Match strings that start and end with the same character

[matias-santurio]

I'm trying to write a grammar rule that matches strings that start and end with the same character, so a rule that matches:

aba
22
zhelloz

but not

abb
21

My first thought was to write something like

TEST = {PUSH(ANY) ~ ANY* ~ POP }

but of course, matching is greedy so ANY* consumes the whole string, is there a way to get around this?

[Tartasprint]

You can avoid consuming the whole string using predicates:

all_but_last = {(!(ANY ~ EOI) ~ ANY)*}

But I'm not sure that this method is very efficient (I think that for every character the parser checks if the next is EOI). Using that I was able to write the grammar you described, see this grammar on the playground.