Stephen J. Turnbull wrote:
> > > If you want to retrieve a filename from the OS then
pass it back at a
> > > later point, you need to retain the raw data. If you can't get at
> > > raw data, you lose.
> > That's exactly the conclusion the Python people just came to.
> Which conclusion? "Retain the raw data" or "you lose"?
"Retain the raw data or lose." There's no third alternative, although
sufficiently creative programmers can (and do) have their data and
lose anyway. ;-)
Oh. I was starting to worry that they had decided that OS functions
were going to start automagically trashing^Wconverting everything for
[Like wxWidgets, which likes to provide its own main() which
"conveniently" converts argv to Unicode.]
The fact is that the problem is the Tower of Babel. One ISO
is not going to turn back God's wrath (in fact, it probably just made
Her madder!) Unicode is a major step toward making the world safe for
low energy/high burden programmers, at least in a restricted area of
multilingual and/or localized text processing. But as usual, the 10%
of corner cases involve 90% of the work, and also as usual, those of
us who care about the corner cases are going to have to bear the
burden of dealing with them.
Surely that doesn't surprise you. ;-)
The problem isn't with having to do stuff myself, it's with being
Once upon a time, if you wanted to hide malicious files, you had to
resort to tricks like putting them in a subdirectory named "..." or
embedding control characters in the name. Nowadays, you just need to
give them a name which isn't valid UTF-8 and many programs will
just pretend they don't exist.
Glynn Clements <glynn(a)gclements.plus.com>
XEmacs-Beta mailing list