organic/src/compare/util.rs

use std::borrow::Cow;
use std::str::FromStr;

use super::compare_field::compare_property_list_of_quoted_string;
use super::compare_field::compare_property_object_tree;
use super::compare_field::compare_property_optional_pair;
use super::compare_field::compare_property_quoted_string;
use super::compare_field::ComparePropertiesResult;
use super::diff::DiffEntry;
use super::diff::DiffStatus;
use super::elisp_fact::GetElispFact;
use super::sexp::Token;
use crate::compare::diff::compare_ast_node;
use crate::compare::sexp::unquote;
use crate::types::AffiliatedKeywordValue;
use crate::types::AstNode;
use crate::types::GetAffiliatedKeywords;
use crate::types::StandardProperties;

/// 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 the byte offset into source that the rust object exists at.
///
/// These offsets are zero-based unlike the elisp ones.
fn get_rust_byte_offsets(original_document: &str, subset: &str) -> (usize, usize) {
    debug_assert!(is_slice_of(original_document, subset));
    let offset = subset.as_ptr() as usize - original_document.as_ptr() as usize;
    let end = offset + subset.len();
    (offset, end)
}

pub(crate) fn compare_standard_properties<
    'b,
    's,
    S: StandardProperties<'s> + GetElispFact<'s> + ?Sized,
>(
    original_document: &'s str,
    emacs: &'b Token<'s>,
    rust: &'b S,
) -> Result<(), Box<dyn std::error::Error>> {
    assert_name(emacs, rust.get_elisp_fact().get_elisp_name())?;
    assert_bounds(original_document, emacs, rust)?;
    assert_post_blank(emacs, rust)?;
    Ok(())
}

fn assert_name<S: AsRef<str>>(
    emacs: &Token<'_>,
    name: S,
) -> Result<(), Box<dyn std::error::Error>> {
    let name = name.as_ref();
    let children = emacs.as_list()?;
    let first_child = children
        .first()
        .ok_or("Should have at least one child.")?
        .as_atom()?;
    if first_child != name {
        Err(format!(
            "AST node name mismatch. Expected a (rust) {expected} cell, but found a (emacs) {found} cell.",
            expected = name,
            found = first_child
        ))?;
    }
    Ok(())
}

/// Assert that the character ranges defined by upstream org-mode's :standard-properties match the slices in Organic's StandardProperties.
///
/// This does **not** handle plain text because plain text is a special case.
fn assert_bounds<'b, 's, S: StandardProperties<'s> + ?Sized>(
    original_document: &'s str,
    emacs: &'b Token<'s>,
    rust: &'b S,
) -> Result<(), Box<dyn std::error::Error>> {
    let standard_properties = get_emacs_standard_properties(emacs)?; // 1-based

    // Check begin/end
    {
        let (begin, end) = (
            standard_properties
                .begin
                .ok_or("Token should have a begin.")?,
            standard_properties.end.ok_or("Token should have an end.")?,
        );
        let (rust_begin, rust_end) = get_rust_byte_offsets(original_document, rust.get_source()); // 0-based
        let rust_begin_char_offset = original_document[..rust_begin].chars().count() + 1; // 1-based
        let rust_end_char_offset =
            rust_begin_char_offset + original_document[rust_begin..rust_end].chars().count(); // 1-based
        if rust_begin_char_offset != begin || rust_end_char_offset != end {
            Err(format!("Rust bounds (in chars) ({rust_begin}, {rust_end}) do not match emacs bounds ({emacs_begin}, {emacs_end})", rust_begin = rust_begin_char_offset, rust_end = rust_end_char_offset, emacs_begin=begin, emacs_end=end))?;
        }
    }

    // Check contents-begin/contents-end
    {
        if let Some(rust_contents) = rust.get_contents() {
            let (begin, end) = (
                standard_properties
                    .contents_begin
                    .ok_or("Token should have a contents-begin.")?,
                standard_properties
                    .contents_end
                    .ok_or("Token should have an contents-end.")?,
            );
            let (rust_begin, rust_end) = get_rust_byte_offsets(original_document, rust_contents); // 0-based
            let rust_begin_char_offset = original_document[..rust_begin].chars().count() + 1; // 1-based
            let rust_end_char_offset =
                rust_begin_char_offset + original_document[rust_begin..rust_end].chars().count(); // 1-based
            if rust_begin_char_offset != begin || rust_end_char_offset != end {
                Err(format!("Rust contents bounds (in chars) ({rust_begin}, {rust_end}) do not match emacs contents bounds ({emacs_begin}, {emacs_end})", rust_begin = rust_begin_char_offset, rust_end = rust_end_char_offset, emacs_begin=begin, emacs_end=end))?;
            }
        } else if standard_properties.contents_begin.is_some()
            || standard_properties.contents_end.is_some()
        {
            Err(format!("Rust contents is None but emacs contents bounds are ({emacs_begin:?}, {emacs_end:?})", emacs_begin=standard_properties.contents_begin, emacs_end=standard_properties.contents_end))?;
        }
    }

    Ok(())
}

/// Assert that the post blank matches between emacs and organic.
///
/// This does **not** handle plain text because plain text is a special case.
fn assert_post_blank<'b, 's, S: StandardProperties<'s> + ?Sized>(
    emacs: &'b Token<'s>,
    rust: &'b S,
) -> Result<(), Box<dyn std::error::Error>> {
    let standard_properties = get_emacs_standard_properties(emacs)?; // 1-based
    let rust_post_blank = rust.get_post_blank();
    let emacs_post_blank = standard_properties
        .post_blank
        .ok_or("Token should have a post-blank.")?;
    if rust_post_blank as usize != emacs_post_blank {
        Err(format!("Rust post-blank {rust_post_blank} does not match emacs post-blank ({emacs_post_blank})", rust_post_blank = rust_post_blank, emacs_post_blank = emacs_post_blank))?;
    }

    Ok(())
}

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,
        }
    })
}

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))?)
}

/// Get a named property from the emacs token.
///
/// Returns Ok(None) if value is nil or absent.
pub(crate) fn get_property<'b, 's>(
    emacs: &'b Token<'s>,
    key: &str,
) -> Result<Option<&'b Token<'s>>, 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 prop = attributes_map.get(key).copied();
    if let Some(Ok("nil")) = prop.map(Token::as_atom) {
        return Ok(None);
    }
    Ok(prop)
}

/// Get a named property containing an unquoted atom from the emacs token.
///
/// Returns None if key is not found.
pub(crate) fn get_property_unquoted_atom<'s>(
    emacs: &Token<'s>,
    key: &str,
) -> Result<Option<&'s str>, Box<dyn std::error::Error>> {
    get_property(emacs, key)?
        .map(Token::as_atom)
        .map_or(Ok(None), |r| r.map(Some))
}

/// Get a named property containing an quoted string from the emacs token.
///
/// Returns None if key is not found.
pub(crate) fn get_property_quoted_string<'s>(
    emacs: &Token<'s>,
    key: &str,
) -> Result<Option<Cow<'s, str>>, Box<dyn std::error::Error>> {
    get_property(emacs, key)?
        .map(Token::as_atom)
        .map_or(Ok(None), |r| r.map(Some))?
        .map(unquote)
        .map_or(Ok(None), |r| r.map(Some))
}

/// Get a named property containing an unquoted numeric value.
///
/// Returns None if key is not found.
pub(crate) fn get_property_numeric<'b, 's, 'x, N: FromStr>(
    emacs: &'b Token<'s>,
    key: &'x str,
) -> Result<Option<N>, Box<dyn std::error::Error + 's>>
where
    <N as FromStr>::Err: std::error::Error,
    <N as FromStr>::Err: 's,
{
    let unparsed_string = get_property(emacs, key)?
        .map(Token::as_atom)
        .map_or(Ok(None), |r| r.map(Some))?;
    let parsed_number = unparsed_string
        .map(|val| val.parse::<N>())
        .map_or(Ok(None), |r| r.map(Some))?;
    Ok(parsed_number)
}

pub(crate) fn compare_children<'b, 's, 'x, RC>(
    source: &'s str,
    emacs: &'b Token<'s>,
    rust_children: &'x [RC],
    child_status: &mut Vec<DiffEntry<'b, 's>>,
    this_status: &mut DiffStatus,
    message: &mut Option<String>,
) -> Result<(), Box<dyn std::error::Error>>
where
    AstNode<'b, 's>: From<&'x RC>,
{
    let emacs_children = emacs.as_list()?;
    let emacs_children_length = emacs_children.len() - 2;
    if emacs_children_length != rust_children.len() {
        *this_status = DiffStatus::Bad;
        *message = Some(format!(
            "Child length mismatch (emacs != rust) {:?} != {:?}",
            emacs_children_length,
            rust_children.len()
        ));
    }
    for (emacs_child, rust_child) in emacs_children.iter().skip(2).zip(rust_children.iter()) {
        child_status.push(compare_ast_node(source, emacs_child, rust_child.into())?);
    }
    Ok(())
}

pub(crate) fn compare_children_iter<'b, 's, RC, RI: Iterator<Item = RC> + ExactSizeIterator>(
    source: &'s str,
    emacs: &'b Token<'s>,
    rust_children: RI,
    child_status: &mut Vec<DiffEntry<'b, 's>>,
    this_status: &mut DiffStatus,
    message: &mut Option<String>,
) -> Result<(), Box<dyn std::error::Error>>
where
    AstNode<'b, 's>: From<RC>,
{
    let emacs_children = emacs.as_list()?;
    let emacs_children_length = emacs_children.len() - 2;
    if emacs_children_length != rust_children.len() {
        *this_status = DiffStatus::Bad;
        *message = Some(format!(
            "Child length mismatch (emacs != rust) {:?} != {:?}",
            emacs_children_length,
            rust_children.len()
        ));
    }
    for (emacs_child, rust_child) in emacs_children.iter().skip(2).zip(rust_children) {
        child_status.push(compare_ast_node(source, emacs_child, rust_child.into())?);
    }
    Ok(())
}

pub(crate) fn assert_no_children(
    emacs: &Token<'_>,
    this_status: &mut DiffStatus,
    message: &mut Option<String>,
) -> Result<(), Box<dyn std::error::Error>> {
    let emacs_children_length = emacs.as_list()?.len();
    // 2, one for the name of the node and one for the properties. Children would come after that.
    if emacs_children_length != 2 {
        *this_status = DiffStatus::Bad;
        *message = Some(format!(
            "Should have no children but emacs has {:?} children.",
            emacs_children_length - 2,
        ));
    }
    Ok(())
}

pub(crate) fn compare_additional_properties<'b, 's, RK, RV, RI>(
    emacs: &'b Token<'s>,
    rust_children: RI,
) -> Result<ComparePropertiesResult<'b, 's>, Box<dyn std::error::Error>>
where
    RK: AsRef<str>,
    RV: AsRef<str>,
    RI: Iterator<Item = (RK, RV)> + ExactSizeIterator,
{
    for (rust_key, rust_value) in rust_children {
        let rust_key = rust_key.as_ref();
        let rust_value = rust_value.as_ref();
        let emacs_value = get_property_quoted_string(emacs, rust_key)?;
        if Some(rust_value) != emacs_value.as_deref() {
            let this_status = DiffStatus::Bad;
            let message = Some(format!(
                "{} mismatch (emacs != rust) {:?} != {:?}",
                rust_key, emacs_value, rust_value
            ));
            return Ok(ComparePropertiesResult::SelfChange(this_status, message));
        }
    }
    Ok(ComparePropertiesResult::NoChange)
}

pub(crate) fn compare_affiliated_keywords<'b, 's, GAK>(
    source: &'s str,
    emacs: &'b Token<'s>,
    rust: &'b GAK,
) -> Result<Vec<ComparePropertiesResult<'b, 's>>, Box<dyn std::error::Error>>
where
    GAK: GetAffiliatedKeywords<'s>,
{
    let mut ret = Vec::new();
    let affiliated_keywords = rust.get_affiliated_keywords();
    for (rust_name, rust_value) in affiliated_keywords.keywords.iter() {
        let emacs_property_name = format!(":{}", rust_name);
        match rust_value {
            AffiliatedKeywordValue::SingleString(rust_value) => {
                let diff = compare_property_quoted_string(
                    source,
                    emacs,
                    rust,
                    emacs_property_name.as_str(),
                    |_| Some(*rust_value),
                )?;
                ret.push(diff);
            }
            AffiliatedKeywordValue::ListOfStrings(rust_value) => {
                let diff = compare_property_list_of_quoted_string(
                    source,
                    emacs,
                    rust,
                    emacs_property_name.as_str(),
                    |_| Some(rust_value.iter()),
                )?;
                ret.push(diff);
            }
            AffiliatedKeywordValue::OptionalPair { optval, val } => {
                let diff = compare_property_optional_pair(
                    source,
                    emacs,
                    rust,
                    emacs_property_name.as_str(),
                    |_| Some((*optval, *val)),
                )?;
                ret.push(diff);
            }
            AffiliatedKeywordValue::ObjectTree(rust_value) => {
                let diff = compare_property_object_tree(
                    source,
                    emacs,
                    rust,
                    emacs_property_name.as_str(),
                    |_| Some(rust_value.iter()),
                )?;
                ret.push(diff);
            }
        };
    }
    Ok(ret)
}

pub(crate) fn affiliated_keywords_names<'s, GAK>(rust: &'s GAK) -> impl Iterator<Item = String> + 's
where
    GAK: GetAffiliatedKeywords<'s>,
{
    rust.get_affiliated_keywords()
        .keywords
        .keys()
        .map(|k| format!(":{}", k))
}