[BUG] InputStream.readNBytes doesn't correctly check for EOF

[Rob Spoor]

On 20/12/2020 19:15, Philippe Marschall wrote:
> 
> 
> On 20.12.20 18:47, Rob Spoor wrote:
>> ...
>>
>> That "> 0" is incorrect here; it's allowed to return 0 before EOF
> 
> I don't think the method is allowed to return 0 before EOF. To quote
> from the method Javadoc.
> 
>  > This method blocks until input data is available, end of file is
> detected, or an exception is thrown.
> 
>  > If no byte is available because the stream is at end of file, the
> value {@code -1} is returned; otherwise, at least one byte is read and
> stored into {@code b}.
> 
> Cheers
> Philippe

Hi Philippe,

You are right about the readNBytes method itself. I was talking about 
the read(byte[], int, int) method that is used internally by readNBytes. 
That method may return 0 before EOF, and if it does then readNBytes 
returns before it should.