in python's xml deserializer (whitespace is significant when
recognizing <None/> nodes, so a general pretty-printer will cause
them to not be recognized.)
This changes empty elements to the more verbose <foo></foo> format,
but that is handled properly by the xml parser.
- Legacy-Id: 382
Only pop if the key is in the dict. (Note: an alternate
implementation would have been pop(arg, '') except the QueryDict
doesn't implement that usage)
- Legacy-Id: 347
* Remove extra "/" in case the database has extra leading or trailing slashes.
* Catch KeyError, not IndexError, since that's what gets thrown in the case I was thinking about.
- Legacy-Id: 335