As the crash logs should indicate I went thought a lot of crashes in a very short period this afternoon as I tried to help resolve a crashing issue for someone else.
The crash was occurring on these lines of code:
field BalDue
formulafill Fee-Paid
I recorded the same as I went through it manually and Panorama came up with:
field BalDue
formulafill Fee-Paid
runmodifyfill
After adding that to the procedure the crashes have ceased.
Nothing in the documentation indicates that runmodifyfill is required and .ModifyFill does not exist in any of these file. I’ve used FormulalFill all over the place without previously seeing issues. Why it (apparently) caused this crash is puzzling.
I started two threads on the assumption that I was dealing with two issues, but I’m suspecting that they’re related - possibly to the Panorama installation.
The database I’m troubleshooting has been running fine for some time on PanX. The user reported that the crashes had just started with b25. I now have it running fine on b25 on the computer on which I swapped versions back and forth. That’s the same machine that’s now running the zlog just fine.
On my other computer, the database continues to crash and the zlog fails. This is also the machine on which I reported the out of synch Datasheet and Text Editors a few days ago. Text Editor Display Issues
Panorama added the runmodifyfill because it is being defensive and it doesn’t know if you might add a .ModifyFill at sometime in the future. But this statement does not in any way change how the formulafill works, and cannot cause a crash in the formulafill statement.
I see five crashes involving the formulafill statement. However, the crash is not in processing the formulafill statement itself, but in updating the display of a Text Display object after the fill operation has completed. It appears that the formula for the Text Display object contains a pattern( function if that helps in any way. Here are the first 14 lines of the stack trace at the time of the crash (out of 61 lines). You also had several similar crashes when the maximum statement was being used (but in those cases the crash was also in the Text Display object).
A possible thing to try would be to use no-show and then not display intermediate changes until the end of the procedure. Let me know if that helps.
You might be seeing it work on the second computer because the form window that is causing problems isn’t open. (To be clear, obviously a form window should not cause a crash like this, but that’s where the bug is – not in the formulafill or maximum statements.)
First of all, I was mistaken that I had restored b25 on the one machine. It was still running b24. That’s where zlog was working and the problematic procedure was not crashing. I added the NoShow and offsetting EndNoShow and it ran much faster.
Running b24 on either machine, the procedure runs fine. Restoring b25 to both machines and it again crashes on both machines.
I’m going to suggest that the user of the database re-install b24 until it’s figured out.
BTW, the Install menu does not reliably update to show the selected version. It tends to say it’s ready to install b25 but it does install the version that had been selected, even though the Version menu has reverted to showing b25.
I had been previously unaware of this ability to choose versions. Thanks, it’s handy.
