nasm-devel — NASM development work

[Nasm-devel] Merging SciTech changes

[Kendall B. <Ken...@sc...>]

Hi All,

Nelson mentioned in a prior email to this list about getting our SciTech 
changes merged into this new SourceForge project. It appears that this has 
now taken place, and I am in the process of syncing up to the newly merged 
sources and will be testing them with our projects.

Thanks Nelson!

Some other comments from what Nelson said earlier:

> I`ve been trying to find people interested in working on Nasm and
> so far I haven`t had much luck. Perhaps this is why the original
> project died, nasm must seem very complete to most developers by
> now. 

We actually also tried to revive the NASM project, and set up a mailing 
list for people interesting in using and enhancing NASM some time ago. 
Right now there are about 40 people subscribed to our mailing list, and I 
sent out an email to that list a few days ago telling them about this new 
project. I hope those people will also subscribe to this new mailing list, 
as I mentioned that we plan to close down our list so all traffic should 
end up here.

> All that I would expect would be that if you make further
> substantial changes to contribute them to the project and as the
> number of active members grow we can all participate in keeping the
> tree sane. 

If we do make changes to NASM I have everything set up with Perforce so 
that I can easily merge my new changes automatically back into the CVS 
respository. Hence I don't develop directly with CVS, but with a branch in 
Perforce. Due to the powerful branching stuff in Perforce (and some 
external programs I have written ;-), it makes it real easy to keep up to 
date with CVS projects and submit changes back.

BTW, has anyone looked into integrating the NASM patches that AMD did to 
include better support for MMX and 3DNow stuff in NASM? I have the latest 
patches that they sent me on the SDK, but have not had a chance to look 
into it. If you want I can post the patches to this list so we can look 
into making sure those changes and updates are included in this new version.

BTW2, perhaps we should change the version on this new NASM to '0.99'?

BTW3, it would be preferrable if our TASM_COMPAT stuff is all compiled in 
by default. I did my best when I added the latest changes (which I 
submitted multiple times to the old developers, but they never got 
integrated) to ensure that regular NASM mode was not changed, and the TASM 
compatible mode is activated with the '-t' command line switch.

Regards,

+---------------------------------------------------------------+
|   SciTech Software - Building Truly Plug'n'Play Software!     |
+---------------------------------------------------------------+
| Kendall Bennett          | Email: Ken...@sc...    |
| Director of Engineering  | Phone: (530) 894 8400              |
| SciTech Software, Inc.   | Fax  : (530) 894 9069              |
| 505 Wall Street          | ftp  : ftp.scitechsoft.com         |
| Chico, CA 95928, USA     | www  : http://www.scitechsoft.com  |
+---------------------------------------------------------------+

[Nasm-devel] documentation

[Frank K. <fbk...@ne...>]

Don't all speak at once!

Nelson Rush said:
> You`ve updated the nasm documents to reflect the changes I presume?

Documents? Updated? What good news *that* would be!

I've "updated" the official Nasm Docs for 0.98 (same as docs for 0.97,
pretty much, I think) to at least list the "KATMAI" and "3DNow!"
instructions supported by 0.98 but not mentioned in the docs. There are
other features that appear in the source but not in the docs - debugger
support for Borland (?) and the "-m" switch, e.g. My progress so far is
at:
http://www.geocities.com/SiliconValley/Heights/1295/

The current html output works on my drive, but not on-line - gotta look
into that.

So, are the following instructions supported in the "new" version? (I
forget who sent me the link, but thanks!)

------------------------------------------------------------
http://homepage1.nifty.com/herumi/soft/gogo2/readme_e.txt

 (9) APPENDIX ( How to make NASM supporting Enhanced 3D Now! )

  Append the following source to insns.dat and make, and you get NASM
which
  can assemble Enhanced 3D Now!.
  At own your RISK, get
http://homepage1.nifty.com/herumi/soft/nasm098-e3dn.lzh
  for Windows. (NO WARRANTY!)

--- x8 --- x8 --- x8 --- x8 --- x8 --- x8 --- x8 --- x8 --- x8 --- x8
--- x8 --- x8 ---
PI2FD     mmxreg,mem          \301\2\x0F\x0F\110\01\x0D    
PENT,3DNOW,SM
PI2FD     mmxreg,mmxreg       \2\x0F\x0F\110\01\x0D         PENT,3DNOW
PFNACC    mmxreg,mem          \301\2\x0F\x0F\110\01\x8A    
PENT,3DNOW,SM
PFNACC    mmxreg,mmxreg       \2\x0F\x0F\110\01\x8A         PENT,3DNOW
PFPNACC   mmxreg,mem          \301\2\x0F\x0F\110\01\x8E    
PENT,3DNOW,SM
PFPNACC   mmxreg,mmxreg       \2\x0F\x0F\110\01\x8E         PENT,3DNOW
PI2FW     mmxreg,mem          \301\2\x0F\x0F\110\01\x0C    
PENT,3DNOW,SM
PI2FW     mmxreg,mmxreg       \2\x0F\x0F\110\01\x0C         PENT,3DNOW
PF2IW     mmxreg,mem          \301\2\x0F\x0F\110\01\x1C    
PENT,3DNOW,SM
PF2IW     mmxreg,mmxreg       \2\x0F\x0F\110\01\x1C         PENT,3DNOW
PSWAPD    mmxreg,mem          \301\2\x0F\x0F\110\01\xBB    
PENT,3DNOW,SM
PSWAPD    mmxreg,mmxreg       \2\x0F\x0F\110\01\xBB        
PENT,3DNOW,SM
FFREEP    fpureg              \1\xDF\10\xC0                
PENT,3DNOW,FPU
--- x8 --- x8 --- x8 --- x8 --- x8 --- x8 --- x8 --- x8 --- x8 --- x8
--- x8 --- x8 ---


  sh...@ku...,
http://homepage1.nifty.com/herumi/gogo_e.html
----------------------------------------------------------------
Is this the same "0.98e3"?

How about that "RETFD" (was that it?) that got left out of 0.98? And
let's see, there's the "ADD" instruction...

Doc says:
"In the forms with an 8-bit immediate second operand and a longer first
operand, the second operand
is considered to be signed, and is sign-extended to the length of the
first operand. In these cases, the
BYTE qualifier is necessary to force NASM to generate this form of the
instruction."

Nasm now generates the "short form" automagicly, right?

I'm not volunteering to be the "Nasm Documentation Project", but I'll
try to help out a little if I can.

Best,
Frank

RE: [Nasm-devel] merging two NASM branches

[Nelson R. <che...@at...>]

I've just begun this project as a Nasm revival. The original site hasn't
been updated in months, and the original authors and maintainers cannot be
reached.

I've been reviewing your diff and changes and it all seems reasonable.
You've updated the nasm documents to reflect the changes I presume?

If you would like to join the project let me know your sourceforge account
name.
If you could let me know how to get in touch with the SciTech crew that is
working on its own Nasm branch I'll talk to them about this as well.

Hopefully we can work together to keep all our changes in a single release.
I've been trying to find people interested in working on Nasm and so far I
haven't had much luck. Perhaps this is why the original project died, nasm
must seem very complete to most developers by now.

I understand that you don't want to be responsible for maintaining the
latest release and I don't want to put pressure on anyone to do so. All that
I would expect would be that if you make further substantial changes to
contribute them to the project and as the number of active members grow we
can all participate in keeping the tree sane.

I will do my best to keep things updated. I really like nasm.

If you would like to forward this message to Kendall it would help. And
mention this mailing list to him as well.

Sincerely,
Nelson Rush

-----Original Message-----
From: nas...@li...
[mailto:nas...@li...]On Behalf Of kd...@ne...
Sent: Wednesday, July 05, 2000 9:12 AM
To: nas...@li...
Subject: [Nasm-devel] merging two NASM branches


Hello nasm-developers,

Just want to point you out that there is a parralel branch
of NASM
developement. SciTech (and particulary Kendall Bennet) are
doing
some improvements. They have their latest sources on their
Perforce
server (this is a kind of ugly<imho> CVS).

Also I am the author of nasm 0.98e that is on
linuxgames.com (you have
a remark about it on your "news" page). I have already
integrated my
latest patches into the SciTech"s version of NASM. I would
like to
propose to join our efforts since there is no much sense in
making
a branch of NASM. Moreover, neither me or SciTech have big
interest
in maintaining the "latest NASM" so if we could donate our
changes
to some team which will take further care of NASM I
personally (and
I"m almost sure Kendall as well) will be happy. We both are
going
to make further improvements to NASM from time to time but
for both
me nor Kendall this is not going to become the MainTime
Project
(its rather a MeanTime project). We just need NASM for our
projects
(I need it for Crystal Space project and Kendall for his
MGL project).

A short description of changes made by scitech:
- they added a TASM-compatible mode, that is, you may
compile
same code (with several precautions) with both TASM and
NASM. It is a
subset of Tasm"s IDEAL mode and is enabled by an additional
command-line switch "-t". Without "-t" NASM acts absolutely
as before.
They need such a mode for their MGL graphics library which
can be
compiled with both TASM and NASM.

On my part of changes, I have fixed some things in OMF
output routines
(I had problems with using Nasm"s object files under OS/2),
added some
new preprocessor defines (I have added %undef long time
before 0.98
came out, also I added %xdefine and %ixdefine which are
like %define
but expands the macros immediately, that is:

%assign a 5
%xdefine b a
%assign a=a+1

This will define "b" to "5", while if I would use simply %
define, it
would expand to latest value of a, in above case to "6".

Also I have added the %+ pseudo-operator for concatenating
substrings
in %define and alike. Example:

%define cextern(x) _ %+ x
cextern (myfunc)

The second line will expand to _myfunc. There was no
universal way to
do above job in NASM 0.98, thus this improvement.

Besides above changes, I have fixed a large number of NASM
bugs I have
encountered while using NASM for Crystal Space project (we
have a lot
of assembly code there, about 150K).

--
Best regards,
 Andrew Zabolotny

P.S. I"m not at home now (I"m even in other city) but my
email
address is bi...@el.... Please answer there as I won"t
have access
to the email account I am writing from. I will be there at
the end of
this week, I hope. I will join this mailing list ASAP, but
until
Monday please don"t reply me on the list.

P.P.S. I will ask also Kendall about merging together our
NASM brand
with your. I don"t think he will object.


_______________________________________________
Nasm-devel mailing list
Nas...@li...
http://lists.sourceforge.net/mailman/listinfo/nasm-devel

[Nasm-devel] merging two NASM branches

[<kd...@ne...>]

Hello nasm-developers,

Just want to point you out that there is a parralel branch 
of NASM
developement. SciTech (and particulary Kendall Bennet) are 
doing
some improvements. They have their latest sources on their 
Perforce
server (this is a kind of ugly<imho> CVS).

Also I am the author of nasm 0.98e that is on 
linuxgames.com (you have
a remark about it on your "news" page). I have already 
integrated my
latest patches into the SciTech"s version of NASM. I would 
like to
propose to join our efforts since there is no much sense in 
making
a branch of NASM. Moreover, neither me or SciTech have big 
interest
in maintaining the "latest NASM" so if we could donate our 
changes
to some team which will take further care of NASM I 
personally (and
I"m almost sure Kendall as well) will be happy. We both are 
going
to make further improvements to NASM from time to time but 
for both
me nor Kendall this is not going to become the MainTime 
Project
(its rather a MeanTime project). We just need NASM for our 
projects
(I need it for Crystal Space project and Kendall for his 
MGL project).

A short description of changes made by scitech:
- they added a TASM-compatible mode, that is, you may 
compile
same code (with several precautions) with both TASM and 
NASM. It is a
subset of Tasm"s IDEAL mode and is enabled by an additional
command-line switch "-t". Without "-t" NASM acts absolutely 
as before.
They need such a mode for their MGL graphics library which 
can be
compiled with both TASM and NASM.

On my part of changes, I have fixed some things in OMF 
output routines
(I had problems with using Nasm"s object files under OS/2), 
added some
new preprocessor defines (I have added %undef long time 
before 0.98
came out, also I added %xdefine and %ixdefine which are 
like %define
but expands the macros immediately, that is:

%assign a 5
%xdefine b a
%assign a=a+1

This will define "b" to "5", while if I would use simply %
define, it
would expand to latest value of a, in above case to "6".

Also I have added the %+ pseudo-operator for concatenating 
substrings
in %define and alike. Example:

%define cextern(x) _ %+ x
cextern (myfunc)

The second line will expand to _myfunc. There was no 
universal way to
do above job in NASM 0.98, thus this improvement.

Besides above changes, I have fixed a large number of NASM 
bugs I have
encountered while using NASM for Crystal Space project (we 
have a lot
of assembly code there, about 150K).

-- 
Best regards,
 Andrew Zabolotny

P.S. I"m not at home now (I"m even in other city) but my 
email
address is bi...@el.... Please answer there as I won"t 
have access
to the email account I am writing from. I will be there at 
the end of
this week, I hope. I will join this mailing list ASAP, but 
until
Monday please don"t reply me on the list.

P.P.S. I will ask also Kendall about merging together our 
NASM brand
with your. I don"t think he will object.

[Nasm-devel] RE: Bug report on your web site :-<

[Nelson R. <che...@at...>]

Oh, I've been busy lately and haven't had time to finish the website.

The NASM project is dead as far as I know. I've tried to get into contact
with the original authors, maintainers, and members and none of them
respond. As a result, I set up a project on SourceForge to continue it. I
really could use some help though, since I really can't focus on keeping it
maintained by myself and no one else has shown any interest.

The original website for NASM is very very broken, and I gleaned as much
information off it for the current site as I could. I've also started to
design the site with PHP and make it more professional.

NASM hasn't been updated in a year and I don't think it is being maintained
any longer. I really hope to revive it with the project on SourceForge,
which has CVS access and mailing lists. I've already imported the v0.98 into
the CVS tree.

If you'd like to join in and help with the website, development,
documentation, or recruiting coders just let me know and I'll add you.


-----Original Message-----
From: ro...@li... [mailto:ro...@li...]On
Behalf Of Adam L.
Sent: Tuesday, June 27, 2000 4:16 PM
To: pal...@us...
Subject: Bug report on your web site :-<


This link doesn't work http://nasm.sourceforge.net/chat.php on the top
menu:

Not Found

The requested URL /chat.php was not found on this server.


Apache/1.3.12 Server at nasm.sourceforge.net Port 80


And what is the status of nasm? Does someone keep running the update?
I didn't see news on nasm for aproximatively 1 year!!!
Is it dead? What's going on?

Thx for porting the web site on sourceforge. :-))))