Allow resizing Strings from GDExtension

[dsnopek]

This adds a new function to gdextension_interface.h to allow us to implement String::resize() in godot-cpp per issue godotengine/godot-cpp#1141

This is added to gdextension_inteface.h rather than ClassDB, because we don't want scripting to have access to it. String::resize() only makes sense if you can access String::ptrw() and directly modify the data at the pointer, which would be impossible from GDScript.

This originally came up when trying to compile the text_server_adv module in Godot as a GDExtension. See: #77532

[bruvzg]

Is it necessary to expose it directly, adding bind_string_method(resize, sarray("size"), varray()); to the variant_call.cpp to expose existing method should work. The other string functions are in the interface due to pointer arguments, but in this case it's only int size.

[bruvzg]

because we don't want scripting to have access to it.

Why? It's exposed for the Arrays, what's the difference, String have writable [] operator exposed, which is the same as ptrw() ?

[dsnopek]

Why? It's exposed for the Arrays, what's the difference, String have writable [] operator exposed, which is the same as ptrw() ?

Ah, that's a good point!

When I get a chance, I'll make a new PR that exposes it normally...

[dsnopek]

I've open PR #79177 which exposes String.resize() the normal way.

Given that String.resize() takes the desired array size of the string (ie. including the null terminator), I'm now less sure that we really want this exposed to GDScript, at least without some additional changes to make it more user-friendly. (Some possible options listed over on the other PR.)

[AThousandShips]

As discussed in #79177 this seems to be the better way to handle this, assuming it doesn't suffer from the same mutability issues

[dsnopek]

Discussed at the GDExtension meeting, and we think it makes sense to add this, however, it would be good to add some additional warnings in the documentation about this being a potentially dangerous operation.

[YuriSizov]

it would be good to add some additional warnings in the documentation about this being a potentially dangerous operation.

Do you mean the online documentation? Or is there something you still want to add to the class references (and therefore need to update this PR)?

[dsnopek]

Do you mean the online documentation?

We were talking about the docs in the comments in gdextension_interface.h. I've already done a push since the meeting that adds this:

 * Warning: This is an error-prone operation - only use it if there's no other
 * efficient way to accomplish your goal.

[YuriSizov]

Thanks!