Am I extracting binary JPEG data from this mysqldump correctly?

I have a very old .sql backup on the vbulletin site that I worked on 8 years ago. I am trying to see the file attachments that are stored in the DB. The script below extracts them all and validates as a hex dumped JPEG and validates the SOI (start of image) and EOI (end of image) bytes (FFD8 and FFD9 respectively) according to the JPEG wiki page .

But when I try to open them with evince I get this message "Error interpreting JPEG image file (JPEG data stream does not contain image)"

What can be done here?

Some background information:

• sqldump is about 8 years old.
• vbulletin 2.x is software that stored information
• Most likely php 4 was used
• most likely mysql 4.0, maybe even 3.x
• the data type of the column in which these attachments are stored in mediatext format

My Python 3.1 script:

#!/usr/bin/env python3.1

import re

trim_l = re.compile(b"""^INSERT INTO attachment VALUES\('\d+', '\d+', '\d+', '(.+)""")
trim_r = re.compile(b"""(.+)', '\d+', '\d+'\);$""")
extractor = re.compile(b"""^(.*(?:\.jpe?g|\.gif|\.bmp))', '(.+)$""")

with open('attachments.sql', 'rb') as fh:
    for line in fh:
        data = trim_l.findall(line)[0]
        data = trim_r.findall(data)[0]
        data = extractor.findall(data)
        if data:
            name, data = data[0]
            try:
                filename = 'files/%s' % str(name, 'UTF-8')
                ah = open(filename, 'wb')
                ah.write(data)
            except UnicodeDecodeError:
                continue
            finally:
                ah.close()

fh.close()

      

        
        
        
      

    

Update The JPEG wiki says that the FF bytes are section markers and the next byte indicates the section type. I see some that are not listed in the wiki page (in particular, I see a lot of 5C bytes, so FF5C). But the list has "common labels", so I'm trying to find a more complete list. Any guidance here would be appreciated as well.