[webkit-reviews] review granted: [Bug 235615] REGRESSION(r287684) speedtest.net uses many GB of memory : [Attachment 450003] Patch

[bugzilla-daemon at webkit.org]

youenn fablet <youennf at gmail.com> has granted Jean-Yves Avenard [:jya]
<jean-yves.avenard at apple.com>'s request for review:
Bug 235615: REGRESSION(r287684) speedtest.net uses many GB of memory
https://bugs.webkit.org/show_bug.cgi?id=235615

Attachment 450003: Patch

https://bugs.webkit.org/attachment.cgi?id=450003&action=review




--- Comment #15 from youenn fablet <youennf at gmail.com> ---
Comment on attachment 450003
  --> https://bugs.webkit.org/attachment.cgi?id=450003
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=450003&action=review

> Source/WebCore/loader/cache/CachedRawResource.cpp:71
> +	   auto incrementalData = data.getSomeData(previousDataSize);

This makes SharedBufferDataView::createSharedBuffer() or more generally
SharedBuffer created from views potentially harmful.

Unrelated but I wonder whether we need to update m_data (makeContiguous call)
in the case of DataBufferingPolicy::DoNotBufferData.
It seems there is a potential to further reduce some memory allocations in that
case: XHR, fetch...