This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: Patch: stl_vector.h and vector.tcc
- From: Matt Austern <austern at apple dot com>
- To: Benjamin Kosnik <bkoz at redhat dot com>
- Cc: libstdc++ at gcc dot gnu dot org, pcarlini at suse dot de, dhruvbird at gmx dot net
- Date: Sun, 1 Aug 2004 16:58:10 -0700
- Subject: Re: Patch: stl_vector.h and vector.tcc
- References: <1089638663.3669.40.camel@localhost.localdomain> <20040730164534.6c124dfe.bkoz@redhat.com> <7D10D886-E273-11D8-B57C-000A95BCF344@apple.com> <20040730225438.7bebbc99.bkoz@redhat.com>
On Jul 30, 2004, at 8:54 PM, Benjamin Kosnik wrote:
I agree, implementing std::tr1::type_traits woulid be a really great
idea.
Other parts too. If we just set up the infrastructure, i.e. the
directory
hierarchy and makefile, then various of us could implement individual
pieces. I'd like to put in hash tables, for example.
Sounds good. I can set up include/tr1 if you'd like, if you have better
ideas let me know. For things that are just include files that should
be
sufficient.
Hm. Hash tables are actually one of the cases where I think we do
need some .cc files. I'd like to have a long list of primes, and
that's a silly thing to put in headers. So I think we should plan
for a source directly immediately.
One other things we should plan for: some of the tr1 extensions
involve new stuff in standard headers. I think the right way to
do this is to have (for example) include/tr1/utility, which would
include the standard <utility> and then add more things. But that
means we need to figure out the mechanics of that inclusion.
Finally: in the long run we probably should have a place for tr1
documentation. Even in the short run, though, we need a place for
tr1 test cases.
-Matt