Merge lp://staging/~robert-ancell/indicator-session/blank-username into lp://staging/indicator-session/14.04

PASSED: Continuous integration, rev:425
http://jenkins.qa.ubuntu.com/job/indicator-session-ci/114/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/indicator-session-trusty-amd64-ci/10
    SUCCESS: http://jenkins.qa.ubuntu.com/job/indicator-session-trusty-armhf-ci/10
        deb: http://jenkins.qa.ubuntu.com/job/indicator-session-trusty-armhf-ci/10/artifact/work/output/*zip*/output.zip

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/indicator-session-ci/114/rebuild

What about using g_unichar_isspace() since we expect the real name to be
a UTF8 string?

I had a long look at what it takes to parse it as UTF-8, I'm not sure if it's worth the effort.
- We really can't be sure it is valid UTF-8 so we have to validate it at the top of the function
- We have to handle invalid UTF8 cases which makes the code more complex
- This only helps if there are only some obscure non-printable UTF8 characters used - it's completely valid to check for the ASCII characters as all the UTF8 will be ignored

So I'd say "don't bother", what do you think?

I guess my concern was that the UTF-8 code would be a multibyte code, where one of those bytes is a space. So we'd improperly detect a space when really it was a valid visible UTF-8 character. But since what we're detecting is all spaces, that would require both bytes to be valid ASCII spaces, and the name to be a single multi-byte character in length. Which seems pretty rare.