float16 in StreamPeer

[theraot]

Currently in StreamPeer I can put and get float and double, and I can put and get int16, but no half precision floating point (float16).

I want to build-in methods to put and get float16 for network traffic and serialization optimization.

Converting GDScript float to float16 stored in an int to use with get_int16 and put_int16 would be error prone and less efficient.

Furthermore, I think the float and double methods could use a rename. So:

• float get_float16(), void put_float16(float value) (currently not existent)
• float get_float32(), void put_float32(float value) (currently float get_float(), void put_float(float value))
• float get_float64(), void put_float64(float value) (currently float get_double(), void put_double(float value))

Notice the mix of language here: the API currently uses the words double and float instead of double and single. Meanwhile the methods working with int say the size in bits.

[Calinou]

I'd suggest get_half() and put_half() instead of renaming the existing methods.

[theraot]

Naming aside the feature is already in the engine. I'm closing this.