Mailing-List: contact cocoon-dev-help@xml.apache.org; run by ezmlm
Precedence: bulk
Reply-To: cocoon-dev@xml.apache.org
Date: Mon, 31 Mar 2003 14:53:02 -0800
From: Rob Johnston <rjohnst@juniper.net>
To: cocoon-dev@xml.apache.org
Subject: Re: Question about TextRecorder in the SAXTransformer
Message-ID: <20030331225302.GJ89266@juniper.net>
References: <20030330171843.GQ26892@juniper.net>
 <LKENJPKDAIKAPKHEOCMMCEOEDCAA.cziegeler@s-und-n.de>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <LKENJPKDAIKAPKHEOCMMCEOEDCAA.cziegeler@s-und-n.de>
User-Agent: Mutt/1.4i

Is the bug the fact that the data is being passed in two [characters]
events or that the .trim() is in TextRecorder.java?  I can nuke the
.trim() locally but don't want to hose anything down the line.

I just want to 'fix' it here locally since I'm on a time constraint, but
I will file a PR for you all also.

thanks,
rob

On Mon, 31 Mar 2003, Carsten Ziegeler wrote:

> Hi Rob,
> 
> this is a bug!
> 
> Carsten
> 
> > -----Original Message-----
> > From: Rob Johnston [mailto:rjohnst@juniper.net]
> > Sent: Sunday, March 30, 2003 7:19 PM
> > To: cocoon-dev@xml.apache.org
> > Subject: Question about TextRecorder in the SAXTransformer
> > 
> > 
> > In the TextRecorder.java, I see the following:
> > 
> > public void characters(char ary[], int start, int length)
> >     throws SAXException {
> >         String value = new String(ary, start, length).trim();
> >         if (value.length() > 0) {
> >             buffer.append(value);
> >         } else {
> >             buffer.append(' ');
> >         }
> >     }
> > 
> > I am wondering what the logic behind having the trim() in there is.
> > Several times I have noticed Cocoon "eating" spaces out of the stream as
> > evidenced in this log:
> > 
> > INFO    2003-03-28 16:05:14.728 [sitemap.transformer.log ] (): 
> > [startElement] 
> > uri=http://xml.juniper.net/cocoon/work-schedule/1.0,local=start,ra
> > w=sched:start
> > DEBUG   2003-03-28 16:05:14.729 [sitemap.transformer.sche] (): 
> > BEGIN startTextRecording
> > DEBUG   2003-03-28 16:05:14.730 [sitemap.transformer.sche] (): 
> > END startTextRecording
> > INFO    2003-03-28 16:05:14.731 [sitemap.transformer.log ] (): 
> > [characters] 03/28/2003 
> > INFO    2003-03-28 16:05:14.733 [sitemap.transformer.log ] (): 
> > [characters] 04:05 PM
> > INFO    2003-03-28 16:05:14.734 [sitemap.transformer.log ] (): 
> > [endElement] 
> > uri=http://xml.juniper.net/cocoon/work-schedule/1.0,local=start,qn
> > ame=sched:start
> > DEBUG   2003-03-28 16:05:14.735 [sitemap.transformer.sche] (): 
> > BEGIN endTextRecording
> > DEBUG   2003-03-28 16:05:14.736 [sitemap.transformer.sche] (): 
> > END endTextRecording text=03/28/200304:05 PM
> > INFO    2003-03-28 16:05:14.737 [sitemap.transformer.log ] (): 
> > [endElement] 
> > uri=http://xml.juniper.net/cocoon/work-schedule/1.0,local=every,qn
> > ame=sched:every
> > ERROR   2003-03-28 16:05:14.739 [sitemap.transformer.sche] (): 
> > URLWorkScheduleTransformer: invalid time: 03/28/200304:05 PM
> > 
> > <sched:start>03/28/2003 04:05 PM</sched:start>
> > 
> > is what is supposed to be being passed to the next transformer, but
> > instead, it is showing up in two different [characters] events, and
> > because of the trim(), it ends up being put back together as
> > "03/28/200304:05 PM" which is invalid.
> > 
> > I was just wondering why the trim() is in there, or if what I am seeing
> > is expected behavior.
> > 
> > thanks,
> > rob