Kevin Cox

If you are interested in merging I gladly will. However it is Dependant on the C Scanner Infrastructure so that will have to be merged first. If that sounds good...

Well I'll try to start porting to current master. Also I realized the reason I wasn't seeing much parallel speedup was because the benchmarks had no limit. Adding a limit...

The rebase was simpler then expected. This is now based on master.

That's odd. I was trying it on a set of ~4M files (my real world use case) and it was significantly faster. Can you explain exactly what your setup is....

Hmm, it actually appears like there was a regression in the cleanup. I'll look into it. However even still the performance is better then master for me with ~2M paths...

Sorry for the delay, life gets in the way. I looked into it and it appears that the reason the benchmarks look very similar is because this new branch is...

>> The benchmark.rb currently benchmarks every prefix of the query > Yep, this is pretty intentional... I figured it was, and I knew about it but it slipped my mind....

> I found in practice that on a large repo (1.5m files) this is significantly slower overall I wonder what makes your large repo test slow. Because I tested with...

I'll add a little checklist for planned improvements: - [ ] Cache results of previous search. This will allow faster pruning when search is strictly more selective. Restricts to a...

> I wouldn't be opposed. I love speed. Most of the cost, however, is probably not in the scanner itself but in the calls to Vim's expand() function, and that...