Links like http://localhost:port/ get rendered incorrectly

[Benno Luthiger]

Problem:
When I enter a link with a URL like http://localhost:8080/something/
this URL is rendered as http://localhost:8082/silva/mysites/mysite/mydir/http://localhost:8080/something/

The problem is caused by
SilvaDocument\mixedcontentsupport.py: line 302 in the _linkHelper() method.

This method should return on line 309 after the URL matches the URL_PATTERN.
However, this pattern doesn't anticipate URLs like http://localhost:8080/x/y/z and therefore the method returns only on line 317 with a unnecessarey prefix added.

The root cause of the problem is that the URL_PATTERN (line 23) wrongly doesn't match a correct url like http://localhost:8080/something/
URL_PATTERN = r'(((http|https|ftp|news)://([A-Za-z0-9%\-_]+(:[A-Za-z0-9%\-_]+)?@)?([A-Za-z0-9\-]+\.)+[A-Za-z0-9]+)(:[0-9]+)?(/([A-Za-z0-9\-_\?!@#$%^&*/=\.]+[^\.\),;\|])?)?|(mailto:[A-Za-z0-9_\-\.]+@([A-Za-z0-9\-]+\.)+[A-Za-z0-9]+))'

I've added the part '(http|https|ftp|news)://localhost(:[0-9]+)?(/([A-Za-z0-9\-_\?!@#$%^&*/=\.]+[^\.\),;\|])?)' as possible fix of this bug:

URL_PATTERN = r'(((http|https|ftp|news)://([A-Za-z0-9%\-_]+(:[A-Za-z0-9%\-_]+)?@)?([A-Za-z0-9\-]+\.)+[A-Za-z0-9]+)(:[0-9]+)?(/([A-Za-z0-9\-_\?!@#$%^&*/=\.]+[^\.\),;\|])?)?|(mailto:[A-Za-z0-9_\-\.]+@([A-Za-z0-9\-]+\.)+[A-Za-z0-9]+)|(http|https|ftp|news)://localhost(:[0-9]+)?(/([A-Za-z0-9\-_\?!@#$%^&*/=\.]+[^\.\),;\|])?))'

As I'm not a regex crack, possibly this could made more efficient.

[todd (todd-infrae)]

Hi Benno(?),

I applied the new regex URL_PATTERN to SilvaDocument-2.1 branch. We should be releasing it soon but I want to confer with Sylvain who may have more changes to add to the release. I've also added a new test to SilvaDocument test_mixedcontentsupport.py that gives you direct access to the link building method.

Todd

[Andy Altepeter (aaltepet)]

Benno, it looks like this was released with as part of the SilvaDocument 2.1.3 tag in svn.