The fact that w3c doesn't support off-line validation of html shows how hard it is to write standards compliant code. Not EVERY html document needs to be publicly accessible, and there are many documents that should NEVER be uploaded to a server.
Conversation
When CLI services are abandoned in preference for web services, especially when it's by what is effectively a non-profit. Like how has access to the data that is logged for research?
w3.org/Consortium/Leg
isn't clear what data is logged by the validators.
1
1
1
So, the information to d/l the checker is here: validator.w3.org/nu/about.html
Thanks to for urging me to look further.
1
2
Quote Tweet
Replying to @encthenet and @DarkainMX
Actually, no, their web checker isn't "available". They haven't done a release recently.
Web site:
Version: 21.4.17
Latest release:
github.com/validator/vali
20.6.30.
Even the latest tag is 21.4.9 which isn't up to date.
1
And w/ the latest release (which this is against) being almost a year out of date, no point in reporting the bugs.
Quote Tweet
Replying to @TychoTithonus
Managed to the the nu validator running. But of course it's out of data, and the results doesn't match the web. Getting this on a CSS file:
error: Element “html” not allowed in this context.
but the web version says the same CSS is 100% fine.
1
Since the --skip-non-XXX and --also-check-XXX don't work, nothing a little find couldn't fix:
java -jar vnu.jar --stdout $(find deploy -name '*.html')
java -jar vnu.jar --css --stdout $(find deploy -name '*.css')
java -jar vnu.jar --svg --stdout $(find deploy -name '*.svg')
1
Replying to
It works with CSS using a single command if you do it this way:
github.com/GrapheneOS/gra
It has a lot of overhead to start running so it's worth setting it up to use a single command.
I remember having to figure out how to work around the terrible way the CLI interface works.
3
1
It's really bad at validating SVG and isn't particularly good at validating CSS.
It honestly isn't very good for HTML either but that's partly the fault of the current HTML living standard. It permits nearly everything so there's not much it can actually validate anymore.
1
1
I recommend eslint for JavaScript and stylelint for CSS.
I started using xmllint for HTML and sticking to valid XML syntax in to catch mistakes like typos. End result is the same due to html-minifier. I don't mind needing to write disabled="" instead of disabled to catch errors.
This Tweet is unavailable. Learn more
It would be nice to have an HTML linter where you could configure the code style to disallow omitting closing tags.
I'd want a clear rule for when to omit quotes, etc. too. html-minifier won't output <link rel href=https://example/> but the seemingly ambiguous /> seems valid...
1