This is the mail archive of the 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: [Patch: mingw/cygwin] Don't strip '@' dangles from stdcall/fastcall dllimport references

On 13 July 2007 11:21, Kai Tietz wrote:

>> const char *
>> i386_pe_strip_name_encoding_full (const char *str)
>> {
>>   const char *p;
>> -  const char *name = default_strip_name_encoding (str);
>> +  const char *name;
>> +
>> +  if (!TARGET_64BIT)
>> +    return name;
>> +
>> +  name = default_strip_name_encoding (str);
>>   /* Strip leading '@' on fastcall symbols.  */
>>   if (*name == '@')
>>     name++;
> AFAICS this seems to be even better, because it avoids diffent
> configuration definitions.

  Runtime tests like this are very much the preferred style to #ifdefs :)

> I am on to add function attribute dependent
> amd64 abi usage 

  Yes, I've been following the thread.

> and there are target specific macros a bit disturbing ;)

  It might be best to tidy up some of these from compile-time tests to runtime tests with a set of preparatory patches to clear the ground nicely before generating the patches that make the ABI switchable by attributes.

Can't think of a witty .sigline today....

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