by functionaltesting » Wed Jul 01, 2009 3:50 pm
Yes, that is what the example illustrates, but not that it does this
by pretending to BE the non-IE browser.
If you have eValid pretend to be browser A, then the server will
deliver pages to it that are tailored for browser A (assuming it LIKES
browser A, which in our example it does not). We've seen that
99% of the material is identical -- basic HTML -- independent of
browser.
The SaveFullHTML command will let you save the exact HTML delivered
to eValid (pretending to be A). You can compare the delivered HTML
(for A vs. IE or vs. B, etc.).
What we have heard that there is a BETTER philosophy: find and remove
cross-browser dependencies...in this case you engineer the "tricks" out
of the JavaScript so you center in on the common subset of known functionality.
The programmers hate this -- they want to exploit this little thing and
that little thing, but in spite of that at the end of the day
you have a version of your application that does not trip up when
the browser changes identity.
The movement clearly is in the direction of making the server response
abide by the common intersection of the most common browsers, and
having ONE browser (and IE clone; where IE = 85% of the market) confirm
correct operation by spoofing others is a distinct testing advantage.
The advantage is that, except for the SetAgent command (which can
be parameterized) the script is the same as you vary the browser
type...
eValid Tech Support Team