>>Is this still true?
>>
>>oldvalue = ""
>>value = "anystring"
>>
>>oldvalue = value (returns .T.)
>>
>>I thought this problem was fixed long ago, but I just ran across a situation where it appears the problem was not fixed.
>>
>>I SET EXACT ON still the cure?
>>
>>regards,
>
>It's always considered as a feature (not bug) and SET EXACT (SET ANSI for Select-SQL) should still care about it.

Hi Edward,

There, now you see how far behind the times I am. Since I could never figure out any conceivable use for this "feature", I alway thought it was a "bug". Foolish me!

A couple of weeks ago I was reading in the MS KB about how to test if a combobox value had been keyed-in rather than selected, and the code specified comparing ".displayvalue != .value". I tried that and it does not work where either .displayvalue or .value is blank. Someone at MS needs to learn about this "feature".

So ("feature" # "bug") = .T. I get it now.

Glad to see the old tried and true workarounds still work.

regards,