I'm trying to land https://codereview.chromium.org/2697013009/, which seems to have changed the expected output of test/inspector/runtime/coverage.js.

I don't see why the output would be affected by my change, but when I tried rebasing the expected output it still doesn't seem consistent: I got different behavior on the v8_win_nosnap_shared_rel_ng_triggered builder: https://build.chromium.org/p/tryserver.v8/builders/v8_win_nosnap_shared_rel_ng_triggered/builds/15724.

Here is the diff between the two:

--- expected-old.txt	2017-02-22 14:50:19.343628974 -0800
+++ expected-current.txt	2017-02-22 14:50:25.483654776 -0800
@@ -208,6 +208,24 @@
                         ranges : [
                             [0] : {
                                 count : 1
+                                endColumnNumber : 11
+                                endLineNumber : 0
+                                startColumnNumber : 0
+                                startLineNumber : 0
+                            }
+                        ]
+                    }
+                ]
+                scriptId : <scriptId>
+                url : 
+            }
+            [1] : {
+                functions : [
+                    [0] : {
+                        functionName : 
+                        ranges : [
+                            [0] : {
+                                count : 1
                                 endColumnNumber : 0
                                 endLineNumber : 9
                                 startColumnNumber : 0
@@ -243,7 +261,14 @@
                 scriptId : <scriptId>
                 url : 4
             }
-            [1] : {
+        ]
+    }
+}
+{
+    id : <messageId>
+    result : {
+        result : [
+            [0] : {
                 functions : [
                     [0] : {
                         functionName : 
@@ -261,14 +286,7 @@
                 scriptId : <scriptId>
                 url : 
             }
-        ]
-    }
-}
-{
-    id : <messageId>
-    result : {
-        result : [
-            [0] : {
+            [1] : {
                 functions : [
                     [0] : {
                         functionName : 
@@ -310,24 +328,6 @@
                 scriptId : <scriptId>
                 url : 4
             }
-            [1] : {
-                functions : [
-                    [0] : {
-                        functionName : 
-                        ranges : [
-                            [0] : {
-                                count : 1
-                                endColumnNumber : 11
-                                endLineNumber : 0
-                                startColumnNumber : 0
-                                startLineNumber : 0
-                            }
-                        ]
-                    }
-                ]
-                scriptId : <scriptId>
-                url : 
-            }
         ]
     }
 }

It's hard to tell from the diff, but it seems to be affecting testBestEffortCoveragePrecise only. The difference is whether the function with "endColumnNumber : 11" is in the 2nd log or the 3rd.