tnagel@ to CC, who argued for not showing type-mismatched policies in a related conversation recently.

After some discussion my (somewhat adjusted) position is: We should log such kind of problems using LOG(ERROR), but that doesn't imply that they shouldn't be surfaced on chrome://policy *additionally*.

I think this is to be considered a regression because technically there is code that should be producing those errors but it never gets called because something in the guts of our policy parsing doesn't even let the ConfigurationPolicyHanlders see those policies and produce the errors it should.

Ah, if that's the case, I have broken it.  :(

Luckily, Anthony has already a CL to fix it: https://codereview.chromium.org/2400703002/