Closed j-mendez closed 1 year ago
We will check it out. Thanks @j-mendez for bringing it to our attention.
Hey @j-mendez, trying it out and had some comments and questions for you:
Thank you and let me know if should post these questions in https://github.com/a11ywatch/github-actions instead.
Hello @dmundra, at the moment the github action does not allow replacing the url, I looks like a worthy add based on the use case above, atm the crawler will only respect links found under the domain of the url passed in. Seems like a reasonable feature just not sure about timeline for adding it. Personally I start the local instance in dev mode and usually all the links are configured dynamically to http://localhost
.
In order to receive the github comments use the config var LIST
set to true and this will output a pass fail list or set the FAIL_ERROR_COUNT
var.
- uses: a11ywatch/github-action@v1.14.0
with:
WEBSITE_URL: ${{ secrets.WEBSITE_URL }}
SITEMAP: true
FAIL_ERRORS_COUNT: 15
LIST: true
I would only use kayle fully by itself for small to medium websites. Integrating with kayle is quite flexible with any headless runner or even non headless as a linter.
@dmundra If you have any more questions feel free to followup on any of the github repos as well. FYI the github action takes a bit longer to start up for the first time since it is downloading and installing, afterwards it should be a lot quicker so the rewards of the speedy testing coming into play. Thanks for checking out the project as well!
Using pa11y leaves out up to 10-35% issues with htmlcs and axe-core opposed to using a11y-js. The repo provides test and benchmarks showing the missing issues not being covered along with drastic speed issues.
I kept seeing this repo popping up on my GitHub feed and it’s an eye sore to see pa11y being used with better alternatives at hand.
Swapping the ci action with https://github.com/a11ywatch/github-actions should be almost a drop in replacement with minor changes and will provide the results with better accuracy with the ability to finish in milliseconds to seconds vs minutes to several hours.