DescriptionWe found a case of a page that uses xhtml (page uses name-space) that has the concat request encoded twice:
& was encoded to &
Looking at the code it seems that there is a special case in the DefaultHtmlSerializer that does not escape '&' when an attribute is considered a url attribute.
Ul attribute is 'src' or 'href' without name-space.
As a result there is an extra encoding code in the ConcatVisitor.
In our case the concat did encoding, and since the doc had name-space, the serializer did another encoding.
The solution is to let the serializer always the encoding as needed and eliminate all the special code.
This will also support the new CajaSerializer that do the same.
Patch Set 1 #
Total comments: 2
Patch Set 2 : Update test case and todo comment #Patch Set 3 : Merge to tip #
MessagesTotal messages: 10
|