Speed up test via synchronous content! calls

Removed 5 `sleep(1)` statements, and reduced test time by 4.5 seconds:

```
$ time julia test/content/api.jl  # Before
julia test/content/api.jl  15.07s user 0.33s system 28% cpu 54.958 total
$ time julia test/content/api.jl  # After
julia test/content/api.jl  15.60s user 0.35s system 31% cpu 51.197 total
```

test/content/api.jl

@@ -3,15 +3,15 @@ using Base.Test
 
 @testset "content! Tests" begin
     w = Window(Blink.@d(:show => false)); sleep(5.0)
-    body!(w, ""); sleep(1.0)  # body! is not synchronous.
+    body!(w, "", async=false);
     @test (@js w document.querySelector("body").innerHTML) == ""
 
     # Test reloading body and a selector element.
     html = """<div id="a">hi world</div><div id="b">bye</div>"""
-    body!(w, html); sleep(1.0)
+    body!(w, html, async=false);
     @test (@js w document.getElementById("a").innerHTML) == "hi world"
     @test (@js w document.getElementById("b").innerHTML) == "bye"
-    content!(w, "div", "hello world"); sleep(1.0)
+    content!(w, "div", "hello world", async=false);
     @test (@js w document.getElementById("a").innerHTML) == "hello world"
     # TODO(nhdaly): Is this right? Should content!(w,"div",...) change _all_ divs?
     @test (@js w document.getElementById("b").innerHTML) == "bye"
@@ -22,14 +22,14 @@ using Base.Test
         # Must create a new window to ensure javascript is reset.
         w = Window(Blink.@d(:show => false)); sleep(5.0)
 
-        body!(w, fadeTestHtml; fade=true); sleep(1.0)
+        body!(w, fadeTestHtml; fade=true, async=false);
         @test (@js w testJS) == "test"
     end
     @testset "Fade False" begin
         # Must create a new window to ensure javascript is reset.
         w = Window(Blink.@d(:show => false)); sleep(5.0)
 
-        body!(w, fadeTestHtml; fade=false); sleep(1.0)
+        body!(w, fadeTestHtml; fade=false, async=false);
         @test (@js w testJS) == "test"
     end
 end