[Solatar]

Thanks, that's a nice start. It could come in handy.

[EdrickV]

Originally posted by RandomTiger
I believe its a seperate compile option thats not been checked for ages so what little functionality was there is probably broken.

I'm not entirely sure if that code will work without major changes or not. (And remember, all my info is based freespace2_public rather then fs2_open.) The demo recording/playback code is marked by ifdefs in some places, other places have checks to make sure it doesn't check certain things while it's playing back a demo. (Said checks seem to exist in the release versions as well, though they'd always do the non-demo playback stuff.) And the demo playback is only enabled for a debug build as is. (See the call to demo_start_playback() from main_hall_do() in MainHallMenu.cpp.)

Setting the demo recorder on or off would be done using the debug console. As I can't compile the code and don't know how to use the debug console, I can't really test it. The demo file itself, as the code stands, would be test.fsd.

To enable demo recording you would have to make a #define for DEMO_SYSTEM and change demo_make to 1 either via code or using the debug console. Demo playback is done by hitting "KEY_DEBUGGED" (defined as 0x8000) plus d. I'm not sure what that debug key actually is because I'm not sure what key that hex value is associated with.

The function calls that I believe are needed to actually record the demo are present in Freespace.cpp, again associated with the DEMO_SYSTEM define. As far as I can tell without trying it, it looks like the demo system's code is still in place and might work. The one thing I am not sure about is game_level_init() which seems to have an integer argument that would only be set by the demo playback code, used to set the random number generator to the same number used in making the demo, but when called the function isn't passed any variables so it would never set that. I'm not sure if that even matters though. (It may be that was from an earlier version of the demo code and isn't used anymore.) In any event, it looks like the demo code could use some cleanup. (Just take a look at the #define for DEMO_DATA and you'll see what I mean. )

[EdrickV]

To correct myself, it seems the DEMO_SYSTEM #define is already present in demo.h.

[Thor]

Thank you.  I'll try it and see how it works.

[YodaSean]

Originally posted by DTP

so re-think this again. Nobody in his or her right mind would do this, for nothing because it is going to take "A LOT" of time.

as an example. :V:'s red faction GEO-mod technology took 2 years to program.

 

I don't understand this.  It didn't take the developers of Worms that long to make their geomod technology...

[RandomTiger]

2D = easy
3D = hard

[Fry_Day]

And besides, the worms geo-mod is totally different tech-wise.
They simply store a large 2d array of pixels,and for each pixel which was destroyed, they simply change the collision flag. That's it. On the other hand, in geo-mod, you have to to a substractive boolean for each explosion with the walls.
A wee bit harder, since you also need to remove unused vertices, so the game won't get bogged down after prolonged play, and a thousand other little details

[Bobboau]

does anyone here know how to do boolean subtraction of polymodels?
...were's that NASA guy
I know, I know, NASA doesn't do that kind of stuff

[RandomTiger]

Um... does anyone realise how difficult that task would be?

VERY!

[EdrickV]

Originally posted by RandomTiger
Um... does anyone realise how difficult that task would be?

VERY!

You mean the polygon boolean subtract aka geomod? Not really, but I imagine it would be harder then anything we, as a group, have tried doing to the source code.

[Fry_Day]

Edrick, but that creates a whole new set of problems - you'd have to store the geometry of each damaged ship, even if they're of the same type. Worse, the BSP tree of a damaged won't be right, so you'd have to rebuild it (There's probably a better solution, like some sort of merge with the tree of the sphere you're substracting, or whatever).
It didn't take volition a long time to get Geo-Mod working for nothing. I think it's a bit beyond the scope of the SCP.

[RandomTiger]

Originally posted by EdrickV


You mean the polygon boolean subtract aka geomod? Not really, but I imagine it would be harder then anything we, as a group, have tried doing to the source code.

It took two years to make that technology apparently, and they are a full time staff of experienced paid codes.

We are not in that kind of position and while I invite anyone to prove me wrong I think there are many other jobs that are a lot easier and would be as rewarding. Theres MOD somewhere out there that boasts this kind of feature using MOD tricks, perhaps a good start to such a project would be to provide him with code support.

[EdrickV]

In otherwords, it is harder then anything we have tried to do. Not that I personally care either way. For scripted mission sequences I can think of ways to fake geomod for a particular situation. For example, having a ship fire a beam through another ship and leaving a hole behind. The trick, of course, is that the model would have to be specially built to do that (destroyable subobject or in game POF swapping) and the ships would have to be lined up just right for it to look as it should. That's nowhere near true geomod stuff, which I've never seen but can imagine, and I'd know I wouldn't attempt to make a full geomod system, it's beyond my expertise.