Thread: Refactor ReindexStmt and its "concurrent" boolean

Hi all,

$subject has been mentioned a couple of times, including today:
https://www.postgresql.org/message-id/20200902010012.GE1489@paquier.xyz

We have a boolean argument in ReindexStmt to control a concurrent
run, and we also have in parallel of that a bitmask to control the
options of the statement, which feels like a duplicate.  Attached is a
patch to refactor the whole, adding CONCURRENTLY as a member of the
available options.  This simplifies a bit the code.

Any thoughts?
--
Michael

On Wed, Sep 2, 2020 at 1:03 PM Michael Paquier <michael@paquier.xyz> wrote:
>
> Hi all,
>
> $subject has been mentioned a couple of times, including today:
> https://www.postgresql.org/message-id/20200902010012.GE1489@paquier.xyz
>
> We have a boolean argument in ReindexStmt to control a concurrent
> run, and we also have in parallel of that a bitmask to control the
> options of the statement, which feels like a duplicate.  Attached is a
> patch to refactor the whole, adding CONCURRENTLY as a member of the
> available options.  This simplifies a bit the code.
>
> Any thoughts?


+1

 struct ReindexIndexCallbackState
 {
- bool concurrent; /* flag from statement */
+ bool options; /* flag from statement */
  Oid locked_table_oid; /* tracks previously locked table */
 };

Shouldn't options be an int?  The rest of the patch looks good to me.

On Wed, Sep 02, 2020 at 01:17:32PM +0200, Julien Rouhaud wrote:
> Shouldn't options be an int?  The rest of the patch looks good to me.

It should, thanks for looking at it.  Let's wait a couple of days and
see if others have any comments.  If there are no objections, I'll try
to commit this one.
--
Michael

On Thu, Sep 03, 2020 at 11:14:38AM +0900, Michael Paquier wrote:
> It should, thanks for looking at it.  Let's wait a couple of days and
> see if others have any comments.  If there are no objections, I'll try
> to commit this one.

And applied.
--
Michael