The future of scriptcore

[dnmr]

This is a question of high importance. It's aimed at the new dev team, but anyone else is free to flame me of course.
The question is as follows: are you guys going to keep the current scriptcore in the soldat dedicated server? Or are you going the change it to something completely different?

This might sound a little weird even to rationally thinking people, but! Most of you probably know that the current scriptcore has a lot of limitations and.. well... "inconveniences". Mainly the fact that some things are implemented, but they do not work properly (for example threads, and according to some people even dynamic arrays are totally fucked). Also, many scripts act differently on different systems - that's pretty much a well known fact. Namely, the server just dies while running a script on one machine, but runs fine on another.
I mean.. The scriptcore was implemented to give some basic scripting functionality from what i gathered. And it just feels like more stuff was being piled up on top of it with time, so it doesn't really work all that well at this point.

Now I'm not sure if the members of the new dev team have experienced the pain with the current scriptcore, but i'm sure there are quite a few coders out here who have, if you haven't. Hell, we even call it shitcore nowadays

So, have you considered replacing the scripting engine with something more stable and convenient, or are we going to stick to the old mess of a thing? I have to say i'm not a full-fledged coder myself and i can't know what the damn is going on inside of the server, so i do accept that it might even be completely impossible to implement a stable enough scripting engine to satisfy everyone, using the dedicated server code you guys have now. I just want to hear about your plans and find out if there is any hope for reliable scripting for this game. Even if the syntax is going to be completely different (actually it might even be a positive thing if the syntax will change).

Thank you for your time, feel free to reply when you have sorted thisngs out

[Falcon`]

Use the Falcon programming language! FTW!
Or no! Use The LOLCode!!!1
</trollpost>
And ye, i agree that current shitcore needs to be TOTALLY changed. But i'm not that skilled programmer to say to what. I guess just expending current pascal to sth err "bigger" may be easiest since soldat is written in delphi (which is object pascal in fact) afaik.

[DorkeyDear]

As far as I can see, there are a few options:

• do nothing
• improve from its current state
• replace the scripting core with the same scripting core, but with features made from scratch (only if its not the actual core causing it)
• replace the scripting core with a same-language scripting core
• replace the scripting core with a different-language scripting core

]

	Advantages	Disadvantages
1.	no work to be done - takes no time to accomplish	not used to the code, and if it is unorganized, it may be difficult to improve
2.	may be a faster solution than others provided (except #1)	may be difficult to work with in its current state (and take longer)
3.	unstable core may be stabilized	unstable core may not be stabilized
4.	can possible make it so it is backwards-capability (scripts that is)	will take longer amount of time than other choices
5.	may be for the better - pascal is an old language (but I don't see anything wrong with using it still)	guarantees old scripts will not work, and must be remade in the new language

If it is completely redone with a different language (not pascal nor delphi, something like lua or gamemonkey or something), many scripters would have to relearn a whole new language. if anything is to be changed, i suggest keeping the sme language, and current functions; we don't really want to break capability too much, since theres a TON of scripts out there, and I would feel like it was all for nothing if a new langauge was to replace what there currently is.
if it is possible to stabize the "crapcore", then that would be what I would try first.. if unsuccessful, then maybe replace it wiht another pascal language engine.. if unsuccessful, then theres not much of a choice: keep it as is (hehe), or replace it with a different language...

and a side note: i do not actually expect any changes to the scripting core in the next update since many peopel don't even use it, and bugs & cheats they care about more ..

[dnmr]

and a side note: i do not actually expect any changes to the scripting core in the next update since many peopel don't even use it, and bugs & cheats they care about more ..

yes, of course nobody would expect anything like that for the next version, that would be nonsense. Just want to find out if them devs have anything like that planned.

[Blue-ninja]

afaik, the moveplayer is pretty screwed up, as scripts like saves and loads often kills you by teleporting you into deadly polygons for an instant on climb maps.

[tk]

actually it might even be a positive thing if the syntax will change).

If it is completely redone with a different language (not pascal nor delphi, something like lua or gamemonkey or something), many scripters would have to relearn a whole new language.

Hell no! The problem is not relearning. And I agree that there are propably better scripting languages than Pascal, however I just don't feel like rewriting my 11k lined LS script to other s**t

//edit

But it's a fact that some things should be changed to increase efficiency and stability. Like adding alternative for all basic functions such as GetPlayer/SpawnStat crap which have a STRING as a parameter, (i can't think of more lame way here). Each time such function is called, a long list of strings is matched... and as I saw the implementation of it, there is no even single 'else'...

[Hacktank]

I agree with dnmr that nothing should be done with the scriptcore in the next update, it and its bug fixes should be hastened in any way possible. But i would like to see it in the update after the next.

Dynamic arrays are totally fu***d, they cause memory leaks...

I personally hate pascal, i would be ecstatic if the scriptcore was replaced by Lua or something similar. Failing this, i think the current scriptcore should be given a complete overhaul, after reading though the script suggestions thread ofc.

[Furai]

For now I'd like to see bugfixes and stabilizing scriptcore. I think that we shouldn't change scripting language. At least not now. Main reason is that Pascal is being taught in many schools + all the scripts are written in it and it will be pain in the ass for all the scripters to rewrite it.

Anyway, as administrator of few servers I've experienced instability of scriptcore. On one machine all scripts were working fine, on other server crashes a lot. I say that bugfixes have to be made before everything. And as new developers get used to the soldat's code then we should think of some radical changes to scriptcore.

If devs will choose to rewrite it and if there is possibility to do so they should rewrite it in another language (I don't know how soldatserver's source looks, maybe it's not possible to have scriptcore in any other lanuage than pascal?).

[dnmr]

I think that we shouldn't change scripting language. At least not now.

the sooner the better, if it's ever going to happen