there's actually 3 grapples coded in and i pulled my hair out trying to track them all down.
there's +grapple which is the normal q2wf grapple
there's grapple, which is a sticky grapple with different physics (still offhand, but grapple cmd turns it on and off)
then there's weapon grapple, which is the on-hand grapple that is straight out of loki's ctf.
the grapple physics are different between both q2wf and q3wfa because of the sum of 3 factors:
1.) my fiddling with the grapple code
2.) my fiddling with collision detection
3.) the differences between q2 and q3 bsp collision maps
I really didn't touch the q2wf grapple code that much except for decreasing it's think time, basically quartering it, and fixing how trigger pushes affect grapplers (wfa bugs revisited). On that note, expect allot of early wfa bugs/crashes to appear, because I could only think of so many, but the way grapples are handled with movers is completely different then wfa.
Collision detection is terribly different between q2 and q3, due to botclips and meshes.
As far as a linux port, i'm keeping up to date on what i can do with cygwin. It compiles under cygwin fine so "I assume" it will compile under straight flavors of *nix.
But i'm also working on a macintosh client port, or actually hunting out people who know what to do so they can do it for me. I actually found a team willing to help, but only osx which is fine.
the ui is going to change a bit more. simtex and a-sniper have tons of proof-of-concept shots, i will be implementing them. but it'll basically be the same structure until the next code change.
i'll be releasing an alpha -client only- soon, as well as running a server at my house that is prepped with visual studio, so don't worry, a test is coming. wish i had a real line, but with 50k up i should be ok to host 16 people on this without lag.
