Re: allow benign typedef redefinitions (C11)

I happened to realize that brin_tuple.h (a somewhat nasty header which
was only supposed to be used by BRIN itself and stuff like amcheck)
recently somehow snuck into tuplesort.h with some nefarious
consequences: that one has polluted execnodes.h, which means that header
now depends on struct definitions that appear in genam.h. We should fix
this. This patch is not the full solution, just a way to show the
problem; for a definitive solution, IMO the definitions of structs
IndexScanInstrumentation and SharedIndexScanInstrumentation should be
elsewhere so that execnodes.h doesn't have to include genam.h.

(gin_tuple.h is also there, but that one's much less of a problem.)

Oh, and replication/conflict.h including all of execnodes.h is kind of
ridiculous. Looks like that's easily solved with something like this
though:

diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index f402b17295c..7bcb4c68e18 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -11,6 +11,7 @@
#ifndef PGSTAT_H
#define PGSTAT_H

+#include "access/transam.h"
#include "datatype/timestamp.h"
#include "portability/instr_time.h"
#include "postmaster/pgarch.h" /* for MAX_XFN_CHARS */
diff --git a/src/include/replication/conflict.h b/src/include/replication/conflict.h
index e516caa5c73..e71b3c8f3d3 100644
--- a/src/include/replication/conflict.h
+++ b/src/include/replication/conflict.h
@@ -9,9 +9,14 @@
#ifndef CONFLICT_H
#define CONFLICT_H

-#include "nodes/execnodes.h"
+#include "access/xlogdefs.h"
+#include "nodes/pg_list.h"
#include "utils/timestamp.h"

+typedef struct EState EState;
+typedef struct ResultRelInfo ResultRelInfo;
+typedef struct TupleTableSlot TupleTableSlot;
+
/*
* Conflict types that could occur while applying remote changes.
*

--
Álvaro Herrera PostgreSQL Developer — https://www.EnterpriseDB.com/
"La espina, desde que nace, ya pincha" (Proverbio africano)