Re: [hatari-devel] Disassembly errors in history listing

[ Thread Index | Date Index | More lists.tuxfamily.org/hatari-devel Archives ]


Hi,

On 08/15/2016 12:22 AM, Roger Burrows wrote:
The disassembly in the listing produced by hi cpu is (occasionally) quite
wrong.  It fails to recognise an instruction and generates a DC.W and then a
number of following instructions can be wrong.
>
For example:
[...]
Some of this may actually be data (that's why I'm debugging), but if Hatari
executes it, the history listing should reflect the actual instructions
executed IMO.

History doesn't store executed instructions, just their addresses.
Disassembly output is just a convenience...

If the content of these addresses is changed, history command
disassembly output isn't correct.  You can check that by having
breakpoint on suspicious address(es):
	b <address> ! <address>

If that's not the case, or you otherwise think/suspect that given
opcode is real instruction, your Hatari is most likely using the CPU
core external disassembler ("ext".  It's the default because most
people like its output more and it supports several disassembly
output options.

Try CPU core internal disassembler ("uae") instead, it has better
instruction support.

See:
	hatari --disasm help


	- Eero





Mail converted by MHonArc 2.6.19+ http://listengine.tuxfamily.org/