Re: another small thing

[Gabriel Dos Reis <gdr at codesourcery dot com>]

Phil Edwards <pedwards@disaster.jaj.com> writes:

| > | but you knew that already.  As far as I've ever been able to tell, local
| > | names don't /really/ have to be uglified,
| >
| > I don't think that is right.
| >
| > Assume we use `t' instead of `__t'; how do you expect this to work
| >
| > 	#define t 89070943
| > 	#include <standard_header>
| 
| I wouldn't expect that to work; here 't' is a global name.
| 
| By "local names" I believed he was referring to this situation:
| 
|     void some_visible_fn (int __name1)
|     {
|         int name2;
|         ...
|     }
| 
| Here, name2 shouldn't strictly be needed to be uglified, since it can't
| collide with names in user land.

I'm still not convinced because __name1 and name2 are subject to the
same evil: there are prepocessed by CPP before the parser sees
anything at all.

-- Gaby
CodeSourcery, LLC                             http://www.codesourcery.com