Bug 156464 – Fix a bug where createNewPage is not working with warmInitialProcess

• Component: WebKit2
• Status: NEW

[Ryuan Choi]

If I am right, createNewPage must create new page with the process of related page regardless of m_haveInitialEmptyProcess.

[Ryuan Choi]

Created attachment 276139 [details]
Patch

[Ryuan Choi]

Created attachment 276146 [details]
Patch

[Alex Christensen]

Comment on attachment 276146 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=276146&action=review

> Tools/TestWebKitAPI/Tests/WebKit2/CreatePageWithWarmInitialProcess.cpp:87
> +    WKPageLoadURL(webView.page(), url.get());
> +    WKContextWarmInitialProcess(context.get());

Why do you need to warm the initial process after loading a URL?  I think the intent of WarmInitialProcess is that it is called before loading a URL.

[Ryuan Choi]

Comment on attachment 276146 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=276146&action=review

>> Tools/TestWebKitAPI/Tests/WebKit2/CreatePageWithWarmInitialProcess.cpp:87
>> +    WKContextWarmInitialProcess(context.get());
> 
> Why do you need to warm the initial process after loading a URL?  I think the intent of WarmInitialProcess is that it is called before loading a URL.

Yes. It is not for the prior Loading(86line) but just to reserve web process before calling createNewPage().
I called it here just to simplify test case.

IMHO, I think that WarmInitialProcess is not only for first process.
So, I called WarmInitialProcess after created every new tab (and without loading a URL).

But in that case, createNewPage() was not working when called window.open() in other tab.

[Ryuan Choi]

Created attachment 277876 [details]
Patch

[Ryuan Choi]

(In reply to comment #5)
> Created attachment 277876 [details]
> Patch

I modified test case little bit with comment.

IMHO, all operations should be fine whenever we call warmInitialProcess.

[Brady Eidson]

Comment on attachment 277876 [details]
Patch

r-, as this has been pending review for over a year now. It is near-impossible that this patch still applies to trunk and unlikely to still be relevant in its current form.