Re: [xml] How to reset an HTML push parser context?

[Daniel Veillard <veillard redhat com>]

On Tue, Sep 11, 2007 at 01:26:30PM +0200, Stefan Behnel wrote:
> Daniel Veillard wrote:
> > On Mon, Sep 10, 2007 at 09:45:10AM +0200, Stefan Behnel wrote:
> > > Hi,
> > >
> > > there isn't currently an API function for resetting a push parser context for
> > > the HTML parser. However, resetting it for reuse doesn't seem to be trivial.
> > > It looks like I have to run htmlCtxtReset() and then create and set up an
> > > input stream (in a pretty ugly way, according to the Create code...). This
> > > could well motivate an official function.
> > >
> > > I also thought about using the xmlCtxtResetPush function, but then I stumble
> > > over things like the spaceTab setup (which is currently a sure crasher for me).
> > >
> > > Is there anything else I have to do to implement this functionality by hand?
> > > And: is there an easier way?
> >
> >   Honnestly I don't know. I don't see why xmlCtxtResetPush() would not
> > work for an html parser context.
>
> In case others are interested, the code below works for me (Pyrex code, but
> should be readable).
>
> Stefan
>
>
> cdef int _htmlCtxtResetPush(xmlparser.xmlParserCtxt* c_ctxt,
>                             char* c_data, int buffer_len,
>                             char* c_encoding, int parse_options) except -1:
>     # libxml2 crashes if spaceTab is not initialised
>     if _LIBXML_VERSION_INT < 20629 and c_ctxt.spaceTab is NULL:
>         c_ctxt.spaceTab = <int*>tree.xmlMalloc(10 * sizeof(int))
>         if c_ctxt.spaceTab is NULL:
>             python.PyErr_NoMemory()
>         c_ctxt.spaceMax = 10

  xmlCtxtResetPush should instead be fixed to cope with that condition.

>     # libxml2 lacks an HTML push parser setup function
>     error = xmlparser.xmlCtxtResetPush(c_ctxt, NULL, 0, NULL, c_encoding)
>     if error:
>         return error
>
>     # fix libxml2 setup for HTML
>     c_ctxt.progressive = 1
>     c_ctxt.html = 1
>     htmlparser.htmlCtxtUseOptions(c_ctxt, parse_options)
>
>     if c_data is not NULL and buffer_len > 0:
>         return htmlparser.htmlParseChunk(c_ctxt, c_data, buffer_len, 0)

  If you think a real C function htmlCtxtResetPush might be useful, then
as usual I take patches ! :-)

Daniel

-- 
Red Hat Virtualization group http://redhat.com/virtualization/
Daniel Veillard      | virtualization library  http://libvirt.org/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine  http://rpmfind.net/