[Boost-commit] svn:boost r43997 - website/public_html/beta/common/code

From: grafikrobot_at_[hidden]
Date: 2008-04-02 09:01:30

Author: grafik
Date: 2008-04-02 09:01:29 EDT (Wed, 02 Apr 2008)
New Revision: 43997
URL: http://svn.boost.org/trac/boost/changeset/43997

Log:
Add logic to attempt to recover content from illegal HTML doc pages.
Text files modified: 
   website/public_html/beta/common/code/boost_archive.php |    10 ++++++++++                              
   1 files changed, 10 insertions(+), 0 deletions(-)

Modified: website/public_html/beta/common/code/boost_archive.php
==============================================================================
--- website/public_html/beta/common/code/boost_archive.php	(original)
+++ website/public_html/beta/common/code/boost_archive.php	2008-04-02 09:01:29 EDT (Wed, 02 Apr 2008)
@@ -284,6 +284,16 @@
         preg_match('@</body>@i',$text,$body_end,PREG_OFFSET_CAPTURE);
         if (!isset($body_begin[0]))
         {
+            //~ Attempt to recover some content from illegal HTML that is missing the body tag.
+            preg_match('@</head>@i',$text,$body_begin,PREG_OFFSET_CAPTURE);
+        }
+        if (!isset($body_begin[0]))
+        {
+            //~ Attempt to recover some content from illegal HTML that is missing the body tag.
+            preg_match('@<html[^>]*>@i',$text,$body_begin,PREG_OFFSET_CAPTURE);
+        }
+        if (!isset($body_begin[0]))
+        {
             return;
         }
         else if (!isset($body_end[0]))