Hey everyone, Thanks for all the information you provided. We were able to reproduce the problem and the engineers have now released a fix. You shouldn't receive any garbled API responses anymore, if you do please let us know.
Best, @themattharris On Jun 14, 7:59 pm, michael hazzard <miramardes...@gmail.com> wrote: > Marc, > Nice theory but I'm not so sure since I'm reloading the same exact tweet (I > think)over and over again and its getting corrupted. > Although I've heard of type of thing before where an application (notepad) > tries to do charset encoding detection but fails and outputs a corrupted > file upon reopening. > On Jun 14, 2011 9:24 PM, "marc fawzi" <marc.fa...@gmail.com> wrote: > > > > > > > > > > > Re: error handling the jsonp "script error" > > > I started using the previously mentioned "window.onerror" mechanism > > wherein I reload the iframe from which I'm initiating the jsonp call, > > up to 10 times, 3 seconds apart (the average delay in getting the > > script back from twitter, which triggers the error which in turn > > starts another jsonp request and so on, up to 10 times) > > > I'm using this approach in addition to my pre-existing "5 retries on > > general error, 10 seconds apart" non-blocking error handler which was > > able to survive the "illegal character" error on its own and recover > > from it, but sometimes it ended up with "Sorry, try again > > later" (twitter recovers after a while on its own and my app continues > > from there but sometimes the error persists with twitter.) The pre- > > existing handler was using more time on average (about 20-50 seconds) > > and the newly added window.onerror mechanism, which runs before the > > self-terminating, non-blocking handler loop, is taking less time on > > average (6-15 seconds) to go thru the same process for the jsonp > > "script error" reported by the browser. However, sometimes, this error > > persists for too many retries and falls out of the window.onerror but > > 99% of the time so far it's getting caught by my pre-existing general > > error handler (now, the second layer of defense) > > > I never thought I had to double up! :) > > > Re: illegal character > > > Wow. Really? you can have the unicode character for "delete" \u0008 in > > the response? I have a full dump of the jsonp response in firebug and > > that's the character that firebug terminates at when reporting the > > error so I assume that's the culprit, but could be others as well. On > > a more relevant note, I don't think the illegal character coming back > > in jsonp have anything to do with the tweet's content. It looks as if > > some tweets contain some combination of characters that cause twitter > > to return a binary dump of some sort (in casual terms, it causes it to > > hallucinate.) > > > Anyway.... intriguing. :) > > > Marc > > On Jun 14, 6:53 pm, Romica <iordan.rom...@gmail.com> wrote: > > > I have the same problem (bad JSON most of the times). The weird thing > > > is that if I access the URL direct from the browser it works (good > > > JSON every time). I'm using jQuery.getJSON and I think that this (or > > > something connected to this) might be the problem. > > > > Unfortunately I cannot use window.onerror (as Michael suggested) > > > because I have multiple asynchronous requests at the same time (it > > > would be tricky and expensive to find out which one is bad - at least > > > I suppose so). > > > > Good luck in solving this > > > -- > > Twitter developer documentation and resources:https://dev.twitter.com/doc > > API updates via Twitter:https://twitter.com/twitterapi > > Issues/Enhancements Tracker: > > https://code.google.com/p/twitter-api/issues/list> Change your membership to > this group: > > https://groups.google.com/forum/#!forum/twitter-development-talk -- Twitter developer documentation and resources: https://dev.twitter.com/doc API updates via Twitter: https://twitter.com/twitterapi Issues/Enhancements Tracker: https://code.google.com/p/twitter-api/issues/list Change your membership to this group: https://groups.google.com/forum/#!forum/twitter-development-talk
