[patch]: fastcall/stdcall and import libraries in ld

[Kai Tietz]

Hi Dave,

2009/11/13 Dave Korn <dave.korn.cygwin@googlemail.com>:
> Kai Tietz wrote:
>
>> This patch makes the described in documentation true, that stdcall
>> functions can be fixuped. It additional implements that fastcall
>> functions can be fixuped, too. Also it enables that functions from
>> import libraries, which have no decoration, can match to
>> fastcall/stdcall functions, when --enable-stdcall-fixup is specified
>> without warnings.
>
>  Looks pretty good, but could really use some testcases, nag nag!

Well, I will sent one for this scenario in a separate patch :)

>> +      && ((*hs == '@' && *string == '_'
>> +        && strncmp (hs + 1, string + 1, sl - 1) == 0)
>> +               || strncmp (hs, string, sl) == 0)
>
>
>  Mixed spaces/TABs at the start of that third line there.  Same in both the
> emulation scripts and the pe-dll.c file.  (BTW, why not factor this out?
> pe_undef_cdecl_match and pe_undef_alias_cdecl_match are almost identical
> except for the global variable they return the found symbol in, the code to do
> the test is a bit hairy, it would sit well as an inline function or macro if
> you could find some sensible header to put it in.)

Well, I thought about this, too. But the pe_undef_alias_cdecl_match is
searching for undefined symbol, but the pe_undef_cdecl_match searches
for defined ones.

>> @@ -1212,6 +1220,9 @@
>>
>>    pe_find_data_imports ();
>>
>> +  if (pe_enable_stdcall_fixup) /* -1=warn or 1=disable */
>> +    pe_fixup_stdcalls ();
>> +
>
>  I think you should add a comment here explaining the rationale for having a
> second pass over the stdcall fixups here.
Ok, I will add.

>> Tested for i686-pc-mingw32, x86_64-pc-mingw32, and i686-pc-cygwin. Ok for apply?
>
>  Builds and tests good here too.  OK by me after those minor tweaks, thank you.
>
>    cheers,
>      DaveK
>
>

Cheers,
Kai

-- 
|  (\_/) This is Bunny. Copy and paste
| (='.'=) Bunny into your signature to help
| (")_(") him gain world domination