Move Network Quality Estimator files to //net/nqe

net/socket/OWNERS

Index: net/socket/OWNERS
diff --git a/net/socket/OWNERS b/net/socket/OWNERS
new file mode 100644
index 0000000000000000000000000000000000000000..ddab0624a174b7979fc2802d142992172279a4ba
--- /dev/null
+++ b/net/socket/OWNERS
@@ -0,0 +1 @@
+per-file socket_performance_watcher*=bengr@chromium.org

[Ryan Sleevi, 2016/04/27 20:51:08]

I would prefer not to introduce a per-file OWNERS here. While I realize Ben is
unquestionably the person most likely to know how this all relates, ideally, we
would have more members of the team who know how to review this code.

Ben, do you feel strongly about this? Did I miss a concern?

[tbansal1, 2016/04/27 20:59:49]

Ben can add more but probably this is my bad. I assumed the OWNERS from the
parent folder would also apply to files in this folder too (because there is no
set noparent here). Is that not the case here?

[tbansal1, 2016/04/27 21:24:31]

According to
https://chromium.googlesource.com/chromium/tools/depot_tools/+/master/owners.py,
if there is no "set noparent" directive, OWNERS from parent folder
(//net/OWNERS) are also owners of all the files (including the files that have
per-file owners) in this folder. Ryan, does that answer the concern?

[Ryan Sleevi, 2016/04/28 16:14:44]

Well, not quite, because I still am not terribly comfortable about per-file
OWNERS, so I wanted to confirm with Ben that he still wanted this. I'm not going
to object, but I'd like to make sure this is something intentional with Ben.
Definitely, the relationship with with the rest of socket APIs is important to
consider, so does it make sense for socket owners to do reviewing?

[tbansal1, 2016/04/28 20:06:45]

Done. I have removed //net/socket/OWNERS.