Bug 26857 – [Chromium] Upstream V8Binding

• Component: WebCore Misc.
• Status: RESOLVED FIXED

[Nate Chapin]

See summary.

[Nate Chapin]

Created attachment 32090 [details]
patch

[Eric Seidel (no email)]

Comment on attachment 32090 [details]
patch

Seems strange that valueToStringWithNullOrUndefinedCheck is the odd-man out here.  All the rest are toWebCoreString() instead of valueTo*

[Dimitri Glazkov (Google)]

Comment on attachment 32090 [details]
patch

Clearing the flag, Nate's working on a bigger/better naming convention. Oh, and it's also faster. Because obviously fastness is important.

[Nate Chapin]

Created attachment 32130 [details]
patch2

[Nate Chapin]

Created attachment 32159 [details]
patch3

Forgot to edit CodeGeneratorV8.pm upstream in previous patches.

[David Levin]

Comment on attachment 32159 [details]
patch3

A few things to address.


> Index: WebCore/bindings/v8/V8Binding.cpp
> +class WebCoreStringResource: public v8::String::ExternalStringResource {

Add a space after WebCoreStringResource.

> +public:
> +    explicit WebCoreStringResource(const String& string)
> +        : m_impl(string.impl()) { }

The { } should go on separate lines at this point.

> +
> +    virtual ~WebCoreStringResource() {}

Add space in {}


> +String v8StringToWebCoreString(v8::Handle<v8::String> v8String, bool externalize)
> +{
> +    WebCoreStringResource* stringResource = static_cast<WebCoreStringResource*>(v8String->GetExternalStringResource());
> +    if (stringResource)
> +        return stringResource->webcoreString();
> +
> +    int length = v8String->Length();
> +    if (length == 0) {

Comparison to 0.


> +String v8ValueToWebCoreString(v8::Handle<v8::Value> object)
> +{
> +    if (object->IsString()) {
> +        v8::Handle<v8::String> v8String = v8::Handle<v8::String>::Cast(object);
> +        String webCoreString = v8StringToWebCoreString(v8String, true);
> +        return webCoreString;

This seems overly verbose plus it may inhibit return value optimization.

I'd suggest just doing:
        return v8StringToWebCoreString(v8String, true);



> +    } else if (object->IsInt32()) {
> +        int value = object->Int32Value();
> +        // Most numbers used are <= 100. Even if they aren't used there's very little in using the space.
> +        const int kLowNumbers = 100;
> +        static AtomicString lowNumbers[kLowNumbers + 1];

Consider using DEFINE_STATIC_LOCAL



> Property changes on: WebCore/bindings/v8/V8Binding.cpp
> ___________________________________________________________________
> Added: svn:executable
>    + *

This should be fixed.



> Index: WebCore/bindings/v8/V8Binding.h
> +#include "config.h"

Don't include config.h in header files.

[Nate Chapin]

Created attachment 32198 [details]
patch4

[David Levin]

Comment on attachment 32198 [details]
patch4

A few last things to address:

> Index: WebCore/bindings/v8/V8Binding.cpp
> +class WebCoreStringResource : public v8::String::ExternalStringResource {
> +public:
> +    explicit WebCoreStringResource(const String& string)
> +        : m_impl(string.impl()) { }

The { } should go on separate lines (since you have initializers).


> +String v8StringToWebCoreString(v8::Handle<v8::String> v8String, bool externalize)
> +{
...
> +
> +    int length = v8String->Length();
> +    if (length) {

if (!length)



> +String v8ValueToWebCoreString(v8::Handle<v8::Value> object)
> +{
...
> +        return webCoreString;
> +    } else {

This else should go away since there is a "return" finishing the previous block.

[Nate Chapin]

Created attachment 32212 [details]
patch5

[Nate Chapin]

Rolled back in r45574 because it was going to be breaking the Chromium canary for too long, will resubmit tomorrow.

[Nate Chapin]

Re-landed at r45603.