[Webkit-unassigned] [Bug 50082] [Qt] Use QNetworkAccessManager zerocopy feature from QNetworkReplyHandler

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=50082





--- Comment #7 from Markus Goetz <markus.goetz at nokia.com>  2010-12-02 08:49:33 PST ---
(From update of attachment 75372)
View in context: https://bugs.webkit.org/attachment.cgi?id=75372&action=review

Uploading new patch in a minute:)

>> WebCore/platform/SharedBuffer.h:45
>> +#endif
> 
> I trust you that this is git showing the diff wrong... if not, you need to fix it

Actually it is not showing it wrong. Just unintuitive. check the whole context.

>> WebCore/platform/network/qt/QNetworkReplyHandler.cpp:237
>> +    // Then this buffer is provided to us in the QNetworkReply
> 
> Where is the dot!

argh

>> WebCore/platform/network/qt/QNetworkReplyHandler.cpp:376
>> +        m_usingZeroCopy = true;
> 
> Please rename m_usingZeroCopy to m_usingZeroCopyFeature

No, with this patch we are always using the zero copy feature.
What we are checking is if the server allowed us to do it by giving a content-length header etc.

>> WebCore/platform/network/qt/ResourceHandleQt.cpp:180
>> +    return true;
> 
> Shouldnt there be a "if usingZeroCopyFeature" here?

No, it's always buffered by QNetworkReplyHandler now.
It's just different ways of buffering depending if zerocopy could be used or not. See QNetworkReplyHandler::sendResponseIfNeeded()

>> WebCore/platform/qt/SharedBufferQt.cpp:50
>> +    SharedBuffer* sb = new SharedBuffer();
> 
> Shouldn't this be RefPtr<SharedBuffer*> sb = new SharedBuffer() ?

I copied/modified that from somewhere else and as far as I understand the refcounting it should be fine..

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.