In some email I received from Andy Doran, sie wrote:
> On Tue, 18 Jan 100, Darren Reed wrote:
> 
> > In some email I received from Luke Mewburn, sie wrote:
> > [...]
> > > /* Make the structure name match the typedef. */
> > > typedef struct _bar {
> > > 	int	level;
> > > } BAR;
> > [...]
> > 
> > I don't particularly like that.  My preference, personally, is to use "_t"
> > to represent typedefs thus:
> > 
> > typedef struct _bar {
> > 	int	level;
> > } bar_t;
> > 
> > At least this way it is distinct from macros and enum's.
> 
> "_t" is reserved by POSIX - we should try to restrict its use.

What's the nature of this reservation ?

For example, we have "uint32_t", "pid_t", etc.

Are these all POSIX types ?

I use "_t" extensively in code I write to indicate typedef's and no
compiler barfs.  However, does that put the namespace I use in danger
of one day colliding with official POSIX types ?

Darren