Re: problem with gnustep on OpenBSD sparc64

[Wolfgang Lux]

Sebastian Reitenbach wrote:

> I extended the output a bit, to show the address used for size  
> (a.k.a. context in parse_array)
>
> $ ./a.out
> sizeof_type: before calling parse_array: element_size: 0, size  
> pointern: 0xfffffffffffcf260
> parse_array: type: [4i]], context a.k.a. size: 0xfffffffffffcf260
> sizeof_type: before calling parse_array: element_size: 0, size  
> pointern: 0xfffffffffffcf0d0
> parse_array: type: i]], context a.k.a. size: 0xfffffffffffcf0d0
> sizeof_type: sizeof(typeName): 4 size pointer: 0xfffffffffffcf0d0
> sizeof_type: size: 32, size pointer: 0xfffffffffffcf0d0
> parse_array: type: ]]
> sizeof_type: after calling parse_array: element_size: 0,  
> element_count: 4, size pointern: 0xfffffffffffcf0d0
> ^^^ as far as I can see, the value on address 0xfffffffffffcf0d0  
> should be 32, but I have no idea, why it is not 32???
> 198                             int element_size = 0;
                                 ^^^^^
The problem seems to be here ...
> (gdb)
> 200     fprintf(stderr, "sizeof_type: before calling parse_array:  
> element_size: %i, size pointern: %p\n", element_size, &element_size);
> (gdb)
> sizeof_type: before calling parse_array: element_size: 0, size  
> pointern: 0xfffffffffffe0340
> 201                             int element_count = parse_array(&t,  
> (type_parser)sizeof_type, &element_size);
                                                                                               ^ 
^^^^^^^^^^^^^^
... and here. The sizeof_type callback expects an argument with type  
size_t *, but &element_size has type int *. This does work on 32-bit  
architectures where sizeof(size_t) = sizeof(int) and it works  
coincidentally on low-endian 64-bit architectures like x86_64, but is  
definitely wrong for big-endian 64-bit machines like sparc.

In order to fix, change line 198 to
  size_t element_size = 0;

Wolfgang