>>>>> "Huong" == Huong Dangminh <huo-dangminh@ys.jp.nec.com> writes:
Huong> Hi,
Huong> I have found a case which could get segmentation fault when
Huong> using GROUPING SETS. It occurs when columns in GROUPING SETS are
Huong> all unsortable but hashable.
Yes, mea culpa. will fix.
BTW, you should have reported this as a bug via the pgsql-bugs list or
the bug submission form.
Huong> Attached grouping_sets_segv.zip include module to reproduce this
Huong> problem.
A much simpler way to reproduce it is to use one of the builtin types
that isn't sortable (I tend to use xid for testing).
Huong> Not yet fully understand the related commit, but I think it is
Huong> fine to put ERRCODE_FEATURE_NOT_SUPPORTED error from
Huong> preprocess_grouping_sets when all columns in GROUPING SETS are
Huong> unsortable. Is that right?
No, that's definitely wrong. The intent is to be able to generate a
hashed path in this case, it's just the logic that tries to prefer
sorting to hashing when the input arrives already sorted is doing the
wrong thing for unsortable data.
--
Andrew (irc:RhodiumToad)