[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
[platform-swt-dev] Announcing new release of SWT Web Automation Framework

My apologies: initial message got sent prematurely. Please find updated version below. Thank you!

Main changes from version below. I actually have _two_ versions on site (being uploaded as we speak).

The stable version was a step back today towards the Javascript-based model, somewhat simplified (uses unique ids
for elements). Relies of sleep statements completely, but works quite well.

The other (unstable) version actually retrieves a full DOM structure for each web page. There are two issues:
1) how to detect when a page has _truly_ loaded in a cross-browser way. I have some ideas with regards to this (setInterval with a function that will check for creation of any new elements whose IDs are not known, say, every 1 second; when no new elements have been created in 1 second, it means we are done)
2) how to detect whether any changes occur after a Javascript statement is executed (perhaps above will work as well?)

In any case, looking much forward to your input.

Highlights:
Stable version - slightly better Javascript version
Unstable version - DOM-based framework, along with tester application (Firebug-like). Unfortunately has some kinks on Windows.

Thank you

Misha Koshelev
---

Dear All:

Please bear with me as I release a new version of the SWT Web Automation Framework at 
www.mkosh.com
now licensed under EPL v1. I would still like to contribute the parts of the framework that are in my
org.eclipse.swt.browser.webdriver
classes back to the SWT project if possible.

Just to remind, this is a cross-platform solution that allows:
* end user web automation applications
* with the ability (that SWT provides) to hide the browser from the user

You will find on the Web site, additionally, a sample application (Facebook Birthday Greeter), as well as a Web Automation Framework
tester.

Notably, the big change from the previous version is that I use NekoHTML to actually parse a DOM structure for the given document. This allows some additional capabilities not present in the prior version, specifically the use of XPath expression querying.

Additionally, it makes the methodology somewhat neater, and, as you can see by the Web Automation Framework Tester tool (which, I have to
say, simulates to some extent the Firebug Firefox Extension - and thank you to Grant Gayed on the SWT forums for all your help.

There is much to be done on the back end side. Roughly, to keep a consistent API, I am following that of WebDriver
http://selenium.googlecode.com/svn/trunk/docs/api/java/index.html
although, as you can see, there are additional features that are not present there.

In any case, I still have much work to do. I have implemented the examples from WebDriver in org.eclipse.swt.browser.webdriver.Test

Additionally, although some keystrokes are still sent incorrectly (e.g., "!"), one really neat thing is that we can even send keystrokes correctly to Internet Explorer with the window _not visible_, which is impossible both using Javascript (I did not find a method to simulate keystrokes that worked well on Internet Explorer) and using WebDriver (as it is impossible to hide the IE window).

In any case, I look forward to your comments/support/encouragement etc.

Thank you!

Misha

p.s. There are some known problems right, especially with detecting mouse move events in the Tester in Windows IE. I will work on fixing this. Thank you!