Non-SFI mode: Enable browser_tests for nacl_helper_nonsfi.

chrome/browser_tests.isolate

 # Copyright (c) 2012 The Chromium Authors. All rights reserved.
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 {
   'conditions': [
     ['OS=="linux"', {
       'variables': {
         'command': [
           '../testing/xvfb.py',
           '<(PRODUCT_DIR)',
@@ skipping 204 matching lines @@
           '<(PRODUCT_DIR)/keyboard_resources.pak',
           '<(PRODUCT_DIR)/libexif.so',
           '<(PRODUCT_DIR)/locales/',
           '<(PRODUCT_DIR)/nacl_helper',
           '<(PRODUCT_DIR)/nacl_irt_x86_64.nexe',
           '<(PRODUCT_DIR)/resources/chromeos/',
           'browser/chromeos/login/test/https_forwarder.py',
         ],
       },
     }],
+    # TODO(hidehiko,mazda): Include nacl_helper_nonsfi for ARM when supported.
+    ['chromeos==1 and (target_arch=="x64" or target_arch=="ia32")', {

[Mark Seaborn, 2014/11/14 16:58:41]

Why restrict to chromeos==1?  We want to run these tests on the normal Linux
bots, don't we?

[hidehiko, 2014/11/14 18:45:31]

This is here, because nacl_helper is listed above section (condition is
chromeos==1). Isn't it needed, too?

csharp@, any thoughts?

[Mark Seaborn, 2014/11/19 06:23:54]

Presumably the tests pass without adding this to browser_tests.isolate?  If so,
can you not add it, please?  This seems to be doubly-redundant:  Gating it on
chromeos==1 means it's a no-op on the main bots, and it seems to be redundant
with the decl in chrome.isolate (as in the earlier comment).

I don't think you should add something unnecessary just to be consistent with a
nearby part of the code. :-)

+      'variables': {
+        'files': [
+          '<(PRODUCT_DIR)/nacl_helper_nonsfi',

[Mark Seaborn, 2014/11/14 16:58:41]

Don't we get this via chrome.isolate below?

[hidehiko, 2014/11/14 18:45:32]

Acknowledged.

+        ],
+      },
+    }],
   ],
   'includes': [
     '../base/base.isolate',
     'chrome.isolate',
   ],
 }