Re: hash, xhash: modernize

[Bruno Haible]

Hi Jim,

> > I tried to make it clear the last time we discussed this (long ago!)
> > that I prefer to keep certain comments very near the function
> > definition (and implementation).

Yes, I sort of remembered this. Therefore I asked for objections, and put
you in CC.

> > I disagree with the premise that hash_delete should be renamed. That's
> > an API-breaking change.

Yes, it's an API change. I applied the usual procedure for API changes in
gnulib: add the new API, then wait for more than a year, before the old API
can be removed.

> That said, I will not object to your normalizing diffs.

I'm confused now. Which of the three patches, or which parts of them,
do you wish to see reverted?

Bruno