From ecbd0c970a2d07dd7df73f0fbfb5c9ae75d557a9 Mon Sep 17 00:00:00 2001
From: jayGattusoNLNZ <jay.gattuso@dia.govt.nz>
Date: Tue, 26 Nov 2013 11:49:15 +1300
Subject: [PATCH] Update arc.py

ARCv 2 files can contain an <arcmetadata> XML block that confuses the header parser when the fileobj is presented.
This edit skips the XML block and allows the first record header to be properly discovered.
---
 warc/arc.py | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/warc/arc.py b/warc/arc.py
index 5889587..1d20916 100644
--- a/warc/arc.py
+++ b/warc/arc.py
@@ -346,6 +346,17 @@ def _read_arc_record(self):
         header = self.fileobj.readline()
         while header and header.strip() == "":
             header = self.fileobj.readline()
+            
+        #JG: this block stops the header parser / reader 
+        #from getting caught on the <arcmetadata> XML lump
+        #that can appear in ARC files
+        
+        if header.startswith("<arcmetadata"):
+            while not header.endswith("</arcmetadata>\n"):
+                header = self.fileobj.readline()
+            header = self.fileobj.readline()
+            header = self.fileobj.readline()
+            
 
         if header == "":
             return None