Contribution description

This PR aims to address the first three tasks in #16855.

This is still WIP, but before I start changing types and documentation, I wanted to have some input.

I followed the following strategies for the wrappers:

• Most of the time the wrapping is straightforward (just call the respective nanocoap function in an inline function)
• due to API inconsistencies in nanocoap sometimes casting is necessary; I aimed to keep the types consistent in the wrappers
• sometimes there are \0-terminated and length variants for adding strings (with the \0-terminated variants calling strlen() most of the time, so I decided to use the length variants only.
• some nanocoap functions use coap_pkt_t, some use coap_hdr_t, some use just raw buffers as the object acted upon. I decided for the wrapper to always use coap_pkt_t as the object acted upon to simplify things towards the front-end. Only exception is gcoap_pkt_build_hdr(), as coap_pkt_init() requires the headers in buf already to be initialized, but it also sets pkt->hdr to the buffer pointer. So requiring pkt->hdr to be already initialized in gcoap_pkt_build_hdr() would lead to pkt->hdr to be initialized twice (once by the caller of gcoap_pkt_build_hdr() and once to in coap_pkt_init(). As such, I decided to have just an opaque buffer be the object acted upon by gcoap_pkt_build_hdr() (while writing this I reallized that as such the name gcoap_build_hdr() might be better suited, but I leave this up to discussion).
• some nanocoap functions are always used with other functions (e.g. setting the payload always adds the payload marker first) in gcoap and its applications), so the wrapper always does this as well.

Future TODOs:

• introduce gcoap-specific type wrappers for nanocoap types
• unify type naming in gcoap types to always be prefixed with gcoap
• use the gcoap-specific types in newly introduced functions
• use the newly introduced functions throughout gcoap and the applications.

Testing procedure

examples/gcoap should still compile. Currently, the new wrappers are not used yet, so nothing to test.

Issues/PRs references

Addresse parts of #16855