Ticket #40 ( Closed )

Short Description invalid screen buffer overwrite
Entered By: PeteL When: 1998-05-29 09:48:28 Build: 1.03.12b
Categories Type: Problem   Department: Product   Category: Not Categorized
I think I've determined what my new timeout problem is - i think that the form fields (at least some of them) are being mapped onto the buffer, but the buffer no longer contains the same screen, due to the timeout. I have a template trace running now, but I don't think it's going to show much, if this is the problem. I don't know why the clocks aren't mismatched. suggestion for better trace?
Append By: PeteL  When: 1998-05-29 09:55:18  New Status: Pending IE
Comment well, the trace showed more than I thought it would, and nicely pinpointed the problem:

Session 1 wrong screen, navigation to flexx.rrn: Successful

there should have been no attempt made to navigate - there are absolutely no PATHTO statements in my environment. In so doing, it apparently thought it found the correct screen, and mapped data over the incorrect screen, into places that weren't input fields, etc. - in short, it got ugly, then drove global.invalid.

Append By: WindSurfer  When: 1998-05-29 10:10:45  New Status: Pending Customer
Comment This is happening, we think, because the code that does the navigation first checks if the desired screen is currently displayed.

This implies that perhaps a <TEENTER> with SKIPSCAN has moved the screen OFF of flexx.rrn, but Screensurfer still thinks that flexx.rrn is displayed.

We have added a recognition engine review of the current screen when the timestamp conflict occurs, rather than use the internal "last recognized". This change is incorporated in build 1.3.13, which will be added to available releases at about 4pm EST today.

A test version of 1.3.13 is now available at ftp://ftp.ieinc2.com/pub/surfer/t1_3_13.zip.

Append By: PeteL  When: 1998-05-29 10:24:24  New Status: Pending IE
Comment no it isn't .... at least not as of 10:21 :)
Append By: WindSurfer  When: 1998-05-29 10:38:03  New Status: Pending Customer
Comment Sorry, changed the naming standards a bit-- if you look closely, you will see that it is now T1_3_13.zip, NOT SS1_3_13.zip.

Versions for testing are now preceeded with T (for test) until they are "versioned", added to the Release Information on the SupportCenter, and posted as SS builds.

A file posted with an SS prefix will have a section in the readme with all changes documented, as well as backing source in our version control system.

This enables us to apply emergency fixes at any build level for live customers without that customer worrying about regressions from accumulated fixes/enhancements.

Append By: PeteL  When: 1998-05-29 11:47:36  New Status: Closed
Comment works fine, thanks