[Bug c/10170] No warning for "if(foo)" instead of "if(foo())"

[Gabriel Dos Reis]

Neil Booth <neil@daikokuya.co.uk> writes:

| gdr at integrable-solutions dot net wrote:-
| 
| > | it.  I just hope the GCC compiler, if it's possible,
| > | will give a warning message as other compilers do, so
| > | users could see their mistakes.  This warning is the
| > | same warning of :
| > |   if ( a_number = another_number) 
| > 
| > I don't think so.
| > And for that warning, we got existing practice. For 
| > 
| >   if(foo)
| > 
| > we've already got existing pratice the other way around.
| 
| For the case that foo is a function name, which is what this whole PR
| is about, it's very rare.  I don't remember ever having had the need to
| write such a line.

so you're not affected ;-).

More seriously:  That is the advertised way to handle (external weak)
functions defined in dynamically loaded library or such.  
you might find the way it works here (in the first paragraph):

  http://sources.redhat.com/ml/binutils/1999-q3/msg00053.html


http://sources.redhat.com/ml/bug-glibc/2001-10/msg00001.html

-- Gaby