diff --git a/cli/loader/src/lib.rs b/cli/loader/src/lib.rs
index e8d6ea9e..d3428dc7 100644
--- a/cli/loader/src/lib.rs
+++ b/cli/loader/src/lib.rs
@@ -503,8 +503,8 @@ impl Loader {
         let library = unsafe { Library::new(&library_path) }
             .with_context(|| format!("Error opening dynamic library {library_path:?}"))?;
         let language = unsafe {
-            let language_fn: Symbol<unsafe extern "C" fn() -> Language> = library
-                .get(language_fn_name.as_bytes())
+            let language_fn = library
+                .get::<Symbol<unsafe extern "C" fn() -> Language>>(language_fn_name.as_bytes())
                 .with_context(|| format!("Failed to load symbol {language_fn_name}"))?;
             language_fn()
         };
diff --git a/cli/src/generate/build_tables/build_lex_table.rs b/cli/src/generate/build_tables/build_lex_table.rs
index 09e8b073..bc65447c 100644
--- a/cli/src/generate/build_tables/build_lex_table.rs
+++ b/cli/src/generate/build_tables/build_lex_table.rs
@@ -26,7 +26,7 @@ pub fn build_lex_table(
         LexTable::default()
     };
 
-    let mut parse_state_ids_by_token_set: Vec<(TokenSet, Vec<ParseStateId>)> = Vec::new();
+    let mut parse_state_ids_by_token_set = Vec::<(TokenSet, Vec<ParseStateId>)>::new();
     for (i, state) in parse_table.states.iter().enumerate() {
         let tokens = state
             .terminal_entries
diff --git a/cli/src/generate/build_tables/build_parse_table.rs b/cli/src/generate/build_tables/build_parse_table.rs
index 30850d3d..a9be76a2 100644
--- a/cli/src/generate/build_tables/build_parse_table.rs
+++ b/cli/src/generate/build_tables/build_parse_table.rs
@@ -455,7 +455,7 @@ impl<'a> ParseTableBuilder<'a> {
         // REDUCE-REDUCE conflicts where all actions have the *same*
         // precedence, and there can still be SHIFT/REDUCE conflicts.
         let mut considered_associativity = false;
-        let mut shift_precedence: Vec<(&Precedence, Symbol)> = Vec::new();
+        let mut shift_precedence = Vec::<(&Precedence, Symbol)>::new();
         let mut conflicting_items = HashSet::new();
         for (item, lookaheads) in &item_set.entries {
             if let Some(step) = item.step() {
diff --git a/cli/src/generate/nfa.rs b/cli/src/generate/nfa.rs
index 2433f520..66f78074 100644
--- a/cli/src/generate/nfa.rs
+++ b/cli/src/generate/nfa.rs
@@ -464,7 +464,7 @@ impl<'a> NfaCursor<'a> {
     fn group_transitions<'b>(
         iter: impl Iterator<Item = (&'b CharacterSet, bool, i32, u32)>,
     ) -> Vec<NfaTransition> {
-        let mut result: Vec<NfaTransition> = Vec::new();
+        let mut result = Vec::<NfaTransition>::new();
         for (chars, is_sep, prec, state) in iter {
             let mut chars = chars.clone();
             let mut i = 0;
diff --git a/cli/src/generate/prepare_grammar/extract_tokens.rs b/cli/src/generate/prepare_grammar/extract_tokens.rs
index 4ba89103..7d87bbd2 100644
--- a/cli/src/generate/prepare_grammar/extract_tokens.rs
+++ b/cli/src/generate/prepare_grammar/extract_tokens.rs
@@ -67,10 +67,10 @@ pub(super) fn extract_tokens(
         .expected_conflicts
         .into_iter()
         .map(|conflict| {
-            let mut result: Vec<_> = conflict
+            let mut result = conflict
                 .iter()
                 .map(|symbol| symbol_replacer.replace_symbol(*symbol))
-                .collect();
+                .collect::<Vec<_>>();
             result.sort_unstable();
             result.dedup();
             result
diff --git a/cli/src/generate/prepare_grammar/process_inlines.rs b/cli/src/generate/prepare_grammar/process_inlines.rs
index c718a429..e37180b3 100644
--- a/cli/src/generate/prepare_grammar/process_inlines.rs
+++ b/cli/src/generate/prepare_grammar/process_inlines.rs
@@ -362,10 +362,10 @@ mod tests {
 
         let inline_map = process_inlines(&grammar, &LexicalGrammar::default()).unwrap();
 
-        let productions: Vec<&Production> = inline_map
+        let productions = inline_map
             .inlined_productions(&grammar.variables[0].productions[0], 1)
             .unwrap()
-            .collect();
+            .collect::<Vec<_>>();
 
         assert_eq!(
             productions.iter().copied().cloned().collect::<Vec<_>>(),
@@ -461,10 +461,10 @@ mod tests {
 
         let inline_map = process_inlines(&grammar, &LexicalGrammar::default()).unwrap();
 
-        let productions: Vec<_> = inline_map
+        let productions = inline_map
             .inlined_productions(&grammar.variables[0].productions[0], 0)
             .unwrap()
-            .collect();
+            .collect::<Vec<_>>();
 
         assert_eq!(
             productions.iter().copied().cloned().collect::<Vec<_>>(),
diff --git a/cli/src/generate/render.rs b/cli/src/generate/render.rs
index f111ed08..b65e2ebe 100644
--- a/cli/src/generate/render.rs
+++ b/cli/src/generate/render.rs
@@ -675,7 +675,7 @@ impl Generator {
 
         // For each lex state, compute a summary of the code that needs to be
         // generated.
-        let state_transition_summaries: Vec<Vec<TransitionSummary>> = lex_table
+        let state_transition_summaries = lex_table
             .states
             .iter()
             .map(|state| {
@@ -732,7 +732,7 @@ impl Generator {
                     })
                     .collect()
             })
-            .collect();
+            .collect::<Vec<Vec<_>>>();
 
         // Generate a helper function for each large character set.
         let mut sorted_large_char_sets = large_character_sets.iter().collect::<Vec<_>>();
@@ -1153,7 +1153,7 @@ impl Generator {
 
             let mut index = 0;
             let mut small_state_indices = Vec::new();
-            let mut symbols_by_value: HashMap<(usize, SymbolType), Vec<Symbol>> = HashMap::new();
+            let mut symbols_by_value = HashMap::<(usize, SymbolType), Vec<Symbol>>::new();
             for state in self.parse_table.states.iter().skip(self.large_state_count) {
                 small_state_indices.push(index);
                 symbols_by_value.clear();
diff --git a/cli/src/parse.rs b/cli/src/parse.rs
index ee395205..4849bda3 100644
--- a/cli/src/parse.rs
+++ b/cli/src/parse.rs
@@ -208,7 +208,7 @@ pub fn parse_file_at_path(parser: &mut Parser, opts: &ParseFileOptions) -> Resul
             let mut indent_level = 0;
             let mut did_visit_children = false;
             let mut had_named_children = false;
-            let mut tags: Vec<&str> = Vec::new();
+            let mut tags = Vec::<&str>::new();
             writeln!(&mut stdout, "<?xml version=\"1.0\"?>")?;
             loop {
                 let node = cursor.node();
diff --git a/cli/src/tests/parser_test.rs b/cli/src/tests/parser_test.rs
index 1df134e3..ecd54fcf 100644
--- a/cli/src/tests/parser_test.rs
+++ b/cli/src/tests/parser_test.rs
@@ -158,10 +158,10 @@ fn test_parsing_with_custom_utf16_input() {
     let mut parser = Parser::new();
     parser.set_language(&get_language("rust")).unwrap();
 
-    let lines: Vec<Vec<u16>> = ["pub fn foo() {", "  1", "}"]
+    let lines = ["pub fn foo() {", "  1", "}"]
         .iter()
-        .map(|s| s.encode_utf16().collect())
-        .collect();
+        .map(|s| s.encode_utf16().collect::<Vec<_>>())
+        .collect::<Vec<_>>();
 
     let tree = parser
         .parse_utf16_with(
@@ -1014,11 +1014,11 @@ fn test_parsing_error_in_invalid_included_ranges() {
 #[test]
 fn test_parsing_utf16_code_with_errors_at_the_end_of_an_included_range() {
     let source_code = "<script>a.</script>";
-    let utf16_source_code: Vec<u16> = source_code
+    let utf16_source_code = source_code
         .as_bytes()
         .iter()
         .map(|c| u16::from(*c))
-        .collect();
+        .collect::<Vec<_>>();
 
     let start_byte = 2 * source_code.find("a.").unwrap();
     let end_byte = 2 * source_code.find("</script>").unwrap();
diff --git a/cli/src/tests/tags_test.rs b/cli/src/tests/tags_test.rs
index 5719269b..6139f732 100644
--- a/cli/src/tests/tags_test.rs
+++ b/cli/src/tests/tags_test.rs
@@ -397,14 +397,14 @@ fn test_tags_via_c_api() {
         })
         .unwrap();
 
-        let syntax_types: Vec<&str> = unsafe {
+        let syntax_types = unsafe {
             let mut len: u32 = 0;
             let ptr =
                 c::ts_tagger_syntax_kinds_for_scope_name(tagger, c_scope_name.as_ptr(), &mut len);
             slice::from_raw_parts(ptr, len as usize)
                 .iter()
                 .map(|i| CStr::from_ptr(*i).to_str().unwrap())
-                .collect()
+                .collect::<Vec<_>>()
         };
 
         assert_eq!(
diff --git a/cli/src/tests/text_provider_test.rs b/cli/src/tests/text_provider_test.rs
index e04beca8..b0b70243 100644
--- a/cli/src/tests/text_provider_test.rs
+++ b/cli/src/tests/text_provider_test.rs
@@ -72,7 +72,7 @@ fn test_text_provider_for_string() {
 
 #[test]
 fn test_text_provider_for_box_of_str_slice() {
-    let text: Box<str> = "// comment".to_owned().into_boxed_str();
+    let text = "// comment".to_owned().into_boxed_str();
 
     check_parsing(text.as_bytes(), text.as_bytes());
     check_parsing(<_ as AsRef<str>>::as_ref(&text), text.as_bytes());
@@ -82,7 +82,7 @@ fn test_text_provider_for_box_of_str_slice() {
 
 #[test]
 fn test_text_provider_for_box_of_bytes_slice() {
-    let text: Box<[u8]> = "// comment".to_owned().into_boxed_str().into_boxed_bytes();
+    let text = "// comment".to_owned().into_boxed_str().into_boxed_bytes();
 
     check_parsing(text.as_ref(), text.as_ref());
     check_parsing(text.as_ref(), &*text);
@@ -91,15 +91,14 @@ fn test_text_provider_for_box_of_bytes_slice() {
 
 #[test]
 fn test_text_provider_for_vec_of_bytes() {
-    let text: Vec<u8> = "// comment".to_owned().into_bytes();
+    let text = "// comment".to_owned().into_bytes();
 
     check_parsing(&*text, &*text);
 }
 
 #[test]
 fn test_text_provider_for_arc_of_bytes_slice() {
-    let text: Vec<u8> = "// comment".to_owned().into_bytes();
-    let text: Arc<[u8]> = Arc::from(text);
+    let text: Arc<[u8]> = Arc::from("// comment".to_owned().into_bytes());
 
     check_parsing(&*text, &*text);
     check_parsing(text.as_ref(), text.as_ref());
@@ -149,7 +148,7 @@ fn test_text_provider_callback_with_owned_bytes_vec_slice() {
                 .unwrap_or_default()
         },
         |_node: Node<'_>| {
-            let slice: Vec<u8> = text.to_owned().into_bytes();
+            let slice = text.to_owned().into_bytes();
             iter::once(slice)
         },
     );
diff --git a/cli/src/wasm.rs b/cli/src/wasm.rs
index 7ecdf50c..11dbb5c5 100644
--- a/cli/src/wasm.rs
+++ b/cli/src/wasm.rs
@@ -48,7 +48,7 @@ pub fn compile_language_to_wasm(
 
     // Exit with an error if the external scanner uses symbols from the
     // C or C++ standard libraries that aren't available to wasm parsers.
-    let stdlib_symbols: Vec<_> = wasm_stdlib_symbols().collect();
+    let stdlib_symbols = wasm_stdlib_symbols().collect::<Vec<_>>();
     let dylink_symbols = [
         "__indirect_function_table",
         "__memory_base",