Dave Cramer <pg(at)fastcrypt(dot)com> writes:
> I've committed the patch with a small formatting difference.
> Also it would be nice if a test case came with this patch.
> Dave Cramer
I'll do that soon, maybe today.
>>> if (length > Integer.MAX_VALUE)
Here I'm a bit concerned the test should be
if ( length > Integer.MAX_VALUE || length < 0 )
since "small enough" negative long values will become positive integers
during the cast.
The question is "how much do we hold the user's hands?" Consider an
infinite stream such as /dev/random with a length of
(long)Integer.MIN_VALUE - 1L.
Here I'm not at all concerned with negative values passed to setBytea()
- handling that is setBytea()'s problem.
Of course we can also test for length < Integer.MIN_VALUE but does it
ever make sense to upload negative length streams?
I can throw in a patch for that when I deliver a unit test, if wanted.
Johann Oskarsson http://www.2ndquadrant.com/ |
PostgreSQL Development, 24x7 Support, Training and Services --+--
In response to
pgsql-jdbc by date
|Next:||From: Kris Jurka||Date: 2012-06-04 14:56:02|
|Subject: Re: REL9_2_DEVEL|
|Previous:||From: Dave Cramer||Date: 2012-06-04 13:00:57|