This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [C++ PATCH] Fix PR/10750 (a regression) - take 2
- From: Gabriel Dos Reis <gdr at integrable-solutions dot net>
- To: Mark Mitchell <mark at codesourcery dot com>
- Cc: Giovanni Bajo <giovannibajo at libero dot it>, gcc-patches at gcc dot gnu dot org, Jason Merrill <jason at redhat dot com>
- Date: 20 Jun 2003 02:38:55 +0200
- Subject: Re: [C++ PATCH] Fix PR/10750 (a regression) - take 2
- Organization: Integrable Solutions
- References: <165a01c3369e$0172cfa0$114e2697@bagio><1056065945.6379.17.camel@minax.codesourcery.com>
Mark Mitchell <mark@codesourcery.com> writes:
| On Thu, 2003-06-19 at 13:04, Giovanni Bajo wrote:
| > Hello,
| >
| > as a follow-up to my "propagate TREE_CONSTANT through CAST_EXPRs" and the
| > replies from Jason and Gaby
| > (http://gcc.gnu.org/ml/gcc-patches/2003-06/msg02107.html), I think I finally
| > sorted out what it needed to be done.
|
| I think I disagree with Jason. :-(
And I will disagree with youd :-)
Consider
template<class Q>
struct X {
typedef typename Q::T T;
static const T a = Q::a;
};
In processing the above template, we can't check that Q::a is a valid
initializer nor that Q::T is a valid type. All that processing needs
to wiat instantiation time.
:
| I liked your earlier patch, and was about to approve it.
|
| Most of the places where we check "processing_template_decl" are
| actually wrong; we should be checking dependency.
|
| So, I'll OK the original patch.
Please hold on until we sort out the issue.
-- Gaby