In message <9485.958198965@coconut.itojun.org>itojun@iijlab.net writes
>
>	It is also true that a developer cannot test code before commit
>	for every possible configuration options.  I believed this is the
>	reason why there's netbsd-current and releaes (1.4.x for example)
>	branch.

Testing IPv6 changes without "options ipv6" is hardly every possible
configuration option, now is it?


>	Though I agree I was not perfect, I feel recent finger-pointing
>	(rather than careful bug reports with enough info to repeat)
>	makes me a little bit unhappy.

And not doing adequate testing before committing changes seems to be
making other people unhappy.  Perhaps all developers working on IPv6
could test their IPv6 changes with*out* "options IPV6", before
committing changes, as part of their normal test cycle?

that way there would be (almost) nothing for fingers to point at.