Re: [Opencxx-users] "make check" fail on branch rel_2_9 with gcc 4.0.1 in C++ headers

[Stefan S. <se...@sy...>]

Gilles J. Seguin wrote:

>>3- is
>> const static size_t __align = (
>>       __alignof__(_Tp) >= sizeof(_Block_record)
>>       ? __alignof__(_Tp)
>>       : sizeof(_Block_record)
>>                               );
> 
> 
> we accept static const
>           static
>           const
> but not   const static

Right, that's another point. occ expects declarations to
start with an optional storage-class-specifier, followed
by a type-id and a declarator-list, while the actual C++ grammar
lets declarations start with a decl-specifier-seq, followed by a
declarator-list.

Thus, 'static' may appear (much to my own surprize when I looked
that up in the spec not long ago !) in the middle of the
decl-specifier-seq, as does 'typedef' and 'mutable'.

Fixing this in the occ parser is almost impossible, and it
would impact the parse tree structure. (meaning all the walkers
need to be updated !)

Luckily, not many will write code like

int typedef Int;

void static foo();

but once you get such code from a std library you are in trouble.

Regards,
		Stefan