talexander/organic
1
0
Fork 0
Code Issues Pull Requests Releases Activity

Move context_many_till to its own combinators file.

Browse Source
This commit is contained in:
Tom Alexander 2022-12-15 23:28:09 -05:00
parent c74b802038
commit 0d2dfe3a14
Signed by: talexander
GPG Key ID: D3A179C9A53C0EDE
3 changed files with 116 additions and 71 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

111
src/parser/combinator.rs Normal file
View File

@ -0,0 +1,111 @@
use crate::parser::parser_with_context::parser_with_context;
use crate::parser::text::paragraph_end;
use super::error::CustomError;
use super::error::MyError;
use super::nom_context::ChainBehavior;
use super::nom_context::ContextElement;
use super::nom_context::ContextTree;
use super::nom_context::ExitMatcherNode;
use super::nom_context::PreviousElementNode;
use super::text::bold_end;
use super::text::bold_start;
use super::text::line_break;
use super::text::link_end;
use super::text::link_start;
use super::text::space;
use super::text::span;
use super::text::symbol;
use super::text::Bold;
use super::text::Link;
use super::text::Res;
use super::text::TextElement;
use super::token::Token;
use super::Context;
use nom::branch::alt;
use nom::bytes::complete::tag;
use nom::bytes::complete::take;
use nom::combinator::cond;
use nom::combinator::eof;
use nom::combinator::map;
use nom::combinator::not;
use nom::combinator::peek;
use nom::combinator::recognize;
use nom::error::ErrorKind;
use nom::error::ParseError;
use nom::error::VerboseError;
use nom::multi::many1;
use nom::multi::many_till;
use nom::sequence::tuple;
use nom::IResult;
use nom::InputLength;
pub fn context_many_till<'r, 's, I, O, E, F, M, T>(
context: Context<'r, 's>,
mut many_matcher: M,
mut till_matcher: T,
) -> impl FnMut(I) -> IResult<I, (Vec<Token<'s>>, F), E> + 'r
where
I: Clone + InputLength,
E: ParseError<I>,
M: for<'x> Fn(Context<'x, 's>, I) -> IResult<I, O, E> + 'r,
T: for<'x> Fn(Context<'x, 's>, I) -> IResult<I, F, E> + 'r,
O: Into<Token<'s>>,
{
move |mut i: I| {
// TODO: Can I eliminate the clone?
let mut current_context = context.clone();
// Despite the clone, the Rc should still point to the same value, otherwise we'll get stuck in an endless loop.
assert!(current_context.ptr_eq(context));
loop {
let len = i.input_len();
match till_matcher(&current_context, i.clone()) {
Ok((remaining, finish)) => {
let mut ret = Vec::new();
while !current_context.ptr_eq(context) {
let (context_element, next_context) = current_context.pop_front();
let context_element = context_element.expect("We only pop off context elements created in this function, so they are all Some()");
current_context = next_context;
match context_element {
ContextElement::ExitMatcherNode(_) => {}
ContextElement::StartOfParagraph => {}
ContextElement::Context(_) => {}
ContextElement::PreviousElementNode(PreviousElementNode {
element: token,
}) => {
ret.push(token);
}
};
}
ret.reverse();
return Ok((remaining, (ret, finish)));
}
Err(nom::Err::Error(_)) => {
match many_matcher(&current_context, i.clone()) {
Err(nom::Err::Error(err)) => {
return Err(nom::Err::Error(E::append(i, ErrorKind::ManyTill, err)))
}
Err(e) => return Err(e),
Ok((remaining, many_elem)) => {
// infinite loop check: the parser must always consume
if remaining.input_len() == len {
return Err(nom::Err::Error(E::from_error_kind(
remaining,
ErrorKind::ManyTill,
)));
}
current_context = current_context.with_additional_node(
ContextElement::PreviousElementNode(PreviousElementNode {
element: many_elem.into(),
}),
);
i = remaining;
}
}
}
Err(e) => return Err(e),
};
}
}
}

1
src/parser/mod.rs
View File

@ -1,3 +1,4 @@
mod combinator;
mod error;
mod list;
mod nom_context;

75
src/parser/text_element_parser.rs
View File

@ -2,6 +2,7 @@
use crate::parser::parser_with_context::parser_with_context;
use crate::parser::text::paragraph_end;
use super::combinator::context_many_till;
use super::error::CustomError;
use super::error::MyError;
use super::nom_context::ChainBehavior;
@ -43,76 +44,6 @@ use nom::InputLength;
type UnboundMatcher<'r, 's, I, O, E> = dyn Fn(Context<'r, 's>, I) -> IResult<I, O, E>;
fn context_many_till<'r, 's, I, O, E, F, M, T>(
context: Context<'r, 's>,
mut many_matcher: M,
mut till_matcher: T,
) -> impl FnMut(I) -> IResult<I, (Vec<Token<'s>>, F), E> + 'r
where
I: Clone + InputLength,
E: ParseError<I>,
M: for<'x> Fn(Context<'x, 's>, I) -> IResult<I, O, E> + 'r,
T: for<'x> Fn(Context<'x, 's>, I) -> IResult<I, F, E> + 'r,
O: Into<Token<'s>>,
{
move |mut i: I| {
// TODO: Can I eliminate the clone?
let mut current_context = context.clone();
// Despite the clone, the Rc should still point to the same value, otherwise we'll get stuck in an endless loop.
assert!(current_context.ptr_eq(context));
loop {
let len = i.input_len();
match till_matcher(&current_context, i.clone()) {
Ok((remaining, finish)) => {
let mut ret = Vec::new();
while !current_context.ptr_eq(context) {
let (context_element, next_context) = current_context.pop_front();
let context_element = context_element.expect("We only pop off context elements created in this function, so they are all Some()");
current_context = next_context;
match context_element {
ContextElement::ExitMatcherNode(_) => {}
ContextElement::StartOfParagraph => {}
ContextElement::Context(_) => {}
ContextElement::PreviousElementNode(PreviousElementNode {
element: token,
}) => {
ret.push(token);
}
};
}
ret.reverse();
return Ok((remaining, (ret, finish)));
}
Err(nom::Err::Error(_)) => {
match many_matcher(&current_context, i.clone()) {
Err(nom::Err::Error(err)) => {
return Err(nom::Err::Error(E::append(i, ErrorKind::ManyTill, err)))
}
Err(e) => return Err(e),
Ok((remaining, many_elem)) => {
// infinite loop check: the parser must always consume
if remaining.input_len() == len {
return Err(nom::Err::Error(E::from_error_kind(
remaining,
ErrorKind::ManyTill,
)));
}
current_context = current_context.with_additional_node(
ContextElement::PreviousElementNode(PreviousElementNode {
element: many_elem.into(),
}),
);
i = remaining;
}
}
}
Err(e) => return Err(e),
};
}
}
}
pub fn document(input: &str) -> Res<&str, Vec<(Vec<TextElement>, &str)>> {
let initial_context: ContextTree<'_, '_> = ContextTree::new();
let paragraph_parser = parser_with_context!(paragraph);
@ -184,7 +115,9 @@ pub fn context_bold_start<'s, 'r>(
recognize(bold_start)(input)
} else {
// TODO: Make this a specific error instead of just a generic MyError
return Err(nom::Err::Error(CustomError::MyError(MyError("Cannot start bold"))));
return Err(nom::Err::Error(CustomError::MyError(MyError(
"Cannot start bold",
))));
}
}