[Uml-devel] Bug fix patch: xmlschemawriter

Brian Thomas thomas at mail630.gsfc.nasa.gov
Tue Mar 18 12:39:01 UTC 2003

On Tuesday 18 March 2003 03:21 pm, Jonathan Riddell wrote:
> On Mon, 17 Mar 2003, Brian Thomas wrote:
> >       I guess I dont understand your point here. Can you explain? As I
> >       stated earlier its simply not valid to create a schema without the
> >       opening xml processing instruction OR having the root node
> >       "<schema>". There is no such thing as a separate header file for
> >       XML schema either, and, to be valid, the header MUST follow the xml
> >       processing instruction.
> It does have the processing instruction and schema.  Have you done a cvs
> update and make install on headers/ directory ?

	Yes, and its wrong, in my opinion, to put these things in a "headings" file. That
	stuff is not supposed to be critical as to whether the code compiles or is
	valid, headings is simply a comment about the file itself. What happens to the file
	if the user doesnt use the system directory for headings (in fact, this is what I did
	just after an update from CVS and ran the xmlschemawriter... ALL my files where
	garbage since they didnt include the kosher headings file!!) 

	Futhermore, under your prescription, we no longer use the package name in the 
	targetnamespace, and xmlns: for the package must be hand edited to make what 
	is there match namespace in the generated file. Admittedly, you probably will be editing 
	this anyways since the URI is so fakey, but in the future, one of the global parameters 
	for xmlschemawriter will be the abilityto supply the URI for the target and package 
	namespaces, so having this in the headings file is again moot.

	 I continue to want the processing instruction and start schema node
	in the xmlschemawriter, not in the headings file. 


> Jonathan Riddell


  * Dr. Brian Thomas 

  * Code 630.1 
  * Goddard Space Flight Center NASA

  *   fax: (301) 286-1771
  * phone: (301) 286-6128

A nuclear war can ruin your whole day.

More information about the umbrello-devel mailing list