Add benchmark for ctags conversion

[jtibshirani]

This change adds a benchmark for the conversion from ctags output to Zoekt
document data, plus a tiny optimization to presize the symbol slices.

[jtibshirani]

I was poking around in this code in case I spotted obvious memory inefficiencies or leaks. I didn't, but figured it would be helpful to commit this.

Benchmark output before vs. after the optimization:

BenchmarkTagsToSections-10    	    7364	    156615 ns/op	   87472 B/op	     461 allocs/op
BenchmarkTagsToSections-10    	    7494	    155150 ns/op	   79296 B/op	     443 allocs/op

[keegancsmith]

Nice! I took a quick look at the implementation and perf and spotted these improvements :) #680

[keegancsmith]

Benchmark output before vs. after the optimization:

FYI benchstat is a great tool for this. My workflow for it something like this (note the importance of the count flag):

$ go test -run ^$ -bench BenchmarkTagsToSections -count 10 > before.txt
$ # hack away at improvements
$ go test -run ^$ -bench BenchmarkTagsToSections -count 10 > after.txt
$ benchstat before.txt after.txt