Upon loading 994.5 my main menu no longer returned the proper exitcodes when clicking on the icon bar. B4 digging thru my code, i went back to 994.4 and it was fine. So there must be something in an xtree mod that created that anamoly.

I tested the vertical srcoll bar in an xtree combo box and it worked... thanx.

Last by not least... im still having sporadic problems having schpat chain to itself... any more ideas on this?

Thanx.

PS: My install tomorrow is going out with 994.2.

(and we will probably stay with that until these minor issues are resolved)

Jack.. any luck reproducing these 2 items?

Thanx.

I just installed this flavor on my laptop and still having the icon / exitcodes issue... so not related to linux/ATE.. (if that helps)

Sorry for the delay on this one - it slipped out of the pile.

There was a subtle change (i.e. fix) to the way XTREE decides which exitcode to return when there are multiple possibilities, like when you click on a button that causes both a validation exitcode (-48) and the exitcode for the button you clicked. Previously, it was returning -48 in that case, but it was decided that it was more important to return the exitcode of the button you clicked (since you can deduce the need to validate from the fact that you are leaving the tree, but you can't deduce which button was clicked from the fact that there is a validation exit).

In order to make that work, XTREE now checks to see if there are any more control codes queued up in the kbd channel before it exits, and if so, it processes them and then exits (which allows the subsequent exitcode to override the previous one).

This change seems to be responsible for the change you are seeing in your main menu, but it seems that it is mainly due to an unnecessary quirk in the way you coded your icon buttons to each send a VK_ sequence followed by a TAB (control+I), e.g.:

"%VK_F1%^i"

What is happening is that the VK_F1 is being processed, but then it is getting overridden by the Control+I, which is generating an Exitcode -9. (You can see this in the XTREE trace.)

My recommendation is to just remove all those ^i suffixes from the buttons on your main menu. I can't see what benefit they provide in any case, and they are clearly out-of-the-mainstream and confusing things. It does not appear that you are using that technique in any of your other icon bars either.

Let me know if that is an acceptable solution.

Admittedly this program predates almost every other program in are arsenal and uses pre-dated techniques. (I even see a JDM hanging around here and there.. )

I have no problems reworking the logic here, but I just dont see where the ^I is getting into the icon command... its being loaded from data statements into the array name ICON'CMD()...

Can u see where the ^I might be getting introduced here? (menu.bas)

***HOLD ON***

I was ready to click on [ADD REPLY] and i had a hunch... the TAB was embedded in the data statement and was getting picked up the the read statement!

The only reason im sending this along is so that others may learn from my mistakes! (as many as there are...!) (no comment, Herman)

Today is one of my nice days, as few as there are, so no nasty comments from me.

Lucky me...

(Actually im impressed that you read this far down the thread... :p )

I read them all. I just hope to understand them one day!