From f500bca0fe1556eb6a776837d64d18f9ae30a093 Mon Sep 17 00:00:00 2001
From: GRiker <griker@hotmail.com>
Date: Wed, 10 Mar 2010 12:06:59 -0700
Subject: [PATCH] GwR revise layout analysis

---
 src/calibre/ebooks/metadata/topaz.py | 24 ++++++++++++++----------
 1 file changed, 14 insertions(+), 10 deletions(-)

diff --git a/src/calibre/ebooks/metadata/topaz.py b/src/calibre/ebooks/metadata/topaz.py
index 4f442a844f..8423c52524 100644
--- a/src/calibre/ebooks/metadata/topaz.py
+++ b/src/calibre/ebooks/metadata/topaz.py
@@ -4,9 +4,10 @@ __copyright__ = '2010, Greg Riker <griker@hotmail.com>'
 __docformat__ = 'restructuredtext en'
 
 ''' Read/write metadata from Amazon's topaz format '''
-import copy, StringIO
+import copy, StringIO, sys
 from struct import pack, unpack
 
+from calibre import prints
 from calibre.ebooks.metadata import MetaInformation
 
 def read_record(raw, name):
@@ -253,7 +254,8 @@ class MetadataUpdater(object):
         offset = 5
         topaz_headers = {}
         dkey_offset = 0
-        highest_payload_offset = 0
+        lowest_payload_offset = sys.maxint
+        lowest_offset_err = None
         for x in range(self.header_records):
             marker = self.data[offset]
             offset += 1
@@ -268,10 +270,10 @@ class MetadataUpdater(object):
                 hdr_offset, consumed = self.decode_vwi(self.data[offset:offset+4])
                 if tag == 'dkey':
                     dkey_offset = hdr_offset
-                if tag != 'metadata':
-                    if hdr_offset > highest_payload_offset:
-                        highest_payload_offset = hdr_offset
-                        #print "adjusting highest_payload_offset to 0x%x (%s)" % (hdr_offset,tag)
+                if tag not in ['dkey','metadata']:
+                    if hdr_offset < lowest_payload_offset:
+                        lowest_payload_offset = hdr_offset
+                        lowest_offset_err = "lowest_payload_offset: 0x%x (%s)" % (hdr_offset,tag)
                 offset += consumed
                 len_uncomp, consumed = self.decode_vwi(self.data[offset:offset+4])
                 offset += consumed
@@ -280,10 +282,13 @@ class MetadataUpdater(object):
                 blocks[val] = dict(hdr_offset=hdr_offset,len_uncomp=len_uncomp,len_comp=len_comp)
             topaz_headers[x] = dict(tag=tag,blocks=blocks)
         self.topaz_headers = topaz_headers
-        self.highest_payload_offset = highest_payload_offset
         self.eod = self.data[offset]
         offset += 1
         self.base = offset
+        self.lowest_payload_offset = lowest_payload_offset + self.base
+        if self.lowest_payload_offset < self.md_header_offset:
+            prints("Unexpected TPZ file layout:\n %s\n       metadata_offset: 0x%x" % (lowest_offset_err, self.md_header_offset))
+            prints("metadata needs to be before payload")
         self.base_value = None
         if dkey_offset:
             self.base_value = self.data[offset:offset + dkey_offset]
@@ -352,8 +357,8 @@ class MetadataUpdater(object):
                     self.metadata[item]['metadata'] = value
                     return
 
-        if self.md_start > self.highest_payload_offset:
-            raise ValueError('Unable to update metadata')
+        if self.md_start > self.lowest_payload_offset:
+            raise ValueError('Unable to update metadata:')
 
         try:
              from calibre.ebooks.conversion.config import load_defaults
@@ -386,7 +391,6 @@ class MetadataUpdater(object):
         self.stream.write(updated_metadata)
         self.stream.write(tail)
 
-
 def set_metadata(stream, mi):
     mu = MetadataUpdater(stream)
     mu.update(mi)