Why not gnat Ada in gcc?

[Robert Dewar]

<<The point was that bulk merges (regardless of where they come from) are a
particularly bad way to sync the official GNU sources with some other
source base.  I was just using the merges you did as an example.  The
bulk merges I used to do from the gcc2 sources before gcc2 deveopment
finally died were just as disruptive.
>>

One clarification that may be useful since it has caused confusion in
at least one previous post, is that there are two issues here.

1. Doing a large bunch of changes at the same time in a coordinated
manner. That can be disruptive certainly, and disruption is to b e
avoided. On the other hand uncoordinated changes done separately
can be even more disruptive. The goal is minimizing disruption, and
generally incremental updates are desirable from this point of view,
no one objects. Indeed we have been maintaining our changes as separate
patches for some time, with the idea that when the GNAT/GNU/Linux folks
put up a tree, we would provide them, but that didn't happen. So now
that will happen in the context of the gcc tree.

2. Combining separate changes into single patches. This is of course
never allowed under any circumstances, and certainly is something we
would expect to avoid unconditionally.