[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Adding MD5 to Image...



	Mmm.. That's closer to what I need, but due to client restraints, I can't have it IN the file (like I had pointed out earlier), since the 'file' is the ISO, and it'll be going straight to disk after it's made. (Or, written STRAIGHT to disk if MKISOFS can print out the MD5SUM itself..

Thanks for the offer tho, looking like a good and useful little utility!
Me.

On vinyl, scdbackup@gmx.net was heard scratching:

Mike 'Fox' Morrey >>

to have mkisofs report the MD5 sum for that image ? ISO Chksum: a69def10...190F5

Lourens >

How about using tee(1) ?

Ashish Rangole >

Having said that it, may be feasible to add it to cdrecord etc.


This is an interesting coincidence. I'm currently pondering
about enhancing this feature in my own software.


Mike, i can offer you a filter which passes the data stream
and appends a MD5 based checksum record to a file on disk.
Like :
  xyz A40813.220939 655192064 47e079ca2d97bef5c782db31e4ce1cec
(name, timestamp, size, MD5)

Download http://scdbackup.webframe.org/scdbackup-0.8.tar.gz
unpack and compile just
  cc -g -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE \
     -o cd_backup_planer cd_backup_planer.c
(or, of course, you may install the package as described in README)

The filter is to be run like this :
  cd_backup_planer -filter_md5 $recordlistfile $recordname

It will append the checksum record to $recordlistfile using
the $recordname which you choose.
See cd_backup_planer -help | less search for "^ -filter_crc32" which explains details.

Example :
  $ echo hello | cd_backup_planer -filter_md5 /tmp/md5test test
  hello
  $ tail -1 /tmp/md5test
  test A40813.223359 6 aee97cb3ad288ef0add6c6b5b5fae48a

It is quite well tested since nearly two years.


Today i had the idea to use this MD5 record for giving sense
to the large padding which i meanwhile append to any write.
(One never knows with future drivers)
A few hundred kB of repeated checksum records at the end of the data
should be recognizable and exploitable for checking purposes.
At least with an error probability that can match the one of MD5.
Well, still some work to do with that.


Have a nice day :)

Thomas



--
___________________________________________________________________________
Mike 'Fox' Morrey - Noble Systems
Director of Mad Science Programming
Head of the Spontaneous Data Generation Department
-> Despite the cost of living, have you noticed how it remains so popular? <-



Reply to: