Possible duplicate for #189

Closing this and replacing with issue #690. The salient concern is overall memory size and not just console data.

My review comments:

Naming/Configuration structure

The current state of this issue defines these two env vars:
SCOPE_LOG_BUFFER_THRESHOLD
SCOPE_LOG_FLUSH_PERIOD
with this structure:

        "libscope": {
          "log": {
            "bufthreshold": 32768,
            "flushperiod": 2000,
            "level": "warning",
            "transport": {
              "type": "file",
              "path": "/tmp/scope.log",
              "buffering": "line"
            }
          },
          "configevent": "true",
          "summaryperiod": 10,
          "commanddir": "/tmp"
        },

What I'd propose is renaming the env vars to:
SCOPE_CONSOLE_BUFFER_THRESHOLD
SCOPE_CONSOLE_FLUSH_PERIOD
with this structure:

        "libscope": {
          "log": {
            "level": "warning",
            "transport": {
              "type": "file",
              "path": "/tmp/scope.log",
              "buffering": "line"
            }
          },
          "console": {
            "bufthreshold": 32768,
            "flushperiod": 2000,
          },
          "configevent": "true",
          "summaryperiod": 10,
          "commanddir": "/tmp"
        },

Comments in scope.yml

We'll need to be clear that these affect two different things:
the handling of 1) file and 2) console watch types

Testing

As part of testing these changes, we'll want to be sure to test "reconfiguration" of a process at least manually for these two new configuration variables. I think things look like they should work, but I wonder if we might be missing a call to ctlDisconnect(g_prevctl, CFG_LS); at the end of doConfig().