Remove wasm_test's dependency on compare module.

src/util/elisp/mod.rs

@@ -0,0 +1,10 @@
+mod sexp;
+mod util;
+
+pub use sexp::sexp;
+pub(crate) use sexp::unquote;
+pub(crate) use sexp::TextWithProperties;
+pub use sexp::Token;
+pub(crate) use util::get_emacs_standard_properties;
+pub(crate) use util::maybe_token_to_usize;
+pub(crate) use util::EmacsStandardProperties;

src/util/elisp/sexp.rs

@@ -0,0 +1,497 @@
+use std::borrow::Cow;
+use std::collections::HashMap;
+
+use nom::branch::alt;
+use nom::bytes::complete::tag;
+use nom::bytes::complete::take_till1;
+use nom::character::complete::anychar;
+use nom::character::complete::digit1;
+use nom::character::complete::multispace0;
+use nom::character::complete::multispace1;
+use nom::character::complete::one_of;
+use nom::combinator::map;
+use nom::combinator::not;
+use nom::combinator::opt;
+use nom::combinator::peek;
+use nom::combinator::recognize;
+use nom::multi::separated_list1;
+use nom::sequence::delimited;
+use nom::sequence::preceded;
+use nom::sequence::tuple;
+
+use crate::error::Res;
+
+const MAX_OCTAL_LENGTH: usize = 3;
+
+#[derive(Debug)]
+pub enum Token<'s> {
+    Atom(&'s str),
+    List(Vec<Token<'s>>),
+    TextWithProperties(TextWithProperties<'s>),
+    Vector(Vec<Token<'s>>),
+}
+
+#[derive(Debug)]
+pub struct TextWithProperties<'s> {
+    pub(crate) text: &'s str,
+    pub(crate) properties: Vec<Token<'s>>,
+}
+
+impl<'s> Token<'s> {
+    pub(crate) fn as_vector<'p>(
+        &'p self,
+    ) -> Result<&'p Vec<Token<'s>>, Box<dyn std::error::Error>> {
+        Ok(match self {
+            Token::Vector(children) => Ok(children),
+            _ => Err(format!("wrong token type, expected vector: {:?}", self)),
+        }?)
+    }
+
+    pub(crate) fn as_list<'p>(&'p self) -> Result<&'p Vec<Token<'s>>, Box<dyn std::error::Error>> {
+        Ok(match self {
+            Token::List(children) => Ok(children),
+            _ => Err(format!("wrong token type, expected list: {:?}", self)),
+        }?)
+    }
+
+    pub(crate) fn as_atom<'p>(&'p self) -> Result<&'s str, Box<dyn std::error::Error>> {
+        Ok(match self {
+            Token::Atom(body) => Ok(*body),
+            _ => Err(format!("wrong token type, expected atom: {:?}", self)),
+        }?)
+    }
+
+    pub(crate) fn as_text<'p>(
+        &'p self,
+    ) -> Result<&'p TextWithProperties<'s>, Box<dyn std::error::Error>> {
+        Ok(match self {
+            Token::TextWithProperties(body) => Ok(body),
+            _ => Err(format!("wrong token type, expected text: {:?}", self)),
+        }?)
+    }
+
+    pub(crate) fn as_map<'p>(
+        &'p self,
+    ) -> Result<HashMap<&'s str, &'p Token<'s>>, Box<dyn std::error::Error>> {
+        let mut hashmap = HashMap::new();
+
+        let children = self.as_list()?;
+        if children.len() % 2 != 0 {
+            return Err("Expecting an even number of children".into());
+        }
+        let mut key: Option<&str> = None;
+        for child in children.iter() {
+            match key {
+                None => {
+                    key = Some(child.as_atom()?);
+                }
+                Some(key_val) => {
+                    key = None;
+                    hashmap.insert(key_val, child);
+                }
+            };
+        }
+
+        Ok(hashmap)
+    }
+}
+
+/// Check if the child string slice is a slice of the parent string slice.
+fn is_slice_of(parent: &str, child: &str) -> bool {
+    let parent_start = parent.as_ptr() as usize;
+    let parent_end = parent_start + parent.len();
+    let child_start = child.as_ptr() as usize;
+    let child_end = child_start + child.len();
+    child_start >= parent_start && child_end <= parent_end
+}
+
+/// Get a slice of the string that was consumed in a parser using the original input to the parser and the remaining input after the parser.
+fn get_consumed<'s>(input: &'s str, remaining: &'s str) -> &'s str {
+    debug_assert!(is_slice_of(input, remaining));
+    let offset = remaining.as_ptr() as usize - input.as_ptr() as usize;
+    &input[..offset]
+}
+
+#[derive(Debug)]
+enum UnquoteState {
+    Normal,
+    Escape,
+    HasEscape {
+        out: Vec<u8>,
+    },
+    HasEscapeEscape {
+        out: Vec<u8>,
+    },
+    Octal {
+        octal_begin_offset: usize,
+        octal: Vec<u8>,
+    },
+    HasEscapeOctal {
+        out: Vec<u8>,
+        octal: Vec<u8>,
+    },
+}
+
+pub(crate) fn unquote(text: &str) -> Result<Cow<'_, str>, Box<dyn std::error::Error>> {
+    if !text.starts_with('"') {
+        return Err("Quoted text does not start with quote.".into());
+    }
+    if !text.ends_with('"') {
+        return Err("Quoted text does not end with quote.".into());
+    }
+    let interior_text = &text[1..(text.len() - 1)];
+    let mut state = UnquoteState::Normal;
+    for (offset, current_char) in interior_text.bytes().enumerate() {
+        // Check to see if octal finished
+        state = match (state, current_char) {
+            (
+                UnquoteState::Octal {
+                    octal_begin_offset,
+                    octal,
+                },
+                b'0'..=b'7',
+            ) if octal.len() < MAX_OCTAL_LENGTH => UnquoteState::Octal {
+                octal_begin_offset,
+                octal,
+            },
+            (
+                UnquoteState::Octal {
+                    octal_begin_offset,
+                    octal,
+                },
+                _,
+            ) => {
+                let octal_number_string = String::from_utf8(octal)?;
+                let decoded_byte = u8::from_str_radix(&octal_number_string, 8)?;
+                let mut out: Vec<u8> = Vec::with_capacity(interior_text.len());
+                out.extend_from_slice(&interior_text.as_bytes()[..octal_begin_offset]);
+                out.push(decoded_byte);
+                UnquoteState::HasEscape { out }
+            }
+            (UnquoteState::HasEscapeOctal { out, octal }, b'0'..=b'7')
+                if octal.len() < MAX_OCTAL_LENGTH =>
+            {
+                UnquoteState::HasEscapeOctal { out, octal }
+            }
+            (UnquoteState::HasEscapeOctal { mut out, octal }, _) => {
+                let octal_number_string = String::from_utf8(octal)?;
+                let decoded_byte = u8::from_str_radix(&octal_number_string, 8)?;
+                out.push(decoded_byte);
+                UnquoteState::HasEscape { out }
+            }
+            (state, _) => state,
+        };
+
+        state = match (state, current_char) {
+            (UnquoteState::Normal, b'\\') => UnquoteState::Escape,
+            (UnquoteState::Normal, _) => UnquoteState::Normal,
+            (UnquoteState::HasEscape { out }, b'\\') => UnquoteState::HasEscapeEscape { out },
+            (UnquoteState::HasEscape { mut out }, _) => {
+                out.push(current_char);
+                UnquoteState::HasEscape { out }
+            }
+            (UnquoteState::Escape, b'n') => {
+                let mut out: Vec<u8> = Vec::with_capacity(interior_text.len());
+                // Subtract 1 from offset to account for backslash.
+                out.extend_from_slice(&interior_text.as_bytes()[..(offset - 1)]);
+                out.push(b'\n');
+                UnquoteState::HasEscape { out }
+            }
+            (UnquoteState::HasEscapeEscape { mut out }, b'n') => {
+                out.push(b'\n');
+                UnquoteState::HasEscape { out }
+            }
+            (UnquoteState::Escape, b'\\') => {
+                let mut out: Vec<u8> = Vec::with_capacity(interior_text.len());
+                // Subtract 1 from offset to account for backslash.
+                out.extend_from_slice(&interior_text.as_bytes()[..(offset - 1)]);
+                out.push(b'\\');
+                UnquoteState::HasEscape { out }
+            }
+            (UnquoteState::HasEscapeEscape { mut out }, b'\\') => {
+                out.push(b'\\');
+                UnquoteState::HasEscape { out }
+            }
+            (UnquoteState::Escape, b'"') => {
+                let mut out: Vec<u8> = Vec::with_capacity(interior_text.len());
+                // Subtract 1 from offset to account for backslash.
+                out.extend_from_slice(&interior_text.as_bytes()[..(offset - 1)]);
+                out.push(b'"');
+                UnquoteState::HasEscape { out }
+            }
+            (UnquoteState::HasEscapeEscape { mut out }, b'"') => {
+                out.push(b'"');
+                UnquoteState::HasEscape { out }
+            }
+            (UnquoteState::Escape, b'0'..=b'7') => {
+                let mut octal = Vec::with_capacity(MAX_OCTAL_LENGTH);
+                octal.push(current_char);
+                // Substract 1 from offset to account for backslash
+                UnquoteState::Octal {
+                    octal_begin_offset: offset - 1,
+                    octal,
+                }
+            }
+            (UnquoteState::HasEscapeEscape { out }, b'0'..=b'7') => {
+                let mut octal = Vec::with_capacity(MAX_OCTAL_LENGTH);
+                octal.push(current_char);
+                // Substract 1 from offset to account for backslash
+                UnquoteState::HasEscapeOctal { out, octal }
+            }
+            (
+                UnquoteState::Octal {
+                    octal_begin_offset,
+                    mut octal,
+                },
+                b'0'..=b'7',
+            ) => {
+                octal.push(current_char);
+                UnquoteState::Octal {
+                    octal_begin_offset,
+                    octal,
+                }
+            }
+            (UnquoteState::HasEscapeOctal { out, mut octal }, b'0'..=b'7') => {
+                octal.push(current_char);
+                UnquoteState::HasEscapeOctal { out, octal }
+            }
+            (state, _) => panic!(
+                "Invalid state unquoting string: {:?} | {} | {:?}",
+                state, offset, interior_text
+            ),
+        };
+    }
+
+    match state {
+        UnquoteState::Normal | UnquoteState::Escape | UnquoteState::Octal { .. } => {
+            Ok(Cow::Borrowed(interior_text))
+        }
+        UnquoteState::HasEscape { out } => Ok(Cow::Owned(String::from_utf8(out)?)),
+        UnquoteState::HasEscapeEscape { mut out } => {
+            out.push(b'\\');
+            Ok(Cow::Owned(String::from_utf8(out)?))
+        }
+        UnquoteState::HasEscapeOctal { mut out, octal } => {
+            out.push(b'\\');
+            out.extend(octal);
+            Ok(Cow::Owned(String::from_utf8(out)?))
+        }
+    }
+}
+
+#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
+pub fn sexp<'s>(input: &'s str) -> Res<&'s str, Token<'s>> {
+    let (remaining, _) = multispace0(input)?;
+    let (remaining, tkn) = token(remaining).map(|(rem, out)| (Into::<&str>::into(rem), out))?;
+    let (remaining, _) = multispace0(remaining)?;
+    Ok((remaining, tkn))
+}
+
+#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
+fn token<'s>(input: &'s str) -> Res<&'s str, Token<'s>> {
+    alt((list, vector, atom))(input)
+}
+
+#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
+fn list<'s>(input: &'s str) -> Res<&'s str, Token<'s>> {
+    let (remaining, _) = tag("(")(input)?;
+    let (remaining, children) = delimited(
+        multispace0,
+        separated_list1(multispace1, token),
+        multispace0,
+    )(remaining)?;
+    let (remaining, _) = tag(")")(remaining)?;
+    Ok((remaining, Token::List(children)))
+}
+
+#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
+fn vector<'s>(input: &'s str) -> Res<&'s str, Token<'s>> {
+    let (remaining, _) = tag("[")(input)?;
+    let (remaining, children) = delimited(
+        multispace0,
+        separated_list1(multispace1, token),
+        multispace0,
+    )(remaining)?;
+    let (remaining, _) = tag("]")(remaining)?;
+    Ok((remaining, Token::Vector(children)))
+}
+
+#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
+fn atom<'s>(input: &'s str) -> Res<&'s str, Token<'s>> {
+    not(peek(one_of(")]")))(input)?;
+    alt((
+        text_with_properties,
+        hash_notation,
+        quoted_atom,
+        unquoted_atom,
+    ))(input)
+}
+
+#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
+fn unquoted_atom<'s>(input: &'s str) -> Res<&'s str, Token<'s>> {
+    let (remaining, body) =
+        take_till1(|c| matches!(c, ' ' | '\t' | '\r' | '\n' | ')' | ']'))(input)?;
+    Ok((remaining, Token::Atom(body)))
+}
+
+#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
+fn quoted_atom<'s>(input: &'s str) -> Res<&'s str, Token<'s>> {
+    let (mut remaining, _) = tag(r#"""#)(input)?;
+    let mut in_escape = false;
+    loop {
+        if in_escape {
+            let (remain, _) = alt((recognize(one_of(r#""n\\"#)), digit1))(remaining)?;
+            remaining = remain;
+            in_escape = false;
+        } else {
+            let end_quote = tag::<_, _, nom::error::Error<_>>(r#"""#)(remaining);
+            if end_quote.is_ok() {
+                break;
+            }
+
+            let escape_backslash = tag::<_, _, nom::error::Error<_>>("\\")(remaining);
+            if let Ok((remain, _)) = escape_backslash {
+                remaining = remain;
+                in_escape = true;
+                continue;
+            }
+
+            let (remain, _) = anychar(remaining)?;
+            remaining = remain;
+        }
+    }
+    let (remaining, _) = tag(r#"""#)(remaining)?;
+    let source = get_consumed(input, remaining);
+    Ok((remaining, Token::Atom(source)))
+}
+
+#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
+fn hash_notation<'s>(input: &'s str) -> Res<&'s str, Token<'s>> {
+    let (remaining, _) = tag("#<")(input)?;
+    let (remaining, _body) = take_till1(|c| matches!(c, '>'))(remaining)?;
+    let (remaining, _) = tag(">")(remaining)?;
+    let source = get_consumed(input, remaining);
+    Ok((remaining, Token::Atom(source)))
+}
+
+fn text_with_properties(input: &str) -> Res<&str, Token<'_>> {
+    let (remaining, _) = tag("#(")(input)?;
+    let (remaining, (text, props)) = delimited(
+        multispace0,
+        tuple((
+            map(quoted_atom, |atom| match atom {
+                Token::Atom(body) => body,
+                _ => unreachable!(),
+            }),
+            preceded(multispace1, opt(separated_list1(multispace1, token))),
+        )),
+        multispace0,
+    )(remaining)?;
+    let (remaining, _) = tag(")")(remaining)?;
+    Ok((
+        remaining,
+        Token::TextWithProperties(TextWithProperties {
+            text,
+            properties: props.unwrap_or(Vec::new()),
+        }),
+    ))
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn simple() {
+        let input = "  (foo bar baz )  ";
+        let (remaining, parsed) = sexp(input).expect("Parse the input");
+        assert_eq!(remaining, "");
+        assert!(match parsed {
+            Token::Atom(_) => false,
+            Token::List(_) => true,
+            Token::TextWithProperties(_) => false,
+            Token::Vector(_) => false,
+        });
+    }
+
+    #[test]
+    fn quoted() {
+        let input = r#"  ("foo" bar baz )  "#;
+        let (remaining, parsed) = sexp(input).expect("Parse the input");
+        assert_eq!(remaining, "");
+        assert!(match parsed {
+            Token::Atom(_) => false,
+            Token::List(_) => true,
+            Token::TextWithProperties(_) => false,
+            Token::Vector(_) => false,
+        });
+        let children = match parsed {
+            Token::List(children) => children,
+            _ => panic!("Should be a list."),
+        };
+        assert_eq!(
+            match children.first() {
+                Some(Token::Atom(body)) => *body,
+                _ => panic!("First child should be an atom."),
+            },
+            r#""foo""#
+        )
+    }
+
+    #[test]
+    fn quoted_containing_paren() {
+        let input = r#"  (foo "b(a)r" baz )  "#;
+        let (remaining, parsed) = sexp(input).expect("Parse the input");
+        assert_eq!(remaining, "");
+        assert!(matches!(parsed, Token::List(_)));
+        let children = match parsed {
+            Token::List(children) => children,
+            _ => panic!("Should be a list."),
+        };
+        assert_eq!(
+            match children.first() {
+                Some(Token::Atom(body)) => *body,
+                _ => panic!("First child should be an atom."),
+            },
+            r#"foo"#
+        );
+        assert_eq!(
+            match children.get(1) {
+                Some(Token::Atom(body)) => *body,
+                _ => panic!("Second child should be an atom."),
+            },
+            r#""b(a)r""#
+        );
+        assert_eq!(
+            match children.get(2) {
+                Some(Token::Atom(body)) => *body,
+                _ => panic!("Third child should be an atom."),
+            },
+            r#"baz"#
+        );
+    }
+
+    #[test]
+    fn string_containing_escaped_characters() {
+        let input = r#"  (foo "\\( x=2 \\)" bar)  "#;
+        let (remaining, parsed) = sexp(input).expect("Parse the input");
+        assert_eq!(remaining, "");
+        assert!(match parsed {
+            Token::Atom(_) => false,
+            Token::List(_) => true,
+            Token::TextWithProperties(_) => false,
+            Token::Vector(_) => false,
+        });
+        let children = match parsed {
+            Token::List(children) => children,
+            _ => panic!("Should be a list."),
+        };
+        assert_eq!(
+            match children.get(1) {
+                Some(Token::Atom(body)) => *body,
+                _ => panic!("First child should be an atom."),
+            },
+            r#""\\( x=2 \\)""#
+        )
+    }
+}

src/util/elisp/util.rs

@@ -0,0 +1,75 @@
+use super::Token;
+
+pub(crate) fn maybe_token_to_usize(
+    token: Option<&Token<'_>>,
+) -> Result<Option<usize>, Box<dyn std::error::Error>> {
+    Ok(token
+        .map(|token| token.as_atom())
+        .map_or(Ok(None), |r| r.map(Some))?
+        .and_then(|val| {
+            if val == "nil" {
+                None
+            } else {
+                Some(val.parse::<usize>())
+            }
+        })
+        .map_or(Ok(None), |r| r.map(Some))?)
+}
+
+pub(crate) struct EmacsStandardProperties {
+    pub(crate) begin: Option<usize>,
+    #[allow(dead_code)]
+    pub(crate) post_affiliated: Option<usize>,
+    #[allow(dead_code)]
+    pub(crate) contents_begin: Option<usize>,
+    #[allow(dead_code)]
+    pub(crate) contents_end: Option<usize>,
+    pub(crate) end: Option<usize>,
+    #[allow(dead_code)]
+    pub(crate) post_blank: Option<usize>,
+}
+
+pub(crate) fn get_emacs_standard_properties(
+    emacs: &Token<'_>,
+) -> Result<EmacsStandardProperties, Box<dyn std::error::Error>> {
+    let children = emacs.as_list()?;
+    let attributes_child = children.get(1).ok_or("Should have an attributes child.")?;
+    let attributes_map = attributes_child.as_map()?;
+    let standard_properties = attributes_map.get(":standard-properties");
+    Ok(if standard_properties.is_some() {
+        let mut std_props = standard_properties
+            .expect("if statement proves its Some")
+            .as_vector()?
+            .iter();
+        let begin = maybe_token_to_usize(std_props.next())?;
+        let post_affiliated = maybe_token_to_usize(std_props.next())?;
+        let contents_begin = maybe_token_to_usize(std_props.next())?;
+        let contents_end = maybe_token_to_usize(std_props.next())?;
+        let end = maybe_token_to_usize(std_props.next())?;
+        let post_blank = maybe_token_to_usize(std_props.next())?;
+        EmacsStandardProperties {
+            begin,
+            post_affiliated,
+            contents_begin,
+            contents_end,
+            end,
+            post_blank,
+        }
+    } else {
+        let begin = maybe_token_to_usize(attributes_map.get(":begin").copied())?;
+        let end = maybe_token_to_usize(attributes_map.get(":end").copied())?;
+        let contents_begin = maybe_token_to_usize(attributes_map.get(":contents-begin").copied())?;
+        let contents_end = maybe_token_to_usize(attributes_map.get(":contents-end").copied())?;
+        let post_blank = maybe_token_to_usize(attributes_map.get(":post-blank").copied())?;
+        let post_affiliated =
+            maybe_token_to_usize(attributes_map.get(":post-affiliated").copied())?;
+        EmacsStandardProperties {
+            begin,
+            post_affiliated,
+            contents_begin,
+            contents_end,
+            end,
+            post_blank,
+        }
+    })
+}

src/util/mod.rs

@@ -1,5 +1,7 @@
 #[cfg(any(feature = "compare", feature = "wasm_test"))]
 pub mod cli;
+#[cfg(any(feature = "compare", feature = "wasm_test"))]
+pub mod elisp;
 #[cfg(any(feature = "compare", feature = "wasm", feature = "wasm_test"))]
 pub mod elisp_fact;
 #[cfg(any(feature = "compare", feature = "wasm_test"))]