Support blank link in plain_list_item_end, move exit matcher to end of loop in plain_list, and maybe consume trailing whitespace in plain_list_item.

This commit is contained in:
Tom Alexander 2023-08-16 17:09:06 -04:00
parent 823c33ef8e
commit 56281633f3
Signed by: talexander
GPG Key ID: D3A179C9A53C0EDE

View File

@ -6,6 +6,7 @@ use nom::character::complete::one_of;
use nom::character::complete::space0; use nom::character::complete::space0;
use nom::character::complete::space1; use nom::character::complete::space1;
use nom::combinator::eof; use nom::combinator::eof;
use nom::combinator::opt;
use nom::combinator::peek; use nom::combinator::peek;
use nom::combinator::recognize; use nom::combinator::recognize;
use nom::combinator::verify; use nom::combinator::verify;
@ -30,6 +31,7 @@ use crate::parser::parser_context::ExitMatcherNode;
use crate::parser::util::blank_line; use crate::parser::util::blank_line;
use crate::parser::util::exit_matcher_parser; use crate::parser::util::exit_matcher_parser;
use crate::parser::util::get_consumed; use crate::parser::util::get_consumed;
use crate::parser::util::maybe_consume_trailing_whitespace_if_not_exiting;
use crate::parser::util::start_of_line; use crate::parser::util::start_of_line;
#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))] #[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
@ -45,10 +47,6 @@ pub fn plain_list<'r, 's>(context: Context<'r, 's>, input: &'s str) -> Res<&'s s
let mut remaining = input; let mut remaining = input;
loop { loop {
let maybe_exit = parser_with_context!(exit_matcher_parser)(&parser_context)(remaining);
if maybe_exit.is_ok() {
break;
}
let list_item = parser_with_context!(plain_list_item)(&parser_context)(remaining); let list_item = parser_with_context!(plain_list_item)(&parser_context)(remaining);
match list_item { match list_item {
Ok((remain, item)) Ok((remain, item))
@ -56,12 +54,16 @@ pub fn plain_list<'r, 's>(context: Context<'r, 's>, input: &'s str) -> Res<&'s s
{ {
remaining = remain; remaining = remain;
children.push(item); children.push(item);
continue;
} }
Ok(_) | Err(_) => { Ok(_) | Err(_) => {
break; break;
} }
}; };
let maybe_exit = parser_with_context!(exit_matcher_parser)(&parser_context)(remaining);
if maybe_exit.is_ok() {
break;
}
} }
if children.is_empty() { if children.is_empty() {
@ -121,6 +123,9 @@ pub fn plain_list_item<'r, 's>(
|(children, _exit_contents)| !children.is_empty(), |(children, _exit_contents)| !children.is_empty(),
)(remaining)?; )(remaining)?;
let (remaining, _trailing_ws) =
maybe_consume_trailing_whitespace_if_not_exiting(context, remaining)?;
let source = get_consumed(input, remaining); let source = get_consumed(input, remaining);
return Ok(( return Ok((
remaining, remaining,
@ -159,7 +164,11 @@ fn plain_list_end<'r, 's>(context: Context<'r, 's>, input: &'s str) -> Res<&'s s
#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))] #[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
fn plain_list_item_end<'r, 's>(context: Context<'r, 's>, input: &'s str) -> Res<&'s str, &'s str> { fn plain_list_item_end<'r, 's>(context: Context<'r, 's>, input: &'s str) -> Res<&'s str, &'s str> {
recognize(parser_with_context!(line_indented_lte)(context))(input) start_of_line(context, input)?;
recognize(tuple((
opt(blank_line),
parser_with_context!(line_indented_lte)(context),
)))(input)
} }
#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))] #[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
@ -169,8 +178,6 @@ fn line_indented_lte<'r, 's>(context: Context<'r, 's>, input: &'s str) -> Res<&'
"Not inside a plain list item", "Not inside a plain list item",
))))?; ))))?;
start_of_line(context, input)?;
let matched = recognize(verify( let matched = recognize(verify(
tuple((space0::<&str, _>, non_whitespace_character)), tuple((space0::<&str, _>, non_whitespace_character)),
// It is fine that we get the indent level using the number of bytes rather than the number of characters because nom's space0 only matches space and tab (0x20 and 0x09) // It is fine that we get the indent level using the number of bytes rather than the number of characters because nom's space0 only matches space and tab (0x20 and 0x09)