RESOLVED FIXED 13631
Page Cache should support pages with frames
https://bugs.webkit.org/show_bug.cgi?id=13631
Summary Page Cache should support pages with frames
Brady Eidson
Reported 2007-05-08 19:50:38 PDT
The back/forward cache needs to work with pages that have frames <rdar://problem/4886592>
Attachments
Refactor related code out of DocumentLoader and enhance logging (8.44 KB, patch)
2009-08-17 17:25 PDT, Brady Eidson
no flags
Cleanup and scaffolding - no change in behavior (9.97 KB, patch)
2009-09-01 10:58 PDT, Brady Eidson
sam: review+
A next step - more of the handling of the "CachedFrame tree" (7.36 KB, patch)
2009-09-02 09:40 PDT, Brady Eidson
sam: review+
Another step - getting close! (9.14 KB, patch)
2009-09-02 16:09 PDT, Brady Eidson
darin: review+
Almost there... (5.33 KB, patch)
2009-09-02 17:19 PDT, Brady Eidson
darin: review-
Address Darin's comments (6.80 KB, patch)
2009-09-02 17:38 PDT, Brady Eidson
darin: review+
Enable back/forward cache for pages with frames (along with some layout tests) (11.15 KB, patch)
2009-09-03 16:39 PDT, Brady Eidson
sam: review+
Brady Eidson
Comment 1 2009-08-17 17:22:36 PDT
Back to working on this lately. It will be coming in baby steps, trying to make the code a little more sane as we get closer to flipping the switch.
Brady Eidson
Comment 2 2009-08-17 17:25:25 PDT
Created attachment 35002 [details] Refactor related code out of DocumentLoader and enhance logging
Eric Seidel (no email)
Comment 3 2009-08-17 18:13:51 PDT
Comment on attachment 35002 [details] Refactor related code out of DocumentLoader and enhance logging It's now possible to test the page cache (with the setting override stuff). Dumi (I think?) added the first page-cache test the other day. Seems we should add one for this change too.
Brady Eidson
Comment 4 2009-08-17 18:24:32 PDT
If he added page cache layout tests, that's great. In this situation, I wish you could've told me *where* he added these, because I personally don't have time to watch all the commits go by on a day to day basis and don't know where to look. Regardless, since this patch is a pure refactoring and has zero change in behavior, testing the change in behavior for regressions is a non-starter. I did run the layout tests and there were no new failures. Putting back up for review.
Brady Eidson
Comment 5 2009-08-17 18:25:14 PDT
Comment on attachment 35002 [details] Refactor related code out of DocumentLoader and enhance logging Layout test not possible, zero behavior change. Putting back up for review.
Brady Eidson
Comment 6 2009-08-17 18:28:38 PDT
Sam Weinig reviewed this over my shoulder.
Brady Eidson
Comment 7 2009-08-17 18:38:55 PDT
This cleanup landed in r47407.
Brady Eidson
Comment 8 2009-09-01 10:30:43 PDT
Updated Radar URL: <rdar://problem/3541409>
Brady Eidson
Comment 9 2009-09-01 10:58:29 PDT
Created attachment 38871 [details] Cleanup and scaffolding - no change in behavior
Brady Eidson
Comment 10 2009-09-01 11:14:31 PDT
Brady Eidson
Comment 11 2009-09-02 09:40:10 PDT
Created attachment 38928 [details] A next step - more of the handling of the "CachedFrame tree" Again, this doesn't change current behavior, and is just another step in landing my work piece-by-piece.
Brady Eidson
Comment 12 2009-09-02 12:33:26 PDT
This morning's patch landed in http://trac.webkit.org/changeset/47985 More coming...
Brady Eidson
Comment 13 2009-09-02 13:17:19 PDT
http://trac.webkit.org/changeset/47989 is related to this work, as well.
Brady Eidson
Comment 14 2009-09-02 16:09:54 PDT
Created attachment 38944 [details] Another step - getting close!
Brady Eidson
Comment 15 2009-09-02 17:08:24 PDT
Brady Eidson
Comment 16 2009-09-02 17:19:12 PDT
Created attachment 38951 [details] Almost there...
Darin Adler
Comment 17 2009-09-02 17:26:17 PDT
Comment on attachment 38951 [details] Almost there... Should use an enum instead of adding a new bool argument. It's true that in one place the argument is a boolean expression, not a constant, but I still think that these confusing functions can be made less so with enums. In fact, I think you could combine the sendUnload and sendPagehide bool arguments into a single enum if you name the values carefully.
Brady Eidson
Comment 18 2009-09-02 17:38:52 PDT
Created attachment 38955 [details] Address Darin's comments
Brady Eidson
Comment 19 2009-09-02 17:56:00 PDT
Landed in http://trac.webkit.org/changeset/48001 I'll flip the switch sometime tomorrow, after I get a chance to update and build on Windows.
Brady Eidson
Comment 20 2009-09-03 16:10:55 PDT
The patch was reviewed over my shoulder, so it's not in the list, but: http://trac.webkit.org/changeset/48034
Brady Eidson
Comment 21 2009-09-03 16:39:13 PDT
Created attachment 39019 [details] Enable back/forward cache for pages with frames (along with some layout tests)
Brady Eidson
Comment 22 2009-09-03 16:57:48 PDT
Note You need to log in before you can comment on or make changes to this bug.