Bug in gcse note hacks...

[David S. Miller]

   Date: Sat, 18 Dec 1999 22:04:10 +0100
   From: "Martin v. Loewis" <martin@loewis.home.cs.tu-berlin.de>

   While we are at it: What happened to the requirement of running the
   full test suite [and a bootstrap?] before submitting a patch? I've
   been ignoring it myself at times, but in general, I think this is a
   good thing. While it slows down the individual, it does not slow
   down the group too much.

I agree with you here.

However, sometimes it is not the case that the developer did not
perform a bootstrap/check pass, he in fact did it on the platform
he works under.  And the bugs show up on another platform.

For example, the failures resulting from missing scan_insn calls
when -O0 in some regclass changes did not cause troubles on the i386
target, but they did cause failures on other targets.

Later,
David S. Miller
davem@redhat.com