From fdf984d07de7f5cdc8e81b97c8c28a2c39e07b34 Mon Sep 17 00:00:00 2001
From: ThunderEX <thunderex@gmail.com>
Date: Wed, 24 May 2017 21:08:30 +0800
Subject: [PATCH] xmltok: Skip comment markup in XML.

This is provisional solution, at later time, comments may become tokenized
and fed to the caller, like other syntactic elements.
---
 xmltok/test.xml  | 1 +
 xmltok/xmltok.py | 8 ++++++++
 2 files changed, 9 insertions(+)

diff --git a/xmltok/test.xml b/xmltok/test.xml
index 1c3607f1..e13f548c 100644
--- a/xmltok/test.xml
+++ b/xmltok/test.xml
@@ -9,5 +9,6 @@
   baz
   
 </u:GetConnectionTypeInfo>
+<!-- Comments, you won't see me after tokenize. -->
 </s:Body>
 </s:Envelope>
diff --git a/xmltok/xmltok.py b/xmltok/xmltok.py
index c7143787..c46f2bd4 100644
--- a/xmltok/xmltok.py
+++ b/xmltok/xmltok.py
@@ -95,6 +95,14 @@ class XMLTokenizer:
                     yield from self.lex_attrs_till()
                     self.expect("?")
                     self.expect(">")
+                elif self.match("!"):
+                    self.expect("-")
+                    self.expect("-")
+                    last3 = ''
+                    while True:
+                        last3 = last3[-2:] + self.getch()
+                        if last3 == "-->":
+                            break
                 else:
                     tag = self.getnsident()
                     yield (START_TAG, tag)