ID:2372951
 
This week started out late because of the holiday, but I had a few bugs to hunt down so a lot of it went into dealing with that.

As it turned out, one of my server improvements for 1425 actually caused a crash on reset: the new bag for ID arrays was not being cleared, as it should have been. Hopefully now with 512.1426 out, Space Station 13 will see some speedups thanks to the server boost.

I also poked the layering bear in response to a bug report from Ter13. Since 455 we've had some behavior in place to handle big icons, where they'll "pull forward" from other icons on the same layer so that if a big-icon turf overlaps other turfs, it will appear correctly. However I'm not sure that logic was ever quite right because of DS's "microlayering" behavior (introduced by Dan in 325), and in the era of visual contents it was failing utterly. In fact I believe there may be some outstanding bug reports that will be fixed as a result of the changes I made. Namely, I switched back to using microlayers entirely but made sure that the microlayer worked better as a floating point number by using powers of 2, and made it just a titch larger so it could be added to fairly normal-range layers without rounding errors getting in the way.

A couple of minor tweaks are in place for 512.1427, including a fix to what appears to be a relatively uncommon icon operation crash (client-side only), but I think those can wait till more fixes and features have piled up.

A tiny little voice is telling me I should bite the bullet and undertake the long-overdue project to switch Dream Maker over to the hardware renderer.

Big thanks to all our Members, donors, and Patrons who contributed last month to help keep BYOND going. Love you guys. I'm always working on stuff to keep this platform moving forward.
"A tiny little voice is telling me I should bite the bullet and undertake the long-overdue project to switch Dream Maker over to the hardware renderer."

What kind of results would we see from this?
In response to Bravo1
Bravo1 wrote:
"A tiny little voice is telling me I should bite the bullet and undertake the long-overdue project to switch Dream Maker over to the hardware renderer."

What kind of results would we see from this?

A more unified means of rendering across the desktop environment, which basically means quicker turnaround on work on the renderer in the future, and better, more responsive tools in Dream Maker. Oh, and fewer limitations on what we can do in terms of rendering.
In response to Ter13
Hmm. Would this solve some of the slowdown issues when working on the icon editor with large icon sizes?
If I can get the software renderer mooted out completely, I will look into 3D matrix support.
In response to Lummox JR
Lummox JR wrote:
If I can get the software renderer mooted out completely, I will look into 3D matrix support.

O:
3D matrix support in BYOND? That would change the ball game quite dramatically... For the better. I never thought I'd say this about an update but I'm actually looking forward to that one.
In response to Lummox JR
lummoxjr are you really working on the webclient like are you adding some new stuff to it?
I've taken the first step towards this new direction by finally, FINALLY moving the core rendering code from the client into byondwin.

Next: Dream Maker needs to start using the same structures for icons that Dream Seeker does. (That's another project.) Then, it needs to switch over to using the rendering code now in byondwin instead of its own. (Also a project.)

But anyway that's one big hurdle down toward 86ing the software renderer; about half a dozen big hurdles to go.
In response to Lummox JR
You're doing God's work, Lummy. <3