[imps] differing results from our v1.3 validator instance and w3c's

Wed Jan 22 11:27:46 PST 2014

[removing support at webdevtutorials.net whatever that is]

Rancid Iodine <rancid.iodine at gmail.com>, 2014-01-22 12:02 +0000:

> We have some instances of w3c validator v1.3 with the (by now of course
> indispensable :) HTML5 validator.nu servlet on the back end, and we're
> wondering why these give a bunch of identical errors in one line for the
> same thing 12 in 12 different places like so:
> ===
> *The value of an attribute contained something that is not allowed by the
> specified syntax for that type of attribute. For instance, the “selected”
> attribute must be either minimized as “selected” or spelled out in full as
> “selected="selected"”; the variant “selected=""” is not allowed.*
> ===

That message doesn't come from the validator.nu backend. I think that's
instead a message the legacy W3C SGML/perl validator emits.

> I ask this because YOUR instance at http://validator.w3.org/ does not give
> the same errors for exactly the same site/source.

What do you mean "YOUR? That URL's for the W3C validator but you
(cross)posted this message to a WHATWG mailing list...

Anyway, of you're using http://validator.w3.org/ to check a document that
has a <!doctype html> doctype, it'll get checked by an instance of the
validator.nu backend, and you won't see that message you quoted above.

> I'm speculating that when we start the backend validator.nu servlet with
> the usual "python build/build.py --w3cbranding all" we may not be getting
> some relevant updates to recent standards,

No, if you do that you'll actually be getting the latest upstream code from
the validator.nu source repository. But like I said, the message you quoted
above is not coming from the validator.nu backend -- it's coming from the
legacy W3C SGML/perl code.

If you have some version of the W3C SGML/perl validator installed on your
system, I recommend you remove it completely and just use the validator.nu
validator.

> if your validator instance is ignoring the same error. Does that sound
> likely to anyone, or is anyone familiar with the error?

Again I don't whose validator your mean when you say "your validator
instance". http://validator.w3.org/ is behaving as expected. Any code you
installed from "python build/build.py --w3cbranding all" is up to date and
behaving as designed.

  --Mike

-- 
Michael[tm] Smith http://people.w3.org/mike