Re: range_agg - Mailing list pgsql-hackers

On Fri, Mar 13, 2020 at 2:39 PM Alvaro Herrera <alvherre@2ndquadrant.com> wrote:
> Here's the rebased version.
>
> I just realized I didn't include the API change I proposed in
> https://postgr.es/m/20200306200343.GA625@alvherre.pgsql ...

Thanks for your help with this Alvaro!

I was just adding your changes to my own branch and I noticed your
v12-0001 has different parameter names here:

diff --git a/src/backend/utils/adt/multirangetypes.c
b/src/backend/utils/adt/multirangetypes.c
index f9dd0378cc..0c9afd5448 100644
--- a/src/backend/utils/adt/multirangetypes.c
+++ b/src/backend/utils/adt/multirangetypes.c
@@ -376,11 +375,11 @@ multirange_typanalyze(PG_FUNCTION_ARGS)
  * pointer to a type cache entry.
  */
 static MultirangeIOData *
-get_multirange_io_data(FunctionCallInfo fcinfo, Oid mltrngtypid,
IOFuncSelector func)
+get_multirange_io_data(FunctionCallInfo fcinfo, Oid rngtypid,
IOFuncSelector func)
 {
     MultirangeIOData *cache = (MultirangeIOData *) fcinfo->flinfo->fn_extra;

-    if (cache == NULL || cache->typcache->type_id != mltrngtypid)
+    if (cache == NULL || cache->typcache->type_id != rngtypid)
     {
         int16       typlen;
         bool        typbyval;
@@ -389,9 +388,9 @@ get_multirange_io_data(FunctionCallInfo fcinfo,
Oid mltrngtypid, IOFuncSelector

         cache = (MultirangeIOData *)
MemoryContextAlloc(fcinfo->flinfo->fn_mcxt,

sizeof(MultirangeIOData));
-        cache->typcache = lookup_type_cache(mltrngtypid,
TYPECACHE_MULTIRANGE_INFO);
+        cache->typcache = lookup_type_cache(rngtypid,
TYPECACHE_MULTIRANGE_INFO);
         if (cache->typcache->rngtype == NULL)
-            elog(ERROR, "type %u is not a multirange type", mltrngtypid);
+            elog(ERROR, "type %u is not a multirange type", rngtypid);

         /* get_type_io_data does more than we need, but is convenient */
         get_type_io_data(cache->typcache->rngtype->type_id,

I'm pretty sure mltrngtypid is the correct name here. Right? Let me
know if I'm missing something. :-)

Yours,
Paul