On Tue, Jun 14, 2016 at 2:48 AM, Jan Nijtmans <jan.nijtm...@gmail.com> wrote:
> > On Mon, Jun 13, 2016 at 1:26 PM, Richard Hipp wrote: > >> So the question becomes: Is your enhancement important enough to > >> delay the release by two months? Or is it better to get 1.35 out > >> Tuesday morning and deal with your enhancement for 1.36?2016-06-13 > 21:33 GMT+02:00 Scott Robison: > On 6/13/16, Scott Robison <sc...@casaderobison.com> wrote: > > Nope, not important enough at all. The current implementation identifies > all > > byte sequences identically to mine. > > I think 1.35 is ready to be released. Actually I like Scott's enhancement > too, a table-based approach can indeed be made faster than custom > byte-level checks, so this is definitely the way to go. The way it > is now (runtime-initialization of a constant table) could be > improved upon further, I'll be happy to have a look at this > after 1.35 is out. > FYI: new commit that hoists the table out of the function and does compile time initialization now. And shrunk the table by 50%. -- Scott Robison
_______________________________________________ fossil-dev mailing list fossil-dev@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/fossil-dev
