Re: What characters can go into a CDATA section and a comment? (I found inconsistencies)

On Wed, 23 Mar 2022 at 11:39, Roger L Costello <[email protected]> wrote:

> Hi Folks,
>
> I did some testing to see what characters can go into a CDATA section and
> a comment. Here's what I found:
>
> Recall the CDATA syntax: <![CDATA[ ... ]]>
>
> Question: Can the content of a CDATA section contain the ] character?
>
> <![CDATA[foo]bar]]>
>
> Answer: Yes
>
> Question: Can the content of a CDATA section contain ]]?
>
> <test><![CDATA[foo]]bar]]></test>
>
> Answer: Yes.
>
> Question: Can the content of a CDATA section contain more than two ]
> characters at the end?
>
> <![CDATA[foo]]]]]>
>
> Answer: Yes.
>
> Question: Can the content of a CDATA section contain more than one [
> character at the start?
>
> <![CDATA[[[foo]]>
>
> Answer: Yes.
>


The perhaps more surprising answer related to this is for the question

Can element content contain the string ]]> without escaping one of the
characters?



> Recall the comment syntax: <!-- ... -->
>
> Question: Can the content of a comment contain a dash character?
>
> <!-- foo - bar -->
>
> Answer: Yes.
>
> Question: Can the content of a comment contain two dashes?
>
> <!-- foo -- bar -->
>
> Answer: No.
>
> Question: Can the content of a comment end with more than two dashes?
>
> <!-- foo----->
>
> Answer: No.
>
> Notice the inconsistency: A CDATA section can end with more than two ]
> characters but a comment cannot end with more than two dashes:
>


>
> <![CDATA[foo]]]]]>      Okay
> <!-- foo----->          Error
>
> Also, a CDATA section can contain ]] but a comment cannot contain --
>
> <![CDATA[foo]]bar]]>
> <!-- foo -- bar -->
>
> Why is there this inconsistency?
>

It's not inconsistency, it's historical legacy.  XML comment syntax is a
simplified but compatible subset of SGML comment syntax which allows
multiple comments within a declaration, marked with --  so

<!ENTITY -- this is a comment -- foo "bar" -- so is this -- >


so if you parsed xml with an sgml parser a -- would end the comment so in
xml it's only allowed at the end.


> Isn't inconsistency in a data format language bad?
>

A world without history would be so dull.


>
> /Roger
>
>
David


>
> _______________________________________________________________________
>
> XML-DEV is a publicly archived, unmoderated list hosted by OASIS
> to support XML implementation and development. To minimize
> spam in the archives, you must subscribe before posting.
>
> [Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
> Or unsubscribe: [email protected]
> subscribe: [email protected]
> List archive: http://lists.xml.org/archives/xml-dev/
> List Guidelines: http://www.oasis-open.org/maillists/guidelines.php
>
>