Do not cache resource metadata if the source code size is small.

Source/bindings/core/v8/V8ScriptRunner.cpp

Index: Source/bindings/core/v8/V8ScriptRunner.cpp
diff --git a/Source/bindings/core/v8/V8ScriptRunner.cpp b/Source/bindings/core/v8/V8ScriptRunner.cpp
index 4bffe58068596dec34b9fbcd24f276be7341af11..80da2d80a79aa14e61f4f149af15b7e3b44751ff 100644
--- a/Source/bindings/core/v8/V8ScriptRunner.cpp
+++ b/Source/bindings/core/v8/V8ScriptRunner.cpp
@@ -280,7 +280,8 @@ PassOwnPtr<CompileFn> selectCompileFunction(V8CacheOptions cacheOptions, CachedM
         // Caching is not available in this case.
         return bind(compileWithoutOptions, V8CompileHistogram::Noncacheable);
 
-    if (cacheOptions == V8CacheOptionsNone)
+    // Do not cache for small scripts, even if caching is available.

[falken, 2015/03/10 00:54:05]

This comment kind of just repeats the code. It'd be nicer to mention the reason
as in the patch description: "For small source code, parsing/compiling may be
faster than deserializing."

+    if (cacheOptions == V8CacheOptionsNone || code->Length() < minimalCodeLength)
         return bind(compileWithoutOptions, V8CompileHistogram::Cacheable);
 
     // The cacheOptions will guide our strategy:
@@ -288,10 +289,6 @@ PassOwnPtr<CompileFn> selectCompileFunction(V8CacheOptions cacheOptions, CachedM
     switch (cacheOptions) {
     case V8CacheOptionsDefault:
     case V8CacheOptionsParseMemory:
-        if (code->Length() < minimalCodeLength) {
-            // Do not cache for small scripts, though caching is available.
-            return bind(compileWithoutOptions, V8CompileHistogram::Cacheable);
-        }
         // Use parser-cache; in-memory only.
         return bind(compileAndConsumeOrProduce, cacheHandler, cacheTag(CacheTagParser, cacheHandler), v8::ScriptCompiler::kConsumeParserCache, v8::ScriptCompiler::kProduceParserCache, false, CachedMetadataHandler::CacheLocally);
         break;