Re: Your changes in revision 106240

[Paul Eggert]

On 10/30/11 01:07, Stefan Monnier wrote:
> While it's generally true that we try to do that,
> we don't do it at any cost in every single possible case.

Yes, avoiding arbitrary limits is a guideline not an absolute
goal, and other goals can override it.  But in the case we're
talking about, avoiding the 32-bit limit is trivial and costs
nothing, and cost concerns do not support overriding the guideline
in this case.

That being said, the length of this discussion suggests that I may
have touched a nerve, which was not my intent.  I'd rather that we
didn't get bogged down on this during the pretest.  How about the
following idea for moving things forward?  I'll stop fixing
newly-introduced integer-width issues in the trunk, and instead will
fold any such issues into the patch for bug 9874, a patch that
won't be applied until after the next release.  We can discuss
matters further then as needed.