update-manpages: explicitly strip color codes

On some systems, help2man produces color codes in the output.  Remove
them to avoid manpage churn.

Also begin adding unit tests.

Change-Id: I3f0204b19d9cae524d3cb5fcfb61ee309b0931fc
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/349655
Tested-by: LaMont Jones <lamontjones@google.com>
Reviewed-by: Xin Li <delphij@google.com>

release/update-manpages

@@ -83,11 +83,6 @@ def main(argv):
     with multiprocessing.Pool() as pool:
       pool.map(partial(worker, cwd=tempdir, check=True), cmdlist)
 
-  regex = (
-      (r'(It was generated by help2man) [0-9.]+', '\g<1>.'),
-      (r'^\.IP\n(.*:)\n', '.SS \g<1>\n'),
-      (r'^\.PP\nDescription', '.SH DETAILS'),
-  )
   for tmp_path in MANDIR.glob('*.1.tmp'):
     path = tmp_path.parent / tmp_path.stem
     old_data = path.read_text() if path.exists() else ''
@@ -95,8 +90,7 @@ def main(argv):
     data = tmp_path.read_text()
     tmp_path.unlink()
 
-    for pattern, replacement in regex:
-      data = re.sub(pattern, replacement, data, flags=re.M)
+    data = replace_regex(data)
 
     # If the only thing that changed was the date, don't refresh.  This avoids
     # a lot of noise when only one file actually updates.
@@ -106,5 +100,25 @@ def main(argv):
       path.write_text(data)
 
 
+def replace_regex(data):
+  """Replace semantically null regexes in the data.
+
+  Args:
+    data: manpage text.
+
+  Returns:
+    Updated manpage text.
+  """
+  regex = (
+      (r'(It was generated by help2man) [0-9.]+', '\g<1>.'),
+      (r'^\033\[[0-9;]*m([^\033]*)\033\[m', '\g<1>'),
+      (r'^\.IP\n(.*:)\n', '.SS \g<1>\n'),
+      (r'^\.PP\nDescription', '.SH DETAILS'),
+  )
+  for pattern, replacement in regex:
+    data = re.sub(pattern, replacement, data, flags=re.M)
+  return data
+
+
 if __name__ == '__main__':
   sys.exit(main(sys.argv[1:]))