This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: Missing explicit instantiations in libstdc++ (was Re: Unreviewed
- From: Benjamin Kosnik <bkoz at redhat dot com>
- To: John David Anglin <dave at hiauly1 dot hia dot nrc dot ca>
- Cc: mark at codesourcery dot com, jason at redhat dot com, libstdc++ at gcc dot gnu dot org
- Date: Tue, 4 Dec 2001 09:45:21 -0800 (PST)
- Subject: Re: Missing explicit instantiations in libstdc++ (was Re: Unreviewed
> You don't fully explain your reasoning. However, this change seems a bit
> pessimistic. If you want to test implicit instantiation, why not just do
> it on systems with weak linking semantics and provide the explicit needed
> when this isn't available? We already have _GLIBCPP_SUPPORTS_WEAK in
> config.h, so this is trivial to implement.
I leave it to the g++ testsuite to test implicit instantiations. It is
not the intention of the 22_locale/*.cc tests to deal with instantiations
of any type. Instead, they should test that custom facets can be compiled
correctly by the headers.
The 22_locale tests I was thinking of changing anyway, and this just
helped convince me that it was a good idea.
In my check in email to gcc-patches, I said that the locale testsuites
were made consistent: some of the facet tests had compile-only, some
didn't. Now they are all compile only.
Sorry if this was unclear.
I'll look at integrating your patch into what was done last night shortly.
-benjamin