Exposed parse method in TLAPM API

[ahelwer]

This adds a simple string -> parse tree method in TLAPM's public API, so it can be hit by non-inline unit tests.

[kape1395]

I added a similar function in the LSP branch.

https://github.com/kape1395/tlapm/blob/0ac8eac06ccb96a4cb138e622e15fa4aaccf41ac/src/tlapm_lib.ml#L617

Mine does more about related modules. Is is meaningful to keep them both?

[ahelwer]

I think so, because it would be nice to just generate a syntax tree without doing anything else. For example, I would want to check how EXTENDS and INSTANCE statements are parsed without having to actually have those modules in existence.

[kape1395]

@ahelwer, the LSP branch has now been merged.
Can you fix the conflict? What do you think about renaming

• my function: module_of_string -> elaborate_module_of_string
• and your: parse_module_from_string -> parse_module_of_string?

[ahelwer]

How about transitive_parse_module_of_string for your method?

[kape1395]

I sent you a PR with an attempt to resolve this naming clash.

[kape1395]

Why have you reverted modctx_of_string to transitive_parse_module_of_string. I think the latter is not very precise.