E-mail List Archives
Thread: Another reason we can't rely on automated scans
Number of posts in this thread: 1 (In chronological order)
From: Jeremy Echols
Date: Fri, Jun 14 2019 3:37PM
Subject: Another reason we can't rely on automated scans
No previous message | No next message
I don't know if anybody else has run into this situation, but we're branding a product that generates the HTML through a terrifying combination of code and XSL transformation. One way in which it holds search form state is via hidden fields. By default, very few form fields are actually given labels, and this has required some tremendously painful hacks to both their core Java code and XSL files.
So we're labeling all the fields, and what do we discover? Though automated tools think the form fields have labels, actual use with NVDA and Firefox suggest the fields do *not* have labels. It turns out the hidden fields which hold onto search state are earlier in the DOM and use the fields' ids.
Of course if we required perfect HTML, this could never happen, but this particular product generates so much bad HTML we don't actually pay any attention to HTML errors anymore - we have no choice but to hope for the best when we do any tweaking.
Long story short: automated tools are still very easy to fool even in fairly trivial cases. Test manually!