Can anybody give me a good reason why on earth we would to keep the
vgrindefs.db Berkeley DB file in /usr/share/misc?  The source file is
less than one 8kB disk block long (of which over a quarter is copyright
notice), and we get a nice 520kB DB file as a benefit(?!), which doesn't
include the copyright notice.  I'm sure the time for getcap(3) to wade
through 12 text entries in less than an 8kB block would be quicker than
the time for the extra read or two that DB would need to locate the data
right across the range of machines we support.

It seems like there was an effort to Berkeley DB any file that was
accessed with getcap(3) simply for the hell of it without putting any
thought into whether or not it was actually a good idea.

Yours in disbelief (unless proven otherwise :-),
Simon.