krb5 thread support and excess support libraries -- seeking opinions, options

[Sam Hartman]

There seem to be a lot of options being tossed around here.  ANd there
are a lot of proposals that may not ever be worth implementing.  I'd
suggest that any additional discussion in this thread be phrased in
terms of tradeoffs between a variety of issues.

1) Making things easy for application developers.

2) Avoiding introducing portability issues for the build system

3) Minimizing complexity

4) Maximizing architectural cleanlyness

5) Minimizing implementation cost


There are a few other potential issues like performance.  Personally I
think that performance is not a significant factor in anything
discussed so far.

I do not think that minimizing number of libraries is a valid goal in
and of itself although I note that it does interact with some of the
goals expressed above.