On 9/25/2012 17:10, Pierre Abbat wrote:
On Tuesday, September 25, 2012 14:34:20 John Marino wrote:
Unknown at this time, but there was significant regression in packages
after moving to gcc 4.4. I don't expect much regression with 4.7
because NetBSD uses 4.5 and SunOS and Linux use 4.6 and even 4.7 now so
others are "blazing the trail" with respect to fixing GCC upgrade
regressions.
This doesn't concern most users because the default compiler will be 4.4
until 4.7 is fully tested. That will take some months and many package
builds.
OK - what about the bug that's been keeping the 32-bit builds from completing?
Pierre
I can only guess that you are referring to the wirecount bug.
That topic has nothing to do with this one.
During my builds a few weeks ago, I no longer get wirecount panics but
it's been replaced by a much less frequent "page not busy.4" panic:
http://bugs.dragonflybsd.org/issues/2296
i386 is much more usable under load now, but there's still a bug or two
loose in there.
