feedburner.com May 27, 2010 10:28 AM   Subscribe

Metafilter's rss feed fails the validator. I've seen feedburner.com feeds often fail on mobile devices in the past, well feedburner.com often does weird stuff.
posted by jeffburdges to Bugs at 10:28 AM (41 comments total)

Yeah, well this page isn't valid either.
posted by blue_beetle at 10:59 AM on May 27, 2010


Is the feed actually failing, or is this a pre-emptive check?
posted by scottreynen at 11:00 AM on May 27, 2010


I will say that I think you have a valid point, jeffburdges; but if I'm honest, I'm only saying that because I'm a bit hungry.
posted by koeselitz at 11:04 AM on May 27, 2010


This is something I would have to have a cool phone to know about, isn't it.
posted by Mister_A at 11:54 AM on May 27, 2010


I'm an old woman and I feel I require an explanation.
posted by Elmore at 12:11 PM on May 27, 2010 [13 favorites]


DennisFilter.
posted by Mister_A at 12:24 PM on May 27, 2010


Does this mean Matt & co can't display one of those self-satisfied "validates" badges somewhere on the page?
posted by maxwelton at 12:35 PM on May 27, 2010 [1 favorite]


Invalid?

YOU'RE INVALID! You're invalid! This entire thread is invalid! They're invalid! *points at moderators while being dragged from the thread*

*voice diminishing into the distance* Hold it! Hold it! I just completed my opening snark!
posted by BeerFilter at 12:53 PM on May 27, 2010


pb does a metric shit ton of filtering on user input, but every week or so there's one new extended character pasted from a Word document we've never filtered for, and he has to add it to the filter.
posted by mathowie (staff) at 12:56 PM on May 27, 2010 [1 favorite]


The validator seems to be complaining about things other than HTML entities this time, though. Specifically, "undefined channel element", "implausible date", "self-reference doesn't match document location", and "unregistered link relationship".

That said, my apparently very fussy RSS reader doesn't seem to mind any of this.
posted by FishBike at 1:15 PM on May 27, 2010


I'm glad I'm not the only one using "metric shit ton". People around here look at me funny when I say it, but it will catch on; I'm sure of it.
posted by mr_crash_davis mark II: Jazz Odyssey at 1:16 PM on May 27, 2010


implausible date

Man, I think we've all been on some of these.
posted by Skot at 1:18 PM on May 27, 2010 [4 favorites]


Feed validators are like emo kids. They're never happy.
posted by inmediasres at 1:19 PM on May 27, 2010 [2 favorites]


Feeds don't fail me now!
posted by dirtdirt at 1:21 PM on May 27, 2010 [2 favorites]




The date thing is because it's DST now and the time zone offset is wrong. It should be -0700 for PDT. So it thinks those posts are in the future.
posted by smackfu at 1:51 PM on May 27, 2010


(And if I pull up the feed in Firefox, those newest entries do show up in the future, in my time zone.)
posted by smackfu at 1:51 PM on May 27, 2010


This thread is making me hungry.
posted by Elmore at 2:00 PM on May 27, 2010 [1 favorite]


The posts are coming from inside the future!
posted by owtytrof at 2:20 PM on May 27, 2010 [1 favorite]


I hate feedburner with the fire of a thousand suns, but I understand that sometimes self hosting a feed can be a pain when there are a lot of people hitting it constantly.
posted by Rhomboid at 2:26 PM on May 27, 2010


You know what I do when I get feedback like this? I delete it. I mean, what other choice do I have? I don't have a pb. I'm not even sure where you get such a thing. I mean, I'd love to snap my fingers and have a pb filter for extended characters (I'm kinda surprised that this hasn't taken care of The Whelk), but every time I try I get nothing!
posted by cjorgensen at 2:38 PM on May 27, 2010 [1 favorite]


pb does a metric shit ton of filtering on user input, but every week or so there's one new extended character pasted from a Word document we've never filtered for, and he has to add it to the filter.

When I was using handrolled blog software, I would have to do this every few weeks after noticing the LJ feed of my blog just plain stopped. I figured the right thing to do would be to find some PHP module that validates or cleans UTF-8 and run that on the input. But then I eventually just gave up and went to Wordpress.

You guys could do it, though! Clean the input, that is, not go to Wordpress.
posted by ignignokt at 2:54 PM on May 27, 2010


That's a new one to me. Looks like feedburner is inserting that tag (feedburner:browserFriendly) that chokes the validator. The rest of the errors are "warnings", meaning that we could fix them to enjoy the widest possible compatibility with feed readers. Which I translate as, "you could fix this if you're obsessive, but you should be ok." Which I'm ok with. This shouldn't break in anyone's newsreader--I think this is an issue with the validator more than anything.
posted by pb (staff) at 2:58 PM on May 27, 2010


I have no idea what anyone is talking about, but I hope UTF-8 is pronounced "oot-fate."
posted by Metroid Baby at 3:00 PM on May 27, 2010


Validator X. I'm the one and only validator. I'm the one and only validator.
posted by Babblesort at 3:02 PM on May 27, 2010 [1 favorite]


Looks like the domain that hosts the feedburner namespace is down, here. That's probably why the validator can't validate those elements. When feedburner gets that fixed up, everything should validate again.
posted by pb (staff) at 3:12 PM on May 27, 2010


The EAsports-meter told me I had an oot-fate of 'LOL'. How the hell am I gonna shed all those meme-troll-GaGanians, Lord Vader Christ-Cat?
posted by Elmore at 3:23 PM on May 27, 2010


I can see this being something my kids say for real, hopefully for the lulz (their lulz, not my lulz).
posted by Elmore at 3:24 PM on May 27, 2010


According to the release version of feedvalidator.org, it is a valid feed with only one warning: see here.

The warning is that the self reference for feeds.feedburner.com/Metafilter does not match feeds2.feedburner.com/Metafilter. Though it looks like both exist.
posted by artlung at 3:27 PM on May 27, 2010


Oh, and I meant to say, the posted validation link was to beta.feedvalidator.org
posted by artlung at 3:28 PM on May 27, 2010


mathowie: "9pb does a metric shit ton of filtering on user input, but every week or so there's one new extended character pasted from a Word document we've never filtered for, and he has to add it to the filter"

Maybe I'm being naive here, but why can't you do something like iconv? You wouldn't actually be converting between encodings, just cleaning up the UTF-8. With "//TRANSLIT//IGNORE" it will transliterate what it can, and ignore (drop) whatever remains.
posted by goodnewsfortheinsane at 5:12 PM on May 27, 2010


Weird. I recognize all the words in goodnewsfortheinsane's comment, but when I put them all together my brain goes "Wuhhh?"
posted by misha at 5:45 PM on May 27, 2010


> think this is an issue with the validator more than anything

Quis validates ipsos validator?
posted by FfejL at 6:15 PM on May 27, 2010


pb does a metric shit ton of filtering on user input

THIS IS AMERICA NONE OF THIS METRIC STUFF PLEASE K THANX
posted by inigo2 at 6:44 PM on May 27, 2010


I hate feedburner with the fire of a thousand suns, but I understand that sometimes self hosting a feed can be a pain when there are a lot of people hitting it constantly.

You know what else can be a pain when there are a lot of people hitting it constantly?
posted by sveskemus at 2:48 AM on May 28, 2010


Yes, I've got an 'emo' rss client that pukes on all feedburner feeds. I'm aware that feedburner has various options, which I cannot find documented anyplace. I've once found one that enabled adding the mefi feed, which I cannot find documented anyplace now, but then the rss reader eventually crashed the phone.
posted by jeffburdges at 5:24 AM on May 28, 2010


Back on the topic of invalid HTML entities breaking RSS feeds from time to time, do the powers that be want us to report these whenever we notice them?
posted by FishBike at 5:35 AM on May 28, 2010


Sure, FishBike. Send me a MeFi Mail if you spot 'em.
posted by pb (staff) at 6:45 AM on May 28, 2010


My suspicion is that invalid RSS is not like invalid HTML. Browsers, as we know, are laboriously engineered not to choke on invalid HTML, even the most hideous kind where a functioning document tree is all but impossible to derive. But RSS feeds are – ostensibly – strict XML, and strict-XML error-handling rules apply.

Now, I would expect that in-browser RSS readers would apply permissive error handling, not draconian, because they’re already doing that for the decoding of HTML content that is their day-to-day activity. But I would also expect that legitimate free-standing RSS readers default to draconian error handling because it is much easier to program for what are usually one-man shops producing this software.

Hence I think RSS validation errors are like CSS validation errors: They’re usually small but have significant impact and should almost always be fixed. All my handmade RSS feeds validate absolutely, but apparently my WordPress feed has minor errors I don’t know how to fix.
posted by joeclark at 10:09 AM on May 28, 2010 [1 favorite]


Don't filter for bad stuff. Only allow goodstuff. [a-Z][0-9][~!@#$%^&*()_+{}|:"<>?[]\;',./]

posted by blue_beetle at 2:27 PM on May 31, 2010


Come on. Accented characters in Latin scripts, or text in other scripts altogether would not be allowed under your suggestion. Please.
posted by goodnewsfortheinsane at 7:11 PM on May 31, 2010


« Older PDX: In Ur City, Drinking Ur Beerz   |   Another Birmingham Baseball Meet-Up Newer »

You are not logged in, either login or create an account to post comments