Mail Archives: djgpp-workers/2003/03/26/05:59:31
Hello.
Eli Zaretskii wrote:
>
> > Date: Wed, 19 Mar 2003 20:46:52 +0000
> > From: Richard Dawe <rich AT phekda DOT freeserve DOT co DOT uk>
> >
> > We could modify stat to use _invent_inode for directories, so that it
> > generates the same inode every time it is called for the directory. We
> > would use _invent_inode, because we cannot rely on the current method
> > (on Win2k/XP) to return the same inode each time.
>
> Are you saying that the current code does _not_ invent inodes on
> Windows/XP? I'd be surprised, I thought we invent inodes for any
> kind of files on all the versions of Windows except 3.1.
It invents inodes on Windows '98 SE. Maybe _ioctl_get_first_cluster can get an
inode on Win2k/XP, but doesn't return consistent results? That would be weird,
but it seems possible.
> > To ensure that we get the same inode, we
> > would also have to ensure that we pass the same filename to _invent_inode.
> > So the directory's filename would need to be fixed using _truename.
>
> I thought `stat' already calls _truename and the result is passed to
> _invent_inode, isn't that true?
Yes, that is true.
I think someone with Win2k/XP needs to investigate this more. The original
e-mail is linked from the DJGPP 2.04 status page:
http://www.phekda.freeserve.co.uk/richdawe/djgpp/2.04/status204.html
Here's the link for convenience:
http://www.delorie.com/archives/browse.cgi?p=djgpp-workers/2002/10/28/17:03:21
I'd do it myself, but I don't have Win2k/XP.
Bye, Rich =]
--
Richard Dawe [ http://www.phekda.freeserve.co.uk/richdawe/ ]
- Raw text -