🚀 go-pugleaf

> I don't know if this is just a pipe dream, or if it can be done. I have a
> BCB5 application. I want to be able to debug it in the IDE, and when I stop
> at a breakpoint, start up some other process (or thread) that I can use to
> help debug the application. I've toyed with using the Open Tools API, and
> still might. I also have used shared memory routines, and although this
> works fine, it is limited (can't "shared" STL classes).


I think that, when your program stops at a breakpoint, all its threads 
are stopped, so it wouldn't be able to actively take part in any IPC 
anymore. But maybe passive IPC, like shared memory, might work (yet I 
doubt that).

On the other hand, you can, of course, create an application for 
maintainability: If you have a number of values that you want to 
observe, simply make sure that any change to those values will be 
noticed by the respective property holder (such as the component), and 
that this change is propagated to a storage block that saves all 
properties you want to monitor.

This is also a simple method to create fault-tolerant software (at a 
very low level; don't expect to much from such a simple effort): Create 
recovery blocks regularly, and save all critical values into it. In case 
of an error, restore the actual properties' values from this block.

Storage and recovery blocks, can, of course, be made available via 
shared memory, or sent via messages, to an observer program.

Regards,
Ron Sommer