Re: UID width

Theodore Y. Ts'o (tytso@MIT.EDU)
Tue, 12 Jan 1999 00:11:50 -0500


Date: Sat, 9 Jan 1999 11:34:26 +1300
From: Chris Wedgwood <chris@cybernet.co.nz>

On Fri, Jan 08, 1999 at 04:37:22PM +0300, Khimenko Victor wrote:

> Ufortunatelly even with glibc2 this transition will not be
> painless. Take a look:

Indeed... I've been trying to work out a sane way of doing this for
some time now. Suggestions welcome.

> -- gnu/types.h --
> typedef unsigned short int __ipc_pid_t;

Is 16-bit -- ouch. It really needs to be 32-bit. I don't get how this
occured, glibc uses 32-bit pid_t everywhere except here.

My best solution so far, is break this API before it gets too
ingrained in code out there (which sucks, because this will break
sybase, which rocks).

We can fix this without forcing people to recompile and without forcing
a major version number bump on the shared library by using ELF symbol
versioning. Eric Youngdale did a presentation on this one or two Linux
Expo's ago, but as far as I know no one has attempted to actually use
this to rev a libc interface to date. This would be a good test case, I
think.

- Ted

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/