Support for Drupal 7 is ending on 5 January 2025—it’s time to migrate to Drupal 10! Learn about the many benefits of Drupal 10 and find migration tools in our resource center.
I'm trying to figure out what I did that might've caused this (possibly deleting and refreshing feed items?) but now when I save my feed node, it instantly forgets the Feed URL, causing it not to be able to add/remove any feed items.
Comment | File | Size | Author |
---|---|---|---|
#10 | feedapi-help-avoid-face-bashing-377382-10.patch | 2.89 KB | webchick |
#9 | validate_turned_off_submodule.patch | 1.77 KB | Aron Novak |
#5 | feedapi-trap-missing-parser-377382-5.patch | 1002 bytes | webchick |
#4 | var_dump.txt | 107.54 KB | webchick |
Comments
Comment #1
Aron NovakCan you paste here the feed URL (but i assume this is not feed-specific problem) and the value of the feedapi_settings_[yourcontentypename] variable?
Do you use the content-type that feedapi creates you at the install?
Comment #2
geodaniel CreditAttribution: geodaniel commentedI've found that this tends to happen if FeedAPI can't fetch the feed on first attempt, especially when there are potential issues with certificates on an https connection, etc.
Comment #3
webchickUgh, I keep getting completely inconsistent results. :\
Here's (roughly) what I did:
* Switched back to Common Syndication Parser (disabled SimplePie).
* Deleted all of my feeds from yesterday thinking something was on crack and it was best to start over.
* Read at #358510: Twitter's user timeline rss feed behaves strangely that Twitter feeds behave stupidly when you request the same feed twice.
* Created a *new* feed with that issue author's twitter feed (http://twitter.com/statuses/user_timeline/17000175.rss) which I definitely have not retrieved before.
* For extra precaution, I checked the box to retrieve the items upon feed creation, so that there was (hopefully) no chance of pinging the feed twice.
* When I'm returned to the feed node, there is *no* "link to this site" link, and when i edit it, my feed URL field is blank. Clicking the Refresh tab doesn't do anything; just refreshes the page. Clicking "Remove items" takes me to a blank page.
* No matter how many times I edit and re-save the node, it will never take my feed URL.
To answer your questions:
1. Yes, I'm using just the "Feed" type that ships with this module, feed item type is set to Story (which was the default)
2. feedapi_settings_feed is:
The feedapi table doesn't have a record in it for this node, so that's obviously the problem.
Comment #4
webchickOk, so I searched for INSERT INTO {feedapi} and the only place this is called is _feedapi_insert(). The code there does:
I changed this to:
This revealed the following:
So I think the code is failing because there is no feed type. I've attached the var_dump text file.
Comment #5
webchickOh, sheesh. Well I don't know if this is "the" problem but it's definitely "a" problem.
I switched from Common Syndication Parser to SimplePie by disabling the former and enabling the latter. However, I did *not* then go under admin/content/types/feed and check the "Enable" box on the SimplePie parser. Therefore, it was defaulting to common_syndication_parser, and thus failed the module_exists() check in _feedapi_call_parsers().
This explains why I ran into it yesterday, since I was futzing with the parsers trying to get it to parse Twitter feeds. At one point I had both the parsers enabled and then thought, "Well that's dumb. I only need one." and disabled the common one. However, since I'd never needed to set those settings at admin/content/types/feed I didn't realize they existed until I read the source code. :P~
How about some error trapping, to save the next poor schlepp 3 hours of her life? ;)
Comment #6
webchickComment #7
webchickHm. Doesn't look like that's the ideal place to put that (it fires on node creation, not on feed refreshing) but hopefully someone who knows what they're doing can put it somewhere where it would be helpful. :)
Comment #8
Aron NovakTotally right, i have never tried to trick FeedAPI with turning off an assigned sub-module ;) Now I'm going to do scrutiny on this area of FeedAPI.
Comment #9
Aron NovakHow about this patch? This utilises FeedAPI already existing validating features for this special case.
Comment #10
webchickWorks great! But let's give them even more help by linking directly to the content type page they need to configure. Also, fixing some minor English issues with the string.
Comment #11
Aron NovakI committed that improved patch, plus when the processors are misconfigured, i corrected the error message according to the parsers' message.
Comment #12
cpelham CreditAttribution: cpelham commentedI am again getting this problem. Updated to the most recent dev version of FeedAPI which should contain the patches mentioned above, right? I then tried re-building my feed, but the URL still does not stick.
This is the feed I am trying to add:
http://www.naturalnews.com/rss.xml
I am using Common Syndication Parser and my feed node type does indeed have the parser turned on.
Comment #13
Aron Novakhttp://www.feedvalidator.org/check.cgi?url=http%3A%2F%2Fwww.naturalnews....
parser common syndication is based on SimpleXML, this means that this parser is more sensitive to bogus feeds.
Comment #14
stevenovy CreditAttribution: stevenovy commentedWe were having the same issue and had also just switched from SimplePie to CSP (we're liking CSP over SimplePie). Turned out to our feed URL was the problem. We tested our URL against a known quantity like http://rss.news.yahoo.com/rss/topstories and that helped us track down the problem.
Comment #15
cpelham CreditAttribution: cpelham commentedHaving identified via feedvalidator.org that our feed was problematic, I wasn't sure what to do about it. But I have found a workaround. I went to feedburner.com and made a feed of the feed. The new feedburner feed saves and imports the feed items successfully. Others may try this approach until CSP becomes more forgiving. If anyone knows of any drawbacks to using feedburner, please let me know.
Comment #17
mattheweigand CreditAttribution: mattheweigand commentedThis is happening to me, the last in a very long line of frustrations with FeedAPI. I consider myself to be a somewhat logical guy, but the behavior of FeedAPI defies all logic. This is happening to me this morning, as I tried to recreate the steps that caused me to make everything work correctly on the test site last night.
My feed url is http://media.digitaltrends.com/feed/korea-it-times/. it contains HTML but simplepie parser was doing fine with processing it last night!! Now it doesn't parse anything and doesn't even save the URL. Same thing happens when I make a feedburner feed of that feed.
Comment #18
mattheweigand CreditAttribution: mattheweigand commentedactivate
Comment #19
cpelham CreditAttribution: cpelham commentedFeedvalidator.org reveals that there is a script tag in this feed. I can't say for sure whether this is why it has stopped working for you within FeedAPI but it might be an issue.
http://www.feedvalidator.org/check.cgi?url=http%3A%2F%2Fmedia.digitaltre...
You might try writing to the webmaster of digitalmediatrends.com and asking if they changed anything about the feed since it stopped working.
Comment #20
mattheweigand CreditAttribution: mattheweigand commentedIt works with common syndication parser. I don't know why. too frustrated and pressed for a deadline to test further.
Comment #21
alex_b CreditAttribution: alex_b commentedPlease refrain from reopening fixed issues. Rather open a new issue with a clear description of your problem and ideally, how to reproduce it. Closing...