On 01/13/2013 02:26 PM, Ulf Zibis wrote:
Am 13.01.2013 13:33, schrieb Remi Forax:
Additionally I'm wondering, whether getNextAttribute() could be
generified.
Yes it probably could however I do no want to do this as part of
this change
I disagree, it should not be generified, getNextAttribute() should
return Object and be casted at callsite,
it will more clear if a ClassCastException occurs.
Relying on inference to ask the compiler to insert the cast is dubious,
I must admit, I do not understand.
What would be wrong with that? :
xxx private <T> T readNext() throws SQLException {
xxx T attrib = (T)getNextAttribute();
xxx lastValueWasNull = attrib == null;
xxx return attrib;
xxx }
811 public String readNString() throws SQLException {
812 return readNext();
813 }
... so why not generifying getNextAttribute() directly?
because it's unclear in readNString() that you can have a
ClassCastException,
there is no cast in readNString.
-Ulf
Rémi
