🚀 go-pugleaf

In article <43kb6s$ntr@news0.rain.rg.net> Paul Long <plong@perf.com> writes:

> kanze@gabi-soft.fr (J. Kanze) wrote:
> >I *think* that the rationale for disallowing defining new types in
> >function prototypes is that there is no way to use such types.
> 
> Right, that's my point.  Other than within the prototype, the new type
> cannot be referrenced. It's practically useless.

I have run into a situation where this construct was necessary.  I was
compiling rpcgen(from Sun's RPC) protocol descriptions into C code
that would be called by Visual Basic.  Because of the way VB works, I
needed to add several new types for each type declared in the protocol
files.  Since I didn't know how the user was naming his types, and
because I was constrained to providing at least a type of the same
name as in the protocol file, I had to define the new types in the
function prototypes.  I also generated VB "prototypes" for the new
functions which passed VB types which were declared outside of the new
functions.  I could do this because I didn't have to pass names
through literally in the VB code.

So this feature does occasionally come in useful.  Whether the
standard ought to cater to this admittedly strange and
system-dependent situation is another matter.

--glenn