[webkit-reviews] review requested: [Bug 39382] Factor PageCache code out of FrameLoader into a PageCacheController : [Attachment 57709] Only try to cache when the main frame is being committed

[bugzilla-daemon at webkit.org]

Nate Chapin <japhet at chromium.org> has asked  for review:
Bug 39382: Factor PageCache code out of FrameLoader into a PageCacheController
https://bugs.webkit.org/show_bug.cgi?id=39382

Attachment 57709: Only try to cache when the main frame is being committed
https://bugs.webkit.org/attachment.cgi?id=57709&action=review

------- Additional Comments from Nate Chapin <japhet at chromium.org>
I hadn't realized that I had changed the assumption of when we would try to
cache.	If we maintain the interface I introduced in this patch
(canCache(Page*) instead of canCache(Frame*)), there's no way for canCache() to
confirm that this was called due to a commit in the main frame.  It makes sense
(to me anyway) that trying to cache from something other than the main frame is
wrong, so this version just adds a !m_frame->tree()->parent() check before
calling canCache() in FrameLoader::commitProvisionalLoad().



Unrelated:  In my digging into isLoadingInAPISense(), I've noticed that it's
recursive and checks all subframes.  Ergo, we're being unnecessarily O(n^2)
when we call it on every subframe as well as the main frame.  Do we care?