Merge branch 'plain_list_perf_investigation'

This is an optimization. When you have something like plain text which ends when it hits the next element, we only need to parse enough to detect that an element is about to occur. For elements like plain lists, this is as simple as parsing a line starting with optional whitespace and then a bullet, which avoids parsing the entire plain list tree. The benefit is most noticeable in deeply nested plain lists.

src/parser/angle_link.rs

@@ -44,7 +44,7 @@ fn path_angle<'r, 's>(
 ) -> Res<OrgSource<'s>, OrgSource<'s>> {
     let parser_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &path_angle_end,
         }));
 

src/parser/citation.rs

@@ -90,7 +90,7 @@ fn global_prefix<'r, 's>(
             depth: 0,
         }))
         .with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &global_prefix_end,
         }));
     let (remaining, (children, _exit_contents)) = verify(
@@ -151,7 +151,7 @@ fn global_suffix<'r, 's>(
             depth: 0,
         }))
         .with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &global_suffix_end,
         }));
     let (remaining, (children, _exit_contents)) = verify(

src/parser/citation_reference.rs

@@ -76,7 +76,7 @@ fn key_prefix<'r, 's>(
             depth: 0,
         }))
         .with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &key_prefix_end,
         }));
     let (remaining, (children, _exit_contents)) = verify(
@@ -101,7 +101,7 @@ fn key_suffix<'r, 's>(
             depth: 0,
         }))
         .with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &key_suffix_end,
         }));
     let (remaining, (children, _exit_contents)) = verify(

src/parser/element_parser.rs

@@ -21,16 +21,19 @@ use super::lesser_block::src_block;
 use super::lesser_block::verse_block;
 use super::org_source::OrgSource;
 use super::paragraph::paragraph;
+use super::plain_list::detect_plain_list;
 use super::plain_list::plain_list;
 use super::source::SetSource;
 use super::util::get_consumed;
 use super::util::maybe_consume_trailing_whitespace_if_not_exiting;
 use super::Context;
+use crate::error::CustomError;
+use crate::error::MyError;
 use crate::error::Res;
 use crate::parser::parser_with_context::parser_with_context;
 use crate::parser::table::org_mode_table;
 
-pub fn element(
+pub const fn element(
     can_be_paragraph: bool,
 ) -> impl for<'r, 's> Fn(Context<'r, 's>, OrgSource<'s>) -> Res<OrgSource<'s>, Element<'s>> {
     move |context: Context, input: OrgSource<'_>| _element(context, input, can_be_paragraph)
@@ -108,3 +111,26 @@ fn _element<'r, 's>(
 
     Ok((remaining, element))
 }
+
+pub const fn detect_element(
+    can_be_paragraph: bool,
+) -> impl for<'r, 's> Fn(Context<'r, 's>, OrgSource<'s>) -> Res<OrgSource<'s>, ()> {
+    move |context: Context, input: OrgSource<'_>| _detect_element(context, input, can_be_paragraph)
+}
+
+#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
+fn _detect_element<'r, 's>(
+    context: Context<'r, 's>,
+    input: OrgSource<'s>,
+    can_be_paragraph: bool,
+) -> Res<OrgSource<'s>, ()> {
+    if detect_plain_list(context, input).is_ok() {
+        return Ok((input, ()));
+    }
+    if _element(context, input, can_be_paragraph).is_ok() {
+        return Ok((input, ()));
+    }
+    return Err(nom::Err::Error(CustomError::MyError(MyError(
+        "No element detected.".into(),
+    ))));
+}

src/parser/exiting.rs

@@ -8,6 +8,9 @@ pub enum ExitClass {
 
     /// Elements who cede priority to alpha elements when matching.
     Beta = 300,
+
+    /// Elements who cede priority to alpha and beta elements when matching.
+    Gamma = 4000,
 }
 
 impl std::fmt::Display for ExitClass {

src/parser/export_snippet.rs

@@ -27,7 +27,7 @@ pub fn export_snippet<'r, 's>(
     let (remaining, backend_name) = backend(context, remaining)?;
     let parser_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &export_snippet_end,
         }));
     let (remaining, backend_contents) = opt(tuple((

src/parser/inline_babel_call.rs

@@ -49,7 +49,7 @@ fn name<'r, 's>(
 ) -> Res<OrgSource<'s>, OrgSource<'s>> {
     let parser_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &name_end,
         }));
     let (remaining, name) = recognize(many_till(
@@ -80,7 +80,7 @@ fn header<'r, 's>(
             depth: 0,
         }))
         .with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &header_end,
         }));
 
@@ -131,7 +131,7 @@ fn argument<'r, 's>(
             depth: 0,
         }))
         .with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &argument_end,
         }));
 

src/parser/lesser_block.rs

@@ -265,13 +265,14 @@ fn _lesser_block_end<'r, 's, 'x>(
     Ok((remaining, source))
 }
 
-fn lesser_block_begin(
+/// Parser for the beginning of a lesser block
-    current_name: &str,
+///
+/// current_name MUST be lowercase. We do not do the conversion ourselves because it is not allowed in a const fn.
+const fn lesser_block_begin(
+    current_name: &'static str,
 ) -> impl for<'r, 's> Fn(Context<'r, 's>, OrgSource<'s>) -> Res<OrgSource<'s>, OrgSource<'s>> {
-    let current_name_lower = current_name.to_lowercase();
+    // TODO: Since this is a const fn, is there ANY way to "generate" functions at compile time?
-    move |context: Context, input: OrgSource<'_>| {
+    move |context: Context, input: OrgSource<'_>| _lesser_block_begin(context, input, current_name)
-        _lesser_block_begin(context, input, current_name_lower.as_str())
-    }
 }
 
 #[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]

src/parser/object_parser.rs

@@ -1,6 +1,5 @@
 use nom::branch::alt;
 use nom::combinator::map;
-use nom::combinator::not;
 
 use super::org_source::OrgSource;
 use super::parser_with_context::parser_with_context;
@@ -34,8 +33,6 @@ pub fn standard_set_object<'r, 's>(
     context: Context<'r, 's>,
     input: OrgSource<'s>,
 ) -> Res<OrgSource<'s>, Object<'s>> {
-    not(|i| context.check_exit_matcher(i))(input)?;
-
     alt((
         map(parser_with_context!(timestamp)(context), Object::Timestamp),
         map(parser_with_context!(subscript)(context), Object::Subscript),
@@ -93,8 +90,6 @@ pub fn minimal_set_object<'r, 's>(
     context: Context<'r, 's>,
     input: OrgSource<'s>,
 ) -> Res<OrgSource<'s>, Object<'s>> {
-    not(|i| context.check_exit_matcher(i))(input)?;
-
     alt((
         map(parser_with_context!(subscript)(context), Object::Subscript),
         map(
@@ -116,7 +111,6 @@ pub fn any_object_except_plain_text<'r, 's>(
     context: Context<'r, 's>,
     input: OrgSource<'s>,
 ) -> Res<OrgSource<'s>, Object<'s>> {
-    // Used for exit matchers so this does not check exit matcher condition.
     alt((
         map(parser_with_context!(timestamp)(context), Object::Timestamp),
         map(parser_with_context!(subscript)(context), Object::Subscript),

src/parser/paragraph.rs

@@ -6,13 +6,13 @@ use nom::multi::many1;
 use nom::multi::many_till;
 use nom::sequence::tuple;
 
+use super::element_parser::detect_element;
 use super::lesser_element::Paragraph;
 use super::org_source::OrgSource;
 use super::util::blank_line;
 use super::util::get_consumed;
 use super::Context;
 use crate::error::Res;
-use crate::parser::element_parser::element;
 use crate::parser::exiting::ExitClass;
 use crate::parser::object_parser::standard_set_object;
 use crate::parser::parser_context::ContextElement;
@@ -28,7 +28,7 @@ pub fn paragraph<'r, 's>(
 ) -> Res<OrgSource<'s>, Paragraph<'s>> {
     let parser_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &paragraph_end,
         }));
     let standard_set_object_matcher = parser_with_context!(standard_set_object)(&parser_context);
@@ -57,7 +57,7 @@ fn paragraph_end<'r, 's>(
     context: Context<'r, 's>,
     input: OrgSource<'s>,
 ) -> Res<OrgSource<'s>, OrgSource<'s>> {
-    let non_paragraph_element_matcher = parser_with_context!(element(false))(context);
+    let non_paragraph_element_matcher = parser_with_context!(detect_element(false))(context);
     alt((
         recognize(tuple((start_of_line, many1(blank_line)))),
         recognize(non_paragraph_element_matcher),

src/parser/parser_context.rs

@@ -62,7 +62,7 @@ impl<'r, 's> ContextTree<'r, 's> {
         //         exit_matcher: ChainBehavior::IgnoreParent(Some(&always_fail)),
         //     }));
 
-        let mut current_class_filter = ExitClass::Beta;
+        let mut current_class_filter = ExitClass::Gamma;
         for current_node in self.iter() {
             let context_element = current_node.get_data();
             match context_element {
@@ -112,9 +112,6 @@ pub enum ContextElement<'r, 's> {
     ExitMatcherNode(ExitMatcherNode<'r>),
     Context(&'r str),
 
-    /// Stores the indentation level of the current list item.
-    ListItem(usize),
-
     /// Stores the name of the greater block.
     GreaterBlock(&'s str),
 

src/parser/plain_link.rs

@@ -113,7 +113,7 @@ fn path_plain<'r, 's>(
     // TODO: "optionally containing parenthesis-wrapped non-whitespace non-bracket substrings up to a depth of two. The string must end with either a non-punctation non-whitespace character, a forwards slash, or a parenthesis-wrapped substring"
     let parser_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &path_plain_end,
         }));
 

src/parser/plain_list.rs

@@ -32,6 +32,27 @@ use crate::parser::util::get_consumed;
 use crate::parser::util::maybe_consume_trailing_whitespace_if_not_exiting;
 use crate::parser::util::start_of_line;
 
+#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
+pub fn detect_plain_list<'r, 's>(
+    _context: Context<'r, 's>,
+    input: OrgSource<'s>,
+) -> Res<OrgSource<'s>, ()> {
+    // TODO: Add support for plain list items that do not have content on the first line.
+    if verify(
+        tuple((start_of_line, space0, bullet, space1)),
+        |(_start, indent, bull, _after_whitespace)| {
+            Into::<&str>::into(bull) != "*" || indent.len() > 0
+        },
+    )(input)
+    .is_ok()
+    {
+        return Ok((input, ()));
+    }
+    return Err(nom::Err::Error(CustomError::MyError(MyError(
+        "No element detected.".into(),
+    ))));
+}
+
 #[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
 pub fn plain_list<'r, 's>(
     context: Context<'r, 's>,
@@ -90,9 +111,6 @@ pub fn plain_list<'r, 's>(
         parser_with_context!(plain_list_item)(&final_item_context)(final_child_start)?;
     children.push((final_child_start, reparsed_final_item));
 
-    let (remaining, _trailing_ws) =
-        maybe_consume_trailing_whitespace_if_not_exiting(context, remaining)?;
-
     let source = get_consumed(input, remaining);
     Ok((
         remaining,
@@ -116,6 +134,7 @@ pub fn plain_list_item<'r, 's>(
         Into::<&str>::into(bull) != "*" || indent_level > 0
     })(remaining)?;
 
+    // TODO: This isn't taking into account items that immediately line break and then have contents
     let maybe_contentless_item: Res<OrgSource<'_>, OrgSource<'_>> =
         alt((eof, line_ending))(remaining);
     match maybe_contentless_item {
@@ -135,12 +154,12 @@ pub fn plain_list_item<'r, 's>(
     };
 
     let (remaining, _ws) = space1(remaining)?;
+    let exit_matcher = plain_list_item_end(indent_level);
     let parser_context = context
         .with_additional_node(ContextElement::ConsumeTrailingWhitespace(true))
-        .with_additional_node(ContextElement::ListItem(indent_level))
         .with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
             class: ExitClass::Beta,
-            exit_matcher: &plain_list_item_end,
+            exit_matcher: &exit_matcher,
         }));
 
     let (remaining, (children, _exit_contents)) = verify(
@@ -194,47 +213,55 @@ fn plain_list_end<'r, 's>(
     )))(input)
 }
 
-#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
+const fn plain_list_item_end(
-fn plain_list_item_end<'r, 's>(
+    indent_level: usize,
+) -> impl for<'r, 's> Fn(Context<'r, 's>, OrgSource<'s>) -> Res<OrgSource<'s>, OrgSource<'s>> {
+    let line_indented_lte_matcher = line_indented_lte(indent_level);
+    move |context: Context, input: OrgSource<'_>| {
+        _plain_list_item_end(context, input, &line_indented_lte_matcher)
+    }
+}
+
+#[cfg_attr(
+    feature = "tracing",
+    tracing::instrument(ret, level = "debug", skip(line_indented_lte_matcher))
+)]
+fn _plain_list_item_end<'r, 's>(
     context: Context<'r, 's>,
     input: OrgSource<'s>,
+    line_indented_lte_matcher: impl for<'rr, 'ss> Fn(
+        Context<'rr, 'ss>,
+        OrgSource<'ss>,
+    ) -> Res<OrgSource<'ss>, OrgSource<'ss>>,
 ) -> Res<OrgSource<'s>, OrgSource<'s>> {
     start_of_line(input)?;
     recognize(tuple((
         opt(blank_line),
-        parser_with_context!(line_indented_lte)(context),
+        parser_with_context!(line_indented_lte_matcher)(context),
     )))(input)
 }
 
-#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
+const fn line_indented_lte(
-fn line_indented_lte<'r, 's>(
+    indent_level: usize,
-    context: Context<'r, 's>,
+) -> impl for<'r, 's> Fn(Context<'r, 's>, OrgSource<'s>) -> Res<OrgSource<'s>, OrgSource<'s>> {
-    input: OrgSource<'s>,
+    move |context: Context, input: OrgSource<'_>| _line_indented_lte(context, input, indent_level)
-) -> Res<OrgSource<'s>, OrgSource<'s>> {
+}
-    let current_item_indent_level: &usize =
-        get_context_item_indent(context).ok_or(nom::Err::Error(CustomError::MyError(MyError(
-            "Not inside a plain list item".into(),
-        ))))?;
 
+#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
+fn _line_indented_lte<'r, 's>(
+    _context: Context<'r, 's>,
+    input: OrgSource<'s>,
+    indent_level: usize,
+) -> Res<OrgSource<'s>, OrgSource<'s>> {
     let matched = recognize(verify(
         tuple((space0::<OrgSource<'_>, _>, 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)
-        |(_space0, _anychar)| _space0.len() <= *current_item_indent_level,
+        |(_space0, _anychar)| _space0.len() <= indent_level,
     ))(input)?;
 
     Ok(matched)
 }
 
-fn get_context_item_indent<'r, 's>(context: Context<'r, 's>) -> Option<&'r usize> {
-    for thing in context.iter() {
-        match thing.get_data() {
-            ContextElement::ListItem(depth) => return Some(depth),
-            _ => {}
-        };
-    }
-    None
-}
-
 #[cfg(test)]
 mod tests {
     use super::*;

src/parser/radio_link.rs

@@ -91,7 +91,7 @@ pub fn radio_target<'r, 's>(
     let (remaining, _opening) = tag("<<<")(input)?;
     let parser_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &radio_target_end,
         }));
 

src/parser/subscript_and_superscript.rs

@@ -162,7 +162,7 @@ fn script_with_braces<'r, 's>(
             },
         ))
         .with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &script_with_braces_end,
         }));
 

src/parser/text_markup.rs

@@ -179,7 +179,7 @@ fn _text_markup_object<'r, 's, 'x>(
     let text_markup_end_specialized = text_markup_end(open.into());
     let parser_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &text_markup_end_specialized,
         }));
 
@@ -229,7 +229,7 @@ fn _text_markup_string<'r, 's, 'x>(
     let text_markup_end_specialized = text_markup_end(open.into());
     let parser_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &text_markup_end_specialized,
         }));
 
@@ -338,7 +338,7 @@ fn _rematch_text_markup_object<'r, 's, 'x>(
     let text_markup_end_specialized = text_markup_end(open.into());
     let parser_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &text_markup_end_specialized,
         }));
 

src/parser/timestamp.rs

@@ -67,7 +67,7 @@ fn sexp<'r, 's>(
 ) -> Res<OrgSource<'s>, OrgSource<'s>> {
     let parser_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &sexp_end,
         }));
 
@@ -99,7 +99,7 @@ fn active_timestamp<'r, 's>(
     let (remaining, _date) = date(context, remaining)?;
     let time_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &active_time_rest_end,
         }));
     let (remaining, _time) =
@@ -132,7 +132,7 @@ fn inactive_timestamp<'r, 's>(
     let (remaining, _date) = date(context, remaining)?;
     let time_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &inactive_time_rest_end,
         }));
     let (remaining, _time) =
@@ -186,12 +186,12 @@ fn active_time_range_timestamp<'r, 's>(
     let (remaining, _date) = date(context, remaining)?;
     let time_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &active_time_rest_end,
         }));
     let first_time_context =
         time_context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &time_range_rest_end,
         }));
     let (remaining, _first_time) =
@@ -247,12 +247,12 @@ fn inactive_time_range_timestamp<'r, 's>(
     let (remaining, _date) = date(context, remaining)?;
     let time_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &inactive_time_rest_end,
         }));
     let first_time_context =
         time_context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &time_range_rest_end,
         }));
     let (remaining, _first_time) =
@@ -303,7 +303,7 @@ fn dayname<'r, 's>(
 ) -> Res<OrgSource<'s>, OrgSource<'s>> {
     let parser_context =
         context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
-            class: ExitClass::Beta,
+            class: ExitClass::Gamma,
             exit_matcher: &dayname_end,
         }));