[twitter-dev] Re: OAuth issues with POST
Hi Matt, Thanks for replying. I had already tried the approach of using the source=test with the signatureBase without success, I still receive 401's for all POST calls, but GETs work fine. I'm suspecting that .NET itself is doing something that perhaps other libraries don't. I'll keep trying some things, including maybe just opening a socket and building an HTTP message myself to trace back steps, but just to confirm: - The encoding of x-www-form-urlencoded is not the same escaping algorithm as URIs (too lazy to look up RFC numbers), so it should be OK for the POST params to use + for spaces and the signatureBase to use %20's for spaces, which is the technically correct way. - Do you know if other libraries, in general, look at a POST request, chop of the URI query, and send that as post parameters after the initial request data? Daniel On Apr 14, 1:55 pm, Matt Sanford wrote: > Hi Daniel, > > While working through issue 433 [1] we've discussed signatures > pretty in-depth. This is an unrelated problem but I think you'll find > the discussion about signatures there helpful, as it has several > examples. The short-short answer is: > > • The string you sign should have the status=testing parameter > appended, like: > > POST&http%3A%2F%2Ftwitter.com%2Fstatuses > %2Fupdate.xml&oauth_consumer_key%3Dw9zJ2JxgWTsJN5OyEHIZjw%26oauth_nonce > %3Ddudialjpcnn4p355%26oauth_signature_method%3DHMAC- > SHA1%26oauth_timestamp%3D1239673686%26oauth_token%3D11173402- > LxhEcmKGl2zKjsbuBE0tdt4UBAxHoR1LFfrs5tTo4%26oauth_version > %3D1.00%26status%3Dtesting > > • When you make the request the URL should > behttp://twitter.com/statuses/update.xml > and the status=testing parameter should be in the body, like a > normal x-www-form-urlencoded POST. > > Thanks; > — Matt Sanford / @mzsanford > > [1] -http://code.google.com/p/twitter-api/issues/detail?id=433 > > On Apr 14, 2009, at 09:28 AM, Dimebrain wrote: > > > > > Hello, > > > I originally commented on issue thread 447 but that issue was closed, > > so I wanted to repost my problem to see if it's something I'm doing > > wrong on my side. > > > I am still failing, but using C# / .NET and a self-authored OAuth > > implementation. > > > My GET calls work correctly, my POST calls 401. > > > Here is the URI: > >http://twitter.com/statuses/update.xml?status=testing > > > Here is my signature base string: > > POST&http%3A%2F%2Ftwitter.com%2Fstatuses > > %2Fupdate.xml&oauth_consumer_key%3Dw9zJ2JxgWTsJN5OyEHIZjw > > %26oauth_nonce > > %3Ddudialjpcnn4p355%26oauth_signature_method%3DHMAC- > > SHA1%26oauth_timestamp%3D1239673686%26oauth_token%3D11173402- > > LxhEcmKGl2zKjsbuBE0tdt4UBAxHoR1LFfrs5tTo4%26oauth_version%3D1.0 > > > And the signature (url encoded): > > E0UkQEmfcaT3DOG7a8L7sImCmVw%3D > > > And the authorization header: > > OAuth > > oauth_consumer_key="w9zJ2JxgWTsJN5OyEHIZjw",oauth_token="11173402- > > LxhEcmKGl2zKjsbuBE0tdt4UBAxHoR1LFfrs5tTo4 > > ",oauth_nonce > > = > > "dudialjpcnn4p355 > > ",oauth_timestamp="1239673686",oauth_signature_method="HMAC- > > SHA1",oauth_signature="E0UkQEmfcaT3DOG7a8L7sImCmVw > > %3D",oauth_version="1.0" > > > And the request header out the door: > > > POST /statuses/update.xml?status=testing HTTP/1.1 > > Content-Type: application/x-www-form-urlencoded > > Authorization: > > Host: twitter.com > > > - I've also tried including "status=testing" in the list of post > > parameters included > > in the signature hashing. I'm not clear on whether it's expected to > > exist there b/c > > of the normalization done to the URI during signature hashing, i.e. > > otherwise it > > would only exist on the POST itself and not in the signature > > > I wonder if someone could clarify a point about POST with status=X, > > and that is, when preparing a POST with the default W3C application/x- > > www-form-urlencoded content type, are we supposed to: > > > a) remove the status=X line from the URI query, and instead write it > > to the request as POST parameters (and therefore include status=X in > > the signature base string as per OAuth 1.0), or > > > b) keep the status=X in the URI itself, and only including other > > parameters in the POST parameters. > > > I think what might be happening is that .NET treats the Query fragment > > of a URI as post parameters, and maybe all approaches with > > application/ > > x-www-form-urlencoded do; in other words it's not really a query as > > per a GET call, it's just the absolute uri path, and then all query > > params are added to the signature base and written in the POST message > > ---> with different URL escaping than a URI (i.e. +'s are part of the > > POST encoding spec as opposed to %20's that are part of the URI > > encoding spec). > > > See, maybe you can't POST because you use %20's in the sig base but > > +'s in the POST params, which is actually according to the spec. > > > I'd love to know if anyone has some insight into this problem. The > > inability to POST w/ OAuth on Twitter is
[twitter-dev] Re: Trending Service for a given set of users
colo application, thanks! On Tue, Apr 14, 2009 at 6:33 PM, Sean Scott wrote: > I'm currently building an AIR twitter client (yes i know yet another one) > and part of the goal is to help users see what is popular (trending) within > their own community. The most popular URLS and topics information exists > for twitter at large, but sometimes what happens in your own group of > cohorts is more interesting and usefull than knowing that 15 million people > are really digging American Idol right now. > So looking for most popular URL RT, most popular topic for a given set of > users. Bonus if the service can also return the same information for the > followers of the user set. > > Hope that helps > > > On Tue, Apr 14, 2009 at 4:21 PM, Carlos Crosetti < > carlos.crose...@gmail.com> wrote: > >> Please can you explain the "trending" output you are looking for? >> >> >> On Tue, Apr 14, 2009 at 2:45 PM, kalisurfer wrote: >> >>> >>> Hi All, >>> >>> Looking for a service where provided a list of users (100+) i can get >>> back the trending URL, topics, hash and RT. Love to be able to access >>> the info via a REST API. >>> >>> Trying to not build it out myself. >>> >>> Thanks, >>> Sean >>> @kalisurfer >>> >> >> >> >> -- >> Carlos Crosetti >> > > > > -- > Sean Scott > cell: 612.867.8133 > portfolio: > http://www.flickr.com/photos/92876...@n00/sets/72157613990263453/ > profile: http://www.linkedin.com/profile?viewProfile=&key=2242610 > blog: http://www.twofortyeight.com/ > other: http://twitter.com/kalisurfer > -- Carlos Crosetti
[twitter-dev] Re: sending DM to all followers?
Your timeline update is doing that already, what is the scenario you are thinking of? On Tue, Apr 14, 2009 at 8:24 PM, Alex wrote: > > I'm wondering if there is a way - or if you would consider adding a > way - to send a DM to all followers via the API? > > Obviously we could grab the followers list and iterate over it to send > the DM to all, though that could require thousands of API calls > depending on the user. (And could therefore take hours to do with the > 100 API query/hour limit.) > > -- Carlos Crosetti
[twitter-dev] Re: Registration of OAuth applications via API
Out of curiosity. What are the use cases for automated application creation? On Tue, Apr 14, 2009 at 18:53, Dr Nic wrote: > > Currently the only way (I think) to register an application to use the > OAuth API is via the web form at http://twitter.com/oauth_clients/new > which requires the user to use login/password on the main site. To > automate this (e.g. move it into a command-line application like the > twitter rubygem) means that the login/password must be stored locally. > Which is fine, until its not. > > The maintainer of the twitter rubygem is/has migrated his code base to > be whole OAuth compatible and removed all notions of storing login/ > password locally to use for Basic Auth on the APIs. This means the > twitter rubygem cannot be extended to support registering OAuth > applications. Which is a pity. > > Can registration of OAuth applications be added to the API, please? > That would be wonderful. > > Cheers > Nic > -- Abraham Williams | http://the.hackerconundrum.com Hacker | http://abrah.am | http://twitter.com/abraham Web608 | Community Evangelist | http://web608.org This email is: [ ] blogable [x] ask first [ ] private.
[twitter-dev] Re: sending DM to all followers?
It's one way of pruning your followers ... :-) --Original Message-- From: Cameron Kaiser Sender: twitter-development-talk@googlegroups.com To: twitter-development-talk@googlegroups.com ReplyTo: twitter-development-talk@googlegroups.com Subject: [twitter-dev] Re: sending DM to all followers? Sent: Apr 14, 2009 7:49 PM > > I'm wondering if there is a way - or if you would consider adding a > > way - to send a DM to all followers via the API? > > I believe that's called "Tweeting" Indeed. Mass DMs are bad manners IMHO. -- personal: http://www.cameronkaiser.com/ -- Cameron Kaiser * Floodgap Systems * www.floodgap.com * ckai...@floodgap.com -- "Endian Little Hate We" -- credits from Connectix Virtual PC 6 for Mac - -- Dossy Shiobara do...@panoptic.com
[twitter-dev] Re: combined DM feed?
+1. Please? Is there an issue opened for this? --Original Message-- From: Alex Sender: twitter-development-talk@googlegroups.com To: Twitter Development Talk ReplyTo: twitter-development-talk@googlegroups.com Subject: [twitter-dev] combined DM feed? Sent: Apr 14, 2009 7:20 PM Was hoping you would consider offering an API call that would return a combined DM feed of both sent and received DMs. Thoughts? -- Dossy Shiobara do...@panoptic.com
[twitter-dev] Re: Registration of OAuth applications via API
> Can registration of OAuth applications be added to the API, please? > That would be wonderful. Not that I disagree with the request, cause I think it'd be a great feature, but John added the Basic Auth support back into the gem about a week later: http://railstips.org/2009/4/5/twitter-gem-reborn-with-fewer-features-and-100-more-oauth#comment-8982 Get 0.6.2 for increased shininess.
[twitter-dev] Registration of OAuth applications via API
Currently the only way (I think) to register an application to use the OAuth API is via the web form at http://twitter.com/oauth_clients/new which requires the user to use login/password on the main site. To automate this (e.g. move it into a command-line application like the twitter rubygem) means that the login/password must be stored locally. Which is fine, until its not. The maintainer of the twitter rubygem is/has migrated his code base to be whole OAuth compatible and removed all notions of storing login/ password locally to use for Basic Auth on the APIs. This means the twitter rubygem cannot be extended to support registering OAuth applications. Which is a pity. Can registration of OAuth applications be added to the API, please? That would be wonderful. Cheers Nic
[twitter-dev] Re: sending DM to all followers?
No, There is no special arrangement needed to send DMs through the API. Doug Williams Twitter API Support http://twitter.com/dougw On Tue, Apr 14, 2009 at 4:58 PM, Alex wrote: > > I understand the negative implications of doing this. The main reason > I ask is that I've had my users ask me. > > There are other sites offering such features, and their web sites > imply they have some kind of special relationship with Twitter that > allows them to do this. I'm not sure if that's true or not? > > > > On Apr 14, 7:46 pm, Doug Williams wrote: > > Alex, > > That sounds very spamish although there are certainly some use cases > where > > it is acceptable. Proceed with caution when sending mass DMs. Ensure the > > messages you are sending are relevant and of value to your followers. > > > > Doug Williams > > Twitter API Supporthttp://twitter.com/dougw > > > > On Tue, Apr 14, 2009 at 4:41 PM, Chad Etzel wrote: > > > > > I believe that's called "Tweeting" > > > -Chad > > > > > On Tue, Apr 14, 2009 at 7:24 PM, Alex wrote: > > > > > > I'm wondering if there is a way - or if you would consider adding a > > > > way - to send a DM to all followers via the API? > > > > > > Obviously we could grab the followers list and iterate over it to > send > > > > the DM to all, though that could require thousands of API calls > > > > depending on the user. (And could therefore take hours to do with the > > > > 100 API query/hour limit.) >
[twitter-dev] Re: sending DM to all followers?
I understand the negative implications of doing this. The main reason I ask is that I've had my users ask me. There are other sites offering such features, and their web sites imply they have some kind of special relationship with Twitter that allows them to do this. I'm not sure if that's true or not? On Apr 14, 7:46 pm, Doug Williams wrote: > Alex, > That sounds very spamish although there are certainly some use cases where > it is acceptable. Proceed with caution when sending mass DMs. Ensure the > messages you are sending are relevant and of value to your followers. > > Doug Williams > Twitter API Supporthttp://twitter.com/dougw > > On Tue, Apr 14, 2009 at 4:41 PM, Chad Etzel wrote: > > > I believe that's called "Tweeting" > > -Chad > > > On Tue, Apr 14, 2009 at 7:24 PM, Alex wrote: > > > > I'm wondering if there is a way - or if you would consider adding a > > > way - to send a DM to all followers via the API? > > > > Obviously we could grab the followers list and iterate over it to send > > > the DM to all, though that could require thousands of API calls > > > depending on the user. (And could therefore take hours to do with the > > > 100 API query/hour limit.)
[twitter-dev] Re: sending DM to all followers?
The only reason I ask is that there are other services that are offering it, and I am wondering how they are doing it. Their web site implies they have some kind of special arrangement with Twitter though I don't know if that's the case. On Apr 14, 7:46 pm, Doug Williams wrote: > Alex, > That sounds very spamish although there are certainly some use cases where > it is acceptable. Proceed with caution when sending mass DMs. Ensure the > messages you are sending are relevant and of value to your followers. > > Doug Williams > Twitter API Supporthttp://twitter.com/dougw > > On Tue, Apr 14, 2009 at 4:41 PM, Chad Etzel wrote: > > > I believe that's called "Tweeting" > > -Chad > > > On Tue, Apr 14, 2009 at 7:24 PM, Alex wrote: > > > > I'm wondering if there is a way - or if you would consider adding a > > > way - to send a DM to all followers via the API? > > > > Obviously we could grab the followers list and iterate over it to send > > > the DM to all, though that could require thousands of API calls > > > depending on the user. (And could therefore take hours to do with the > > > 100 API query/hour limit.)
[twitter-dev] Re: sending DM to all followers?
The only reason I ask is that there are other services that are offering it, and I am wondering how they are doing it. Their web site implies they have some kind of special arrangement with Twitter though I don't know if that's the case. On Apr 14, 7:46 pm, Doug Williams wrote: > Alex, > That sounds very spamish although there are certainly some use cases where > it is acceptable. Proceed with caution when sending mass DMs. Ensure the > messages you are sending are relevant and of value to your followers. > > Doug Williams > Twitter API Supporthttp://twitter.com/dougw > > On Tue, Apr 14, 2009 at 4:41 PM, Chad Etzel wrote: > > > I believe that's called "Tweeting" > > -Chad > > > On Tue, Apr 14, 2009 at 7:24 PM, Alex wrote: > > > > I'm wondering if there is a way - or if you would consider adding a > > > way - to send a DM to all followers via the API? > > > > Obviously we could grab the followers list and iterate over it to send > > > the DM to all, though that could require thousands of API calls > > > depending on the user. (And could therefore take hours to do with the > > > 100 API query/hour limit.)
[twitter-dev] Re: sending DM to all followers?
> > I'm wondering if there is a way - or if you would consider adding a > > way - to send a DM to all followers via the API? > > I believe that's called "Tweeting" Indeed. Mass DMs are bad manners IMHO. -- personal: http://www.cameronkaiser.com/ -- Cameron Kaiser * Floodgap Systems * www.floodgap.com * ckai...@floodgap.com -- "Endian Little Hate We" -- credits from Connectix Virtual PC 6 for Mac -
[twitter-dev] Re: sending DM to all followers?
Alex, That sounds very spamish although there are certainly some use cases where it is acceptable. Proceed with caution when sending mass DMs. Ensure the messages you are sending are relevant and of value to your followers. Doug Williams Twitter API Support http://twitter.com/dougw On Tue, Apr 14, 2009 at 4:41 PM, Chad Etzel wrote: > > I believe that's called "Tweeting" > -Chad > > On Tue, Apr 14, 2009 at 7:24 PM, Alex wrote: > > > > I'm wondering if there is a way - or if you would consider adding a > > way - to send a DM to all followers via the API? > > > > Obviously we could grab the followers list and iterate over it to send > > the DM to all, though that could require thousands of API calls > > depending on the user. (And could therefore take hours to do with the > > 100 API query/hour limit.) > > > > >
[twitter-dev] Re: sending DM to all followers?
Please don't. @Jesse On Tue, Apr 14, 2009 at 5:24 PM, Alex wrote: > > I'm wondering if there is a way - or if you would consider adding a > way - to send a DM to all followers via the API? > > Obviously we could grab the followers list and iterate over it to send > the DM to all, though that could require thousands of API calls > depending on the user. (And could therefore take hours to do with the > 100 API query/hour limit.) > >
[twitter-dev] Re: combined DM feed?
There has been some chatter before of offering combined feeds (user_timeline, friends_timeline, direct_messages, etc...) in a single call. However, we have recommended people build a proxy to do such aggregation until time and resources allow us to look into it. Doug Williams Twitter API Support http://twitter.com/dougw On Tue, Apr 14, 2009 at 4:20 PM, Alex wrote: > > Was hoping you would consider offering an API call that would return a > combined DM feed of both sent and received DMs. > > Thoughts? >
[twitter-dev] Re: sending DM to all followers?
I believe that's called "Tweeting" -Chad On Tue, Apr 14, 2009 at 7:24 PM, Alex wrote: > > I'm wondering if there is a way - or if you would consider adding a > way - to send a DM to all followers via the API? > > Obviously we could grab the followers list and iterate over it to send > the DM to all, though that could require thousands of API calls > depending on the user. (And could therefore take hours to do with the > 100 API query/hour limit.) > >
[twitter-dev] Re: Deprecation of source parameter registration
We've finished the removal of this functionality from the site so third-party registration is no longer supported. If you feel that you have an extraordinary need for explicit source parameter registration, please email a...@twitter.com. Include why you need to register an application manually and why OAuth will not work in your case. Thanks, Doug Williams Twitter API Support http://twitter.com/dougw On Apr 12, 7:39 am, Sam Johnston wrote: > Hi Doug, > > On Apr 9, 4:14 am, Doug Williams wrote: > > > > > Beginning late this week or early next week, application developers will no > > longer be able to request APIsourceparameters. Instead, newsource > > parameters will only be available for OAuth applications, and will be > > managed by the developer through the registration and management interface > > (http://twitter.com/oauth_clients). > > This seems a little premature don't you think? Thesourceparameter is > important for marketing applications as well as gauging popularity - > OAuth is both still in beta and unsupported by many clients/ > applications. > > I've just started developing a Twitter app using python-twitter that > is to run on Google AppEngine, but it will be some time after the next > release before we have OAuth support (and even then it's dependent on > a "major overhaul of the HTTP layer")[1]. > > Sam > > 1.http://code.google.com/p/python-twitter/issues/detail?id=37&q=oauth#c4
[twitter-dev] sending DM to all followers?
I'm wondering if there is a way - or if you would consider adding a way - to send a DM to all followers via the API? Obviously we could grab the followers list and iterate over it to send the DM to all, though that could require thousands of API calls depending on the user. (And could therefore take hours to do with the 100 API query/hour limit.)
[twitter-dev] combined DM feed?
Was hoping you would consider offering an API call that would return a combined DM feed of both sent and received DMs. Thoughts?
[twitter-dev] Re: Automating twitter posts
http://easytweets.com On Apr 14, 4:20 pm, Beier wrote: > Try HootSuite.com, not only will it automate your RSS tweets, it will > also automate you whole Twitter accounts > > On Apr 14, 7:50 am, dondmcg wrote: > > > I work for a newspaper and we have more and more limited resources and > > four flags newspapers. I am wondering if there is any way o automate > > twitter posts off an RSS feed or newsletter? In other words if I set > > up four twitter accounts to the four papers is there a way I can > > populate htose accounts with posts without adding to my workload? > > > Sorry if this has already been asked I searched but did not find > > anything.
[twitter-dev] API Changes for April 14, 2009
Hello, Two bug fixes today, one of which was introduced by the OAuth signature change on April 9th. * Fixed (OAuth): non-ASCII characters in POST parameters were incompatible afte the April 9th change. That incompatibility was corrected. » See Also: http://code.google.com/p/twitter-api/issues/detail?id=433 * Fixed (REST): The since_id parameter now works on the / direct_messages/sent method » See Also: http://code.google.com/p/twitter-api/issues/detail?id=431 Thanks; — Matt Sanford / @mzsanford
[twitter-dev] IP Address range
I've got a python script that is monitoring the playlist for our local public radio station, and tweeting when new tracks come up. It is using @wdav as the twitter ID (although that is not relevant to this question...) I am using the twitter.py library to wrap the twitter api. Runs fine on my local laptop, but when I deployed it to my hosted server I had to tell them an IP address it was posting to so they could implement a firewall rule to let the traffic through. I gave them 128.121.146.100, since that's what comes back from a ping to twitter.com. The problem is that it seems the script is frequently trying to use other ip addresses to reach twitter. Is there a range of IP addresses that might be valid Twitter endpoints, that I need to pass on to the hosted server admin team? Any help greatly appreciated! Bill
[twitter-dev] Best way to check for errors on status update using Twitter OAuth code
Quick question: I am implementing OAuth and am using the sample PHP code Twitter gave out. My question is what is the best way to determine an error occurred with the status update using $content = $to->OAuthRequest('https://twitter.com/statuses/ update.xml', array('status' => 'Test OAuth update. #testoauth'), 'POST'); $content contains different values depending on the type error. Probably an easy answer here but I'm not finding it. Thanks for the help
[twitter-dev] Re: Trending Service for a given set of users
I'm currently building an AIR twitter client (yes i know yet another one) and part of the goal is to help users see what is popular (trending) within their own community. The most popular URLS and topics information exists for twitter at large, but sometimes what happens in your own group of cohorts is more interesting and usefull than knowing that 15 million people are really digging American Idol right now. So looking for most popular URL RT, most popular topic for a given set of users. Bonus if the service can also return the same information for the followers of the user set. Hope that helps On Tue, Apr 14, 2009 at 4:21 PM, Carlos Crosetti wrote: > Please can you explain the "trending" output you are looking for? > > > On Tue, Apr 14, 2009 at 2:45 PM, kalisurfer wrote: > >> >> Hi All, >> >> Looking for a service where provided a list of users (100+) i can get >> back the trending URL, topics, hash and RT. Love to be able to access >> the info via a REST API. >> >> Trying to not build it out myself. >> >> Thanks, >> Sean >> @kalisurfer >> > > > > -- > Carlos Crosetti > -- Sean Scott cell: 612.867.8133 portfolio: http://www.flickr.com/photos/92876...@n00/sets/72157613990263453/ profile: http://www.linkedin.com/profile?viewProfile=&key=2242610 blog: http://www.twofortyeight.com/ other: http://twitter.com/kalisurfer
[twitter-dev] Re: Trending Service for a given set of users
I would surmise: http://apiwiki.twitter.com/Search+API+Documentation#Trends On Tue, Apr 14, 2009 at 16:21, Carlos Crosetti wrote: > Please can you explain the "trending" output you are looking for? > > > On Tue, Apr 14, 2009 at 2:45 PM, kalisurfer wrote: > >> >> Hi All, >> >> Looking for a service where provided a list of users (100+) i can get >> back the trending URL, topics, hash and RT. Love to be able to access >> the info via a REST API. >> >> Trying to not build it out myself. >> >> Thanks, >> Sean >> @kalisurfer >> > > > > -- > Carlos Crosetti > -- Abraham Williams | http://the.hackerconundrum.com Hacker | http://abrah.am | http://twitter.com/abraham Web608 | Community Evangelist | http://web608.org This email is: [ ] blogable [x] ask first [ ] private. Sent from Madison, Wisconsin, United States
[twitter-dev] Re: Trending Service for a given set of users
Please can you explain the "trending" output you are looking for? On Tue, Apr 14, 2009 at 2:45 PM, kalisurfer wrote: > > Hi All, > > Looking for a service where provided a list of users (100+) i can get > back the trending URL, topics, hash and RT. Love to be able to access > the info via a REST API. > > Trying to not build it out myself. > > Thanks, > Sean > @kalisurfer > -- Carlos Crosetti
[twitter-dev] Re: Combination of search operators returns 404
Hi Daniel, The issue here is the until: sate being too far in the past and combined with filter:links. This does not seems like it should 404 so I'll try and correct that. Thanks; — Matt On Apr 13, 2009, at 07:34 PM, Dimebrain wrote: This query contains two documented search operators, and it bombs with a 404 if they're called together, but works if called separately. Previous unit tests allowed multiple operators. Is 404 the correct response for "no results"? It returns as HTML. http://search.twitter.com/search.json?q=twitter until:2009-01-04 filter:links
[twitter-dev] Re: Automating twitter posts
Try HootSuite.com, not only will it automate your RSS tweets, it will also automate you whole Twitter accounts On Apr 14, 7:50 am, dondmcg wrote: > I work for a newspaper and we have more and more limited resources and > four flags newspapers. I am wondering if there is any way o automate > twitter posts off an RSS feed or newsletter? In other words if I set > up four twitter accounts to the four papers is there a way I can > populate htose accounts with posts without adding to my workload? > > Sorry if this has already been asked I searched but did not find > anything.
[twitter-dev] Re: autenticating user using OAuth
Also applications that support multiple twitter accounts will probably have to have their own local user authentication system, unless anyone has figured out a better way to deal with this? -Chad On Tue, Apr 14, 2009 at 3:27 PM, Doug Williams wrote: > That is a design decision we leave up to the application developer. So long > as you aren't storing credentials (OAuth applications never do) then we are > happy campers. > > It makes sense for a lot of Twitter applications to use the Twitter user id > as the application's internal account id. However it can make growing > difficult if the service will move to support other services other than > Twitter. When developing an application that will likely support more than > one service, a level of indirection (application dependent user ids) would > be recommended. > > Thanks, > Doug Williams >
[twitter-dev] Re: Trending Service for a given set of users
On Tue, Apr 14, 2009 at 10:45 AM, kalisurfer wrote: > > Hi All, > > Looking for a service where provided a list of users (100+) i can get > back the trending URL, topics, hash and RT. Love to be able to access > the info via a REST API. That seems like a good idea and I'm hoping to take TwURLed News ( http://twurlednews.com) in that direction... but I can't say how long it might be before I get there. I'd like to see it too, if I don't manage to build it first. Nick
[twitter-dev] Re: autenticating user using OAuth
That is a design decision we leave up to the application developer. So long as you aren't storing credentials (OAuth applications never do) then we are happy campers. It makes sense for a lot of Twitter applications to use the Twitter user id as the application's internal account id. However it can make growing difficult if the service will move to support other services other than Twitter. When developing an application that will likely support more than one service, a level of indirection (application dependent user ids) would be recommended. Thanks, Doug Williams Twitter API Support http://twitter.com/dougw On Tue, Apr 14, 2009 at 9:08 AM, Arik Fraimovich wrote: > > On Apr 13, 10:37 pm, Matt Sanford wrote: > > It is currently the case that you will get the Accept/Deny page > > every time. We're working on a redirect like Zachary mentioned and > > hope to have it out by the end of the week. > > Is this kind of usage you will encourage or is it still be better to > create our own user login system? > > Thanks > Arik >
[twitter-dev] Re: Em-dash, accented characters in updates: Failed to validate...
Hey Abraham, Checkout the dreaded issue 433 as we found out there is a bug in that RESERVED_CHARACTERS depending on the $KCODE variable. Thanks; — Matt On Apr 14, 2009, at 11:35 AM, Abraham Williams wrote: I'll try to have a look at this tonight. Abraham On Fri, Apr 10, 2009 at 19:02, Matt Sanford wrote: Hi Isaac, The change we implemented seems to have fixed people using a newer version of the Ruby oauth gem but it sounds like it may have broken some other libraries. My guess is that there is an encoding normalization difference between the PHP library and the newest Ruby version right NOW. The spec does have some normalization information [1] [2] and it seems like the Ruby code is doing that work via [3]: def escape(value) URI::escape(value.to_s, OAuth::RESERVED_CHARACTERS) end And that constant is defined as: # reserved character regexp, per section 5.1 RESERVED_CHARACTERS = /[^\w\d\-\.\_\~]/ I know Abraham reads this Google Group so perhaps he and I can get together on Monday and compare some pre-signature strings between Ruby and PHP to figure out what's going on. If you see this Abraham drop me an email otherwise you can expect one from me on Monday to sort this out. Thanks; — Matt [1] - http://oauth.net/core/1.0/#anchor14 [2] - http://oauth.net/core/1.0/#encoding_parameters [3] - http://github.com/mojodna/oauth/blob/dbd946790a882d91fb111ba64331de003c135872/lib/oauth/helper.rb#L9 [4] - http://github.com/mojodna/oauth/blob/dbd946790a882d91fb111ba64331de003c135872/lib/oauth/oauth.rb On Apr 10, 2009, at 03:47 PM, Isaac wrote: Hi, I've been using the em-dash when posting status updates, and apparently this now causes the Twitter API to return "Failed to validate..." errors. This just started happening within the last 24-48 hours. I am using OAuth (specifically, the PHP library written by Abraham Williams), and my site is http://dora.fm. I noticed in the API changelog that accented characters -- and I'm assuming, characters like the em-dash -- have supposedly been "fixed" (Fixed (OAuth): Accented characters in statuses were causing signature error for OAuth clients. This has been corrected.) But I'm finding that this is broken NOW and was working fine before. Suggestions? Thanks. -- Abraham Williams | http://the.hackerconundrum.com Hacker | http://abrah.am | http://twitter.com/abraham Web608 | Community Evangelist | http://web608.org This email is: [ ] blogable [x] ask first [ ] private. Sent from Madison, Wisconsin, United States
[twitter-dev] Re: Em-dash, accented characters in updates: Failed to validate...
I'll try to have a look at this tonight. Abraham On Fri, Apr 10, 2009 at 19:02, Matt Sanford wrote: > Hi Isaac, > The change we implemented seems to have fixed people using a newer > version of the Ruby oauth gem but it sounds like it may have broken some > other libraries. My guess is that there is an encoding normalization > difference between the PHP library and the newest Ruby version right NOW. > The spec does have some normalization information [1] [2] and it seems like > the Ruby code is doing that work via [3]: > > def escape(value) > URI::escape(value.to_s, OAuth::RESERVED_CHARACTERS) > end > > And that constant is defined as: > > # reserved character regexp, per section 5.1 > RESERVED_CHARACTERS = /[^\w\d\-\.\_\~]/ > > I know Abraham reads this Google Group so perhaps he and I can get > together on Monday and compare some pre-signature strings between Ruby and > PHP to figure out what's going on. If you see this Abraham drop me an email > otherwise you can expect one from me on Monday to sort this out. > > Thanks; > — Matt > > > [1] - http://oauth.net/core/1.0/#anchor14 > [2] - http://oauth.net/core/1.0/#encoding_parameters > [3] - > http://github.com/mojodna/oauth/blob/dbd946790a882d91fb111ba64331de003c135872/lib/oauth/helper.rb#L9 > [4] - > http://github.com/mojodna/oauth/blob/dbd946790a882d91fb111ba64331de003c135872/lib/oauth/oauth.rb > > On Apr 10, 2009, at 03:47 PM, Isaac wrote: > > > Hi, > > I've been using the em-dash when posting status updates, and > apparently this now causes the Twitter API to return "Failed to > validate..." errors. This just started happening within the last 24-48 > hours. > > I am using OAuth (specifically, the PHP library written by Abraham > Williams), and my site is http://dora.fm. > > I noticed in the API changelog that accented characters -- and I'm > assuming, characters like the em-dash -- have supposedly been > "fixed" (Fixed (OAuth): Accented characters in statuses were causing > signature error for OAuth clients. This has been corrected.) > > But I'm finding that this is broken NOW and was working fine before. > > Suggestions? > > Thanks. > > > -- Abraham Williams | http://the.hackerconundrum.com Hacker | http://abrah.am | http://twitter.com/abraham Web608 | Community Evangelist | http://web608.org This email is: [ ] blogable [x] ask first [ ] private. Sent from Madison, Wisconsin, United States
[twitter-dev] Re: How to get a user's friends?
Thanks guys. Sorry if repetitive. On Apr 14, 11:02 am, Matt Sanford wrote: > Hi Brandon, > > You can get the full user objects by paginating the /statuses/ > friends method [1]. The ids method was added for the exact cache usage > you describe, but only works well if you need to examine a users > network information (things like followers-in-common). The reason we > returns ids is because we cache that information and it can be > returned very quickly [2]. Returning anything more would essentially > be a very expensive version of the /statuses/friends method … even > just screen names would be expensive. The idea of returning screen > names has come up a few times [3][4][5] but it's not something we can > support and still keep the site running. > > Thanks; > — Matt Sanford / @mzsanford > > [1] -http://apiwiki.twitter.com/REST+API+Documentation#statuses/friends > [2] -http://groups.google.com/group/twitter-development-talk/msg/efa45ceb8... > [3] -http://groups.google.com/group/twitter-development-talk/browse_frm/th... > [4] -http://groups.google.com/group/twitter-development-talk/browse_frm/th... > [5] -http://code.google.com/p/twitter-api/issues/detail?id=265 > > On Apr 14, 2009, at 09:31 AM, Brandon Geiger wrote: > > > > > Right now we are using the method: > > >http://twitter.com/friends/ids.json > > > to get a user's friends, but it doesn't have the friend's screen_name, > > which is obviously how users identify their friends. So what we are > > currently having to do, is store the id's in our database then loop > > through and use the method: > > >http://twitter.com/users/show/.json > > > to figure out the screen_name associated to the id, and store that in > > our system as well. This is a failed system because users can change > > their screen names and it burns through API requests to fast. > > > API DEV TEAM: anyway you guys can add screen_name in the response for > > the methodhttp://twitter.com/friends/ids.json? > > > That would solve this issue. > > > If I'm missing something obvious, or if anyone has a better approach > > please reply. Thanks!
[twitter-dev] Trending Service for a given set of users
Hi All, Looking for a service where provided a list of users (100+) i can get back the trending URL, topics, hash and RT. Love to be able to access the info via a REST API. Trying to not build it out myself. Thanks, Sean @kalisurfer
[twitter-dev] Updating icon for OAuth applications
I have trouble updating the icon for my OAuth applications. I tried several different GIF and PNG files, much smaller than 700k and always get the error message: Your application was registered, but there was a problem with your application image. Probably too big. Has anyone else had this problem, or is it just that every image file on my computer is corrupted? Guan
[twitter-dev] Re: Request for documentation review
Ahmed and Chad, I'm rolling in all of those suggestions right now. Thank you both for thorough perusal. Any REST guys out there that can offer the similar insight? Doug Williams Twitter API Support http://twitter.com/dougw On Tue, Apr 14, 2009 at 9:50 AM, Chad Etzel wrote: > > A note on the "Trends" page: > > / QUOTE ***/ > > Twitter Search API Method: trends > > Returns the top ten queries that are currently trending on Twitter. > The response includes the time of the request, the name of each > trending topic, and the url to the Twitter Search results page for > that topic. > > / END QUOTE ***/ > > > "Returns the top ten *queries* that are currently trending on Twitter" > is inaccurate. They are the top ten trends being posted by people on > twitter. It should really be the "... top ten (phrases/words/topics) > that are currently trending on Twitter". > > -Chad > > > On Tue, Apr 14, 2009 at 12:40 PM, Chad Etzel wrote: > > More feedback about the Search API section: > > > > # Mentioning a user: > http://search.twitter.com/search.atom?q=to%3Amzsanford > > > > "to:user" in Search-land means that the tweet physically begins with > > @user. This would be the stricter "reply" definition. I would change > > "Mentioning" to "Replies to a user" or "Strict Replies to a user" > > > > > > # Referencing a user: http://search.twitter.com/search.atom?q=%40biz > > > > "Referencing" should really be "Mentioning" here since that is in line > > with the nomenclature of "mentions" in Twitter/REST-land. > > > > # Combine any of the operators together: > > http://search.twitter.com/search.atom?q="happy+hour"+near%3ASF > > > > The Search API does *NOT* support the "near:location" operator (even > > though the website does). You must send in a geocoded lat/lang/radius > > style argument in order to accomplish this with the API. Ask Matt > > about the official format. > > > > # Originating from an application: > > http://search.twitter.com/search.atom?q=landing from:tweetie > > > > 4 things: > > 1) searching for an application uses the "source:" operator, not "from:" > > 2) you should mention somewhere that you *cannot* just search for > > tweets from a source, but that you must at least pass in some other > > keyword to search along with the source: operator. > > 3) On a more general note, the spaces should be urlencoded as well. > > It would probably help to mention somewhere in the doc that all > > queries must be urlencoded before sending them. > > 4) For some reason the link in this example does not actually go to that > query. > > Other thoughts: > > > > Where are the Optional Arguments like "since_id", "max_id", "rpp", and > "page" ? > > > > I believe I was told by Matt that queries are actually limited to 138 > > characters since it counts the "q=" characters in the request as part > > of the query. That should probably be noted. > > > > I'll have a look at the REST sections later, but I don't exercise > > those as heavily, so I might not be the best to review it. > > > > -Chad > > > > On Mon, Apr 13, 2009 at 7:11 PM, Doug Williams wrote: > >> Coderz, > >> I am refactoring the API documentation [1] to make it friendlier on the > eyes > >> and easier on new developers. Please give it a once over and reply with > any > >> of errors, addition requests, or suggestions. We will be using these new > >> docs officially in a few days (linked from the apiwiki.twitter.com) > once any > >> glaring omissions and/or errors are addressed. > >> > >> My goal is to make the documentation more robust and example laden to > help > >> new developers get going without needing help. Suggestions and critism > along > >> those lines is welcome. > >> > >> A final plea for help: > >> If you have some spare time and examples to borrow from, please send > >> a...@twitter.com any code/instructions that I can then paste into method > body > >> to help new devs get started. Basic Hello World! style usage examples > for > >> each method would go a long way to curbing new developer problems. Be > sure > >> to include a link to any libraries your example is using. Working full > >> length example code hosted on github would be preferred. > >> > >> 1. http://apiwiki.twitter.com/Twitter-API-Documentation > >> > >> Thanks, > >> Doug Williams > >> Twitter API Support > >> http://twitter.com/dougw > > >
[twitter-dev] Re: Request for documentation review
Ahmed and Chad, I'm rolling in all of those suggestions right now. Thank you both for thorough perusal. Any REST guys out there that can offer the similar insight? Doug Williams Twitter API Support http://twitter.com/dougw On Tue, Apr 14, 2009 at 9:50 AM, Chad Etzel wrote: > > A note on the "Trends" page: > > / QUOTE ***/ > > Twitter Search API Method: trends > > Returns the top ten queries that are currently trending on Twitter. > The response includes the time of the request, the name of each > trending topic, and the url to the Twitter Search results page for > that topic. > > / END QUOTE ***/ > > > "Returns the top ten *queries* that are currently trending on Twitter" > is inaccurate. They are the top ten trends being posted by people on > twitter. It should really be the "... top ten (phrases/words/topics) > that are currently trending on Twitter". > > -Chad > > > On Tue, Apr 14, 2009 at 12:40 PM, Chad Etzel wrote: > > More feedback about the Search API section: > > > > # Mentioning a user: > http://search.twitter.com/search.atom?q=to%3Amzsanford > > > > "to:user" in Search-land means that the tweet physically begins with > > @user. This would be the stricter "reply" definition. I would change > > "Mentioning" to "Replies to a user" or "Strict Replies to a user" > > > > > > # Referencing a user: http://search.twitter.com/search.atom?q=%40biz > > > > "Referencing" should really be "Mentioning" here since that is in line > > with the nomenclature of "mentions" in Twitter/REST-land. > > > > # Combine any of the operators together: > > http://search.twitter.com/search.atom?q="happy+hour"+near%3ASF > > > > The Search API does *NOT* support the "near:location" operator (even > > though the website does). You must send in a geocoded lat/lang/radius > > style argument in order to accomplish this with the API. Ask Matt > > about the official format. > > > > # Originating from an application: > > http://search.twitter.com/search.atom?q=landing from:tweetie > > > > 4 things: > > 1) searching for an application uses the "source:" operator, not "from:" > > 2) you should mention somewhere that you *cannot* just search for > > tweets from a source, but that you must at least pass in some other > > keyword to search along with the source: operator. > > 3) On a more general note, the spaces should be urlencoded as well. > > It would probably help to mention somewhere in the doc that all > > queries must be urlencoded before sending them. > > 4) For some reason the link in this example does not actually go to that > query. > > Other thoughts: > > > > Where are the Optional Arguments like "since_id", "max_id", "rpp", and > "page" ? > > > > I believe I was told by Matt that queries are actually limited to 138 > > characters since it counts the "q=" characters in the request as part > > of the query. That should probably be noted. > > > > I'll have a look at the REST sections later, but I don't exercise > > those as heavily, so I might not be the best to review it. > > > > -Chad > > > > On Mon, Apr 13, 2009 at 7:11 PM, Doug Williams wrote: > >> Coderz, > >> I am refactoring the API documentation [1] to make it friendlier on the > eyes > >> and easier on new developers. Please give it a once over and reply with > any > >> of errors, addition requests, or suggestions. We will be using these new > >> docs officially in a few days (linked from the apiwiki.twitter.com) > once any > >> glaring omissions and/or errors are addressed. > >> > >> My goal is to make the documentation more robust and example laden to > help > >> new developers get going without needing help. Suggestions and critism > along > >> those lines is welcome. > >> > >> A final plea for help: > >> If you have some spare time and examples to borrow from, please send > >> a...@twitter.com any code/instructions that I can then paste into method > body > >> to help new devs get started. Basic Hello World! style usage examples > for > >> each method would go a long way to curbing new developer problems. Be > sure > >> to include a link to any libraries your example is using. Working full > >> length example code hosted on github would be preferred. > >> > >> 1. http://apiwiki.twitter.com/Twitter-API-Documentation > >> > >> Thanks, > >> Doug Williams > >> Twitter API Support > >> http://twitter.com/dougw > > >
[twitter-dev] Re: public_timeline, invalid profile_image?
Yes. We're testing a system that will replace the public timeline method for frequent requesters. On Mon, Apr 13, 2009 at 21:44, Eric Martin wrote: > > Alex - seems to still be an issue. Any updates on when we might see a > fix? > > On Mar 30, 12:49 pm, Alex Payne wrote: >> We're on this. Thanks for the reports. >> >> >> >> On Sun, Mar 29, 2009 at 18:27, Gary Zhao wrote: >> > I'm seeing it too. >> >> > 2009/3/29 Günter Grodotzki >> >> >> since some days I am always getting: >> >> >>http://static.twitter.com/images/default_profile_normal.png >> >> >> as profile-image from the user via public-timeline (Data-mining-feed). >> >> >> I checked the announcement-google-group + twitter.com/twitterapi >> >> (subscribed to feed anyway ;) ) but could not see any change. >> >> >> The site affected:www.geoheartbeat.com >> >> > -- >> > Gary >> >http://twitter.com/garyzhao >> >> -- >> Alex Payne - API Lead, Twitter, Inc.http://twitter.com/al3x > -- Alex Payne - API Lead, Twitter, Inc. http://twitter.com/al3x
[twitter-dev] Re: update_profile_image not updating user's profile_image_url
What Cameron said. The first 401 is our servers saying, "you need to identify yourself". Once that happens, the request seems to go through. On Tue, Apr 14, 2009 at 08:36, Cameron Kaiser wrote: > >> Thanks, I have sent you a Charles bebug output. >> Interesting to see is that with a request I first get a 401 >> unauthorised error. > > Strictly speaking, per the HTTP spec, clients should not reply with > authentication information to a resource that has not first requested it. > The 401 is to make that possible. > > -- > personal: http://www.cameronkaiser.com/ > -- > Cameron Kaiser * Floodgap Systems * www.floodgap.com * ckai...@floodgap.com > -- Never eat more than you can lift. -- Miss Piggy > > -- Alex Payne - API Lead, Twitter, Inc. http://twitter.com/al3x
[twitter-dev] Re: How to get a user's friends?
Hi Brandon, You can get the full user objects by paginating the /statuses/ friends method [1]. The ids method was added for the exact cache usage you describe, but only works well if you need to examine a users network information (things like followers-in-common). The reason we returns ids is because we cache that information and it can be returned very quickly [2]. Returning anything more would essentially be a very expensive version of the /statuses/friends method … even just screen names would be expensive. The idea of returning screen names has come up a few times [3][4][5] but it's not something we can support and still keep the site running. Thanks; — Matt Sanford / @mzsanford [1] - http://apiwiki.twitter.com/REST+API+Documentation#statuses/friends [2] - http://groups.google.com/group/twitter-development-talk/msg/efa45ceb8bed71b4 [3] - http://groups.google.com/group/twitter-development-talk/browse_frm/thread/2d013e8fc6d9a71e/ef907e1ca06b97ff [4] - http://groups.google.com/group/twitter-development-talk/browse_frm/thread/98f4c4d13954e8bf/4b511a0455f8bffe [5] - http://code.google.com/p/twitter-api/issues/detail?id=265 On Apr 14, 2009, at 09:31 AM, Brandon Geiger wrote: Right now we are using the method: http://twitter.com/friends/ids.json to get a user's friends, but it doesn't have the friend's screen_name, which is obviously how users identify their friends. So what we are currently having to do, is store the id's in our database then loop through and use the method: http://twitter.com/users/show/.json to figure out the screen_name associated to the id, and store that in our system as well. This is a failed system because users can change their screen names and it burns through API requests to fast. API DEV TEAM: anyway you guys can add screen_name in the response for the method http://twitter.com/friends/ids.json? That would solve this issue. If I'm missing something obvious, or if anyone has a better approach please reply. Thanks!
[twitter-dev] Re: How to get a user's friends?
there are several threads discussing this if you search the group. There may even be an open issue. On Tue, Apr 14, 2009 at 11:31, Brandon Geiger wrote: > > Right now we are using the method: > > http://twitter.com/friends/ids.json > > to get a user's friends, but it doesn't have the friend's screen_name, > which is obviously how users identify their friends. So what we are > currently having to do, is store the id's in our database then loop > through and use the method: > > http://twitter.com/users/show/.json > > to figure out the screen_name associated to the id, and store that in > our system as well. This is a failed system because users can change > their screen names and it burns through API requests to fast. > > API DEV TEAM: anyway you guys can add screen_name in the response for > the method http://twitter.com/friends/ids.json? > > That would solve this issue. > > If I'm missing something obvious, or if anyone has a better approach > please reply. Thanks! -- Abraham Williams | http://the.hackerconundrum.com Hacker | http://abrah.am | http://twitter.com/abraham Web608 | Community Evangelist | http://web608.org This email is: [ ] blogable [x] ask first [ ] private. Sent from Madison, Wisconsin, United States
[twitter-dev] Re: Request for documentation review
More feedback about the Search API section: # Mentioning a user: http://search.twitter.com/search.atom?q=to%3Amzsanford "to:user" in Search-land means that the tweet physically begins with @user. This would be the stricter "reply" definition. I would change "Mentioning" to "Replies to a user" or "Strict Replies to a user" # Referencing a user: http://search.twitter.com/search.atom?q=%40biz "Referencing" should really be "Mentioning" here since that is in line with the nomenclature of "mentions" in Twitter/REST-land. # Combine any of the operators together: http://search.twitter.com/search.atom?q="happy+hour"+near%3ASF The Search API does *NOT* support the "near:location" operator (even though the website does). You must send in a geocoded lat/lang/radius style argument in order to accomplish this with the API. Ask Matt about the official format. # Originating from an application: http://search.twitter.com/search.atom?q=landing from:tweetie 4 things: 1) searching for an application uses the "source:" operator, not "from:" 2) you should mention somewhere that you *cannot* just search for tweets from a source, but that you must at least pass in some other keyword to search along with the source: operator. 3) On a more general note, the spaces should be urlencoded as well. It would probably help to mention somewhere in the doc that all queries must be urlencoded before sending them. 4) For some reason the link in this example does not actually go to that query. Other thoughts: Where are the Optional Arguments like "since_id", "max_id", "rpp", and "page" ? I believe I was told by Matt that queries are actually limited to 138 characters since it counts the "q=" characters in the request as part of the query. That should probably be noted. I'll have a look at the REST sections later, but I don't exercise those as heavily, so I might not be the best to review it. -Chad On Mon, Apr 13, 2009 at 7:11 PM, Doug Williams wrote: > Coderz, > I am refactoring the API documentation [1] to make it friendlier on the eyes > and easier on new developers. Please give it a once over and reply with any > of errors, addition requests, or suggestions. We will be using these new > docs officially in a few days (linked from the apiwiki.twitter.com) once any > glaring omissions and/or errors are addressed. > > My goal is to make the documentation more robust and example laden to help > new developers get going without needing help. Suggestions and critism along > those lines is welcome. > > A final plea for help: > If you have some spare time and examples to borrow from, please send > a...@twitter.com any code/instructions that I can then paste into method body > to help new devs get started. Basic Hello World! style usage examples for > each method would go a long way to curbing new developer problems. Be sure > to include a link to any libraries your example is using. Working full > length example code hosted on github would be preferred. > > 1. http://apiwiki.twitter.com/Twitter-API-Documentation > > Thanks, > Doug Williams > Twitter API Support > http://twitter.com/dougw
[twitter-dev] Re: autenticating user using OAuth
On Apr 13, 10:37 pm, Matt Sanford wrote: > It is currently the case that you will get the Accept/Deny page > every time. We're working on a redirect like Zachary mentioned and > hope to have it out by the end of the week. Is this kind of usage you will encourage or is it still be better to create our own user login system? Thanks Arik
[twitter-dev] Re: OAuth issues with POST
Hi Daniel, While working through issue 433 [1] we've discussed signatures pretty in-depth. This is an unrelated problem but I think you'll find the discussion about signatures there helpful, as it has several examples. The short-short answer is: • The string you sign should have the status=testing parameter appended, like: POST&http%3A%2F%2Ftwitter.com%2Fstatuses %2Fupdate.xml&oauth_consumer_key%3Dw9zJ2JxgWTsJN5OyEHIZjw%26oauth_nonce %3Ddudialjpcnn4p355%26oauth_signature_method%3DHMAC- SHA1%26oauth_timestamp%3D1239673686%26oauth_token%3D11173402- LxhEcmKGl2zKjsbuBE0tdt4UBAxHoR1LFfrs5tTo4%26oauth_version %3D1.00%26status%3Dtesting • When you make the request the URL should be http://twitter.com/statuses/update.xml and the status=testing parameter should be in the body, like a normal x-www-form-urlencoded POST. Thanks; — Matt Sanford / @mzsanford [1] - http://code.google.com/p/twitter-api/issues/detail?id=433 On Apr 14, 2009, at 09:28 AM, Dimebrain wrote: Hello, I originally commented on issue thread 447 but that issue was closed, so I wanted to repost my problem to see if it's something I'm doing wrong on my side. I am still failing, but using C# / .NET and a self-authored OAuth implementation. My GET calls work correctly, my POST calls 401. Here is the URI: http://twitter.com/statuses/update.xml?status=testing Here is my signature base string: POST&http%3A%2F%2Ftwitter.com%2Fstatuses %2Fupdate.xml&oauth_consumer_key%3Dw9zJ2JxgWTsJN5OyEHIZjw %26oauth_nonce %3Ddudialjpcnn4p355%26oauth_signature_method%3DHMAC- SHA1%26oauth_timestamp%3D1239673686%26oauth_token%3D11173402- LxhEcmKGl2zKjsbuBE0tdt4UBAxHoR1LFfrs5tTo4%26oauth_version%3D1.0 And the signature (url encoded): E0UkQEmfcaT3DOG7a8L7sImCmVw%3D And the authorization header: OAuth oauth_consumer_key="w9zJ2JxgWTsJN5OyEHIZjw",oauth_token="11173402- LxhEcmKGl2zKjsbuBE0tdt4UBAxHoR1LFfrs5tTo4 ",oauth_nonce = "dudialjpcnn4p355 ",oauth_timestamp="1239673686",oauth_signature_method="HMAC- SHA1",oauth_signature="E0UkQEmfcaT3DOG7a8L7sImCmVw %3D",oauth_version="1.0" And the request header out the door: POST /statuses/update.xml?status=testing HTTP/1.1 Content-Type: application/x-www-form-urlencoded Authorization: Host: twitter.com - I've also tried including "status=testing" in the list of post parameters included in the signature hashing. I'm not clear on whether it's expected to exist there b/c of the normalization done to the URI during signature hashing, i.e. otherwise it would only exist on the POST itself and not in the signature I wonder if someone could clarify a point about POST with status=X, and that is, when preparing a POST with the default W3C application/x- www-form-urlencoded content type, are we supposed to: a) remove the status=X line from the URI query, and instead write it to the request as POST parameters (and therefore include status=X in the signature base string as per OAuth 1.0), or b) keep the status=X in the URI itself, and only including other parameters in the POST parameters. I think what might be happening is that .NET treats the Query fragment of a URI as post parameters, and maybe all approaches with application/ x-www-form-urlencoded do; in other words it's not really a query as per a GET call, it's just the absolute uri path, and then all query params are added to the signature base and written in the POST message ---> with different URL escaping than a URI (i.e. +'s are part of the POST encoding spec as opposed to %20's that are part of the URI encoding spec). See, maybe you can't POST because you use %20's in the sig base but +'s in the POST params, which is actually according to the spec. I'd love to know if anyone has some insight into this problem. The inability to POST w/ OAuth on Twitter is effecting everyone that uses my library. Daniel
[twitter-dev] Re: Request for documentation review
A note on the "Trends" page: / QUOTE ***/ Twitter Search API Method: trends Returns the top ten queries that are currently trending on Twitter. The response includes the time of the request, the name of each trending topic, and the url to the Twitter Search results page for that topic. / END QUOTE ***/ "Returns the top ten *queries* that are currently trending on Twitter" is inaccurate. They are the top ten trends being posted by people on twitter. It should really be the "... top ten (phrases/words/topics) that are currently trending on Twitter". -Chad On Tue, Apr 14, 2009 at 12:40 PM, Chad Etzel wrote: > More feedback about the Search API section: > > # Mentioning a user: http://search.twitter.com/search.atom?q=to%3Amzsanford > > "to:user" in Search-land means that the tweet physically begins with > @user. This would be the stricter "reply" definition. I would change > "Mentioning" to "Replies to a user" or "Strict Replies to a user" > > > # Referencing a user: http://search.twitter.com/search.atom?q=%40biz > > "Referencing" should really be "Mentioning" here since that is in line > with the nomenclature of "mentions" in Twitter/REST-land. > > # Combine any of the operators together: > http://search.twitter.com/search.atom?q="happy+hour"+near%3ASF > > The Search API does *NOT* support the "near:location" operator (even > though the website does). You must send in a geocoded lat/lang/radius > style argument in order to accomplish this with the API. Ask Matt > about the official format. > > # Originating from an application: > http://search.twitter.com/search.atom?q=landing from:tweetie > > 4 things: > 1) searching for an application uses the "source:" operator, not "from:" > 2) you should mention somewhere that you *cannot* just search for > tweets from a source, but that you must at least pass in some other > keyword to search along with the source: operator. > 3) On a more general note, the spaces should be urlencoded as well. > It would probably help to mention somewhere in the doc that all > queries must be urlencoded before sending them. > 4) For some reason the link in this example does not actually go to that > query. > Other thoughts: > > Where are the Optional Arguments like "since_id", "max_id", "rpp", and "page" > ? > > I believe I was told by Matt that queries are actually limited to 138 > characters since it counts the "q=" characters in the request as part > of the query. That should probably be noted. > > I'll have a look at the REST sections later, but I don't exercise > those as heavily, so I might not be the best to review it. > > -Chad > > On Mon, Apr 13, 2009 at 7:11 PM, Doug Williams wrote: >> Coderz, >> I am refactoring the API documentation [1] to make it friendlier on the eyes >> and easier on new developers. Please give it a once over and reply with any >> of errors, addition requests, or suggestions. We will be using these new >> docs officially in a few days (linked from the apiwiki.twitter.com) once any >> glaring omissions and/or errors are addressed. >> >> My goal is to make the documentation more robust and example laden to help >> new developers get going without needing help. Suggestions and critism along >> those lines is welcome. >> >> A final plea for help: >> If you have some spare time and examples to borrow from, please send >> a...@twitter.com any code/instructions that I can then paste into method body >> to help new devs get started. Basic Hello World! style usage examples for >> each method would go a long way to curbing new developer problems. Be sure >> to include a link to any libraries your example is using. Working full >> length example code hosted on github would be preferred. >> >> 1. http://apiwiki.twitter.com/Twitter-API-Documentation >> >> Thanks, >> Doug Williams >> Twitter API Support >> http://twitter.com/dougw >
[twitter-dev] How to get a user's friends?
Right now we are using the method: http://twitter.com/friends/ids.json to get a user's friends, but it doesn't have the friend's screen_name, which is obviously how users identify their friends. So what we are currently having to do, is store the id's in our database then loop through and use the method: http://twitter.com/users/show/.json to figure out the screen_name associated to the id, and store that in our system as well. This is a failed system because users can change their screen names and it burns through API requests to fast. API DEV TEAM: anyway you guys can add screen_name in the response for the method http://twitter.com/friends/ids.json? That would solve this issue. If I'm missing something obvious, or if anyone has a better approach please reply. Thanks!
[twitter-dev] Re: OAuth "Failed to validate oauth signature and token" if application already authorized
This behavior resolved as of this morning; no need to investigate further, and no changes made. On Apr 13, 1:31 pm, Dimebrain wrote: > Hello, > > Recently, I have noticed that if I attempt to do a request token / > access token exchange, i.e. a new application OAuth workflow, I always > fail with "Failed to validate oauth signature and token" from Twitter > if the application has already been successfully authorized in a > previous workflow. If I revoke the application and try again, it works > fine. > > Now I understand the motive for that, to avoid unnecessary load of > having apps that perform the OAuth exchange on every user session, but > this was not how it was working before. Did I miss an API > announcement? My app persists the token, so for me this isn't the end > of the world. > > First I'd like to know if this is expected behavior. I can faithfully > reproduce the scenario of no access, gain access, try to gain access > again and fail, revoke access and try again to succeed. My expectation > is the OAuth process works even if the user has already granted access > to the user. > > Anyone experiencing this right now?
[twitter-dev] OAuth issues with POST
Hello, I originally commented on issue thread 447 but that issue was closed, so I wanted to repost my problem to see if it's something I'm doing wrong on my side. I am still failing, but using C# / .NET and a self-authored OAuth implementation. My GET calls work correctly, my POST calls 401. Here is the URI: http://twitter.com/statuses/update.xml?status=testing Here is my signature base string: POST&http%3A%2F%2Ftwitter.com%2Fstatuses %2Fupdate.xml&oauth_consumer_key%3Dw9zJ2JxgWTsJN5OyEHIZjw%26oauth_nonce %3Ddudialjpcnn4p355%26oauth_signature_method%3DHMAC- SHA1%26oauth_timestamp%3D1239673686%26oauth_token%3D11173402- LxhEcmKGl2zKjsbuBE0tdt4UBAxHoR1LFfrs5tTo4%26oauth_version%3D1.0 And the signature (url encoded): E0UkQEmfcaT3DOG7a8L7sImCmVw%3D And the authorization header: OAuth oauth_consumer_key="w9zJ2JxgWTsJN5OyEHIZjw",oauth_token="11173402- LxhEcmKGl2zKjsbuBE0tdt4UBAxHoR1LFfrs5tTo4",oauth_nonce="dudialjpcnn4p355",oauth_timestamp="1239673686",oauth_signature_method="HMAC- SHA1",oauth_signature="E0UkQEmfcaT3DOG7a8L7sImCmVw %3D",oauth_version="1.0" And the request header out the door: POST /statuses/update.xml?status=testing HTTP/1.1 Content-Type: application/x-www-form-urlencoded Authorization: Host: twitter.com - I've also tried including "status=testing" in the list of post parameters included in the signature hashing. I'm not clear on whether it's expected to exist there b/c of the normalization done to the URI during signature hashing, i.e. otherwise it would only exist on the POST itself and not in the signature I wonder if someone could clarify a point about POST with status=X, and that is, when preparing a POST with the default W3C application/x- www-form-urlencoded content type, are we supposed to: a) remove the status=X line from the URI query, and instead write it to the request as POST parameters (and therefore include status=X in the signature base string as per OAuth 1.0), or b) keep the status=X in the URI itself, and only including other parameters in the POST parameters. I think what might be happening is that .NET treats the Query fragment of a URI as post parameters, and maybe all approaches with application/ x-www-form-urlencoded do; in other words it's not really a query as per a GET call, it's just the absolute uri path, and then all query params are added to the signature base and written in the POST message ---> with different URL escaping than a URI (i.e. +'s are part of the POST encoding spec as opposed to %20's that are part of the URI encoding spec). See, maybe you can't POST because you use %20's in the sig base but +'s in the POST params, which is actually according to the spec. I'd love to know if anyone has some insight into this problem. The inability to POST w/ OAuth on Twitter is effecting everyone that uses my library. Daniel
[twitter-dev] Re: Geocoding posts
Get all posts? Nope. Determine geolocation if you had all posts? Not likely. Possible, with some percentage accuracy or likelihood, but in general, just not likely, unless you're parsing for lat/long coords specifically ... On Tue, Apr 14, 2009 at 11:59 AM, djMax wrote: > > And there's no way for me to get all posts to try and do this myself > right? > > On Apr 14, 11:43 am, Abraham Williams <4bra...@gmail.com> wrote: > > As far as I know Twitter does not save any geo data for individual > tweets. > > None is available through the API anyways. > > > > On a side note: Laconi.ca is working on adding geolocations for each > notice. > > This could be incredibly useful information especially with the increased > > penetration of geo aware phones. > > > > On Mon, Apr 13, 2009 at 22:26, djMax wrote: > > > > > How does Twitter GeoCode tweets OTHER than the users location (if at > > > all)? I'd like to include short urls for geolocation, but want to > > > understand how Twitter uses them. Thanks. > > > > -- > > Abraham Williams |http://the.hackerconundrum.com > > Hacker |http://abrah.am|http://twitter.com/abraham > > Web608 | Community Evangelist |http://web608.org > > This email is: [ ] blogable [x] ask first [ ] private. > > Sent from Madison, Wisconsin, United States >
[twitter-dev] Re: Geocoding posts
And there's no way for me to get all posts to try and do this myself right? On Apr 14, 11:43 am, Abraham Williams <4bra...@gmail.com> wrote: > As far as I know Twitter does not save any geo data for individual tweets. > None is available through the API anyways. > > On a side note: Laconi.ca is working on adding geolocations for each notice. > This could be incredibly useful information especially with the increased > penetration of geo aware phones. > > On Mon, Apr 13, 2009 at 22:26, djMax wrote: > > > How does Twitter GeoCode tweets OTHER than the users location (if at > > all)? I'd like to include short urls for geolocation, but want to > > understand how Twitter uses them. Thanks. > > -- > Abraham Williams |http://the.hackerconundrum.com > Hacker |http://abrah.am|http://twitter.com/abraham > Web608 | Community Evangelist |http://web608.org > This email is: [ ] blogable [x] ask first [ ] private. > Sent from Madison, Wisconsin, United States
[twitter-dev] Re: Geocoding posts
As far as I know Twitter does not save any geo data for individual tweets. None is available through the API anyways. On a side note: Laconi.ca is working on adding geolocations for each notice. This could be incredibly useful information especially with the increased penetration of geo aware phones. On Mon, Apr 13, 2009 at 22:26, djMax wrote: > > How does Twitter GeoCode tweets OTHER than the users location (if at > all)? I'd like to include short urls for geolocation, but want to > understand how Twitter uses them. Thanks. > -- Abraham Williams | http://the.hackerconundrum.com Hacker | http://abrah.am | http://twitter.com/abraham Web608 | Community Evangelist | http://web608.org This email is: [ ] blogable [x] ask first [ ] private. Sent from Madison, Wisconsin, United States
[twitter-dev] Re: update_profile_image not updating user's profile_image_url
> Thanks, I have sent you a Charles bebug output. > Interesting to see is that with a request I first get a 401 > unauthorised error. Strictly speaking, per the HTTP spec, clients should not reply with authentication information to a resource that has not first requested it. The 401 is to make that possible. -- personal: http://www.cameronkaiser.com/ -- Cameron Kaiser * Floodgap Systems * www.floodgap.com * ckai...@floodgap.com -- Never eat more than you can lift. -- Miss Piggy
[twitter-dev] Re: update_profile_image not updating user's profile_image_url
Thanks, I have sent you a Charles bebug output. Interesting to see is that with a request I first get a 401 unauthorised error. Bart On Apr 13, 7:14 pm, Alex Payne wrote: > The output you sent me doesn't include your original request and > doesn't have HTTP headers in the response. Try using a proxy like > Charles (http://www.charlesproxy.com/) to get the complete debugging > output. > > > > > > On Sun, Apr 12, 2009 at 13:34, bart wrote: > > > I have same problem, any news yet? > > > On 2 apr, 19:38, Alex Payne wrote: > >> It is most likely a Twitter bug. We've been oscillating between two > >> image uploading libraries, trying to deal with issues in each. > > >> On Wed, Apr 1, 2009 at 22:50, pianoben wrote: > > >> > I've Googled around a bit, and haven't found anything that talks about > >> > this issue, so I humbly submit my problem to the Twitter wizards. > > >> > Here's the problem - when I call the API function > >> > update_profile_image, the upload succeeds, and Twitter returns status > >> > code 200. If I go to my actual profile image page (http://twitter.com/ > >> > account/profile_image/pianoben), the updated image (let's say > >> > "O_RLY.jpg") is present. However, the rest of my profile still > >> > retains links to the scaled-down version of the former profile image, > >> > i.e. instead of "http:///o_rly_normal.jpg", everywhere a > >> > profile image is needed still links to "http:/// > >> > old_picture_normal.jpg". The latter is the URL returned as > >> > profile_image_url when calling /users/show/pianoben.xml. > > >> > I'm pretty sure I've got my library code down pat, as the image itself > >> > verifiably uploads - it looks like Twitter is barfing on updating my > >> > user info after the API call completes. It is worth noting that none > >> > of this happens if I upload an image via the web interface. > > >> > Has anyone else encountered this? What did you do to solve this? Or > >> > is this a Twitter bug? > > >> > Thanks, > >> > Ben > > >> -- > >> Alex Payne - API Lead, Twitter, Inc.http://twitter.com/al3x > > -- > Alex Payne - API Lead, Twitter, Inc.http://twitter.com/al3x- Hide quoted text > - > > - Show quoted text -
[twitter-dev] Re: Automating twitter posts
http://twitterfeed.com/ On Tue, Apr 14, 2009 at 09:50, dondmcg wrote: > > I work for a newspaper and we have more and more limited resources and > four flags newspapers. I am wondering if there is any way o automate > twitter posts off an RSS feed or newsletter? In other words if I set > up four twitter accounts to the four papers is there a way I can > populate htose accounts with posts without adding to my workload? > > Sorry if this has already been asked I searched but did not find > anything. > -- Abraham Williams | http://the.hackerconundrum.com Hacker | http://abrah.am | http://twitter.com/abraham Web608 | Community Evangelist | http://web608.org This email is: [ ] blogable [x] ask first [ ] private.
[twitter-dev] Automating twitter posts
I work for a newspaper and we have more and more limited resources and four flags newspapers. I am wondering if there is any way o automate twitter posts off an RSS feed or newsletter? In other words if I set up four twitter accounts to the four papers is there a way I can populate htose accounts with posts without adding to my workload? Sorry if this has already been asked I searched but did not find anything.
[twitter-dev] Re: autenticating user using OAuth
On Apr 14, 12:41 am, Doug Williams wrote: > Where are you storing the secret upon authorization? Is it going into > persistent storage (such as a database) for future use? > Yes, now storing the secret in persistent db so and using internal authentication so user does not have to go to the accept/deny page everytime. When the user logs in, I use the secret from db to make calls to twitter. Thanks, Hameedullah Khan.
[twitter-dev] Re: autenticating user using OAuth
On Apr 14, 12:37 am, Matt Sanford wrote: > Hi Hameedullah, > > It is currently the case that you will get the Accept/Deny page > every time. We're working on a redirect like Zachary mentioned and > hope to have it out by the end of the week. > Thats really nice to know. Thank you guys. Thanks, Hameedullah Khan.
[twitter-dev] Re: Is there a way to tell if a tweet has been favorited (regardless of user)?
2009/4/13 Damon Clinkscales > http://twitfave.com/ (and I think they recently said they are offering an > API) > I built and run twitfave - thanks for mentioning it ! We do actually offer an knocked-together API, see http://twitfave.com/home/about/api - I'm very keen to improve that though, as I'm basically serialising the django objects that I'd be sending to views and throwing them straight to the web (hey, this is a one person project in my spare time ;) ). Feedback is very much appreciated. Note especially the section about "Timeliness" - obviously there's only so much information we can crawl from Twitter, and so we prioritise 1) people who score highly in the apps scoring index and 2) followers of people who follow the twitfave user. (As a side note, I'd still love a firehose of all favouriting activity, but I've mentioned that to Doug and Alex, and it's understandably not high on their priority list currently). Mark
[twitter-dev] Re: Request for documentation review
Hi Doug, I had a look at the SEARCH section of the documentation (not the trends) since the application I am developing primarily uses the search API. First I do like the extra use of examples and I think this will help the new developers out. I have a few observations. 1 - It says that the maximum number of ORs allowed in search is 5. Currently my app uses 13 ORs and it seems to be working fine. Is this an error in the documentation, or is it a planned change to the API? I hope that this is an error, since I would then need to make more calls to the API to achieve the same result. 2 - I believe one thing that should be in the documentation is the fact that the from_user_id is different than the actual twitter user_id. This is fairly important information that I only discovered when browsing through the support tickets. 3 - The new document does not mention that we have access to 4 months of data from search like the current doc does. Is this still the case? 4 - In the current document the Optional URL parameters are very clear. However in the new document they are under the heading Limitations which for someone looking at it for the first time, will surely be confused. 5 - The optional URL parameters under limitations seem to be inaccurate. Isn't SINCE deprecated? and there is no mention of since_id, rpp, page etc. Thanks On Apr 13, 4:11 pm, Doug Williams wrote: > Coderz, > I am refactoring the API documentation [1] to make it friendlier on the eyes > and easier on new developers. Please give it a once over and reply with any > of errors, addition requests, or suggestions. We will be using these new > docs officially in a few days (linked from the apiwiki.twitter.com) once any > glaring omissions and/or errors are addressed. > > My goal is to make the documentation more robust and example laden to help > new developers get going without needing help. Suggestions and critism along > those lines is welcome. > > A final plea for help: > If you have some spare time and examples to borrow from, please send > a...@twitter.com any code/instructions that I can then paste into method body > to help new devs get started. Basic Hello World! style usage examples for > each method would go a long way to curbing new developer problems. Be sure > to include a link to any libraries your example is using. Working full > length example code hosted on github would be preferred. > > 1.http://apiwiki.twitter.com/Twitter-API-Documentation > > Thanks, > Doug Williams > Twitter API Supporthttp://twitter.com/dougw