Re: [RFC patch]: Adjust the use of 'long' type in dwarf2.h header

["H.J. Lu" <hjl dot tools at gmail dot com>]

On Thu, Feb 24, 2011 at 3:33 AM, Pierre Muller
<pierre.muller@ics-cnrs.unistra.fr> wrote:
>> #if __GNUC__ >= 2
>> /* Define BFD64 here, even if our default architecture is 32 bit ELF
>> ? ?as this will allow us to read in and parse 64bit and 32bit ELF
>> files.
>> ? ?Only do this if we believe that the compiler can support a 64 bit
>> ? ?data type. ?For now we only rely on GCC being able to do this. ?*/
>> #define BFD64
>> #endif
>>
>> ....
>> #include "dwarf.h"
>> ---
>>
>> dwarf_vma should be the same inside and outside of readelf.c.
>
>
> ?This means, if I understand this correctly that
> if I compile a 32 bit readelf executable
> bfd_vma will be a 64-bit integer inside readelf.c,
> but a 32-bit integer in dwarf.c.
>
> ?dwarf.c still mainly uses bfd_vma
> and has dwarf_vma only for one function
> get_encoded_value...
>
> ?Should read_leb128 (and maybe other functions within
> dwarf.c) also be changed to use
> dwarf_vma rather than bfd_vma in that case?
>

Probably since readelf.c has

static unsigned long
read_uleb128 (unsigned char *data, unsigned int *length_return)
{
  return read_leb128 (data, length_return, 0);
}



-- 
H.J.