[Freeciv-Dev] Re: [FreeCiv-Cvs] per: Fix a division by zero that happens if we reque...

["Per I. Mathisen" <per@xxxxxxxxxxx>]

On Fri, 11 Oct 2002, Raimar Falke wrote:
> > Not sure if you suggested that, but I'm totally opposed to adding this
> > patch before release. Adding asserts to a problem that is not properly
> > understood is arecipe for bugs. My fix has the advantage of being
> > guaranteed not to introduce new bugs, yet avoiding the core dump.
>
> Just adding these asserts doesn't introduce bugs. The problem is that
> the issue is rare. So we can only catch it if we:
>  - add some something which can catch it (if only your patch is
> applied this isn't the case) AND
>  - we expose the changes to a lot of games/people (before a release is
> the best time).

> > After the release, however, I'm for it.
>
> Then we will forget and your band-aid will stay there eternal. we both
> know the laziness involved in such issues.

No. I categorically refuse to add patches that _increase_ the chance of
coredumps happening just before we make a new release just because we
_may_ forget to commit the patch later. Or because we want players to core
their games so that we can have the pleasure of finding perfect fixes for
imperfect code.

Send it to jitterbug. If we forget good patches that are registered in
jitterbug, we have bigger problems. And we don't solve these by making
buggy releases.

  - Per