default output device would be /dev/dsp,
and the default output frequency would be 44100 Hz. Would the default
resolution be 8 or 16 bits? Then, we would need a mute option as opposed
to a '-s' sound option.

	Next, GTuxNES doesn't really reflect the current state of the
sound, does it? Does TuxNES still feature an echo option? And since we
have proper sound support, we no longer have (or need) the "simulate
square wave" option, right? I've been using GTuxNES to help me learn GTK
programming so I night be able to modify it myself.

	Also, I notice that we still have this option:
    O           - Toggle between sinusoidal and square wave synthesis
which doesn't seem to do anything. Get rid of it?

	Just doing some housekeeping/spring cleaning.
-- 
	-Mike Melanson

[Tuxnes-devel] zip file support

[Mike M. <mel...@pc...>]

Hi team,
	I need some advice on incorporating a patch sent to me awhile
back. A user needed support for zip files since his entire ROM collection
was contained within zip files, 1 by 1. I tried it out tonight and it
appears to work all right. However, when he first sent me the patch, the
idea struck me as unintuitive since a zip file can contain many files, as
opposed to a gzip file which can only contain one file.

	Like I said, it works quite well, and I guess it makes some
intelligent decisions. If there's one .nes file in the archive, it uses
that file. If there's more than one .nes file, it appears to use the 1st
alphabetical one. If there are no .nes files, it exits gracefully with
"unrecognized format".

	So, should we go ahead and incorporate the change and trust that
users won't try to open up multi-ROM zip archives to play select
ROMs? Or should we figure out some new CLI option to be able to select
certain files inside an archive? Gee, the latter option sounds like work
to me...:) It's great that the zip file support meets one user's
need. Maybe we should release the option as it is (and document its
precise functionality) and add more advanced handling if users ask for it?

	Or maybe I'm just making a mountain out of a molehill.

	Thanks...
-- 
	-Mike Melanson

Re: [Tuxnes-devel] CPU core graft

[Mike M. <mel...@pc...>]

On Mon, 26 Feb 2001, Mike Melanson wrote:

> 	Come to think of it, though, Nosefart is GPL'd and it relies on a
> 6502 emulator core. Have you checked that out?

	Quick follow-up: I just checked the Nosefart source (get it at
http://www.nofrendo.org/nosefart/ ) and Matt Conte's nes6502 CPU core
bears the GPLv2 notice on top.

-- 
	-Mike Melanson

Re: [Tuxnes-devel] CPU core graft

[Mike M. <mel...@pc...>]

On Fri, 23 Feb 2001, Jim Ursetto wrote:

> For this reason, I have grafted a portable cpu core
> (currently Marat's M6502) into tuxnes.  Since I just
> got it running yesterday, I haven't completely
> genericized the interface to allow plugging in any cpu
> core.  However, any core with the equivalent of read,
> write, and loop (irq check) operations should now be
> really easy to port.

	This sounds like a really interesting project, and one that I've
entertained myself. I'm glad to see someone is trying it out.

> I'm pretty sure Marat's license is incompatible with
> the GPL, but the glue between tuxnes and the new cpu
> core, being written by me, is. ;)

	I don't mean to discourage you from writing a new 6502 core, if
that's what you really want to do. But I think there are several open
source cores out there which have much looser licenses that Marat's (I
think Loopy has a core that's close to public domain). Again, if you'd
like to write a new 6502 core, more power to you, and if it was GPL'd, I
would say, "FINALLY!"...:) (a GPL'd 6502 core, that is).

	Come to think of it, though, Nosefart is GPL'd and it relies on a
6502 emulator core. Have you checked that out?

> Let me know if anybody is interested in looking at the
> code, even in its kind of haphazard state.

	I would be very interested to see it and eventually incorporate it
into the codebase. We may even be able to use it to emulate games with 
certain mappers that can't be emulated under the current NEStra/TuxNES
model.

-- 
	-Mike Melanson

[Tuxnes-devel] Moving CVS tree

[Mike M. <mel...@pc...>]

Hi team,
	It's really annoying that the TuxNES source code is in the CVS
tree root. I just put in a support request for them to move it if
possible, so just a heads-up in case it happens.

-- 
	-Mike Melanson

[Tuxnes-devel] Weird configure thinghy

[Jeroen Ruigrok/A. <as...@wx...>]

Why oh why does automake generate a CFLAGS in Makefile with -O2?

I don't mind optimisations, but experience has taught me to avoid any
optimisation level above -O.  It is even known to produce broken code on
Alpha platforms if set to -O2.

If anyone has the experience to fix this please set it to -O by default.
Otherwise I will if not objected against and will document how people
can tweak TuxNES to squeak out the last remaining few cycles.

-- 
Jeroen Ruigrok vd Werven/Asmodai    asmodai@[wxs.nl|bart.nl|freebsd.org]
Documentation nutter/C-rated Coder BSD: Technical excellence at its best  
	  D78D D0AD 244D 1D12 C9CA  7152 035C 1138 546A B867
I'm a child of the air, I'm a witch of the wind...

[Tuxnes-devel] CPU core graft

[Jim U. <ji...@3e...>]

Hello,

I'm looking at porting tuxnes to the Sega Dreamcast.
The first stumbling block is of course the heavy
dependence on the x86 instruction set, due to the
dynamic recompiler.

For this reason, I have grafted a portable cpu core
(currently Marat's M6502) into tuxnes.  Since I just
got it running yesterday, I haven't completely
genericized the interface to allow plugging in any cpu
core.  However, any core with the equivalent of read,
write, and loop (irq check) operations should now be
really easy to port.

The changes to get Marat's core to work were: implement
Read6502, Write6502, and Loop6502 operations; create C
functions to perform certain calculations previously
done (in assembly language) in INPUT:, OUTPUT:, NMI:,
and the MMC asm linkage functions; and a one-line
change to M6502.c for convenience.  There might be
other minor stuff.

I'm pretty sure Marat's license is incompatible with
the GPL, but the glue between tuxnes and the new cpu
core, being written by me, is. ;)

I've seen slight graphical glitches (the last line of
the status bar in Zelda sometimes gets corrupted, and
the map in Wizards and Warriors is slightly corrupted).
These may be my fault, or the tuxnes code may have been
dependent on peculiarities in the old asm code [tuxnes
has really WEIRD timing code, so I wouldn't rule this
out].

Let me know if anybody is interested in looking at the
code, even in its kind of haphazard state.

Jim

-- 
ji...@3e... / 0x43340710 / 517B C658 D2CB 260D 3E1F  5ED1 6DB3 FBB9 4334 0710

[Tuxnes-devel] CVS has been moved to Sourceforge

[Mike M. <mel...@pc...>]

Hi team,
	Sourceforge has finally imported the TuxNES CVS tree. If
anyone checked in files since last Friday and didn't tell the list, you'll
have to re-commit the changes because the tree snapshot was made last
Friday night. Click on the CVS link from the TuxNES website to find out
how to access the tree. Oh, something they forget to mention is that the
module name is ".".

	Paul, I've added you to the project but I can't tell if you have
write access to the repository. Other current developers, you still need
to register and send me your IDs.

-- 
	-Mike Melanson

[Tuxnes-devel] Here's the new list

[Mike M. <mel...@pc...>]

Hi team,
	Scratch that last message about manually subscribing; I figured
out how to transfer the whole list. Welcome to the new tuxnes-devel list,
hosted by Sourceforge.

-- 
	-Mike Melanson