b.location is just the string used by the user to set the breakpoint not its address, so the current code relies on the user using an hexadecimal address and doesn't handle breakpoints set on a function.

Would it be worth it to parse info breakpoints to extract all the addresses instead of ignoring some breakpoints ?

I also considered parsing b.location and look for the address from the symbol name, but I still have troubles with that part of the API and we would have to handle ourselves the differences between declarations such as b <fun>, b *<fun> and b *'<fun>'+<offset>.

GDB 13.1 did introduce Breakpoint.locations that can return the exact address, but I don't think it's an option since it relies on a recent version of gdb.