RE: cvs via Cygwin (W98) to FAT to Linux - permissions

[Steve Jorgensen <jorgens at coho dot net>]

-----Original Message-----
From:	Robert Collins [SMTP:robert.collins@itdomain.com.au]
Sent:	Wednesday, June 27, 2001 7:33 PM
To:	jorgens@coho.net; Cygwin List (E-mail)
Subject:	RE: cvs via Cygwin (W98) to FAT to Linux - permissions

> -----Original Message-----
> From: Steve Jorgensen [mailto:jorgens@coho.net]
> Sent: Thursday, June 28, 2001 12:34 PM
> To: Cygwin List (E-mail)
> Subject: RE: cvs via Cygwin (W98) to FAT to Linux - permissions
>
>
> It seems to me there are some big potential dangers to a
> Cygwin UMSDOS
> implementation.  The main issue is that UMSDOS expects that
> it is the only
> thing that will be modifying anything in a synchronized
> directory while it
> is mounted.  I don't think this can conceivably be enforced inside a
> running Windows session short of holding locks on every single file.

I'm to quite sure what you mean here. Cygwin1.dll will be providing the
only access to the files in the UMSDOS system, so syncronisation
shouldn't be an issue. Sure someone deleting a file out from within a
win32-only program could cause confusion, but thats not a UMSDOS only
issue :}.

Could you clarify a little?

[Steve Jorgensen]  Perhaps you're right, I didn't think that through all 
the way.  After all, UMSDOS only cares that the mangled filename be 
preserved, and code could probably be added easily enough to handle files 
that go missing because someone deleted them through Explorer or something.

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Bug reporting:         http://cygwin.com/bugs.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/