Bug 12368
Summary: | Amazon.com "click to see price" window partially left behind when closed | ||
---|---|---|---|
Product: | WebKit | Reporter: | Alice Liu <alice.barraclough> |
Component: | Evangelism | Assignee: | Nobody <webkit-unassigned> |
Status: | RESOLVED FIXED | ||
Severity: | Normal | CC: | astrange, bdakin, ddkilzer, kday, koivisto, markmalone, webkit |
Priority: | P2 | Keywords: | HasReduction, InRadar |
Version: | 420+ | ||
Hardware: | Mac | ||
OS: | OS X 10.4 |
Alice Liu
Go to an amazon.com product page where you have to click a link to see the sale price. (right now, you can search for s4695d and find a samsung tv that will do the job, but that won't be a valid statement forever)
click on the link that says "click to see the price" and a window will popup, and press the window's close button. Not all of it goes away.
The range of regression is between r16987(works) and r16994(broken). It is most certainly changeset 16988 since its the only one in the range pertaining to rendering. That fix was for bug 4377.
<rdar://problem/4887255>
Attachments | ||
---|---|---|
Add attachment proposed patch, testcase, etc. |
Alice Liu
When i said "most certainly" I meant "almost certainly". I haven't investigated at a finer resolution at this time.
David Kilzer (:ddkilzer)
(In reply to comment #0)
> Go to an amazon.com product page where you have to click a link to see the sale
> price. (right now, you can search for s4695d and find a samsung tv that will
> do the job, but that won't be a valid statement forever)
Could you post a review of the TV on this bug if you decide to buy one? ;)
ToT crashes when opening amazon.com's home page (Bug 12369).
Antti Koivisto
This is not a bug. Amazon.com actually relies on broken behavior that was fixed in bug 4377. Specifically tiger webkit would not paint any children of a layer with visibility:none, even if visible. The script
http://g-images.amazon.com/images/G/01/nav2/gamma/n2CoreLibs/n2CoreLibs-simplePopover-41389.js
has the following check in hide function (and similar for show)
if (goN2U.bIsSafari) {
goN2U.hide(this.myID);
} else {
goN2U.undisplay(this.myID);
}
goN2U.hide() sets visibility:hidden for the top level abs positioned div, but that does not anymore hide the children, leaving some user visible carbage. Removing the Safari check and using generic undisplay() path (which sets display:none) here makes the popup hiding work as expected (seemingly even with tiger webkit).
David Kilzer (:ddkilzer)
Per Comment #3, this is now an evangelism issue.
Mark, do you have a contact at Amazon.com?
Mark Malone
Perhaps Todd can lend a hand...
Kevin McCullough
Radar: 4887255
Ken Day
My name is Ken Day. I'm a development engineer on Amazon's JavaScript Framework. This bug report was brought to our attention.
Is this bug reproducable in a released Safari/WebKit combination?
-- Ken Day
kday@amazon.com
Beth Dakin
Hi Ken! Thanks for giving this bug some attention! The bug is not present in any released version of Safari/WebKit, but it exists in our Nightly builds. You can download our Nightlies at nightly.webkit.org
Alexander Strange
This bug is visible on Safari 3 (see bug #14427).
David Kilzer (:ddkilzer)
*** Bug 14427 has been marked as a duplicate of this bug. ***
Robert Blaut
I'm unable to reproduce the described problem. I think the amazon.com was redesigned since time of reporting this bug.
Alexander Strange
I can't find an Amazon page using "click to see price" but the problem still exists (as above).
David Kilzer (:ddkilzer)
(In reply to comment #12)
> I can't find an Amazon page using "click to see price" but the problem still
> exists (as above).
Searching for "Sony XBR" returns many hits that have "Click to see price" links. For example:
http://www.amazon.com/Sony-Bravia-KDL-52XBR6-52-Inch-1080p/dp/B001F7HJL0/ref=pd_bbs_sr_1?ie=UTF8&s=electronics&qid=1230121778&sr=8-1
I don't see this issue anymore with Safari 3.2.1 on Mac OS X 10.4.11. It seems that Amazon has fixed the issue (or perhaps that JavaScript code doesn't run anymore for recent Safari versions).
Alexander Strange
That works; I guess #14427 isn't a duplicate anymore, then.