This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ 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: More test results...



> Right but why were there implicit instantiations being generated?  If there
> hadn't been, the problem wouldn't have been noticed.

The original code wasn't posted, but it was likely to not be using 
specializations, and thus implicitly generating instantiations of 
numeric_limts<string> from the generic numeric_limits template definition.

The details are archived in the thread, I believe.

Anyway...... on to another, separate issue.

> > If you've actually fixed the static const issue, then perhaps the stl 
> > changes to enums can be reverted. Thoughts?
> 
> I haven't been following that discussion, sorry.

It's the whole

%cat > 1.cc
const int i = 5; 

%nm 1.o 
00000000 r i

thing

You've said in private email that:

> With -O or -fno-keep-static-consts, it goes away.  Perhaps C++ should
> default to -fno-keep-static-consts.

...which I can verify does what I'd expect. Any hope to make this the 
default?

thanks,
benjamin


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