Re: Update A File After A Couple Other Chains

Then any number of things could be happening:

- You are not doing the chain
if moonBlue();
  chain FileInQuestion;
EndIf;  // but a much more complicated if structure that you get lost in 
the details

- The chain is not successful
chain (e) FileInQuestion;
// "assume" success.

- You are chaining with the no lock option;

- You have a hidden update that is lost in the complexity of the code.

- Some version of RCLRSC or some other such thing.

- You are using the old two file technique.  One file to read for filling 
subfiles, etc thus avoiding record locks and that rot.  Another file, 
overridden to the first, to actually update.  And you are updating the 
second when you've only read the first.

Here's a strange debugging technique.
Turn on journalling on the file for one iteration.  Include all entries - 
including open/close entries.  Look the at the journal entries created. 
But that might not show you anything.  For instance if you do have the 
repeating updates then it would show you the first and the second one that 
fails would not go to the journal.  But if you set a breakpoint just 
before the update in question, it might.

Rob Berendt