Bug 6270 – Some content is rendered twice

• Component: Layout and Rendering
• Status: RESOLVED DUPLICATE

[John Gruber]

This is a rendering bug I encountered on the front page of Daring Fireball today. Sometimes -- not 
every time but most times, reload a few times if you don't see the problem at first -- the sidebar 
content (nav menu and sponsored by ad) is incorrectly rendered a second time, about 50px or so below 
the first (correct) rendered instance. It's like a double picture.

Note that the repeated content is not repeated in the HTML markup. This is clearing a rendering bug.

The bug is triggered by some combination of the photo of my son in the main column, and the 
"SidebarPromo" <div> in the sidebar. If I take out the photo, the glitch disappears. If I take out the 
SidebarPromo <div>, the glitch disappears.

Further, if I move the SidebarPromo <div> closer to the rest of the sidebar (my byline credit), the glitch 
doesn't appear. If you look at the linked CSS file for the bug demonstration (sidebar-fubar.css) and 
change the margin-top: on line 488 from 7em to 3em (or less), the bug disappears.

Lastly, when trying to reproduce this bug with recent nightly builds (using WebKit.app), Safari crashes 
when it loads this page.

[John Gruber]

Created attachment 5339 [details]
Screenshot demonstrating the rendering glitch in this bug report.

[John Gruber]

Created attachment 5340 [details]
Crash log for nightly build rendering test page.

[Eric Seidel (no email)]

There was, unfortunately at least one new crasher in last night's nightly:
http://bugzilla.opendarwin.org/show_bug.cgi?id=6269
Please accept our appologies.

This definitely looks like a bug.  However, most helpful to us would be a reduced test case (demonstrating 
the failure) and attached to the bug.  That will probably be easiest to do after we provide a new nightly 
which doesn't crash. :)

[Geoffrey Garen]

After landing the fix for 6269, I can't reproduce the crash. A tip of tree build, now that it doesn't crash, 
renders this page correctly. So I think this bug is fixed.

[Eric Seidel (no email)]

This bug was not about the crash, as I read it.  Reopening.

[Eric Seidel (no email)]

I think this bug is a dup of:
http://bugzilla.opendarwin.org/show_bug.cgi?id=3509

[Eric Seidel (no email)]

ggaren and I talked over IRC.  This doesn't seem to reproduce on tip of tree.  Please try again with the 
latest nightly (this next one shouldn't crash!)

[John Gruber]

When I load the test page with the current (at this writing) WebKit.app nightly (Wed Dec 28 13:02:30 GMT 
2005), I get a reproducible crash after reloading the page. Load once, works (with no repainting display 
glitch), but reload, and it crashes. I'm not sure if this build is TOT or not.

If I go back to a build from two weeks ago (WebKit-CVS-2005-12-12 20-20-08 GMT), the problem goes 
away. No repainting glitch, no crasher.

[John Gruber]

I can confirm that with the nightly build dated "Thu Dec 29 0:34:27 GMT 2005", this is fixed: no repainting 
glitch, no crash. Thanks!

[John Gruber]

Wait. I spoke too soon.

Using the current nightly (Thu Dec 29 0:34:27 GMT 2005), if I load my test page (http://daringfireball.net/
misc/2005/12/sidebar-fubar.html), the glitch doesn't appear. But if I leave that test page open in a tab 
and switch to another tab in the same browser window, then switch back to the test page tab, the 
repainted sidebar content appears. I.e. the problem now only appears when switching to a tab with the 
test page, not when the test page first appears.

[Alexey Proskuryakov]

I couldn't reproduce the problem via switching tabs on ToT, but could reproduce with different steps:
1) opened http://daringfireball.net/misc/2005/12/sidebar-fubar.html in a new window
2) opened http://daringfireball.net in a second tab
3) switched tabs back and forth
4) clicked on COLOPHON link

The results looked similar to the screenshot. Didn't reproduce on second attempt, though.

[Joost de Valk (AlthA)]

I can confirm this in latest ToT. Steps to reproduce: load page, reload a couple of times and scroll down. 
You get the exact same result als in screenshot.

[Daniele Metilli]

Works for me with r13244.

[Alexey Proskuryakov]

Still happens for me (r13236).

[Ingo Chao]

Created attachment 7026 [details]
Reduction

Reduction. Steps to reproduce (WebKit-SVN-r13261): Hit refresh rapidly (maybe 10-20x or more). Sometimes, the image will not appear. The next refresh brings it back, but the red bordered float will be duplicated. 

The image has no dimensions applied; the red float is shorter than the image. The sidebar and the footer establish enough margin to hold the duplication.

[Ingo Chao]

Created attachment 7028 [details]
Screenshot of the reduction

Screenshot of the reduction

Based on the reduction, I think this bug is very likely the same as bug 6795.

Confirmed that this and bug 6795 are the same. Marking this earlier report as the duplicate, since the other one shows some progress towards a fix.

*** This bug has been marked as a duplicate of 6795 ***