Adjust when _setPrivacyProxyFailClosedForUnreachableNonMainHosts is called
Created attachment 454406 [details] Patch
<rdar://problem/89972004>
Comment on attachment 454406 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=454406&action=review r=me > Source/WebKit/NetworkProcess/cocoa/NetworkDataTaskCocoa.mm:346 > + || request.url().host() == parameters.topOrigin->host()) Is there a condition where (isMainFrameNavigation && parameters.topOrigin && request.url().host() != parameters.topOrigin->host())? If not, maybe we don't need to check isMainFrameNavigation.
Good point. I also need to modify web sockets.
parameters.topOrigin is sometimes null and sometimes contains an origin with an empty host. I'm going to stay on the safe side and keep the main frame main resource check.
Created attachment 454526 [details] Patch
Comment on attachment 454526 [details] Patch I verified that this does not re-introduce the original issue. Intuitively this should fix the perf regression, and I'll check it in and see if the bots agree.
Committed r291206 (248361@main): <https://commits.webkit.org/248361@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 454526 [details].
Re-opened since this is blocked by bug 237829
Created attachment 454998 [details] Patch
Comment on attachment 454998 [details] Patch r=me Not sure what was wrong with the previous patch, or how this fixes it. Can you add that note in the ChangeLog?
I'm under the impression that there was nothing wrong with the previous patch, but it was rolled out before the bots caught up to the fact that it fixed a performance regression.
http://trac.webkit.org/r291598
This was hopefully the final fix for rdar://88965550
Reverting yet again in https://bugs.webkit.org/show_bug.cgi?id=238842
Reopening to attach new patch.
Created attachment 457579 [details] Patch
Created attachment 457580 [details] Patch
r292846
*** Bug 237296 has been marked as a duplicate of this bug. ***