This is the mail archive of the
mailing list for the GCC project.
Re: Ping for build-with-cxx patches
- From: Mark Mitchell <mark at codesourcery dot com>
- To: Ian Lance Taylor <iant at google dot com>
- Cc: gcc-patches at gcc dot gnu dot org, jason at redhat dot com
- Date: Sun, 28 Jun 2009 14:43:53 -0700
- Subject: Re: Ping for build-with-cxx patches
- References: <firstname.lastname@example.org>
Ian Lance Taylor wrote:
> This is a ping for some C++ frontend patches. I believe that these are
> the only remaining patches required for --enable-build-with-cxx to work.
> * Add --static-libstdc++ option to the g++ driver.
> * Avoid inappropriate -Wsign-compare warnings.
+/* Even unsigned enum types promote to signed int. We don't want to
This comment is a bit inaccurate. I believe that unsigned enum types
promote to signed int only if the range of the enumerators is
sufficiently small that a signed int can hold all of them, right? If
so, please clarify the comment; perhaps just by saying "Even some
unsigned enum ...".
Also, we will still get warnings about "-1 < E" (where "E" is an
enumerator) being always true if we set TREE_NO_WARNING here?
> * Avoid incorrect "no effect" warnings.
OK, but with a comment before the setting of "side_effects". Perhaps:
/* Macros like:
#define ABORT_IF(X) (X) ? abort() : (void)0;
are common. In this case, only one of the conditionally-executed
expressions has side effects. To avoid spurious warnings, we only warn
if neither conditionally-executed expression has side effects. */
> * Don't warn about a logical op RHS which is never executed.
(650) 331-3385 x713