keepalive: ensure `close_all()` actually closes all cached connections
While debugging why LFS blob downloads are getting corrupted with workers, I
noticed that prior to spinning up the workers, the ConnectionManager has 2
connections to the server and calling `KeepAliveHandler.close_all()` left one
behind. The reason is the value component of `self._cm.get_all().items()` is a
list, and `self._cm.remove()` modifies said list while the caller is iterating
over it. Now `get_all()` is a deep copy of both the dict and lists in all
cases.
$ hg init
$ echo a > a
$ hg ci -Am t
adding a
$ hg mv a b
$ hg ci -Am t1
$ hg debugrename b
b renamed from a:b789fdd96dc2f3bd229c1dd8eedf0fc60e2b68e3
$ hg mv b a
$ hg ci -Am t2
$ hg debugrename a
a renamed from b:37d9b5d994eab34eda9c16b195ace52c7b129980
$ hg debugrename --rev 1 b
b renamed from a:b789fdd96dc2f3bd229c1dd8eedf0fc60e2b68e3