Greetings:

I installed the Import module on my Drupal test bed
(running Drupal 4.4.0), and generally like it.
However, it seems to have a problem with certain
XML/RSS feeds that do not have a complete <title>
element.

The problem could be due, in part, to the way Moveble
Type sites gennerate XML.  If an MT user saves an
article with the title field blank, MT only writes the
ending <title /> tag and not the beginning. Example:

... 

<item>
  <title /> 
<link>http://www.somesite.com/archives/015809.php</link>
  <description>Blah blah blah ...


I think the Import module is seeing this as a
poorly-formed feed and is ignoring it. Is there a bug
fix or a workaround that I can apply so I can get it
to recognize feeds like this? (Also, how are null titles
handled by Drupal generally?)

Thanks in advance for your help.


(Sorry for the php tags. I don't know how to post code in here without the angle brackets disappearing.)

Comments

gruffbear’s picture

gruffbear commented

"What's 'the problem,'" you ask? The feeds just don't appear at all. I apologize for not making that clear. Thanks.

marky’s picture

marky commented
Priority: Critical » Minor

I'd suggest contacting the feed authors and pointing out their feeds are invalid. The <title> is a *required* element in rss 1 & 2, so I'm not convinced we should "break" Drupal to accommodate feed authors who produce invalid feeds. The fact that MT allows you to produce broken feeds is beyond the scope of this issue, and I'd recommend the developers set this to "won't fix".

See http://web.resource.org/rss/1.0/spec#s5.3.1 and http://feedvalidator.org/docs/rss2.html#requiredChannelElements for more...

gruffbear’s picture

gruffbear commented
Priority: Minor » Critical

Regarding the feed in question, when the site has 100,000+ hits per day, I have no influence over the guy who produces this rather popular blog; nor do I have any influence over the people over at MT. To them, I'm nobody. (And apparently to you as well.)

So if you were designing automobiles, you would build them without springs or shock absorbers -- and then blame the government for not building roads that are smooth enough?

The blogger in question, and MT, and you can all have the attitude that "it's not my issue." That's all well and good; but I need a robust application that works with news feeds that are found in the real world.

Thank you for your insights, which I will take to heart.

gruffbear’s picture

gruffbear commented

To clarify: This has to do with the Import MODULE, not the syndication feed handler in core Drupal. The module allows use of the feed as a Class A node, which is why I need it. The syndication feed handler works perfectly with the apparently malformed Title block; so as to "breaking Drupal," I'm not exactly sure what you mean. Cheers.

marky’s picture

marky commented

It's completely immaterial how popular the feed is - it doesn't change the fact that it's broken as indicated by your own admission.

There are literally hundreds of potential reasons for a feed to be broken or invalid, and I hope you would agree it would be impossible to code for every possible eventuality.

There are well established standards for rss as indicated in the links I supplied, and this system (by and large) abides by them. These standards exist for a reason - by following the standards you maximize your chances of producing compatible feeds. Your feed author chose not to follow these standards. How is that the fault of this system? How can it be the responsibility of this piece of software to compensate for the failings of systems outside it's control?

The <title> has been a required element as far back as HTML 2.0. If MT allows the author to create invalid markup then that's a failing of MT, and nothing to do with Drupal.

I notice you reset this report to critical. Please explain how it can be considered a critical bug in a piece of software to ignore invalid input. While you're at it, please explain how it's even a bug in the first place. If you'd like to see more generous feed parsing in the import.module, then feel free to open a 'feature request' and ask for it. IMHO if this bug report belongs anywhere it belongs on the MT site.

My original suggestion still stands - contact the author and indicate his feed is invalid. You can (or should) include the references I outlined above. I would guess that Drupal probably isn't the only software that cannot read his feeds. Fixing his markup would more than likely increase his readership. You never know, he may even be grateful you made the effort... ;)

gruffbear’s picture

gruffbear commented

It's critical to me because the issue is a show-stopper on my project. It obviously isn't critical to you because you don't give a damn.

In my daytime life I work with a commercial CMS (Stellent) that can handle virtually any feed you throw at it. Based on that experience, I guess my expectations for Drupal were way too high.

In particular, I don't understand how come the malformed Title element is happily accepted by the core Drupal application, but not in the Import module. Wouldn't that tend to suggest that there's a bug in the Import module?

Maybe you should submit a bug report to "fix" core Drupal so that it barfs on malformed Title elements too. That would make your position internally consistent and defensible. :)

If you post anything else to this thread, you'll have the last word. I give up. It's a drag to have to deal with people who would rather be "right" than be excellent.

"A foolish consistency is the hobgoblin of little minds." --Ralph Waldo Emerson

Bèr Kessels’s picture

Bèr Kessels commented

Wow, gruffbear, calm down please!

Indeed, if you would have looked around, there are a lot of people looking for improvement of this module. This import module (now named node aggregator) is *way* too tight in validating. This needs to be changed.

And Last but not least: If someone says something to you: don't shout at the whole of drupal. It might make people like mee feel annoyed. I am working hard to get this module up and running, to submit patches and to reply to feature requests! So if you say to the whole of drupal that they can stick the module up somewhere (you never used those words, but they summarize your last comment) that includes me. I as maintainer, however agree ewith you, and like to get input.

For now I am marking this duplicate, because already two issues are dealing with the too tight validation-issue.

Regards, Ber

Bèr Kessels’s picture

Bèr Kessels commented
Status: Active » Closed (won't fix)

This module is deprecated in favour of http://drupal.org/project/feedparser