Extra major complaint (Page 2 of 3)

The problem here is, that with this feature, you make it easy for the bad guys, to get their malicious code executed on the victims machines. We would need to find some way to "sandbox" these "mods" so they are very limited in the damage they can do. On the other hand, this might again limit the possibilities of those client mods. However, all of this would be a lot easier in LUA, and the other good thing about lua is that it is a interpreted language, which means you won't get a binary as mod, but you will have the source files, which means you could actually have a look at them and check if they do anything bad, before you execute them. I still think this would be quite a good approach, but it requires a lot of work, and I think we need to do more important stuff first.

This will always need client mods, and if we implement a client mod downloader, well see above what I wrote about that. You said earlier, that the protocol is quite restrictive which makes it hard to implement some things. Could you please name some examples and maybe give some possible solutions? I don't think it would hurt anybody to alter the protocol a bit so that it is easier to implement new things (this would actually be quite nice for vanilla development, too, because we would not need to break compatibility with the old version so fast, when developing).

Thanks, this is all I wanted to see. Working together to try come up with a solution to the problem (and admittedly a very hard one at that), rather than just dismiss it without giving it much thought.

I'm not hugely knowledgeable on either of these languages (only started on C++ this year), but could a temporary solution to the malicious code problem but that you only able to load mods that are from an official list of trusted developers, whose code is publicly available. Such as the DDrace mod.

In terms of the sandboxing approach, which would definitely be the safer option, you said  that it would take a lot of time, but what if someone else offered to do the programming of it; freeing you guys up to work on other more important things. The only thing you would need to do is just to go through the code to make sure that it is legit and is up to standard.

Fair enough, I think the sandboxing approach would be the best bet or something else if anyone has a better idea.

EDIT:
I think another thread should be made focusing, on the mod loader as the other pointless "issues" and solutions mentioned in the initial post detract from the main problem at hand. Also should be titled differently maybe "Mod Loader, possible solutions?" rather than the current title as I don't see this as a complaint but instead a suggestion.

Thats great to hear, but Ghost please could you find out by speaking with oy or whoever else, what their thoughts are on this and if they approve of the feature. Because taking the magnitude of programming this, it would be a complete waste if people go through all this work for it not to be included. Or tell me the proper channels to speak with them in regards to this, because I am sure they are too busy to answer every pm they receive.

No, I want the opposite.  I want to give more power to the people who make mods.

I want to make mod makers to make changes into Teeworlds protocol. For example, there could be mods with more than five weapons, mods with vehicles that look like this:

and not ones made from hearts or armors -.-

Vehicles in Teeworlds! Totally teeish!

It could be! But it wont... But I would like to give more power to people to encourage them to make new maps and mods. Maybe make a tutorial in which the most important step is to instruct people to come up with original and awesome mods like openFNG and zCatch.

var a = "I am a string!";  
alert(typeof a); // Will output "string"  
var b = new String("I am also a string!");  
alert(typeof b); // Will output "object"

Omg... Who made JavaScript?

And standard Teeworlds is a boring game deliberately crafted against players. The only thing that saves it is laser, without which the only thing people can do is to camp.

Teeworlds need a panic button... i know that this button can destroy all the universe and i like it alot! ^^

I dont read the whole thread but i know both sides.
On the one hand mods like nodes could add very nice gametypes to teeworlds. on the other hand the devs dont want to support this stuff. I know that it is hard to provide a good modding api. As Dune and Ghost said before you have to fix security issues and keep the mod running in a sandbox. Also you need to make sure that the scripts work on every OS which isn't easy if the api grows. The selection of the API is a very important step.

I decided to use lua due to it's easy implementation. But i noticed that lua is just amazing. It supports everything you need for modding. including an simple c/c++ interface to add new functions and of course multiple return values which makes functions like GetCharacterPos very easy:
x, y = GetCharacterPos(id)

This is one of the main advantage. Javascript is a good scripting language too but i do not think that javascript would be good for teeworlds.

The Lua API that i provide grows to about 15'000-20'000 lines of c/c++ code. Additionally we added sqlite3 to provide an easy way to store data. The whole client is about 3 times larger (line number) as the normal client. This is caused due to the additional libs needed to provide simple and great modding possibility. We are also working to make the lua api safer. This includes to deny system calls, etc.

To conclude: I dont know if these kind of modding would be good for teeworlds. due to this fact i launched and kept this project as an unofficial client/server mod. I know that there are good reason for these kind of modding but there are also good reasons to KEEP IT SIMPLE (kiss).

I think a scripting language would not be added before 0.9.0