Path: utzoo!mnetor!tmsoft!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!uunet!mcsun!ukc!strath-cs!stl!robobar!ronald From: ronald@robobar.co.uk (Ronald S H Khoo) Newsgroups: comp.lang.perl Subject: Re: lockf mods Message-ID: <1991Feb4.214044.2422@robobar.co.uk> Date: 4 Feb 91 21:40:44 GMT References: <1991Jan31.203029.12599@sdd.hp.com> <1991Feb1.074602.25583@robobar.co.uk> <1991Feb02.132117.26077@convex.com> Organization: Robobar Ltd., Perivale, Middx., ENGLAND. Lines: 14 tchrist@convex.COM (Tom Christiansen) writes: > I'm a bit nervous about using lockf() in all possible applicaltions. It > requires (on Suns and Convexen at least) rpc.lockd to be running. flock() > on the other hand is quick and can be run single user. Hmm.. Did I say use lockf() in all applications ? I musta been asleep :-) I probably meant to say "possibly include a flock() emulator written in terms of lockf() for sites without flock()" so that perl programs written with the *perl* flock() function have some chance of working on stock System V machines ? Is this reasonable? Does flock() map onto lockf() well enough for this approach to work ? -- Ronald Khoo +44 81 991 1142 (O) +44 71 229 7741 (H)