5250 Bridge Timing

[adelgadodotorg]

I'm scripting 5250 sessions in JD Edwards ERP and I'm running into problems with its timing on redirections between different screen states and programs. So far I've been able to conditionally deal with any problems. e.g. If a blank screen shows up intermittently between two programs then I'll check for that possibility after enough failures are documented.

In the 5250 JDE interface some function keys generate a box over the rest of the screen similar to a dialog window. On one particular screen one of these optional windows is opened automatically when a user enters the primary screen that envelopes it. It appears that the user enters the program/screen (tightly coupled code) before the program loops back on itself to a state where the dialog window is displayed.

The 5250 Bridge sometimes catches a snapshot of the program at the point that the user has entered it but it has not yet looped back on itself. I can move the process along by "pressing" a key. However, I'm in a state that is only made unique by the fact that it may or may not happen -- something that by its nature I can't readily identify -- so I basically have to find a key that doesn't mean anything in any of the states involved or I may trigger yet another screen. This is something that I am still working through and that I'll likely encounter in other programs.

Has anyone else run into something similar? How have you dealt with it? i.e. Have you been able to effectively add a delay into your code to allow the 5250 screens to catch up? etc.

[rodflohr]

To add a delay to your script, try using usleep():

http://us2.php.net/usleep

This will let you set the sleep time in microseconds. You can also consider setting a loop to check for the correct screen to be present, and sleep some fraction of a second during each iteration of the loop. The loop should be finite, perhaps 5 or 10 iterations.

[shlomov]

You mean the windows repositioning is done once as default and than it acts on the DSPF parameters WINDOW(&STLIN &STPOS) ??

[adelgadodotorg]

You mean the windows repositioning is done once as default and than it acts on the DSPF parameters WINDOW(&STLIN &STPOS) ??

Schlomov, I'm not fluent in RPG III or DDS so I can't add much to the discussion on what's happening in JDE. I described what I'm experiencing while utilizing the 5250 Bridge to automate some tasks normally carried out by users.

Rod, any suggestions on how long to sleep for? Have you seen similar reports elsewhere? I slogged through the first iteration without adding any delays and only conditionally catching problem spots.

[shlomov]

The Bridge Demo was desisgned to capture the default i5 menues as presented with no tailored changes.

The 5250 Bridge API interacts with the display via the IBM Web Facing service. If the application captured presents an un counted for screen the 5250 Brige does not ignore it and presents it as a blank screen and goes into a halt where a user intervention is needed.

In this case, you will need to have your PHP script check for the uncounted customized vendor application display screen or blank screen, and wait for it to be replaced with a screen requiring input or automated your response with a function key.

[adelgadodotorg]

The 5250 Bridge API interacts with the display via the IBM Web Facing service. If the application captured presents an un counted for screen the 5250 Brige does not ignore it and presents it as a blank screen and goes into a halt where a user intervention is needed.

In this case, you will need to have your PHP script check for the uncounted customized vendor application display screen or blank screen, and wait for it to be replaced with a screen requiring input or automated your response with a function key.

Yes, this is as I've encountered and countered. Can you speak to the intermittent nature of those blank or sometimes partially built screens?

Alfred

[mooseisacan]

The 5250 Bridge API interacts with the display via the IBM Web Facing service. If the application captured presents an un counted for screen the 5250 Brige does not ignore it and presents it as a blank screen and goes into a halt where a user intervention is needed.

Is there any way to know that this happened? It would be nice to know when the output is clean.

[ridgo]

The 5250 Bridge API interacts with the display via the IBM Web Facing service. If the application captured presents an un counted for screen the 5250 Brige does not ignore it and presents it as a blank screen and goes into a halt where a user intervention is needed.

Is there any way to know that this happened? It would be nice to know when the output is clean.

i came across this issue discussed in here zend.com/en/company/news/event/webinar-bridge-and-platform