This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Question about __attribute__ (mis)handling


At 11:13 AM -0800 12/13/01, Aldy Hernandez wrote:

>  > There is no warning.  That test case I showed illustrates it is not
>>  even being handled.  What I believe is that there is a missing call
>  > to decl_attributes() for inline members.
>
>put a break on decl_attributes().  if it never gets called, fix it :)

I did.  It doesn't :-)

Here's a reduced test case for verifying this:

struct T2 {
   void __attribute__ ((this_should_also_be_an_warning)) t2_member1() {}
   void t2_member2();
};

If you put a stop in the for-loop in decl_attributes(), i.e., where 
it knows it has a non-null attributes list you will see it stops only 
for __builtin_printf and __builtin_fprintf for a format attribute.

Fix it?  I was hoping you would :-)

Ira


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]