Could there be an issue with tokenizing either the predicate or the
searched text? The tweet is just broken by spaces and common


On Oct 12, 8:32 am, Fabien Penso <> wrote:
> On Mon, Oct 12, 2009 at 5:28 PM, John Kalucki <> wrote:
> > Unicode is untested on the Streaming API, but UTF-8 really should be
> > supported. I've filed an internal ticket to dig into this more, but I
> > can't promise a resolution.
> Yes I had Korean users with no issues.
> > In the mean time, I'd try putting the track parameter in the POST
> > header (e.g. use curl -d @file as recommended in the Wiki) on the off
> > chance that the webserver is mangling the unicode in the URL. Please
> > report back to the group on the success or failure of this approach.
> I already use POST, the curl was just to show a easy to reproduce test.

Reply via email to