Closed
Bug 275153
Opened 21 years ago
Closed 21 years ago
apparently-valid rss feed will not verify in TB (derstandard.at)
Categories
(MailNews Core :: Feed Reader, defect)
MailNews Core
Feed Reader
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: gottfried, Assigned: mscott)
References
Details
(Keywords: fixed-aviary1.0)
Attachments
(4 files, 3 obsolete files)
3.23 KB,
patch
|
Details | Diff | Splinter Review | |
5.22 KB,
patch
|
mscott
:
superreview+
|
Details | Diff | Splinter Review |
29.29 KB,
image/jpeg
|
Details | |
20.02 KB,
text/xml
|
Details |
i have problems with loading the rss feed from derstandard.at. there are couple
of them available and a list of the rss feeds can be found at
http://derstandard.at/standard.asp?channel=WEBSTANDARD
after subscribing e.g. http://derStandard.at/?page=rss&ressort=Newsroom
thunderbird tries to verify the rss feed, but it never succeeds in doing this.
it seems that thunderbird hangs somewhere and the only way is to cancel the
subscription process.
Comment 1•21 years ago
|
||
On testing, I see the same thing: TB 1.0, Win2K.
Note that subscribing to one of those feeds using the Firefox "live feed"
feature seems to work just fine.
Status: UNCONFIRMED → NEW
Ever confirmed: true
Hardware: Macintosh → All
Summary: derstandard.at rss feed → apparently-valid rss feed will not verify in TB (derstandard.at)
Comment 2•21 years ago
|
||
Two issues: download callback and invalid date fields.
1.) The callback was failing when the host name contained capital letters. This
fix uses the normalized value from XmlHttpRequest instead of the RDF resource
value, and now finds the correct feed in gFzFeedCache.
2.) The date field was invalid, resulting in dates at the beginning of the
epoch (1969...). This patch checks for a valid date, and if invalid, sets it to
the current time.
Comment 3•21 years ago
|
||
the first regex wasn't catching leading garbage, now it must match from the
beginning. I'm already trimming whitespace before sending.
Attachment #171418 -
Attachment is obsolete: true
Comment 4•21 years ago
|
||
bug278694 had a feed sending a timestamp. this adds a routine to check for
that, and uses the resulting date if it seems sane (less than three years old).
Attachment #171446 -
Attachment is obsolete: true
Assignee | ||
Comment 5•21 years ago
|
||
Comment on attachment 171539 [details] [diff] [review]
add timestamp catching routine, for bug278694
Thanks for making this patch. I think the following change is going to cause
problems:
- gFzFeedCache[this.url] = this;
+ var uri = this.request.channel.originalURI.spec;
+ gFzFeedCache[uri] = this;
I'm pretty sure this is going to break code later on where we look up a URI in
the feed cache to see if a download is already in progress (Bug #267139).
Particularly the cases where we later do:
gFzFeedCache[feed.url] = "";
and
if (gFzFeedCache[this.url])
return; // don't do anything, the feed is already in use
the date change looks ok though if you want to make a separate patch with just
that for now I'll drive it into the tree.
Comment 6•21 years ago
|
||
OK, this patch contains the dates only.
At least the URI problem is correctly diagnosed now. Perhaps we could implement
a couple methods on gzFeedCache instead of using a simple Object?
Attachment #171539 -
Attachment is obsolete: true
Comment 7•21 years ago
|
||
Assignee | ||
Comment 8•21 years ago
|
||
the date patch has been checked in. Thanks for the patch. It's great to have
someone working on the RSS stuff.
Comment 9•21 years ago
|
||
*** Bug 261013 has been marked as a duplicate of this bug. ***
Comment 10•21 years ago
|
||
*** Bug 278694 has been marked as a duplicate of this bug. ***
Comment 11•21 years ago
|
||
Assignee | ||
Comment 12•21 years ago
|
||
Comment on attachment 171815 [details] [diff] [review]
normalize host name in feed cache
I'll check this in as this patch looks good to me and it passes my regression
tests.
I can't say how exciting it is to have someone pushing the RSS stuff again :)
Keep up the good work.
Attachment #171815 -
Flags: superreview+
Assignee | ||
Comment 13•21 years ago
|
||
fixed
Status: NEW → RESOLVED
Closed: 21 years ago
Keywords: fixed-aviary1.0
Resolution: --- → FIXED
Target Milestone: --- → Thunderbird1.1
Assignee | ||
Comment 14•21 years ago
|
||
Robert, I think we may have broken date handling with this patch. I'm noticing
that the date for feed entries is always the date I downloaded the feed and not
the date the article was first written like it used to be.
For instance:
http://weblogs.mozillazine.org/ben/index.rdf
We don't seem to be extract <dc:date> anymore for the articles.
Assignee | ||
Comment 15•21 years ago
|
||
actually I broke the date field as part of my code cleanup for Bug #279150
Comment 16•20 years ago
|
||
There is an inconsistant display of dates in the messages vs. display in the
'Date' Column in the folder.
The feed in question is http://www.php.net/news.rss
They have an item that refers to an archive of old stories, and they have dated
it using dc:date as 1969-12-31. The XML is as follows:
<item rdf:about="http://www.php.net/news-2004.php">
<title/>
<link>http://www.php.net/news-2004.php</link>
<description>News Archive</description>
<dc:date>1969-12-31</dc:date>
</item>
<!-- / RSS-Items PHP/RSS -->
This is the last item of the feed.
I haven't been able to test the bug in any CVS / Development version ... I am
using version 1.0.2 (20050317) on Windows XP, SP2.
The issue is this:
The display of the message header (in the message preview portion) is correct,
it shows '12/30/1969 7:00 PM'. However, the Date column of the folder shows the
message dated '2/5/2106 1:28 AM'. So, instead of setting this item as the oldest
item, it will be the newest item for the next 100 or so years ...
I'll upload a copy of the rss feed, plus a small screenshot.
If this has already been fixed in code, my apologies. Otherwise, this could make
a good test case.
Comment 17•20 years ago
|
||
A screenshot of the date display inconsistancy.
Comment 18•20 years ago
|
||
Component: RSS → Feed Reader
Product: Thunderbird → MailNews Core
Target Milestone: Thunderbird1.1 → ---
You need to log in
before you can comment on or make changes to this bug.
Description
•