Update malachite-bigint

* Fix inconsistencies with cpython while parsing format strings which contain colons inside square brackets

Co-authored-by: Jeong, YunWon <69878+youknowone@users.noreply.github.com>

.github/workflows/ci.yaml

@@ -63,7 +63,7 @@ jobs:
       - name: install ruff
         run: python -m pip install ruff
       - name: run python lint
-        run: ruff --ignore=E501 ast --show-source
+        run: ruff check ast
 
       - name: spell checker
         uses: streetsidesoftware/cspell-action@v2

Cargo.toml

@@ -1,5 +1,5 @@
 [workspace.package]
-version = "0.3.1"
+version = "0.4.0"
 authors = ["RustPython Team"]
 edition = "2021"
 rust-version = "1.72.1"
@@ -15,12 +15,12 @@ members = [
 ]
 
 [workspace.dependencies]
-rustpython-parser-vendored = { path = "vendored", version = "0.3.1" }
-rustpython-ast = { path = "ast", default-features = false, version = "0.3.1" }
-rustpython-parser-core = { path = "core", features = [], version = "0.3.1" }
-rustpython-literal = { path = "literal", version = "0.3.1" }
-rustpython-format = { path = "format", default-features = false, version = "0.3.1" }
-rustpython-parser = { path = "parser", default-features = false, version = "0.3.1" }
+rustpython-parser-vendored = { path = "vendored", version = "0.4.0" }
+rustpython-ast = { path = "ast", default-features = false, version = "0.4.0" }
+rustpython-parser-core = { path = "core", features = [], version = "0.4.0" }
+rustpython-literal = { path = "literal", version = "0.4.0" }
+rustpython-format = { path = "format", default-features = false, version = "0.4.0" }
+rustpython-parser = { path = "parser", default-features = false, version = "0.4.0" }
 
 anyhow = "1.0.45"
 bitflags = "2.4.0"
@@ -32,7 +32,7 @@ log = "0.4.16"
 num-complex = "0.4.0"
 num-bigint = "0.4.3"
 num-traits = "0.2"
-malachite-bigint = "0.2.0"
+malachite-bigint = "0.2.3"
 memchr = "2.5.0"
 rand = "0.8.5"
 serde = { version = "1.0.133", default-features = false }

ast/src/builtin.rs

@@ -128,10 +128,10 @@ pub enum Constant {
 
 impl Constant {
     pub fn is_true(self) -> bool {
-        self.bool().map_or(false, |b| b)
+        self.bool().is_some_and(|b| b)
     }
     pub fn is_false(self) -> bool {
-        self.bool().map_or(false, |b| !b)
+        self.bool().is_some_and(|b| !b)
     }
     pub fn complex(self) -> Option<(f64, f64)> {
         match self {

ast/src/lib.rs

@@ -20,7 +20,7 @@ mod generic;
 mod impls;
 mod ranged;
 #[cfg(feature = "unparse")]
-mod unparse;
+pub mod unparse;
 
 #[cfg(feature = "malachite-bigint")]
 pub use malachite_bigint as bigint;

ast/src/source_locator.rs

@@ -156,11 +156,11 @@ impl crate::fold::Fold<TextRange> for LinearLocator<'_> {
         let context = self.will_map_user(&range);
 
         let name = self.fold(name)?;
+        let type_params = self.fold(type_params)?;
         let bases = self.fold(bases)?;
         let keywords = self.fold(keywords)?;
         let body = self.fold(body)?;
         let range = self.map_user(range, context)?;
-        let type_params = self.fold(type_params)?;
 
         Ok(crate::StmtClassDef {
             name,
@@ -190,11 +190,11 @@ impl crate::fold::Fold<TextRange> for LinearLocator<'_> {
         let context = self.will_map_user(&range);
 
         let name = self.fold(name)?;
+        let type_params = self.fold(type_params)?;
         let args: Box<crate::Arguments<SourceRange>> = self.fold(args)?;
         let returns = self.fold(returns)?;
         let body = self.fold(body)?;
         let type_comment = self.fold(type_comment)?;
-        let type_params = self.fold(type_params)?;
         let range = self.map_user(range, context)?;
         Ok(crate::StmtFunctionDef {
             name,
@@ -225,11 +225,11 @@ impl crate::fold::Fold<TextRange> for LinearLocator<'_> {
         let context = self.will_map_user(&range);
 
         let name = self.fold(name)?;
+        let type_params = self.fold(type_params)?;
         let args: Box<crate::Arguments<SourceRange>> = self.fold(args)?;
         let returns = self.fold(returns)?;
         let body = self.fold(body)?;
         let type_comment = self.fold(type_comment)?;
-        let type_params = self.fold(type_params)?;
         let range = self.map_user(range, context)?;
         Ok(crate::StmtAsyncFunctionDef {
             name,
@@ -274,6 +274,34 @@ impl crate::fold::Fold<TextRange> for LinearLocator<'_> {
             keywords,
         })
     }
+
+    fn fold_pattern_match_mapping(
+        &mut self,
+        node: crate::PatternMatchMapping<TextRange>,
+    ) -> Result<crate::PatternMatchMapping<Self::TargetU>, Self::Error> {
+        let crate::PatternMatchMapping {
+            keys,
+            patterns,
+            rest,
+            range,
+        } = node;
+        let context = self.will_map_user(&range);
+
+        let mut located_keys = Vec::with_capacity(keys.len());
+        let mut located_patterns = Vec::with_capacity(patterns.len());
+        for (key, value) in keys.into_iter().zip(patterns.into_iter()) {
+            located_keys.push(self.fold(key)?);
+            located_patterns.push(self.fold(value)?);
+        }
+        let rest = self.fold(rest)?;
+        let range = self.map_user(range, context)?;
+        Ok(crate::PatternMatchMapping {
+            keys: located_keys,
+            patterns: located_patterns,
+            rest,
+            range,
+        })
+    }
 }
 
 struct LinearLookaheadLocator<'a, 'b>(&'b mut LinearLocator<'a>);

ast/src/unparse.rs

@@ -379,12 +379,7 @@ impl<'a> Unparser<'a> {
             }
             Expr::Subscript(crate::ExprSubscript { value, slice, .. }) => {
                 self.unparse_expr(value, precedence::ATOM)?;
-                let mut lvl = precedence::TUPLE;
-                if let Expr::Tuple(crate::ExprTuple { elts, .. }) = slice.as_ref() {
-                    if elts.iter().any(|expr| expr.is_starred_expr()) {
-                        lvl += 1
-                    }
-                }
+                let lvl = precedence::TUPLE;
                 self.p("[")?;
                 self.unparse_expr(slice, lvl)?;
                 self.p("]")?;

format/src/format.rs

@@ -863,17 +863,46 @@ impl FormatString {
     }
 
     fn parse_part_in_brackets(text: &str) -> Result<FormatPart, FormatParseError> {
-        let parts: Vec<&str> = text.splitn(2, ':').collect();
+        let mut chars = text.chars().peekable();
+
+        let mut left = String::new();
+        let mut right = String::new();
+
+        let mut split = false;
+        let mut selected = &mut left;
+        let mut inside_brackets = false;
+
+        while let Some(char) = chars.next() {
+            if char == '[' {
+                inside_brackets = true;
+
+                selected.push(char);
+
+                while let Some(next_char) = chars.next() {
+                    selected.push(next_char);
+
+                    if next_char == ']' {
+                        inside_brackets = false;
+                        break;
+                    }
+                    if chars.peek().is_none() {
+                        return Err(FormatParseError::MissingRightBracket);
+                    }
+                }
+            } else if char == ':' && !split && !inside_brackets {
+                split = true;
+                selected = &mut right;
+            } else {
+                selected.push(char);
+            }
+        }
+
         // before the comma is a keyword or arg index, after the comma is maybe a spec.
-        let arg_part = parts[0];
+        let arg_part: &str = &left;
 
-        let format_spec = if parts.len() > 1 {
-            parts[1].to_owned()
-        } else {
-            String::new()
-        };
+        let format_spec = if split { right } else { String::new() };
 
-        // On parts[0] can still be the conversion (!r, !s, !a)
+        // left can still be the conversion (!r, !s, !a)
         let parts: Vec<&str> = arg_part.splitn(2, '!').collect();
         // before the bang is a keyword or arg index, after the comma is maybe a conversion spec.
         let arg_part = parts[0];
@@ -1168,6 +1197,34 @@ mod tests {
         );
     }
 
+    #[test]
+    fn test_square_brackets_inside_format() {
+        assert_eq!(
+            FormatString::from_str("{[:123]}"),
+            Ok(FormatString {
+                format_parts: vec![FormatPart::Field {
+                    field_name: "[:123]".to_owned(),
+                    conversion_spec: None,
+                    format_spec: "".to_owned(),
+                }],
+            }),
+        );
+
+        assert_eq!(FormatString::from_str("{asdf[:123]asdf}"), {
+            Ok(FormatString {
+                format_parts: vec![FormatPart::Field {
+                    field_name: "asdf[:123]asdf".to_owned(),
+                    conversion_spec: None,
+                    format_spec: "".to_owned(),
+                }],
+            })
+        });
+
+        assert_eq!(FormatString::from_str("{[1234}"), {
+            Err(FormatParseError::MissingRightBracket)
+        });
+    }
+
     #[test]
     fn test_format_parse_escape() {
         let expected = Ok(FormatString {

literal/src/escape.rs

@@ -232,7 +232,7 @@ impl UnicodeEscape<'_> {
     }
 }
 
-impl<'a> Escape for UnicodeEscape<'a> {
+impl Escape for UnicodeEscape<'_> {
     fn source_len(&self) -> usize {
         self.source.len()
     }
@@ -254,24 +254,6 @@ impl<'a> Escape for UnicodeEscape<'a> {
     }
 }
 
-#[cfg(test)]
-mod unicode_escape_tests {
-    use super::*;
-
-    #[test]
-    fn changed() {
-        fn test(s: &str) -> bool {
-            UnicodeEscape::new_repr(s).changed()
-        }
-        assert!(!test("hello"));
-        assert!(!test("'hello'"));
-        assert!(!test("\"hello\""));
-
-        assert!(test("'\"hello"));
-        assert!(test("hello\n"));
-    }
-}
-
 pub struct AsciiEscape<'a> {
     source: &'a [u8],
     layout: EscapeLayout,
@@ -391,7 +373,7 @@ impl AsciiEscape<'_> {
     }
 }
 
-impl<'a> Escape for AsciiEscape<'a> {
+impl Escape for AsciiEscape<'_> {
     fn source_len(&self) -> usize {
         self.source.len()
     }
@@ -439,3 +421,21 @@ impl std::fmt::Display for BytesRepr<'_, '_> {
         self.write(formatter)
     }
 }
+
+#[cfg(test)]
+mod unicode_escape_tests {
+    use super::*;
+
+    #[test]
+    fn changed() {
+        fn test(s: &str) -> bool {
+            UnicodeEscape::new_repr(s).changed()
+        }
+        assert!(!test("hello"));
+        assert!(!test("'hello'"));
+        assert!(!test("\"hello\""));
+
+        assert!(test("'\"hello"));
+        assert!(test("hello\n"));
+    }
+}

literal/src/float.rs

@@ -6,6 +6,33 @@ pub fn parse_str(literal: &str) -> Option<f64> {
     parse_inner(literal.trim().as_bytes())
 }
 
+fn strip_underlines(literal: &[u8]) -> Option<Vec<u8>> {
+    let mut prev = b'\0';
+    let mut dup = Vec::<u8>::new();
+    for p in literal {
+        if *p == b'_' {
+            // Underscores are only allowed after digits.
+            if !prev.is_ascii_digit() {
+                return None;
+            }
+        } else {
+            dup.push(*p);
+            // Underscores are only allowed before digits.
+            if prev == b'_' && !p.is_ascii_digit() {
+                return None;
+            }
+        }
+        prev = *p;
+    }
+
+    // Underscores are not allowed at the end.
+    if prev == b'_' {
+        return None;
+    }
+
+    Some(dup)
+}
+
 pub fn parse_bytes(literal: &[u8]) -> Option<f64> {
     parse_inner(trim_slice(literal, |b| b.is_ascii_whitespace()))
 }
@@ -15,10 +42,10 @@ fn trim_slice<T>(v: &[T], mut trim: impl FnMut(&T) -> bool) -> &[T] {
     // it.take_while_ref(&mut trim).for_each(drop);
     // hmm.. `&mut slice::Iter<_>` is not `Clone`
     // it.by_ref().rev().take_while_ref(&mut trim).for_each(drop);
-    while it.clone().next().map_or(false, &mut trim) {
+    while it.clone().next().is_some_and(&mut trim) {
         it.next();
     }
-    while it.clone().next_back().map_or(false, &mut trim) {
+    while it.clone().next_back().is_some_and(&mut trim) {
         it.next_back();
     }
     it.as_slice()
@@ -28,11 +55,16 @@ fn parse_inner(literal: &[u8]) -> Option<f64> {
     use lexical_parse_float::{
         format::PYTHON3_LITERAL, FromLexicalWithOptions, NumberFormatBuilder, Options,
     };
+
+    // Use custom function for underline handling for now.
+    // For further information see https://github.com/Alexhuszagh/rust-lexical/issues/96.
+    let stripped = strip_underlines(literal)?;
+
     // lexical-core's format::PYTHON_STRING is inaccurate
     const PYTHON_STRING: u128 = NumberFormatBuilder::rebuild(PYTHON3_LITERAL)
         .no_special(false)
         .build();
-    f64::from_lexical_with_options::<PYTHON_STRING>(literal, &Options::new()).ok()
+    f64::from_lexical_with_options::<PYTHON_STRING>(&stripped, &Options::new()).ok()
 }
 
 pub fn is_integer(v: f64) -> bool {

parser/src/python.lalrpop

@@ -370,13 +370,14 @@ MatchStatement: ast::Stmt = {
             .last()
             .unwrap()
             .end();
+        let subject_range = (subjects.first().unwrap().start()..subjects.last().unwrap().end()).into();
         ast::Stmt::Match(
             ast::StmtMatch {
                 subject: Box::new(ast::Expr::Tuple(
                     ast::ExprTuple {
                         elts: subjects,
                         ctx: ast::ExprContext::Load,
-                        range: (location..end_location).into()
+                        range: subject_range,
                     },
                 )),
                 cases,

parser/src/python.rs

@@ -1,5 +1,5 @@
 // auto-generated: "lalrpop 0.20.0"
-// sha3: de5ffc51d44962eb297cbbf668fae33652eed69586405ebef2229fc02d183bc8
+// sha3: c2ba3f0f3de013733a18ba664f36f7f587254cc430656e553ceada96d33c409b
 use crate::{
     ast::{self as ast, Ranged, bigint::BigInt},
     lexer::{LexicalError, LexicalErrorType},
@@ -30842,13 +30842,14 @@ fn __action82<
             .last()
             .unwrap()
             .end();
+        let subject_range = (subjects.first().unwrap().start()..subjects.last().unwrap().end()).into();
         ast::Stmt::Match(
             ast::StmtMatch {
                 subject: Box::new(ast::Expr::Tuple(
                     ast::ExprTuple {
                         elts: subjects,
                         ctx: ast::ExprContext::Load,
-                        range: (location..end_location).into()
+                        range: subject_range,
                     },
                 )),
                 cases,

parser/src/snapshots/rustpython_parser__parser__tests__patma.snap

@@ -3783,7 +3783,7 @@ expression: parse_ast
             range: 2720..2760,
             subject: Tuple(
                 ExprTuple {
-                    range: 2720..2760,
+                    range: 2726..2730,
                     elts: [
                         Name(
                             ExprName {

parser/src/soft_keywords.rs

@@ -132,8 +132,8 @@ where
             }
         }
 
-        self.start_of_line = next.as_ref().map_or(false, |lex_result| {
-            lex_result.as_ref().map_or(false, |(tok, _)| {
+        self.start_of_line = next.as_ref().is_some_and(|lex_result| {
+            lex_result.as_ref().is_ok_and(|(tok, _)| {
                 #[cfg(feature = "full-lexer")]
                 if matches!(tok, Tok::NonLogicalNewline | Tok::Comment { .. }) {
                     return self.start_of_line;

vendored/src/text_size/mod.rs

@@ -23,8 +23,6 @@ mod range;
 mod size;
 mod traits;
 
-#[cfg(feature = "schemars")]
-mod schemars_impls;
 #[cfg(feature = "serde")]
 mod serde_impls;
 

vendored/src/text_size/size.rs

@@ -5,7 +5,6 @@ use {
         fmt, iter,
         num::TryFromIntError,
         ops::{Add, AddAssign, Sub, SubAssign},
-        u32,
     },
 };