RE: Re: Full screen application Specs

Dave
 
Back in the days of VM/SP I had a lot of problems with hung users an I made
some changes to CP - 
Only in one or two places and as I remeber when I did that I no longer had
Hung users or Hung dialed terminals - But that was way before TCPIP
I can still find the fixes - but not sure that they would be relevant to
VM/6.
 
Any thoughts about looping thru the grafs instead of assigning the first one
that Hecules thinks is available.
 
I just tried that  on a Z/OS system - TN3270E where I can see the terminal
in the VTAM screen and I always get the next one in the pool.
even if I finsih correctly? 
 
I guess that logic prevents people from getting stuck on a stuck terminal
:-)
 
Roc


  _____  

From: [email protected] [mailto:[email protected]] On Behalf Of
Dave Wade
Sent: Sunday, August 01, 2010 20:11 
To: [email protected]
Subject: RE: [H390-VM] Re: Full screen application Specs


  

I think this needs Ivan or PaulG. I know there was an issue with
VM370/Hercules when we first started with VM/370 and screens were not
dropped when the IP connection was dropped and some devices were changes to
cope with this. 

I also did some wick checks and CPWATCH seems to handle this OK so it may be
worth look at that, but I could see where it did it at a brief glance...

Dave Wade G4UGM
Illegitimi Non Carborundum

> -----Original Message-----
> From: [email protected] <mailto:H390-VM%40yahoogroups.com>  
> [mailto:[email protected] <mailto:H390-VM%40yahoogroups.com> ] On
Behalf Of Rocky
> Sent: 01 August 2010 17:39
> To: [email protected] <mailto:H390-VM%40yahoogroups.com> 
> Subject: [H390-VM] Re: Full screen application Specs
> 
> 
> Ivan - Dave - if you saee this - pls give an opinion - 
> Problem down below
> 
> Paul in reply to your Email- NO REST FOR THE WEARY 
> 
> Today is Sunday and I am adopting your day of Rest - I need 
> it:-) Used my day of rest up yesterday. :-)
> 
> Here is an update ------ 
> 
> ALL
> 
> I made the fixes you guys pointed out and the site is now up 
> with the newer version.
> 
> That was one big BUG - The line 18 - also fixed the 5th 
> position in the action field
> 
> More important I am trying to figure out why the treminals 
> are getting stuck.
> 
> I have a theory and I am waiting for IVAN or DAVE to tell me 
> why it's wrong.Or maybe I got it right this time.
> 
> If a user gets discoonncted for some reason - unrelated to 
> the application - or if he just leaves the application in the 
> middle by closing his emulator - What I expect to happen, is 
> that Hercules will somehow notify CP and CP will reset the 
> graf and the graf at least will get the VM LOGON - 
> 
> The appication should also get an interrupt of some kind - 
> Not exactly sure - IVAN ???? any ideas what I should expect 
> to see in the CSW
> 
> Depending on where we are in the application - Right now If I 
> am in a wait state on the device, and I get an interrupt - in 
> most cases, I probably treat it as an attention interrupt. 
> 
> So I go do a read which will eventually fail and then either 
> I am probably going back to wait or send a message that an 
> invalid key was pressed.
> 
> Need a better definition of how to handle and identify this 
> interrupt if I am getting it. 
> 
> But from what we see is that the USERS are getting to 
> hercules and no further- 
> 
> They get the hercules SCREEN - which means that Hercules 
> accepted them, but they can't get to VM because VM still 
> thinks the graf that hercules is choosing is DIALED to the 
> application.
> 
> So three questions here.
> 
> 1) why is VM not resetting the graph?
> 2) what is the exact flavor of the CSW when this interrupt occurrs.
> 3) Should Hercules move on to the next device in the pool if 
> no response from this one.
> 
> I am pretty sure that modern ZOS systems loop thru the 
> terminals in the TCPIP pool, even if a session was completed 
> successfully - 
> 
> You dont keep getting the same terminal but you usually get 
> the next one in the pool.
> 
> Meanwhile I added an inactivity timer to the program - Made 
> it short 15 minutes on the EDIT screens and 20 on the BROWSE 
> (If you don't press something it will throw you out - No warning)
> 
> On the FLIST 10 minutes and the other small screens at the 
> beginning I think only 5,
> 
> So hopefully - if it works = It will do the reset to the graf 
> from the application. and after that Hercules should be able 
> to find VM again - That's ok for now but we need a another 
> solution that is not a work-around but that analyzes the CSW 
> if thats possible.
> 
> All of the above is theory - Not prooven fact yet.
> 
> I am pretty sure the application is running- because I 
> configured my emulator to use the LUNAME option and went to 
> a diffeernt graf - not the first one but a specific one at 
> the end of list and all worked while the first prt was stuck 
> on the Hercules screen
> 
> Grafs 0C0-0CF are defined - so if anybody gets stuck on the 
> Hercules screen I'd appreciate if you could try one the 
> higher numbers and see if that works, and of course let us 
> know what happens.
> 
> Probably need some INPUT from IVAN here.
> 
> As far as the application is concerned I have decided that if 
> there are no drastic changes or bugs I am ready to go to the 
> BETA stage - 
> 
> For me that means writing output to the disk.
> 
> Please - everybody try playing around with it - and BUGS or 
> ERRORS are appreiciated - Other suggestions also.
> 
> Significant things that are missing that I can think of are:
> 
> CHANGE COMMAND
> FIND as opposed to SEARCH which we have.
> HEX 
> 
> (And some highlighting and cursor field settings for certain 
> screens) - all comments are welcome.
> 
> I'll wait a few days and if nothing big comes up I'll add the 
> write. That will be the BETA
> 
> I quess we can accept some LIMITATIONS in the functionality 
> that we can fix in the future as long as it does not crash 
> and destroy anything.
> 
> We can live with LIMITATIONS, but not ERRORS
> 
> As usual - advice and suggestions are appreciated - If you 
> don't know how to get in just ask
> 
> If anybody wants the current AWS tape with the applicatin I 
> can send it - Rather not send sources yet unless you want 
> them for developing something else or just to look - 
> 
> They are also on the server as well and you can look at them 
> there with the full screen browse or editior - 
> 
> Suggest that for deep understanding that you look at the 
> listings because most of the good stuff is in MACROS which 
> don't show up in the code but are genned in the listing.
> 
> ( Logic of the app is being changed dynamically and I don't 
> want to answer questions about yesterdays application LOGIC).
> 
> But the IO and tasking Routines are pretty stable 
> 
> If you want the tape you can install it locally and have fun 
> or try looking at the listings
> 
> Source will be out on the publc site in the future -
> 
> 
> Roc
> 
> 
> 
> 
> 
> --- In [email protected] <mailto:H390-VM%40yahoogroups.com> ,
"paoloG" <crc@...> wrote:
> >
> > 
> > 
> > 
> > 
> > --- In [email protected] <mailto:H390-VM%40yahoogroups.com> ,
"Rocky" <rocsystems@> wrote:
> > >
> > > thanks for the kudos
> > > 
> > > Well the sabbath is a day of rest so I only worked 3 hiours today 
> > > ---
> > 
> > Rocky,
> > 
> > Sabbath should be nearly finished...If you can see at least three 
> > stars in the sky you can go back to your programming job :-)
> > 
> > On the contrary soon I'll start our rest day (in Italian 'Domenica' 
> > that means more or less : 'Our Lord's Day'..;-)
> > 
> > Regards.
> > 
> > Paul
> >
> 
> 
> 
> 
> ------------------------------------
> 
> Yahoo! Groups Links
> 
> 
>