Bug 6389 – REGRESSION: Problem loading porland.citysearch.com home page

• Status: RESOLVED
• Resolution: DUPLICATE
• Component: Layout and Rendering

[Shawn Smith]

Going to http://portland.citysearch.com shows the header then it begins to flash as what looks like a 
rendering loop starts.  The page never seems to complete rendering.

[Shawn Smith]

Still broken in the January 6th Nightly.

[Joost de Valk (AlthA)]

Confirmed, this one needs reduction.

[David Kilzer (:ddkilzer)]

This is definitely NOT a rendering loop.  After the page is initially loaded, a new URL is then continually reloaded via JavaScript: http://portland.citysearch.com/1

Furthermore, I tried loading the page using Firefox, saving it as "web page, complete" to disk, then loading the copy on disk in Safari.  Guess what?  No weird reload bug, although I did see this in the JavaScript console:

Undefined value
file:///tmp/bug-6389/portland-citysearch-com_files/menu_builder_yp.js
Line 21

This smells like a browser sniffing bug to me.

[David Kilzer (:ddkilzer)]

Actually, this is a regression, too.  Works for me in Safari 2.0.3 (417.8) on 10.4.4.

Adding Regression keyword, bumping to priority P1, updating summary.

[David Kilzer (:ddkilzer)]

Interesting.  If you stop the insane reloads and bring up the Web Inspector, the very last element that the inspector shows is the following:

        <script language="JavaScript"><!--
          insertLocationsOptions('savedLocs');
          insertLocationsOptions('recLocs');
        //--></script>

This JS function is defined in the menu_builder_yp.js file, and the function is using a local variable named "location".  I wonder if the scope of this variable isn't being obeyed and the window.location variable is getting written to?  That would explain the reloads, and why the page only renders so far each time.

What I don't understand is why the page loads successfully if it's loaded from disk?

[Alexey Proskuryakov]

Woks for me in ToT. Fixed as bug 5210, perhaps?

[David Kilzer (:ddkilzer)]

I *knew* that JavaScript variable bug sounded familiar.  Thanks, Alexey!

Works for me as of ToT r12574.  Marking RESOLVED/DUPLICATE of Bug 5210.


*** This bug has been marked as a duplicate of 5210 ***

[David Kilzer (:ddkilzer)]

I just reverted my Subversion tree back to r12564 and rebuilt to confirm that the fix to Bug 5210 really is what fixed this bug.