Success report on Linux/PPC, small Ada problem

[Robert Dewar]

>>They need to set a goal of once a week, come hell or high water.

That's certainly the goal, as I mentioned earlier, part of the trouble is
that we are not at the starting point yet of having a reliable GCC 3 version
of GNAT, but we are extremely close, especially on GNU/Linux (Solaris is not
far behind, Tru64 has a bit further to go, HPUX is also in reasonable shape).

<<I wish I had contributed the merge script... so I could point them at
it.  Maybe Jason has a copy of it.
>>

We are developing our own scripts, we have some special needs. For instance
our revision histories are required internally to be keyed to the tracking
numbers of the related problem reports or enhancement requests to which they
refer, but these numbers have no relevance in the FSF tree, and so must be
stripped. 

We will also need to automate any agreed on differences between the two
trees (e.g. header formats that mention ACT etc).