Gerrit: Reliability and fidelity of verification results

I'd like to rise a question about reliability and fidelity of Jenkins
Gerrit change verification results. Consider this change: 16877.

* On patch set 8: [1] TB 66 voted VRFY+1 [2].
* On patch set 9: [3] TB 60 voted VRFY1-1 [4].

There weren't any changes in the affected file textdoc.cxx between those
two patch sets. Moreover, while I don't have access to Mac OS X, I
verified that the patch set 9 works as expected on Linux, with most
recent clang version available. Excerpt from my autogen.input:

  CXX=/home/davido/pgm/llvm-clang-trunk/inst/bin/clang++

And the version of clang:

  $ clang++ --version
  clang version 3.7.0 (trunk 238405)

Now what?

[1] https://gerrit.libreoffice.org/16877/8
[2]
http://ci.libreoffice.org/job/lo_gerrit_master/4083/Gerrit=Gerrit,Platform=MacOSX/consoleFull
[3] https://gerrit.libreoffice.org/16877/9 
[4]
http://ci.libreoffice.org/job/lo_gerrit_master/4097/Gerrit=Gerrit,Platform=MacOSX/consoleFull