Re: About adding a new filed to a struct in primnodes.h - Mailing list pgsql-hackers

What I mean here is something like below. 

diff --git a/src/backend/nodes/read.c b/src/backend/nodes/read.c
index 8c1e39044c..c3eba00639 100644
--- a/src/backend/nodes/read.c
+++ b/src/backend/nodes/read.c
@@ -29,6 +29,7 @@

 /* Static state for pg_strtok */
 static const char *pg_strtok_ptr = NULL;
+static const char *pg_strtok_extend(int *length, bool testonly);

 /* State flag that determines how readfuncs.c should treat location fields */
 #ifdef WRITE_READ_PARSE_PLAN_TREES
@@ -102,6 +103,20 @@ stringToNodeWithLocations(const char *str)
 #endif


+const char*
+pg_strtok(int *length)
+{
+ return pg_strtok_extend(length, false);
+}
+
+/*
+ * Just peak the next filed name without changing the global state.
+ */
+const char*
+pg_peak_next_field(int *length)
+{
+ return pg_strtok_extend(length, true);
+}
 /*****************************************************************************
  *
  * the lisp token parser
@@ -149,7 +164,7 @@ stringToNodeWithLocations(const char *str)
  * as a single token.
  */
 const char *
-pg_strtok(int *length)
+pg_strtok_extend(int *length,  bool testonly)
 {
  const char *local_str; /* working pointer to string */
  const char *ret_str; /* start of token to return */
@@ -162,7 +177,8 @@ pg_strtok(int *length)
  if (*local_str == '\0')
  {
  *length = 0;
- pg_strtok_ptr = local_str;
+ if (!testonly)
+ pg_strtok_ptr = local_str;
  return NULL; /* no more tokens */
  }

@@ -199,7 +215,8 @@ pg_strtok(int *length)
  if (*length == 2 && ret_str[0] == '<' && ret_str[1] == '>')
  *length = 0;

- pg_strtok_ptr = local_str;
+ if (!testonly)
+ pg_strtok_ptr = local_str;

  return ret_str;
 }

-- the below is a demo code to use it. 
diff --git a/src/backend/nodes/readfuncs.c b/src/backend/nodes/readfuncs.c
index 76ab5ae8b7..c19cd45793 100644
--- a/src/backend/nodes/readfuncs.c
+++ b/src/backend/nodes/readfuncs.c
@@ -689,13 +689,27 @@ _readFuncExpr(void)

  READ_OID_FIELD(funcid);
  READ_OID_FIELD(funcresulttype);
- READ_BOOL_FIELD(funcretset);
+ token = pg_peak_next_field(&length);
+ if (memcmp(token, ":funcretset", strlen(":funcretset")) == 0)
+ {
+ READ_BOOL_FIELD(funcretset);
+ }
+ else
+ local_node->funcretset = false;
+
  READ_BOOL_FIELD(funcvariadic);
  READ_ENUM_FIELD(funcformat, CoercionForm);
- READ_OID_FIELD(funccollid);
+   READ_OID_FIELD(funccollid);
  READ_OID_FIELD(inputcollid);
  READ_NODE_FIELD(args);
- READ_LOCATION_FIELD(location);
+
+    token = pg_peak_next_field(&length);
+ if (memcmp(token, ":location", strlen(":location")) == 0)
+ {
+ READ_LOCATION_FIELD(location);
+ }
+ else
+ local_node->location = -1;

  READ_DONE();
 }

After writing it, I am feeling that this will waste a bit of performance

since we need to token a part of the string twice.  But looks the overhead

looks good to me and can be avoided if we refactor the code again with 

"read_fieldname_or_nomove(char *fieldname,  int *length) " function.