Re^3: flock and read-ahead buffering on input

Perl won't do any read-ahead buffering of the file until you perform the first read action. If the kernel performs read-ahead, it has a global picture of the file and and all processes, and will be capable of invalidating any buffered data if necessary.

So that code is perfectly safe for your needs.

Dave.

Comment on Re^3: flock and read-ahead buffering on input

Replies are listed 'Best First'.
Re^4: flock and read-ahead buffering on input by dbooth (Novice) on Apr 18, 2014 at 17:46 UTC
Thanks, that's the key piece of information that I needed. Unfortunately it is not documented in the sysopen, flock or read documentation. Can you point me to where that is documented?	[reply]
Re^5: flock and read-ahead buffering on input by dave_the_m (Monsignor) on Apr 18, 2014 at 22:47 UTC
Thanks, that's the key piece of information that I needed. Unfortunately it is not documented in the sysopen, flock or read documentation. Can you point me to where that is documented? Which are you referring to? That perl won't read-ahead, or that the kernel knows what it's doing? For the former, the whole purpose of sysopen/sysread etc is to avoid buffer layers; the man page for sysread specifically states this. As for the kernel, this is basic; if the kernel didn't do that it would be broken beyond belief. Dave.	[reply]
Re^6: flock and read-ahead buffering on input by dbooth (Novice) on Apr 19, 2014 at 05:03 UTC
I was looking for evidence that perl won't read-ahead when sysopen is called. The perldoc for sysopen says nothing about buffering one way or another, nor does the perldoc for read. I see now that the perdoc for sysread says that it "bypassess buffered IO", though I hadn't looked there before, since I'm using read -- not sysread. Thanks very much for your help!	[reply]