From 02710dc4126abcd9929f80c11417033afd6b646a Mon Sep 17 00:00:00 2001
From: rzvxa <rzvxa@protonmail.com>
Date: Wed, 28 Aug 2024 18:56:33 +0330
Subject: [PATCH] feat(regular_expression): implement `Display` for
 `RegularExpression` type.

---
 .../src/body_parser/mod.rs                    |   3 +-
 crates/oxc_regular_expression/src/display.rs  | 374 ++++++++++++++++++
 crates/oxc_regular_expression/src/lib.rs      |   1 +
 3 files changed, 377 insertions(+), 1 deletion(-)
 create mode 100644 crates/oxc_regular_expression/src/display.rs

diff --git a/crates/oxc_regular_expression/src/body_parser/mod.rs b/crates/oxc_regular_expression/src/body_parser/mod.rs
index be0e18bd65dcf8..0e1135778cbf05 100644
--- a/crates/oxc_regular_expression/src/body_parser/mod.rs
+++ b/crates/oxc_regular_expression/src/body_parser/mod.rs
@@ -2,9 +2,10 @@ mod diagnostics;
 mod parser;
 mod reader;
 mod state;
-mod unicode;
 mod unicode_property;
 
+pub(crate) mod unicode;
+
 pub use parser::PatternParser;
 
 #[cfg(test)]
diff --git a/crates/oxc_regular_expression/src/display.rs b/crates/oxc_regular_expression/src/display.rs
new file mode 100644
index 00000000000000..014877837853df
--- /dev/null
+++ b/crates/oxc_regular_expression/src/display.rs
@@ -0,0 +1,374 @@
+use std::fmt::{self, Display};
+
+#[allow(clippy::wildcard_imports)]
+use crate::ast::*;
+use crate::body_parser::unicode::{is_lead_surrogate, is_syntax_character, is_trail_surrogate};
+
+impl<'a> Display for RegularExpression<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "/{}/{}", self.pattern, self.flags)
+    }
+}
+
+impl Display for Flags {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        let mut flags = String::with_capacity(8);
+        macro_rules! if_true_append {
+            ($flag:ident, $char:literal) => {
+                if self.$flag {
+                    flags.push($char);
+                }
+            };
+        }
+
+        // write flags in the order they are described in the `MDN`
+        // <https://developer.mozilla.org/en-US/docs/Web/JavaScript/Guide/Regular_expressions#advanced_searching_with_flags>
+        if_true_append!(has_indices, 'd');
+        if_true_append!(global, 'g');
+        if_true_append!(ignore_case, 'i');
+        if_true_append!(multiline, 'm');
+        if_true_append!(dot_all, 's');
+        if_true_append!(unicode, 'u');
+        if_true_append!(unicode_sets, 'v');
+        if_true_append!(sticky, 'y');
+
+        write!(f, "{flags}")
+    }
+}
+
+impl<'a> Display for Pattern<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "{}", self.body)
+    }
+}
+
+impl<'a> Display for Disjunction<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write_join(f, "|", &self.body)
+    }
+}
+
+impl<'a> Display for Alternative<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write_join(f, "", &self.body)
+    }
+}
+
+impl<'a> Display for Term<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        match self {
+            Self::BoundaryAssertion(it) => write!(f, "{it}"),
+            Self::LookAroundAssertion(it) => write!(f, "{}", it.as_ref()),
+            Self::Quantifier(it) => write!(f, "{}", it.as_ref()),
+            Self::Character(it) => write!(f, "{it}"),
+            Self::Dot(it) => write!(f, "{it}"),
+            Self::CharacterClassEscape(it) => write!(f, "{it}"),
+            Self::UnicodePropertyEscape(it) => write!(f, "{}", it.as_ref()),
+            Self::CharacterClass(it) => write!(f, "{}", it.as_ref()),
+            Self::CapturingGroup(it) => write!(f, "{}", it.as_ref()),
+            Self::IgnoreGroup(it) => write!(f, "{}", it.as_ref()),
+            Self::IndexedReference(it) => write!(f, "{it}"),
+            Self::NamedReference(it) => write!(f, "{}", it.as_ref()),
+        }
+    }
+}
+
+impl Display for BoundaryAssertion {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "{}", self.kind)
+    }
+}
+
+impl Display for BoundaryAssertionKind {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        match self {
+            Self::Start => write!(f, "^"),
+            Self::End => write!(f, "$"),
+            Self::Boundary => write!(f, r"\b"),
+            Self::NegativeBoundary => write!(f, r"\B"),
+        }
+    }
+}
+
+impl<'a> Display for LookAroundAssertion<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "({}{})", self.kind, self.body)
+    }
+}
+
+impl Display for LookAroundAssertionKind {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        match self {
+            Self::Lookahead => write!(f, "?="),
+            Self::NegativeLookahead => write!(f, "?!"),
+            Self::Lookbehind => write!(f, "?<="),
+            Self::NegativeLookbehind => write!(f, "?<!"),
+        }
+    }
+}
+
+impl<'a> Display for Quantifier<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "{}", self.body)?;
+
+        match (self.min, self.max) {
+            (0, None) => write!(f, "*")?,
+            (1, None) => write!(f, "+")?,
+            (0, Some(1)) => write!(f, "?")?,
+            (min, Some(max)) if min == max => write!(f, "{{{min}}}",)?,
+            (min, max) => {
+                let max = max.map_or_else(String::default, |it| it.to_string());
+                write!(f, "{{{min},{max}}}",)?;
+            }
+        }
+
+        if !self.greedy {
+            write!(f, "?")?;
+        }
+
+        Ok(())
+    }
+}
+
+impl Display for Character {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        let cp = self.value;
+        // escape syntax characters
+        if is_syntax_character(cp) {
+            write!(f, r"\")?;
+        }
+
+        if is_lead_surrogate(cp) || is_trail_surrogate(cp) {
+            write!(f, r"\u{cp:X}")
+        } else {
+            // TODO: use `self.kind` to print the correct representation.
+            let Some(ch) = char::from_u32(cp) else { return Err(fmt::Error) };
+            write!(f, "{ch}")
+        }
+    }
+}
+
+impl Display for Dot {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, ".")
+    }
+}
+
+impl Display for CharacterClassEscape {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "{}", self.kind)
+    }
+}
+
+impl Display for CharacterClassEscapeKind {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        match self {
+            Self::D => write!(f, r"\d"),
+            Self::NegativeD => write!(f, r"\D"),
+            Self::S => write!(f, r"\s"),
+            Self::NegativeS => write!(f, r"\S"),
+            Self::W => write!(f, r"\w"),
+            Self::NegativeW => write!(f, r"\W"),
+        }
+    }
+}
+
+impl<'a> Display for UnicodePropertyEscape<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        if self.negative {
+            write!(f, r"\P")?;
+        } else {
+            write!(f, r"\p")?;
+        }
+
+        if let Some(value) = &self.value {
+            let name = &self.name;
+            write!(f, "{{{name}={value}}}")
+        } else {
+            write!(f, "{{{}}}", self.name)
+        }
+    }
+}
+
+impl<'a> Display for CharacterClass<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "[")?;
+
+        if !self.body.is_empty() {
+            if self.negative {
+                write!(f, "^")?;
+            }
+            let sep = match self.kind {
+                CharacterClassContentsKind::Union => "",
+                CharacterClassContentsKind::Subtraction => "--",
+                CharacterClassContentsKind::Intersection => "&&",
+            };
+            write_join(f, sep, &self.body)?;
+        }
+
+        write!(f, "]")
+    }
+}
+
+impl<'a> Display for CharacterClassContents<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        match self {
+            Self::CharacterClassRange(it) => write!(f, "{}", it.as_ref()),
+            Self::CharacterClassEscape(it) => write!(f, "{it}"),
+            Self::UnicodePropertyEscape(it) => write!(f, "{}", it.as_ref()),
+            Self::Character(it) => write!(f, "{it}"),
+            Self::NestedCharacterClass(it) => write!(f, "{}", it.as_ref()),
+            Self::ClassStringDisjunction(it) => write!(f, "{}", it.as_ref()),
+        }
+    }
+}
+
+impl Display for CharacterClassRange {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "{}-{}", self.min, self.max)
+    }
+}
+
+impl<'a> Display for ClassStringDisjunction<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "\\q{{")?;
+        write_join(f, "|", &self.body)?;
+        write!(f, "}}")
+    }
+}
+
+impl<'a> Display for ClassString<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write_join(f, "", &self.body)
+    }
+}
+
+impl<'a> Display for CapturingGroup<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        let body = &self.body;
+        if let Some(name) = &self.name {
+            write!(f, "(?<{name}>{body})")
+        } else {
+            write!(f, "({body})")
+        }
+    }
+}
+
+impl<'a> Display for IgnoreGroup<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        fn write_flags(
+            f: &mut fmt::Formatter<'_>,
+            prefix: char,
+            flags: &ModifierFlags,
+        ) -> fmt::Result {
+            if flags.ignore_case {
+                write!(f, "{prefix}i")?;
+            }
+            if flags.sticky {
+                write!(f, "{prefix}y")?;
+            }
+            if flags.multiline {
+                write!(f, "{prefix}m")?;
+            }
+            Ok(())
+        }
+
+        write!(f, "(?")?;
+        if let Some(enabling) = &self.enabling_modifiers {
+            write_flags(f, '\0', enabling)?;
+        }
+        if let Some(disabling) = &self.disabling_modifiers {
+            write_flags(f, '-', disabling)?;
+        }
+        write!(f, ":{})", self.body)
+    }
+}
+
+impl Display for IndexedReference {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "\\{}", self.index)
+    }
+}
+
+impl<'a> Display for NamedReference<'a> {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, r"\k<{}>", self.name)
+    }
+}
+
+fn write_join<S, I, E>(f: &mut fmt::Formatter<'_>, sep: S, items: I) -> fmt::Result
+where
+    S: AsRef<str>,
+    E: Display,
+    I: IntoIterator<Item = E>,
+{
+    let sep = sep.as_ref();
+    let mut iter = items.into_iter();
+
+    if let Some(first) = iter.next() {
+        write!(f, "{first}")?;
+    }
+
+    for it in iter {
+        write!(f, "{sep}{it}")?;
+    }
+
+    Ok(())
+}
+
+#[cfg(test)]
+mod test {
+    use oxc_allocator::Allocator;
+
+    static CASES: &[(&str, /* expected display */ Option<&str>)] = &[
+        ("/ab/", None),
+        ("/abc/i", None),
+        ("/a*?/i", None),
+        ("/emo👈🏻ji/u", None),
+        ("/ab|c/i", None),
+        ("/a|b+|c/i", None),
+        ("/(?=a)|(?<=b)|(?!c)|(?<!d)/i", None),
+        (r"/(cg)(?<n>cg)(?:g)/", None),
+        (r"/^(?=ab)\b(?!cd)(?<=ef)\B(?<!gh)$/", None),
+        (r"/^(?<!ab)$/", None),
+        (r"/[abc]/", None),
+        (r"/[a&&b]/v", None),
+        (r"/[a--b]/v", None),
+        (r"/[^a--b--c]/v", None),
+        (r"/[a[b[c[d[e[f[g[h[i[j[k[l]]]]]]]]]]]]/v", None),
+        (r"/[\q{abc|d|e|}]/v", None),
+        (r"/\p{Basic_Emoji}/v", None),
+        // we lose the flags ordering
+        ("/abcd/igv", Some("/abcd/giv")),
+        (r"/\d/ug", Some(r"/\d/gu")),
+        // we always display "syntax characters" as their escaped form.
+        (r"/c]/", Some(r"/c\]/")),
+        (r"/[|\]]/", Some(r"/[\|\]]/")),
+        ("/a{0}|b{1,2}|c{3,}/i", None),
+        // NOTE: surrogated characters can't be displayed without access to the next character so we
+        // can't print them correctly via `Display` trait. Instead we print their unicode code point
+        (r"/Em🥹j/", Some(r"/Em\uD83E\uDD79j/")),
+        // TODO: currently we do **NOT** respect the `Character::kind` field so everything is
+        // lowered to their final form.
+        (r"/\n\cM\0\x41\./", Some("/\n\r\0A\\./")),
+        (r"/\n\cM\0\x41\u1234\./u", Some("/\n\r\0Aሴ\\./u")),
+        (r"/\n\cM\0\x41\u{1f600}\./u", Some("/\n\r\0A😀\\./u")),
+        // TODO: `\b` is parsed as `\u{8}` which is wrong.
+        // (r"/[\bb]/", Some(r"/[\bb]/")),
+    ];
+
+    fn test_display(allocator: &Allocator, source: &str, expect: Option<&str>) {
+        use crate::{Parser, ParserOptions};
+        let expect = expect.unwrap_or(source);
+        let parsed = Parser::new(allocator, source, ParserOptions::default()).parse().unwrap();
+        assert_eq!(expect, parsed.to_string());
+    }
+
+    #[test]
+    fn test() {
+        let allocator = &Allocator::default();
+        CASES
+            .iter()
+            .for_each(|(source, expect)| test_display(allocator, source, expect.as_deref()));
+    }
+}
diff --git a/crates/oxc_regular_expression/src/lib.rs b/crates/oxc_regular_expression/src/lib.rs
index ae2f1c0a58bc0f..5b2d4758915efe 100644
--- a/crates/oxc_regular_expression/src/lib.rs
+++ b/crates/oxc_regular_expression/src/lib.rs
@@ -2,6 +2,7 @@
 
 pub mod ast;
 mod body_parser;
+mod display;
 mod flag_parser;
 mod literal_parser;
 mod options;