Mail Archives: djgpp/1995/08/02/14:22:23
> This is about the 2nd most ridiculous post I've seen on such subjects.
> This fellow has hard facts to support a claim. You present nothing but
> arrogance. Present some numbers to back up your claim that with
Actually, I felt the original post was a bit arrogant (I wondered if the
guy worked for Watcom, actually). OK, I don't have Watcom, so maybe it
is the best thing out there. But lets do a real comparison, instead of
a meaningless blurb.
Proposed test:
1) Image size comparison must include all code needed to run in the following
environments: Naked config.sys, himem only, emm386, windows, on a
distributed image.
2) Image should call a minimum of 10 external functions and use the
argc, argv. Wildcard expansion should be supported in both or a
non-glob setup be used in V2 image. Environment variables and signals
should be used.
BTW, I seldom use the C++ portion, but I was very surprised at the size
bloat there (I think something is wrong).
With a different libc, I have written V2 images which happily run with
a total size of around 18K on the hello-world thing, so I know its a
feature issue.
But I can do the same functionality as the "hello world" type prog in a
50 *byte* .com file, so does that mean that assembler is always 500 times
smaller than Watcom? See how ridiculous these comparisons can be?
- Raw text -