Rule with optional suffix, suffix must end in a specific character

[filippodebortoli]

I am currently trying to encode the following rule using pest (an extremely simplified version of this):

ASCII_ALPHA ((ASCII_ALPHANUMERIC|'.')* ASCII_ALPHANUMERIC)?

My pest rule looks as follows:

test_rule = @{ASCII_ALPHA ~ ((ASCII_ALPHANUMERIC|".")* ~ ASCII_ALPHANUMERIC)? }

My expectation is that MyParser::parse(Rule::test_rule, "test") should execute successfully and consume the entire string.
What is actually happening is that the parser only matches the first character "t".
By poking around the rule, I found out that this is the one part that prevents this parser from matching the whole string:

test_rule = @{ASCII_ALPHA ~ ((ASCII_ALPHANUMERIC|".")* ~ ASCII_ALPHANUMERIC)? }
                                                       ^------------------^

Have I encountered a bug, or is there a specific reason this rule is not working as I am expecting?

I thank you very much for this fantastic crate and the hard work you put into it! :-)

[birkenfeld]

I think the problem here is that matching is greedy and does not backtrack. I.e. all the est are consumed in the (ASCII_ALPHANUMERIC|".")* part and there is nothing left for the final ASCII_ALPHANUMERIC. Therefore, the whole (...)? part fails.

I think this rule should work instead:

test_rule = @{ASCII_ALPHA ~ ("."* ~ ASCII_ALPHANUMERIC)* }

[filippodebortoli]

The rule you are suggesting works and seems to be equivalent to the one I need (after a quick check).
Thank you very much for this nice reply and for explaining why the original rule does not work as intended!

[birkenfeld]

You're welcome! I opened #744 to make this clearer in the docs.