I repeat again: you described a situation is possible only in the case where the start is made with the normal download links (turbobit.net), then download suspended implemented through proxy access to the country where the experiment is conducted, for example, Germany, and there is an attempt to continue downloading what happens with a message about the absence of a name / file length. If an attempt is made to start downloading in the same Germany (more precisely, with the IP of the range, distributed in Germany), the FRD simply does not recognize the URL, as processed by the program. In the first case, when the file download has already started, the program willy-nilly trying to do something, and because you get an error.
I think the best method to overcome this situation is to add plug-in on URL new.turbobit.net with processing logic similar to the normal URL turbobit.net (more precisely, the two plug-ins - the ordinary and premium). But what can be done with the "hovering" download, I do not imagine.
PS My English is really not perfect, because it is the result of Google Translator. My native language is Russian, English skills are limited to school and university courses after the application was not special, but after perezda to Germany were driven more necessary to communicate German so that even tests MCSD 10 years ago gave up on the German ...