Quoting Rob Kendrick <[EMAIL PROTECTED]>: > On Wed, 05 Nov 2008 11:58:10 +0000 > dave higton <[EMAIL PROTECTED]> wrote: > > > Quoting "David J. Ruck" <[EMAIL PROTECTED]>: > > > > > Richard Porter <[EMAIL PROTECTED]> wrote: > > > > I can replicate this problem to order. I open a window on NetSurf > > > > which loads my local home page, http://localhost/ > > > > (internet.websites.index/html), from WebJames. > > > > > > You are attempting to use a cooperatively multitasked operating > > > system to both serve and display content, expect such problems even > > > if the browser is working acceptibly with external servers. > > > > You're being unnecessarily negative, Dave. Many of us see similar > > single- tasking behaviour on Windows and Linux from time to time. It > > isn't just a case of cooperative versus pre-emptive; it's how the app > > is written that counts. > > His point is that in a PMT system, such deadlocks cannot occur, but in > a CMT system, enormous care must be taken to stop them being > commonplace. Also, the RISC OS IP stack does not make this care easy > (especially given its lack of any meaningful select() call.).
My point is that deadlocks /do/ occur in Windows and Linux, two of the most widely known PMT systems. We've all experienced them. PMT, in and of itself, is not sufficient to prevent them. Well, I suppose now you'll tell me that you never have... Dave ------------------------------------------------- Visit Pipex Business: The homepage for UK Small Businesses Go to http://www.pipex.co.uk/business-services
