nnrss-request-article produces multipart articles

nnrss-request-article produces multipart articles

[Klaus Straubinger]

Due to some last-minute changes before the release of Gnus 5.10.1, the
function nnrss-request-article now produces multipart/alternative
articles. The reason given for this change was to avoid some
unspecified problems with w3m. In my view, this is no sufficient reason
for such a change. nnrss should always produce text/html articles since
RSS sites use XML/HTML as output format.

Furthermore, the created HTML code in the text/html part is illegal HTML:
text is added after the closing "</html>" tag.

The nnrss-content-function now has a difficult task to do something
sensible like adding the content of the included link.

Some stylistic issues also come to mind:
  - some code lines are longer than 80 characters,
  - the use of "when" instead of "if" in combination with "progn" seems
    preferable,
  - instead of producing multipart headers manually, an analogon to
    the function mm-insert-multipart-headers should be created and used.

-- 
Klaus Straubinger

Re: nnrss-request-article produces multipart articles

[Mark A. Hershberger]

Klaus Straubinger <KSNetz@Arcor.DE> writes:

> Due to some last-minute changes before the release of Gnus 5.10.1, the
> function nnrss-request-article now produces multipart/alternative
> articles. The reason given for this change was to avoid some
> unspecified problems with w3m. In my view, this is no sufficient reason
> for such a change. nnrss should always produce text/html articles since
> RSS sites use XML/HTML as output format.

Yay! Competing feature requests.

RSS feeds do not always contain embedded HTML.  In fact, the 
many people strongly urge that <description> elements not contain
anything except plain text[1].  Other nnrss users complained when I
made the default a single text/html.  The multipart/alternative lets
you choose.

FWIW, there are at least four different ways that content is put into an
RSS feed.

<description> -- Strongly suggested that new feeds just put plain
  text here, but historical use means there are a lot of feeds with
  encoded HTML.  Often this is just an excerpt of the full piece.
  Some people put a short summary here.

<content:*>, especially <content:encoded> -- Encoded markup.  Often
  contains the full content of the item.

<xhtml:body> -- The full content in all its un-encoded XHTML glory.

<content:items> containing <rdf:value> -- The full content in all its
  un-encoded XML glory.

I'd love for nnrss to support all of these.  In fact, since many feeds
contain <description> (plain-text) and one of the other methods, I'd
like to put <description> in a text/plain part and the others in a
text/html part if more than just the <description> is present.

>   - instead of producing multipart headers manually, an analogon to
>     the function mm-insert-multipart-headers should be created and used.

Thanks for letting me know.  This is the sort of thing where a Gnus
developer manual would help.

I'll fix some of the problems you mentioned tonight.

Mark.

Footnotes: 
[1]  http://www.textism.com/article/553/
     http://philringnalda.com/blog/2002/09/rss_10_contentencoded.php
     http://www.peerfear.org/rss/permalink/1028943207.shtml
     http://revjim.net/archives/2002/09/25/c3818.php

-- 
You are a mystery as deep as the sea; the more I search, the more
I find, and the more I find the more I search for you.
	    -- St. Catherine of Siena