On Thu, Aug 18, 2016 at 4:15 PM, Andres Freund <andres@anarazel.de> wrote:
> On 2016-08-18 16:11:27 -0400, Robert Haas wrote:
>> On Wed, Aug 17, 2016 at 11:18 PM, Andres Freund <andres@anarazel.de> wrote:
>> > On August 17, 2016 8:15:56 PM PDT, Michael Paquier <michael.paquier@gmail.com> wrote:
>> >
>> >>+ { /* pg_ctl command w path, properly quoted */
>> >>+ PQExpBuffer pg_ctl_path = createPQExpBuffer();
>> >>+ printfPQExpBuffer(pg_ctl_path, "%s%spg_ctl",
>> >>+ bin_dir,
>> >>+ (strlen(bin_dir) > 0) ? DIR_SEP : ""
>> >>+ );
>> >>+ appendShellString(start_db_cmd, pg_ctl_path->data);
>> >>+ destroyPQExpBuffer(pg_ctl_path);
>> >>+ }
>> >>
>> >>This is not really project-style to have an independent block. Usually
>> >>those are controlled by for, while or if.
>> >
>> > Besides the comment positioning I'd not say that that is against the usual style, there's a number of such blocks
already. Don't think it's necessarily needed here though...
>>
>> Really? I'd remove such blocks before committing anything, or ask for
>> them to be removed, unless there were some special reason for having
>> them.
>
> Well, reducing the scope of variables *can* be such a reason, no? As I
> said, I don't see any reason here, but in general, it's imo a reasonable
> tool on one's belt.
I think it's worth reducing the scope of variables when that's as
simple as putting them inside a block that you have to create anyway,
but I'm skeptical about the idea that one would create a block just to
reduce the scope of the variables. I don't think that's our usual
practice, and I would expect the compiler to detect where the variable
is referenced first and last anyway.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
