[E2] Built-in ranger changes?

[Divran]

I noticed there are a lot of more data you can get out of a ranger trace which isn't accessible through E2.

I've already made these and they will be committed:

• ranger:fraction()
• ranger:hitSky()
• ranger:hitWorld()
• ranger:matType()
• ranger:toTable()
• ranger:hitGroup()

I'm going to make these:

• None atm. They have all been made See above

I'm thinking of making these:

• ranger:fractionLeftSolid()
• ranger:positionLeftSolid()

And I'm having compatibility issues with these:

• ranger:position() and ranger:fractionLeftSolid()/ranger:leftSolidPosition()
  Currently, ranger:position() returns the position where it leaves the solid area if it was created inside a solid area. (By solid, I mean the world. Props don't count)
  This, however, makes fractionLeftSolid and leftSolidPosition pointless. Therefore, I'm considering changing ranger:position() so that it returns the hit position like Lua does (ie if the ranger starts in the world, and exits the solid brush, continues on, and hits another solid face, it should return THAT hit position. That's how the regular Lua traces work). I can then add the other two functions and they will then have a purpose.

The reason I'm posting this is for you to discuss and suggest what you think should be done about this/these compatibility issues (By compatibility issues I'm referring to the fact that old E2s may break). Also created a poll.

EDIT: Also, here's the Gmod Wiki page which explains what all of these things do: http://wiki.garrysmod.com/?title=Traceres

EDIT: REMEMBER it will only be a compatibility problem for old E2s who need to scan the ground (for simple obstacle - gaps/chasms in this case - avoidance or similar), because it only affects traces which were started inside the world and exits it.

[Spoon God]

E2 Rangers are awesome...this is good!
If you're worried about old E2s breaking with ranger:fractionLeftSolid()/ranger:leftSolidPosition() you should just make another function that does what you were thinking of making ranger:position() do. Something like ranger:forceHitPos() or ...i don't know, ranger:position2()...or maybe even ranger:pos() since it DOESN'T EXIST YET! >_<
Maybe a ranger:bounce() would be useful, which would return a direction vector facing how the ranger would bounce off where it hit (as though it were a laser hitting a mirror)

EDIT:
If you decide to use ranger:fractionLeftSolid()/ranger:leftSolidPosition() anyway, my vote goes to using ranger:leftSolidPosition()

[Colonel Thirty Two]

Don't think you need RD:hitNonWorld unless it does not equal !RD:hitWorld()

[Divran]

Don't think you need RD:hitNonWorld unless it does not equal !RD:hitWorld()

I'm not sure, but it's possible it does equal !hitWorld.

[Grocel]

A ranger always fires from the e2 unless some thing else is set with rangerOffset(V,V).

If I want to fire a ranger from a entity else then the e2 is quite annoying to set it up with rangerOffset and the E:pos() stuff, because with E:pos() it doesn't work because the ranger trace, will be blocked by the E entity used by E:pos().

So can we get a RD = E:ranger(N) function?
I would be like ranger(N) but instant of the e2 entity it uses a another entity.

And can you make us able to do some thing like RD:start(V) and RD:end(V), this changes the start and the end of the ranger trace, it would be useful.

[Divran]

E:ranger(N) is certainly possible, but

And can you make us able to do some thing like RD:start(V) and RD:end(V), this changes the start and the end of the ranger trace, it would be useful.

this would mean you retrace again every time you change the start or end... thus it's useless because you may as well run another rangerOffset with new arguments.

[Grocel]

E:ranger(N) is certainly possible

Nice.

this would mean you retrace again every time you change the start or end... thus it's useless because you may as well run another rangerOffset with new arguments.

Oh, I didn't know, nevermind. ^_^

[Divran]

Bump
I've added another option to the poll which I thought should work.

[Tolyzor]

Do any of these give a hit water output?

Another thing that would be useful (but is likely not possible) would be a leaveObject() output, so that if the trace is started inside a prop, or underwater, it returns the position where it leaves the object.

[Divran]

I made a hitWater ranger function using pointContents, but you can now just use the matType function (returns "slosh" )

Don't think it's possible to make a function that returns the position where it left the (non-world) object, without making it expensive