I already asked the same question over at dev.twitter.com, however, I didn't get an answer there. So maybe someone here on SO ran into the same issue and has an answer.
In my application I count the length of the characters the user enters to compose a tweet. However, if the user enters an URL, this will be shortened automatically (by Twitter's API) when posting the tweet. So I have to replace the length of the URL with the length of the resulting t.co URL in my character counter.
However, the problem is now, what is Twitter's definition of an URL so that I know when you adapt my character counter and when not. For example www.verylongexampleurl.de gets shortened, while verylongexampleurl.de (without the www) doesn't, but verylongexampleurl.com does get shortened again.
I couldn't find any documentation, but maybe I missed it. All hints are appreciated.
Quoting from dev.twitter.com:
Need help parsing tweet text?
Take a look on the Twitter text processing library we’re using for auto linking and extraction of usernames, lists & hashtags.
Ruby: https://github.com/twitter/twitter-text-rb
Java: https://github.com/twitter/twitter-text-java
Javascript: https://github.com/twitter/twitter-text-js
The actual specification (tests) can be found here: https://github.com/twitter/twitter-text-conformance/blob/master/autolink.yml
Related
I have an API that returns URL that render an image from Adobe Scene7. The API is doubling up, or repeating, the URL in the xml return and is causing parsing issues in my presentation. Is there a technical term for this repeated URL? I need to know how to report this to my developers.
Here is an example, in the url can see where the protocol starts and repeats the exact same URL...
<d1p1:ProductImageUrl>http://s7d7.scene7.com/is/image/BBE004?&op_colorize=0f5196http://s7d7.scene7.com/is/image/BBE004?&op_colorize=0f5196</d1p1:ProductImageUrl>
I think the technical term for this is "invalid URL".
I have a page tab app that I am hosting. I have both http and https supported. While I receive a signed_request package as expected, after I decode it does not contain page information. That data is simply missing.
I verified that like schemes are being used (https) among facebook, my hosted site and even the 'go between'-- facebook's static page handler.
Also created a new application with page tab support but got the same results-- simply no page information in the signed_request.
Any other causes people can think of?
I add the app to the page tab using this link:
https://www.facebook.com/dialog/pagetab?app_id=176236832519816&next=https://www.intelligantt.com/Facebook/application.html
Here is the page tab I am using (Note: requires permissions):
https://www.facebook.com/pages/School-Auction-Test-2/154869721351873?id=154869721351873&sk=app_176236832519816
Here is the decoded signed_request I am receiving:
{"algorithm":"HMAC-SHA256","code":!REMOVED!,"issued_at":1369384264,"user_id":"1218470256"}
5/25 Update - I thought maybe the canvas app urls didn't match the page tab urls so I spent several hours going through scenarios where they both had a trailing slash or not. Where they both had a trailing ? or not, with query parameters or not.
I also tried changing the 'next' value when creating the page tab to the canvas app url and the page tab url.
No success on either count.
I did read where because I'm seeing the 'code' value in the signed_request it means Facebook either couldn't match my urls or that I'm capturing the second request. However, I given all the URL permutations I went through I believe the urls match. I also subscribed to the 'auth.authResponseChange' which should give me the very first authResponse that should contain the signed_request with page.id in it (but doesn't).
If I had any reputation, I'd add a bounty to this.
Thanks.
I've just spent ~5 hours on this exact same problem and posted a prior answer that was incorrect. Here's the deal:
As you pointed out, signed_request appears to be missing the page data if your tab is implemented in pure javascript as a static html page (with *.htm extension).
I repeated the exact same test, on the exact same page, but wrapped my html page (including js) within a Perl script (with *.cgi extension)... and voila, signed_request has the page info.
Although confusing (and should be better documented as a design choice by Facebook), this may make some sense because it would be impossible to validate the signed_request wholly within Javascript without placing your secretkey within the scope (and therefore revealing it to a potential hacker).
It would be much easier with the PHP SDK, but if you just want to use JavaScript, maybe this will help:
Facebook Registration - Reading the data/signed request with Javascript
Also, you may want to check out this: https://github.com/diulama/js-facebook-signed-request
simply you can't get the full params with the javascript signed_request, use the php sdk to get the full signed_request . and record the values you need into javascript variabls ...
with the php sdk after instanciation ... use the facebook object as following.
$signed_request = $facebook->getSignedRequest();
var_dump($signed_request) ;
this is just to debug but u'll see that the printed array will contain many values that u won't get with js sdk for security reasons.
hope that helped better anyone who would need it, cz it seems this issue takes at the min 3 hours for everyone who runs into.
I am currently working on a Flattr plugin for a popular open-source RSS reader (tiny tiny RSS).
I am using the lookup API for the first time and am unsure why I am getting mixed results.
So I'm unsure if I use the API correctly and want to confirm with you experts if I got something basic wrong.
First, let's see if I can come up with an API call that looks up a thing successfully. I look at the Flattr page of thing 1066706 (I can't post the whole URL here as SO only allows me two URLs for this whole post). On that page, I find the official URL which Flattr stores for that thing and look that up with the API:see here
This returns {"type":"thing","resource":"https:\/\/api.flattr.com\/rest\/v2\/things\/1066706", ... so that's good.
But it seems this method is not a sure way to test if things exist. Here is an example that doesn't work: I open the Flattr page of thing e7579b349cb7b319b28d883cd4064e1e.
That URL I find on that page is indeed the URL of that article and I don't see any other URL it might have. I look that up in the same way as above:check this
Alas, I get {"message":"not_found","description":"No thing was found"}
(I also tried both of these with encoded URLs, but got the same result. I figured this is easier to read for you.)
So, why would that second thing not be found? Thanks for any enlightenment.
The id "e7579b349cb7b319b28d883cd4064e1e" is not a real thing id but a hash that identifies a temporary thing for a not yet submitted thing - it's part of Flattr's autosubmit functionality: http://developers.flattr.net/auto-submit/
So the system is very correct in telling you that a thing for that URL doesn't exist - someone would need to flattr that thing for it to become submitted for real and created in the system with a real id to it.
(Just for reference - for some URL:s, like Twitter URL:s, Flattr can actually answer that the URL is flattrable even though it can't find it in the system: {"message": "flattrable", "description": "Thing is flattrable "} That way you can now that it is possible to flattr that thing without you having to use any kind of flattr-button/url supplied by the author to be able to flattr the URL)
Also - if you don't know it yet then for a RSS reader you should primarily be looking for rel-payment links to find out whether an entry is flattrable or not, see http://developers.flattr.net/feed/ and http://relpayment.com/
I have tagged over 700 blog posts with tags containing hyphens, and these tags suddenly stopped working in 2011, because Tumblr decided (without any notice) to forbid hyphens in tags (I guess hyphens are blocked now, because spaces in tags (which are allowed) get changed to hyphens.). Unfortunately, Tumblr is not willing to globally rename all tags containg hyphens (although these tags are of no use anymore → 404).
Now I want to rename my tags myself.
I tried to do it with the "Mass Post Editor" (tumblr.com/mega-editor), but it's not possible to select posts by tag. I'd have to manually select post after post and look if a certain tag was used, and if so, delete it and add a new one instead. This would be a huge job (700 tagged posts, but more than 1000 in total).
So I thought that the Tumblr API might help me. I'm no programmer, but I'd be willing to dig into it, if I could get some help here as a starting point.
I think I need the following process:
select all posts that are tagged with x (= a tag containing hyphens)
tag all these posts with y (= a tag without hyphens)
delete the tag x on all these posts
I'd start this process for every affected tag manually.
I see that the method (or whatever you call it) /post knows the request parameter tag:
Limits the response to posts with the specified tag
(I guess I can only hope that this works for tags containing hyphens, too.)
After that I'd need a way to add and remove tags from that result set. /post/edit doesn't say anything about tags. Did I miss something? Isn't it possible to add/remove tags with the API?
Have you an idea how I could "easily" rename my tags?
Is it possible with the API? Could you give me a starting point, tip etc. how I could manage to do it?
I don't know if this might be helpful, but I noticed that the search function is still able to find posts "tagged" with tags that contain hyphens.
Example: let's say I have the tag foo-bar. It is linked with /tagged/foo-bar (→ 404). I can find the posts with /search/foo-bar (but this is of course not ideal because it might also find posts that contain (in the body text) words similar/equal to the tag name).
I tried to encode the hyphen (/tagged/foo%2Dbar), but no luck.
just for the record, because this is a popular google search: i've done it! you can use it at http://dev.goose.im/tags/.
i used a combo of PHP and jquery, basing my jquery off of a previous tumblr api script i wrote a year or two ago, and used this tumblr php oauth script for the authentication. if anyone wants me to put up the source code, i'd be happy to.
If you aren't a programmer, how much is your time is worth to you? As they say, time is money. Not only do you have to figure out how to use the API, but choose a language and learn to write in it. That's no small task. You could higher a freelancer for $50 for an hour worth of work.
To answer your question, yes it is possible to do this with the API. It mentions "These parameters are used for /post, /post/edit and /post/reblog methods." and tags is mentioned as a string of comma separated words.
What you want to do is get a listing of every single blog post using the /posts method. You'll want to look at the "Request" section to figure out the criteria to pass to this URL. You want it to be as general as possible to get a complete listing of all your posts.
After you get a listing of posts you'll want to iterate over it and modify the tags parameter provided in the response for each post. You'll want to use the id paramater along with /post/edit, which again takes tags as a string.
The simplest language you can use for this task is PHP. You'll want to look at the curl extension to make your requests. You'll want to read up on arrays as you'll be using them a lot. You'll also need to look at explode, implode, str_replace (for the dashes), and foreach for iterating over the result.
When you do this I would highly recommend you use break at the end of your foreach loop so it only affects one post at first. Testing it first will be important, as you don't want to accidentally erase your tags/posts. print and var_dump are good ways to help you debug the code. xdebug is a nice extension that allows you to step through the code line by line as it runs. Netbeans is an IDE that has good xdebug support.
There's also a nice page here to get you started with PHP. You'll need to install PHP on your machine. You don't need to install a web server - for this PHP-CLI (command line) sapi is good enough.
The Red Cross has a locator page where you can submit a zip code by means of a form and that runs through a JSP to return contact information for your local Red Cross office.
From an iOS app, I have the user's zip code and would like to run this process and get that contact information back... or at least the appropriate URL to link to. I was hoping I could find some way to pass the zip code by URL but it doesn't look like it's going to be that easy.
Could anyone offer some direction as to how to go about this? I've done some simple things with forms before but I'm not quite sure where to start with this one.
According to the <form> action parameter, that's the URL you have to post your data to:
http://www.redcross.org//portal/site/en/template.MAXIMIZE/ziplocator/;jsessionid=MWwKPvjSWmvz8p4XrRtNLVL0VCTM7fcwfnFnKHpwRhJTnwLMMDcv!-1938881463!-1334769155?javax.portlet.tpst=2bd907ea326f7e9e934afa36c23f78a0_ws_MX&javax.portlet.prp_2bd907ea326f7e9e934afa36c23f78a0_viewID=result&javax.portlet.begCacheTok=com.vignette.cachetoken&javax.portlet.endCacheTok=com.vignette.cachetoken&vgnextoid=6d65e821cbdf9110VgnVCM1000002bf3870aRCRD
The field name is zipcode. The AFNetworking documentation is quite okay explaining how to send a POST request. I would also recommend the nsscreencasts series, he has an episode on AFNetworking as well.
I hope this helps. :)
You can use NSURLConnection to execute a post request and return the result, you may have to do some parsing of the result though depending on the format it is returned in.
Here is an example:
http://forums.macrumors.com/showthread.php?t=689884
NSURLConnection documentation:
https://developer.apple.com/library/mac/#documentation/Cocoa/Reference/Foundation/Classes/nsurlconnection_Class/Reference/Reference.html