Re: svn commit: r607188 - /xmlgraphics/fop/branches/Temp_ImagePackageRedesign/src/java/org/apache/fop/fo/FOTreeBuilder.java

2008-01-07 Thread Vincent Hennebert
Hi,

Max Berger wrote:
> Jeremias,
> 
> not that fop is on 1.4: AFAIK this is what assert statements are for:
> Add checks for things which *should not happen*, such as contract
> violations.

Yeah, I would probably add an assert statement, and if asserts are not 
enabled a NPE will be thrown anyway. But like for IllegalStateException 
that will probably also cause any prior exception to be swallowed by 
Xalan? So I guess the current situation is ok.



Vincent


-- 
Vincent HennebertAnyware Technologies
http://people.apache.org/~vhennebert http://www.anyware-tech.com
Apache FOP Committer FOP Development/Consulting


Re: svn commit: r607188 - /xmlgraphics/fop/branches/Temp_ImagePackageRedesign/src/java/org/apache/fop/fo/FOTreeBuilder.java

2008-01-07 Thread Max Berger
Jeremias,

not that fop is on 1.4: AFAIK this is what assert statements are for:
Add checks for things which *should not happen*, such as contract
violations.

Max

On Mon, 2008-01-07 at 09:21 +0100, Jeremias Maerki wrote:
> It could be null if the caller violates the ContentHandler contract. The
> question should be: Should we validate the ContentHandler contract with
> check code of our own? But that's an academic question because the check
> is barely noticeable at runtime.
> 
> On 02.01.2008 12:18:18 Vincent Hennebert wrote:
> > Hi,
> > 
> > > URL: http://svn.apache.org/viewvc?rev=607188&view=rev
> > > Log:
> > > Throwing IllegalStateException causes a prior exception to be swallowed 
> > > by Xalan-J. Need to throw a SAXException instead.
> > > Instead of logging an error about the element mismatch throw a 
> > > SAXException because the logging only confuses the user as it's 
> > > practically always a follow-up exception of an exception happening 
> > > earlier in the respective startElement() event.
> > > Modified: 
> > > xmlgraphics/fop/branches/Temp_ImagePackageRedesign/src/java/org/apache/fop/fo/FOTreeBuilder.java
> > 
> > > +/** [EMAIL PROTECTED] */
> > >  public void endElement(String uri, String localName, String 
> > > rawName)
> > >  throws SAXException {
> > >  if (currentFObj == null) {
> > 
> > How can currentFObj be null in the first place? Normally it was set in 
> > the previous startElement event (or a ValidationException has been throw 
> > anyway)?
> > 
> > Vincent
> > 
> > 
> > -- 
> > Vincent HennebertAnyware Technologies
> > http://people.apache.org/~vhennebert http://www.anyware-tech.com
> > Apache FOP Committer FOP Development/Consulting
> 
> 
> 
> 
> Jeremias Maerki
> 



Re: svn commit: r607188 - /xmlgraphics/fop/branches/Temp_ImagePackageRedesign/src/java/org/apache/fop/fo/FOTreeBuilder.java

2008-01-07 Thread Jeremias Maerki
It could be null if the caller violates the ContentHandler contract. The
question should be: Should we validate the ContentHandler contract with
check code of our own? But that's an academic question because the check
is barely noticeable at runtime.

On 02.01.2008 12:18:18 Vincent Hennebert wrote:
> Hi,
> 
> > URL: http://svn.apache.org/viewvc?rev=607188&view=rev
> > Log:
> > Throwing IllegalStateException causes a prior exception to be swallowed by 
> > Xalan-J. Need to throw a SAXException instead.
> > Instead of logging an error about the element mismatch throw a SAXException 
> > because the logging only confuses the user as it's practically always a 
> > follow-up exception of an exception happening earlier in the respective 
> > startElement() event.
> > Modified: 
> > xmlgraphics/fop/branches/Temp_ImagePackageRedesign/src/java/org/apache/fop/fo/FOTreeBuilder.java
> 
> > +/** [EMAIL PROTECTED] */
> >  public void endElement(String uri, String localName, String 
> > rawName)
> >  throws SAXException {
> >  if (currentFObj == null) {
> 
> How can currentFObj be null in the first place? Normally it was set in 
> the previous startElement event (or a ValidationException has been throw 
> anyway)?
> 
> Vincent
> 
> 
> -- 
> Vincent HennebertAnyware Technologies
> http://people.apache.org/~vhennebert http://www.anyware-tech.com
> Apache FOP Committer FOP Development/Consulting




Jeremias Maerki



Re: svn commit: r607188 - /xmlgraphics/fop/branches/Temp_ImagePackageRedesign/src/java/org/apache/fop/fo/FOTreeBuilder.java

2008-01-02 Thread Vincent Hennebert
Hi,

> URL: http://svn.apache.org/viewvc?rev=607188&view=rev
> Log:
> Throwing IllegalStateException causes a prior exception to be swallowed by 
> Xalan-J. Need to throw a SAXException instead.
> Instead of logging an error about the element mismatch throw a SAXException 
> because the logging only confuses the user as it's practically always a 
> follow-up exception of an exception happening earlier in the respective 
> startElement() event.
> Modified: 
> xmlgraphics/fop/branches/Temp_ImagePackageRedesign/src/java/org/apache/fop/fo/FOTreeBuilder.java

> +/** [EMAIL PROTECTED] */
>  public void endElement(String uri, String localName, String rawName)
>  throws SAXException {
>  if (currentFObj == null) {

How can currentFObj be null in the first place? Normally it was set in 
the previous startElement event (or a ValidationException has been throw 
anyway)?

Vincent


-- 
Vincent HennebertAnyware Technologies
http://people.apache.org/~vhennebert http://www.anyware-tech.com
Apache FOP Committer FOP Development/Consulting