I made 15mins to go test this for my own peace of mind.
I tested on V4.6 and my green-screen test behaved as you described. Also, I noticed that the non-key-file-variable value was already 'lost' in the exit logic of the called LS (LS2) but that if I LSEXITed from the after-read-no-error logic of LS2 the variable was NOT lost. This suggests to me that the 'loss' of the value occurs somehow when ProIV handles "end-of-selected-records" in an LSUPDATE.
For me this is a clear-cut bug. But then I have no code that relies on it..
FWIW I tested an update function and there is NO loss of values when you use an LSCALL in a similar way to the screen example (ie. ProIV is not consistent about this).
LSUpdates : bug?
16 replies to this topic
Posted 02 December 2005 - 02:11 PM
It is for this reason that I always call true global updates in screens. That way I know what is being passed in and what is being passed out. Also, a global update allows for clearing of files, control breaks etc that a "screen update" does not. A screen update is emulating an update, but it still has all the behavior of a paging screen.
Reply to this topic
0 user(s) are reading this topic
0 members, 0 guests, 0 anonymous users