Re: RCS and checksum

[Guido Sette]

On 2010, Sep 14, at 19:53 , Paul Eggert wrote:

> On 09/14/10 10:45, Guido Sette wrote:
>> Any advice on how to improve that approximative code?
>> 
>> Now, if only I could find a way to automatically verify the checksum at 
>> checkout...
> 
> The code looks all right, for starters.  You can verify the checksum
> by using "rlog" to find out what the checksum should be, and using
> "md5sum" to verify the checksum.
> 
> By the way, md5sum has been cracked to some extent; if you want
> a checksum that is cryptographically secure, I suggest SHA-2,
> e.g., sha224sum or better.


I am sincerely very grateful for your time and pointers.

For the reasons I wrote in my initial post (being a "one-member team" writer in 
humanities), I am not concerned too much with security, but of course it 
bothers me to use an insecure tool nevertheless, so I will use sha224sum.

I think I will try to use awk to parse the checksum so that it can be verified.

That will probably allow me to go the whole kit and kaboodle (probably saying 
"the whole shebang" would be more appropriate in this context!) and also allow 
me to use other diff tools, more easily understood by me.

Right now I am finishing a long article, and I am afraid to weaken my 
concentration on that side if I throw myself in the awk depths, but I will 
submit my solution to this second aspect as soon as I can hatch it (probably 
embarrassingly naive).

I am so satisfied with RCS, now even more.

Thank you once again,

Guido Sette