Re: [gobo-eiffel-develop] Problem with XM_TREE_CALLBACKS_PIPE and Namespaces?

[Franck A. <fr...@ne...>]

Colin Paul Adams:

> Franck, could you please explain the significance of the changes for
> an application? I'm looking at CVS online (diff for 1.17 to 1.18 of 
> XM_NAMESPACE_RESOLVER), and I see that the basic change is to call
> attributes_force instead on next.on_attribute, but I don't understand
> the significance.

It's as I explained earlier: the on_attribute event for xmlns:xxx 
attribute is sent delayed with the element and other attributes 
(which are delayed because it's required for resolution), rather 
than out of order. 

No change when set_forward_xmlns (False) which is the default.

That is:

<doc xmlns='blah' b='c'/>

Event sequence before:

- @xmlns=blah
- doc
- @b=c

and after:
- doc
- @xmlns=blah
- @b=c

> (what do I need to do to revert to the old version to confirm this? 

cd $GOBO/library/xml/event
cvs update -p -r1.17 xm_namespace_resolver.e > xm_namespace_resolver.e

(plain 'cvs update' to get back to CVS head, cvs update -p only 
prints on the stdout without changing your CVS state.)