Bug 64378

Summary:

REGRESSION (r88601): Screen elements don't render in Lord of Ultima

Product:

WebKit

Reporter:

Michael Miller <michaelmiller>

Component:

Layout and Rendering

Assignee:

Nobody <webkit-unassigned>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

aestes, bfulgham, dominicc, inferno, mikelawther, sam, simon.fraser

Priority:

Keywords:

InRadar, NeedsReduction, Regression

Version:

528+ (Nightly build)

Hardware:

Mac (Intel)

OS:

OS X 10.6

URL:

http://lordofultima.com

Attachments:

Description	Flags
How WebKit renders the map	none
How Safari renders the map	none

Michael Miller

Reported 2011-07-12 11:47:31 PDT

Created attachment 100533 [details] How WebKit renders the map Seen in Webkit r90760 as well as the current Canary build as of this write-up, 14.0.818.0. Certain elements in the Lord of Ultima map do not render properly. Black is displayed instead. See the attached screenshots which shows how it's displayed in the current Safari release (5.0.5), and in Canary.

Attachments
How WebKit renders the map (656.22 KB, image/png) 2011-07-12 11:47 PDT, Michael Miller	no flags	Details
How Safari renders the map (1.09 MB, image/png) 2011-07-12 11:48 PDT, Michael Miller	no flags	Details
View All Add attachment proposed patch, testcase, etc.

Michael Miller

Comment 1 2011-07-12 11:48:04 PDT

Created attachment 100535 [details] How Safari renders the map

Dominic Cooney

Comment 2 2011-07-12 19:24:34 PDT

Two things that you could do to make this easier to fix are: 1. Find out which revision range it regressed in. You can download archived nightly builds for Mac at <http://nightly.webkit.org/builds/trunk/mac/1> and Windows at <http://nightly.webkit.org/builds/trunk/win/1> 2. Try to build a HTML page that demonstrates the problem and attach it to this bug.

Alexey Proskuryakov

Comment 3 2011-07-12 21:27:20 PDT

<rdar://problem/9765590>

Michael Miller

Comment 4 2011-07-12 22:25:14 PDT

Dominic, I downloaded r87771 (Jun 1 2011) and the problem does not occur. Over the next couple of days I'll attack from both sides and try to find the date it broke.

Andy Estes

Comment 5 2011-07-13 18:16:24 PDT

I bisected locally. Looks like this is a regression due to <http://trac.webkit.org/changeset/88601>.

Abhishek Arya

Comment 6 2011-07-13 19:16:37 PDT

(In reply to comment #5) > I bisected locally. Looks like this is a regression due to <http://trac.webkit.org/changeset/88601>. Do you have a testcase/reduction. I think it might due to these two line 123, 124 - http://trac.webkit.org/browser/trunk/Source/WebCore/html/HTMLBodyElement.cpp?rev=88601#L123. If you have a debug build, you can try removing these 2 lines to see if it fixes. Without a repro, hard to say anything.

Andy Estes

Comment 7 2011-07-13 19:22:06 PDT

A reduction is being worked on, but I don't have one yet. Let me try your suggestion in a debug build.

Abhishek Arya

Comment 8 2011-07-13 19:32:25 PDT

(In reply to comment #7) > A reduction is being worked on, but I don't have one yet. Let me try your suggestion in a debug build. Thanks a lot Andy.

Andy Estes

Comment 9 2011-07-13 19:33:26 PDT

(In reply to comment #6) > Do you have a testcase/reduction. I think it might due to these two line 123, 124 - http://trac.webkit.org/browser/trunk/Source/WebCore/html/HTMLBodyElement.cpp?rev=88601#L123. No, removing these two lines has no effect.

Andy Estes

Comment 10 2011-07-13 19:33:50 PDT

A coworker of mine at Apple is working on a reduction. I'll post it here when we have it.

Andy Estes

Comment 11 2011-07-19 14:05:40 PDT

Looks like this site has regressed even further due to <http://trac.webkit.org/changeset/91206>. With that change in place the game doesn't even load. This deserves its own bug.

Andy Estes

Comment 12 2011-07-19 14:10:34 PDT

Filed <https://bugs.webkit.org/show_bug.cgi?id=64825> to track this separate issue.

Sam Weinig

Comment 13 2011-07-28 17:12:37 PDT

I reopened https://bugs.webkit.org/show_bug.cgi?id=62230 to roll it out. If a patch is known to cause a regression and is not dealt with in a timely manner, it should be rolled out.

Sam Weinig

Comment 14 2011-07-28 17:42:13 PDT

(In reply to comment #13) > I reopened https://bugs.webkit.org/show_bug.cgi?id=62230 to roll it out. If a patch is known to cause a regression and is not dealt with in a timely manner, it should be rolled out. The offending patch was rolled out in http://trac.webkit.org/changeset/91965. Perhaps we should close this bug since the other bug is now reopened?

Abhishek Arya

Comment 15 2011-08-05 07:02:17 PDT

Repro from http://code.google.com/p/chromium/issues/detail?id=91558. I can take a look early next week, right now i am traveling back from a conference. What steps will reproduce the problem? 1. Run this code and click on the image: <!DOCTYPE html> <html> <body> </body> <script> d = document.body.appendChild(document.createElement("div")) d.style.width = "100px" d.style.height = "100px" d.style.backgroundImage = "url('a.png')" d.onclick = function () { document.body.removeChild (d); d.style.backgroundImage = "url('b.png')"; document.body.appendChild(d); } </script> </html> What is the expected result? The image change to b.png What happens instead? The image is still a.png Please provide any additional information below. Attach a screenshot if possible. The same problem cause display corruptions in the browser game “Lord of Ultima“ and Chrome 13.

Brent Fulgham

Comment 16 2022-02-12 18:27:21 PST

This issue was fixed when r88601 was rolled out.

Note You need to log in before you can comment on or make changes to this bug.