On Tue, 20 May 2008, Eric Blake wrote:
| What qualifies a platform as not being worth porting to?
It's somewhat subjective. The gnulib mailing list has come up with an
ad-hoc definition of any platform released within the last 3-5 years, and
preferably where the vendor has not declared the platform end-of-life,
while also supporting older platforms where the vendor is still actively
providing support. So, as an example, Solaris 4 and Windows 95 are
generally not viable porting targets (their vendors no longer support
Clearly these are not suitable rules for libtool. Autoconf and
libtool need to be dicated by whether the features they need to
execute are present (might require augmentation such as replacement
compiler and shell), and not how old the target is, or if its
manufacturer still cares about it.
I am quite certain that Solaris 2.4 is a solid POSIX-compliant OS
which is capable of performing useful tasks today just as it did when
it was released. It does not have POSIX threads, but it has most
other useful Unix things that we enjoy today. Some Y2K and
daylight-savings bugs may be present but this is not necessarily cause
to throw the hardware away.
To me, a 5 year old system has reached middle age and a 3 year old
system is practically new. Ten years is senior citizen and should
seek retirement.
Regardless, we should not be changing the libtool 2.2.X install
footprint for the sake of change. There should be a good reason.
Bob
======================================
Bob Friesenhahn
[EMAIL PROTECTED], http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer, http://www.GraphicsMagick.org/
