Internally ignores /cdn-cgi/ urls, which look like bad links but are a Cloudflare 'feature' related to hiding email addresses. Improves help dialog for cases where the scan stalls early.

Splits 'flag missing link url' into two separate options - flag href="#" and flag href=""

- Fixes some issues with 'archive during scan', particularly with regard to scanning a list of links. - Fixes issue with 'include pdfs in Sitemap' option

Now treats an invalid certificate as a bad link. A new option "accept invalid certificate and continue" on the Advanced tab may be checked if you're testing a website that doesn't yet have a certificate, or which you know is expired. Now capable of finding lazy-loaded images or videos whose url exists within a data-src attribute of a div tag. The 'check for lazy load images' option must be on.

Fixes failure to launch on older systems (previous version should have run on 10.14-10.15)

new build of 12.11.1, very minor update

Adds ability to add a rule which can whitelist an additional domain. This is useful for testing links within pdf/Word documents (Integrity Pro feature) which are hosted externally, or for including a subsite which happens to be on another domain. Improves handling of urls enclosed in smart quotes. Previously these were marked as 'unable to form url' because the smart quotes weren't being trimmed and therefore being treated as part of the url. Such urls are now more helpfully reported as 'Smart quotes detected' and marked bad, although the urls themselves may be OK

Fixes bug that could happen under unlikely circumstances. If starting url redirects to another domain, and that page doesn't have a base href, then relative urls on it could be constructed using the original url rather than the redirect one (and therefore probably reported as broken). This wouldn't have caused a problem if any other internal page redirected to an external one, because it would have then been treated as an external url and only tested, not crawled. It's recommended anyway that if you know that your starting url redirects to a different domain, that you use the target url, because otherwise it leaves ambiguity as to what is internal and what is external.

- Allows customisation of coloured row backgrounds in html exports. Uses the styles in Preferences>Reports>Styles. For existing users who have already made changes there, it'll be necessary to add the styles for .badbg and .warningbg - Looks for role=presentation in image tags, if present then the warning for no alt text is suppressed. - Fixes issue with encoded characters in urls as discovered on the page - It wasn't always ignoring "data:image..." but is now - removes jsessionid from a url as this can cause an infinite loop

- Previously failed to parse a page if the page contained zero hyperlinks. Now will parse a page with zero hyperlinks if it has at least one image on it - improves the parsing of an image srcset where unencoded commas are present within urls - adds a warning if an srcset contains width descriptor(s) as well as pixel density descriptor(s).

Fixes a possible crash when using 'highlight' feature from within Link Inspector. Updates the list of user-agent strings so that they're current versions.

Important fix for all users. Using 'Limit crawl based on robots.txt' could cause the scan to stall if the user is blocking AI training bots. This is because Integrity was misinterpreting "User-agent: Google-Extended" (Bard) as one of the web indexing Google bots.

- Fixes a problem with the link inspector not displaying all instances of the link url if it appears multiple times on a particular page - Now generates a warning if a link that appears to be internal (from its domain) but is marked rel=external. This may be deliberate and is legal but there are serious SEO implications if it happens unintentionally. - Despite collecting the data and possibly generating a warning, Integrity does not *observe* the rel=external at this point. ie if it is internal according to its domain, integrity does not mark it as an external link or treat it as external, ie not following it. This does raise some questions and the matter is left open at this point. - Now observes rel=nofollow in a link, if at the point the engine comes to follow that link, *all instances discovered so far* are marked 'rel=nofollow'. In other words if an instance has been discovered at that point which doesn't have the nofollow keyword in its rel, then the link will be followed.

Fixes a possible issue with character encoding

- More clearly marks image tags with empty source. Previously 'flag missing link url' would also pick up img tags with missing src, if image check was switched on, but would confusingly mark them with the status "missing link url". Now the images are given the status "missing image src" so that they can be distinguished from missing link urls in the By Status and other views. - Fixes a bug that caused 'urls' containing a bunch of javascript to appear as bad links under very specific and unlikely circumstances.

Adds Expand All and Collapse All to View menu with keyboard shortcuts, and buttons for those functions to the toolbar palette. These will expand / collapse all items in the current view, if the view is expandable, eg by page, by status, Spelling by word. Fixes bug that may have prevented pdfs from correctly being included in the sitemap if the preference was checked or caused them to be included even if the preference was unchecked. Now this behaves as expected according to the preference.

- Two settings were not being saved. Rules / Consider http and https versions of the same url to be the same and Rules / Consider http pages external - When scanning locally (file://) and with the 'test anchors' setting switched on, could go into a loop.

Fixes the issues with sorting some of the columns of the 'All links' table

Small fix for a problem that will affect very few users, but will be a very important fix for those users. If the starting url has a path, and that path contains encoded characters (eg %20) and relative links on the page, then the crawl may not have proceeded past the first page.

- Adds 'redirect count' as an optional column to the All links table. Sorting this table will find links with the longest redirect chains. This appeared with the SEO results in version 10 and earlier. However, redirect count isn't a property of pages but link urls (different link urls may arrive at the same page via different numbers of redirects.) - At the end of a scan, if user hasn't added a name for the configuration, "A new website" is automatically replaced with the first word or two from the starting url's meta title.

Fixes a problem that could cause repetition or even a loop with certain urls that contain percent-encoded characters Fixes an issue causing some missing anchors to not be reported When testing anchors, if the anchor fragment of a url isn't found on the target page, link is now coloured red in the views and treated as a bad link, rather than being coloured orange as before.

- Reinstates 'Export Image XML Sitemap' which was a feature of v10 but has been unintentionally missing from v12 to date. - Tidies up the image checking functionality internally. Images must be included in the scan if images are to be included in the sitemap xml or an image sitemap, so image checking is forced on if either of these settings are enabled. To disable image checking (for efficiency with certain operations) then all of these settings need to be switched off. - Fixes empty lastmod tags appearing in sitemaps where no date is available. Tags are now correctly left out if no date is available (optional tag). NB Google says, if no date is available, don't insert date of sitemap generation

- Some fixes to the 'check anchors' functionality. Note that the status will be the http status regardless of whether the anchor is good or bad, ie '200 no error' even if the anchor doesn't appear on the target page. In the case of a problem with an anchor, a warning is created on that url which can be highlighted in orange depending on your Preferences, or shown in the warnings table in the case of Integrity Pro and Scrutiny. - Fixes the option Preferences>Views>Display labels>'Display labels for redirected URLs'. - An efficiency improvement which may help with a slowing-down problem experienced by some users and will generally make the app more efficient for medium to large sites.

- Fixes double-quotes in url (even if percent-encoded) corrupting the exported bad links (by link) csv - Fixes crash which happened under unlikely circumstances  - Fixes a bug that could have caused spurious statuses for certain urls (caused if url contained a percent symbol which isn't part of a percent-encoding, which is a bad practice anyway.) - Fixes some urls with unlikely character combinations in the querystring or path to misleadingly display in tables as the domain only - Some social and calendar 'add this' links are now listed but not checked. It's potentially not a good thing to request these urls to check them. The ones in question are major domains (eg Google, Outlook) and tend to be generated by plugins, and so more likely to give a false positive when testing than to actually be broken.

Better handling of html5 picture tag Preferences for Sitemap (whether to include unique image urls, audio, video, pdf) are now all switched on by default