This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
PR 2828: Following up
- To: gcc-patches at gcc dot gnu dot org
- Subject: PR 2828: Following up
- From: Mark Mitchell <mark at codesourcery dot com>
- Date: Fri, 18 May 2001 17:29:26 -0700
- Organization: CodeSourcery, LLC
Well, the ENABLE_CHECKING assertion didn't work, even on a i386
bootstrap. It seems that we sometimes call this function with bogus
GLOBAL_LIVE_AT_START information in the first place, and are hoping it
will get updated correctly here. I can't fully convince myself that
this approach is necessarily guaranteed to terminate, but I can't
construct a counter-example off the top of my head. One wants to
prove that GLOBAL_LIVE_AT_START will only be shrunk finitely many
times.
Anyhow, more practically, please test the patch without the
ENABLE_CHECKING hunk. (On the GCC 3.0 branch, ENABLE_CHECKING is
turned off, so it won't matter.) The rest of the patch should still
be OK.
--
Mark Mitchell mark@codesourcery.com
CodeSourcery, LLC http://www.codesourcery.com