Bug in i386_process_record?

[Michael Snyder]

Hui Zhu wrote:
> On Fri, Aug 28, 2009 at 11:06, Michael Snyder<msnyder@vmware.com> wrote:
>> Hui Zhu wrote:
>>> On Fri, Aug 28, 2009 at 09:35, Michael Snyder<msnyder@vmware.com> wrote:
>>>> Hui Zhu wrote:
>>>>> On Thu, Aug 27, 2009 at 09:43, Hui Zhu<teawater@gmail.com> wrote:
>>>>>> On Thu, Aug 27, 2009 at 08:28, Michael Snyder<msnyder@vmware.com>
>>>>>> wrote:
>>>>>>> Do you think you could add some new tests to i386-reverse.exp,
>>>>>>> to verify the string instructions?
>>>>>>>
>>>>>>> Thanks,
>>>>>>> Michael
>>>>>>>
>>>>>> OK. I will do it.
>>>>>>
>>>>>> Thanks,
>>>>>> Hui
>>>>>>
>>>>> Hi Michael,
>>>>>
>>>>> I make a patch to add the test for string insn.
>>>>>
>>>>> Please help me review it.
>>>> Good start -- but you need to write some expect script to go with it!
>>>> ;-)
>>> Hi Michael,
>>>
>>> This patch can make inferior without string_insn_patch get fail in:
>>> gdb_test "continue" \
>>>    " end of main .*" \
>>>    "continue to end of main"
>>> Prec will get error in asm volatile("rep\n" line when continue.
>>>
>>> Do you think I need make string_insn test  divide with inc_test in
>>> expect script?
>> My intention when I wrote the i386-reverse test was
>> that it should be extended with more tests over time.
>>
>> In fact, I had this one in mind.   ;-)
>>
>> Don't worry about "without string_insn_patch", since you will
>> check it in tomorrow.  You will check in this test later than
>> that, so nobody will get this test unles they already have the
>> string_insn_patch.
> 
> OK.
> 
> I make a new version patch that divide the inc_test and
> string_insn_test.  Please help me review it.

So far, so good.   ;-)

Now, you need to actually test something.
Like, record the value of the memory before and after the instruction,
then reverse step the instruction, and test that the value of the
memory actually reverts to its earlier value.