James Carlson <[EMAIL PROTECTED]> wrote:
> Joerg Schilling writes:
> > James Carlson <[EMAIL PROTECTED]> wrote:
> > > Right ... plus, if it were important enough to contend with, I'd think
> > > we could create (yet another) new stat interface to deal with the
> > > problem.
> >
> > I would be really happy with this!
> >
> > I would propose to make it effective for all newly compiled programs that
> > request the large file interface. The old interface remains for binary
> > compatilility for programs that use the Solaris-2.6 largefile stat
> > interface.
>
> Sounds mostly good, but I'd like to see the compatibility issues
> worked through in a little more detail. (As part of a project,
> probably.)
To be honest, this would only work if time(3), gettimeofday(2), utimes(2), ...
are changed also together with all time_t users.
I was amazed to see how many places are affected when I convinced the FreeBSD
people to use a 64 bit time_t instead of a 32 bit one in the 64 bit sparc
version of FreeBSD a few years ago. This is no simple change as e.g. even
timeouts from init and svc.startd are affected. The result on FreeBSD was that
cleanly written code only needed a recompile. In case if a 32 bit OS, things
look different as more potential printf formats are affected.
Jörg
--
EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin
[EMAIL PROTECTED] (uni)
[EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/
URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily
_______________________________________________
opensolaris-discuss mailing list
[email protected]
