[patch] Re: Another weird sparc solaris 2.8 compiler error

Thu Feb 1 15:29:00 GMT 2001

Tom Tromey writes:
> Does anybody know why we had an error() function here in the first
> place?

It not something that we've been using recently, or even ever. Maybe
it was satisfying a symbol while linking along with some other .os
from gcc. Oh, I see: we used to link jvgenmain against mangle.o, and
mangle.o always used `error.' We don't anymore because I couldn't get
it quickly to work when I was working on the new mangling scheme. Now,
maybe it was a bit short sighted from me to do that, especially with
the UTF8-in-mangled-name-if-supported that we'll put in shortly, plus
the fact that we aren't able to generate `main' for a class bearing
unicodes in its name anymore :-(
I'll put that on my todo list. 
./A