Add a recipe to identify culprits for chromium compile failures.

scripts/slave/recipe_modules/filter/api.py

Index: scripts/slave/recipe_modules/filter/api.py
diff --git a/scripts/slave/recipe_modules/filter/api.py b/scripts/slave/recipe_modules/filter/api.py
index 059f848883355bbb2281a61ae6e3cd4220d6deb2..38497b7893b9a48f365ef073799d2de9c22b91e3 100644
--- a/scripts/slave/recipe_modules/filter/api.py
+++ b/scripts/slave/recipe_modules/filter/api.py
@@ -159,14 +159,25 @@ class FilterApi(recipe_api.RecipeApi):
         # Ensure that mb runs in a clean environment to avoid
         # picking up any GYP_DEFINES accidentally.
         del kwargs['env']
+      # We need to use the actual mastername and buildername we're running on,
+      # because it may be configured with different MB settings from those for
+      # the continuous builder a trybot may be configured to match.
+      # However, use target_mastername and target_buildername instead if they
+      # are available in the build properties. This gives a recipe running on a
+      # trybot the flexibility to run MB in exactly the same configuration as
+      # the continuous builder it tries to match.
+      target_mastername = self.m.properties.get('target_mastername') or \
+          self.m.properties['mastername']
+      target_buildername = self.m.properties.get('target_buildername') or \
+          self.m.properties['buildername']
       step_result = self.m.python(
           'analyze',
           self.m.path['checkout'].join('tools', 'mb', 'mb.py'),
           args=['analyze',
                 '-m',
-                self.m.properties['mastername'],
+                target_mastername,
                 '-b',
-                self.m.properties['buildername'],
+                target_buildername,
                 '-v',
                 build_output_dir,
                 self.m.json.input(analyze_input),