[this is a complete aside] On Wed, 9 Jun 2010, Aggelos Economopoulos wrote:
> I fully agree that tokens are a more convenient API than, say, lockmgr locks, For what it's worth, let me second this: providing you understand the approach, they're astonishingly convenient. I've got a largish deployment and configuration-management system that's largely built out of straightforward unix scripts here at work; I occasionally need some mutual exclusion in it (I _have_ actually been bitten by race conditions when multiple sets of scripts were executed at the same time). As luck would have it, Matt gave the original description of his token idea at about the time I was casting around for a cheap solution that would let my users write their own locked scripts using something that wasn't overly difficult to use. I knocked up an implementation and we've been happily using it ever since :-) -- jan grant, ISYS, University of Bristol. http://www.bris.ac.uk/ Tel +44 (0)117 3317661 http://ioctl.org/jan/ I'm the dandy information superhighwayman.
