Prev: -1%N = -1
Next: communication between classes
From: Oncaphillis on 17 Apr 2008 02:39
Hi all,

Working with g++ 3.* I have to include 3rd
party C-headers which tend to use the same
name for typedefs and
variables

e.g:

<snip>
/* test.h
*/

typedef int bar;

struct foo {
bar bar;
};

</snip>

....which seems to be valid C-code. When I include
these with:

<snip>
extern "C" {
#include <test.h>
};

</snip>

g++ chokes on these identical IDs.

I help myself with including

#ifdef __cplusplus
typedef int bar_t;
struct foo {
bar_t bar;
};

#else
typedef int bar;
struct foo {
bar bar;
};
#endif

...which is cumbersome and ugly.

1. Is this behavior part of the standard or a defect
of GNU C++ ?

2. Is there a more elegant (may be even portable) way
to compile these C-headers without having to change
them.

Thanks you for your attention

O.


--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

From: red floyd on 17 Apr 2008 21:13
Oncaphillis wrote:
> Hi all,
>
> Working with g++ 3.* I have to include 3rd
> party C-headers which tend to use the same
> name for typedefs and
> variables
>
> e.g:
>
> <snip>
> /* test.h
> */
>
> typedef int bar;
>
> struct foo {
> bar bar;
> };
>

Find the guy who wrote that code, and shoot him.

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

From: Marco Manfredini on 18 Apr 2008 08:15
Oncaphillis wrote:

[does not work:]
> typedef int bar;
>
> struct foo {
> bar bar;
> };
> 1. Is this behavior part of the standard or a defect

I think this is standard behavior. There is a rule for class-scope
declarations in 3.3.6 which basically says, that a declaration is only
valid, if it yields the same declaration after seeing all of the class.
This prohibits redefining a symbol after using it.

>
> 2. Is there a more elegant (may be even portable) way
> to compile these C-headers without having to change
> them.

Not really. It might be a bit more elegant not to duplicate everything
in the headers, but just to wrap all occurrences of a typename in a
macro:
#ifdef __cplusplus__
#define TYPESYM(X) X##_t
#else
#define TYPESYM(X)
#endif

typedef int TYPESYM(bar);
struct TYPESYM(foo) {
TYPESYM(bar) bar;
};



--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

From: Oncaphillis on 18 Apr 2008 08:20
red floyd wrote:

> Oncaphillis wrote:
>> Hi all,
>>
>> Working with g++ 3.* I have to include 3rd
>> party C-headers which tend to use the same
>> name for typedefs and
>> variables
>>
>> e.g:
>>
>> <snip>
>> /* test.h
>> */
>>
>> typedef int bar;
>>
>> struct foo {
>> bar bar;
>> };
>>
>
> Find the guy who wrote that code, and shoot him.
>

Ok.. that was my plan B. But since Mom and Dad always
told me that it is something you shouldn't do I tried
to find something else.

So is this kind of typdef and variable name mix up
forbidden by C++ or is it just g++ that barks at
me ?

Cheers

O.


--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

 | 
Pages: 1
Prev: -1%N = -1
Next: communication between classes