Support positional parameters as integer literals

[viktormalik]

In all places where an integer literal is expected, we now also accept an integer positional parameter.
These are namely:

• builtin function calls (lhist, buf, signal, strncmp)
• array indices.

Fixes #1980.

Note: the only function that is not supported is str, since the solution conflicts with the way str currently handles positional parameters.

Checklist

• Language changes are updated in man/adoc/bpftrace.adoc and if needed in docs/reference_guide.md
• User-visible and non-trivial changes updated in CHANGELOG.md
• The new behaviour is covered by tests

[fbs]

I think we should drop this error, i see every user of this already has its own error handling. Adding an extra error without location on top of that doesn't sound helpful.

Or we could return the error too and append that to the error message, that would be useful but somewhat tricky to do. Std::variant works for it but quite verbose

[viktormalik]

I agree with dropping the error, it doesn't really say much.

Returning the error and appending it to the message could be useful for the other error in this function (line 2215) as it always prints to the default stderr even if the function is called from semantic analyser (which normally prints to its own err_ stream). The question is whether it is worth the effort.