> If Linus would only test the code to make sure it compiles in different
> configurations, that would catch many of the stupid little "bugs" we
> always see.
And then we'd get about one new kernel release per decade. There are
too many possible configurations to test them all, and you'd need a
warehouse to hold all the hardware.
*WE* are the people who should be testing each new kernel. Yes, 2.0
is supposed to be stable, but I won't consider it so until 2.1 is out.
Since many people seem to ignore the development kernels, I think the
current state of affairs is nearly unavoidable. We could have tested
pre2.0 until blue in the face, but as soon as 2.0 got released all
these 1.2.x sleepers grabbed it and exposed it to many many more
hardware configurations than were available to the 1.3.x developers.
/Of course/ a few new problems came to light.
Once these reports die down, we can consider 2.0.x stable and move on
-- ------------------ Linux hackers do it in protected mode ------------------- | Christopher Biggs - Software Engineer, Stallion Technologies, Australia | | email@example.com - Ph+61-7-3270-4266 - Fx+61-7-3270-4245 - PGP&MIME | ------- Mathematics and alcohol don't mix --- Never drink and derive -------