can't match ANY.

[shi-yan]

trying to create a parser for plantuml, ANY doesn't match alphabets,

https://plantuml.com/sequence-diagram

sample code:
test.pest

WHITESPACE = _{ " " | "\t" }
startuml_sequence = {"@startuml" ~ "sequence" ~ NEWLINE}
enduml = {"@enduml"}
sequence_content = {ANY+ ~ NEWLINE}
sequence = {SOI ~ startuml_sequence ~ sequence_content ~ enduml ~ EOI}

main.rs

extern crate pest;
#[macro_use]
extern crate pest_derive;

use pest::Parser;

#[derive(Parser)]
#[grammar = "test.pest"]
struct IdentParser;

fn main() {
    let pairs = IdentParser::parse(Rule::sequence, "@startuml sequence
    test
    @enduml").unwrap_or_else(|e| panic!("{}", e));

    for pair in pairs {
        println!("Rule:    {:?}", pair.as_rule());
        println!("Span:    {:?}", pair.as_span());
        println!("Text:    {}", pair.as_str());

        for inner_pair in pair.into_inner() {
            println!("Rule:    {:?}", inner_pair.as_rule());
        }
    }
}

output

thread 'main' panicked at ' --> 2:5
  |
2 |     test␊
  |     ^---
  |
  = expected sequence_content', src/main.rs:14:34

if I change ANY to ASCII_ALPHA, it would work. I don't understand. looks like a bug to me.

[CAD97]

ANY+ eats the entire input, thus cannot be followed by NEWLINE (or anything else other than EOI).

PEG parsing is eager, and each match eats the maximal much, without any backtracking within a rule.

You likely want to match {everything other than NEWLINE} rather than truly ANY as the head of sequence_content.

[shi-yan]

Thank you, I figured out too. but I later tried

(!NEWLINE)+

            |
          6 | sequence_content = { (!NEWLINE)+ }␊
            |                      ^---------^
            |
            = expression inside repetition is non-progressing and will repeat infinitely

and !NEWLINE+

thread 'main' panicked at ' --> 1:1
  |
1 | @startuml sequence␊
  | ^---
  |
  = expected sequence', src/main.rs:14:34

[birkenfeld]

See https://pest.rs/book/grammars/syntax.html#predicates - !NEWLINE does not consume anything but is like a lookahead assertion in regexes, so you need (!NEWLINE ~ ANY)+.