Page 34 of the QL Technical Guide; 6.3.3 Scheduler Loop Tasks states that: "Calls from the scheduler loop do not interrupt atomic tasks. This means that operations such as allocating or releasing memory can be performed safely."
Now admittedly it doesn't say " This means that operations such as allocating or releasing memory AND starting jobs " but, conversely, the definitive QDOS reference written by the author of QDOS himself doesn't say anywhere else that you can't. But anyway it's all a word game. We're probably doing things now with the QL "platform" that were never even imagined back when the machine and its original OS were first produced. In the driver I have decided to implement an "OS Capability" byte, which has bits set to indicate that the current OS is capable (or not capable) of doing things in a certain way. Bit 0 just got allocated to the function: "Can launch tasks from the scheduler loop" Maybe I'll be able to set it high for Minerva as well ;) In any event, I'm sure there will be a few more bits allocated before this driver makes it out of beta. I suspect that if I'd taken the path of simply not supporting QDOS (which, believe me, I would have loved to have done) the resulting product would be too restrictive to be useful. Anyway, I have no criticisms of SMSQ and the way it does things, nor of QPC2 - which has enabled me to develop this code so quickly. Adrian -----Original Message----- From: ql-users-boun...@lists.q-v-d.com [mailto:ql-users-boun...@lists.q-v-d.com] On Behalf Of Marcel Kilgus Sent: 23 February 2011 12:21 To: ql-us...@q-v-d.com Subject: Re: [Ql-Users] Ranting: A Story of two Operating Systems Where does it say it should work under QDOS? I guess this wasn't consciously "fixed" in SMSQ/E, it's probably more a side effect of the new implementation. And documenting all possible side effects borders on the impossible, unfortunately. _______________________________________________ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm
