[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: What to escape when serializing XML

• From: richard@i... (Richard Tobin)
• To: xml-dev@l...
• Date: Wed, 3 Jan 2007 16:19:30 +0000 (GMT)

In article <00e301c72f37$7dcad4e0$6401a8c0@turtle> you write:

>I think there's a reasonable expectation (and certainly the QT serialization
>spec assumes it) that if there's a DTD or schema that causes whitespace to
>be condensed in this way, that's because there's no semantic difference
>between different whitespace sequences, and that therefore it's OK for the
>serializer to assume that any whitespace sequence will do.

XML parsing ensures that an attribute declared with tokenised type
will be normalised to a sequence of tokens separated by single space
characters, whether the attribute is valid or not.  If it's not valid,
the tokens may not be NMTOKENs (they may even contain tabs, linefeeds
etc), but you can rely on them being separated by single spaces.

So on the one hand there's nothing you can do in serialisation to
maintain spaces within the tokens, but on the other hand if your data
came from parsing an XML document the situation is never going to
arise.

-- Richard

• References:
  • RE: What to escape when serializing XML
    • From: "Michael Kay" <mike@s...>