Re: Small update to reversed_comparison_code

[Geoff Keating <geoffk at geoffk dot org>]

kenner@vlsi1.ultra.nyu.edu (Richard Kenner) writes:

>     Forfcing the issue with a build breakage is tecnical extortion, and
>     something that worries people a lot (including Mark and me!)
> 
> I do understand the concern, but GCC is a volunteer project, so motivation to
> fix things is something that's not necessarily bad.  Sure, if this were a
> commercial project with normal industrial management, this would be a very
> bad thing, but the role of that management is precisely to make sure that
> things like this don't fall on the floor.  In an unmanaged project there's
> nobody to do that, so this sort of "extortion" has to play that role.

The question is whether this particular problem is more important than
the other things that people could be working on.  For instance,
having to fix this bug on an emergency basis probably delayed the GCC
3.0 release by a week or so, and arguably it would have been better to
get the release out first and make the fix in 3.0.1.

The worst thing is not this (because the decision on which was the
higher priority could have gone either way), it is that the way you
are proposing that such decisions be made is not a good way; the
effect of your proposal is that such decisions be made by an
individual acting alone in checking a patch into the compiler.

-- 
- Geoffrey Keating <geoffk@geoffk.org>