Update comment about set_join_pathlist_hook().
authorEtsuro Fujita <efujita@postgresql.org>
Thu, 21 Sep 2023 10:45:00 +0000 (19:45 +0900)
committerEtsuro Fujita <efujita@postgresql.org>
Thu, 21 Sep 2023 10:45:00 +0000 (19:45 +0900)
The comment introduced by commit e7cb7ee14 was a bit too terse, which
could lead to extensions doing different things within the hook function
than we intend to allow.  Extend the comment to explain what they can do
within the hook function.

Back-patch to all supported branches.

In passing, I rephrased a nearby comment that I recently added to the
back branches.

Reviewed by David Rowley and Andrei Lepikhov.

Discussion: https://postgr.es/m/CAPmGK15SBPA1nr3Aqsdm%2BYyS-ay0Ayo2BRYQ8_A2To9eLqwopQ%40mail.gmail.com

src/backend/optimizer/path/joinpath.c

index 821d282497bc4396de41f36c38e5491bc1089707..b2916ad6901ab92c7e27026d292dbe52b5704270 100644 (file)
@@ -333,7 +333,10 @@ add_paths_to_joinrel(PlannerInfo *root,
                                                                                                 jointype, &extra);
 
        /*
-        * 6. Finally, give extensions a chance to manipulate the path list.
+        * 6. Finally, give extensions a chance to manipulate the path list.  They
+        * could add new paths (such as CustomPaths) by calling add_path(), or
+        * add_partial_path() if parallel aware.  They could also delete or modify
+        * paths added by the core code.
         */
        if (set_join_pathlist_hook)
                set_join_pathlist_hook(root, joinrel, outerrel, innerrel,