[ovs-dev] Development branch pushed to master?

[Ben Pfaff]

On Tue, Oct 30, 2018 at 03:25:19PM +0300, Ilya Maximets wrote:
> Hello Ben,
> 
> I assume that you pushed your development branch to the master.
> There are a lot of not reviewed and "work in progress" patches.
> Many of them doesn't have even a Sign-off and breaks a lot of
> unit tests.
> 
> Current master looks like this:
> 
> * cb59ce30f 2018-10-29 | manpages: Include ovs.tmac in most man roots. (upstream/master) [Ilya Maximets]
> 
>    ^--- My patch. OK. Reviewed.
> 
> * c2ab9948b 2015-07-06 | Add oftest tests to testsuite. [Ben Pfaff]
> * 0f8ab4e4b 2018-09-27 | netdev-dummy: Implement queue stats. [Ben Pfaff]
> * 79de2c37b 2015-05-24 | Implement emergency flow cache. [Ben Pfaff]
> * b157e9214 2015-05-25 | Preserve priorities for OF1.0 exact-match flows. [Ben Pfaff]
> * 51809c1d7 2015-07-29 | ofproto-dpif: Make barriers more effective. [Ben Pfaff]
> * 40dff9ede 2015-07-04 | ofproto-dpif-upcall: Increment dump_seq before instead of after dump. [Ben Pfaff]
> 
>   ^--- Bunch of patches that never(?) was on mail-list.
>        Some of them breaks a lot of unit test.
>        Most of them has no Sign-off's.
> 
> * 51643ee47 2018-10-09 | ofproto: Don't always treat passive controllers as "equal". [Ben Pfaff]
> * d51e5b1a2 2018-10-26 | vswitchd: Allow user to configure controllers as "primary" or "service". [Ben Pfaff]
> * f58e3cec2 2018-10-24 | connmgr: Make treatment of active and passive connections more uniform. [Ben Pfaff]
> * f122a5b87 2018-10-24 | connmgr: Improve interface for setting controllers. [Ben Pfaff]
> * b7e77e073 2018-10-29 | rconn: New function rconn_is_reliable(). [Ben Pfaff]
> * 53348f4b0 2018-10-25 | connmgr: Modernize coding style. [Ben Pfaff]
> 
>   ^--- Bunch of just submitted to mail-list patches. Not reviewed yet, I guess.
> 
> 
> I think, that we need to revert at least the second group of patches, because
> they're breaking the tests and, probably, the functionality.
> Not sure about the third group, but they needs to be reviewed.
> 
> Force-push is also an option, probably.

Oops.  Hasn't happened before, weird that I screwed up like that after
10 years.

I decided to just force-push the correct commits in this case.