>I believe you're confusing Glade
with AT-SPI.
Yup - I just realized that. But the nice thing about that is what
I want to do can still be tested under automatic mode and this seems
to be a very valuable thing as a lot of people are creating web based applications
and this is a very valuable tool for doing that.
I have one request of the developers who a working
on AT-SPI tool sniff. At the bottom of the AT-SPI browser (sniff) you list
"Name, Role, Description, Action". Could you add the full
element pathname so that some of us who have a problem remembering our names
can use this information to generate testing sequences and be able to know that we have
the correct element.
>There isn't a direct correspondence
between the GTK/rich-client world
>and the HTML world, so I believe
you won't find this approach to be
>meaningful or useful. (It
might be quite a fun hack though).
Anything that will allow me to create testing scripts
with data control files will work just fine to test any web-based application.
>It sounds like you're talking about
generating AT-SPI information. This
>should be generated by the browser
in response to the DOM tree of the
>web page being viewed. If
you follow standard accessibility guidelines
>in the design of your website,
a browser should be able to give better
>information to AT-SPI (and hence
to Dogtail).
The DOM tree is being created in epiphany but not
as well in firefox so the easy solution is to use epiphany for testing and any other browser for
final verification. Also thank you for the reference to the accessibility guides as my developers have
never seen them before (so they said).
