C - Menus and multiple selections...

[desiv]

I am confused (not a surprise) on multiple selections...
Long story short, I found some code I wrote 32 years ago and have been playing with it.  It's 1.3 code, using DICE on my Amiga.
Working great, but I remembered a bug/programming issue I had seen back then and never addressed.

Basically, I have an "Options" menu with four items that are CHECKIT|CHECKED|MENUTOGGLE entries.

My program uses IntuitMessage GetMsg to read the Class and Code.

I then do switches on the Class and then on the MENUPICK (Code) and have case entries for each.
For troubleshooting, I added a "puts ("Sound");" in one and a "puts("Extra");" in the other in my case statements so I could see on the shell when each registers.
And as long as I do a normal select (RMB, to the Options menu, hover over one of them and let go) it works fine.  Checkmark goes away, my variables get set, I see the puts proper result in my shell.
I can do the same thing when I RMB to select the item while using the RMB to get to the menu option.

What doesn't work is if I am holding the RMB and then use the LMB to select one AND THEN another one (while still holding the RMB).
When I let go, only the first item has my variables get set and shows results of the puts command.
BUT, both items lose their checkmarks (or gain them...)
i.e. the checkmark shows up or goes away with each LMB click, but only the first one registers any activity in the case statement.

I know I'm doing something wrong, but not sure what...

Pointers on places I can look for help?  I'm not yet finding what I think I need in the books/PDFs/text files (Amiga C Manual) I have been looking at...

Thanx,

[desiv]

Thanx!
Still poking and cleaning things up (and trying to fully understand it), but it's working.
DICE is giving me an "int-ptr conversion" warning on this line:
struct MenuItem* item = ItemAddress(&my_menu, code);
(maybe I should have split them up, but I was being lazy)
But I think that makes sense, and ItemAddress is an int...

Gotta get back into C...
It's been quite a while... Lots to refresh and learn...

[desiv]

Probably because it is. You should #include <proto/intuition.h> to let dcc know about the return type of the ItemAddress function.

Thanx, I haven't looked into prototyping yet...  It's on my list...

Also put -proto on the dcc command line to get an error for each undeclared function.

Oh, that just sounds like something asking for trouble and more work...  ;-)

[desiv]

And no, it does not ask for trouble, it helps you to avoid trouble.

You are making it tough for me to justify being lazy. ;-)
And you are right. 
This question came up as I found a printout (just paper) from a solitaire game I wrote using DICE in 92 or so.  OCRd it and went about fixing the code, to get it to compile.
Finally got there, but one of the OCR errors that took me a bit to find, I would have found much quicker if I had prototyped that function, if I understand it correctly.

OCR just dropped an entry it couldn't read, so there was one fewer item than it really needed...

I'll make sure to add prototyping as a move forward... 
Thanx!

[desiv]

I knew it was going to be trouble.. ;-)

I added:
#include <proto/intuition.h>

And now I get an error on compile, on my OpenScreen command?
DC1: "klondike1.11.c" L:233 Fatal:104 Pragma argument count conflict with prototype: OpenScreen

And that line just looks like this:
my_screen= (struct Screen *) OpenScreen( &my_new_screen );

As far as I can tell, there should only be one argument (not sure yet what a "pragma argument" is) for OpenScreen and it should be a pointer to a screen structure., I think.

Maybe the 1.3 includes don't like the proto/intuition.h that came with DICE (I looked at that .h file and it was pretty empty mostly pointing to other files, and as I don't fully understand prototyping and pragmas... ;-)   (obviously)
The DICE setup info had me copy over the NDK amiga13 includes, but that was into the amiga13 directory, and that didn't have any proto directory, so it's using the one that came with DICE.  Not sure if that matters.

But that's OK.  Just means I have more C to learn about... ;-)

[desiv]

I know, talking to myself... ;-)
I just figured that I should update this as I forgot to post on my status...
I got the multiple menu selections fully working, but I basically re-wrote the way it tried to do it...
I was doing switches on which menu tree to be in and then checked within each menu match as to which menu item was selected.  (Or something like that...  Don't have all the code in front of me...)
That wasn't working for me if you selected (LMB) multiple menu items during the same RMB action...

I took out the switches and changed it to IF statements that check for MENUNUM && MENUITEM and that fixed it.
I don't have that many menus and items, so the lack of efficiency shouldn't be a problem.  ;-)

If you want to see the (not great, but working) code, I put it (source and program) on Aminet, called Klondike92...  (92 being the year I stopped working on it)

Also, if you want to read a bit about my thought process while I was working on it (I don't recommend it; I'm not a blogger), I did keep a mini journal thing...
https://retrofiend.blogspot.com/2023/04/rescuing-my-old-commodore-amiga.html