Fixed issue where tags at the beginning of summary not parsed correctly.

.gitmodules

@@ -4,3 +4,6 @@
 [submodule "tests/test_repo2"]
 	path = tests/test_repo2
 	url = https://github.com/saschagrunert/test2
+[submodule "tests/test_repo3"]
+	path = tests/test_repo3
+	url = git@github.com:jwir3/ghp-release-cleaner

src/lib.rs

@@ -1010,6 +1010,20 @@ mod tests {
         assert!(journal.generate_template().is_ok());
     }
 
+    #[test]
+    fn output_correctly_categorized() {
+      let mut journal = GitJournal::new("./tests/test_repo3").unwrap();
+      assert!(journal
+          .parse_log("444b3a64355557844beb848d80051ada2c00176a..HEAD", "rc", 0, true, false, None, None)
+          .is_ok());
+      assert_eq!(journal.parser.result.len(), 4);
+      assert_eq!(journal.parser.result[0].name, "Unreleased");
+      assert_eq!(journal.parser.result[0].commits.len(), 2);
+      assert_eq!(journal.parser.result[0].commits[0].summary.tags.len(), 1);
+      assert!(journal.print_log(false, None, Some("CHANGELOG.md")).is_ok());
+      assert!(journal.print_log(true, None, Some("CHANGELOG.md")).is_ok());
+    }
+
     #[test]
     fn path_failure() {
         assert!(GitJournal::new("/etc/").is_err());

src/parser.rs

@@ -776,7 +776,7 @@ pub struct FooterElement {
 }
 
 lazy_static! {
-    static ref RE_TAGS: Regex = Regex::new(r"[ \n]:(.*?):").unwrap();
+    static ref RE_TAGS: Regex = Regex::new(r"[ \n]*:(.*?):").unwrap();
     static ref RE_FOOTER: Regex = RegexBuilder::new(r"^([\w-]+):\s(.*)$")
         .multi_line(true)
         .build()
@@ -827,12 +827,25 @@ impl Parser {
     }
 
     fn parse_summary<'a>(&mut self, input: &'a [u8]) -> ParserResult<'a, SummaryElement> {
+        // Before we do anything, let's parse the entire summary line for tags so that we don't
+        // skip any tags that happen before, say, the category.
+        let entire_string = str_or_empty(input);
+        let mut tags = vec![];
+        for cap in RE_TAGS.captures_iter(entire_string) {
+          tags.push((&cap[1]).to_string());
+        }
+
         let (input, p_prefix) = opt(separated_pair(alpha1, char('-'), digit1))(input)?;
         let (input, _) = space0(input)?;
         let (input, p_category) = self.parse_category(input)?;
         let (input, _) = space1(input)?;
         let (input, p_tags_rest) = map(rest, Self::parse_and_consume_tags)(input)?;
 
+        let p_tags_rest_vec = p_tags_rest.0;
+        tags.extend(p_tags_rest_vec);
+        tags.sort();
+        tags.dedup();
+
         Ok((
             input,
             SummaryElement {
@@ -841,7 +854,7 @@ impl Parser {
                     format!("{}-{}", str_or_empty(p.0), str_or_empty(p.1))
                 }),
                 category: p_category.to_owned(),
-                tags: p_tags_rest.0,
+                tags: tags,
                 text: p_tags_rest.1,
             },
         ))
@@ -850,6 +863,7 @@ impl Parser {
     fn parse_and_consume_tags(input: &[u8]) -> (Vec<String>, String) {
         let string = str_or_empty(input);
         let mut tags = vec![];
+
         for cap in RE_TAGS.captures_iter(string) {
             if let Some(tag) = cap.get(1) {
                 tags.extend(
@@ -888,6 +902,7 @@ impl Parser {
             .next()
             .ok_or_else(|| format_err!("Summar line parsing: Commit message length too small."))?
             .trim();
+
         let mut parsed_summary = match self.clone().parse_summary(summary_line.as_bytes()) {
             Ok((_, parsed)) => parsed,
             _ => bail!("Summary parsing failed: '{}'", summary_line),