Factor out a perf test result printer method.

chrome/chrome_browser.gypi

Index: chrome/chrome_browser.gypi
diff --git a/chrome/chrome_browser.gypi b/chrome/chrome_browser.gypi
index c2b9674aeb1f95d518e373e1e4b1fdfbdf917f93..e64d579c2dc90a3b359886ee056c1aa4428ebeda 100644
--- a/chrome/chrome_browser.gypi
+++ b/chrome/chrome_browser.gypi
@@ -2491,6 +2491,7 @@
             '../components/components.gyp:navigation_interception',
             '../net/net.gyp:net_with_v8',
             '../printing/printing.gyp:printing',
+            '../testing/perf/perf_test.gyp:*',

[Nico, 2013/08/27 03:49:03]

This is a bit unfortunate. I don't think we want to have 'browser' depend on a
target in the testing/ folder. testing/ looks too much like test-only code, so
people will add stuff to this target, and then it ends up being compiled into
your shipping binary.

Is this dependency needed? What for?

[tonyg, 2013/08/27 04:08:31]

Oh jeez, this is ridiculous. I've just learned that apparently we link the page
cyclers into the chrome binary to support a --visit-urls chrome flag which runs
the page cyclers.

This never should have landed and I'd like to rip it out:
https://chromiumcodereview.appspot.com/10386134

For the purposes of this patch, I'm not making the dependency any worse. Would
you be okay if I landed this patch as-is then put together a subsequent patch to
rip the page cyclers out of the chrome binary.

[Nico, 2013/08/27 15:09:21]

Sounds good, please add a comment with a link to the bug you filed to this line
too (if it's still needed after your other CL).

[tonyg, 2013/08/27 15:23:59]

Done.

             '../third_party/adobe/flash/flash_player.gyp:flapper_version_h',
             '../third_party/expat/expat.gyp:expat',
             '../third_party/hunspell/hunspell.gyp:hunspell',