Eduard Bagdasaryan

> split in two or three parts I separated the first part as [PR756](https://github.com/squid-cache/squid/pull/756).

> why Optional is necessary here You are right that we could reach the same goal with the default constructed Ip::Address object (it lacks isEmpty(), though) but in this case...

> Please separate the flags refactoring from the interception refactoring from the PROXY protocol addition I separated all the refactoring required for the new feature in a dedicated [PR835.](https://github.com/squid-cache/squid/pull/835) >...

> it is existing, working code rather than something that still needs to be designed and written You are right, initially, this PR832 encompassed all the required changes. After discussing...

Since there are not open change requests here - I think we shoud remove S-waiting-for-author until we discuss/agree on possible solution(s).

> Perhaps we should add S-waiting-for-prerequisite? Or an even more general S-waiting-for-other IMO both 'prerequisite' and 'other' are quite general - e.g., 'reviewer' can fall into these categories. I would...

> As far as I know there is "master merge" option in plugin I know nothing about this option. AFAIK, the plugin merges PR on its parent branch automatically and...

I found another [issue](https://github.com/jenkinsci/ghprb-plugin/issues/563#issuecomment-325833595) explaining this 'double' meaning of 'sha1' variable, which correlates with my tests. Obviously, this is a documentation flaw: such things should be warned about (at least)....

> Of cause when you clone just branch and build just in branch - it knows nothing about merge status. To be more precise, the plugin does not clone the...

> Could you please add a test Ok, added a separate `decorations.test`. I copied the basic `RUN` commands from basic.test - but perhaps there is a better way to avoid...