Users reported that their router was "hosed" when pushing their client
application based on the library into production because the library was
creating a new socket connection for every request. It turns the problem was
that they were not properly closing the connection for every request. It
appears to be worse with NetHttpTransport than ApacheHttpTransport, but either
way it is a problem. The fix is to be extra careful to always close the
connection for both success and error responses, either by calling
response.parseAs*(...) or response.ignore().
Since the library is working as intended, we should improve the documentation.
Perhaps this can go into a new trouble-shooting section.
Original issue reported on code.google.com by yan...@google.com on 25 Jul 2011 at 11:07
Original issue reported on code.google.com by
yan...@google.com
on 25 Jul 2011 at 11:07