On Sat, 17 Oct 2009 20:45:39 -0700
Philip Guenther <[email protected]> wrote:

Many solid points and:

>If the answers to those questions are "yes", then you should say so
>and back it up.  If not, why should we apply a diff to remove code
>from someone who doesn't know why the code is there?  As is, I believe
>you don't understand why that code was added, don't understand when it
>was triggered, and don't know whether or not it could still be
>triggered.

Although I'd rather crawl under a rock, I'll respond.

You're correct. I got started on this train of thought after reading
an old USENIX paper[1]. The line I suggested for deletion is identical
to what section 1.2.1 of that paper recommended should always be
included. Recent material continues to recommend such code, but not
specifically in OpenBSD[2],[3].

Since such code is only rarely used in OpenBSD -- 9 occurrences in the
whole tree -- I thought it might be an obsolete or inapplicable
recommendation. One such occurrence was in m4/main.c, so I contacted the
developer who had recently checked in that file.

I did read manuals, study code, consider, and test first. But I knew my
expertise was limited, and that there could be other considerations.
I never said imperatively "delete this". I asked Stuart if it was a
valid change. To you, I said it was "possible cruft".

It's good that you are circumspect. Frankly, I was expecting the answer
to be known. But if it's doubtful, the change has to be rejected.

[1] http://www.darwinsys.com/history/canthappen.html
[2] http://oreilly.com/catalog/9780596009588
[3] 
http://www.cs.princeton.edu/courses/archive/spr06/cos217/lectures/23signals.pdf
-- 
Matt Fisher <[email protected]>

Reply via email to