You are viewing a plain text version of this content. The canonical link for it is here.
Posted to general@xerces.apache.org by Clemens Cap <ca...@informatik.uni-rostock.de> on 2000/02/16 10:18:44 UTC

Re: encoding not supported

Thanx, Scott, for your help on this !

On Fri, Feb 11, 2000 at 03:00:44PM -0500, Scott Boag/CAM/Lotus wrote:
> > Several hours ago I discovered a nice solution to the problem: Setting
> the
> > JVM to encoding UTF-8, everything is fine and there is no problem.
> 
> Hmm... it looks like the problem is probably in FormatterToXML
> getMimeEncoding, where it does encoding = System.getProperty
> ("file.encoding"), but I'm not totally sure... I would need to retrace the
> execution path.  It could also come from StylesheetRoot if xsl:output
> method="text", or it could come from the Assaf's OutputFormat class as a
> default value, though this would surprise me a bit.

At present, the workaround of setting the encoding of the JVM to UTF-8
and using UTF-8 seems to solve my problem. However, I do not understand
fully the logic of the code and guess, as you suggest, that only extensive
retracing of the execution path could explain the situation.

> > The exception is raised below, where it says catch (Exception e3). Now
> the logic
> > of below exception code is not very clean. If javaEncoding is ISO8859-1
> > then it becomes "ISO" alone and is used in new OutputStreamWriter where
> > it most likely raises an exception. What is the intention of this ?
> 
> The problem is that the various Java VMs are not using encoding names
> consistenantly... some of them use the ISO prefix and some do not.

Mine uses ISO8859-1 instead of ISO8895_1 :)

> (Obviously, I should have commented the code.)  So, I first try creating a
> writer with "8859_1", and if that does not work, I try creating a writer
> with "ISO8859_1".  The funny thing is, this should all be covered in
> convertMime2JavaEncoding, but perhaps I was desperate at the time to get
> this to work from several levels.
> 

> >    1) It always can be detected easily which part of the program raises
> the exception
> >   (eg by dumping the stack somewhere)
> 
> If you use -edump you should be able to do this from the command line.  Not
> sure if this addresses this or if I am missing your statement.

RTFM :) I was not aware of that feature and it definitely helps me ! Thanx.

Regards,
Clemens.

-- 
Prof.Dr.Clemens H. CAP      
Dept.of Computer Science    office phone               +49-(0)381-498/3390
University of Rostock       office fax                 +49-(0)381-498/3440
Albert Einstein Strasse 21  mailto:cap@informatik.uni-rostock.de
D-18051 Rostock, Germany    http://wwwtec.informatik.uni-rostock.de/iuk

*** Please read this link  http://www.freepatents.org