On Tue, Mar 12, 2019 at 10:27 AM Michael Paquier <mich...@paquier.xyz>
wrote:

> It could be an idea to split the patch in two pieces:
> - One patch which refactors the code for the new option in
> PostgresNode.pm
> - Second patch for the new test with integration in RewindTest.pm.
> This should touch different parts of the code, so combining both would
> be fine as well for me :)
> --
> Michael
>

Have updated the patch doing as you suggested


On Wed, Mar 20, 2019 at 7:45 AM Michael Paquier <mich...@paquier.xyz> wrote:

> I would have avoided
> extra routines in the patch, like what you have done with
> create_standby_tbl_mapping(), but instead do something like
> init_from_backup() which is able to take extra parameters in a way
> similar to has_streaming and has_restoring.  However, the trick with
> tablespace mapping is that the caller of backup() should be able to
> pass down multiple tablespace mapping references to make that a
> maximum portable.
> --
> Michael


Also updated the patch to achieve your suggestion.

Attachment: 0001-Refactors-the-code-for-the-new-option-in-PostgresNod.patch
Description: Binary data

Attachment: 0002-Add-new-test-with-integration-in-RewindTest.pm.patch
Description: Binary data

Reply via email to