Closed Bug 297906 Opened 19 years ago Closed 11 years ago

Thunderbird displays only first article of RSS feed when all item <link>s are the same and no <guid> is present

Categories

(MailNews Core :: Feed Reader, defect)

defect
Not set
normal

Tracking

(Not tracked)

RESOLVED DUPLICATE of bug 264482

People

(Reporter: jure.sah, Unassigned)

References

()

Details

(Keywords: helpwanted)

Attachments

(1 file, 1 obsolete file)

User-Agent:       Mozilla/5.0 (Windows; U; Win98; sl-SI; rv:1.7.8) Gecko/20050511 Firefox/1.0.4
Build Identifier: Mozilla/5.0 (Windows; U; Win98; sl-SI; rv:1.7.8) Gecko/20050511 Firefox/1.0.4

The adding process goes okay, indicating there are 14 articles being verified,
but when displaying the RSS feed only the first article from the XML is visible
(regardless of date). New articles download normally. Deleting and re-adding has
no effect. RSS feed is validated and perfectly valid.

Reproducible: Sometimes

Steps to Reproduce:
1. Add http://dustworld.dyndns.org:32500/RSS+XML
2. View articles

Actual Results:  
Only first item of the XML in the RSS feed displayed.

Expected Results:  
Displaying all of the articles from the feed.
this feed works for me on the trunk. which thunderbird version are you using?
I can confirm this bug with Thunderbird 1.0.2 (DE)
I have been able to consitently reproduce this bug on 2 windows xp pro machines, Thunderbird 1.0.7.  The original feed cited is no longer valid, but I can reproduce this bug with any valid feed from my windows 2003 machine, iis 6. I can even copy a the source of 3rd party feeds which are currently displaying correctly in Thunderbird, into an xml file on my server, and it causes the same problem. This leads me to suspect it is some kind of encoding problem in iis6 which is triggering this issue. Here is an example feed, which validates fine on feedvalidator.org:

http://www.thegreensheet.com/rss.aspx?class=335&city=hou

Again- I'm having the exact same issue. When I add this feed, the counter shows that it is downloading the correct number of items, but when I click on the feed to view the items, only the first item appears in the message list.  Feeds from other sources show all items correctly.
I too have an IIS, and Thunderbird 1.5. feedvalidator.org says valid for my self-programmed rss 2.0 feed. But it only loads the first article into the list.

My first assumption was that thunderbird does more restrictive checks on the feed and stops after the first block. But after throwing out everything unneeded (optional rss elements), it still didn't work. 

I'm trying the newest nightly build now (hope it won't destroy my data).
Nah, doesn't work either version 1.5 (20051214). I'm open to suggestions and beta-tests. (And yes, the feed works in e.g. FeedReader)
The problem for my particular feed was resolved when I realized that the item link is treated as a unique field.  I had been using a generic link. When I added an id to the querystring of the link, the problem was immediately resolved.  I'm not sure if this applies to all the feeds I had tested, but this did resolve any immediate problems I was having. Doh!!
Comment on attachment 207977 [details]
RSS feed that shows the bug on my machine

I have the same problem. Here is my configuration:
Windows XP SP2 w/ Thunderbird 1.0.7 EN
The RSS feed comes from the Serendipity blogging engine. It generates (at least it seems) valid XML. The main RSS feed works correctly, but the RSS comment feed is broken.
Google Reader (www.google.com/reader) can read both RSS normaly, but Thunderbird only shows the latest entry. 
A friend of mine using 1.0.6 FR also has the same problem.

Link to the broken RSS feed:
http://tibosoft.dnsalias.org/blog/index.php?/feeds/comments.rss2
If <link> element is the same for more than one <item> elements, only one item is shown in Thundebird, in spite of presence of different values of <pubDate> and <guid>

Using TB 1.5 on Linux (so please change OS from Windows 98 to All)
I have this problem with several feeds. I'm using Thunderbird 1.5.0.2 with Windows XP Home Edition. I'm also trying to use it for podcasts. It seems to consistently show only the first file on feeds that have only one webpage even if they have multiple articles or mp3 files. I also have this problem inconsistently on other feeds.
Always have the problem with: http://www.npr.org/rss/podcast.php?id=35
and http://www.enn.com/player2/enn_radio_podcast.php, which counts through about 27 downloads but then only shows one file.
Similarly, I can't get updates from: http://www.loe.org/podcast.rss which only has one mp3 file at a time, but won't update.
I have also had this problem intermitently with: http://www.enn.com/player2/podcast/ennradiopodcast.xml 
and some craigslist RSS feeds for searches, although both of these are currently working.
(In reply to comment #10)
Actually, now that I've gone through more of the bug list, I think this the same as bug 264482.
(In reply to comment #11)
> (In reply to comment #10)
> Actually, now that I've gone through more of the bug list, I think this the
> same as bug 264482.
Yeah, giving my some GUIDs worked for me too. Maybe TB should generate internal GUIDs by date/subject? Sadly, not all feeds are conform, but TB should be able to handle those if they are not too broken.
Here's another example of the problem:
http://www.ci.austin.tx.us/qact/qact_rss.cfm

I assume it's because there is no GUID.  

TB 1.5.0.5 (20060719)
XP 5.1.2600 SP2 Build 2600
Summary: Thunderbird downloads only first article of valid RSS feed → Thunderbird downloads only first article of valid RSS feed when no GUID defined
Status: UNCONFIRMED → NEW
Ever confirmed: true
QA Contact: rss
Flags: wanted1.8.1.x?
Flags: wanted1.8.0.x?
Flags: wanted-thunderbird3?
Flags: blocking1.9.0.3?
Flags: blocking1.9.0.2?
Flags: blocking1.8.1.17?
Flags: blocking1.8.0.15?
Flags: blocking-thunderbird3?
Flags: blocking-thunderbird3.0b1?
 I can reproduce this bug in Thunderbird 2.0.0.16 and in the Thunderbird 3.0a1
pre nightly (21 August 2008) with the feed from Nature
<http://www.nature.com/news/rss/all_index.rdf>.

 Interestingly, in Firefox 2.0.0.16 the live bookmark of the feed also shows
only the first item /but/ Firefox 3.0.1 will show all items in the live
bookmark. Perhaps something was changed/fixed in Firefox 3 (a cursory search of
bugzilla didn't turn up anything obvious)?
Please do not randomly request blocking flags, certainly not all of them at the same time, one is enough to nominate it for the next release and drivers can always set flags appropriately.

Thunderbird and Firefox currently have different back-end code so that change is most likely due to a Firefox fix.
Assignee: mscott → nobody
Flags: wanted1.8.0.x?
Flags: blocking1.9.0.3?
Flags: blocking1.9.0.2?
Flags: blocking1.8.1.17?
Flags: blocking1.8.0.15?
Flags: blocking-thunderbird3?
Flags: blocking-thunderbird3.0b1?
Flags: blocking-thunderbird3.0b1-
Flags: blocking-thunderbird3-
Keywords: helpwanted
OS: Windows 98 → All
Hardware: PC → All
As is usually the case, this bug seems to have suffered from a lot of piling-on with things that are other bugs.

Impossible to say what was up with the feed from comment 0, since it's gone and was apparently a problem in 1.0.x only.

As comment 6 says, the feed from comment 3 *was* what this bug has apparently become about: if there's no <guid> in your RSS 2.0 feed, then we expect the <link> to be what identifies a post, so that when the author edits the title, or changes the post date, we don't treat something that's already been seen as a new post.

The attachment from comment 8 works fine for me on the trunk, and since it has <guid>s, is not what this bug is about; nor are the feeds in comment 10, though the feed in comment 13 is.

jwq: I can reproduce your problem with the Nature feed, but that also is not this bug (it's an RSS 1.0 feed with what certainly appear to be unique <link>s and rdf:abouts for each item) - if you didn't already, please file that as a new bug.
Flags: wanted1.8.1.x?
Summary: Thunderbird downloads only first article of valid RSS feed when no GUID defined → Thunderbird displays only first article of RSS feed when all item <link>s are the same and no <guid> is present
Version: unspecified → Trunk
Attached file Minimal testcase
Fixing this isn't likely to be very straightforward: not only is pubDate not a perfect indication of identity, it's also optional (as is title) and it's not uncommon to change it (or to change the title) for a post that's still "the same thing" so you can only use either one or both as identity for feeds in this situation, where all the links are the same.

If someone wants to play around with testcase feeds that have identical links but different 1) titles, 2) pubDates, 3) descriptions to see what the big online readers and any other client readers you have around do with them, that would be helpful.
Attachment #207977 - Attachment is obsolete: true
I get this behaviour on a feed that has items like:

    <item>
      <title>app-accessibility/gok-2.24.0</title>
      <description>                      Added to tree:&lt;br/&gt;
                      2008-10-11&lt;br /&gt;
                      Open bugs:&lt;br /&gt;
                      &lt;a href=&quot;https://bugs.gentoo.org/buglist.cgi?quicksearch=gok&quot;&gt;
                      Quick Seach gok on bugs.gentoo.org
                      &lt;/a&gt;

</description>
      <pubDate>Tue, 11 Nov 2008 02:00:00 +0200</pubDate>
      <guid isPermaLink="false">app-accessibility/gok-2.24.0:1::gentoo/20081011</guid>
    </item>

So there is no link item at all (standard says it's optional) as the items are supposed to be complete themselves. Thunderbird 2.0.0.17.
Component: RSS → Feed Reader
Product: Thunderbird → MailNews Core
I noticed the same problem with the gimp-news-feed: http://www.gimp.org/news.rdf

I'm using thunderbird 7.0.1 on a WinXP SP3 machine. I already tried to create a new "blog & news"-account and got the same result. While adding the feed thunderbird shows that it imports 21 items (downloading feed xx from 21) but I can see only the oldest.

Importing the feed into opera shows me all items.
The content of the link-tag is allways the same link to gimp.org and there is not guid-tag.
TB 10, win7 x64
I have a rss feed where occasionally some items have only <title> and <description> tags. No <link>, no <guid>, no <time>, only plain title and contents of a article.
These items were creating duplicates on each rss check.
Checking source, I found the place where rss was parsed and added a final touch - if parser could not get id for an item, it substitutes it with title or description. I am not a programmer so this is the best I could do for myself.
Thinking about it, rss parser should do this final check - if it has no tags to get unique ID from then it should generate some kind of hash based on title and/or description. When later the feed is checked and parsed, this id would be the same for unchanged items and they would not be shown as 'new'. However, any other item with different title and/or changed description would get a new hash and show in feeds as a new item.
That's a suggestion I hope someone would read and consider for implementing.
RSS 2.0 specs says that *all* tags in <item> are optional except either title or description must be present. That's why online feed validators could say the feed is OK while TB wouldn't be able to parse it correctly.
There seem to be a number of problems with RSS feeds.  Is there a tracking bug covering all of them?
(In reply to David E. Ross from comment #25)
> There seem to be a number of problems with RSS feeds.  Is there a tracking
> bug covering all of them?

why would any tracking bug be needed?  feed issues should be filed in this one component, and you can watch all bugs via watching the component
Status: NEW → RESOLVED
Closed: 11 years ago
Resolution: --- → DUPLICATE
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: