Make token / detoken errors more verbose

Debugging token issues such a missing ones can be quite a pain, so after failing to see the issue of my script using debug build (stacktrace) and even --scripts I've decided to add this following code piece to the error string on the detoken functions:

error(err .. " in token: " .. token .. "\npossible tokens:\n" .. table.tostring(e, 1), 0)
...
error("Token returned nil, it may not exist: " .. token .. "\npossible tokens:\n" .. table.tostring(e, 1), 0)

suddenly things made sense and I was able to easily fix my script. (hint: cfg didn't even exist in the context I was using it in)

So why didn't I just make a PR you ask?
Because this is a quick dirty hack and could use some work, as you can quickly grasp by looking at the output:

But it's a start, it's better than the cryptic error message you can see in the upper part of the output...

Start by locating the detoken functions and reading how their current error strings expose missing or nil tokens. Reproduce the issue with a debug build and --scripts, then improve the diagnostic output so the failing token and possible tokens are readable; verify the resulting message against the reported cryptic output.