Bug fixes

• a11ym-sniffers does not strictly expect the node_modules to be in the same directory. A new findNodeModules function is now used to look for the closest appropriated node_modules directory.

Features

• Support asynchronous reporters (if it returns a promise).

Features

• Report can have options, with --report-options (CLI) or reportOptions (configuration file).
• Can use report by path, i.e. you can create your own reporter and use them by specifying: --report path/to/my/report.js (CLI, same with the configuration file by using report: path/to/my/report.js).

Bug fixes

• Be less strict regarding types since the introduction of the configuration file (null, undefined, empty strings, are now all treated equally).

Features

Global

• a11ym can be used as a library,
• const a11ym = require('the-a11y-machine') will just work,
• Can use a bootstrap/configuration file, see a11ym.yml.example,
• Filter by error level with the --error-level CLI option, or the errorLevel configuration entry,
• PhantomJS is now a dependency of the project, no need to install it before.

Crawler

• The crawler handles most redirections automatically.

Dashboard

• Can be used as an HTTP server, or can generate static files with the --static-outout CLI option,
• Add the --open CLI option to directly open the fresh report in your favorite browser.

Testers

• Default viewport is now 1280x720 (likely to serve a desktop version),
• Ensure no proxy is used in some cases,
• VNU has been updated to 16.6.29, will fix several HTML test issues, for instance with TLS support.

Bug fixes

Report

• The default report output directory a11ym_output is created in the current working directory (CWD) instead of the root of the project.

Dashboard

• Fix header line-height.

Bug fix

• Fix the dependencies:
  • pa11y has merged our PR (https://github.com/pa11y/pa11y/pull/124), remove our fork,
  • HTML_CodeSniffer has merged our PR (https://github.com/squizlabs/HTML_CodeSniffer/pull/151), remove our fork but keep the master,
  • simplecrawler 1.* serie has been released, stick on it, but the API has change, fix it.

Bug fix

• Fix the progression maximum number.

Bug fix

• Fix a bug when creating the report directory (it's now synchronous).

Features

• Even a better exploration algorithm: Move the algorithm in the core of the crawler, get even better results ealier,
• HTML validation: We now use The Nu Html Checker to check HTML validation, in addition to the accessibility validation. Simply add the HTML in the --standards option, like --standards WCAG2AA,HTML to enable HTML validation:

All validations are still local, nothing sent over the network.

Bug fix

• Fix a regression in the exit code,
• Fix an error in an error message.

Bug fixes

• Be consistent with the new pa11y naming (working on our own fork).

Features

• New algorithm to explore resources: Provide more interesting results earlier,
• Tests happen in parallel: Introduce the -w/--workers options to control the number of workers,
• New a11ym-sniffers command to build sniffers, bye bye Grunt,
• New HTTP authentication: Introduce the --http-auth-user and --http-auth-password options to authenticate HTTP requests,
• New error and 404 logs from the crawler,
• Better logs with nyancat full of colours,
• New -V/--no-verbose to control the log verbosity,
• Add <base /> tag in the index.html report: Useful when a sysadmin would like to move this index.

Bug fixes

• Ctrl-C now works much better,
• Avoid using the crawler when --maximum-urls is set to 1,
• role="presentation" is partially supported (please, see https://github.com/squizlabs/HTML_CodeSniffer/issues/149),
• A11ym now uses a specific user-agent: liip/a11ym, it avoids being blocked by some firewalls.