On bash I do

curl theWebPage.html | java -jar $HOME/tagsoup-1.2.1.jar --nons  | java
-jar $HOME/saxon9he.jar -s:- -xsl:yourXSLTFile.xsl

which pipes the web page under test into tagsoup which converts it to well
formed XML which I then pipe into an XSL transformation.

I don't bother with things like Selenium for exactly the reasons you are
complaining about but of course your team may not buy into that.


On Mon, Aug 29, 2016 at 7:25 PM, Hank Ratzesberger xml@xxxxxxxxxxxx <
xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:

> Hi XSL List,
>
> I am hoping to improve our test automation built on Selenium. The xpath to
> elements in our tests is complicated. Any changes break the workflow and
> fixing the xpath is manual process and slow.
>
> If, in the process of running a test, if the web page was scrapped and put
> into an xml file, or even a text file, with xpath to all inputs and other
> controls, differences could be reported, and those differences might even
> be able to be cut and pasted to fix the test in the next update.
>
> In any case, processing this way could rationalize / normalize the xpath
> to all controls.  This way, developers don't have to keep deciphering when
> pages change.
>
> Has anyone here seen something like this? It would seem to be something
> xslt was made for.
>
> Best,
> Hank
>
> --
> Hank Ratzesberger
> XMLWerks.com
> XSL-List info and archive <http://www.mulberrytech.com/xsl/xsl-list>
> EasyUnsubscribe <-list/601651> (by
> email <>)