LinuxSA Mailing list archives

  From: Alan Kennington <akenning@topology.org>
  To  : David Lloyd <lloy0076@rebel.net.au>
  Date: Tue, 22 May 2001 17:24:05 +0930

Re: GZIP Catastrophe

On Tue, May 22, 2001 at 05:11:14PM +0930, David Lloyd wrote:
> 
> I have a gzipp'ed file that apparently has a CRC error (1). Is there any
> way at all to recover it?
> 


David,

I don't really understand at which level the CRC error is happening.
I take it the gzip application is reporting that it's one and only
file-wide CRC is wrong.
If so, there's the obvious grunge method of recovery, on the assumption that
the error could be a single bit:

For all bits in the file:
	invert the bit
	if CRC is correct, this might be the right error correction
	end

For all candidates out of this for-loop, try it out with
gunzip and see if it looks sensible.

Alternatively, if that doesn't work, which means that more than
one bit is damaged, you could try the super-grunge method,
which is to try all alterations of all bytes, which would
take about 256/8 times as long, clearly.
This could produce more false candidates.

If you try 32-bit error corrections, you'll get so many
false candidates, it won't be worth the bother.

Unfortunately, with a compressed file, the information density
is so high that you can't use internal redundancy to solve the problem.
Maybe one day, gzip will include FEC fields.

Cheers,
Alan Kennington.

-- 
LinuxSA WWW: http://www.linuxsa.org.au/  IRC: #linuxsa on irc.linux.org.au
To unsubscribe from the LinuxSA list:
  mail linuxsa-request@linuxsa.org.au with "unsubscribe" as the subject