Thread: Review for pg_dump: Sort overloaded functions in deterministic order

Review for pg_dump: Sort overloaded functions in deterministic order

From
Joachim Wieland
Date:
Patch looks good, all concerns that had been raised previously have
been addressed in this version of the patch.

The only thing that IMO needs to change is a stylistic issue:

if (fout->remoteVersion >= 80200)
{   [...]   (fout->remoteVersion >= 80400) ?
"pg_catalog.pg_get_function_identity_arguments(oid)" : "NULL::text",   [...]
}

Please just create a whole new

if (fout->remoteVersion >= 80400)
{  [...]
}

here.

Other than that, the feature works as advertised and does not
negatively affect runtime or memory consumption (the new field is only
added to functions / aggregates).

Please send a new version of the patch that changes the above
mentioned item, the patch also needs rebasing (off by a couple of
lines).



Re: Review for pg_dump: Sort overloaded functions in deterministic order

From
Joel Jacobson
Date:
Hi Joachim,

Attached, please find new patch. Test unchanged.

Best regards,

Joel

> Patch looks good, all concerns that had been raised previously have
> been addressed in this version of the patch.
>
> The only thing that IMO needs to change is a stylistic issue:
>
> if (fout->remoteVersion >= 80200)
> {
>     [...]
>     (fout->remoteVersion >= 80400) ?
> "pg_catalog.pg_get_function_identity_arguments(oid)" : "NULL::text",
>     [...]
> }
>
> Please just create a whole new
>
> if (fout->remoteVersion >= 80400)
> {
>    [...]
> }
>
> here.
>
> Other than that, the feature works as advertised and does not
> negatively affect runtime or memory consumption (the new field is only
> added to functions / aggregates).
>
> Please send a new version of the patch that changes the above
> mentioned item, the patch also needs rebasing (off by a couple of
> lines).

Attachment

Re: Review for pg_dump: Sort overloaded functions in deterministic order

From
Alvaro Herrera
Date:
Joel Jacobson wrote:
> Hi Joachim,
>
> Attached, please find new patch. Test unchanged.

This was committed, as discussed in the original patch's thread.

It would be great if reviewers could reply to the email that submits the
patch, instead of creating a thread of their own.  It helps keep things
better organized.

Thanks for the review.

--
Álvaro Herrera                http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services