Re: [xml] Enhancement request: xmllint verbosity control

[Michael Smith <smith xml-doc org>]

Daniel, you wrote:

> On Mon, Nov 19, 2001 at 02:31:26PM +0900, Michael Smith wrote:
> > ... it'd be nice perhaps to have the option of emitting less
> > verbose output (that is, without the content models) similar to what
> > nsgmls emits:
> >
> >   $ nsgmls -wxml -s c:/cygwin/usr/local/xml/xml/xml.dcl xmllinttest.xml
> >   nsgmls:c:/cygwin/usr/local/xml/xml/xml.dcl:1:W: SGML declaration was not implied
> >   nsgmls:xmllinttest.xml:9:11:E: end tag for "chapter" which is not finished
> >   $
>
>   Not printing the full content model is easy. Saying an element is not
> finished is a very different kind of analysis. The validator check if
> the content model matches the DTD. But it doesn't try to make 'guesses'
> about about the reason, i.e. it doesn't try to anticipate over the
> author intent.
>   If you are looking for the first point, well I could do this. If you are
> looking for the second, I can't without spending a lot of time (I don't
> currently have) on this issue. Can you clarify ?

Just the option of not having xmllint emit the full content models for
each error would be a big help in troubleshooting, so it'd be nice for
now to have the capability.

As far as level of specificity in the validity error messages, I can
see it's a lot more work to implement, and perhaps nsgmls validity
message are more specific than most parsers. I used the example from
ngslms mostly just to show how it puts each message on a single line.

Regards,

  --Mike