Обсуждение: Fix typo with logical connector (src/backend/commands/vacuumparallel.c)

On Fri, Aug 19, 2022 at 5:35 PM Ranier Vilela <ranier.vf@gmail.com> wrote:
>
> Hi,
>
> At function parallel_vacuum_process_all_indexes there is
> a typo with a logical connector.
>
> I think that correct is &&, because both of the operators are
> bool types [1].
>
> As a result, parallel vacuum workers can be incorrectly enabled.
>
> Attached a trivial fix.
>
> [1] https://wiki.sei.cmu.edu/confluence/display/c/EXP46-C.+Do+not+use+a+bitwise+operator+with+a+Boolean-like+operand

Good catch! Patch LGTM.

-- 
Bharath Rupireddy
RDS Open Source Databases: https://aws.amazon.com/rds/postgresql/

On Fri, Aug 19, 2022 at 5:40 PM Bharath Rupireddy
<bharath.rupireddyforpostgres@gmail.com> wrote:
>
> On Fri, Aug 19, 2022 at 5:35 PM Ranier Vilela <ranier.vf@gmail.com> wrote:
> >
> > Hi,
> >
> > At function parallel_vacuum_process_all_indexes there is
> > a typo with a logical connector.
> >
> > I think that correct is &&, because both of the operators are
> > bool types [1].
> >
> > As a result, parallel vacuum workers can be incorrectly enabled.
> >
> > Attached a trivial fix.
> >
> > [1]
https://wiki.sei.cmu.edu/confluence/display/c/EXP46-C.+Do+not+use+a+bitwise+operator+with+a+Boolean-like+operand
>
> Good catch! Patch LGTM.
>

+1. This looks fine to me as well. I'll take care of this early next
week unless someone thinks otherwise.

-- 
With Regards,
Amit Kapila.

Ranier Vilela <ranier.vf@gmail.com> writes:
> At function parallel_vacuum_process_all_indexes there is
> a typo with a logical connector.
> I think that correct is &&, because both of the operators are
> bool types [1].
> As a result, parallel vacuum workers can be incorrectly enabled.

Since they're bools, the C spec requires them to promote to integer
0 or 1, therefore the & operator will yield the desired result.
So there's not going to be any incorrect behavior.  Nonetheless,
I agree that && would be better, because it would short-circuit
the evaluation of parallel_vacuum_index_is_parallel_safe() when
there's no need.

            regards, tom lane

On Fri, Aug 19, 2022 at 7:45 PM Ranier Vilela <ranier.vf@gmail.com> wrote:
>
> Em sex., 19 de ago. de 2022 às 10:28, Tom Lane <tgl@sss.pgh.pa.us> escreveu:
>>
>> Ranier Vilela <ranier.vf@gmail.com> writes:
>> > At function parallel_vacuum_process_all_indexes there is
>> > a typo with a logical connector.
>> > I think that correct is &&, because both of the operators are
>> > bool types [1].
>> > As a result, parallel vacuum workers can be incorrectly enabled.
>>
>> Since they're bools, the C spec requires them to promote to integer
>> 0 or 1, therefore the & operator will yield the desired result.
>> So there's not going to be any incorrect behavior.
>
>
> So, my assumption is incorrect.
>

Right, but as Tom pointed it is still better to change this. However,
I am not sure if we should backpatch this to PG15 as this won't lead
to any incorrect behavior.

--
With Regards,
Amit Kapila.

Amit Kapila <amit.kapila16@gmail.com> writes:
> Right, but as Tom pointed it is still better to change this. However,
> I am not sure if we should backpatch this to PG15 as this won't lead
> to any incorrect behavior.

If that code only exists in HEAD and v15 then I'd backpatch.
It's a very low-risk change and it might avoid merge problems
for future backpatches.

            regards, tom lane

On Sat, Aug 20, 2022 at 10:04 AM Tom Lane <tgl@sss.pgh.pa.us> wrote:
>
> Amit Kapila <amit.kapila16@gmail.com> writes:
> > Right, but as Tom pointed it is still better to change this. However,
> > I am not sure if we should backpatch this to PG15 as this won't lead
> > to any incorrect behavior.
>
> If that code only exists in HEAD and v15 then I'd backpatch.
> It's a very low-risk change and it might avoid merge problems
> for future backpatches.
>

Okay, done that way. Thanks!

-- 
With Regards,
Amit Kapila.