Publish version 0.1.9.

Since we only support org-mode tables, type is always org. Value seems to always be nil, not sure why.

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "organic"
-version = "0.1.8"
+version = "0.1.9"
 authors = ["Tom Alexander <tom@fizz.buzz>"]
 description = "An org-mode parser."
 edition = "2021"

org_mode_samples/lesser_element/comment/indented.org

@@ -1,4 +1,5 @@
 # Comment
+#   
   # indented line
 #        At the top of the file
 

src/compare/diff.rs

@@ -86,6 +86,7 @@ use crate::types::Superscript;
 use crate::types::Table;
 use crate::types::TableCell;
 use crate::types::TableRow;
+use crate::types::TableRowType;
 use crate::types::Target;
 use crate::types::Time;
 use crate::types::TimeUnit;
@@ -1172,9 +1173,20 @@ fn compare_footnote_definition<'b, 's>(
 ) -> Result<DiffEntry<'b, 's>, Box<dyn std::error::Error>> {
     let children = emacs.as_list()?;
     let mut child_status = Vec::new();
-    let this_status = DiffStatus::Good;
+    let mut this_status = DiffStatus::Good;
-    let message = None;
+    let mut message = None;
-    // TODO: Compare :label :pre-blank
+    // TODO: Compare :pre-blank
+
+    // Compare label
+    let label = get_property_quoted_string(emacs, ":label")?
+        .ok_or("Footnote definitions should have a name.")?;
+    if label != rust.label {
+        this_status = DiffStatus::Bad;
+        message = Some(format!(
+            "Label mismatch (emacs != rust) {:?} != {:?}",
+            label, rust.label
+        ));
+    }
 
     for (emacs_child, rust_child) in children.iter().skip(2).zip(rust.children.iter()) {
         child_status.push(compare_ast_node(source, emacs_child, rust_child.into())?);
@@ -1197,9 +1209,20 @@ fn compare_comment<'b, 's>(
     rust: &'b Comment<'s>,
 ) -> Result<DiffEntry<'b, 's>, Box<dyn std::error::Error>> {
     let child_status = Vec::new();
-    let this_status = DiffStatus::Good;
+    let mut this_status = DiffStatus::Good;
-    let message = None;
+    let mut message = None;
-    // TODO: Compare :value
+
+    // Compare value
+    let value =
+        get_property_quoted_string(emacs, ":value")?.ok_or("Comments should have a value.")?;
+    let rust_value = rust.get_value();
+    if value != rust_value {
+        this_status = DiffStatus::Bad;
+        message = Some(format!(
+            "Value mismatch (emacs != rust) {:?} != {:?}",
+            value, rust_value
+        ));
+    }
 
     Ok(DiffResult {
         status: this_status,
@@ -1219,9 +1242,19 @@ fn compare_drawer<'b, 's>(
 ) -> Result<DiffEntry<'b, 's>, Box<dyn std::error::Error>> {
     let children = emacs.as_list()?;
     let mut child_status = Vec::new();
-    let this_status = DiffStatus::Good;
+    let mut this_status = DiffStatus::Good;
-    let message = None;
+    let mut message = None;
-    // TODO: Compare :drawer-name
+
+    // Compare drawer-name
+    let name =
+        get_property_quoted_string(emacs, ":drawer-name")?.ok_or("Drawers should have a name.")?;
+    if name != rust.name {
+        this_status = DiffStatus::Bad;
+        message = Some(format!(
+            "Name mismatch (emacs != rust) {:?} != {:?}",
+            name, rust.name
+        ));
+    }
 
     for (emacs_child, rust_child) in children.iter().skip(2).zip(rust.children.iter()) {
         child_status.push(compare_ast_node(source, emacs_child, rust_child.into())?);
@@ -1269,10 +1302,33 @@ fn compare_node_property<'b, 's>(
     rust: &'b NodeProperty<'s>,
 ) -> Result<DiffEntry<'b, 's>, Box<dyn std::error::Error>> {
     let child_status = Vec::new();
-    let this_status = DiffStatus::Good;
+    let mut this_status = DiffStatus::Good;
-    let message = None;
+    let mut message = None;
 
-    // TODO: Compare :key :value
+    // Compare key
+    let key =
+        get_property_quoted_string(emacs, ":key")?.ok_or("Node properties should have a key.")?;
+    if key != rust.name {
+        this_status = DiffStatus::Bad;
+        message = Some(format!(
+            "Key mismatch (emacs != rust) {:?} != {:?}",
+            key, rust.name
+        ));
+    }
+
+    // Compare value
+    let value = get_property_quoted_string(emacs, ":value")?;
+    match (value.as_ref(), rust.value) {
+        (None, None) => {}
+        (Some(emacs_value), Some(rust_value)) if emacs_value == rust_value => {}
+        _ => {
+            this_status = DiffStatus::Bad;
+            message = Some(format!(
+                "Value mismatch (emacs != rust) {:?} != {:?}",
+                value, rust.value
+            ));
+        }
+    }
 
     Ok(DiffResult {
         status: this_status,
@@ -1333,7 +1389,23 @@ fn compare_table<'b, 's>(
         }
     }
 
-    // TODO: Compare :type :value
+    // Compare type
+    let table_type = get_property_unquoted_atom(emacs, ":type")?.expect("Table should have a type");
+    if table_type != "org" {
+        this_status = DiffStatus::Bad;
+        message = Some(format!(
+            "Table type mismatch (emacs != rust) {:?} != {:?}",
+            table_type, "org"
+        ));
+    }
+
+    // Compare value
+    let value = get_property(emacs, ":value")?;
+    if value.is_some() {
+        // I don't know what :value is for, but it seems to always be nil. This is here to alert me to value being non-nil so I can investigate.
+        this_status = DiffStatus::Bad;
+        message = Some(format!("Non-nil value {:?}", value))
+    }
 
     for (emacs_child, rust_child) in children.iter().skip(2).zip(rust.children.iter()) {
         child_status.push(compare_ast_node(source, emacs_child, rust_child.into())?);
@@ -1357,12 +1429,23 @@ fn compare_table_row<'b, 's>(
 ) -> Result<DiffEntry<'b, 's>, Box<dyn std::error::Error>> {
     let children = emacs.as_list()?;
     let mut child_status = Vec::new();
-    let this_status = DiffStatus::Good;
+    let mut this_status = DiffStatus::Good;
-    let message = None;
+    let mut message = None;
 
-    // TODO: Compare :type
+    // Compare type
-    //
+    let row_type = get_property_unquoted_atom(emacs, ":type")?;
-    // :type is an unquoted atom of either standard or rule
+    let rust_row_type = rust.get_type();
+    match (row_type, &rust_row_type) {
+        (Some("standard"), TableRowType::Standard) => {}
+        (Some("rule"), TableRowType::Rule) => {}
+        _ => {
+            this_status = DiffStatus::Bad;
+            message = Some(format!(
+                "Type mismatch (emacs != rust) {:?} != {:?}",
+                row_type, rust_row_type
+            ));
+        }
+    }
 
     for (emacs_child, rust_child) in children.iter().skip(2).zip(rust.children.iter()) {
         child_status.push(compare_ast_node(source, emacs_child, rust_child.into())?);

src/parser/comment.rs

@@ -1,18 +1,19 @@
 use nom::branch::alt;
-use nom::bytes::complete::is_not;
 use nom::bytes::complete::tag;
+use nom::character::complete::anychar;
 use nom::character::complete::line_ending;
 use nom::character::complete::space0;
-use nom::character::complete::space1;
 use nom::combinator::eof;
 use nom::combinator::not;
-use nom::combinator::opt;
+use nom::combinator::recognize;
 use nom::multi::many0;
+use nom::multi::many_till;
 use nom::sequence::preceded;
 use nom::sequence::tuple;
 
 use super::org_source::OrgSource;
 use super::util::get_consumed;
+use super::util::org_line_ending;
 use crate::context::parser_with_context;
 use crate::context::ContextElement;
 use crate::context::RefContext;
@@ -38,15 +39,29 @@ pub(crate) fn comment<'b, 'g, 'r, 's>(
     let parser_context = context.with_additional_node(&parser_context);
     let comment_line_matcher = parser_with_context!(comment_line)(&parser_context);
     let exit_matcher = parser_with_context!(exit_matcher_parser)(&parser_context);
-    let (remaining, _first_line) = comment_line_matcher(input)?;
+    let (remaining, first_line) = comment_line_matcher(input)?;
-    let (remaining, _remaining_lines) =
+    let (remaining, mut remaining_lines) =
         many0(preceded(not(exit_matcher), comment_line_matcher))(remaining)?;
 
     let source = get_consumed(input, remaining);
+    let mut value = Vec::with_capacity(remaining_lines.len() + 1);
+    let last_line = remaining_lines.pop();
+    if let Some(last_line) = last_line {
+        value.push(Into::<&str>::into(first_line));
+        value.extend(remaining_lines.into_iter().map(Into::<&str>::into));
+        let last_line = Into::<&str>::into(last_line);
+        // Trim the line ending from the final line.
+        value.push(&last_line[..(last_line.len() - 1)])
+    } else {
+        // Trim the line ending from the only line.
+        let only_line = Into::<&str>::into(first_line);
+        value.push(&only_line[..(only_line.len() - 1)])
+    }
     Ok((
         remaining,
         Comment {
             source: source.into(),
+            value,
         },
     ))
 }
@@ -57,14 +72,13 @@ fn comment_line<'b, 'g, 'r, 's>(
     input: OrgSource<'s>,
 ) -> Res<OrgSource<'s>, OrgSource<'s>> {
     start_of_line(input)?;
-    let (remaining, _indent) = space0(input)?;
+    let (remaining, _) = tuple((space0, tag("#")))(input)?;
-    let (remaining, (_hash, _leading_whitespace_and_content, _line_ending)) = tuple((
+    if let Ok((remaining, line_break)) = org_line_ending(remaining) {
-        tag("#"),
+        return Ok((remaining, line_break));
-        opt(tuple((space1, is_not("\r\n")))),
+    }
-        alt((line_ending, eof)),
+    let (remaining, _) = tag(" ")(remaining)?;
-    ))(remaining)?;
+    let (remaining, value) = recognize(many_till(anychar, org_line_ending))(remaining)?;
-    let source = get_consumed(input, remaining);
+    Ok((remaining, value))
-    Ok((remaining, source))
 }
 
 #[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]

src/parser/property_drawer.rs

@@ -101,7 +101,7 @@ fn node_property<'b, 'g, 'r, 's>(
     context: RefContext<'b, 'g, 'r, 's>,
     input: OrgSource<'s>,
 ) -> Res<OrgSource<'s>, NodeProperty<'s>> {
-    let (remaining, (_start_of_line, _leading_whitespace, _open_colon, _name, _close_colon)) =
+    let (remaining, (_start_of_line, _leading_whitespace, _open_colon, name, _close_colon)) =
         tuple((
             start_of_line,
             space0,
@@ -120,6 +120,7 @@ fn node_property<'b, 'g, 'r, 's>(
                 remaining,
                 NodeProperty {
                     source: source.into(),
+                    name: Into::<&str>::into(name),
                     value: None,
                 },
             ))
@@ -132,6 +133,7 @@ fn node_property<'b, 'g, 'r, 's>(
                 remaining,
                 NodeProperty {
                     source: source.into(),
+                    name: Into::<&str>::into(name),
                     value: Some(value.into()),
                 },
             ))

src/parser/table.rs

@@ -3,8 +3,8 @@ use nom::bytes::complete::is_not;
 use nom::bytes::complete::tag;
 use nom::character::complete::line_ending;
 use nom::character::complete::space0;
-use nom::combinator::eof;
 use nom::combinator::not;
+use nom::combinator::opt;
 use nom::combinator::peek;
 use nom::combinator::recognize;
 use nom::combinator::verify;
@@ -17,6 +17,7 @@ use super::keyword::table_formula_keyword;
 use super::object_parser::table_cell_set_object;
 use super::org_source::OrgSource;
 use super::util::exit_matcher_parser;
+use super::util::org_line_ending;
 use crate::context::parser_with_context;
 use crate::context::ContextElement;
 use crate::context::ExitClass;
@@ -105,7 +106,7 @@ fn org_mode_table_row_rule<'b, 'g, 'r, 's>(
     input: OrgSource<'s>,
 ) -> Res<OrgSource<'s>, TableRow<'s>> {
     start_of_line(input)?;
-    let (remaining, _) = tuple((space0, tag("|-"), is_not("\r\n"), line_ending))(input)?;
+    let (remaining, _) = tuple((space0, tag("|-"), opt(is_not("\r\n")), org_line_ending))(input)?;
     let source = get_consumed(input, remaining);
     Ok((
         remaining,
@@ -125,7 +126,7 @@ fn org_mode_table_row_regular<'b, 'g, 'r, 's>(
     let (remaining, _) = tuple((space0, tag("|")))(input)?;
     let (remaining, children) =
         many1(parser_with_context!(org_mode_table_cell)(context))(remaining)?;
-    let (remaining, _tail) = recognize(tuple((space0, alt((line_ending, eof)))))(remaining)?;
+    let (remaining, _tail) = recognize(tuple((space0, org_line_ending)))(remaining)?;
     let source = get_consumed(input, remaining);
     Ok((
         remaining,

src/types/greater_element.rs

@@ -94,6 +94,7 @@ pub struct PropertyDrawer<'s> {
 #[derive(Debug)]
 pub struct NodeProperty<'s> {
     pub source: &'s str,
+    pub name: &'s str,
     pub value: Option<&'s str>,
 }
 
@@ -110,6 +111,12 @@ pub struct TableRow<'s> {
     pub children: Vec<TableCell<'s>>,
 }
 
+#[derive(Debug)]
+pub enum TableRowType {
+    Standard,
+    Rule,
+}
+
 impl<'s> StandardProperties<'s> for PlainList<'s> {
     fn get_source<'b>(&'b self) -> &'s str {
         self.source
@@ -193,3 +200,13 @@ impl<'s> PlainListItem<'s> {
             .map(|(checkbox_type, _)| checkbox_type)
     }
 }
+
+impl<'s> TableRow<'s> {
+    pub fn get_type(&self) -> TableRowType {
+        if self.children.is_empty() {
+            TableRowType::Rule
+        } else {
+            TableRowType::Standard
+        }
+    }
+}

src/types/lesser_element.rs

@@ -12,6 +12,7 @@ pub struct Paragraph<'s> {
 #[derive(Debug)]
 pub struct Comment<'s> {
     pub source: &'s str,
+    pub value: Vec<&'s str>,
 }
 
 #[derive(Debug)]
@@ -209,3 +210,16 @@ impl<'s> StandardProperties<'s> for LatexEnvironment<'s> {
         self.source
     }
 }
+
+impl<'s> Comment<'s> {
+    pub fn get_value(&self) -> String {
+        // TODO: maybe we should handle parsing here instead of storing the parsing result in the AST since I imagine getting the value of comments won't be a common operation.
+        let final_size = self.value.iter().map(|line| line.len()).sum();
+        let mut ret = String::with_capacity(final_size);
+        for line in &self.value {
+            ret.push_str(line);
+        }
+
+        ret
+    }
+}

src/types/mod.rs

@@ -35,6 +35,7 @@ pub use greater_element::QuoteBlock;
 pub use greater_element::SpecialBlock;
 pub use greater_element::Table;
 pub use greater_element::TableRow;
+pub use greater_element::TableRowType;
 pub use lesser_element::BabelCall;
 pub use lesser_element::Clock;
 pub use lesser_element::Comment;