Maybe the docs you're looking at -- or at least the sections that you quoted -- for this particular software package (whatever it is) were originally written for people using solaris systems, and have not been updated for the readers who use FreeBSD/MacOSX.

Apparently this software program was able to modify perl 5.x to use flock "correctly" by "performing the final link of the perl executable using the BSD compatibility library by using /usr/ucb/cc as the linker" for building perl5 on Solaris 2.5 ... which they further asset that such linking can not be done unless one purchases Sun's "BSD compatability package."

Actually, I don't think that software package, in itself, affects a given user's installation of perl at all. The doc here is saying that if the user -- meaning, I think, the user on a solaris system -- happens to be using a perl interpreter whose installation did not meet this certain condition, this software might not work as intended, since the "correct" flock support might not have been properly compiled into the interpreter.

Of course, if you were on a solaris box, and had installed perl using gcc instead of /usr/ucb/cc, flock might still work quite well -- the quoted material is not specific about conditions that are known to do the wrong thing.

Since you are on a BSD system, you should have no need whatsoever for a "BSD compatability package" -- that sort of thing would only make sense on non-BSD flavors of unix (e.g. solaris). "Correct" file locking should be a native resource for you.

In any case, the proof is in the locking. Have you tested flock to see that it works? Have you tried installing this software and running its tests to see if they work?

In case it helps, I posted a simple "semaphore file" module here (Re: Replacing a string in a file), which I copied out of a Sean Burke article in The Perl Journal a few years back. I have used it to good effect, including as an easy way to test whether flock works as intended on various platforms.