That's what I was expecting to see. However, I have a user who's update made it to his timeline twice. I see that we sent the request twice, 5 seconds apart, because the first one didn't complete. The second request returned successful.
The user's timeline is protected, but the messages are id 1440033342 and 1440033271. I log the ids of successful posts and, in my logs, I see the higher id (1440033342). --Eric On Apr 2, 2009, at 6:03 PM, Doug Williams wrote: > If your application tries to update the status of the same account > within a short period of time, Twitter will ignore the update. As > the statuses/update method returns the status object, in the case > where the message was ignored, the previously successful update > (with the same) text will be returned. > > You can confirm this behavior yourself. Try to update an account's > status with two requests back to back containing the same text: > > $ curl -u user:password -d "status=test" > http://twitter.com/statuses/update.xml > > You will see that the first update is successful. The second request > will then return the same status as the first update (verify by id). > > Doug Williams > Twitter API Support > http://twitter.com/dougw > > > On Thu, Apr 2, 2009 at 11:58 AM, Eric Blair <eric.s.bl...@gmail.com> > wrote: > > Just got a report from one of my users that a message he posted > through our app made it through to his Twitter timeline twice. Looking > at our server logs, I can see that when he posted, we got a timeout > from Twitter and successfully tried to repost. My guess is that the > timed-out post actually went through, as did our report. > > We don't want to be hitting Twitter with duplicate posts, which is why > we're careful about when we retry. However, I've seen references to > Twitter filtering out duplicates, so I was under the impression that > Twitter would detect and reject the repost message in this case. [1] > > [1]: > http://groups.google.com/group/twitter-development-talk/browse_thread/thread/fdaf7454be8f9006/acc53333323f664a?lnk=gst&q=duplicate#acc53333323f664a > > Am I understanding this correctly or should I be more concerned about > duplicate posts making it through my retry code? > > --Eric >