Re: [Monotone-devel] Re: monotone & CVS import

[Nathaniel Smith]

On Thu, Nov 13, 2003 at 01:23:13AM +0200, Ori Berger wrote:
> Nathaniel Smith wrote:
> 
> >On the other hand, Windows filesystems will refuse to create files
> >with ':'s in them...  (I wonder what happens if you have a multibyte
> >UTF8 character, one of whose bytes is 0x3A == ':'...)
> 
> This can't happen. In a UTF-8 string, all non ascii chars have bit 7 
> set. The first byte also has a distinct range from the other bytes 
> in a multibyte string.
> 
> And it is done _exactly_ because of these issues.

Ah, for some reason I was thinking that the last character might not
have the top bit set.  I have no idea why.

Thanks for clarifying.

So it sounds like the only issue with UTF-8 is whether bytes with
their high bit set can cause security problems, and it seems doubtful
to me -- the only security problem I'm aware of is that we can't have
paths containing nasty metacharacters like "..", and what system would
have metacharacters with the high bit set?  Surely all systems will
either pass any such character through uninterpreted or reject it
altogether, right?

-- Nathaniel

-- 
The Universe may  /  Be as large as they say
But it wouldn't be missed  /  If it didn't exist.
  -- Piet Hein