Should std.log be used in the library?

[natecraddock]

See this discussion: #55 (comment)

[VisenDev]

My motivation for this is I think at a bare minimum we should be able to tell the user why something fails. If parsing fails, I think it should have a way to tell the user where it failed and why.

Relevant ziglang/zig#17995

This is how std.json does this I believe
https://ziglang.org/documentation/master/std/#A;std:json.Diagnostics

I added logging as a simple way to achieve this functionality, since I hadn't looked into adding more complete diagnostics functionality yet

Perhaps we could look into replicating some of the functionality from std.json?

[nurpax]

I'd be a bit against logging by default from library code. At least there must be a way to turn it off. (Or preferably it should be off by default).

I think a motivating use-case where logging is undesirable is where someone is building a console application (f.ex. something like pandoc) using Ziglua. If the inputs to the application fail, it should be the application printing an error message and there should be no errors logged from Ziglua in this case.

The diagnostics pattern in Zig is good though. But I don't think that's logging anything using std.log.