Suppress "never useful" precedence warning in modular parser specifications?

Question

Is it possible to suppress unused precedence warnings in menhir?

Background:

I have a core parser Lib.mly with several rules and, separately, a host of additional parsers (A.mly, B.mly, ...) which use definitions from Lib.mly. To support using a single lexer for all the languages, all tokens are defined in Lib.mly and we use external_tokens(Lib) annotations in _tags and have extended myocamlbuild.ml to suppress unused tokens warnings with --unused-tokens which helps significantly. However, I still get many warnings of the form:

File "parsers/ParserLib.mly", line 126, characters 0-9:                     
Warning: the precedence level assigned to FIX is never useful.              

These make it hard to see other warnings I actually want to address. Different subsets of the operators are used in various languages, so I need to have a single "global" precedence that orders them appropriately. Thanks for any tips!

Mach

Answer 1

As of today, Menhir has a new option --unused-precedence-levels.

If you have examples of grammars that are split over several files, I might be interested to look at them (and possibly include some of them in Menhir's test suite, if permitted).

Turboprop

This works perfectly! Thank you very much for the support and for making menhir such a stellar tool. Our parsers are not yet public, but I will check with my colleagues and send you a tarball. – Mach

Answer 2

Not sure if this is what you're searching for but the two following options may be useful :

--unused-token <token>           Do not warn that <token> is unused
--unused-tokens                  Do not warn about any unused token

Samarskite

Yes, those have been helpful for avoiding the unused token warnings. Unfortunately, they don't seem to turn off the "precedence never useful" warnings for said unused tokens :\ I think the next step is filing a menhir issue and just using awk to filter unwanted warnings for now, but ugh. – Mach