Re: DOSEMU/DOSBOX question

[Wally Bass wbass@xxxxxxxx]

Flash wrote:
>Wally Bass wrote:
>>... you should probably start thinking about the form of your
>>data now.

>Good thinking. But is this necessary for xy4 with the
>appropriate updates from Robt., Carl, and co.? Turbo xy4 can
>read long file names (one of the reasons to upgrade from xy3).
>Or did I mess something earlier in this thread?
I haven't looked at functions you are referring to, but my guess is that
they don't solve much of the problem that we are talking about here.
The XY4 documentation never mentions long filename, so I presume that the
the XY4 code uses only the old DOS calls for IO, which intrinsically deal
only with 8.3 uppercase filenames. Unless the functions to which you are
referring are binary (e.g., not XPL) functions which actually effectively
change the DOS file system calls that XY4 is doing, way down in guts of
EDITOR.EXE, they won't solve the problems being discussed here. Read Paul's
posts, in particular.
Consider the case where I have a file with a non-8.3 name -- say
"ratherlong.txt". There are mostly two questions:
The first would be, can I open the file with XY for reading. To to that, I
must know the mangled name that DOSEMU has constructed for the file --
let's say it's "RATHE~AB.TXT". The updates you mention may very well help
in that they could either tell me the "RATHE~AB.TXT" name so that I know
what to enter, or they could allow me to enter "ratherlong.txt" on the
command (or select it from a list), and then map that to "RATHE~AB.TXT" in
constructing the XY command that opens the file. But unless there are
binary mods inside the guts of XY4, the file is going to get opened using
the name RATHE~AB.TXT in the open call.
The second question, then, is the behavior when the file is saved, after
some editing. In particular, will the "ratherlong.txt" name be preserved
for use by Linux applications which reference the file. For a DOS window
under MS Windows, the answer is "yes" the long filenames will be preserved
(for the Windows apps). But the answer under DOSEMU by virtue of Paul's
testing is "no", and unless the functions mentioned accomplish what seems
to be the impossible, they aren't going help with this part of the problem.
Perhaps someone familiar with the functions that Flash describes would
indicate whether the above assessment is correct.

Wally Bass