lukem@wasabisystems.com (Luke Mewburn) writes:
> > Hmm. Personally, I for one would rather see this feature go away
> > and we should use existing mechanisms. 

as would I.  it's broken.

There's the issue of _ and .

There's also the issue of expected behaviour of you say:

	include FOOBAR

in FOOBAR.DEBUG.

All of a sudden, the code paths which used to apply to FOOBAR no
longer apply to the kernel you just built -- which from the
configuration should be logically the same.

Unless i'm misremembering, _old_ config used to -D the kernel name,
and that wasn't done in new config.  Instead, it could be done with
options.


> And as per usual in these situations, a couple of ideas are put up, no
> one comments one way or another about one of the points (and get off
> topic on the other point), you wait a couple of days and commit the
> change (modulo some minor feedback), and then every man and his dog
> jumps up [...]

(1) A couple of days are adequate time to discuss random technical
issues?  If you want to contribute to the kernel discussion in NetBSD,
does that mean that you must stay within 2 days of current with
tech-kern, and not take any time to consider the issues?

(2) It's not clear that a couple of days is adequate time to consider
many issues.  e.g. the latter issue mentioned above only occurred to
me in the shower this morning -- and i'd read the original post when
it was posted.

(3) if, even after a change is committed, it's seen to be logically
and/or technically flawed, as i believe this change is, it's not at
all clear that it should be left in the source tree.


Finally, your statement above about waiting a couple of days appears
to be simple rubbish in this case.

It looks like you posted your original message at:

  by mail.netbsd.org with SMTP; 11 Dec 2000 19:28:20 -0000

Right now it is:

	Wed Dec 13 17:59:57 GMT 2000

so your "couple of days" of comments should still be open, even if
that were all you were going to allow.  (I don't see an obvious thread
which earlier presented the suggestion for the change, but I'll admit
i didn't search the mail archives very hard.

And the change was committed at:

date: 2000/12/12 17:49:20;  author: lukem;  state: Exp; lines: +2 -2
change KERNEL_foo -> _KERNEL_foo, as suggested by matt green

(that time is GMT, AFAIK), i.e. less than a day after it was proposed.


So, about this change, please don't get on your high-horse about how
people dare to comment on changes after they've been committed after
"adequate" consideration time.

You, as a member of core, should do better.

Further, as far as i'm concerned, you should go out of your way to
_encourage_ more discussion, and take action in the case where bogus
changes are made _without_ adequate discussion.

(Things that come to mind recently have been the line discipline
changes, and the watchdog interface.  The latter, especially, annoys
me, because (1) as far as I can tell, it simply wasn't discussed; (2)
as you'll see if you read the ex post facto discussion here, from
people who've been playing with watchdogs for a while, it's simply not
sufficient.)

You, as someone in a position of authority in the project, should, as
far as i'm concerned, be using your participation to forward the goals
of the project.  That should be to foster discussion, to encourage
technically correct, well thought out changes, etc.  That's not what
your message did; it basically tried to slap down discussion that you
agreed with about an ill-conceived, inadequately-discussed change to
the source tree that you happen to be fairly fond of.

Bad form.



cgd