Re: [gcc-in-cxx] Fix top-level configury

[Ian Lance Taylor <iant at google dot com>]

Eric Botcazou <ebotcazou@adacore.com> writes:

>> It also comments out some code which added
>> -fkeep-inlin-functions to stage 1.  When building with C++ this breaks
>> the inline functions which are provided by <gmp.h>, as they wind up
>> getting included in many .o files leading to duplicate definition
>> errors at link time.
>
> This was added precisely because of that, i.e. because non-GCC compilers tend 
> not to eliminate inline functions when not optimizing.  So, if you plan to 
> support other C++ compilers then G++, that's a step backwards.

I don't think so.  The C++ inline semantics are like the C99
semantics: a function which is inline, but neither static nor extern,
must not be included in the output file.  You must be talking about
C90 compilers which implement inline as a hint, and keep the function
if it is not inlined everywhere--e.g., the C++ or C99 extern inline
semantics.  Passing -fkeep-inline-functions to g++ changes the inline
semantics to keep inline functions even when C++ rules say they must
be discarded.  This is rarely correct, and it legitimately breaks when
using <gmp.h> which expects the normal C++ inline semantics.  So I
think my patch is correct when building with C++.

Ian