Re: [PATCH v4 29/34] qapi: Implement deprecated-output=hide for QMP events

[Markus Armbruster]

Eric Blake <address@hidden> writes:

> On 3/17/20 6:54 AM, Markus Armbruster wrote:
>> This policy suppresses deprecated bits in output, and thus permits
>> "testing the future".  Implement it for QMP events: suppress
>> deprecated ones.
>>
>> No QMP event is deprecated right now.
>>
>> Signed-off-by: Markus Armbruster <address@hidden>
>> ---
>
>> @@ -140,6 +141,23 @@ static void test_event_d(TestEventData *data,
>>       qobject_unref(data->expect);
>>   }
>>   +static void test_event_deprecated(TestEventData *data, const void
>> *unused)
>> +{
>> +    data->expect = qdict_from_jsonf_nofail("{ 'event': 
>> 'TEST-EVENT-FEATURES1' }");
>> +
>> +    memset(&compat_policy, 0, sizeof(compat_policy));
>> +
>> +    qapi_event_send_test_event_features1();
>> +    g_assert(data->emitted);
>> +
>> +    compat_policy.deprecated_output = COMPAT_POLICY_OUTPUT_HIDE;
>
> Umm, did you forget to set compat_policy.has_deprecated_output = true?

I did.  Works anyway, because I don't bother checking it.  I'll clean it
up.

> (proof that we really want QAPI defaults to be working, to get rid of
> pointless has_* members...)
>
>> +    data->emitted = false;
>> +    qapi_event_send_test_event_features1();
>> +    g_assert(!data->emitted);
>> +
>> +    qobject_unref(data->expect);
>> +}
>> +
>
>> +++ b/scripts/qapi/events.py
>> @@ -61,7 +61,8 @@ def gen_param_var(typ):
>>       return ret
>>     -def gen_event_send(name, arg_type, boxed, event_enum_name,
>> event_emit):
>> +def gen_event_send(name, arg_type, features, boxed,
>> +                   event_enum_name, event_emit):
>>       # FIXME: Our declaration of local variables (and of 'errp' in the
>>       # parameter list) can collide with exploded members of the event's
>>       # data type passed in as parameters.  If this collision ever hits in
>> @@ -86,6 +87,14 @@ def gen_event_send(name, arg_type, boxed, 
>> event_enum_name, event_emit):
>>           if not boxed:
>>               ret += gen_param_var(arg_type)
>>   +    if 'deprecated' in [f.name for f in features]:
>> +        ret += mcgen('''
>> +
>> +    if (compat_policy.deprecated_output == COMPAT_POLICY_OUTPUT_HIDE) {
>> +        return;
>
> Here, you took the shortcut that if we always 0-initialize,
> deprecated_output will never be true except when has_deprecated_output
> is also true.  But the test above violated that rule of thumb.

Generated code ensures FOO is zero when !has_FOO.  Manual code should do
the same.

The pedantically correct check

    has_FOO && FOO == some_non_zero_value

can then safely be abbreviated to just

    FOO == some_non_zero_value

which I find less turing and more more legible.

> Otherwise, this patch makes sense.

Thanks!