288110 - Some RSS items cause a corruption in the message display window

Status: RESOLVED FIXED

(MailNews Core :: Feed Reader, defect)

Description

[Sahab Yazdani]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.5) Gecko/20041109 Firefox/1.0 (MOOX M3)
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.5) Gecko/20041109 Firefox/1.0 (MOOX M3)

On this particular RSS feeds, whenever the author posts a picture on her blog,
the RSS reader in Thunderbird puts out a corrupted message display.  This
shouldn't happen and should instead display it normally as it would appear in
Firefox.

Reproducible: Sometimes

Steps to Reproduce:
1. subscribe to http://www.deadjournal.com/users/punky_inky/data/rss using
Thunderbird
2. wait for the author to post a picture on her blog
3. try to read that posting using Thunderbird

Actual Results:  
The message display pane in Thunderbird's output is the message formatted in a
very peculiar way.

Expected Results:  
It should display similar to all the other messages.

Comment 1

[Sahab Yazdani]

Attachment: The current RSS feed which contains an offending item.

This attachment contains the RSS feed from the user's blog site which
demonstrates the broken rendering. The item labeled "tis me. <br /><br /> ...."
does it.

Comment 2

[Sahab Yazdani]

Attachment: What Thunderbird displays on my computer for the offending test case

This is the output that Thunderbird is giving me just in case the RSS feed
doesn't work for people for some reason.

Comment 3

[Sahab Yazdani]

Sorry for spamming, but I forgot to mention that I am using the recently
released Thunderbird 1.0.2 and that the output of the message pane snould look
like this link: http://www.deadjournal.com/users/punky_inky/394037.html#comments

Comment 4

[Mike Cowperthwaite]

The feed works, but that 'tis me' item is no longer available.  I found another 
item with a posted JPEG ("omg it <i>wasn't</i> april fool's"), but that one 
displayed normally.

Comment 5

[Sahab Yazdani]

(In reply to comment #4)
> The feed works, but that 'tis me' item is no longer available.  I found another 
> item with a posted JPEG ("omg it <i>wasn't</i> april fool's"), but that one 
> displayed normally.
Hmm, that one seems to be gone, but there is another one now.  She posted on:
April 3rd at 10:05.  It doesn't have a subject since it is just two seperate
pictures.  Once again, I've attached a screenshot of what it looks like on my
system to demonstrate the problem.

Comment 6

[Sahab Yazdani]

Attachment: Better example of what Thunderbird displays on my computer for the offending RSS feed.

Comment 7

[Mike Cowperthwaite]

(In reply to comment #6)
> Better example of what Thunderbird displays on my computer for the offending
> RSS feed.

A "better" example?  The original screenshot showed the picture displayed with a 
heavy pixelation effect that definitely looked nothing like the same image 
loaded from the website.  In the screenshot you just posted, the image looks OK; 
but, there is the weird CSS blurb stuck in the middle there, and instead of the 
second picture, you have that scrolling thing that looks like an <iframe>.  
These are two distinct problems.

I loaded the feed up again, but when I view that item, I see the two pictures 
and no CSS blurb.  Assuming you still have that item in your RSS folder, please 
save it to disk (as a .EML file) and attach the .EML to this bug.


PS: In the future, rather than posting a screenshot of the entire 3-pane window, 
and going to the effort of blurring it out, why don't you instead crop the 
screenshot, or capture a standalone window instead of the 3pane.

Comment 8

[Sahab Yazdani]

> A "better" example?  The original screenshot showed the picture displayed with a 
> heavy pixelation effect that definitely looked nothing like the same image 
> loaded from the website.  In the screenshot you just posted, the image looks OK; 
You are absolutely right and I am sorry about that original misleading JPEG
attachment.  The CSS blurb thing that I demonstrated in the second JPEG
attachment is the problem that I wanted to report.  The first one also
demonstrated the CSS problem, so I've also attached the EML file for that one.

> but, there is the weird CSS blurb stuck in the middle there, and instead of the 
> second picture, you have that scrolling thing that looks like an <iframe>.  
> These are two distinct problems.
> 
> I loaded the feed up again, but when I view that item, I see the two pictures 
> and no CSS blurb.  Assuming you still have that item in your RSS folder, please 
> save it to disk (as a .EML file) and attach the .EML to this bug.
Will do!

> 
> 
> PS: In the future, rather than posting a screenshot of the entire 3-pane window, 
> and going to the effort of blurring it out, why don't you instead crop the 
> screenshot, or capture a standalone window instead of the 3pane.
Sorry about that.  I will remember that helpful little hint :-)

P.S. Once again, sorry about the confusion.  Can I use a "I'm new here" just
this once? ;-)

Comment 9

[Sahab Yazdani]

Attachment: The EML file corresponding to JPEG attachment #179550.

Comment 10

[Sahab Yazdani]

Attachment: This is the EML file corresponding to the original JPEG attachment #178896

Comment 11

[Mike Cowperthwaite]

Attachment: Screen shot of sample

This is what I see with attachment 179586 [details] loaded into my install of TB.
You seem to have something else going on, that's blocking the background page.
The screenshot you attached appears as if you're using "Show article summary",
but the message corresponding to it is a regular, show-the-web-page item.  And
View | Message Body As | Plain	suppresses the load of the images.  So what are
you doing?  

Do you have any extensions installed that might be interfering?  If so, try
running TB in Safe Mode and see if the problem persists.


> sorry about the confusion.

Don't sweat it.  At least you figured out how to use attachments, which is much
better than pasting message source into a comment.

Comment 12

[Sahab Yazdani]

> This is what I see with attachment 179586 [details] [edit] loaded into my install of TB.
Yes, that is what it should like I suppose.

> You seem to have something else going on, that's blocking the background page.
> The screenshot you attached appears as if you're using "Show article summary",
> but the message corresponding to it is a regular, show-the-web-page item.  And
> View | Message Body As | Plain	suppresses the load of the images.  So what are
> you doing?  
I checked around my settings, and nothing out of the ordinary seems to be there.
 Namely, "Show the article summary instead of loading web-page" is OFF and "View
| Message Body As |" is set to "Original HTML".  Then again, I haven't really
played with the settings in Thunderbird much, just changed the colours of the
labels around since I installed it.
> 
> Do you have any extensions installed that might be interfering?  If so, try
> running TB in Safe Mode and see if the problem persists.
I have "QuoteCollapse" installed, but I tried in Safe Mode as well and the same
results.

I asked the author of the blog to subscribe to her own feed using Thunderbird
and she says that the post corresponding to attachment #179550 [details] is corrupted like
my display: "picture, weird CSS, frame".
I also tried it out on one of the university machines that runs Thunderbird 1.0
and ran into the same problem.

I took a look at the internals of the EML files and I believe I've found the
offending error in it.  Namely if you look at attachment #179550 [details], the TITLE tag
in it contains an opened IMG tag that never closes (line 16):
>     <title><img
src="http://individual.utoronto.ca/punky_inky/PICTURES/Me/Jpegs/me17Smaller.jpg"><br
/><br /><img src="http://individual.utoronto.ca/punky_inky/P</title>

if I remove that bad IMG tag and make the line:
>     <title><img
src="http://individual.utoronto.ca/punky_inky/PICTURES/Me/Jpegs/me17Smaller.jpg"><br
/><br /></title>

it will render properly.

If there is need, I can attach the DIFF or a "fixed" version of the EML file.

Comment 13

[Mike Cowperthwaite]

(In reply to comment #12)
> I took a look at the internals of the EML files and I believe I've found the
> offending error in it.  Namely if you look at attachment #179550 [details] [edit], the
> TITLE tag in it contains an opened IMG tag that never closes (line 16):

(for brevity:)

     <title><img src="..."><br /><br /><img src="...</title>
           tag not closed here----------------------|


Nice catch!  Oddly enough, I see the same thing in the version on my system, but 
it renders correctly.  I wonder what the difference is.

Confirming bug based on details of report, altho I can't reproduce it and don't 
know why.

Probably the correct thing to do is, when creating that item, the text in the 
<title> needs to be escaped with &gt; and &lt;  (etc.).

Comment 14

[Robert Sayre]

Attachment: strip tags before assigning item/description to title

We need to strip tags before we excerpt a description from RSS feeds that don't
have titles. That's what was happening here. Another feed with messed up titles
was http://scripting.com/rss.xml.

Comment 15

[Scott MacGregor]

thanks a lot Robert.