Re: GCC 3.0 Release Criteria

[Jeffrey A Law <law at cygnus dot com>]

  In message <20000427090938X.mitchell@codesourcery.com>you write:
  > >>>>> "Bryan" == Bryan W Headley <bheadley@prismalink.com> writes:
  > 
  >     Bryan> Next thing is: use Linux 2-2-14 as the criteria? Let's get
  > 
  > I'm happy to use a different kernel -- but I think it should be a
  > known stable kernel that can be run in place of the kernels shipped
  > with the supported versions of the RedHat and Debian distributions
  > without too much hassle.
Agreed.  This is similar to the situation we ran into in previous releases.
The way to go is to settle on one or two versions of the kernel, preferably
stable ones.

If we do need to test something from the unstable series, then we need to
pick one and stick with it for a noteworthy amount of time during the
GCC release cycle.  Obviously since we really haven't started the GCC
release cycle, we can't select any particular kernel from the unstable
tree yet.

If linux-2.4 gets released early enough, then we may be able to include it
in addition to 2.2.14 at your discretion.


  > It's hard enough to ask people to actually try running a new kernel on
  > their machines -- that pretty much means they have to not care if the
  > machine gets hosed.  And, I don't want to ask GCC testers to be
  > tracking down kernel bugs at the same time as GCC bugs.  I picked
  > 2.2.14 because it shipped with RedHat 6.2, I believe.
Right.


  > Independently of which kernel version we choose for testing, I would
  > certainly encourage the kernel folks to test their prereleases with
  > development snapshots of current GCC.
Agreed.  I've done this in the past myself and it has paid dividends in
the hunt for compiler bugs ;-)
jeff