[fortran, docs] Add embryonic gfc-internals.texi manual.

[Brooks Moses]

Toon Moene wrote:
> Tobias Schlüter wrote:
>> How about putting this into the wiki?  I don't think it makes much sense 
>> to distribute this documentation to the user, nor does it make much 
>> sense to have different versions for different future versions / 
>> branches.  Also the entry barrier with .texi is higher (well, IMO).

One of the reasons that I was wanting to do this as a .texi document is 
that for me personally the entry barrier is lower -- especially when I'm 
working without a net connection, which is often the case.  Having a 
scratch copy in which I can write guesses that might be entirely wrong 
and then edit and correct before pushing them "live" works better for 
me.  But that's all personal preference, and I know it's not universal.

Another advantange of putting this in SVN in some form is that 
(regardless of how the approval is done) it goes through the usual patch 
submission process, so all the changes are posted to the mailing list, 
which I hope will make it easy for people to give them a quick look to 
make sure they look correct (and there's an automatic forum for asking 
questions about anything that's unclear, too).

Having things posted to the mailing list also means that I have more of 
an opportunity to learn from new content.  :)

> Why not considering it an equivalent for GNU Fortran as 
> gcc/doc/gccint.texi is for GCC ?

That was somewhat my thought, yes.  (Admittedly, gccint.texi predates 
Wikis by a fair bit, so when it was written a Wiki wasn't an option.)

> And as a corollary - why not use the 
> standard approval rules for it ?

My original thought for not using the standard approval rules was that 
it would, as a temporary measure, lower the entry barriers for adding 
content.  However, I agree with FX's point that it's probably better to 
address that that for non-maintainers with blanket permissions on an 
individual basis, and it's a moot point for maintainers, so I'll 
withdraw the proposal for using non-standard rules.

- Brooks