David and John, I agree, I like IBM's implementation, what experience I have had so far with it. Being an ex-VMer, its comfortable and easy to use ..but not everyone has that experience, per se.
Scott On Wed, Mar 30, 2016 at 9:27 AM, David Crayford <dcrayf...@gmail.com> wrote: > On 30/03/2016 8:22 PM, John McKown wrote: > >> Simpler, easier to use API. No need to create a file system object and >>> >ftok a token, you can just use a namespace with mq_open() and features >>> that >>> >don't exist >>> >in System V such as mq_notify() spring to mind. In in nutshell it's a >>> >better design. Having said that the System V message queues are better >>> than >>> >nothing and you >>> >don't have to be authorized to use them which is goodness. >>> >> An. I see what you're getting at. It wasn't, as I had thought, that >> message queues are bad because they are SYSV, but that IBM should have >> implemented the POSIX version of message queues instead of the SYSV >> version. Too much gaming on my part lately. It rots the brain. >> > > In fact the z/OS implementation is decent. You can use select() and poll() > which you can't on some other Unixes. I suppose > POSIX message queues are quite new and IBM implemented the System V > version because they were part of the standard at that time. > > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
