v3.9.0

• Allow missing hrefs: #475

v3.7.0

• Rubocop! #431
• Security update for Nokogiri: #414
• Handle bases without href: #420
• Support top links: #421
• SRI/CORS fixes: #412

v3.6.0

• Documentation updates! #396, #403, #404
• Escape quotes in anchors: #401
• Make cache directory configurable: #395

v3.5.0

• Lots of README improvements: #372, #375, #383, #384, #391
• Option to enforce SRI/CORS in link and script: #388
• Skip dns-prefetch links: #390
• Let the data-proofer-ignore attribute trickle down: #392
• Remove excess slashes: #393
• Fix URL normalization problems: #394

v3.4.0

• Ignore missing alt text for aria-hidden attributes
• Ensure that hash tags on other pages are still referenced

Both came from #368.

v3.3.0

• Support OpenGraph parsing: #359

v3.2.0

• Support the base tag: #358

v3.1.0

• Improvements to CLI link parsing: #320
• Use regex for colon escape: #350
• Support newer ActiveSupport: #355

v3.0.0

A massive release! This version changes many of the uglier, inconsistent aspects of HTMLProofer, while packing it with many more features.

• #276 - HTML 5 documents do not need href for anchors
• #288 - href_ignore and href_swap are now the more generic url_ignore and url_swap, applicable to as and scripts.
• #291 - Rename binary to htmlproofer and the library can now be accessed as require 'html-proofer', which is much more consistent Ruby.
• #292 - Allows you to ignore specific HTTP status codes
• #296 - Add new file, directory, directories, and links methods for running HTML-Proofer
• #300 - check_html validations are all now opt-in, since Nokogiri reports a lot of bogus cruft
• #301 - Verbosity and log levels are now working for the CLI in a sane manner
• #302 - The cache directory is no longer created every time
• #305 - Caching handles hash links better
• #307 - Namespaces can be ignored in check_html validations
• #311 - Allow extensionless URLs. Thanks @Floppy!

v2.6.1

• Fix an issue with HTML validation: #281