Subject Re: Patch release IBO 4.2Ic
Author Marco Menardi
--- In, "Jason Wharton" <jwharton@i...> wrote:
> Marco,
> The fix to IB Combo Box has not been tested enough by me to clear it for
> release. It is still on my list of things to do. Others aren't
begging for
> it so I'm not looking at that as a critical thing to introduce. I'm also
> concerned it may actually impact existing applications and just may
hold on
> to your changes until a more significant (major) release cycle. You
have the
> fix for your application and I could be wrong but I don't see others
> noticing it. If I'm wrong I'd like to know about that from others.

Well, I'm very happy you are working on them :)
I'm not surprised if I'm the only one reporting those problems.
Usually few people report problems, and less of them can identify the
source of it, so maybe my problems can have other workarounds in the
program leaving untouched IBO, but I prefer having IBO work propery :)
there are 3 issues:
a) IB_ComboBox destroy problem under particular situations
b) IB_ComboBox and DropDown style new value cleared if the same value
is entered
c) IB_Text text position problem if borders are used

b) can be easely reproduced with the steps I gave you. I'm sure few
people noticed, because it's strange that you enter the same value
already in the IB_ComboBox, if you are a programmer. Users are not so
"cleaver" and I'm sure have reported this problem but, until hardly to
remember and reproduce, sure none of IBO programmers have figured it
out. Neverless, looking at the source with my hints you clearly see
the problem.

a) this is also a complex (uncommon) situation and with my patch, yes,
I'm not 100% sure it will not break anything, even if I'm using the
patch since july 2002. What I don't like in my patch is that
"DataSource := nil;" is moved from the original sequence to the new
position. I've not revised the code, but maybe we could better
intercept TIB_customComboBox.GetItemIndex and check here for
assigned(DataSource) and avoid TcustomCombo.GetItemIndex. Let me know.

c) In normal use (thin borders) , there is no problem (I've not
noticed problems when I developed the border addition). But the
problem exists, as I've discovered months later, and the fact that
right now there are no people incurring in it is not a good argument
not to fix it :) This fix, AFAIR, has no collateral effects, why not

Jason, I know your position is a difficoult one. It's not easy give
you a patch instead of just scream "bug!", but I do with pleasure as
much as I can. I understand that it's also much more easy fix a bug
that "works for me" than test it and be sure that will not break
something else.
Just hope you understand that heving no feedback on bugs/patches
reported is very demotivating.
Why not have a sort of "beta tester" IBO users to try to test each
other's fix? Enlarging the user base will more easely surface
collateral problems.
best regards
Marco Menardi