Re: Could timestamps be replaced with MD5?

[Tobias Weingartner]

On Tuesday, April 18, Pavel Roskin wrote:
> > Timestamps should be checked first.  If they mis-match, then MD5.  I'd
> > say this should be enough of an "optimization" to keep *most* MD5
> > computations within check.  Oh, and if the timestamp is the same, but
> > the file has changed (not firing the MD5 check), a quick "touch" of the
> > file should fix all...
> 
> Make no mistake! Don't touch the file! Adjust the timestamp in
> CVS/Entries.
> 
> Touching the file may cause major problems, especially when it's touched
> "backwards".

How, pray tell, would that be any different than the 2 scenarios that I
wrote about in my previous e-mail.

1) timestamps differ (not less, not more, differ), MD5 sums are the same,
   files are not the same.
2) timestamps don't differ, MD5 sums are not the same.
   files are not the same.

Those are the "two" potential "problem" areas of using a "sum" to represent
the contents of a file.  For both problems, I gave viable solutions (based
on the fact that the probability of this happening were minute).

--Toby.