This is the mail archive of the libc-alpha@cygnus.com mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: struct stat


> Date: Sat, 13 Mar 1999 13:43:44 -0500
> From: "David A. Greene" <greened@eecs.umich.edu>

> Geoff Keating wrote:
> > If you really really must do this sort of thing, use 'double'.
> > Fortunately, only an arithmetic type is required, according to the
> > standard.  You would have to ensure that inode numbers can't be NaN,
> > or -0, but this is a very easy thing to check, and it still leaves you
> > with more possible inodes than you can fit in a 2^64-byte filesystem.

Actually, I meant 'device', not 'inode' numbers.  Inode numbers are
still 32 bits at present.

> > (Of course, you'd use a real `long long' in the kernel and libc.)
> 
> When you say "you," I hope you mean "libc developers..."  :)

As in, "You would have..."?

I think that needs to be done by kernel developers.  It would be
inefficient to do in the libc.  I believe the kernel doesn't yet use
64-bit device numbers, so it's just another thing to take into account
when making the change.

-- 
Geoffrey Keating <geoffk@ozemail.com.au>


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]