re module does not handle backreferences in replacement text.

[pyjsorg]

The re module for pyjs does not handle backreferences in replacement
text.

For example, regexes to do simple reformatting of text, as in:

Text between single asterisks will be emphasized.
Text between double asterisks will be bolded.

Can be handled by these compiled regular expressions:

B_re = re.compile(r'\*\*(.*?)\*\*', re.DOTALL)
EM_re = re.compile(r'\*(.*?)\*', re.DOTALL)

and later used like this:

txt = myTextArea.getText()
txt = B_re.sub(r'<STRONG>\1</STRONG>', txt)
txt = EM_re.sub(r'<EM>\1</EM>', txt)

This works great with pyjd, but with pyjs, you get "\1" in the
replacement text, not the backreference.

Full example is attached.

Original issue: http://code.google.com/p/pyjamas/issues/detail?id=495 (September 18, 2010 16:02:23)

[chozabu]

Blast - just encountered this myself.

[chozabu]

my workaround is along these lines:

    otext = "text with a http://chozabu.org link"
    newtext=""+otext
    exp= r"((https?|ftp)://[^ ]+)"
    linklist=re.findall(exp, otext)

    for link in linklist:
        newlink = "<a href="+link[0]+">"+link[0]+"</a>"
        newtext = newtext.replace(link[0], newlink)
    self.add(HTML(newtext))