IDL compiler: mostly merge Non-wrapper types and RefPtr conversion exceptions

IDL compiler: mostly merge Non-wrapper types and RefPtr conversion exceptions

This makes no changes to generated code, and
reduces the special cases to 2 types:
DOMStringList and XPathNSResolver,
each used a handful of places.

Uses:

DOMStringList:
Location.idl
SecurityPolicy.idl
IDBDatabase.idl
IDBObjectStore.idl

XPathNSResolver:
DocumentXPathEvaluator.idl (partial for Document)
XPathEvaluator.idl


Background:
In the rewrite, I noticed that the list of exception in GetNativeType
(where the native Blink C++ type is |RefPtr<$idlType>|)
largely overlapped with nonWrapperTypes, and haraken requested that I
merge these if possible.

BUG=239771
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=160010

[Nils Barth (inactive), 2013-10-19 08:27:53]

Hi haraken,
Here's the "merge NON_WRAPPER_TYPES and REFPTR_CONVERSION_TYPES" CL
(Perl CG changes).

There are 2 types for which a full merge changes the generated code,
though it does still compile.
How does it look?

[haraken, 2013-10-19 08:31:26]

Would you upload a diff of the generated code?

[Nils Barth (inactive), 2013-10-19 10:12:57]

On 2013/10/19 08:31:26, haraken wrote:
> Would you upload a diff of the generated code?

Done, uploaded unified .diff.

[haraken, 2013-10-19 10:49:45]

On 2013/10/19 10:12:57, Nils Barth wrote:
> On 2013/10/19 08:31:26, haraken wrote:
> > Would you upload a diff of the generated code?
> 
> Done, uploaded unified .diff.

Thanks. As far as I see the diff, you're changing web-exposed behavior. For
example:

  46  static void transactionMethod(const v8::FunctionCallbackInfo<v8::Value>&
args)
  47  {
  48 -    if ((args.Length() == 1 && (V8DOMStringList::HasInstance(args[0],
args.GetI
     solate(), worldType(args.GetIsolate())))) || (args.Length() == 2 &&
(V8DOMString
     List::HasInstance(args[0], args.GetIsolate(),
worldType(args.GetIsolate()))))) {
  49 +    if (args.Length() == 1 || args.Length() == 2) {
  50          transaction1Method(args);
  51          return;
  52      }

and:

 163 -    // Custom Signature 'createExpression'
 164 -    const int createExpressionArgc = 2;
 165 -    v8::Handle<v8::FunctionTemplate>
createExpressionArgv[createExpressionArgc]
      = { v8::Handle<v8::FunctionTemplate>(), v8::Handle<v8::FunctionTemplate>()
};
 166 -    v8::Handle<v8::Signature> createExpressionSignature =
v8::Signature::New(de
     sc, createExpressionArgc, createExpressionArgv);
 167 -    proto->Set(v8::String::NewSymbol("createExpression"),
v8::FunctionTemplate:
     :New(XPathEvaluatorV8Internal::createExpressionMethodCallback,
v8Undefined(), cr
     eateExpressionSignature, 0));
 168 -

etc.

Overall, I think you're doing a right thing in that you're getting rid of weird
special casing. However, you need to do it more incrementally, verifying that
each change doesn't change web-exposed behavior or the behavior change is OK.

[Nils Barth (inactive), 2013-10-19 10:57:11]

On 2013/10/19 10:49:45, haraken wrote:
> Thanks. As far as I see the diff, you're changing web-exposed behavior.
> 
> Overall, I think you're doing a right thing in that you're getting rid of
weird
> special casing. However, you need to do it more incrementally, verifying that
> each change doesn't change web-exposed behavior or the behavior change is OK.

Got it; uploaded patch with no changes to generated code, and reduces us to 2
special cases.
We can then check these individually.

[haraken, 2013-10-19 11:02:48]

LGTM

[commit-bot: I haz the power, 2013-10-19 11:05:33]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/nbarth@chromium.org/28863007/130001

[commit-bot: I haz the power, 2013-10-19 12:22:29]

Retried try job too often on linux_blink_rel for step(s) webkit_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_blin...

[commit-bot: I haz the power, 2013-10-19 13:52:48]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/nbarth@chromium.org/28863007/130001

[commit-bot: I haz the power, 2013-10-19 14:33:35]

Change committed as 160010