Bug 26670 - [Win] REGRESSION (r37793): Hitting back button on Google often gives non-interactive Google page
Summary: [Win] REGRESSION (r37793): Hitting back button on Google often gives non-inte...
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Page Loading (show other bugs)
Version: 528+ (Nightly build)
Hardware: All Windows Vista
: P2 Normal
Assignee: Oliver Hunt
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2009-06-23 19:49 PDT by Alice Liu
Modified: 2009-07-02 16:56 PDT (History)
3 users (show)

See Also:


Attachments
patch authored by oliver (1.28 KB, patch)
2009-06-23 19:50 PDT, Alice Liu
no flags Details | Formatted Diff | Diff
fix with changelog (1.61 KB, patch)
2009-07-02 16:51 PDT, Oliver Hunt
andersca: review+
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Alice Liu 2009-06-23 19:49:23 PDT
* SUMMARY
If I search for something on Google, then click the back button, I am often shown a "non-interactive" Google - i.e., it behaves like a screenshot of Google.

* STEPS TO REPRODUCE (using Safari 4.0 on Windows)
1. Open http://www.google.com/
2. Type "test" into the search field
3. If Google Suggest comes up, dismiss it by clicking outside of the search field
4. Click the "Google Search" button to search
5. Click the "Back" button in Safari

* RESULTS
Often, I am given a "non-interactive" page, as if it were just a screenshot of the Google page.

* NOTES
I can't get this to reproduce if, instead of step 3, I click a search term in Google suggest or click the "Google Search" button in the Google Suggest window - only if I dismiss the search suggestions.

Refreshing the page causes Google to load properly.
Comment 1 Alice Liu 2009-06-23 19:50:48 PDT
Created attachment 31766 [details]
patch authored by oliver

Posting Oliver Hunt's patch for this bug.
Comment 2 Alice Liu 2009-06-23 19:51:32 PDT
<rdar://problem/6939593>
Comment 3 Maciej Stachowiak 2009-06-23 20:49:04 PDT
Patch needs a ChangeLog, and ideally some form of test case.
Comment 4 Maciej Stachowiak 2009-06-23 20:50:11 PDT
(I can't tell if the concepts of the patch are correct on the substance.)
Comment 5 Dave Hyatt 2009-07-02 13:51:32 PDT
Comment on attachment 31766 [details]
patch authored by oliver

Why is the setNeedsLayout() necessary? I'd have thought the forceLayout would take care of that.
Comment 6 Oliver Hunt 2009-07-02 16:51:13 PDT
Created attachment 32211 [details]
fix with changelog
Comment 7 Anders Carlsson 2009-07-02 16:51:59 PDT
Comment on attachment 32211 [details]
fix with changelog

r=me
Comment 8 Oliver Hunt 2009-07-02 16:56:41 PDT
Committing to http://svn.webkit.org/repository/webkit/trunk ...
	M	WebKit/win/ChangeLog
	M	WebKit/win/WebCoreSupport/WebFrameLoaderClient.cpp
Committed r45508