Beefy Boxes and Bandwidth Generously Provided by pair Networks
Pathologically Eclectic Rubbish Lister
 
PerlMonks  

Control the Web browser from the client-side?

by Maestro_007 (Hermit)
on Mar 07, 2002 at 21:06 UTC ( #150129=perlquestion: print w/replies, xml ) Need Help??

Maestro_007 has asked for the wisdom of the Perl Monks concerning the following question:

I'm trying to figure out if it's possible to write a perl script that opens a browser (Netscape), and can click on links, fill out forms, etc.

My company is deploying a web-based application and I'm on the team responsible for automating the GUI testing. We've looked at Rational and Silk, and one of our guys has fashioned a magic script-driven test harness (using VB of all things!) to control IE. It's pretty cool, but our application supports Netscape too.

I want to find a Perl solution, because that's the Lingua Franca of our department, so I ask the question here: can I control Netscape from the client side using Perl modules that already exist?

I've searched both the Monastery and CPAN, but I can only find things that deal with Mozilla and Netscape from a rather more peripheral view (i.e. manipulating Bookmarks and history lists, etc). I've used LWP and all sorts of HTTP and HTML modules for processing requests and responses, but that's not the problem. The problem is seeing if Netscape will correctly render the output from the application.

I'd supply sample code, but that presupposes that I know whether such a thing is possible!

  • Comment on Control the Web browser from the client-side?

Replies are listed 'Best First'.
Re: Control the Web browser from the client-side?
by perrin (Chancellor) on Mar 07, 2002 at 21:46 UTC
    First, to control Netscape you can use the OLE interface on Win32 and the remote control interface on unix systems.

    However, you probably don't want to do this. The type of testing most people do is along the lines of fetching a page and running a regex against the returnded HTML to check for correctness. There's no need to use Netscape or IE for this, since Perl is perfectly capable of doing it.

    There are dozens of modules for automated web testing on CPAN. You might look at webchatpp, HTTP::WebTest, or HTTP::MonkeyWrench. Lots of others have been announced on the mod_perl list, but I can't remember all of their names now.

      Yeah, I'm afraid for now the mandated solution is to perform all actions through a session of Netscape. I've suggested just parsing the HTML with the various Perl mods out there, and we do that to an extent when testing the actual application, but this task involves specifically testing how things are rendered on the Major Browsers.

      It's the same problem as if we were writing a Java client that called an API. We'd want a way to test that Java client, both through the GUI and through the API directly.

      In our case, however, the browser IS the GUI, and given that, we need to do more than parse the HTML according to how the Perl mods do it, we need to make sure IE and Netscape parse it the way we think they do.

      I'm definitely going to look into OLE to work with Netscape. I wasn't sure it could be used since I saw no direct mention of it here or at CPAN. Ultimately I hope it does just come down to parsing the HTML using the standard Perl modules, but in the meantime I gotta do what the boss says. In any event, I don't think rolling my own regexen are the way to go when parsing HTML.

      Update: added 'rolling my own' after perrin's most astute comments

      thanks!! MM

        In any event, I don't think regexen are the way to go when parsing HTML

        Most of the HTML parsing modules for perl actually do use regexes. However, what I meant was a regex for checking correctness of content, as in /Hello $username/ or /order number: $order_id/.

Re: Control the Web browser from the client-side?
by Ryszard (Priest) on Mar 07, 2002 at 23:31 UTC

    For checking NS renders, I dont see how automating would really help, as you have have a human there to make a subjective decision. Surely automating it would take longer than the QA dept trawling thru' a few web screens (depending on the size of the app).

    If your html engine is centralised (ie all the html code gen passes thru' the one spot), I would be inclinded to take a sample of perhaps 10 different web pages (or a significant sample) and verify those.

    You can then present your UAT doc to mgmt. If mgmt are particularly anal about these things, I reckon including a disclaimer such as "All HTML is produced via a centralised HTML engine."

    If you have a large site, you'll be getting into the land diminishing returns which is obviously just not efficient.

    I personally think you would be after a process template, rather than automating a point and click solution. With the process template, you'd have a bit of doco that has all the tests you require, all you have to do is perform the tests on "the next platform" and put the results in your document (word templates work well for me).

Re: Control the Web browser from the client-side?
by rbc (Curate) on Mar 07, 2002 at 21:50 UTC
    I haven't done much with LWP but I imagine that you could
    have a perl script that used LWP to navigate links to get the
    URL's and then do a system( netscape, $url ) ...

    psuedo code:
    ... for ( how ever long ) { @links = getlinks( $theWebPage ); foreach $link (@link) { system( netscape, $link ); } } ...
    I dunno just a guess.

Log In?
Username:
Password:

What's my password?
Create A New User
Domain Nodelet?
Node Status?
node history
Node Type: perlquestion [id://150129]
Approved by root
help
Chatterbox?
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others musing on the Monastery: (6)
As of 2023-02-02 14:43 GMT
Sections?
Information?
Find Nodes?
Leftovers?
    Voting Booth?
    I prefer not to run the latest version of Perl because:







    Results (19 votes). Check out past polls.

    Notices?