GitLab

[755-Kea-DHCP-servers-observations-should-be-reset] implement required changes to dhcp4 and dhcp6 servers

[755-Kea-DHCP-servers-observations-should-be-reset-to-some-initial-values-upon-restart-or-reconfiguration] add initialization of all statistisc during set up

[575-cb-cmds-unknown-fields-in-arguments-should-raise-an-error] Added metadata and fixed server tests

[616-error-msgs-contain-references-to-config-file-while-config-backend-is-used] Improved messages (but still not found the bug)

Previously it was stored in TEST_DATA_BUILDDIR which resides inside repo sources.
Due to the fact that when repo was located in deep patch creating socket was
failing as max socket patch is about 100 characters. Now it is located
in temp folder managed by Sandbox class. The sandbox directory is created
in test constructor and deleted in destructor. As the temp directory
is in form /tmp/kea-XXXXXX the lnegth is always lower than 1000, so running
the unit tests never fails.

doc/guide/ctrl-channel.xml
    Minor update to using socat

src/bin/dhcp4/tests/ctrl_dhcp4_srv_unittest.cc
src/bin/dhcp4/tests/ctrl_dhcp4_srv_unittest.cc
    Updated testing for server response to server
    side control channel timeouts

src/lib/cc/json_feed.h
        JSONFeed::getProcessedText() - new method that
        returns a copy of the current accumulation of
        accepted text

src/lib/config/config_messages.mes
src/lib/config/command_mgr.cc
    Connection::receiveHandler() - added log info about
    discarded data when client closes connection

    Connection::timeoutHandler() - added info about discarded
    data to server's response to client on server side timeout

It gives unpredictable results on OpenBSD. Okayed on jabber.

This is trivial fix which I am committing on my discretion.

Also, when timeout occurs in control channel connection, all async tasks
are cancelled.