Adding upstream version 26.2.1.

Signed-off-by: Daniel Baumann <daniel@debian.org>

sqlglotrs/Cargo.lock

@@ -448,6 +448,12 @@ version = "0.8.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "2b15c43186be67a4fd63bee50d0303afffcef381492ebe2c5d87f324e1b8815c"
 
+[[package]]
+name = "rustc-hash"
+version = "2.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c7fb8039b3032c191086b10f11f319a6e99e1e82889c5cc6046f515c9db1d497"
+
 [[package]]
 name = "ryu"
 version = "1.0.18"
@@ -497,10 +503,11 @@ dependencies = [
 
 [[package]]
 name = "sqlglotrs"
-version = "0.3.4"
+version = "0.3.5"
 dependencies = [
  "criterion",
  "pyo3",
+ "rustc-hash",
  "serde",
  "serde_json",
  "sqlglotrs",

sqlglotrs/Cargo.toml

@@ -1,12 +1,13 @@
 [package]
 name = "sqlglotrs"
-version = "0.3.4"
+version = "0.3.5"
 edition = "2021"
 license = "MIT"
 
 [lib]
 name = "sqlglotrs"
 crate-type = ["cdylib", "rlib"]
+bench = false
 
 [[bench]]
 name = "long"
@@ -19,6 +20,7 @@ profiling = ["serde", "serde_json"]
 
 [dependencies]
 pyo3 = {version ="0.22.6", features = ["auto-initialize"]}
+rustc-hash = { version = "2.1" }
 
 # Optional dependencies used for profiling
 serde = { version = "1", features = ["derive"] , optional = true }

sqlglotrs/src/settings.rs

@@ -1,6 +1,6 @@
-use std::collections::{HashMap, HashSet};
-
 use pyo3::prelude::*;
+use rustc_hash::FxHashMap as HashMap;
+use rustc_hash::FxHashSet as HashSet;
 
 pub type TokenType = u16;
 

sqlglotrs/src/token.rs

@@ -49,13 +49,11 @@ impl Token {
     pub fn append_comments(&self, comments: &mut Vec<String>) {
         Python::with_gil(|py| {
             let pylist = self.comments.bind(py);
-            for comment in comments.iter() {
+            for comment in comments.drain(..) {
                 if let Err(_) = pylist.append(comment) {
                     panic!("Failed to append comments to the Python list");
                 }
             }
         });
-        // Simulate `Vec::append`.
-        let _ = std::mem::replace(comments, Vec::new());
     }
 }

sqlglotrs/src/tokenizer.rs

@@ -23,14 +23,11 @@ pub struct Tokenizer {
 impl Tokenizer {
     #[new]
     pub fn new(settings: TokenizerSettings, token_types: TokenTypeSettings) -> Tokenizer {
-        let mut keyword_trie = Trie::new();
-        let single_token_strs: Vec<String> = settings
-            .single_tokens
-            .keys()
-            .map(|s| s.to_string())
-            .collect();
-        let trie_filter =
-            |key: &&String| key.contains(" ") || single_token_strs.iter().any(|t| key.contains(t));
+        let mut keyword_trie = Trie::default();
+
+        let trie_filter = |key: &&String| {
+            key.contains(" ") || settings.single_tokens.keys().any(|&t| key.contains(t))
+        };
 
         keyword_trie.add(settings.keywords.keys().filter(trie_filter));
         keyword_trie.add(settings.comments.keys().filter(trie_filter));
@@ -114,7 +111,7 @@ impl<'a> TokenizerState<'a> {
 
     fn tokenize(&mut self) -> Result<Vec<Token>, TokenizerError> {
         self.scan(None)?;
-        Ok(std::mem::replace(&mut self.tokens, Vec::new()))
+        Ok(std::mem::take(&mut self.tokens))
     }
 
     fn scan(&mut self, until_peek_char: Option<char>) -> Result<(), TokenizerError> {
@@ -146,7 +143,7 @@ impl<'a> TokenizerState<'a> {
             }
 
             if !self.settings.white_space.contains_key(&self.current_char) {
-                if self.current_char.is_digit(10) {
+                if self.current_char.is_ascii_digit() {
                     self.scan_number()?;
                 } else if let Some(identifier_end) =
                     self.settings.identifiers.get(&self.current_char)
@@ -205,7 +202,7 @@ impl<'a> TokenizerState<'a> {
     }
 
     fn char_at(&self, index: usize) -> Result<char, TokenizerError> {
-        self.sql.get(index).map(|c| *c).ok_or_else(|| {
+        self.sql.get(index).copied().ok_or_else(|| {
             self.error(format!(
                 "Index {} is out of bound (size {})",
                 index, self.size
@@ -237,7 +234,7 @@ impl<'a> TokenizerState<'a> {
             self.column,
             self.start,
             self.current - 1,
-            std::mem::replace(&mut self.comments, Vec::new()),
+            std::mem::take(&mut self.comments),
         ));
 
         // If we have either a semicolon or a begin token before the command's token, we'll parse
@@ -503,7 +500,7 @@ impl<'a> TokenizerState<'a> {
         let mut scientific = 0;
 
         loop {
-            if self.peek_char.is_digit(10) {
+            if self.peek_char.is_ascii_digit() {
                 self.advance(1)?;
             } else if self.peek_char == '.' && !decimal {
                 if self.tokens.last().map(|t| t.token_type) == Some(self.token_types.parameter) {
@@ -537,8 +534,7 @@ impl<'a> TokenizerState<'a> {
                             .numeric_literals
                             .get(&literal.to_uppercase())
                             .unwrap_or(&String::from("")),
-                    )
-                    .map(|x| *x);
+                    ).copied();
 
                 let replaced = literal.replace("_", "");
 
@@ -607,8 +603,7 @@ impl<'a> TokenizerState<'a> {
             } else {
                 self.settings
                     .keywords
-                    .get(&self.text().to_uppercase())
-                    .map(|x| *x)
+                    .get(&self.text().to_uppercase()).copied()
                     .unwrap_or(self.token_types.var)
             };
         self.add(token_type, None)
@@ -718,13 +713,13 @@ impl<'a> TokenizerState<'a> {
             if i == 0 {
                 self.is_alphabetic_or_underscore(c)
             } else {
-                self.is_alphabetic_or_underscore(c) || c.is_digit(10)
+                self.is_alphabetic_or_underscore(c) || c.is_ascii_digit()
             }
         })
     }
 
     fn is_numeric(&mut self, s: &str) -> bool {
-        s.chars().all(|c| c.is_digit(10))
+        s.chars().all(|c| c.is_ascii_digit())
     }
 
     fn extract_value(&mut self) -> Result<String, TokenizerError> {

sqlglotrs/src/trie.rs

@@ -1,6 +1,6 @@
-use std::collections::HashMap;
+use rustc_hash::FxHashMap as HashMap;
 
-#[derive(Debug)]
+#[derive(Debug, Default)]
 pub struct TrieNode {
     is_word: bool,
     children: HashMap<char, TrieNode>,
@@ -35,21 +35,12 @@ impl TrieNode {
     }
 }
 
-#[derive(Debug)]
+#[derive(Debug, Default)]
 pub struct Trie {
     pub root: TrieNode,
 }
 
 impl Trie {
-    pub fn new() -> Self {
-        Trie {
-            root: TrieNode {
-                is_word: false,
-                children: HashMap::new(),
-            },
-        }
-    }
-
     pub fn add<'a, I>(&mut self, keys: I)
     where
         I: Iterator<Item = &'a String>,
@@ -59,7 +50,7 @@ impl Trie {
             for c in key.chars() {
                 current = current.children.entry(c).or_insert(TrieNode {
                     is_word: false,
-                    children: HashMap::new(),
+                    children: HashMap::default(),
                 });
             }
             current.is_word = true;