| Newsgroups |
gmane.comp.emulators.hercules390.general |
| Message-ID |
<[email protected]> |
Joe,
I am with Paul. Whilst I don’t mis-treat MVS in the same way Paul does, there are times when things go pear shaped, and when I type “quit” on the Hercules command line I expect it to quit.
Sometimes, when running VM rather than MVS this doesn’t happen.
Paul,
For something like MVS “deterministic” is a pretty useless term. Because timings on any OS depend on what else is happening, I am pretty sure Hercules is, given any set of initial conditions, including the state of the other tasks deterministic.
However, things like processing broadcast packets from the network can change the subtle timings, and subtle timings are what is driving this.
Dave
From: [email protected] <[email protected]>
Sent: 23 June 2019 00:31
To: [email protected]
Subject: Re: [hercules-390] Re: quit files
"The flush is probably happening prematurely."
Dude. Come on. Its the next-to-last line before exit(0). There's no way it could be premature.
Joe
On Sat, Jun 22, 2019 at 7:27 PM kerravon86-/[email protected] <mailto:[email protected]> [hercules-390] <[email protected] <mailto:[email protected]> > wrote:
---In [email protected] <mailto:[email protected]> , <joemonk64@... <mailto:joemonk64@...> > wrote :
> The reason I dont believe you is that this is the end of the code:
> fflush( stdout ); exit(0);
> There is an explicit flush of stdout as you can see...
The flush is probably happening prematurely.
> this is why I am telling you its not a Hercules problem.
The incomplete and random log file cannot be
anything BUT a Hercules problem. It's not
because I typed quit too slowly or too fast.
quit is quit and Hercules should terminate
gracefully, writing out a full log file every
time. All applications should terminate
gracefully when asked to do so. ALL of
them. Even ones that start with the
letter "H".
BFN. Paul.