Assignment: rss & atom test cases

zem · 2005-09-22 03:49:12

The Feed Validator is helpful in knowing if RSS and Atom output is correct, but it only tells part of the story. It doesn’t detect multiple-encoding problems, or tell us if the output as displayed by a feed reader looks the way it should.

So, what we need is:

1. A collection of articles to use as standard test cases, stored in a mysqldump file. They should include every conceivable tricky situation: utf-8 characters, named and numeric entities, raw ampersands, raw html, textpattern tags, missing titles, etc.

2. A snapshot of each feed as it should be – saved in a file, with any errors corrected manually.

The idea is that we can load up the mysql dump file, fetch the feed output, and compare it with the snapshot. We can also view the snapshot in a variety of feed readers, to make sure subscribers are seeing what we thing they’re seeing.

Bonus points for taking a copy of the RSS snapshot, and manually transforming it into RSS 2.0 format.

ubernostrum · 2005-09-25 02:22:30

The validator’s code comes with a fairly comprehensive test suite… before I think about how that’d best be packaged for this sort of use, is there any reason why it wouldn’t make a good set from which to test?

zem · 2005-09-25 04:56:41

We can already use feedvalidator.org to check the syntax of feeds, I’m not sure that the suite would provide much more than a few ideas for edge cases. Duplicating tests that are already done by feedvaildator.org is a waste of effort.

What we need is something to check the content, not the syntax: to make sure that the output, when decoded by a feed reader or parser, correctly corresponds to the input. i.e. if I put some *textiled* "words" in an article (or excerpt, or title), what should the feed output be, for Atom, RSS 0.92, and (ideally) RSS 2.0? <content>some textiled words</content> would be valid, but it’s not what we want.

jiblet · 2005-09-30 22:12:31

Quick question, since I am largely ignorant of RSS matters… since Textile is already geared towards outputting valid XHTML, can we not just wrap the content in a CDATA and just output it verbatim? I believe that’s how WordPress does it. Sorry if this issue has already been discussed.

phiw13 · 2005-10-01 00:46:05

I guess you know that, but the RSS feed from the Dev Weblog comes out very bad, especially code blocks [1]. That is something I can reproduce with the latest rev. If you want the parsed html code as seen by NNW, just ask.

[1] screenshot from this article, seen in NetNewsWire.

zem · 2005-10-02 01:12:56

Here’s a starting point: test cases.

The feed output is as generated by 4.0.1 – that is to say, sometimes it’s probably encoded incorrectly, and might have other errors also.

Still to do:

Examine the test output errors, and manually correct them (i.e. produce a copy of that file as it should be, not as it is generated currently)
Create new test cases (articles containing relative links, inline images, utf-8, conditional tags, other tricky stuff), and the corresponding expected output
Create expected output snapshots, from the same input data, for RSS 0.92, 2.0

We could definitely use some help here. Doing this sort of thing ties up the dev team, and keeps us from working on new features.

colak · 2005-10-02 09:16:23

<blockquote> zem wrote:

> Here’s a starting point: test cases.

The feed output is as generated by 4.0.1 – that is to say, sometimes it’s probably encoded incorrectly, and might have other errors also.</blockquote>
I would very much like to help here as I do have some <a href=“http://forum.textpattern.com/viewtopic.php?pid=79419”>unresolved</a> issues regarding rss (and atom). Alex, how did you generate that rss? is that just a normal mySQL dump of the db?

<blockquote>We could definitely use some help here. Doing this sort of thing ties up the dev team, and keeps us from working on new features.</blockquote>
I am volunteering (I have no php/mySQL experience). Do let me know of steps I can take and i will be posting my results.

zem · 2005-10-03 22:14:42

Alex, how did you generate that rss? is that just a normal mySQL dump of the db?

The article data is a quick summary of the important bits of the article table, generated in PHP. If you create a bunch of articles containing the same title/body/excerpt, then the Atom feed output you get should be identical (other than timestamps, article ID’s etc).

There are two copies of the output: one with the feed set to include a summary only, and one with it set to include the entire article.

Do let me know of steps I can take and i will be posting my results.

What we need people to do is:

1. examine the feed output in that file, with respect to the article data, and find places where there’s something wrong in the feed. Perhaps it’s encoded incorrectly, or something is missing.

2. suggest new articles to include in the tests. Articles containing relative links, for example.

3. show us what the feed output for those same articles should look like for RSS 0.92 and/or 2.0.

colak · 2005-10-04 07:26:19

Hi alex
Here is a sample of the rss with notes…
<a href=“http://neme.org/main/nemerss.xml.txt”>nemerss.xml.txt</a>

hope it will be of help

Last edited by colak (2005-10-04 07:57:34)

darrepac · 2005-10-07 07:13:30

Hello,

On my side, sharpreader detect often error on my rss feed…when I launch feed validator I have every time the following error :
line 14, column 79: Missing channel element: description

Pac

zem · 2005-10-08 03:02:39

To recap:

The test case file lists the content of some articles (title/excerpt/body), and a snapshot of the Atom feed output currently produced by Textpattern for those articles. There are two versions of the feed: one with summaries only, and one with article bodies.

In some cases, the output is wrong: incorrect encoding, validation problems, (perhaps) incorrect content.

What we need is for some people to examine the output, check it against the input, and tell us where it’s wrong, and what the correct output should be.

Ideally, also show us what the correct output would be for RSS 0.92 and RSS 2.0 feeds produced from the same article data.

colak · 2005-10-08 08:03:35

Hi Alex. maybe <a href=“http://forum.textpattern.com/viewtopic.php?pid=80207#p80207”>this</a> post might be relevant to this topic.

Textpattern CMS

Textpattern CMS support forum

#1 2005-09-22 03:49:12

Assignment: rss & atom test cases

#2 2005-09-25 02:22:30

Re: Assignment: rss & atom test cases

#3 2005-09-25 04:56:41

Re: Assignment: rss & atom test cases

#4 2005-09-30 22:12:31

Re: Assignment: rss & atom test cases

#5 2005-10-01 00:46:05

Re: Assignment: rss & atom test cases

#6 2005-10-02 01:12:56

Re: Assignment: rss & atom test cases

#7 2005-10-02 09:16:23

Re: Assignment: rss & atom test cases

#8 2005-10-03 22:14:42

Re: Assignment: rss & atom test cases

#9 2005-10-04 07:26:19

Re: Assignment: rss & atom test cases

#10 2005-10-07 07:13:30

Re: Assignment: rss & atom test cases

#11 2005-10-08 03:02:39

Re: Assignment: rss & atom test cases

#12 2005-10-08 08:03:35

Re: Assignment: rss & atom test cases

Board footer