Ted, thanks. I have not yet tested this approach in production so I can’t comment on the performance. What I can say is that the approach seems to work, i.e. the functional expression defined library code is parsed only once at the first trigger execution and never after that. It is actually behaving so well that it is hard to redefine the library - I usually comment out the “if ( …” test and next trigger execution redefines the lib. Re-saving the trigger, disabling/enabling it does not do much, I think even deleting and re-defining the trigger did not work but I have not spent much time on this (the 30s timing may be the issue here, I usually do that faster ).
How is “cache” implemented? Where I could find more on the trigger mechanism architecture?
will try to put my object to cache, should not be too much problem.
Thanks for your time, appreciate your replies! A.