W3C home > Mailing lists > Public > www-validator@w3.org > April 2007

validator.w3.org / xml.resource.org

From: Frank Ellermann <nobody@xyzzy.claranet.de>
Date: Tue, 24 Apr 2007 19:37:19 +0200
To: www-validator@w3.org
Message-ID: <462E404F.56C4@xyzzy.claranet.de>
Cc: xml2rfc@lists.xml.resource.org

Hi, I tried to use validator-test.w3.org with an xml2rfc source,
and got tons and tons of errors, so I guess that's an issue or
feature of this test version.

To check what's going on I then tested validator.w3.org, and I
*think* (could be wrong) that this used to report no errors.

Today I get various errors, some of them in the style 

| Error Line 547 column 8: error connecting to 
| "xml.resource.org" (Connection refused).

That's for the tenth external entity &rfc3977; in the validated
document, is there some limit how many entities the validator
can fetch from xml.resource.org, or how fast it should do that ?

The other errors with http://validator.w3.org are in the style 

| Warning Line 239 column 9: cannot generate system identifier
| for general entity "nbsp".

Maybe the 2 *.ent files with the entities are not more available
where they used to be available (?), nbsp was defined in file

Some errors are rather cryptic:

| Line 762, Column 45: reference to non-existent ID "RFC3977". 

That's an <xref target="RFC3977" />, the DTD says it's an IDREF,
the validator never got the external entity &rfc3977;, and so it
reports a missing ID for the IDREF.  

I think the new validator could abort all efforts to validate a
document, if anything with fetching external documents doesn't 
work.  It should report what went wrong with fetching external
documents as good as possible.  It should not say that a source
is "invalid" only because it was unable to fetch the required
external documents, IMO that's "inconclusive", not "invalid".

BTW, we're not talking about hundreds of external documents in
this example, it's one DTD, two *.ent files, and about twenty
"bibxml" snippets.  An approach like "fetch at most one file
per server and second" would be fine, if that doesn't trigger
the xml.resource.org limit.

Received on Tuesday, 24 April 2007 17:40:13 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:59:01 UTC