This is the mail archive of the
mailing list for the GCC project.
Re: Patch for PR 14635 (regression: nan not in C90)
- From: Roger Sayle <roger at eyesopen dot com>
- To: "Joseph S. Myers" <jsm at polyomino dot org dot uk>
- Cc: Richard Henderson <rth at redhat dot com>, <gcc-patches at gcc dot gnu dot org>
- Date: Fri, 19 Mar 2004 06:34:46 -0700 (MST)
- Subject: Re: Patch for PR 14635 (regression: nan not in C90)
On Fri, 19 Mar 2004, Joseph S. Myers wrote:
> But this is another simple regression bug. I've now reported it as
> middle-end/14649. Evaluated function calls are not permitted in constant
> expressions (even as an extension). There should only ever be exceptions
> for those __builtin_ cases specifically documented as such, such as
> __builtin_constant_p and __builtin_nan.
This is not a middle-end bug! There are some languages that allow
function calls in global variable initializers, and permit their
optimization by evaluating them at run-time. C++ for example.
Please change the bugzilla PR to c/14649.