Sorry, I forgot this part. This is a very low risk change. This is a change only to the way in which attribute values that contain charCode 13 are serialized. Since literal newlines in attribute values are dicey at best, this seems like a very compatible change. There's a limited possibility that this will expose a difference between a JavaScript carriage return and an XML carriage return (some native line ending thing) but I think this is unlikely.
A On Mar 23, Jim Grandy wrote: > Adam, > > What is the risk of regression with this change? Any potential for unintended > side effects? In other words: how isolated is the change? > > On Mar 23, 2007, at 1:00 PM, Adam Wolff wrote: > > > svn change 4462 fixes XML serialization of newlines in > > attribute values. Currently newlines are escaped as \r in XML > > serialization, which is generally interpreted by XML parsers > > as a space. The W3C spec for XML serialization recommends > > using the canonicalized form: > > http://www.w3.org/TR/1999/WD-xml-c14n-19991109.html#charescaping > > > > > > A >
