This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: A (somewhat) strange Sparc library issue...
- To: mrs at wrs dot com
- Subject: Re: A (somewhat) strange Sparc library issue...
- From: "David S. Miller" <davem at dm dot cobaltmicro dot com>
- Date: Thu, 27 Aug 1998 11:49:23 -0700
- CC: egcs at cygnus dot com, rth at cygnus dot com
- References: <199808271844.LAA23681@kankakee.wrs.com>
Date: Thu, 27 Aug 1998 11:44:52 -0700
From: mrs@wrs.com (Mike Stump)
I'd just change it and not look back. First, sparc linux is
relatively new.
Yes but I have users, and I have concerns that their systems will
break mysteriously if I choose some of the options I have available.
Consider anyone who had been using "long double" in some library of
some sort, and compiled it with the old compiler (or even better, they
didn't compile it, it's a core library on their system).
Look at it this way, the people that wanted longer than double doubles
and used long doubles will never know that yesterday it was broken and
you fixed it. If you fail to fix it, they will just submit bug
reports complaining about it.
They will know, inadvertantly, because today they will get failures
when linking newly compiled code with older libraries having been
compiled previously before egcs was "upgraded".
Back to reality, what I am doing in my spare time is trying to make
sure a major library component of sparc linux systems won't have this
issue (ie. glibc etc.). Based upon what I find there, I'll have a
good idea of what avenues are safe to take.
Later,
David S. Miller
davem@dm.cobaltmicro.com