Make vecdraw:SCALE work only on :VECTOR, not :START

[Dunbaratu]

(I've known this was wrong for a while but thought it was too late to fix it because it had been in public use. But now I think it's impossible to use it anyway the way it is now so nobody has been using it. Users keep giving each other advice of "leave this at 1, don't change it because it messes up" and I suspect nobody is using it at all. Which makes it okay to change it to something more sane.)

What it does now: When you change the :SCALE of a vecdraw, it multiplies the scale by BOTH the :START and the :VEC vectors, causing the vector to not only get bigger, as expected, but also to change where the tail is drawn, which isn't expected.

Proper behavior: Just use :SCALE on the :VEC only, not the :START position.

[space-is-hard]

I'm no vector expert, but I've never come across a situation where the current implementation of :SCALE has been useful anyways.

+1

[hvacengi]

Agreed. As is, I just multiply the vector by my scale factor, and leave the specified scale factor set to 1. I think the current implementation is only useful if you are trying to visualize how vectors stack together (or instance current velocity plus delta v) at a larger scale.

[ozin370]

I highly support this change. When I started using vecdraws and tried to use the scale param to adjust the vector scale, it really confused me and I thought I was doing something wrong.

[ozin370]

Actually, come to think of it.. it would be great if scale could be changed into thickness. Depending on circumstances, vecdraws are often too thin or thick compared to the length, sometimes making them a bit useless.

[Dunbaratu]

Sort of halfway between bug and enhancement. Although the current behavior is consistent, it's consistently stupid and useless. I think it needs fixing.

[Dunbaratu]

I like the idea of a width, but not for the scale value because that would mess things up in existing scripts.

[Dunbaratu]

Whoa.. there already IS a :width suffix, but it's totally undocumented. I forgot all about it.

[ozin370]

Hah! Good to know

[Dunbaratu]

My PR above documents :WIDTH, as well as implementing some new rules for how scaling works. I'm hoping it's entirely compatible with existing scripts, except for the rare exception where someone actually used both :SCALE and a nonzero :START location, which I'm thinking is going to be super super rare because that's exactly the situation that made this previous type of scaling so broken.

[Dunbaratu]

Warning, if you start using :width now, before my pull request above becomes public, you'll be using a system that's about to change its interpretation slightly. I figured it was safe to change how it works given that it wasn't officially promised before (by not being documented).