Re: [Rfc-markdown] empty include files

[Carsten Bormann <[e-mail]>]

Hi Michael.

> Line 351 says:
>        <artwork name="" type="" align="left" alt=""><![CDATA[
> ]]></artwork>
> 
> As a result of:
> ~~~~
> {::include yang/ietf-voucher-tree-latest.txt}
> ~~~~

The kramdown-rfc part of this works as designed.

> Putting some content ("HELLO") into yang/ietf-voucher-tree-latest.txt makes
> xml2rfc happy.
> 
> This feels like an XML2RFC bug in that it can't cope with an empty artwork.
> The diagnostics from xml2rfc seems off too, since the input doesn't have what
> was provided.
> 
> But, maybe kramdown-rfc2629 is at fault for daring to include no content.

First of all, there is nothing wrong with the {::include} mechanism here.

Without any use of {::include}, an empty code block as in...

~~~
~~~

…will give you…

<figure><artwork><![CDATA[
]]></artwork></figure>

(Note that kramdown-rfc prepends a newline to every piece of artwork that doesn’t already start with a newline because of earlier xml2rfc idiosyncrasies; I don’t know if this is strictly necessary any more.)

Second, the definition of <artwork> [1] does not contain any text that would declare empty artwork undesirable.  It should probably rule out blank (not just empty) artwork, but it doesn’t.  In any case, kramdown-rfc gave xml2rfc what you specified it should.

Yes, there is xml2rfc behavior here that could be called a bug, but then if it silently took blank artwork that wouldn’t be too helpful either.  Kramdown-rfc does not attempt to mirror all the “helpfulness” of xml2rfc, so there you are.

Grüße, Carsten

[1]: https://xml2rfc.tools.ietf.org/xml2rfc-doc.html#name-artwork-2