Merge lp://qastaging/~robertcarr/mir/add-log-all-option into lp://qastaging/mir

PASSED: Continuous integration, rev:1203
http://jenkins.qa.ubuntu.com/job/mir-team-mir-development-branch-ci/354/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-trusty-i386-build/149
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-trusty-amd64-build/146
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-team-mir-development-branch-trusty-amd64-ci/83
        deb: http://jenkins.qa.ubuntu.com/job/mir-team-mir-development-branch-trusty-amd64-ci/83/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-team-mir-development-branch-trusty-armhf-ci/83
        deb: http://jenkins.qa.ubuntu.com/job/mir-team-mir-development-branch-trusty-armhf-ci/83/artifact/work/output/*zip*/output.zip

Click here to trigger a rebuild:
http://10.97.0.26:8080/job/mir-team-mir-development-branch-ci/354/rebuild

It might be useful, but there's an elegance to not specifying different sets of logging options to turn on.

What about turning on all the lttng options?

I think there's a "slippery slope" we don't want to start down.

Perhaps a better/more generic way to express this would be a shortcut to apply a report type to all reports that accept it?

--all-reports={log|lttng|...}

"Needs discussion"

Maybe --all-reports is better...the thing is we don't really seem to be using lttng a whole bunch.

I've found very frequently in cases of remote debugging, especially with people not on the Mir team who don't know as quickly where to look goes like:

1. Ok can you run things with MIR_SERVER_LEGACY_INPUT_REPORT=log MIR_SERVER_INPUT_REPORT=log
2. Oh...strange, let's add the MSG_PROCESSOR and SESSION_MEDIATOR report.
3. Oh! We need the connector report.

So it's gotten to where I always ask for people to set all the reports to log as a quite common case.

What do people think of the --log-all option?

> Maybe --all-reports is better...the thing is we don't really seem to be using
> lttng a whole bunch.
>
> I've found very frequently in cases of remote debugging, especially with
> people not on the Mir team who don't know as quickly where to look goes like:
>
> 1. Ok can you run things with MIR_SERVER_LEGACY_INPUT_REPORT=log
> MIR_SERVER_INPUT_REPORT=log
> 2. Oh...strange, let's add the MSG_PROCESSOR and SESSION_MEDIATOR report.
> 3. Oh! We need the connector report.
>
> So it's gotten to where I always ask for people to set all the reports to log
> as a quite common case.
>
> What do people think of the --log-all option?

Perhaps the best way to consolidate the options discussed above would be to have:

--reports-default arg Default setting for all reports.
                              [{log,lttng,off}:default=off]

(I'm not sure how to document what happens if "lttng" is selected as default as many of the reports don't currently support this option.)

I think --reports-default is kind of misleading as you point out about LTTNG. Realistically though I would feel confident guessing that the logs are used over lttng with a factor of something like 100:1 so this is the main case to optimize for.