[rfc-i] CDATA, was: Fwd: New Version Notification for draft-hoffman-xml2rfc-06.txt

Julian Reschke julian.reschke at gmx.de
Wed Apr 30 09:51:46 PDT 2014


On 2014-04-30 18:41, Paul Kyzivat wrote:
> On 4/30/14 12:19 PM, Dale R. Worley wrote:
>>> From: Paul Kyzivat <pkyzivat at alum.mit.edu>
>>>
>>> I have had occasion to include blocks of XML as artwork in a draft.
>>> Doing this using escaping is really unpleasant. Using CDATA helps a lot.
>>> But of course you can still get in trouble with ]]>. I wish there was
>>> some better solution for those cases, that would allow copy/paste of the
>>> xml source.
>>
>> As Paul Hoffman noted, you can clean up the copy&paste by replacing
>>
>>      ]]>
>>
>> with
>>
>>      ]]]]><![CDATA[>
>>
>> throughout.  Your editor should have a command that can do that.  All
>> solutions will require escaping *something*, because otherwise there's
>> no way to detect the end of the text to be interpreted literally.
>
> Yes, I know. I was thinking of boundaries in mime multipart, where I can
> specify an arbitrary delimiter. But this is really a XML issue, not an
> xml2rfc issue.
>
> What about having an include mechanism that indicates the included
> content is to be treated as plain text, not xml? And when a processor
> *generates* a canonical XML without external references, it could then
> do the needed escaping.

I'm using a shell script that I use for including either XML or plain 
text into XML (the XML is always self-contained; the shell script 
refreshes the document on demand). See 
<http://trac.tools.ietf.org/tools/xml2rfc/trac/browser/vocabulary/v2/latest/refresh-inclusions.sh>.

If there's sufficient interest I can make it part of the rfc2629.xslt 
distribution.

Best regards, Julian



More information about the rfc-interest mailing list