Climbing on one-sided linedefs

Does what it says on the tin. You couldn't before, now you can.

With this branch, thok barriers are now completely optional for maps with no aesthetic need for sky-topped walls. All the bugs and complications that exist with climbing right now still exist, except now there's one less issue making them suck forever and now Nev3r will be very happy.

Do I need a testing .wad? It's... not that hard to create a very small map and go nuts with Knuckles in it. I have one, but uploading it to the ftp almost feels insulting.

Added 1 new commit:

• f93b31f0 - FUCK C90

Less cursing of C standards and more compiling with error/warning flags, for future reference? =V

I personally compile with ERRORMODE=1 and GCC48=1 flags (higher GCCxx version flags exist mind)

you know that's probably wise honestly, it's just i'm very sleeb-feeling right now so let me curse

Status changed to closed

Status changed to reopened

mentioned in commit a938efa3

Status changed to merged

http://i.imgur.com/0TlqIcb.png

i hate everything, do we revert this or do we try and get P_IsPointInSubsector to work in OpenGL (because it is this, swapping P_PointInSubsector out for that was the only thing i changed meaningfully in the climbing code)

...what's going on here exactly? What does OpenGL have to do with this exactly? Is it a renderer-specific issue?

it's renderer specific

knuckles' wall climbing code uses P_PointInSubsector in 2.1.15 and earlier, whilst this branch uses P_IsPointInSubsector, and for SOME reason this change means Knuckles stays climbing forever, even if he's climbed into the air to the side of/above/below a latchable wall, in OPENGL ONLY. in software it behaves exactly as you would expect.

That is bizarre, huh