On Tue, Mar 24, 2009 at 10:11 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Sigh ... I could've sworn I tested that code path, but evidently not,
> 'cause it's broken as can be.
If it's a code path not exercised by any regression test, is it worth
it to add one or we don't have any chance to break this code again
later?
--
Guillaume