On Wed, Mar 6, 2013 at 6:30 PM, Matt L. Miller <[log in to unmask]> wrote:

> While I was doing some failure-mode testing on some code that uses
> I have noticed that ShapeFileImporter.read(URL, GeomVectorField, Bag,
> Class)
> has a problem with its exception handling.
> The method declares that it throws a FileNotFoundException; however, it
> actually never does so -- the entire IO section is surrounded in a try ...

Actually, it does throw a FileNotFoundException:

    250         try
    251         {
    252             FileInputStream shpFileInputStream = new
    254             if (shpFileInputStream == null)
    255             {
    256                 throw new FileNotFoundException(shpFile.getFile());
    257             }

> catch that catches all IOExceptions and prints an error message, but does
> NOT propagate the exception back up to the caller. Thus, you get an error
> message on STDERR, but then your code continues blithely along.
> I would recommend either adding a
> throw e;
You're right.  It's bad form to catch exceptions without propagating them.
 I've committed changes whereby the exceptions are rethrown with
appropriate changes to the function signatures.