Re: [Sdcc-user] simulating

[Raphael N. <rn...@we...>]

> Instructions counter: 94
> Clock cycles counter: 484
> Last instruction: call 0xAC

OK, I reproduced the error using oshonsoft's PIC simulator. I solved it 
by selecting the correct PIC (16f84) rather than using the simulator's 
default (16f877), which messes up the software (pseudo) stack: the 
software stack is placed by the linker into an address range that is 
plain memory on the 16f84, but occupied by SFRs on the 16f877. As a 
result, reading from code memory (during initialization) does not jump 
to the RETLW instructions, but always to address 0 (the fixed? contents 
of the SFRs), the RESET vector. This causes an infinite loop, occupying 
two HW-stack slots in each iteration (CALL __gptrget2 and a nested CALL 
__codeptrget1).

So, no bug here after all.

Regards,
Raphael