"Cor Ligthert [MVP]" <no************@planet.nl> wrote in message
news:uD**************@TK2MSFTNGP11.phx.gbl...
Greg,
I did not look to it before, however in my opinion can you say that the
1.1 way is a bug.
What happens in your routine is:
There is a sort order in the dv that should be dynamic.
In 1.1 it does not react on it if you change inside the for loop
In 2.0 it changes the order direct when there is a change in the
sortorder.
I agree with you that the last can give weird results.
Cor
I would agree also it was probably a bug. (Definately poorly written code on
my part.) But the lesson here is, if you recompile you need to do some
thorough testing.
In fact I forgot about this test I did and reported in another thread...
"I just ran my test app (the 1.1 version) from that thread on a machine with
nothing but 2.0 on it. I got the 2.0 output (not suprising)."
In other words, if 1.1 is not available to the program it will run using 2.0
framework. (kinda obvious when you think about it) And any hidden gotchas
will appear WITHOUT recompiling! Anybody else see a problem here? :^)
You would think .NET apps would REQUIRE the framework they were complied
against since they obvioulsy may no't work the same on newer builds. (I
know there is a way to enforce this in a *.config file, but maybe it should
be the default behavior, no?)
Greg