jwasm v2.02

[japheth]

Hello,

JWasm v2.02 was just released.

No new features, just bug fixes. For Windows programming, at least one - reported by drizz - is important:

Code Select Expand


   -  runtime conditionals: the 'logical NOT' operator ! did not
      invert logical operators && and ||.


So switch to it ASAP!

[jcfuller]

japheth,
  404 on the change log link.

James

[japheth]

Thanks! fixed.

[jj2007]

Not sure if this requires a fix... it does assemble with ml.exe, though.

Code Select Expand

include \masm32\include\masm32rt.inc

SwitchArg = 1 ; the option we choose
UseNothing = 0 ; option 0: do nothing
UseEdit = 1 ; option 1: create an edit control
UseListbox = 2 ; option 2: create a listbox

_Switch equ if 0
_Case equ elseif SwitchArg eq
_Endsw equ endif

.code
start:
%_Switch
%_Case UseEdit
print "Case Edit", 13, 10
%_Case UseListbox
print "Case Listbox", 13, 10
%_Case UseNothing
print "Case nothing", 13, 10
%_Endsw
exit

end start


[jj2007]

On my Celeron, the RichMasm source (12,000 lines) assembles & links in 800 milliseconds with ml.exe, and 660 milliseconds with Jwasm.exe of January 1, 2010. With the new version, it takes 2,700 ms... an extra pass??

Note that other sources have not slowed down. Maybe it's something very peculiar to this one.

[japheth]

Not sure if this requires a fix... it does assemble with ml.exe, though.

You're probably the only one which uses this peculiarity. Exploring wild lands is sometimes slightly uncomfortable.

On my Celeron, the RichMasm source (12,000 lines) assembles & links in 800 milliseconds with ml.exe, and 660 milliseconds with Jwasm.exe of January 1, 2010. With the new version, it takes 2,700 ms... an extra pass??

One or two extra passes won't take that long. With JWasm, usually the first pass is the one which takes the lion part of total assembly time. If you are able to supply the source, I'll take a look.

[jj2007]

If you are able to supply the source, I'll take a look.

Check mail@japheth.de

Cheers,
Jochen

[japheth]

Check mail@japheth.de

Won't work. I removed the email program from japheth.de. Too much spam. Please use the contact address found at japheth.de!

[japheth]

JJ,

> One or two extra passes won't take that long.

Yes, but JWasm v2.02 needs 42 passes instead of 5 to assembly the source. This results in a significantly longer total assembly time. It's a "backpatching" issue. Fixing this issue was on the agenda for v2.02, but was postponed and is now scheduled for v2.03.

the current preliminary v2.03 has the "backpatching" status reverted to the one of 1. Jan., so it will do better.

[jcfuller]

Slashdot posting: http://developers.slashdot.org/story/10/01/29/1632253/x86-Assembler-JWASM-Hits-Stable-Release?art_pos=14

James

[sinsi]

http://www.asmcommunity.net/board/index.php?topic=29797.0