On Thu, 17 Dec 2020, tom ehlert wrote:
DOES THIS DIFFER FROM MSDOS BEHAVIOUR?
This.
When I wrote some DOS elements of my own about 20 years ago, the first
priority was compatibility. There were some places where I thought MS-DOS
behavior was daft, and offered an option to disable it, but went out of
the way to make sure it was as bug-for-bug compatible with MS-DOS as,
within the limits of my ability to write it, I could achieve. Even some
obscure or outright bizarre stuff was put in, and special checking for
very specific MS-DOS behavior which I knew to be relied on (e.g.,
accepting oddities like "cd\" as equivalent to "cd \") was required. To
me, "this is daft, I'm not implementing this" or "this is broken, I'm
doing something saner" weren't options, if they broke compatibility.
I think at some point here among some people the idea of being an
open-source reimplementation of DOS compatible with MS-DOS 3.31 (or 6.22)
was lost, but I am pretty sure that was always the ultimate goal. In that
regard, I believe that if a change causes programs to break because they
expect MS-DOS to do X, and FreeDOS does Y, it is a regression.
I am pretty sure that the only MS-DOS program where "abcdefghijk" as a
filename is interpreted as "abcdefgh.ijk" is Microsoft BASIC
(gwbasic/qbasic etc.). I may be wrong in this regard, it is documented
behavior there, but I know of no other software for DOS that does this.
-uso.
_______________________________________________
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel
