[Bug target/49826] [4.7/4.8 Regression] Symbols are not decorated with attribute stdcall and -mrtd

[ktietz at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49826

Kai Tietz <ktietz at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P2                          |P4

--- Comment #8 from Kai Tietz <ktietz at gcc dot gnu.org> 2013-01-15 17:54:02 UTC ---
(In reply to comment #4)
> (In reply to comment #3)
> 
> The linker does not work very well with undecorated stdcall-s. For example:
> <http://sourceware.org/PR13495>. But if that bug will be fixed, then this one
> can be closed.

Hmm, can't reproduce what your issue is here actual.  Nevertheless seems to me
that patch plain wrong.  For -mrtd the default-calling convention is stdcall,
so to mark a symbol explicit with an stdcall-attribute is somewhat a
nop-expression.

I understand that you would like to have an extension that gcc is able to use
for user-attributed stdcall functions instead the @<n> decoration.

The suggested patch looks wrong, as it doesn't check if symbol has
user-attribute on symbol.  Nevertheless this bug is IMHO an extension and not a
bug.