Greg Hurrell
Greg Hurrell
@kevincox: would you be able to rebase this against "master"?
Yeah, definitely interested. I just want to confirm your benchmark results on top of current master before going too far.
Thanks, @kevincox. So I had a chance to play around with this. The "big" matcher benchmark is indeed a bit faster, but I found in practice that on a large...
Interesting. Thanks for the analysis. > The benchmark.rb currently benchmarks every prefix of the query Yep, this is pretty intentional because it reflects what I think is a fairly common...
> I'm not sure it is. I very rarely use non-selective queries. Also I suspect the less selective your query the more you are relying on scoring to suface an...
Thanks for this. Can you explain a little bit about how/why this fixes things?
Cool, thanks for the explanation. Before we merge this, I think we need to modify it so that it works in environments that aren't monkey-patched. For example, on my system...
Ha, I take it back. Ruby doesn't work the way I think it should. ``` begin rescue Gem::Ext::RandomGarbage end ``` Works.
Actually, we're going to have to try again. Had to revert this because it was blowing up for me in Ruby 2.3.3-p222 ([commit](https://github.com/wincent/command-t/commit/ce9daf8d792c945d32d0bf8878149eaa8aee2b81)).
Sigh, well that is subtle. I had tested both: ``` begin rescue Gem::Ext::RandomGarbage end ``` And: ``` begin raise 'something' rescue RuntimeError, RandomCrap puts 'rescued' end ``` which also worked......