From 8376f40054002a2dbba7f718aef82c6a8c5f7bc6 Mon Sep 17 00:00:00 2001
From: William G Underwood <42812654+WGUNDERWOOD@users.noreply.github.com>
Date: Wed, 1 May 2024 22:03:32 -0400
Subject: [PATCH] Simplified comment matching regex

---
 src/regexes.rs         | 3 +--
 src/subs.rs            | 3 +--
 tests/comments_in.tex  | 4 ++++
 tests/comments_out.tex | 4 ++++
 4 files changed, 10 insertions(+), 4 deletions(-)

diff --git a/src/regexes.rs b/src/regexes.rs
index 5621f8b..b73350b 100644
--- a/src/regexes.rs
+++ b/src/regexes.rs
@@ -7,8 +7,7 @@ lazy_static! {
     pub static ref RE_NEWLINES: Regex = Regex::new(r"\n\n\n+").unwrap();
     pub static ref RE_TABS: Regex = Regex::new(r"\t").unwrap();
     pub static ref RE_TRAIL: Regex = Regex::new(r" +\n").unwrap();
-    pub static ref RE_PERCENT: Regex = Regex::new(r"\\\%").unwrap();
-    pub static ref RE_COMMENT: Regex = Regex::new(r"\%.*").unwrap();
+    pub static ref RE_COMMENT: Regex = Regex::new(r"(?P<text>.*^|[^\\])(?P<comm>\%.*$)").unwrap();
     pub static ref RE_ITEM: Regex = Regex::new(r".*\\item.*").unwrap();
     pub static ref RE_DOCUMENT_BEGIN: Regex =
         Regex::new(r".*\\begin\{document\}.*").unwrap();
diff --git a/src/subs.rs b/src/subs.rs
index a847460..dd2d8d4 100644
--- a/src/subs.rs
+++ b/src/subs.rs
@@ -42,6 +42,5 @@ pub fn remove_trailing_spaces(file: &str) -> String {
 //}
 
 pub fn remove_comment(line: &str) -> String {
-    let new_line = RE_PERCENT.replace_all(line, "").to_string();
-    RE_COMMENT.replace_all(&new_line, "").to_string()
+    RE_COMMENT.replace_all(line, "$text").to_string()
 }
diff --git a/tests/comments_in.tex b/tests/comments_in.tex
index 0fba314..c6ffac9 100644
--- a/tests/comments_in.tex
+++ b/tests/comments_in.tex
@@ -15,4 +15,8 @@
 % Similarly they might contain \begin{align}
 % unmatched environment tags.
 
+This is a percent sign \% and not a comment
+
+Some lines might have both \% percents % and comments \end{align}
+
 \end{document}
diff --git a/tests/comments_out.tex b/tests/comments_out.tex
index 853727a..bc25171 100644
--- a/tests/comments_out.tex
+++ b/tests/comments_out.tex
@@ -15,4 +15,8 @@
 % Similarly they might contain \begin{align}
 % unmatched environment tags.
 
+This is a percent sign \% and not a comment
+
+Some lines might have both \% percents % and comments \end{align}
+
 \end{document}