Re: Success report on Linux/PPC, small Ada problem

[dewar at gnat dot com (Robert Dewar)]

<<While there are separate trees, it isn't meaningful to say that "current"
only applies to the ACT one.  I'm quoting the *current* FSF documentation;
you're referring (I suppose) to the current ACT documentation.  If the ACT
documentation is more appropriate for a GNU manual than the FSF
documentation, then perhaps a swap between the manuals in the two trees is
needed.  If the docs in GCC are "obsolete", they are still the current GCC
docs.
>>

Indeed, we are doing major updating and reorganization work on the
documentation right now, one component of which is making changes to
make the manuals more conformant with the FSF environment. Hopefully we
can get versions of these available as soon as possible (that is actually
by far the most important point in this discussion). The documentation
that is there now with the FSF version is obviously inadequate, no argument
about that.


One problem that is significant is the VMS issue. The current manual is
significantly affected by the requirement of producing an acceptable
VMS version. We have a preprocessor that generates either a Unix version
or a VMS version (with no "nasty unixisms").

Now of course we could simply put the Unix output of the preprocessor
into the FSF version, but the problem with that is that it would make
it impractical to feed back any modifications on the FSF version into
the master sources, and again would create divergence, which is not
per se a disadvantage to ACT (just to emphasize that it is not ACT's
interests that I am concerned with here), but it would be to the
disavantage of the FSF version (though of course the ACT versions are
also freely available, just at different sites). 

I really want to try to avoid divergence as much as possible. The VMS
issue with the manuals is a difficult one to solve, but we do have
some ideas.