WWW::Mechanize / javascript pointer

[spstansbury]

I've got a Mechanize script that logs on to a website, navigates through a couple of pages and prepares a report.

The resulting report is listed in a table with a download button in the last cell. Using WWW::Mechanize::Shell -> links shows the download button as a link, but executing an open on that link doesn't cause anything to happen.

The relevant code snippet is:

<a href="MySurvey_SummaryDownload.aspx?sm= yUsTZIyAVJQqSsTu3TSbdT0ZMzsfbk0cbSdAOVkNUDY%3d" class="itBtn">

Running HTTP::Recorder only shows:$agent->follow_link(n => '25');for this step.

Wireshark shows a:GET http://www.surveymonkey.com/MySurvey_SummaryDownload.aspx?sm=yUsTZIyAVJQqSsTu3TSbdT0ZMzsfbk0cbSdAOVkNUDY%3d

What's the syntax for getting Mechanize to parse the "sm" so I can construct this GET?

As always, thanks for you input!

[Corion]

Why would you want to "parse the sm"?

The link you showed and the resulting URL are identical unless the whitespace in the link you showed actually exists:

<a href="MySurvey_SummaryDownload.aspx?sm= yUsTZIyAVJQqSsTu3TSbdT0ZMzs
+fbk0cbSdAOVkNUDY%3d" class="itBtn">
        /MySurvey_SummaryDownload.aspx?sm=yUsTZIyAVJQqSsTu3TSbdT0ZMzsf
+bk0cbSdAOVkNUDY%3d

Anyway, the general rule of thumb when automating things is to compare what your script does against what the reference does. So, compare what Wireshark gets from your browser against what it gets when you use WWW::Mechanize. Also test the site with Javascript disabled.

[spstansbury]

Thanks for the response.

I want the ability to parse the "sm=" because it is generated for each report and since this is javascript, I don't have a button to click, so therefore I need to parse the report identifier and build a GET statement using it.

[Corion]

If the site uses Javascript, you'll have to parse it, even though your HTML looks good enough. But personally, I'd use my module WWW::Mechanize::FireFox, which understands Javascript by virtue of automating FireFox.