NWNX4 Prototype 1.08 beta 2 ready

[Gryphyn]

When I same GameObject I was referring to the object-data, not the string-data which available. In your context, GameObject being the 'string-data', this is still available, it's just been given a new name.
It's also been made semi-safe by providing immutability of the data by using internal memory management to accept a return result.

Sure, there's no reason why BOTH methods could not be available to plugin developers. Papillon made that call when the new NWScript functions became available, he made the justifications then.

If no-one stirs the waters we'd just stagnate.

Cheers
Gryphyn

[virusman]

[Papillon]

The disadvantage of not having the GameObject pointer anymore is a valid point, and it's not the way I want it to be.

At the time I was about to finalize 1.08, nobody was using it (AFAIK) and I was thinking that the NWNXSetObject and NWNXGetObject functions would be available soon - which was not the case. Those functions would offer a much cleaner and powerful interface to manipulating objects, but well - we do not have them.

Currently, I am not sure if those functions will ever make it. I'll ask them again, and I hope I will receive an answer this time (the problem might have been a SPAM filter the last time). If we will not get those functions, we should start thinking about hooking SCORCO again, and maybe other functions that would be appropriate for getting the object pointer.

The thing about the virtual functions was already discussed between us and will be fixed. The point about the prototype not beeing backward compatible to the final version is... uhmm backward .
_________________
Papillon

[virusman]

NWNX4 Functions and NWNX4 Leto were using the GameObject pointer. NWNX4 Leto requires the old interface as well (I don't know Delphi, so I have to work with what dragonsong gave me).
I doubt that waiting half a year each time for Obsidian to implement 1-2 functions for us is a right thing to do, considering that the same can be done with hooking in 2-3 days. I know that you want to get rid of MadCHook to make NWNX4 x64-compatible, but maybe there is another solution to this?

[Papillon]

I agree that hooking all other functions expect those OE gave us already, and the (future) object functions, can and should be done without OE by just hooking them. The NWNX* and SCORCO functions are so fundamental, though, that I'd like to have official hooks for them.

This has nothing to do with madcodehook yes or no, though. I can "hook" the functions in 1.08 without madcodehook, but I can not load the NWNX DLL into the server's process space. Well actually I can, but because of some nit-picky problem which I can not quite remember right now, starting the IPC Server does not work once my DLLMain functions executes inside the server process. I guess it is because you are quite limited in what you can do within DLLMain (loader lock and such), and madcodehook probably uses another way to inject the DLL. I am so || close to removing the dependency on madcodehook, and than such a stupid problem gets in the way.. bleh.

_____

Side note: If I see comments like "huh, this is easy, why not just this way or that way"... I can not help but think: Yeah, right, present me a solution. Nothing is holding you back.

Side note 2: Is it just a coincidence that this discussion about 1.08 final takes place in the beta thread ? Since, you know, 1.08 was in beta for quite some time, and comments about it's very core parts would have been somewhat more helpful during that time, but really much less a couple of weeks later.
_________________
Papillon

[virusman]

[Papillon]

[GodBeastX]