Add support for Erlang Process Labels

[ToddG]

WHAT
Expose Erlang Process Labels in Gleam.

WHY
Labeling gleam processes enables displaying processes in the Erlang Process Observer with those labels. This greatly assists in debugging and documentation.

EXAMPLE
Note in the attached pic how the labels make it easy to reason about the labeled parts of the app...whereas those parts with just PIDs are basically opaque (to me, anyway).

CONTEXT
https://discordapp.com/channels/768594524158427167/1397286856290340894/1418251111499956324

IMAGE(S)

[lpil]

Sounds good to me.

Are there any restrictions we'd want to add on what values can be used as labels?

[ToddG]

The docs state:

The process label can be any term, for example {worker_process, 1..N}.

Although my personal pref would be to limit my terms to [a-zA-Z-_] length ~12 or 15 characters, that's just me. My vote would be to follow what proc_lib allows.

[lpil]

I wasn't asking what Erlang can do, I was asking what we should do! We commonly do different things to Erlang in order to provide the developer experience we want.

I'm leaning towards permitting anything, but seeing as we can never change these APIs after we add them it is always important to consider all the options before making code changes.

[karlsson]

Maybe this issue should go into gleam_erlang application with a set_label function added to the process module?

[lpil]

Smart, I agree!