I add here the response of the CW expert team regarding the Breakpoint IDs.
Breakpoint IDs should be treated as an opaque handle, you should not expect them to have a specific value.
There are multiple reasons:
- there would be id conflicts when having two simultaneous debug session. E.g. which project bp #1 would refer when debugging two projects in in parallel?
- Breakpoints set by user in source files before debug session (or left from previous debug session) precedes the ones that would be set by script. You would have random behavior depending on intentional/unintentional set of breakpoints before starting debug session.
Automation solution should rely on the fact that “bp” command will return bp id. You can assign the return value to a TCL variable and use it when referring to that variable.
E.g.
set bp_id [bp -auto main.c 21]
I have tried a similiar TCL statement but my syntax was apparently in error, so i missed that.
The automated script could be modified as follows (only related part is displayed, see previous posts):
bp all off
set bp_id [bp -auto SysTick.c 89]
bp #$bp_id ignore 3
This closes the automation part.
I will try to compile CppUTest for coldfire to close the process. I will post the results.
Thanks,
Ilias