ID:2322935
 
It's been an extremely fruitful couple of weeks for optimizations.

This week, I turned my focus to the client side, because there are some outstanding requests to look into that and some of them come with purse strings attached. So on the heels of last week's server modifications (released early this week), I dug deep into the client code to see where things could potentially get faster. Boy howdy did I find things to trim. The result of this is that performance for the routines that go into turning atoms into icons should be massively improved. Several of the major workhorse routines should see substantial improvement, including the all-important GetCappearanceItileList() which is like the beating heart of the client.

Also on the client, plane masters are much better about what gets clipped and what doesn't. Bravo is working on parallax scrolling, but turfs in a scaled-down plane were being cutoff prematurely. This at last has a fix, and boy was it a difficult one to manage. Multiple parts of the code fought me on this one, but now a lot better results can be had by messing with planes.

On a similar note, I finally fixed an issue in 512 which was bugging me, that being that objects with filters could suddenly have the overflow from those filters appear or disappear just because the object itself went out of draw bounds. Now the overflows from filters are taken into account.

Time to turn attention back to the webclient, I do believe.

Christmas is almost upon us! Get your online shopping on and while you're at it, don't forget to $ay thank$ to BYOND for all the development this year. For all who've become Members or donated, your support is hugely appreciated.
Can't wait to play with byond again during the winter break
Will these performance upgrades make the webclient run better?