From d0b4eb76001049a23fe038025cfa56472703fa70 Mon Sep 17 00:00:00 2001
From: Michael Williamson <mike@zwobble.org>
Date: Thu, 18 May 2017 20:12:25 +0100
Subject: Avoid parsing URI when replacing fragment

---
 mammoth/compat.py            |  9 ---------
 mammoth/docx/body_xml.py     |  5 ++---
 mammoth/docx/uris.py         |  7 +++++++
 tests/docx/body_xml_tests.py | 12 ++++++++++++
 4 files changed, 21 insertions(+), 12 deletions(-)
 delete mode 100644 mammoth/compat.py

diff --git a/mammoth/compat.py b/mammoth/compat.py
deleted file mode 100644
index 5fb9acf..0000000
--- a/mammoth/compat.py
+++ /dev/null
@@ -1,9 +0,0 @@
-import sys
-
-if sys.version_info < (3, ):
-    from urlparse import urldefrag
-else:
-    from urllib.parse import urldefrag
-    
-    
-__all__ = ["urldefrag"]
diff --git a/mammoth/docx/body_xml.py b/mammoth/docx/body_xml.py
index 0724ac8..cc4e98f 100644
--- a/mammoth/docx/body_xml.py
+++ b/mammoth/docx/body_xml.py
@@ -4,11 +4,10 @@ import re
 from .. import documents
 from .. import results
 from .. import lists
-from ..compat import urldefrag
 from . import complex_fields
 from .xmlparser import node_types, XmlElement
 from .styles_xml import Styles
-from .uris import uri_to_zip_entry_name
+from .uris import replace_fragment, uri_to_zip_entry_name
 
 
 def reader(
@@ -296,7 +295,7 @@ def _create_reader(numbering, content_types, relationships, styles, docx_file, f
         if relationship_id is not None:
             href = relationships[relationship_id].target
             if anchor is not None:
-                href = urldefrag(href)[0] + "#" + anchor
+                href = replace_fragment(href, anchor)
             
             return children_result.map(lambda children: documents.hyperlink(href=href, children=children))
         elif anchor is not None:
diff --git a/mammoth/docx/uris.py b/mammoth/docx/uris.py
index 59f3ffe..892078e 100644
--- a/mammoth/docx/uris.py
+++ b/mammoth/docx/uris.py
@@ -3,3 +3,10 @@ def uri_to_zip_entry_name(base, uri):
         return uri[1:]
     else:
         return base + "/" + uri
+
+
+def replace_fragment(uri, fragment):
+    hash_index = uri.find("#")
+    if hash_index != -1:
+        uri = uri[:hash_index]
+    return uri + "#" + fragment
diff --git a/tests/docx/body_xml_tests.py b/tests/docx/body_xml_tests.py
index 1277238..cccf472 100644
--- a/tests/docx/body_xml_tests.py
+++ b/tests/docx/body_xml_tests.py
@@ -623,6 +623,18 @@ class HyperlinkTests(object):
             _read_and_get_document_xml_element(element, relationships=relationships)
         )
         
+    @istest
+    def existing_fragment_is_replaced_when_anchor_is_set_on_external_link(self):
+        relationships = Relationships({
+            "r42": Relationship(target="http://example.com/#previous")
+        })
+        run_element = xml_element("w:r")
+        element = xml_element("w:hyperlink", {"r:id": "r42", "w:anchor": "fragment"}, [run_element])
+        assert_equal(
+            documents.hyperlink(href="http://example.com/#fragment", children=[documents.run([])]),
+            _read_and_get_document_xml_element(element, relationships=relationships)
+        )
+        
     @istest
     def hyperlink_is_read_as_internal_hyperlink_if_it_has_an_anchor_attribute(self):
         run_element = xml_element("w:r")
-- 
cgit v1.2.3