x11-calc Updated

[Mike T.]

(05-03-2024 06:38 PM)Thomas Klemm Wrote:  Meanwhile I've built the binaries with debug information: make DEBUG=-g
I used lldb and set a breakpoint at x11-calc.c:781:

fprintf(stdout, "ROM Size: %4u words \n", ROM_SIZE);

And then just single stepped through the code until the exit of the program.
The output can be found in the attachment debug-x11-calc-35.zip.

I hope this is useful.

Thank you that is very informative. Somewhere after line 818 (in the while loop) b_abort is being set.

I think that this should only happen in one of two places.

• When the code checks to see if a Ctrl-Z was pressed.
  
  
  if (h_keyboard->key == (XK_Z & 0x1f)) /* Ctrl-Z to exit */
  b_abort = True;
  
  
  In the past I have had issues with random keystrokes being received by an the application on startup (which is why there is a 200ms delay to 'debounce' the keyboard at about line 660), but that only matters if the program checks for keystrokes, which I don't think is the case here as I don't think keyboard shortcuts are enabled.
  .
• Or if the window manager sent an event that tells the program to close the window:
  
  
  case ClientMessage : /* Message from window manager */
  if (x_event.xclient.data.l[0] == wm_delete) b_abort = True;
  break;
  
  
  This would suggest that the window manager is asking the application to close almost as soon as it is opened, possibly because it is off screen? However, if x11-calc-35 exits normally even when you have commented this out, something else is setting b_abort which is a bit odd.
  

There have been a lot of changes to the makefile incorporated in the testing branch to address some issues in the build process (that I've not seen myself). It is just possible that you are seeing a related problem and switching to the testing branch might help.