Re: sjlj-exceptions ICE. simplified test case

[Mike Stump <mrs at windriver dot com>]

> Date: Tue, 25 Apr 2000 20:49:27 -0700
> From: Richard Henderson <rth@cygnus.com>

> I wonder if we shouldn't work to ban sjlj for gcc 3.0.  Even on
> a.out targets we can support dwarf2 unwind tables through collect2.

If people want to fix bugs by supporting other (dwarf2) EH schemes
than the sjlj scheme, that does seem entirely reasonable.  If people
want to discourage sjlj that also seems reasonable, but I'd vote for
leaving in sjlj for the time being.  It was put in for ports that
didn't have dwarf2 work done for them.  It was put in so that people
could have a system to throw through third party code, should they
need to.  It was put in so that C++ EH could be turned on be default
on for all targets, so that we have uniform support for C++, across
all targets.

I think table size is in part a red herring.  There are various table
compaction strategies I think could be used that are not used
currently.