On Wed, Jul 05, 2017 at 01:48:06PM +0200, Ulrich Windl wrote:

Ondrej Kuzník ondra@mistotebe.net schrieb am 28.06.2017 um 17:51 in

...

Hi, this is intended and how the issue in ITS#6545 has been fixed.

The operation you are looking at has removed some values, then added other values for the same attribute straight away as another modify within the same op.

What would break if those "Og==" lines were omitted?

ITS#6545: interrupted replication/potential desynchronisation, depending on the operation. See the related changes to the test suite (test063).

Regards,