Current codepath works like: * Send request to read the back buffer to a pixel-buffer * Create native image from pixel-buffer * Draw native image to context 2D We should try to do better especially when GPUP is enabled because this leads to: * WP: Send request to read the back buffer to a pixel buffer * GPUP: glReadPixels to a malloc buffer (WebCore::PixelBuffer) * GPUP: Send the pixel buffer to WP * WP: Creates native image out of that, sends that to GPUP * GPUP: draws that native image to Context2D back-buffer
<rdar://87173127>
Pull request: https://github.com/WebKit/WebKit/pull/10560
Committed 262518@main (50ae4e81ccce): <https://commits.webkit.org/262518@main> Reviewed commits have been landed. Closing PR #10560 and removing active labels.
Regressed: Bug 254942 – REGRESSION(262518@main) [cairo] Crash under GraphicsContextGL::paintToCanvas