//! Validation of char literals

use std::u32;

use arrayvec::ArrayString;

use crate::{
    string_lexing::{self, StringComponentKind},
    TextRange,
    SyntaxError,
    SyntaxErrorKind::*,
    SyntaxToken,
};

pub(super) fn validate_char_node(node: SyntaxToken, errors: &mut Vec<SyntaxError>) {
    let literal_text = node.text();
    let literal_range = node.range();
    let mut components = string_lexing::parse_char_literal(literal_text);
    let mut len = 0;
    for component in &mut components {
        len += 1;
        let text = &literal_text[component.range];
        let range = component.range + literal_range.start();
        validate_char_component(text, component.kind, range, errors);
    }

    if !components.has_closing_quote {
        errors.push(SyntaxError::new(UnclosedChar, literal_range));
    }

    if let Some(range) = components.suffix {
        errors.push(SyntaxError::new(InvalidSuffix, range + literal_range.start()));
    }

    if len == 0 {
        errors.push(SyntaxError::new(EmptyChar, literal_range));
    }

    if len > 1 {
        errors.push(SyntaxError::new(OverlongChar, literal_range));
    }
}

pub(super) fn validate_char_component(
    text: &str,
    kind: StringComponentKind,
    range: TextRange,
    errors: &mut Vec<SyntaxError>,
) {
    // Validate escapes
    use self::StringComponentKind::*;
    match kind {
        AsciiEscape => validate_ascii_escape(text, range, errors),
        AsciiCodeEscape => validate_ascii_code_escape(text, range, errors),
        UnicodeEscape => validate_unicode_escape(text, range, errors),
        CodePoint => {
            // These code points must always be escaped
            if text == "\t" || text == "\r" || text == "\n" {
                errors.push(SyntaxError::new(UnescapedCodepoint, range));
            }
        }
        StringComponentKind::IgnoreNewline => { /* always valid */ }
    }
}

fn validate_ascii_escape(text: &str, range: TextRange, errors: &mut Vec<SyntaxError>) {
    if text.len() == 1 {
        // Escape sequence consists only of leading `\` (only occurs at EOF, otherwise e.g. '\' is treated as an unclosed char containing a single quote `'`)
        errors.push(SyntaxError::new(EmptyAsciiEscape, range));
    } else {
        let escape_code = text.chars().skip(1).next().unwrap();
        if !is_ascii_escape(escape_code) {
            errors.push(SyntaxError::new(InvalidAsciiEscape, range));
        }
    }
}

pub(super) fn is_ascii_escape(code: char) -> bool {
    match code {
        '\\' | '\'' | '"' | 'n' | 'r' | 't' | '0' => true,
        _ => false,
    }
}

fn validate_ascii_code_escape(text: &str, range: TextRange, errors: &mut Vec<SyntaxError>) {
    // An AsciiCodeEscape has 4 chars, example: `\xDD`
    if !text.is_ascii() {
        // FIXME: Give a more precise error message (say what the invalid character was)
        errors.push(SyntaxError::new(AsciiCodeEscapeOutOfRange, range));
    } else if text.chars().count() < 4 {
        errors.push(SyntaxError::new(TooShortAsciiCodeEscape, range));
    } else {
        assert_eq!(
            text.chars().count(),
            4,
            "AsciiCodeEscape cannot be longer than 4 chars, but text '{}' is",
            text,
        );

        match u8::from_str_radix(&text[2..], 16) {
            Ok(code) if code < 128 => { /* Escape code is valid */ }
            Ok(_) => errors.push(SyntaxError::new(AsciiCodeEscapeOutOfRange, range)),
            Err(_) => errors.push(SyntaxError::new(MalformedAsciiCodeEscape, range)),
        }
    }
}

fn validate_unicode_escape(text: &str, range: TextRange, errors: &mut Vec<SyntaxError>) {
    assert!(&text[..2] == "\\u", "UnicodeEscape always starts with \\u");

    if text.len() == 2 {
        // No starting `{`
        errors.push(SyntaxError::new(MalformedUnicodeEscape, range));
        return;
    }

    if text.len() == 3 {
        // Only starting `{`
        errors.push(SyntaxError::new(UnclosedUnicodeEscape, range));
        return;
    }

    let mut code = ArrayString::<[_; 6]>::new();
    let mut closed = false;
    for c in text[3..].chars() {
        assert!(!closed, "no characters after escape is closed");

        if c.is_digit(16) {
            if code.len() == 6 {
                errors.push(SyntaxError::new(OverlongUnicodeEscape, range));
                return;
            }

            code.push(c);
        } else if c == '_' {
            // Reject leading _
            if code.len() == 0 {
                errors.push(SyntaxError::new(MalformedUnicodeEscape, range));
                return;
            }
        } else if c == '}' {
            closed = true;
        } else {
            errors.push(SyntaxError::new(MalformedUnicodeEscape, range));
            return;
        }
    }

    if !closed {
        errors.push(SyntaxError::new(UnclosedUnicodeEscape, range))
    }

    if code.len() == 0 {
        errors.push(SyntaxError::new(EmptyUnicodeEcape, range));
        return;
    }

    match u32::from_str_radix(&code, 16) {
        Ok(code_u32) if code_u32 > 0x10FFFF => {
            errors.push(SyntaxError::new(UnicodeEscapeOutOfRange, range));
        }
        Ok(_) => {
            // Valid escape code
        }
        Err(_) => {
            errors.push(SyntaxError::new(MalformedUnicodeEscape, range));
        }
    }
}

#[cfg(test)]
mod test {
    use crate::{SourceFile, TreeArc};

    fn build_file(literal: &str) -> TreeArc<SourceFile> {
        let src = format!("const C: char = '{}';", literal);
        SourceFile::parse(&src)
    }

    fn assert_valid_char(literal: &str) {
        let file = build_file(literal);
        assert!(file.errors().len() == 0, "Errors for literal '{}': {:?}", literal, file.errors());
    }

    fn assert_invalid_char(literal: &str) {
        let file = build_file(literal);
        assert!(file.errors().len() > 0);
    }

    #[test]
    fn test_ansi_codepoints() {
        for byte in 0..=255u8 {
            match byte {
                b'\n' | b'\r' | b'\t' => assert_invalid_char(&(byte as char).to_string()),
                b'\'' | b'\\' => { /* Ignore character close and backslash */ }
                _ => assert_valid_char(&(byte as char).to_string()),
            }
        }
    }

    #[test]
    fn test_unicode_codepoints() {
        let valid = ["Ƒ", "バ", "メ", "﷽"];
        for c in &valid {
            assert_valid_char(c);
        }
    }

    #[test]
    fn test_unicode_multiple_codepoints() {
        let invalid = ["नी", "👨‍👨‍"];
        for c in &invalid {
            assert_invalid_char(c);
        }
    }

    #[test]
    fn test_valid_ascii_escape() {
        let valid = [r"\'", "\"", "\\\\", "\\\"", r"\n", r"\r", r"\t", r"\0"];
        for c in &valid {
            assert_valid_char(c);
        }
    }

    #[test]
    fn test_invalid_ascii_escape() {
        let invalid = [r"\a", r"\?", r"\"];
        for c in &invalid {
            assert_invalid_char(c);
        }
    }

    #[test]
    fn test_valid_ascii_code_escape() {
        let valid = [r"\x00", r"\x7F", r"\x55"];
        for c in &valid {
            assert_valid_char(c);
        }
    }

    #[test]
    fn test_invalid_ascii_code_escape() {
        let invalid = [r"\x", r"\x7", r"\xF0"];
        for c in &invalid {
            assert_invalid_char(c);
        }
    }

    #[test]
    fn test_valid_unicode_escape() {
        let valid = [r"\u{FF}", r"\u{0}", r"\u{F}", r"\u{10FFFF}", r"\u{1_0__FF___FF_____}"];
        for c in &valid {
            assert_valid_char(c);
        }
    }

    #[test]
    fn test_invalid_unicode_escape() {
        let invalid = [
            r"\u",
            r"\u{}",
            r"\u{",
            r"\u{FF",
            r"\u{FFFFFF}",
            r"\u{_F}",
            r"\u{00FFFFF}",
            r"\u{110000}",
        ];
        for c in &invalid {
            assert_invalid_char(c);
        }
    }
}