RDF4J crash when loading ntriples file

[abrokenjester]

(Migrated from https://openrdf.atlassian.net/browse/SES-2256 )

I tried to load a file with triples, like this http://es.dbpedia.org/data/Madrid.ntriples

When loading through the workbench I got this error:
ClientHTTPException: MALFORMED DATA: Element type "http:" must be followed by either attribute specifications, ">" or "/>". [line 1, column 8]"

And when loading through the console, this happened:

mvp> load /media/Madrid.ntriples
Loading data...
Closing repository 'mvp'...
Disconnecting from http://localhost:8080/openrdf-sesame
Exception in thread "main" java.lang.NullPointerException
at org.openrdf.http.client.RDF4JSession.upload(RDF4JSession.java:681)
at org.openrdf.http.client.RDF4JSession.upload(RDF4JSession.java:672)
at org.openrdf.repository.http.HTTPRepositoryConnection.add(HTTPRepositoryConnection.java:469)
at org.openrdf.repository.http.HTTPRepositoryConnection.add(HTTPRepositoryConnection.java:393)
at org.openrdf.console.Load.addData(Load.java:155)
at org.openrdf.console.Load.load(Load.java:98)
at org.openrdf.console.Load.execute(Load.java:79)
at org.openrdf.console.Console.executeCommand(Console.java:305)
at org.openrdf.console.Console.start(Console.java:277)
at org.openrdf.console.Console.main(Console.java:135)

Renaming the file from Madrid.ntriples to Madrid.nt all worked.

Here a possible clue by Dale Visser:
I'm guessing that RIO failed to guess the correct serialization format, and somehow ended up crashing with a very uninformative NullPointerException. I don't know if the responsibility for the problem lies on the HttpClient side (e.g., wrong content type header) or the server side, or if they somehow "collaborated"

[barthanssens]

I tested this in workbench 3.0 and while the file was now uploaded without issues in the workbench, viewing the uploaded triples doesn't appear to be successful (i.e. trying to view results 100-150 yields in an empty page)