Subj : Re: Crc utility
To   : mark lewis
From : Tony Langdon
Date : Sat Oct 01 2016 08:13 am

-=> mark lewis wrote to Tony Langdon <=-

 ml> there shouldn't really be any TIC software that /requires/ the CRC...
 ml> it's only
 ml>  real use is to ensure that the file is not changed in transit...

Actually, it is a required field in the TIC specification, and the way one
hatched files in Mystic, it is effectively "in transit" from a dummy node.

 ml> depending on your OS, you may have the necessary tools already at
 ml> hand...

 ml>   cksum

That one didn't work for me.

 ml>   crc32

Don't have that one and can't find it for Raspian (though other distros seem to
have it available).

 ml> cksum or crc32 is the one you want but you should check with a known
 ml> file that has a known CRC value and see which program outputs the same
 ml> match... i say this because at least crc32 hashes can have pre- and
 ml> post- conditioning applied
 ml>  to them...

 ml>  i had to initialize the crc variable to $ffffffff (as a signed
 ml> longint) and then feed each character individually to the crc

That sounds similar to the TIC specs.  

 ml> routine... i could have fed the entire blob of data and let the routine
 ml> walk through it but that gave a different and wrong result... at the
 ml> end, we did not XOR with $ffffffff but we do have to invert the result
 ml> so the bytes are reversed since to match what's in
 ml>  the JAM data files... this byte reversal is because the legacy method
 ml> used is for traditional analogue modems and the UART wants everything
 ml> reversed for transmission... granted a UART is not used in this case
 ml> but it is the way the sofware is written and there's a lot of software
 ml> that uses this same formulation method...

Yes, again that rings a bell with the specs, which I did look up.  There is
some code there showing how the CRC32 is generated, just prefer not to have to
re-learn file handling in Pascal just yet. :)

 ml> [time passes]

 ml> i just checked what i was checking for JAM with the cksum and crc32
 ml> tools but neither returned the desired results... i'll have to go play,
 ml> now, with checking a file and its TIC to see if i can determine which
 ml> is correct if either one is... pre- and post- conditioning and whether
 ml> to flip the bytes or not make a huge difference in the results...

I had no joy with cksum, and I don't have crc32 on my system.


.... I hit the CTRL key but I'm still not in control!
--- MultiMail/Win32 v0.49
 * Origin: Freeway BBS - freeway.apana.org.au (3:633/410)

.