From a00831c847b7efd81be520ea9b5d02f70555351f Mon Sep 17 00:00:00 2001
From: Ed Page <epage@duosecurity.com>
Date: Thu, 31 Oct 2019 16:02:41 -0600
Subject: [PATCH 1/2] fix: Ignore numbers as identifiers

---
 typos/src/tokens.rs | 27 +++++++++++++++++++++++----
 1 file changed, 23 insertions(+), 4 deletions(-)

diff --git a/typos/src/tokens.rs b/typos/src/tokens.rs
index acc1a36..fa2c755 100644
--- a/typos/src/tokens.rs
+++ b/typos/src/tokens.rs
@@ -78,23 +78,33 @@ impl Parser {
     }
 
     pub fn parse<'c>(&'c self, content: &'c str) -> impl Iterator<Item = Identifier<'c>> {
-        let ignore_hex = self.ignore_hex;
         self.words_str
             .find_iter(content)
-            .filter(move |m| !ignore_hex || !is_hex(m.as_str().as_bytes()))
+            .filter(move |m| self.accept(m.as_str().as_bytes()))
             .map(|m| Identifier::new_unchecked(m.as_str(), m.start()))
     }
 
     pub fn parse_bytes<'c>(&'c self, content: &'c [u8]) -> impl Iterator<Item = Identifier<'c>> {
-        let ignore_hex = self.ignore_hex;
         self.words_bytes
             .find_iter(content)
-            .filter(move |m| !ignore_hex || !is_hex(m.as_bytes()))
+            .filter(move |m| self.accept(m.as_bytes()))
             .filter_map(|m| {
                 let s = std::str::from_utf8(m.as_bytes()).ok();
                 s.map(|s| Identifier::new_unchecked(s, m.start()))
             })
     }
+
+    fn accept(&self, contents: &[u8]) -> bool {
+        if is_number(contents) {
+            return false;
+        };
+
+        if self.ignore_hex {
+            return !is_hex(contents);
+        }
+
+        true
+    }
 }
 
 impl Default for Parser {
@@ -103,6 +113,15 @@ impl Default for Parser {
     }
 }
 
+fn is_number(ident: &[u8]) -> bool {
+    lazy_static::lazy_static! {
+        // `_`: number literal separator in Rust and other languages
+        // `'`: number literal separator in C++
+        static ref DIGITS: regex::bytes::Regex = regex::bytes::Regex::new(r#"^[0-9_']+$"#).unwrap();
+    }
+    DIGITS.is_match(ident)
+}
+
 fn is_hex(ident: &[u8]) -> bool {
     lazy_static::lazy_static! {
         // `_`: number literal separator in Rust and other languages

From 68cd36d0de90226dbc9d31c2ce6d8bf6b69adb5c Mon Sep 17 00:00:00 2001
From: Ed Page <epage@duosecurity.com>
Date: Fri, 1 Nov 2019 16:25:19 -0600
Subject: [PATCH 2/2] perf: Only do hex check if digits are in identifiers

---
 typos/src/tokens.rs | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/typos/src/tokens.rs b/typos/src/tokens.rs
index fa2c755..883f1a5 100644
--- a/typos/src/tokens.rs
+++ b/typos/src/tokens.rs
@@ -50,7 +50,7 @@ impl ParserBuilder {
         Parser {
             words_str,
             words_bytes,
-            ignore_hex: self.ignore_hex,
+            ignore_hex: self.ignore_hex && self.include_digits,
         }
     }
 }