This is the mail archive of the
mailing list for the GCC project.
Re: Ping: Patch for 21562
On May 26, 2005, at 8:18 PM, Dale Johannesen wrote:
On May 26, 2005, at 5:08 PM, Jeffrey A Law wrote:
On Mon, 2005-05-23 at 10:08 -0700, Dale Johannesen wrote:
It really seems to me that the problem is verify_flow and expand
really need to be fixed. Any code that assumes block 0 is a
successor of ENTRY is broken.
Is there some fundamental reason why we can't fix verify_flow and
Not that I know of. Is there consensus that this is the right thing
On IRC, RTH said it would be hard not to have the successors of ENTRY
be BB zero in RTH. Jan said was going to get a patch which changes
cfgexpand so that we always get BB 0 as the successor of ENTRY.