This is the mail archive of the
mailing list for the GCC project.
Re: .weak vs. C++ templates. was: SCO 5.0.2 error
- To: jason at cygnus dot com
- Subject: Re: .weak vs. C++ templates. was: SCO 5.0.2 error
- From: Mike Stump <mrs at windriver dot com>
- Date: Wed, 3 Nov 1999 13:21:55 -0800 (PST)
- Cc: egcs at egcs dot cygnus dot com, robertlipe at usa dot net, wrobbie at my-Deja dot com
> To: Mike Stump <firstname.lastname@example.org>
> From: Jason Merrill <email@example.com>
> Date: 02 Nov 1999 23:38:10 -0800
> The thing is, the change also benefits systems that do support weak
> followed by strong symbols; now all template instantiations will be
> .linkonce, so we'll only get one copy. Before, if there was an
> explicit instantiation, we might have gotten two.
Ah, interesting feature, thanks for the explanation. Could we have
the explicit instantiation in a link once section, with a real
definition (non-weak)? I guess I was thinking that would be possible,
but I've never tried it.
Anyway, this is a very small issue that I don't have a strong
preference for. I defer to you and Mark on template stuff.