Re: Bug in 9.0Alpha4

Hi,
   I noticed a problem with the source code of 9.0Alpha 4. In parse_agg.c, there is a call made to transformSortClause.

00098     torder = transformSortClause(pstate,


00099                                  agg->aggorder,
00100                                  &tlist,


00101                                  true /* fix unknowns */ ,
00102                                  true /* force SQL99 rules */ );


00103 

   Here agg->aggorder should be a List of SortGroupClause pointers, whereas transformSortClause expects the second argument as a list of SortBy pointers. I verified the doxygen code by downloading the 9.0alpha4 version. I am trying to understand this piece of code, while i thought i should report this bug.

Thanks,
Gokul.

transformSortClause is passed the untransformed aggorder list, which is
in fact a list of SortBy nodes, and it returns the transformed list
(SortGroupClause nodes), which is stored back into the aggorder field
a bit further down.

There are a number of regression tests that would fail in obvious ways
if this code didn't work.

When we were doing the ordered-aggregates patch, I considered passing
all those values as explicit parameters to transformAggregateCall,
and having it build the Aggref node from scratch and return it.
However having seven or eight parameters to transformAggregateCall
(and more in future if we ever add more features here) didn't really
seem to be better style than abusing Aggref a bit.  But maybe it is
the best way after all.  Thoughts?