After initial connection: "cancelled"

I am finding that since early this morning (Jan 25th) WebPageTest has been behaving a little differently that usual.

Firstly, When the test starts I still get the brown rectangle telling me that my test is waiting to start instead of the yellow rectangle telling me that the test is under way.

Secondly, also since this morning WebPageTest has been unable to download images from my site generated by phpThumb during testing:

[url=http://www.webpagetest.org/result/110125_E5_1KEA/1/details/#request19]http://www.webpagetest.org/result/110125_E5_1KEA/1/details/#request19[/url]

I can download the images OK in my browser and I cannot find any other tools (pingdom, zoompf and others) that have the same problem.

I’m not saying that these two issues are connected, only that I noticed both changes in behaviour at the same time.

The only change I have made the the site under test recently was some lines in .htaccess allowing me to refer to the phpthumb images using urls without a querystring and this worked perfectly well for about 48 hours before the problem in WebPageTest showed up.

Any help appreciated.

Thanks

Have you tried loading the page in IE? I just tried IE8 on my desk and got a blank page (along with javascript errors).

btw, the error I’m getting is:

‘_mF’ is undefined main.js, line 3 character 220

and the main.js in this case is for Google maps. Did you change anything about the code and how you integrate with Google maps?

Yep, something is very broken - this is what I get when I “view source” on the page:

You trying to inject the js to load asynchrounously? Looks like it is overwriting the page contents.

Never seen that before!

I’ve removed ‘defer=“defer”’ from the initial google maps script which may have resulted in scripts executing in the wrong order under some circumstances.

Thanks for that alert.

The phpThumb-generated images seem to be loading OK now (But I don’t think it’s because of anything I did!).

I expect they were getting canceled because of the DOM changing when the script replaced the html body (the browser had started to request the images but no longer needed them). When you fixed the script problem it also addressed the canceled requests for the images.

-Pat

That all sounds logical — thanks for all your help.

-Steve