GitLab

The first line of answers contains text and bolding that's calculated
server-side. That calculation is nearly identical to what we do Chrome-
side for non-answers. This change makes it so that we use the Chrome
calculation to avoid the problem were an old-but-still-matching answer
could be copied to the what-you-typed match and have the wrong bolding.

Because the first line of answers can have some additional text besides
the query, that additional text is added in a separate step.

Because we're no longer using the server-provided bolding, which used
HTML <b> tags, we no longer need the parsing of those tags that we were
previously using.

Finally, since there was code moving around in omnibox_text_view.cc
anyway, I made some simplifications based on the fact that this class
only ever has one render text.

Bug: 721820
Change-Id: I816c55cf9ff680dd922a18ef4cb9301543e1de25
Reviewed-on: https://chromium-review.googlesource.com/1070506Reviewed-by: Dave Schuyler <dschuyler@chromium.org>
Commit-Queue: Justin Donnelly <jdonnelly@chromium.org>
Cr-Commit-Position: refs/heads/master@{#567248}