Okay, I tested it with the Java test case, and well, it's still unusably slow when I duplicate it 100 times, but it's indeed incredibly faster than it was before :) (e.g. I actually saw the file after a few seconds, instead of getting bored and hitting ^C) And well, 350100 tags is a bit too much anyway so "a few seconds" seem reasonable. Especially as with "only" 35010 it's close to usable, so I guess it's just great.
And as you mention, it scales a lot better: 350000 tags is only slightly slower than 35000, while it used to be a *lot* slower (I didn't even care to wait long enough). --- Reply to this email directly or view it on GitHub: https://github.com/geany/geany/pull/797#issuecomment-166696813
