Thanks for the report. I may have misunderstood, but I cannot reproduce this.

https://codereview.chromium.org/1880043002/ introduces a passing test that shows "[*.]google.com" is a PREDECESSOR of (i.e. more specific than) "*".

Commented on patch.

I made a mistake with my debugging, sorry for the extra work. But it did reveal a question about how I'm matching contentsettingspatterns, which I left for you on the other CL. Thanks for all your help!

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/69745e4858c876f6c2329266c28f01a35a41ea6d

commit 69745e4858c876f6c2329266c28f01a35a41ea6d
Author: msramek <msramek@chromium.org>
Date: Wed Apr 13 08:19:05 2016

Add a ContentSettingsPattern test for "*" vs. a site-scoped pattern.

We have tests for "*".Compare(Wildcard()) == IDENTITY and for
Wildcard().Compare("[*.]google.com") == SUCCESSOR. Assuming the comparison
is transitive, it should hold that "*".Compare("[*.]google.com") == SUCCESSOR.

This CL adds an explicit test case, so that we know that the comparison is,
in this case, in fact transitive.

BUG= 602470 

Review URL: https://codereview.chromium.org/1880043002

Cr-Commit-Position: refs/heads/master@{#386936}

[modify] https://crrev.com/69745e4858c876f6c2329266c28f01a35a41ea6d/components/content_settings/core/common/content_settings_pattern_unittest.cc