Re: [xml] xmllint schema validation reports line numbers above 65535 incorrectly (was: Re: Important: possible incompatible changes ahead for 2.9.0 !)

[Daniel Veillard <veillard redhat com>]

On Fri, Aug 10, 2012 at 09:24:56PM +0200, Johan Corveleyn wrote:
> On Fri, Aug 10, 2012 at 11:27 AM, Daniel Veillard <veillard redhat com> wrote:
> > On Fri, Aug 10, 2012 at 10:03:39AM +0200, Johan Corveleyn wrote:
> > > On Fri, Aug 10, 2012 at 9:43 AM, Daniel Veillard <veillard redhat com> wrote:
> > [...]
> > > >  That said there are ways to get "long" line numbers stored in adjacent
> > > > nodes c.f. xmlGetLineNo()
> > > >   But not knowing how you actually use the parser nor how you
> > > > use the schemas validation, I cant tell if you can extract full line
> > > > numbers. It should be possible in the xmlReader for example.
> > > > But I don't know how you validate, maybe this could be added, but
> > > > I have no details from you to reproduce your specific situation so ...
> > >
> > > Thanks for the information.
> > >
> > > The problem I'm seeing is with schema validation of xmllint. Schema
> > > validation errors happening after line 65535 are always reported by
> > > xmllint as line 65535. If you'd like I can draw up a quick example, or
> > > write a script to generate such a file and then reproduce it.
> > >
> > > So maybe that particular problem is only in xmllint's usage of the
> > > libraries (when doing schema validation), and maybe that can be fixed
> > > there by accessing those long line numbers?
> >
> > an instance would compress very well I assume :) send that, the schemas
> > and show me how you're running xmllint. XSD can be validated though a
> > normal tree, the reader or SAX. there might be ways to get extended line
> > informations in all case in a controlled environment like xmllint.
>
> Ok, here is an example in attachment (it compressed very well indeed :-)).
>
> The xml file has a schema validation error at line 72720 (level1
> element without level2 child).
>
> When I run:
>
>     xmllint --noout --schema test_schema.xsd test_xml.xml
>
> I get:
>
>     test_xml.xml:65535: element level1: Schemas validity error :
> Element 'level1': Missing child element(s). Expected is ( level2 ).
>     test_xml.xml fails to validate

  Sigh ... okay, fixed for xmllint with tree support:

thinkpad:~/XML -> xmllint --noout --schema test_schema.xsd test_xml.xml
test_xml.xml:72721: element level1: Schemas validity error : Element
'level1': Missing child element(s). Expected is ( level2 ).
test_xml.xml fails to validate
thinkpad:~/XML ->

  see the commit for details:

  http://git.gnome.org/browse/libxml2/commit/?id=968a03a2e54f5bcf53089f5e3c8f790dbe0bf824

but that doesn't fix it for the reader or SAX mode, I would have
to investigate further

thinkpad:~/XML -> xmllint --stream --noout --schema test_schema.xsd
test_xml.xml
Schemas validity error : Element 'level1': Missing child element(s).
Expected is ( level2 ).
test_xml.xml fails to validate
thinkpad:~/XML -> xmllint --sax --noout --schema test_schema.xsd
test_xml.xmlSchemas validity error : Element 'level1': Missing child
element(s). Expected is ( level2 ).
test_xml.xml fails to validate
thinkpad:~/XML ->

> Interestingly, the same problem is not present for parser errors. If
> there is a parser error around that same line, the line number is
> reported correctly.

  yeah when parsing the parser context is given back to __xmlRaiseError
and it extracts an exact line number

Daniel

-- 
Daniel Veillard      | libxml Gnome XML XSLT toolkit  http://xmlsoft.org/
daniel veillard com  | Rpmfind RPM search engine http://rpmfind.net/
http://veillard.com/ | virtualization library  http://libvirt.org/