Re: reducing the footprint of ScanKeyword (was Re: Large writable variables)

On Wed, Jan 9, 2019 at 5:33 PM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> really need here. We could go with "[no-]case-insensitive", perhaps.
> Or "[no-]case-fold", which is at least a little shorter and less
> double-negative-y.

I'd be in favor of --[no-]case-fold.

On Tue, Jan 8, 2019 at 5:53 PM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> I improved the comment about how come the hash table entry assignment
> works.

I've gone over the algorithm in more detail and I don't see any nicer
way to write it. This comment in PerfectHash.pm:

(It does change '_', else we could just skip adjusting
# $cn here at all, for typical keyword strings.)

...seems a bit out of place in the module, because of its reference to
keywords, of interest right now to its only caller. Maybe a bit of
context here. (I also don't quite understand why we could
hypothetically skip the adjustment.)

Lastly, the keyword headers still have a dire warning about ASCII
order and binary search. Those could be softened to match the comment
in gen_keywordlist.pl.

--
John Naylor https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services