From b54347f8d1ec97ef9a53d96255676a82ec685ec7 Mon Sep 17 00:00:00 2001
From: chdorner <christof@chdorner.com>
Date: Wed, 1 Jun 2016 11:21:55 +0200
Subject: [PATCH] Allow rewriting of empty srcset attributes

Strictly speaking a `srcset` attribute must consist of one or more
strings
(http://w3c.github.io/html/semantics-embedded-content.html#element-attrdef-img-srcset)
However are websites out there that specify an empty string as the
value.

This commit makes sure that the rewriting does not break and just
returns an empty string.
---
 pywb/rewrite/html_rewriter.py           | 1 +
 pywb/rewrite/test/test_html_rewriter.py | 4 ++++
 2 files changed, 5 insertions(+)

diff --git a/pywb/rewrite/html_rewriter.py b/pywb/rewrite/html_rewriter.py
index e57f8591..171b4e69 100644
--- a/pywb/rewrite/html_rewriter.py
+++ b/pywb/rewrite/html_rewriter.py
@@ -211,6 +211,7 @@ class HTMLRewriterMixin(object):
     def _rewrite_srcset(self, value, mod=''):
         values = value.split(',')
         values = map(lambda x: self._rewrite_url(x.strip()), values)
+        values = [v for v in values if v is not None]
         return ', '.join(values)
 
     def _rewrite_css(self, css_content):
diff --git a/pywb/rewrite/test/test_html_rewriter.py b/pywb/rewrite/test/test_html_rewriter.py
index 818bd114..13063936 100644
--- a/pywb/rewrite/test/test_html_rewriter.py
+++ b/pywb/rewrite/test/test_html_rewriter.py
@@ -115,6 +115,10 @@ r"""
 >>> parse('<img srcset="//example.com/1x 1x, //example.com/foo 2x, https://example.com/bar 4x">')
 <img srcset="/web/20131226101010///example.com/1x 1x, /web/20131226101010///example.com/foo 2x, /web/20131226101010/https://example.com/bar 4x">
 
+# empty srcset attrib
+>>> parse('<img srcset="">')
+<img srcset="">
+
 # Script tag
 >>> parse('<script>window.location = "http://example.com/a/b/c.html"</script>')
 <script>window.WB_wombat_location = "/web/20131226101010/http://example.com/a/b/c.html"</script>