[mod_python] zero terminated strings and req.write

Erik Stephens at <mod_python at 24ksoftware.com> replied:
> On Mon, 8 Mar 2004, Frederick Grim wrote:
> 
> > def gen_xml():
> > 	return """<?xml ... Tons of xhtml stuff here
> >
> > </html>"""
> >
> > def main(req, ERR='0'):
> > 	re.content_type = 'text/xml'
> > 	req.write(gen_xml())
> > 	return apache.OK
> >
> > The problem here is that python puts null bytes at the end of a
> > string and req.write fails with the error that there is a 0 after
> > </html>.
> 
> How does it fail?  Raises an exception?  What is the actual exception
> type and value?  It sounds strange to me cuz I've never had to worry
> about null bytes in strings in my Python travels.

okay I have tracked the problem down.  It seems that return apache.OK
adds a 0 to the output that the browser picks up so somewhere between function and
browser mod_python leaves the zero at the end of the string.  The problem
manifested itself with a failure when content_type was set to 'text/xml' since I
guess mod_python shoved the output through a validating parser.

	1. Is this a correct analysis
	2. Is this a bug or my misuse of mod_python
	if is_bug:
		print 'Has this been addressed?'
	else:
		fred_file_bug_report()

	heh :)
		Fred
> 
> 
> > How do I get this failure to not happen.  Even if i set content_type
> > to 'text/html' there appears a little zero at the end of my
> > script. It seems like a bug? maybe.  I am using mod_python 3.1.3
> > apache 2.0.47, and python 2.3
> 
> Dumb questions maybe, but what handler are you using?  If you're using
> the publisher handler, then maybe that zero at the end is what
> apache.OK equals?
> 
> 
> Best regards,
> Erik
> 
> _______________________________________________
> Mod_python mailing list
> Mod_python at modpython.org
> http://mailman.modpython.org/mailman/listinfo/mod_python

-- 
"Windows Longhorn will be like having a little cop right inside your computer"
	-- Jorge Lopez, MCSE