Can XML comments go anywhere?

Question

I wrote a simple tool to generate a DBUnit XML dataset using queries that the user enters. I want to include each query entered in the XML as a comment, but the DBUnit API

Answer 1

The first example is not valid XML, the declaration has to be the first thing in a XML document.

But besides that, comments can go anywhere else.

Correcting your first example:

<?xml version="1.0" encoding="UTF-8"?>
<!-- Queries used: ... -->
<dataset>
</dataset>

Answer 2

The processing instruction must be the very first thing in the XML content (see XML comment and processing instructions). The following should work:

<?xml version='1.0' encoding='UTF-8'?>
<!-- Queries used: ... -->
<dataset>
  ...
</dataset>

Answer 3

According to the XML specification, a well-formed XML document is:

document ::= prolog element Misc*

where prolog is

prolog ::= XMLDecl? Misc* (doctypedecl Misc*)?

and Misc is

Misc ::= Comment | PI | S

and

XMLDecl ::= '<?xml' VersionInfo EncodingDecl? SDDecl? S? '?>'

which means that, if you want to have comments at the top, you cannot have an XML type declaration.

You can, however, have comments after the declaration and outside the document element, either at the top or the bottom of the document, because Misc* can contain comments.

The specification agrees with Wikipedia on comments:

2.5 Comments

[Definition: Comments may appear anywhere in a document outside other markup; in addition, they may appear within the document type declaration at places allowed by the grammar. They are not part of the document's character data; an XML processor MAY, but need not, make it possible for an application to retrieve the text of comments. For compatibility, the string "--" (double-hyphen) MUST NOT occur within comments.] Parameter entity references MUST NOT be recognized within comments.

All of this together means that you can put comments anywhere that's not inside other markup, except that you cannot have an XML declaration if you lead with a comment.

However, while in theory theory agrees with practice, in practice it doesn't, so I'd be curious to see how your experiment works out.

Answer 4

Thanks for the answers everyone!

As it turns out, the comment ahead of the file seemed to work, but when I delved into the DBUnit source, it is because validation is turned off.

I did try a simple document load via:

DocumentBuilderFactory factory = DocumentBuilderFactory.newInstance();
DocumentBuilder builder = factory.newDocumentBuilder();
Document document = builder.parse(new File("/path/to/file"));

and this fails with an exception because the XML Declaration is not the first thing (as others indicated would be the case).

So, while DBUnit would work, I prefer to have valid XML, so I moved the comment to the end (since DBUnit generates the XML Declaration, it is not an option to place the comment below it, even though I would prefer that... at least not without modifying the XML after the fact, which would be more work than it is worth).