Re: [xml] xmlHashScan and *data argument

[Daniel Veillard <veillard redhat com>]

On Mon, Oct 20, 2003 at 08:50:23AM +0800, William M. Brack wrote:
> Melvyn Sopacua said:
> > On Sunday 19 October 2003 23:43, Daniel Veillard wrote:
> > > On Sun, Oct 19, 2003 at 11:31:54PM +0200, Melvyn Sopacua wrote:
> <snip>
>
> > I still don't get why xmlDumpElementDecl accepts buf as first
> > argument while
> > it is an xmlHashScanner on an xmlElementTable - the same as I'm
> > doing.
> >
> > I won't dwell on it, but if you can explain it, I can sleep better
> > ;)
> >
> > > > Further investigation shows that xmlSchemaTypeDump in
> > > xmlschema.c indeed
> > > > has the argument order you list and swapping the arguments in my
> > > function
> > > > works.
> > >
> > >   Okay, normal. There is actually  3rd argument passed if you look
> > > at the
> > > declaration of xmlHashScanner() in libxml/hash.h
>
> hmmmm..... This thread has gotten my poor mind all twisted.  It
> certainly seems to me that Melvyn has a "valid" point concerning the
> scan callbacks for dumping data in valid.c.  Let's see... the
> function def for the scanner callback is:

  Hum, maybe I didn't put enough attention in checking the actual
code. I looked at the code posted in the mail, and assumed the one
in valid.c worked (maybe it's simply not used anymore) because that
kind of error should lead to an immediate crash if called.
  Maybe bugzilla it so this issue gets checked for good (might
be as simple as xmllint'in a document with an internal subset and
putting a breakpoint in the given function).

Daniel

-- 
Daniel Veillard      | Red Hat Network https://rhn.redhat.com/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/