I wanted to preserve newlines in the generated webpage. I had a
solution with split in my template. Something like:
#####
<element py:strip="" py:for="line in blogEntry.entryText.split('\n')">
    ${line}<br/>
</element>
#####

I made the following function:
#####
def splitLines(thisText):
    if thisText == None:
        return []
    thisText = thisText.replace('&', '&amp;')
    thisText = thisText.replace('<', '&lt;')
    thisText = thisText.replace('>', '&gt;')
    return '<br/>'.join(thisText.split('\n'))
#####

and now I can put the following in my template:
#####
${XML(tg.splitLines(blogEntry.entryText))}
#####

This is more clear and I am rid of the <br/> at the end.

--
Cecil Westerhof

You might want to add &quot; to your list. And if in the majority of
cases thisText doesn't contain the special characters the following
will be better performance-wise:

 def splitLines(thisText):
     if thisText == None:
         return []
     if '&' in thisText: thisText = thisText.replace('&', '&amp;')
     if '<' in thisText: thisText = thisText.replace('<', '&lt;')
     if '>' in thisText: thisText = thisText.replace('>', '&gt;')
     if '"' in thisText: thisText = thisText.replace('"', '&quot;' )
     return '<br/>'.join(thisText.split('\n'))

And if this part of your code becomes the real performance bottleneck
you might want to do the search/replace using the re module. Take a
look at xml.etree.ElementTree in the stdlib for an example (if you
have python > 2.4).

Cheers,
Daniel
--
Psss, psss, put it down! - http://www.cafepress.com/putitdown

Daniel Fetchinson schrieb:

There is also a function cgi.escape for doing this in the standard lib.

-- Christoph

2008/5/13 Daniel Fetchinson <fetchin...@googlemail.com>:

I do not think this will be the bottleneck. Displaying is offcourse
what is done most, but retreiving the values from the database is more
expensive as the replacements I think.
But it never hurts to think about performance I think. So I made it like:
#####
startSpaces = re.compile(' +')
def splitLines(thisText):
    if thisText == None:
        return None
    thisText = xml.etree.ElementTree._encode_entity(thisText)
    lines    = thisText.split('\n')
    for i in range(len(lines)):
        match = startSpaces.match(lines[i])
        if match:
            lines[i] = '&nbsp;' * match.end() + lines[i][match.end():]
    return '<br/>'.join(lines)
#####

With xml.etree.ElementTree._encode_entity the conversion is done more
efficiënt and at the same time there is more done.
I find the spaces at the start of a line significant. That is why I
added the for-loop. In this way the spaces at the start of a line are
replaced with &nbsp; and is the indent of a line saved.

--
Cecil Westerhof

2008/5/13 Cecil Westerhof <cldwester...@gmail.com>:

     if thisText == None:
         return None
should be changed to
     if thisText == None:
         return ''

Otherwise XML crashes.

--
Cecil Westerhof