Shared library annoyance with gcc-3_0-branch

[dewar@gnat.com]

<<There have been many examples in these mail lists of people
being encouraged to submit partial solutions to hard problems
that can be perfected later.  That's precisely what I was trying
to do.  Perhaps my rhetorical skills were weak.
>>

Sorry, I definitely do not want to discourage that. I think two things
are important:

1. Partial solutions should not break existing code

2. There needs to be someway to distinguish real releases from
releases which contain partial solutions of this type.